Non-Nutritive Sweetened Beverages and Cardiometabolic Risk
by
Néma McGlynn
A thesis submitted in conformity with the requirements for the degree Master of Science
Department of Nutritional Sciences University of Toronto
© Copyright by Néma McGlynn 2020
ii
Non-Nutritive Sweetened Beverages and Cardiometabolic Risk
Néma McGlynn
Master of Science
Department of Nutritional Sciences
University of Toronto
2020
Abstract
Whether non-nutritive sweetened beverages (NSBs) improve cardiometabolic risk factors similar
to water in their intended substitution for sugar sweetened beverages (SSBs) is unclear. To assess
the effect of 3 prespecified substitutions (NSBs for SSBs, NSBs for water and water for SSBs)
on body weight and cardiometabolic risk factors, we undertook a systematic review and network
meta-analysis. Fourteen trials (n=1530) predominantly in people with overweight/obesity at risk
for or with diabetes were identified. The substitution of NSBs for SSBs improved body weight,
BMI, body fat, triglycerides and intrahepatocellular lipids supporting the use of NSBs, as a
viable alternative to water, for SSBs. Due to imprecision in the estimate and concerns that NSBs
adversely affect glucose tolerance through compositional gut microbiome changes, we provided
a rationale, design and baseline characteristics report for the STOP Sugars NOW trial to
determine the effect of our 3 prespecified substitutions on gut microbiome and glucose tolerance.
Words: 150
iii
Acknowledgments
I would sincerely like to thank the following individuals whose support and guidance were
pivotal in the completion of this research and attainment of my master’s degree!
First of all, to my sister Natalie, for putting up with losing her spot at the dining room table to
two computer screens, and gently reminding me when my random notes started taking over the
living room. I know it was driving you crazy little cat so thank you for biting your tongue! To
my mother and father for supporting my decision to leave full time employment to pursue
graduate studies. Your Costco deliveries of toilet paper, paper towel and nuts were a lifesaver!
To the Toronto 3D lab crew, including the MSBers, I want to thank each of you for answering all
my questions, even if they were ridiculous. I always looked forward to coming into the lab and
working with you all. Andrea Glenn, Andreea Zarbau, Rodney (little big bro) Au-Yeung,
Meaghan Kavanagh, Jarvis Noronha, Catherine Braunstein, Dr. Laura Chiavaroli, Sonia Blanco
Mejia, Maxine Seider, Sabrina Ayoub-Charette, Stephanie Nishi, Effie Viguiliouk, Danielle Lee,
Annette Cheung, Amna Ahmed, Darshna Patel, Sandhya Sahya-Pudaruth, Melanie Paquet,
Bashyam Balachandran and Stefan Kabisch you all made this journey so much fun!
Dr. David Jenkins, our conversations during my weekends in the lab were a HUGE pick-me-up. I
always loved hearing your thoughts on the many topics we discussed. You were my favourite
part of the day.
A BIG thank you to my committee members Dr. Elena Comelli and Dr. Tom Wolever. Your
advice and direction both in between and during my committee meetings really helped me stay
focused and on track. I would also like to acknowledge Dr. Anthony Hanley, first for guiding me
through your nutritional epidemiology course, it was a rough ride but truly rewarding, and for
offering your time to be my examiner. You have a very positive reputation among the students in
the Nutritional Sciences department and now I know why!
A special acknowledgment to Louisa Kung, our graduate administrator, who went above and
beyond to help me on several different occasions. You are the backbone of the nutritional
sciences department and someone I could always count on.
iv
Dr. Tauseef Ahmad Khan, words cannot express the gratitude I feel towards your endless support
and guidance in every aspect of my research projects. I leaned on you a lot and will be forever
grateful for your patience and understanding during my fumbles through this process. You are a
wonderful educator who really knows how to get the most out of the “copy paste” method.
Finally, a very special thank you to the one and only Dr. John Sievenpiper. Your endless drive
and determination have given so many students an opportunity to learn from the best! Thank you
for pushing me out of my comfort zone. Your unwavering dedication to upholding scientific
rigor in search of the “truth” despite ALL the obstacles that have been thrown your way is
incredibly inspirational. I don’t know how you do it but I’m incredibly proud to say that you
were my supervisor. YOU are the trailblazer.
v
Table of Contents Acknowledgments.......................................................................................................................... iii
Table of Contents ............................................................................................................................ v
List of Tables ................................................................................................................................. ix
List of Figures ................................................................................................................................ xi
List of Abbreviations .................................................................................................................... xv
Chapter 1: Introduction ................................................................................................................... 1
Chapter 2: Literature Review .......................................................................................................... 3
2.1 Obesity and Diabetes ................................................................................................................ 3
2.1.2 Overview of Obesity and Diabetes..................................................................................... 3
2.1.3 Prevalence and Economic Burden...................................................................................... 3
2.1.4 Risk Factors and Complications ......................................................................................... 3
2.1.5 Measures of Overweight and Obesity ................................................................................ 5
2.1.6 Diagnosis and Targets of Control for Type 2 Diabetes ...................................................... 6
2.1.7 Prevention and Management .............................................................................................. 7
2.2 Sugar-Sweetened Beverages ..................................................................................................... 7
2.2.1 Worldwide Intakes of Sugar-Sweetened Beverages .......................................................... 9
2.2.2 Guidelines for Sugar-Sweetened Beverage Consumption ............................................... 10
2.2.3 Replacements for Sugar-Sweetened Beverages in Guidelines ......................................... 10
2.3 Non-Nutritive Sweeteners ....................................................................................................... 13
2.3.1 Definition of Non-Nutritive Sweeteners .......................................................................... 13
2.3.2 Safety and Approval Process............................................................................................ 13
2.3.3 Sweetening Intensity ........................................................................................................ 16
2.3.4 Non-Nutritive Sweetener Consumption Trends ............................................................... 16
2.3.5 Absorption, Digestion, Metabolism and Excretion of Non-Nutritive Sweeteners ........... 16
2.3.5.1 Aspartame .................................................................................................................. 16
2.3.5.2 Sucralose .................................................................................................................... 17
2.3.5.3 Acesulfame Potassium and Saccharin ....................................................................... 17
2.3.6 Potential Mechanisms of Non-Nutritive Sweeteners on Body Weight and Glucose
Control ....................................................................................................................................... 19
2.3.6.1 Activation of Sweet Taste Receptors ......................................................................... 19
2.3.6.2 Changes in Taste Preferences and Dietary Intake ..................................................... 19
2.3.6.3 Gut Microbiome ........................................................................................................ 20
vi
2.3.7 Systematic Reviews and Meta-Analyses of Prospective Cohort Studies Investigating the
Association between Non-Nutritive Sweeteners and Obesity/ Type 2 Diabetes Mellitus Risk 21
2.3.8 Systematic Reviews and Meta-Analyses of Randomized Controlled Trials Investigating
the Effect of Non-Nutritive Sweeteners on Obesity/ Type 2 Diabetes Mellitus Risk............... 22
2.3.9 Network Meta-Analyses ................................................................................................... 26
Chapter 3: Rationale and Objectives............................................................................................. 27
3.1 Rationale.............................................................................................................................. 27
3.2 Objectives ............................................................................................................................ 27
Chapter 4: Effect of Non-Nutritive Sweetened Beverages as a Replacement Strategy for Sugar-
Sweetened Beverages on Body Weight and Cardiometabolic Risk: A Network Meta-Analysis of
Randomized Controlled Trials ...................................................................................................... 29
4.1 Abstract ................................................................................................................................... 30
4.2 Introduction ............................................................................................................................. 31
4.3 Methods................................................................................................................................... 32
4.3.1 Design............................................................................................................................... 32
4.3.2 Data Sources and Searches ............................................................................................... 32
4.3.3 Study Selection ................................................................................................................. 33
4.3.4 Data Extraction ................................................................................................................. 33
4.3.5 Outcomes .......................................................................................................................... 33
4.3.6 Risk of Bias Assessment .................................................................................................. 34
4.3.7 Data Synthesis .................................................................................................................. 34
4.3.8 Grading of the Evidence ................................................................................................... 35
4.3.9 Patient Involvement.......................................................................................................... 36
4.4 Results ..................................................................................................................................... 37
4.4.1 Search Results .................................................................................................................. 37
4.4.2 Available Data .................................................................................................................. 37
4.4.3 Trial Characteristics ......................................................................................................... 37
4.4.4 Risk of Bias ...................................................................................................................... 38
4.4.5 Effect of Substitution of NSBs for SSBs ......................................................................... 38
4.4.6 Effect of Substitution of Water for SSBs ......................................................................... 38
4.4.7 Effect of Substitution of NSBs for Water ........................................................................ 39
4.4.8 Inconsistency .................................................................................................................... 39
4.4.9 Intransitivity (a domain of indirectness, in the indirect estimates) .................................. 39
vii
4.4.10 Sensitivity Analyses ....................................................................................................... 39
4.4.11 Subgroup Analyses ......................................................................................................... 39
4.4.12 Publication Bias.............................................................................................................. 40
4.4.13 Grading of the Evidence ................................................................................................. 40
4.5 Discussion ........................................................................................................................... 40
4.5.1 Findings in the Context of Existing Studies ..................................................................... 41
4.5.2 Strengths and Limitations................................................................................................. 43
4.5.3 Implications ...................................................................................................................... 46
4.5.4 Conclusions ...................................................................................................................... 46
4.5.5 Funding Statement............................................................................................................ 47
4.5.6 Conflict of Interest ........................................................................................................... 48
4.5.7 Ethics Approval ................................................................................................................ 51
4.5.8 Data Sharing ..................................................................................................................... 51
Chapter 5: Rationale, Design and Baseline Characteristics Assessing the Effect of Substituting
NSBs versus Water for SSBs on Gut Microbiome, Glucose Tolerance, and Cardiometabolic Risk
Factors: Strategies To OPpose SUGARS with Non-nutritive sweeteners Or Water trial (STOP
Sugars NOW) ................................................................................................................................ 98
5.1 Abstract ................................................................................................................................... 99
5.2 Introduction ........................................................................................................................... 100
5.3 Objective and Hypotheses..................................................................................................... 101
5.4 Methods................................................................................................................................. 102
5.4.1 Study Design .................................................................................................................. 102
5.4.2 Blinding .......................................................................................................................... 103
5.4.3 Participants ..................................................................................................................... 103
5.4.4 Sample Size (Power) Calculation ................................................................................... 104
5.4.5 Recruitment, Consent and Screening ............................................................................. 105
5.4.6 Run-In Phase .................................................................................................................. 106
5.4.7 Randomization ............................................................................................................... 106
5.4.8 Intervention .................................................................................................................... 106
5.4.9 Study Visits .................................................................................................................... 107
5.4.10 Washout phase.............................................................................................................. 108
5.4.11 Outcomes ...................................................................................................................... 109
5.4.12 Adherence Assessment ................................................................................................. 109
viii
5.4.13 Analytical Techniques .................................................................................................. 109
5.4.13.1 Anthropometric Analyses ...................................................................................... 109
5.4.13.2 Biochemical Analyses ........................................................................................... 110
5.4.13.3 Biomarkers of Adherence ...................................................................................... 111
5.4.14 Antibiotic Use .............................................................................................................. 111
5.4.15 Compensation ............................................................................................................... 112
5.4.16 Statistical Analyses ...................................................................................................... 112
5.4.16.1 Subgroup Analysis ................................................................................................. 112
5.4.17 Adverse Effects ............................................................................................................ 112
5.5 Results ................................................................................................................................... 113
5.5.1 Participant Flow ............................................................................................................. 113
5.5.2 Baseline Characteristics ................................................................................................. 113
5.6 Discussion ............................................................................................................................. 114
5.7 Conclusion ............................................................................................................................ 115
5.8 Acknowledgements and Funding .......................................................................................... 116
5.9 Statement of Contribution ..................................................................................................... 116
Chapter 6: General Discussion.................................................................................................... 125
6.1 Summary ............................................................................................................................... 125
6.2 Strengths and Limitations ..................................................................................................... 126
6.3 Clinical Implications ............................................................................................................. 128
6.4 Future Directions .................................................................................................................. 129
Chapter 7: Conclusions ............................................................................................................... 130
References ................................................................................................................................... 132
ix
List of Tables
Chapter 2
Table 2.1 Harmonized Criterion for the Diagnosis of Metabolic Syndrome[35] ............................. 4
Table 2.2 BMI Classifications[36] ................................................................................................... 5
Table 2.3 Diagnostic Criterion for Type 2 Diabetes Mellitus [12] .................................................. 7
Table 2.4 Non-Nutritive Sweetener Guideline Recommendations .............................................. 12
Table 2.5 Common Non-Nutritive Sweeteners Used in Foods and Beverages in Canada[82, 91, 92]
....................................................................................................................................................... 15
Table 2.6 Systematic Reviews and Meta-Analyses of Randomized Controlled Trials
Investigating the Effect of Non-Nutritive Sweeteners on Adiposity ............................................ 23
Table 2.7 Systematic Reviews and Meta-Analyses of Randomized Controlled Trials
Investigating the Effect of Non-Nutritive Sweeteners on Diabetes Risk ..................................... 25
Chapter 4
Table 4.1: Trial Characteristics .................................................................................................... 52
Appendix Table 4.1: PRISMA-NMAa Checklist ........................................................................ 58
Appendix Table 4.2: Search Strategy.......................................................................................... 62
Appendix Table 4.3: PICOTSb Framework ................................................................................ 63
Appendix Table 4.4: Side-Splitting Approach for Inconsistency for Body Weight ................... 64
Appendix Table 4.5: Side-Splitting Approach for Inconsistency for BMI ................................. 64
Appendix Table 4.6: Side-Splitting Approach for Inconsistency for Body Fat % ..................... 64
Appendix Table 4.7: Side-Splitting Approach for Inconsistency for Waist Circumference ...... 64
Appendix Table 4.8: Side-Splitting Approach for Inconsistency for HbA1c ............................. 65
Appendix Table 4.9: Side-Splitting Approach for Inconsistency for FPG ................................. 65
Appendix Table 4.10: Side-Splitting Approach for Inconsistency for 2h-PG ............................ 65
Appendix Table 4.11: Side-Splitting Approach for Inconsistency for FPI ................................. 65
Appendix Table 4.12: Side-Splitting Approach for Inconsistency for HOMA-IR ..................... 66
Appendix Table 4.13: Side-Splitting Approach for Inconsistency for LDL-C ........................... 66
Appendix Table 4.14: Side-Splitting Approach for Inconsistency for Non-HDL-C .................. 66
Appendix Table 4.15: Side-Splitting Approach for Inconsistency for TGs................................ 66
Appendix Table 4.16: Side-Splitting Approach for Inconsistency for HDL-C .......................... 67
x
Appendix Table 4.17: Side-Splitting Approach for Inconsistency for TC ................................. 67
Appendix Table 4.18: Side-Splitting Approach for Inconsistency for SBP ............................... 67
Appendix Table 4.19: Side-Splitting Approach for Inconsistency for DBP ............................... 67
Appendix Table 4.20: Side-Splitting Approach for Inconsistency for IHCL ............................. 68
Appendix Table 4.21: Side-Splitting Approach for Inconsistency for ALT ............................... 68
Appendix Table 4.22: Side-Splitting Approach for Inconsistency for AST ............................... 68
Appendix Table 4.23: Side-Splitting Approach for Inconsistency for Uric Acid....................... 68
Appendix Table 4.24: Loop-Specific Approach for Inconsistency ............................................ 69
Chapter 5
Table 5.1: Study Exclusion Criteria1.......................................................................................... 118
Table 5.2 Power Calculation for Primary Outcomes Between Water and NSB Arms .............. 119
Table 5.3: Latin Square Randomization for Three Group Crossover Study1 ............................ 120
Table 5.4: Complete List of All Available Study Beverages ..................................................... 120
Table 5.5: Participant Visit Schedule ......................................................................................... 121
Table 5.6 Baseline Characteristics of Randomized Participants1 .............................................. 123
xi
List of Figures
Chapter 2
1Figure 2.1 SSB consumption and the development of obesity, T2DM and Cardiometabolic
Risk[49] ............................................................................................................................................. 9
2Figure 2.2 Absorption, Digestion, Metabolism and Excretion of Acesulfame Potassium,
Saccharin, Aspartame and Sucralose[91] ........................................................................................ 18
3Figure 2.3 Network Meta-Analysis Diagram Combining Direct and Indirect Evidence[141] ....... 26
Chapter 4
Figure 4.1 Literature Search for RCTs of NSBs reporting on Adiposity, Glycemic Control,
Blood Lipids, Blood Pressure, NAFLD and Uric Acid. ............................................................... 54
Figure 4.2 Network Results: Substitution of NSBs for SSBs ...................................................... 55
Figure 4.3: Network Results: Substitution of Water for SSBs .................................................... 56
Figure 4.4 Network Results: Substitution of NSBs for Water ..................................................... 57
Appendix Figure 4.1: Cochrane Risk of Bias Summary for all Included Trials......................... 70
Appendix Figure 4.2: Risk of Bias Proportion for all Included Trials ....................................... 71
Appendix Figure 4.3: Transitivity Analysis_Box Plots Showing the Distribution of the Mean
Age (Years) of the Trials Across the Available Direct Comparisons ........................................... 72
Appendix Figure 4.4: Transitivity Analysis_Box Plots Showing the Distribution of the study
Length (Weeks) of the Trials Across the Available Direct Comparisons .................................... 73
Appendix Figure 4.5: Transitivity Analysis_Box Plots Showing the Distribution of the Sample
Size of the Trials Across the Available Direct Comparisons ....................................................... 74
Appendix Figure 4.6: Transitivity Analysis_Box Plots Showing the Distribution of the % Males
of the Trials Across the Available Direct Comparisons ............................................................... 75
Appendix Figure 4.7: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Adiposity Outcomes................................................................................................................. 76
Appendix Figure 4.8: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Glycemic Outcomes ................................................................................................................. 77
Appendix Figure 4.9: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Blood Lipid Outcomes ............................................................................................................. 78
xii
Appendix Figure 4.10: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Blood pressure, NAFLD and Uric Acid Outcomes ................................................................. 79
Appendix Figure 4.11: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Adiposity Outcomes................................................................................................................. 80
Appendix Figure 4.12: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Glycemic Outcomes ................................................................................................................. 81
Appendix Figure 4.13: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Blood Lipid Outcomes ............................................................................................................. 82
Appendix Figure 4.14: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Blood pressure, NAFLD and Uric Acid Outcomes ................................................................. 83
Appendix Figure 4.15: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
on Adiposity Outcomes................................................................................................................. 84
Appendix Figure 4.16: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
....................................................................................................................................................... 85
Appendix Figure 4.17: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
on Blood Lipid Outcomes ............................................................................................................. 86
Appendix Figure 4.18: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
on Blood pressure, NAFLD and Uric Acid Outcomes ................................................................. 87
Appendix Figure 4.19: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Body
Weight ........................................................................................................................................... 88
Appendix Figure 4.20: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on BMI ..... 88
Appendix Figure 4.21: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Body Fat
% ................................................................................................................................................... 89
Appendix Figure 4.22: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Waist
Circumference ............................................................................................................................... 89
xiii
Appendix Figure 4.23: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on HbA1c . 90
Appendix Figure 4.24: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on FPG ..... 90
Appendix Figure 4.25: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on 2h-PG .. 91
Appendix Figure 4.26: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on FPI ....... 91
Appendix Figure 4.27: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on HOMA-IR
....................................................................................................................................................... 92
Appendix Figure 4.28: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on LDL-C . 92
Appendix Figure 4.29: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Non-HDL-
C .................................................................................................................................................... 93
Appendix Figure 4.30: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on TGs ...... 93
Appendix Figure 4.31: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on HDL-C 94
Appendix l Figure 4.32: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on TC ....... 94
Appendix Figure 4.33: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on SBP ..... 95
Appendix Figure 4.34: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on DBP ..... 95
Appendix Figure 4.35 Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on IHCL ... 96
Appendix Figure 4.36: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on ALT ..... 96
Appendix Figure 4.37: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on AST ..... 97
Appendix Figure 4.38: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on uric acid97
xiv
Chapter 5
Figure 5.1: Study Design1 .......................................................................................................... 117
Figure 5.2: Trial Flow: Screening and Randomization .............................................................. 122
xv
List of Abbreviations
2h-PG = 2h plasma glucose
Ace-K = acesulfame potassium
ADI = Acceptable Daily Intake
ADME = Absorption, Distribution, Metabolism, and Excretion
ALT = alanine transaminase
AST = aspartate transaminase
BMI = body mass index
CHD = coronary heart disease
CI = confidence interval
CIHR = Canadian Institutes of Health Research
CVD = cardiovascular disease
CAC = Codex Alimentarius Commission
CPIR = cephalic phase insulin response
CPR = cephalic phase response
DBP = diastolic blood pressure
DNSG = Diabetes and Nutrition Study Group
EASD = European Association for the Study of Diabetes
FAO = Food and Agriculture Organisation
FPI = fasting plasma insulin
GDM = gestational diabetes
GDP = gross domestic product
GI = gastrointestinal (GI)
GL = glycemic load
GLP-1 = glucagon-like peptide-1
GRADE = Grading of Recommendations Assessment, Development, and Evaluation
HbA1c = hemoglobin A1c
HDL-C = HDL-cholesterol
HOMA-IR = Homeostatic Model Assessment of Insulin Resistance
xvi
iAUC = incremental area under the curve
IDF = International Diabetes Federation
IFG = impaired fasting glucose
IGT = impaired glucose tolerance
IHCL = intra-hepatocellular lipid
JECFA = Joint Expert Committee on Food Additives
LDL-C = LDL-cholesterol
MD = mean difference
MetS = metabolic syndrome
NAFLD = non-alcoholic fatty liver disease
NCEP-ATP III = National Cholesterol and Education Program Adult Treatment Panel III
NNS = non-nutritive sweetener
NOAEL = No Observed Adverse Effect Level
Non-HDL-C = Non-HDL-cholesterol
NSB = non-nutritive sweetened beverage
OGTT = oral glucose tolerance test
PCS = prospective cohort study
RCTs = randomized controlled trials
SBP = systolic blood pressure SBP
SMD = standardized mean difference
SRMA = systematic reviews and meta-analysis
SSB = sugar-sweetened beverage
STOP Sugars NOW = Strategies To OPpose SUGARS with Non-nutritive sweeteners Or Water
STR = sweet taste receptor
T2DM = type 2 diabetes mellitus
TC = total cholesterol
TGs = triglycerides
WHO = World Health Organization
1
Chapter 1: Introduction
Rates of obesity worldwide are increasing and have nearly tripled since 1975. In 2016, over
650 million adults, 18 years of age and older, were living with obesity [1]. These numbers
are concerning as obesity is a well-known risk factor for many debilitating chronic diseases
including type 2 diabetes mellitus (T2DM) [2-4], which has increased in tandem with global
obesity rates [5].
While multiple lifestyle and dietary factors have been implicated in the etiology of obesity
and T2DM, dietary sugars have emerged as the dominant nutrient of concern as their excess
intakes have been associated with weight gain, T2DM and downstream complications
including hypertension and coronary heart disease [6-9]. As a result, the World Health
Organization has called for reductions in free sugars to ≤5-10% of energy [10], with a focus
on sugar-sweetened beverages (SSBs) due to their substantially high free sugar content [5].
This call to action has provided an opportunity for non-nutritive sweetened beverages
(NSBs). Despite the potential benefits of NSBs, recommendations for their consumption are
inconsistent across dietary and clinical practice guidelines [11-18], as concerns exist that they
contribute to an increased risk of the very diseases they were designed to prevent. The
preferred replacement for SSBs in these guidelines is water.
Recent systematic reviews and meta-analyses show an association between NSBs and
increased risk of weight gain, diabetes and cardiometabolic risk [7, 19-23] in prospective
cohort studies, with no effect on these outcomes in the higher quality evidence from
randomized controlled trials (RCTs) [19-21, 24, 25]. The majority of these RCT syntheses,
however, have failed to account for the nature of the comparator, an important determinant in
assessing the effectiveness of NSBs. When non-nutritive sweeteners (NNSs), including
NSBs, displace energy from caloric controls, such as SSBs, the weight of the evidence
indicates benefit for NSBs on several cardiometabolic outcomes [19-21, 24, 26]. As a result,
this thesis will explore the effect of NSBs, as a replacement for SSBs, through 3 prespecified
substitutions: NSBs for SSBs (intended substitution with caloric displacement), water for
2
SSBs (“standard of care” substitution with caloric displacement) and NSBs for water
(matched substitution without caloric displacement) on measures of cardiometabolic risk.
To address this gap in knowledge, this thesis will 1) provide a brief overview of obesity and
diabetes, their relationship to the metabolic syndrome, diagnostic criteria, and management
and prevention; describe how SSBs act as an underlying driver to the development of
cardiometabolic risk factors; SSB consumption trends and recommended intakes; and
summarize the safety, biological pathways and concerns of NNSs on adiposity and glucose
control; 2) report results from a systematic review and network meta-analysis on the effect of
NSBs, as a replacement for SSBs, through 3 prespecified substitutions: NSBs for SSBs,
water for SSBs, and NSBs for water on body weight and cardiometabolic risk; and 3) provide
a rationale, design and baseline characteristics report assessing the effect of substituting
NSBs for SSBs, water for SSBs, and NSBs for water on gut microbiome, glucose tolerance,
and cardiometabolic risk factors: Strategies To OPpose SUGARS with Non-nutritive
sweeteners Or Water trial (STOP Sugars NOW).
3
Chapter 2: Literature Review
2.1 Obesity and Diabetes
2.1.2 Overview of Obesity and Diabetes
Both obesity and diabetes are considered progressive chronic conditions and independent risk
factors for cardiovascular disease [12, 27, 28]. Characterized by abnormal or excessive fat
accumulation, obesity is a major risk factor for type 2 diabetes mellitus (T2DM) [27]. Diabetes is
characterized by the presence of hyperglycemia resulting from impaired and/or defective insulin
secretion and/or action, and represents a group of metabolic disorders that can lead to long-term
microvascular and macrovascular complications [12]. Diabetes can be classified into three main
categories: type 1 diabetes mellitus, T2DM and gestational diabetes (GDM) [29]. Of these types,
the most common is T2DM, representing approximately 90% of people living with diabetes [29].
2.1.3 Prevalence and Economic Burden
The prevalence of obesity and T2DM are increasing. Since 1975 obesity has nearly tripled with
13% of the world’s adult population living with obesity in 2016 at an estimated 650 million [1].
In parallel with the diabetes epidemic is an obesity epidemic and it’s estimated that 80% to 90%
of individuals living with T2DM are overweight or obese [12]. In 2017, an estimated 425 million
adults were living with diabetes and this number is expected to rise to 629 million by 2045 [29].
An aging population coupled with increasing urbanization and economic development are
implicated in the development of these chronic conditions as they are linked with sedentary
lifestyles and poor dietary habits [30]. The economic implications are immense. The global
economic cost of obesity was estimated at USD 2 trillion (2.8% of the gross domestic product
(GDP)) [31], in 2014 and USD 1.3 trillion (1.8% of the GDP) for diabetes in 2017 [32].
2.1.4 Risk Factors and Complications
Risk factors for obesity, although complex, are linked to those that lead to an imbalance of
energy consumed to energy expended, including biological, behavioral, social and environmental
factors [28]. Excess body fat is the number one driver for T2DM, as well as a combination of
genetic and metabolic features, and modifiable lifestyle factors. These include a family history of
the disease, an unhealthy diet, physical inactivity, advancing age, hypertension, cultural origin,
4
impaired glucose tolerance (IGT), smoking, a history of GDM and poor nutrition during
pregnancy [5, 29, 33].
Individuals living with prediabetes, a condition characterized by hyperglycemia below diagnostic
thresholds, are at an increased risk of developing T2DM. Diagnosis of prediabetes is made by the
presence of IGT, impaired fasting glucose (IFG) or a hemoglobin A1c (HbA1c), of 6.0% to
6.4%. IGT occurs when glucose levels are between 7.8-11.0 mmol/L two-hours after an oral
glucose tolerance test (OGTT), whereas IFG is a higher than normal fasting glucose level in the
range of 6.1-6.9 mmol/L [12, 29]. The International Diabetes Federation (IDF) estimates that the
number of people worldwide with IGT was 352.1 million (global prevalence of 7.3%) in 2017
[29].
Included in the risk factors for T2DM is the metabolic syndrome (MetS), a multifaceted and
highly prevalent clustering of abnormalities including central obesity, hypertension, dyslipidemia
and elevated blood glucose [12]. Not only is obesity a significant risk factor in the development
of MetS, but individuals with this syndrome are at significant risk for developing T2DM and
cardiovascular disease (CVD) [34]. A harmonized definition for the diagnosis of the MetS was
established in 2009 and is based on having a minimum of 3 or more criteria (Table 2.1) [35].
1Table 2.1 Harmonized Criterion for the Diagnosis of Metabolic Syndrome[35]
Measure Criteria Cut Points
Men Women
Elevated waist circumference (cm)1
• Canada; USA ≥102 ≥88
• Europids; Middle-Eastern; Sub-Saharan African;
Mediterranean
≥94 ≥80
• Asians; Japanese; South and Central Americans ≥90 ≥80
Elevated TGs (mmol/L) (drug treatment is an alternate indicator2) ≥1.7
Reduced HDL-C (mmol/L) (drug treatment is an alternate
indicator2)
<1.0 <1.3
Elevated BP (mmHg) (antihypertensive drug treatment in an
individual with a history of hypertension is an alternate indicator)
Systolic ≥130 and/or
diastolic ≥85
Elevated FPG (mmol/L) (drug treatment is an alternate indicator3) ≥5.6 BP, blood pressure; FPG, fasting plasma glucose; HDL-C, high-density lipoprotein-cholesterol; T2DM, type 2
diabetes mellitus; TGs, triglycerides 1Population and country specific cut points 2Fibrates and nicotinic acid are commonly used drugs for elevated triglycerides and reduced HDL-C. Individuals
taking one of these drugs are presumed to have high triglycerides and low HDL-C. Omega-3 fatty acids at high
doses indicated high triglycerides
5
3Most individuals with T2DM will have the metabolic syndrome
Aside from T2DM, obesity can lead to the development of several other comorbidities including
coronary artery disease, stroke, osteoarthritis, certain cancers, obstructive sleep apnea and non-
alcoholic fatty liver disease (NAFLD) [36]. T2DM can lead to significant microvascular and
cardiovascular complications. It is one of the leading causes of CVD, blindness, kidney failure,
amputation of the lower-limb and is associated with increased rates of cancer, physical and
cognitive disability, tuberculosis and depression [5, 12]. Eventually, these complications can
result in premature death. The IDF estimated that the number of deaths resulting from diabetes
was 4 million in 2017, accounting for 10.7% of global all-cause mortality among individuals
between 20-79 years [29].
2.1.5 Measures of Overweight and Obesity
The two most common measures for overweight and obesity are through the assessment of body
mass index (BMI) and waist circumference. Taken together, both these measures identify the
level and distribution of adiposity among individuals [36]. As different populations vary in their
level of risk due to differences in fat distributions, the cut-offs for BMI and waist circumference
vary by ethnicity. In general, BMI cut-offs identify individuals at an increased risk for morbidity
and mortality (Table 2.2) [36]. For Asian Americans, lower BMI cut-offs are recommended as
they are at risk for obesity-related diseases below BMI classification values used for non-Asians
[37, 38]. Current population-specific waist circumference cut-offs are the same as those used for
the diagnosis of the MetS as outlined in Table 2.1 [36]. These population-specific waist
circumference values are recommended by the IDF to align the World Health Organization
(WHO) guidelines with the National Cholesterol and Education Program Adult Treatment Panel
III (NCEP-ATP III) guidelines [36, 39-41].
2Table 2.2 BMI Classifications[36]
Classification1 BMI (kg/m²) Disease Risk2
Underweight <18.5 Increased
Normal Weight 18.5-24.9 Least
Overweight 25-29.9 Increased
Obesity ≥30
Class I 30-34.9 High
Class II 35-39.9 Very High
Class II ≥40 Extremely High
6
BMI, body mass index; T2DM, type 2 diabetes mellitus 1For Asian and South Asian individuals BMI classifications are recommended as ≥23 for overweight and ≥25 for
obese [37] 2For T2DM, hypertension and cardiovascular disease relative to normal weight and waist circumference [36]
Even though BMI is often used as the conventional method for assessing adiposity due to its
low-cost and convenience, it cannot distinguish between lean and fat mass or give insight into
the distribution of body fat [42]. Although waist circumference is a better estimate of fat
distribution, particularly visceral fat and its relationship to risk of myocardial infarction (MI) and
CVD, results from INTERHEART’s case-control study of 52 countries by Yusuf et al. [43],
showed that waist-to-hip ratio was a better predictor of both. According to the WHO, waist-to-
hip ratios of ≥0.9cm in men and ≥0.85cm in women substantially increases the risk of metabolic
complications including decreased glucose tolerance, reduced insulin sensitivity and adverse
lipid profiles, all risk factors for T2DM and CVD [44].
Other methods of measuring body fat distribution include skinfold thickness measurements,
bioelectrical impedance, underwater weighing and dual energy x-ray absorptiometry (DXA).
Despite providing more accurate measures of the amount and distribution of body fat, they are
more cumbersome and/or expensive. As a result, BMI, waist circumference and waist-to-hip
ratio are more commonly used [42].
2.1.6 Diagnosis and Targets of Control for Type 2 Diabetes
The three most-common threshold parameters for detecting elevated blood glucose levels are
fasting plasma glucose (FPG), HbA1c, and a 75g 2-hour plasma glucose (2h-PG) OGTT.
According to the 2018 Diabetes Canada guidelines, the diagnostic criteria for T2DM (Table 2.3)
are based on glycemic thresholds associated with microvascular disease, particularly retinopathy,
from venous blood samples and laboratory methods. If one test comes back in the diagnostic
range, the same test should be repeated another day, except in the case of random plasma glucose
where an alternate test should confirm this value among asymptomatic individuals. This
approach results in a confirmatory diagnosis in around 40% to 90% of people presenting with an
initial positive test [12].
7
3Table 2.3 Diagnostic Criterion for Type 2 Diabetes Mellitus [12]
Parameter Diagnostic Criteria for T2DM
FPG (mmol/L) ≥7.0 mmol/L
HbA1c (%) ≥6.5% (adults)
2h-PG in a 75g OGTT (mmol/L) ≥11.1 mmol/L
Random PG (mmol/L) ≥11.1 mmol/L
2h-PG, 2-hour plasma glucose; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; OGTT, oral glucose
tolerance test; PG, plasma glucose; T2DM, type 2 diabetes mellitus.
Achieving optimal glycemic control is the cornerstone of T2DM management and HbA1c is the
standard test for determining glycated hemoglobin. Evidence indicates that improved glycemic
control (A1C ≤7.0%) reduces the risk of both microvascular and CV complications. To achieve
this target, people living with T2DM should aim for a FPG of 4.7-7.0 mmol/L and a 2h-PG level
of 5.0-10.0 mmol/L [12].
2.1.7 Prevention and Management
Several strategies exist in the prevention and management of obesity and T2DM. The main
preventative public health initiatives are centered around maintaining a healthy body weight,
physical activity and nutrition [12, 28]. Obesity and T2DM management target the adoption of a
healthy lifestyle with a focus on dietary, physical activity, and weight loss approaches, including
behavioural interventions. [12, 28, 29]. Pharmacologic interventions for the treatment of obesity
may be implemented in adults who are unable to attain weight loss through diet and exercise
[28]. Individuals with a BMI ≥40 are candidates for bariatric surgery [36]. For T2DM, the
addition of pharmacotherapy is considered when glycemic control is not achieved within 3
months of behavioral interventions [12]. Given the fiscal costs and personal stigma associated
with these diseases, focusing on prevention will provide substantial economic benefits, while
enhancing quality of life.
2.2 Sugar-Sweetened Beverages
Although several dietary factors are linked to increased body weight and T2DM risk,
considerable evidence indicates an association between excess sugar-sweetened beverage (SSB)
intake and increased risk of obesity, T2DM and cardiometabolic risk [6, 7, 45-48]. The
8
mechanism for these outcomes is likely mediated through an excess intake of energy and through
metabolic effects of certain sugars present in SSBs [49].
Considering that a standard 355mL (12 fl oz) can of regular soda provides ~35-42g of sugar and
140-160 calories, it is likely that SSBs contribute to weight gain by reduced compensation of
subsequent energy intake coupled with a decreased level of satiety [49]. Increased caloric
consumption and weight gain with energy-matched intakes of beverages compared to solid
foods, have been demonstrated in short-term feeding trials suggesting that liquid calories lack
appetite-suppressing characteristics [50-53].
Simultaneously, the glycemic index values of SSBs are moderate-to-high [54], and rapidly spike
postprandial blood glucose and insulin levels [55, 56]. When consumed in large amounts, SSBs
contribute to a high dietary glycemic load (GL), which is of relevance as high-GL diets have
been shown to stimulate insulin resistance [57], aggravate inflammatory biomarkers [58] and are
linked to a higher risk of T2DM and coronary heart disease (CHD) [49].
Furthermore, SSBs contain fructose as a constituent of sucrose or high fructose corn syrup [49].
A moderate consumption of fructose is converted to glucose, lactate and fatty acids in the liver
[49]. However, excess intakes can lead to increased hepatic de novo lipogenesis, triglycerides,
cholesterol, glucose, insulin resistance, uric acid and NAFLD markers [49, 59-63]. A visual
representation of excess SSB consumption and the development of obesity, T2DM and
cardiometabolic risk is shown in Figure 2.1.
9
1Figure 2.1 SSB consumption and the development of obesity, T2DM and Cardiometabolic
Risk[49]
2.2.1 Worldwide Intakes of Sugar-Sweetened Beverages
Despite decreased consumption of SSBs in North America, overall intakes are still above current
WHO [10] recommendations of a maximum of 5-10% of total energy [49, 64-66]. According to
the latest Canadian Community Health Survey Nutrition data, SSBs were the most important
source of free/added sugars among all age groups, where regular soft drinks accounted for 8.9%
of total sugars in 2015 compared to 14.6% in 2004 [64]. Similarly, the most recent results from
the 2011 to 2014 National Health and Nutrition Examination Survey, showed that SSBs provided
6.5% of total calories among adults in the United States [67].
Contrary to high-income countries like Canada and the United States, the increase in
urbanization and beverage marketing in several low and middle-income countries are implicated
in their rising intakes of SSBs [49]. Results from a 2010 survey by Singh et al. [68] of 187
countries found adults living in upper-middle-income and lower-middle-incomes countries had
higher SSB intakes compared to those living in high-income and low-income countries. Of the
21 regions surveyed, the Americas, especially parts of Latin America and the Caribbean, had the
10
highest SSB intakes compared to East Asia which had the lowest [68]. Similarly, in a 2016
review investigating global SSB sales based on calories per person per day, found that per capital
sales of SSBs had increased in lower-middle-income countries and decreased in some high-
income countries [69].
In regard to age and gender, Singh et al. [68] report that adults over the age of 20 consumed an
average of 0.58 (8oz) servings of SSBs per day. Highest SSB intakes were among men aged 20-
39 (1.7 servings per day), whereas women aged 60 plus had the lowest intakes (0.53 servings per
day) [68]. Inequalities in SSB consumption run parallel with inequalities in obesity and T2DM
prevalence in that groups of lower socio-economic status tend to have higher SSB intakes as well
as a higher risk for obesity and T2DM [49].
2.2.2 Guidelines for Sugar-Sweetened Beverage Consumption
Current recommendations set by the WHO and endorsed by national authorities including
Canada, the United States and the United Kingdom, call for reducing free/added sugars to less
than 5-10% of total calories, particularly those coming from SSBs [10, 11, 13, 70]. In line with
these recommendations, public polices exist which aim towards reducing intakes of SSBs
through replacement strategies, taxation, marketing restrictions, limiting availability in schools,
public awareness campaigns and front-of-pack labelling [49]. Of these policies, SSB taxation
shows global momentum as the funds raised can be put toward medical, and public health
initiatives and interventions that target obesity [71]. Evidence from Mexico’s National Institute
of Public Health showed that two years after the Mexican government levied an SSB tax of 1
peso (about $0.07) per liter ($0.02 per ounce) in January 2014, SSB purchases decreased an
average of 7.6% for the general population and by 11.7% among households with the fewest
resources [71, 72].
2.2.3 Replacements for Sugar-Sweetened Beverages in Guidelines
When considering replacement beverages for SSBs, the consensus among several health
authorities is to consume water [11, 13, 14, 73-78]. Non-nutritive sweetened beverages (NSBs)
mimic the sweet taste of SSBs without the associated energy. Despite the potential benefits of
NSBs as displacements of excess calories from SSBs, concerns exist that non-nutritive
sweeteners (NNSs) contribute to an increased risk of obesity and T2DM [19, 21]. Implicated
biological mechanisms for these concerns include impaired sensory and endocrine signaling
11
mediated by the sweet taste receptor [79, 80] and changes to the gut microbiome [80, 81]. As a
result, recommendations for their consumption are inconsistent across dietary and clinical
practice guidelines [11-18]. The US dietary guidelines for Americans and Canada’s Food Guide
discourage the use of NSBs, recommending that water and not NSBs replace SSBs [11, 13]. The
American Heart Association, including diabetes associations in the UK, US, and Canada support
NSBs insofar as they are used to displace calories from sugars and sugar-sweetened beverages
[12, 15, 16, 18]. The European Association for the Study of Diabetes (EASD), on the other hand,
has not made any specific recommendations about NNSs or NSBs [17]. Table 2.4 summarizes
the NNS recommendations in the guidelines from these authoritative bodies.
12
4Table 2.4 Non-Nutritive Sweetener Guideline Recommendations
Guideline Non-Nutritive Sweetener Recommendations
Canada's Food Guide, 2018[11] • “Sugar substitutes do not need to be consumed to reduce intake of free sugars.”
• “No well-established health benefits associated with the intake of sweeteners.”
• “Water, unsweetened milk or fortified soy beverage, and fruit should be offered instead.”
Diabetes Canada, 2018[12] • “The evidence from SRMAs of RCTs,…have shown a weight loss benefit when NNSs are used to
displace excess calories from added sugars (especially from SSBs) in overweight children and
adults without diabetes, a benefit…similar to that seen with… water.”
Dietary Guidelines for
Americans, 2015[13]
• “Replacing added sugars with high-intensity sweeteners may reduce calorie intake in the short-
term, yet questions remain about their effectiveness as a long-term weight management strategy.”
• “…added sugars should be reduced in the diet and not replaced with NNSs, but rather with healthy
options, such as water in place of SSBs.”
American Diabetes
Association, 2020[15]
• “For those who consume SSBs regularly, a low-calorie or NSB may serve as a short-term
replacement strategy, but overall, people are encouraged to decrease both sweetened and NSBs
and use other alternatives, with an emphasis on water intake.”
American Heart Association,
2018[18]
• “Prudent to advise against prolonged consumption of NSBs by children.”
• “For adults who are habitually high consumers of SSBs, …NSBs may be a useful replacement
strategy to reduce intake of SSBs…”
• …“the use of other alternatives to SSBs, with a focus on water (plain, carbonated, and
unsweetened flavored), should be encouraged.”
Diabetes UK, 2018[16] • “For people who are accustomed to sugar sweetened products, NNSs have the potential to reduce
overall energy and carbohydrate intake and may be preferred to sugar when consumed in
moderation”
NSBs, non-nutritive sweetened beverages; NNS, non-nutritive sweetener; SRMA, systematic review and meta-analysis; UK, United Kingdom; WHO, World
Health Organization.
13
2.3 Non-Nutritive Sweeteners
2.3.1 Definition of Non-Nutritive Sweeteners
Non-nutritive sweeteners (NNSs) are sugar substitutes and are also referred to as high-intensity
sweeteners, artificial sweeteners or low-calorie sweeteners [12]. Sugar alcohols, which are also
sugar substitutes, belong to a family of sweeteners called “polyols”. They have a chemical
structure that’s similar to sugar but contain less calories and are only partly absorbed by the
body. As a result, when consumed in amounts greater than 10g, gastrointestinal (GI) discomfort
can occur [82, 83]. NNSs, are significantly sweeter than sucrose (table sugar) and unlike sugar
alcohols their use is not self-limiting from GI symptoms [12, 84]. This distinction is important as
sugar alcohols, aside from erythritol, don’t have Acceptable Daily Intake (ADI) values because
of their potential for GI distress [12]. Even though sugar alcohols are sugar substitutes they
contain more calories than high-intensity sweeteners and, therefore, will not be included in this
thesis when referring to NNSs [85].
2.3.2 Safety and Approval Process
The safety of NNSs has been approved by multiple regulatory bodies, including the European
Food Safety Authority, U.S. Food and Drug Administration and Health Canada [86-88].
Internationally, the Joint Food and Agriculture Organisation (FAO)/WHO Expert Committee on
Food Additives (JECFA) and the Codex Alimentarius Commission (CAC), act to harmonize the
use of NNSs around the world. Comprised of a group of international scientific experts, the
JECFA conducts risk assessments and evaluates the safety of food additives in order to advise
the CAC, FAO, WHO, and its member countries [89] including Canada [90].
NNSs are considered one of the most widely researched food additives and have undergone
extensive toxicology testing in both human and animal safety studies [91]. These toxicology
studies provide the necessary information to identify any potential hazards associated with a
NNS and it’s corresponding No Observed Adverse Effect Level (NOAEL). The NOAEL is the
dose of a substance that does not cause any adverse effects when consumed daily and is
established from long-term, repeated-dose animal studies. Derived from the NOAEL is the ADI
which is calculated by dividing the NOAEL by a typical safety factor of 100. The ADI was
developed by the JECFA and is defined as “the amount of a food additive, expressed on a body
14
weight basis, that can be consumed daily over a lifetime without appreciable health risk”. The
ADI then acts as a safeguard to protect an entire population, including the most susceptible and
those with the highest potential for exposure by ensuring a wide margin of safety [91].
In Canada, NNSs are regulated as food additives under the Food and Drug Regulations and
associated Marketing Authorizations [89]. Only those permitted for use are listed under Health
Canada’s “List of Permitted Sweeteners” [92]. There are currently 9 NNSs (advantame,
acesulfame potassium (Ace-K), aspartame, saccharin, monk fruit extract, neotame, steviol
glucosides, sucralose and thaumatin) approved for use in foods, beverages and/or as table-top
sweeteners by Health Canada [92]. Table 2.5 lists the most common NNSs used in foods,
beverages or as table-top sweeteners in Canada. Of these, aspartame, Ace-K and sucralose are
the most common NNSs found in NSBs [93, 94].
15
5Table 2.5 Common Non-Nutritive Sweeteners Used in Foods and Beverages in Canada[82, 91, 92]
Sweetener Common Forms & Uses Brand Name
Examples
Sweetness
Intensity1
ADI2
Approx. Amount
(mg) in NSBs
No. of NSBs, Tablets or
Packets Equivalent to
ADI3
Acesulfame
Potassium
(Ace-K)
• Although permitted, currently not available
as a table-top sweetener
• Foods and beverages
• Not
applicable
200x 15 42 24 NSBs
Aspartame • Table-top sweetener
• Foods and beverages
• NutraSweet®
Equal®
• Private label
brands4
200x 40
200 14 NSBs
Saccharin • Approved for use in foods and beverages
• Only available in pharmacies as a table-top
sweetener
• Hermesetas® 300x 5
12 (one tablet
Hermesetas®)
28 tablets
Steviol
glycosides
(Stevia)
• Table-top sweetener
• Foods and beverages
• Truvia®
• PureVia®
• Sugar Twin®
Stevia
• Private label
brands4
200-300x 4
>230mg stevia
leaf extract
(Zevia®)5
1 NSB
Sucralose • Table-top sweetener
• Foods and beverages
• Splenda®
• Private label
brandsⱡ
600x 56
18 (Diet
Mountain
Dew®)[95]
19 NSBs
Cyclamate • Table-top sweetener
• Not permitted to be used in foods or
beverages
• Sugar Twin®
• Sweet'N
Low®
• Private label
brands4
30x[96] 11
264 (one packet
Sugar Twin®)
3 packets
ADI, acceptable daily intake; mg, milligram, No., number; NSB, non-nutritive sweetened beverage 1Gram-for-gram sweetness compared to sucrose (table sugar). 2ADI established by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) and reflect those used by Health Canada. Based on mg per kg of body
weight per day. 3Number of NSBs, tablets or packets a 68.2kg (150 pound) person would need to consume to reach the ADI. 4Private label brands include those such as No Name®, President’s Choice®, IrresistiblesTM and Compliments. 5 Received from Zevia customer relations representative. Due to proprietary issues, unable to provide exact quantities of stevia leaf extract per can. 6The ADI for sucralose is 5mg in Canada vs 15mg as per the JECFA.
16
2.3.3 Sweetening Intensity
The sweetening intensity of NNSs are several times higher than sucrose. As a result, very little is
required to elicit a sweet taste. Furthermore, NNSs produce bitter or metallic “off tastes” and are
often found as blends of NNSs to enhance the sweetness intensity and taste profile of a food or
beverage, which further limit the amount of a sweetener that can be used [91]. The sweetening
intensity of NNSs are shown in Table 2.5, along with the number of NSBs, packets or tablets a
68kg (150lbs) person would need to consume in order to reach its respective ADIs.
2.3.4 Non-Nutritive Sweetener Consumption Trends
A 2016 review by Sylvetsky et al. [85] on NNS consumption trends indicated that NSBs are
consumed in the largest proportion of total NNSs globally, exist in many food products, and are
often added to foods and beverages by consumers and product developers. In addition to their
widespread use, the demand for NSBs, is expected to rise [97]. Reasons for the demand are
attributed to increasing awareness and preference for SSB alternatives, and to combat the rise in
health-related problems associated with an excess intake of SSBs [97].
2.3.5 Absorption, Digestion, Metabolism and Excretion of Non-Nutritive
Sweeteners
The biological pathways of NNSs include absorption, distribution, metabolism, and excretion
(ADME), and are established through extensive animal studies in order to determine their safety.
There are considerable differences in the ADME patterns of different NNSs. Knowing the
ADME of each is important in determining if their purported biological effects are indicative of
potential health consequences. A visual representation of the similarities and differences in the
ADME pathways of aspartame, sucralose, Ace-K and saccharin are shown in Figure 2.2.
2.3.5.1 Aspartame
Aspartame is completely digested by digestive enzymes in the GI tract into methanol and the
amino acids phenylalanine and aspartic acid. These byproducts are in the same form as those
derived from foods such as fruits, vegetables and dietary proteins including, meat, fish, poultry,
eggs, diary and legumes. The amino acids are used for protein synthesis and metabolism. Any
excess is excreted in the urine. Methanol is converted to formaldehyde in the liver and used by
17
the body or converted to formic acid and excreted in the urine or broken down to carbon dioxide
and water [91].
2.3.5.2 Sucralose
Sucralose, although similar in structure to sucrose, is not digested into monosaccharides and
undergoes little to no metabolism, thereby providing no calories and no impact on blood glucose.
Only a small amount is absorbed. It is primarily excreted, unchanged, in the feces with
approximately 9-22% excreted in the urine as unchanged sucralose [98]. Due to its low level of
absorption and systemic clearance, the likelihood of sucralose accumulation in the body from
chronic consumption is low [91].
2.3.5.3 Acesulfame Potassium and Saccharin
Ace-K and saccharin are the only two NNSs that are absorbed as intact molecules. Ace-K is
quickly absorbed and distributed to tissues throughout the body via the bloodstream and is
primarily eliminated in the urine with less than 1% excreted in the feces [91].
About 85% to 95% of saccharin is absorbed and excreted in the urine and the remainder is
excreted in the feces [91]. Saccharin was approved for use in foods and beverages in Canada
since 2016, yet is currently only available as the table-top sweetener Hermesetas® [99]. Even
though saccharin is not currently found in beverages in Canada, it’s ADME pathway is worth
mentioning as it has been implicated in the development of glucose intolerance through
alterations in the gut microbiome [81].
18
2Figure 2.2 Absorption, Digestion, Metabolism and Excretion of Acesulfame Potassium,
Saccharin, Aspartame and Sucralose[91]
19
2.3.6 Potential Mechanisms of Non-Nutritive Sweeteners on Body Weight and
Glucose Control
Several potential biological mechanisms suggest an association between NNS intake, weight
gain, glucose intolerance and other negative outcomes. As the ADME pathways vary between
NNSs, so too may their physiological and behavioural effects. These mechanisms, which have
mainly been explored in animal studies, include the activation of sweet taste receptors, changes
in taste preferences, and alterations in the gut microbiome [79, 80, 100].
2.3.6.1 Activation of Sweet Taste Receptors
Sweet taste receptors (STRs), a heterodimeric G-protein coupled receptor (GPCR) comprised of
taste 1 receptor member 2 and taste 1 receptor member 3 subtypes, are activated by sweet tasting
compounds, including NNSs. When activated in the oral cavity, STRs release neurotransmitters
alerting the brain to sweetness [79, 80], while signaling cephalic phase responses (CPRs) which
prime the body to optimize food digestion and nutrient absorption [80]. The most investigated
CPR is the cephalic phase insulin response (CPIR), characterized by a small, neutrally-mediated
release of insulin prior to nutrient absorption [101]. Saccharin is the only NNSs that has been
reported to stimulate the CPIR in humans [102]. Although it’s been postulated that the repeated
activation of the CPIR by saccharin may weaken the body’s ability to respond to sugar
appropriately [103], studies to date in humans indicate non-significant effects on glucose
homeostasis [104].
Aside from the transduction of taste in the mouth, results from animal studies and human cell
lines indicate that the activation of STRs in the pancreas increases insulin secretion [105-107],
and when activated in the intestine, increase the rate of glucose absorption [108-111] and
glucagon-like peptide-1 (GLP-1) [112]. In human studies, when NNSs were administered with
oral glucose, an augmentation of insulin and/or GLP-1 was demonstrated [113-116]. These
hormonal responses, however, were not observed in the majority of human studies that
administered NNSs alone [79].
2.3.6.2 Changes in Taste Preferences and Dietary Intake
As there is an innate preference for sweetness [79, 117] it has been theorized that NNSs, due to
their sweetness intensity, elicit greater preference for sweet tastes leading to poor dietary
choices, increased energy intake, and resultant obesity [79]. This hypothesis has been observed
20
in rodent studies where early-life exposure resulted in a greater preference for sweetness [118].
Overall evidence from adult cross-sectional studies comparing NSB to SSB consumers show an
improvement in diet quality among NSB consumers [117, 119-121]. Furthermore, a 2017 cross-
sectional analysis by Leahy et al. [122] comparing NSB to water consumers showed higher NSB
intakes were associated with a lower intake of carbohydrates, total sugars and added sugars
compared to higher water consumers, suggesting that NSBs, like water, can be a viable option
for reducing sugar intake. Data from intervention trials among adults and children, show that
when sweetness with calories (such as SSBs) is replaced by sweetness without calories (such as
NSBs), energy intake is reduced, and its effect on appetite and energy intake is comparable to
water [20, 117, 123].
2.3.6.3 Gut Microbiome
Several rodent studies have indicated a link between NNSs and alterations of the gut
microbiome, particularly with the NNSs saccharin and sucralose [81, 124-129]. Data on the
effects of NNSs on the human intestinal microbiota, however, are limited and have only been
reported for saccharin, sucralose and steviol glycosides [81, 130, 131].
The most influential human intervention on NNSs and gut microbiome changes was a small,
non-randomized 2014 study by Suez et al. [81]. In this trial, 7 healthy volunteers were given the
max ADI of saccharin (5mg/kg/d) every day for a week and monitored by continuous glucose
measurements and daily OGTTs. Participants that developed poor glycemic responses were
classified as “responders” (n=4), whereas those who did not were classified as “non-responders”
(n=3). Stool from both “responders” and “non-responders” were transferred to germ-free mice.
The mice that received stool from the “responders” developed glucose intolerance suggesting a
causal role for saccharin-induced microbiota alterations. Given the small sample size (n=7), lack
of control group and dose of saccharin given (an amount not found in “real-world” conditions),
the interpretability of this study is severely limited, yet citations for this paper have exceeded
1000 [130].
21
2.3.7 Systematic Reviews and Meta-Analyses of Prospective Cohort Studies
Investigating the Association between Non-Nutritive Sweeteners and Obesity/
Type 2 Diabetes Mellitus Risk
Prospective cohort studies (PCS) consistently demonstrate a positive association for harm with
NNS intake (mainly from NSBs) on weight gain, obesity, T2DM and other cardiometabolic
outcomes including MetS, hypertension, stroke and CVD [7, 19, 20, 22, 132, 133]. Although
PCS provide the highest level of observational evidence to assess the relationship between NSBs
and obesity/T2DM risk, their results may be complicated by reverse causality, residual
confounding or mediation. With reverse causality, many individuals who consume NSBs may be
doing so because they are overweight or obese and, therefore, at an elevated risk for T2DM and
other cardiometabolic risk factors [18, 134]. An emerging method used to address this issue is by
using change in intake to change in outcome analysis [18, 134]. When this type of method is
applied, negative associations are seen between NNS intake and body weight [18, 134-136].
Residual confounding, another inherent limitation with prospective cohort studies, could also
explain the observed association between NNSs and increased risk for T2DM. Even when
relevant covariates such as diet or physical activity are adjusted for, findings may be biased due
to residual confounding [134]. With mediation, an intervening variable or mediator is influenced
by the exposure which in turn influences the outcome [137]. In a 2015 systematic review and
meta-analysis (SRMA) by Imamura et al. [7], the association between a higher intake of NSBs
and risk of T2DM was attenuated from 25% (RR=1.25, 95% CI 1.18 to 1.33) to 8% (RR=1.13,
85% CI 1.02 to 1.15) when adjusted for adiposity, suggestive of mediation for the observed
association as obesity lies on the causal pathway between NSBs and T2DM.
22
2.3.8 Systematic Reviews and Meta-Analyses of Randomized Controlled Trials
Investigating the Effect of Non-Nutritive Sweeteners on Obesity/ Type 2 Diabetes
Mellitus Risk
Contrary to PCSs, RCTs provide the greatest level of protection against bias from confounding
and reverse causality. Results from syntheses of RCTs, however, are not consistent indicating
non-significant or beneficial effects of NNSs on outcomes of adiposity (Table 2.6) and glucose
control (Table 2.7) [19-21, 24, 26, 138]. A major limitation with the majority of these analyses is
the failure to account for the nature of the comparator. When NNSs are compared against energy
matched controls such as water or another NNS, non-significant effects are seen, as these
comparators do not allow for the displacement of energy by the NNS. On the other hand, when
NNSs were compared against sugary controls, including SSBs, a benefit on outcomes of
adiposity and glucose control were observed. Failing to account for the nature of the comparator
resulted in conclusions of “no compelling evidence” for the “intended benefits” of NNSs in a
WHO-commissioned analysis by Toews et al. [21] and a previous SRMA by Azad et al. [19]. In
contrast, a 2020 review by Laviada-Molina et al. [25] and 2016 review by Rogers et al. [20],
distinguished between the effects of NNSs based on the nature of the comparator. In the SRMA
by Rogers et al. [20], although outcomes of glucose control were not assessed, a reduction in
body weight was observed with NNSs compared to sugar-sweetened products among 8 RCTs in
adults (-1.41kg, 95% CI -2.62 to -0.20). When used as a replacement for excess sugars, the
weight of the evidence from SRMAs of RCTs shows benefit with NNS intake on body weight
[20, 21] and fasting glucose [21, 26], with no adverse effect on outcomes of adiposity or glucose
control in individuals with and without diabetes [19-21, 24, 26, 123, 138], a benefit that has
shown to be comparable to displaced excess energy from other interventions, such as water [12].
Other limitations of these reviews include an overly strict selection criteria resulting in the
exclusion of studies with relevant data that could have been included. Azad et al. [19] restricted
studies to ≥6 months, Toews et al. [21] only included studies that specified the NNS, Onakpoya
et al. [26] limited the NNS in their review to stevia-only, while Santos et al. [24] only assessed
aspartame intake, and Wiebe et al. [138] excluded trials with placebo controls. The results from
these reviews were also based on a limited number of studies, thereby reducing their external
validity.
23
6Table 2.6 Systematic Reviews and Meta-Analyses of Randomized Controlled Trials Investigating the Effect of Non-Nutritive
Sweeteners on Adiposity
Study No. of RCTs (participants/population) NNS
Source
Comparator Relevant Results
Laviada-
Molina et
al. 2020
4 (554/ OW, OB adults) NNSs Water • Non-significant effects on BW/BMI with NNSs (SMD=-
0.20, 95% CI -0.62 to -0.23)
13 (1997/ NW, OW, OB children,
adolescents and adults)
NNSs Sugary controls1 • Significant reduction in BW/BMI with NNSs (SMD=-
0.56, 95% CI -0.79 to -0.34)
Toews et
al. 2019
5 (356/ NW, OW, OB adults)
NNS Placebo or sugary
control1
• Non-significant effects on BW with NNSs (MD=1.29kg,
95% CI -2.80 to 0.21)
3 (246/ OW, OB adults)
NNS Sugary control1 • Significant reduction in BW with NNSs (MD=-1.99kg,
95% CI -2.84 to -1.14)
2 (174/ OW, OB adults)
NNS Sugary control1 • Reduced BMI with NNSs (MD=-0.60kg/m2, 95% CI -1.19
to -0.01)
Santos et
al. 2018
3 (179/ OW, OB adults, with or without
T2DM)
NNS
(Asp)
Sugary control1 • Non-significant effects on BW with NNSs (MD=-5.0kg,
95% CI -11.56 to 1.56)
2 (179/ OW, OB adults with or without
T2DM)
NNS
(Asp)
Sugary control1 • Non-significant effects on BW with NNSs (MD=-3.78kg,
95% CI -9.74 to 2.18)
Azad et al.
2017
5 (791/ OW, OB adults)
NSBs, or
mixture
of NNSs
Water or asp avoidance
• Non-significant effects on BW with NNSs (SMD=-0.17,
95% CI -0.54 to 0.21)
3 (242/ NW, OW, OB, or mild HTN
adults)
NSBs, or
stevioside
capsules
Water or placebo • Non-significant effects on BMI with NNSs (MD=-
0.37kg/m2, 95% CI -1.10 to 0.36)
3 (683/ OW, OB adults on weight-loss
program)
NSBs Water
• Non-significant effects on WC with NSBs (SMD=-0.16,
95% CI -0.56 to 0.25)
1 (25/ OW, OB adults on weight-loss
program)
NSBs Water
• Non-significant effects on % body fat with NSBs (MD=-
1.01%, 95% CI -3.01 to 0.99)
Rogers et
al. 2016
8 (1332/ OW, OB adults)
NNSs Sugary control1 • Significant reduction in BW with NNSs (MD=-1.41kg,
95% CI -2.62 to -0.20)
1 (641/ NW children) NSBs SSBs
• Significant reduction in BW with NSBs (MD=-1.02kg,
95% CI -1.52 to -0.52)
24
Miller at al.
2014
10 (796/ NW, OW, OB adults) NNS Sugary controls1 • Significant reduction in BW with NNSs (WGMD=-
0.72kg, 95% CI -1.15 to -0.30)
5 (475/ NW, OW, OB adolescents and
adults)
NNS Sugary controls1 • Significantly reduced BMI with NNSs (WGMD=-
0.24kg/m2, 95% CI -0.41 to -0.07)
5 (1000/ NW, OW, OB children,
adolescents and adults)
NNS Sugary controls1 • Significantly reduced fat mass with NNSs (WGMD=-
1.10kg, 95% CI -1.77 to -0.44)
3 (981/ NW, OW, OB children, and
adults)
NSBs SSBs • Significantly reduced WC with NNSs (WGMD=-0.83cm,
95% CI -1.29 to -0.37)
Asp, aspartame; CI, confidence interval; BMI, body mass index; BW, body weight; HTN, hypertension; NNSs, non-nutritive sweeteners; No., number; NSBs,
non-nutritive sweetened beverages; NW, normal weight; OB, obese; OW, overweight; T2DM, type 2 diabetes mellitus; WC, waist circumference; WGMD,
weighted group mean difference. 1Sugary controls include those coming from foods and/or beverages.
25
7Table 2.7 Systematic Reviews and Meta-Analyses of Randomized Controlled Trials Investigating the Effect of Non-Nutritive
Sweeteners on Diabetes Risk
Study No. of RCTs
(participants/population)
NNS Source Comparator Relevant Results
Toews et al.
2019
2 (52/ OW adults)
NNSs Sugary controls1 • Significant reduction in FBG with NNSs (MD = 0.16
mmol/L, 95% CI -0.26 to -0.06)
2 (66/ OW adults)
NNSs Sugary controls1 • No difference observed for plasma insulin (MD = -1.60
pmol/L, 95% CI -8.39 to 5.19)
2 (66/ OW adults)
NNSs Sugary controls1 • No difference observed for HOMA-IR (MD = -0.14, 95% CI
-0.38 to 0.10)
Azad et al.
2017
3 (99/ OW adults)
1 (62/ OW females)
NNSs
NSBs
Water or placebo
Water
• Non-significant effects for insulin resistance (HOMA-IR)
(MD = 0.10, 95% CI, -0.57 to 0.76)
• Non-significant effects on HbA1c (MD = 0.07%, 95% CI -
0.00 to 0.14). Follow-up = 24 weeks
Santos et al.
2018
5 (119/ NW, OW, OB adults with
or without T2DM)
NNS (Asp) Control • Non-significant effects on fasting blood glucose (MD = -
0.03, 95% CI -0.14 to 0.21)
3 (58/ NW, OW, OB adults) NNS (Asp) Sucrose • Non-significant effects on fasting blood glucose (MD = -
0.13, 95% CI -0.95 to 0.69)
4 (101/ NW, OW, OB adults with
or without T2DM)
NNS (Asp) Control • Non-significant effects on insulin levels (MD = -0.13, 95%
CI -0.95 to 0.69)
2 (40/ NW, OW, OB adults)
NNS (Asp) Sucrose • Non-significant effects on insulin levels (MD = -2.54, 95%
CI -11.37 to 6.29)
Wiebe et al.
2011
1 (10/T1DM adults)
NNSs Sugary controls1 • No effect of sweetener type on HbA1c (MD = -0.02%, 95%
CI -0.40 to 0.30) over 4 weeks in a cross-over study
1 (41/OW adults)
NNSs Sugary controls1 • No effect of sweetener type on HOMA-IR (MD = -0.20, 95%
CI -0.58 to 0.18) over 10 weeks in a parallel trial
Onakpoya
et al. 2015
6 (551/ OW, OB adults with or
without T1DM, T2DM, HTN and
hyperlipidemia)
Stevioside
capsules
Matching placebo or
nothing
• Significant reduction in FBG in favour of steviol glycoside
(MD = -0.63 mmol/L, 95% CI -0.90 to -0.36)
Asp, aspartame; BG, blood glucose; CI, confidence interval; FBG, fasting blood glucose; HbA1c, hemoglobin A1c; HOMA-IR, homeostatic model assessment
for insulin resistance; HTN, hypertension; IGT, impaired glucose tolerance; MetS, metabolic syndrome; NNSs, non-nutritive sweeteners; NSBs, non-nutritive
sweetened beverages; NW, normal weight; OB, obese; OW, overweight; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus. 1Sugary controls include those coming from foods and/or beverages
26
2.3.9 Network Meta-Analyses
Compared to standard pairwise meta-analyses, network meta-analyses (NMAs) are extensions
that can be used to simultaneously compare multiple treatments that have not been directly
compared to one another in trials [139]. By comparing multiple treatments in the same model,
NMAs yield more information on the comparative effects of all interventions by using both
direct comparisons of interventions within RCTs and indirect comparisons across trials based on
a common comparator [140]. Figure 2.2 provides a simplistic representation of how direct and
indirect evidence are combined in a NMA [141].
While previous SRMAs have explored NNSs, none to date have simultaneously compared the
effects of NSBs, SSBs and water. Due to the importance of the nature of the comparator in
drawing inferences on the effects of NSBs, NMA techniques provide a statistical means of
leveraging all 3 beverage comparisons that were not directly compared head-to-head in trials.
3Figure 2.3 Network Meta-Analysis Diagram Combining Direct and Indirect Evidence[141]
27
Chapter 3: Rationale and Objectives
3.1 Rationale
Health authorities discourage consumption of SSBs as they have been consistently associated
with weight gain, type 2 diabetes mellitus (T2DM) and downstream complications including
hypertension and coronary heart disease [6-9]. Non-nutritive sweetened beverages (NSBs)
provide a viable alternative for SSBs, yet dietary and clinical practice guidelines are inconsistent
in their recommendations for NSBs out of concern that they do not have established benefits and
recommend that water replace SSBs [11-14, 142]. Systematic reviews and meta-analyses
assessing the totality of evidence from prospective cohort studies on the associations of NSBs on
obesity, diabetes and cardiometabolic risk suffer from inherent limitations of reverse causality,
residual confounding and mediation [18, 134, 137]. Meanwhile, syntheses from RCTs failed to
consider the nature of the comparator by evaluating non-nutritive sweeteners (NNSs) to sugar
(SSBs, sugary foods, etc.) or calorie-matched controls (another NNS, water, placebo, etc.) [19,
21, 24, 26, 138]. There also exists concern that NNSs may induce glucose intolerance through
changes in the gut microbiome [81]. To date, no syntheses of randomized controlled trials
(RCTs) has assessed the effect of substituting NSBs for SSBs, while comparing their
effectiveness to water. Whether NSBs effect cardiometabolic risk factors similar to water in their
intended substitution for SSBs remains unclear.
3.2 Objectives
The overall objective of this thesis is to assess the effect of NSBs as a replacement for SSBs
through 3 prespecified substitutions: NSBs for SSBs (intended substitution with caloric
displacement), water for SSBs (“standard of care” substitution with caloric displacement), and
NSBs for water (matched substitution without caloric displacement) on measures of body
weight, gut microbiome, glucose tolerance and cardiometabolic risk factors.
Specific objectives include:
1. Conduct a systematic review and network meta-analysis of RCTs assessing the effect of
NSBs, as a replacement for SSBs, through 3 prespecified substitutions (NSBs for SSBs,
water for SSBs, and NSBs for water) on measures of body weight and cardiometabolic
28
risk factors using Grading of Recommendations Assessment, Development, and
Evaluation (GRADE) to assess the certainty of evidence.
2. Provide a rationale, design and baseline characteristics report investigating the effect of
NSBs, as a replacement for SSBs, through 3 prespecified substitutions (NSBs for SSBs,
water for SSBs, and NSBs for water), on gut microbiome, glucose tolerance and
cardiometabolic risk factors over 4-weeks in overweight/obese participants who are
regular SSB drinkers: Strategies To OPpose SUGARS with Non-nutritive sweeteners Or
Water trial (STOP Sugars NOW).
29
Chapter 4: Effect of Non-Nutritive Sweetened Beverages as a
Replacement Strategy for Sugar-Sweetened Beverages on Body Weight
and Cardiometabolic Risk: A Network Meta-Analysis of Randomized
Controlled Trials
Néma D McGlynn1,2, Tauseef A Khan1,2, Lily Wang1,3, Roselyn Zhang1,4, Laura Chiavaroli1,2,
Fei Au-Yeung1,2, Elena M Comelli2,5, Vasanti S Malik3,6, James O Hill7, Lawrence A
Leiter1,2,8,9,10, Arnav Agarwal3, Per B Jeppesen11, Dario Rahelić12,13,14, Hana Kahleová15,16, Jordi
Salas-Salvadó17,18,19, Cyril WC Kendall1,2,20, John L Sievenpiper1,2,8,9,10
1Toronto 3D Knowledge Synthesis and Clinical Trials Unit, Clinical Nutrition and Risk Factor Modification Centre,
St. Michael's Hospital, Toronto, ON, Canada; 2Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 3Department of Medicine, Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 4Applied Human Nutrition, Mount Saint Vincent University, Halifax, NS, Canada; 5Joannah and Brian Lawson Centre for Child Nutrition, Toronto, ON, Canada; 6Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA; 7Department of Nutrition Sciences, The University of Alabama at Birmingham, Birmingham, Alabama, USA; 8Division of Endocrinology and Metabolism, Department of Medicine, St. Michael’s Hospital, Toronto, ON,
Canada; 9Department of Medicine, Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 10Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada; 11Department of Clinical Medicine, Aarhus University, Aarhus University Hospital, Aarhus, Denmark; 12Vuk Vrhovac University Clinic for Diabetes, Endocrinology and Metabolic Diseases, Merkur University Hospital,
Zagreb, Croatia; 13University of Zagreb School of Medicine, Zagreb, Croatia; 14University of Osijek School of Medicine, Osijek, Croatia; 15Institute for Clinical and Experimental Medicine, Diabetes Centre, Prague, Czech Republic; 16Physicians Committee for Responsible Medicine, Washington, DC, USA; 17Universitat Rovira i Virgili, Departament de Bioquímica i Biotecnologia, Unitat de Nutrició Humana, Reus, Spain; 18Consorcio CIBER, M.P. Fisiopatología de la Obesidad y Nutrición (CIBERObn), Instituto de Salud Carlos III,
Madrid, Spain; 19Institut d'Investigació Pere Virgili (IISPV), Reus, Spain; 20College of Pharmacy and Nutrition, University of Saskatchewan, Saskatoon, SK, Canada
Corresponding Author: John L Sievenpiper MD, PhD, FRCPC, St. Michael's Hospital, #6138-61
Queen Street East, Toronto, ON, M5C 2T2, CANADA, Tel: 416 867 7475, Fax: 416 867 7495,
email: [email protected]
Keywords: non-nutritive sweetener, low-calorie sweetener, sugar-sweetened beverage, water, cardiometabolic risk,
systematic review, meta-analysis, network meta-analysis
30
4.1 Abstract
Objective: Concerns exist that non-nutritive sweetened beverages (NSBs) do not have
established benefits, with major dietary guidelines recommending that water and not NSBs
replace sugar-sweetened beverages (SSBs). Whether NSBs improve cardiometabolic risk factors
similar to water in their intended substitution for SSBs is unclear. To inform the update of the
European Association of the Study of Diabetes (EASD) clinical practice guidelines for nutrition
therapy, we conducted a systematic review and network meta-analysis to assess the effect of 3
prespecified substitutions (NSBs for SSBs, water for SSBs and NSBs for water) on body weight
and cardiometabolic risk factors in adults with and without diabetes.
Methods: We searched MEDLINE, Embase and the Cochrane Library through March 2019.
Randomized controlled trials (RCTs) with ≥3week interventions comparing NSBs, SSBs, and/or
water were included. The primary outcome was body weight. Secondary outcomes were other
measures of adiposity, glycemic control, blood lipids, blood pressure, nonalcoholic fatty liver
disease and uric acid. Two independent reviewers extracted data and assessed risk of bias. A
network meta-analysis was performed with data expressed as mean (MD) or standardized mean
(SMD) differences with 95% confidence intervals (CI). GRADE assessed certainty of evidence.
Results: We identified 14 RCTs of 21 trial comparisons, (n=1530) substituting NSBs for SSBs
(7 trials, n=467), water for SSBs (2 trials, n=270) and NSBs for water (7 trials, n=838)
predominantly in adults with overweight/obesity at risk for or with diabetes. Substitution of
NSBs for SSBs reduced body weight (MD,−1.11 kg [95% CI,−1.90, –0.32 kg]), as well as BMI
(−0.32 kg/m2 [−0.58, –0.07 kg/m2]), body fat (−0.60% [−1.03, –0.18%]), triglycerides (−0.24
mmol/L [−0.45, -0.02 mmol/L]), and intrahepatocellular lipid (−0.44 SMD [95% CI, −0.69, –
0.19 SMD]). Substitution of water for SSBs reduced only uric acid (-0.05 mmol/L [-0.08, -0.01
mmol/L]). There was no effect of substituting NSBs for water on any outcome except HbA1c
(+0.21% [+0.02, +0.40%]). The certainty of the evidence was moderate (NSBs for SSBs) and
low (water for SSBs and NSBs for water) for body weight and ranged from low to high for all
other outcomes across all substitutions.
Conclusions: The intended substitution of NSBs for SSBs improves body weight and
cardiometabolic risk factors, showing similar benefits to water and without evidence of harm.
The available evidence supports the use of NSBs as an alternative replacement strategy for SSBs
31
in overweight/obese adults at risk for or with diabetes, over the moderate term. There is a need
for more high-quality RCTs.
Protocol registration: ClinicalTrials.gov identifier, NCT02879500
4.2 Introduction
Sugars have emerged as an important public health concern. The evidence on which this concern
is based derives almost exclusively from sugar-sweetened beverages (SSBs) with excess intakes
of SSBs associated with weight gain and its downstream complications including risk of type 2
diabetes mellitus (T2DM), hypertension, and coronary heart disease (CHD) [6-9]. There is an
urgent need for SSB reduction strategies. Whether non-nutritive sweetened beverages (NSBs) as
a replacement strategy for SSBs have the intended benefits remains unclear. Recent systematic
reviews and meta-analyses, including a WHO-commissioned review [21], have shown NSBs to
be associated with an increased risk of the conditions that they are intended to prevent such as
weight gain, diabetes, and cardiovascular disease (CVD) in prospective cohort studies [19] and
have failed to show the expected weight loss and downstream improvements in cardiometabolic
risk factors in randomized controlled trials (RCTs) [19, 21]. Biological mechanisms involving
impaired sensory and endocrine signaling mediated by the sweet taste receptor [79, 80] and
changes to the microbiome [80, 81] have been offered in support of these observations.
A number of methodological considerations, however, have been raised that limit the inferences
that can be drawn from these data. The available prospective cohort studies are at high risk of
reverse causality (people may consume NSBs because they are already overweight/obese or are
at high risk for weight gain) [130, 134, 143]. Furthermore, the syntheses of RCTs have failed to
account fully for the calories available to be displaced by NSBs with caloric (e.g. SSBs) and
noncaloric (e.g. water, placebo, weight loss diets) comparators pooled together, or noncaloric
comparators used as the sole comparator leading to an underestimation of the effect of NSBs in
their intended substitution for SSBs [130, 134, 143].
The prevailing uncertainties have led to mixed recommendations from authoritative bodies.
Neither the US dietary guidelines for Americans nor Canada’s Food Guide recommend NSBs,
and instead recommend that water and not NSBs replace SSBs [11, 13]. The American Heart
Association (AHA) supports a narrow indication for NSBs, recommending that NSBs replace
SSBs only in adults who are habitual consumers of SSBs with an emphasis on water or
32
unsweetened alternative [18]. Similarly, diabetes associations in the UK, US, and Canada
support NSBs insofar as they are used to displace calories from sugars and SSBs [12, 15, 16].
The European Association for the Study of Diabetes (EASD) has not made any specific
recommendations about non-nutritive sweeteners (NNSs) or NSBs [17]. To update the
recommendations of the EASD, the Diabetes and Nutrition Study Group (DNSG) of the EASD
commissioned a systematic review and meta-analysis to summarize the available evidence from
RCTs of the effect of NSBs, the most important source of NSSs in the diet and a single food
matrix, on intermediate cardiometabolic outcomes. Strength of the evidence was assessed using
the Grading of Recommendations Assessment, Development, and Evaluation (GRADE)
approach [144].
Because of the importance of the comparator in drawing inferences about the effects of NSBs,
we chose to conduct network-meta-analyses as opposed to traditional pairwise meta-analyses to
assess the effect of NSBs through 3 prespecified substitutions: NSBs for SSBs (intended
substitution with caloric displacement), water for SSBs (“standard of care” substitution with
caloric displacement), and NSBs for water (matched substitution without caloric displacement).
4.3 Methods
4.3.1 Design
Our systematic review and network meta-analysis was conducted according to the Cochrane
Handbook for Systematic Reviews of Interventions [145] and reported according to the Preferred
Reporting Items for Systematic Reviews and Meta-Analyses Involving a Network Meta-analysis
(PRISMA-Network Meta-Analysis) [146] (Appendix Table 4.1). The protocol is registered at
ClinicalTrials.gov (identifier, NCT02879500; Results).
4.3.2 Data Sources and Searches
We searched MEDLINE, Embase, and the Cochrane Library from inception through 28 March
2019. The full search strategy is presented in Appendix Table 4.2. Briefly, we searched using
variations of the exposure terms (NSBs, SSBs), outcome terms (adiposity, glycemia, blood
lipids, blood pressure, non-alcoholic fatty liver disease (NAFLD) and uric acid) and study design
terms (randomized controlled trial, randomized, placebo). The search was limited to human
studies with no language restriction. Additionally, reference lists of selected studies and reviews
33
were searched, experts in the field were contacted and Google Scholar searches were conducted
to identify any additional articles.
4.3.3 Study Selection
Appendix Table 4.3 shows the summary of our study selection using the PICOTS framework
[146]. We included RCTs of at least 3-weeks duration that investigated the effect of one of three
beverages (NSBs, SSBs or water) on cardiometabolic risk factors compared to another among
the three beverages. Trials were included if the intervention arm assessed the effect of NSBs,
SSBs or water consumed alone or while on a weight-loss or nutrition education program in
adults. Trials were excluded if they had non-usual intake method (e.g. used NNS capsules that
bypass the oral taste receptors); had a mixed intervention; had a duration of <3 weeks; did not
use a comparator arm of NSBs, SSBs or water; included children, pregnant or breastfeeding
women; or did not provide suitable end-point data. We only studied NNSs in liquid (beverage
trials) as these allow for clear comparisons. Studies of NNSs added to foods, fortified beverages
or nutrient-dense beverages (e.g. milk, juice, etc.) were excluded due to the presence of other
macronutrients. When multiple publications existed for the same study, the article with the most
updated information was included. Published abstracts were not included.
4.3.4 Data Extraction
Two independent reviewers (NM and RZ) assessed the titles and abstracts of all identified
studies and reviewed and extracted relevant data from each report, including study design,
blinding, sample size, participant characteristics, follow-up duration, identification of NSBs,
SSBs and water as an intervention and/or comparator, beverage dosages, outcome data and
funding source. In those trials where the data were included in figures and not provided
numerically, we extracted data using Plot Digitizer V.2.6.8 (http://plotdigitizer.sourceforge.net/).
Additional information was requested from the authors when necessary. Disagreement were
resolved by consensus or where necessary by a third author (TAK).
4.3.5 Outcomes
Outcomes were intermediate measures of cardiometabolic risk as determined by the Clinical
Practice Guidelines Committee of the DNSG of the EASD. The primary outcome was body
weight. Secondary outcomes were other measures of adiposity (body mass index (BMI), % body
34
fat and waist circumference); glycemic control (HbA1c, fasting plasma glucose (FPG), 2h
plasma glucose (2h-PG) during a 75g oral glucose tolerance test (75g-OGTT), fasting plasma
insulin (FPI) and Homeostatic Model Assessment of Insulin Resistance (HOMA-IR)); blood
lipids ((low-density lipoprotein-cholesterol (LDL-C), Non-high-density lipoprotein-cholesterol
(Non-HDL-C), triglycerides (TGs), high-density lipoprotein-cholesterol (HDL-C) and total
cholesterol (TC)); blood pressure (systolic blood pressure (SBP) and diastolic blood pressure
(DBP)); measures of NAFLD (intra-hepatocellular lipid [IHCL], alanine transaminase (ALT)
and aspartate transaminase (AST)); and uric acid. Change-from baseline differences were
preferred over end differences; if not provided, these were calculated from the available data
using published formulas [145].
4.3.6 Risk of Bias Assessment
Risk of bias for each included trial was assessed by the two independent reviewers using the
Cochrane Risk of Bias tool [145]. Assessment was done across 5 domains of bias (sequence
generation, allocation concealment, blinding of participants and personnel, incomplete outcome
data and selective reporting). The risk of bias was assessed as either low (proper methods taken
to reduce bias), high (improper methods creating bias) or unclear (insufficient information
provided to determine the bias level). Difference between reviewers was resolved by consensus.
4.3.7 Data Synthesis
Network meta-analysis, based on a frequentist framework, was conducted using the "network"
suite of commands available in STATA version 15 (College Station, TX: StataCorp LP). The
network meta-analysis synthesized all of the available evidence (direct and indirect effects) and
quantified the pooled network effect of each intervention against every other intervention. We
reported our results as mean differences (MDs) and 95% confidence intervals (CIs). To display
the results for outcomes on the same plot, standardized mean differences (SMDs) were
calculated and pseudo 95% CIs, whereby the SMD 95% CIs were proportionally scaled to the
MD 95% CIs.
We performed a random-effects network meta-analysis for each outcome to compare the three
interventions simultaneously (NSBs, SSBs and water) in a single analysis by combining both
direct and indirect evidence across the selected network of studies. We used change from
baseline values from each study to calculate the mean differences between treatments for each
35
substitution (NSBs for SSBs, water for SSBs and NSBs for water), otherwise we used post-
intervention values. The network diagrams were generated to show the interactions among the
studies included in the network meta-analysis and to illustrate the available direct comparisons
between treatments [147].
Inconsistency was assessed in the direct, indirect, and network estimates. Wassessed interstudy
heterogeneity in the direct and indirect (each pairwise comparison arm) estimates using the
Cochran Q statistic with quantification by the I2 statistic, where an I2≥50%, P<0.10 was
considered an indication of substantial interstudy heterogeneity. We measured incoherence in
the network estimates using both local (loop-specific and side-splitting) and global (design-by-
treatment interaction model) approaches to evaluate the presence of incoherence. The loop-
specific approach looked at the inconsistency in each closed loop in the network [147] while the
side-splitting approach detected direct estimate comparisons that disagreed with the indirect
evidence from the entire network [148]. The design-by-treatment interaction model was applied
as a global approach to simultaneously check for inconsistency from all possible sources in the
network [149]. To explore sources of inconsistency, we conducted sensitivity analyses among
the direct and network comparisons, with the systematic removal of each trial and recalculation
of the pooled estimate. An influential trial was considered a study whose removal changed the
magnitude of the pooled effect by >10% significant. If ≥10 comparisons were available, then we
also conducted a priori subgroup analyses by age, study duration, type of design, disease status,
risk of bias and funding source.
Indirectness was assessed in the indirect comparisons by evaluation of intransitivity (age, study
length, sample size and percentage of males) across the pairwise comparisons comprising the
indirect estimates.
Publication bias was assessed if ≥10 trial comparisons were available, We used comparison-
adjusted funnel plots to assess funnel plot asymmetry [147]
4.3.8 Grading of the Evidence
We assessed the certainty of the evidence using the Grading of Recommendations Assessment,
Development, and Evaluation (GRADE) system [150] with an extension for network meta-
analyses [144] and other recent guidance [151-153] from the GRADE Working group. Evidence
36
was graded as high, moderate, low or very low certainty. Network estimates of RCTs and the
direct and indirect estimates that make-up these network estimates started at high certainty of
evidence and were downgraded by established criteria. Criteria to downgrade included serious
risk of bias (weight of RCTs in the direct estimates) show high risk of bias assessed by the
Cochrane Risk of Bias tool), inconsistency (substantial unexplained heterogeneity [I2≥50%,
P<0.10] in the direct estimates and incoherence in the network estimates by the loop-specific
[p<0.05], side-splitting [p<0.05] and the design-by-treatment interaction model [p<0.10]),
indirectness (presence of factors related to the participants, interventions, and study conditions
that limited the generalizability of the results among the direct estimates and intransitivity in the
indirect estimates), imprecision (the 95% CIs for effect estimates were wide and crossed
prespecified minimally important differences (MIDs) for benefit and/or harm in the direct and
network estimates) and publication bias (evidence of adjusted funnel plot asymmetry).
We used a step-wise approach to apply these downgrades [144], in which the certainty of
evidence from the direct and indirect estimates inform the network estimates. The first step
involved applying the downgrades to the direct estimates. The second step involved applying the
downgrades to the indirect estimates. As the indirect estimates were each comprised of two
pairwise estimates that shared a common comparator, we assessed the certainty of evidence as
the lowest certainty of evidence in the first-order loop. The final step involved grading the
network estimates. If only one of the direct or indirect estimates was available for a given
outcome, then the starting point for the certainty of evidence for the network estimates was based
on that estimate. If both the direct and indirect estimates were available, then the starting point
for the certainty of evidence was based on the estimate that contributed the most weight to the
network estimate. If incoherence and/or imprecision were detected in the network estimates, then
further downgrades were applied for serious inconsistency and serious imprecision, in the
absence of the same downgrades in the direct or indirect estimates used to inform the certainty of
evidence for the network estimate.
4.3.9 Patient Involvement
No patients were directly involved in the development of the research question, selection of the
outcome measures, design and implementation of the study, or interpretation of the results
37
4.4 Results
4.4.1 Search Results
Figure 4.1 shows the flow of the literature. We identified 4541 reports of which 9 met our
eligibility criteria. An additional 5 were found through manual searching. A total of 16 reports of
14 RCTs with 21 trial comparisons (n=1530 participants) were included that assessed the effect
of NSBs for SSBs, water for SSBs, or NSBs for water on body weight, other measures of
adiposity, and cardiometabolic risk [154-169]. Of these, 14 trials reported adiposity (14 for body
weight, 9 for BMI, 5 for % body fat and 6 for waist circumference) [154, 156-162, 164-169], 10
trials reported glycemic control and insulin resistance measurements (10 for FPG, 7 for FPI, and
5 for HOMA-IR, HbA1c and 2hPG [154, 156-162, 164, 168], 8 trials reported blood lipids (7 for
LDL-C and 8 for TGs, HDL-C and TC) [156-162, 164], 5 trials reported blood pressure (4 for
DBP) [156-158, 164, 168], 3 trials reported NAFLD outcomes (2 for IHCL fat, ALT and AST)
[156, 159, 163] and 2 trials reported uric acid concentrations [155, 156].
4.4.2 Available Data
Study and population variables including age, sample size, % females, population description
and intervention details were presented for all RCTs included. The available direct comparisons
between each beverage were represented as a network diagram for each outcome (Appendix
Figures 4.19-4.38).
4.4.3 Trial Characteristics
Table 4.1 shows the key characteristics of all included RCTs, for each beverage comparison, and
the effect on cardiometabolic risk.
The 16 reports of 14 RCTs consisted of 1530 participants (median 72; range 27 to 308) with a
median follow-up of 12 weeks (range 3 to 52). The median age of participants was 34 years
(range 23 to 48). Most participants were overweight or obese with a median BMI of 31 kg/m2
(range 22 to 36). There were more women than men (% men to women, 23:77). Only 8 trials
reported the type of NNS used in the NSBs. Six trials had NSBs with aspartame only. One trial
had NSBs with aspartame and acesulfame potassium (Ace-K) combined and one trial had 4
separate arms with saccharin, rebaudioside A (RebA), sucralose or aspartame as the NNS in the
38
NSBs. The other 6 trials did not report which NNSs were present in the NSBs. The median
beverage doses were 1000 mL/day (range 250 to 2000) for NSBs, 1000 mL/day (range 250 to
1750) for SSBs and 580 mL/day (range 250 to 2000) for water. Nine trials included participants
who were overweight and/or obese. Thirteen trials included healthy individuals, and one with
participants with T2DM. Twelve trials had a parallel design and two were cross-over. Most of
the trials were conducted in Europe (6) and North America (5). Two trials were conducted in the
Middle East and one in South America. Seven trials were funded by agencies (government, not-
for-profit health agency, or university sources), four by industry, and three had agency and
industry funding combined.
4.4.4 Risk of Bias
Appendix Figures 4.1 and 4.2 shows the individual Cochrane Risk of Bias tool assessments for
each of the included trials on body weight and other measures of adiposity, glycemic control,
established blood lipid targets, blood pressure, markers of NAFLD and uric acid. Eight trial
comparisons were rated as unclear risk of bias and 11 were rated as low risk of bias. None were
identified as having a high risk of bias.
4.4.5 Effect of Substitution of NSBs for SSBs
Figure 4.2 shows the network meta-analyses of the effect of the substitution of NSBs for SSBs
on body weight, other measures of adiposity, and cardiometabolic risk factors. The substitution
of NSBs for SSBs reduced the primary outcome body weight (MD= −1.11 kg; 95% CI −1.90, –
0.32 kg) and secondary outcomes BMI (MD= −0.32 kg/m2 ; 95% CI −0.58, –0.07 kg/m2, %
body fat (MD= −0.60%; 95% CI −1.03, –0.18%), TGs [(MD= −0.2 mmol/L; 95% CI −0.45, -
0.02 mmol/L)] and IHCL [(MD= −0.44 SMD; 95% CI −0.69, –0.19 SMD)]. There were no
significant differences for other outcomes.
4.4.6 Effect of Substitution of Water for SSBs
Figure 4.3 shows the network meta-analyses of the effect of the substitution of water for SSBs
on body weight, other measures of adiposity, and cardiometabolic risk factors. The substitution
of water for SSBs did not reduce the primary outcome body weight (MD= −0.09 kg; 95% CI
39
−1.35, 1.17 kg) but did reduce the secondary outcome uric acid (MD= -0.05 mmol/L; 95% CI -
0.08, -0.01 mmol/L). There were no differences in any of the other secondary outcomes.
4.4.7 Effect of Substitution of NSBs for Water
Figure 4.4 shows the network analyses of the effect of the substitution of NSBs for water on
body weight, other measures of adiposity, and cardiometabolic risk factors. The substitution of
NSBs for water did not affect the primary outcome body weight (MD= −1.02 kg; 95% CI −2.15,
0.11 kg) but did increase the secondary outcome HbA1c (MD=0.21%; 95% CI 0.02, 0.40%).
There were no significant differences for any other secondary outcomes.
4.4.8 Inconsistency
Appendix Tables 4.4 to 4.23 shows the side-splitting, and Appendix Table 4.24 shows the
loop-specific assessments of inconsistency (incoherence) in the network estimates. No
significant incoherence was observed by either approach across the 3 substitutions (p<0.05),
except for TGs in the loop-specific assessment (p=0.04). The design-by-treatment model did not
show any significant global inconsistency for any outcome (p<0.10).
4.4.9 Intransitivity (a domain of indirectness, in the indirect estimates)
Appendix Tables 4.3-4.6 shows the evaluation of intransitivity (a domain of indirectness)
among the indirect comparisons by comparing the distribution of the potential effect modifiers
across the available direct comparisons for age, study length, sample size and percentage males.
The assumption of transitivity was met for all indirect comparisons.
4.4.10 Sensitivity Analyses
The systematic removal of each trial among the direct comparisons did not modify the direction,
magnitude or significance of the effect estimates or the evidence of heterogeneity for any of the
outcomes among our 3 prespecified comparisons. We did not conduct sensitivity analyses in the
network as several of the RCTs contained multiple comparison arms in which the removal of one
RCT resulted in the removal of more than one comparison arm.
4.4.11 Subgroup Analyses
As <10 trials were available for any outcome, subgroup analyses was not undertaken.
40
4.4.12 Publication Bias
As <10 trials were available for any outcome, publication bias analyses were not undertaken.
4.4.13 Grading of the Evidence
Appendix Figures 4.7 to 4.18 shows the GRADE assessment for the direct, indirect, and
network meta-analysis. The certainty of the evidence for the substitution of NSBs for SSBs was
moderate for the improvement in the primary outcome body weight owing to a sole downgrade
for imprecision and from low to high for the secondary adiposity and cardiometabolic outcomes
owing to downgrades for inconsistency (waist circumference), indirectness (waist circumference,
HbA1c, 2h-PG, ALT and AST), and imprecision (BMI, waist circumference, HbA1c, 2h-PG,
FPI, LDL-C, Non-HDL-C, TGs, HDL-C, TC, SBP, DBP, IHCL, ALT, AST and uric acid). The
certainty of the evidence for the substitution of water for SSBs was low for the improvement in
the primary outcome body weight owing to a downgrade for inconsistency and imprecision, and
from low to high for the secondary adiposity and cardiometabolic outcomes owing to
downgrades for inconsistency (HbA1c and FPI), indirectness (waist circumference, HbA1c, 2h-
PG, FPI, HOMA-IR, IHCL, ALT, AST and uric acid), and imprecision (BMI, waist
circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-HDL-C, TGs, HDL-C, TC, SBP, DBP, IHCL,
ALT, AST and uric acid). The certainty of the evidence for the substitution of NSBs for water
was low for the improvement in the primary outcome body weight owing to a downgrade for
inconsistency and imprecision, and from low to high for the secondary adiposity and
cardiometabolic outcomes owing to downgrades for inconsistency (waist circumference, HbA1c,
FPI, HOMA-IR, Non-HDL-C and TGs), indirectness, (IHCL, ALT, AST and uric acid), and
imprecision (BMI, waist circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-HDL-C, TGs, TC,
SBP, DBP, IHCL, ALT, AST and uric acid).
4.5 Discussion
The present systematic review and network-meta-analysis of 14 RCTs of 21 trial comparisons
involving 1530 predominantly overweight or obese adult participants who were at risk for or had
diabetes assessed the effect of NSBs on body weight, other measures of adiposity, glycemic
control, blood lipids, blood pressure, markers of NAFLD and uric acid in 3 prespecified
substitutions: NSBs for SSBs (intended substitution with caloric displacement), water for SSBs
41
(“standard of care” substitution with caloric displacement), and NSBs for water (matched
substitution without caloric displacement). The substitution of NSBs for SSBs improved body
weight, BMI, body fat, TGs and IHCL, whereas the substitution of water for SSBs improved
only uric acid. There was no effect of the substitution of NSBs for water with the exception of a
small increase in HbA1c.
4.5.1 Findings in the Context of Existing Studies
We are not aware of any other systematic review and network-meta-analyses that simultaneously
assessed the effects of NSBs, SSBs, and water on intermediate outcomes of cardiometabolic risk.
While several systematic reviews and meta-analyses have assessed the effect of the substitution
of NNSs for caloric comparators in the form of sugary foods/beverages (usually SSBs) either
alone [20] or pooled with substitutions with matched non-caloric comparators such as water,
placebo, or calorie-matched weight loss diets on cardiometabolic risk factors [19-21, 24, 26],
none have quantitatively assessed the effect of NSBs relative to that of water for the intended use
of SSB reduction.
Our findings are in agreement with other systematic reviews and meta-analyses that provided
analyses which allowed for the interpretation of results by comparator. The finding that NSBs
reduce body weight, BMI and body fat in substitution for SSBs with caloric displacement and
have neutral effects on these outcomes in substitution for water without caloric displacement
have been shown by other systematic reviews and meta-analyses of RCTs. Rogers et al. [20]
showed that NNSs reduced body weight in analyses of the substitution of NNSs for a caloric
comparator (sugars in food and beverages) in RCTs involving participants who were
predominantly overweight or obese. Similarly, Toews et al. [21] showed a small improvement in
BMI in analyses of the substitution of NNSs for a caloric comparator (sucrose in food and
beverages) in RCTs involving participants who were predominantly healthy, while Miller et al.
[23] showed these reductions in body weight and BMI, as well as fat mass and waist
circumference in analyses of the substitution of NNSs for the caloric comparator sugars in food
and beverages in RCTs involving participants who were predominantly overweight or obese. A
recent analysis by Laviada-Molina et al. [25] further support our findings where they showed a
significant reduction in unified measurements of body weight and BMI (expressed as SMDs)
with NNSs in substitution for caloric comparators (sugars in foods and beverages) in
42
predominantly overweight or obese participants. In another broader analysis of the effect of
NNSs in substitution for a combination of caloric and non-caloric comparators by Toews et al.
[21] and another analysis that restricted the effect of NSBs in substitution for matched non-
caloric comparators (placebo, water, or weight loss diet) by Azad et al. [19] found no differences
in body weight with NNSs in RCTs involving participants who were predominantly overweight
or obese.
Overall, these findings are consistent with the mechanism that NSBs lead to weight loss through
contributing to a reduction in net energy intake. Although the displacement of calories using
water in substitution for SSBs did not result in the same significant reductions in body weight or
adiposity as seen with NSBs, there were no differences between NSBs and water across these
outcomes, suggesting that lack of differences were likely owing to imprecision.
Downstream improvements in other intermediate cardiometabolic outcomes would be expected
through displacement of calories from SSBs. In addition to weight gain [170], fructose-
containing sugars providing excess calories especially in beverage form have been shown to
increase TGs [59, 60], glucose [61], insulin [61], uric acid [62], and NAFLD markers [63].
Previous systematic reviews and meta-analyses of RCTs suggest that displacement of calories
from sugars using NSBs or water may improve these outcomes. Toews et al. [21] showed that
NNSs in substitution for caloric sugars (sucrose) in food and beverages reduced blood pressure,
being the effect greater when the substitution of NSBs for a non-caloric placebo was accounted
for. We showed that NSBs and water in substitution for SSBs, with the intended displacement of
calories, improve several intermediate cardiometabolic outcomes, including TGs and IHCL for
NSBs, and uric acid for water. Although we failed to show an effect on blood pressure (SBP,
DBP) or other aspects of the lipid profile (LDL-C, non-HDL-C, HDL-C, TC), uric acid, or other
markers of NAFLD (ALT, AST), the direction of the effects favored NSBs or water with 95%
CIs that were imprecise for benefit but excluded meaningful harm based on the prespecified
MIDs. There were also no differences between the two active treatments, NSBs or water, in these
outcomes.
There is a particular concern that NNSs may induce and promote glucose intolerance. Much of
this concern derives from a single intervention study that showed a decrease in 75g-OGTT
derived glucose tolerance after 6 days of saccharin in an NSB format among 7 healthy
43
individuals with subsequent transfer of the phenotype to germ free mice by fecal transplant. This
study, however, had many sources of bias: it saw the effect only with saccharin (an atypical
sweetener not used in NSBs with the exception of TaB); the NNS was given at an extreme dose
(at the maximum ADI of 5mg per kg of body weight/d, given in 3 divided doses per day,
equivalent to 120mg of saccharin or 4 TaB beverages [95]); there was no control group (making
it difficult to disentangle the effect of the trial protocol [6 consecutive days of 75g-OGTTs] from
that of the sweetener); and did a post hoc separation of participants into responders (n=4) and
non-responders (n=3) [81]. It also did not test the question of whether NSBs in the intended
substitution for SSBs would have the same effect. Given the weight loss seen with caloric
displacement, it would be expected that NSBs in substitution for SSBs would improve markers
of glycemic control. This suggestion is supported by the existing systematic reviews and meta-
analyses that allow for interpretation by comparator. Toews et al. [21] showed an improvement
in fasting plasma glucose without a change in fasting insulin or insulin sensitivity (HOMA-IR)
for the substitution of NNSs for caloric sugars (sucrose in food and beverages), whereas Azad et
al. [19] showed no effect on HbA1c or HOMA-IR of the substitution of NSBs for non-caloric
comparators (water or placebo) in overweight, obese or hypertensive participants. We did not see
evidence of benefit or harm across all glycemic control outcomes for caloric displacement with
the substation of either NSBs for SSBs or water for SSBs. We also did not see differences for the
substitution of NSBs for water with the exception of an increase in HbA1c. This difference,
however, is questionable due to imprecision detected in the estimate with two of the three
included direct trials showing improvements in HbA1c for both the water and NSBs arms [161,
162] and the third trial showing no effect of either the water or the NSBs arm, however the
direction of point estimate favored water [159]. More high-quality trials are needed to improve
these estimates.
4.5.2 Strengths and Limitations
This systematic review and network-meta-analysis has several strengths, the first being the use of
network-meta-analysis that unlike traditional pairwise meta-analyses allowed for the assessment
of multiple comparisons by leveraging direct and indirect comparisons with a common
comparator. By using a network meta-analysis it also allowed for the simultaneous assessment of
the 3 prespecified substitutions (NSBs for SSBs, water for SSBs and NSBs for water). Other
strengths include a comprehensive literature search inclusion of only RCTs which provide the
44
greatest protection against bias; no evidence of serious risk of bias among the included trials; and
use of the GRADE approach to assess the certainty of the estimates.
Several of our analyses presented limitations. First, evidence of serious inconsistency was
present on the secondary adiposity outcome of waist circumference in the substitution of NSBs
for SSBs; on the secondary cardiometabolic outcomes of HbA1c and FPI in the substitution of
water for SSBs; and on the primary outcome of body weight and several secondary adiposity
(waist circumference) and cardiometabolic (HbA1c, FPI, HOMA-IR, Non-HDL-C and TGs)
outcomes in the substitution of NSBs for water. We detected substantial heterogeneity (I2≥50%,
P<0.10) in all of these outcomes that could not be explained fully by sensitivity analyses within
the direct pairwise comparisons. We therefore downgraded the evidence for serious
inconsistency.
Second, there was evidence of serious indirectness in several of our analyses. As only one RCT
or less of direct comparisons was available for several secondary adiposity and cardiometabolic
outcomes in the analyses for the substitution of NSBs for SSBs (waist circumference, HbA1c,
2h-PG, ALT and AST), water for SSBs (waist circumference, HbA1c, 2h-PG, FPI, HOMA-IR,
IHCL, ALT, AST and uric acid) and NSBs for water (IHCL, ALT, AST and uric acid), we
downgraded the evidence in all cases for serious indirectness. The moderate median follow-up of
12-weeks was considered another potential source of indirectness across our analyses. Although
there is some uncertainty as to whether the benefits and lack of harm seen with NSBs extend
beyond our 12-week medium follow-up, we felt that any harms would have manifest within 12-
weeks when participants tend to be most adherent. Our analyses also included RCTs with up to 1
year of follow-up that showed no evidence of harm and even benefit. The longest RCT by Peters
et al. [164] of a 12-month intervention of NSBs substituted for water in the context of a 12-
month behavioral weight loss treatment program among participants with overweight and obesity
showed superiority of NSBs for weight loss and weight maintenance. Other large RCTs in
children (not captured in our analyses that were restricted to adults) offer further evidence of
durable benefit. The largest, longest, and highest quality RCT to date which investigated the
double-blinded substitution of NSBs for SSBs among 641 normal-weight children (ages, 4-11
years) over 18 months showed reductions in weight gain and body fat gain [171]. Adherence to
the NSBs in this trial was also confirmed by a urinary biomarker (urinary sucralose). Based on
45
these findings, we chose not to downgrade the evidence for serious indirectness owing to a lack
of long-term follow up and chose instead to make all of our conclusions specific to the moderate
term.
Third, there was evidence of serious imprecision in several of our pooled estimates. The 95%
CIs for the primary outcome of body weight and several secondary adiposity and
cardiometabolic outcomes in the analyses for the substitution of NSBs for SSBs (BMI, waist
circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-HDL-C, TGs, HDL-C, TC, SBP, DBP, IHCL,
ALT, AST and uric acid); water for SSBs (body weight, BMI, waist circumference, HbA1c, 2h-
PG, FPI, LDL-C, Non-HDL-C, TGs, HDL-C, TC, SBP, DBP, IHCL, ALT, AST and uric acid);
and NSBs for water (body weight, BMI, waist circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-
HDL-C, TGs, TC, SBP, DBP, IHCL, ALT, AST and uric acid) crossed our prespecified MIDs.
Accordingly, we downgraded the evidence in all cases for serious imprecision.
Finally, we were also unable to assess for publication bias or conduct subgroup analyses as there
were less than 10 direct RCT comparisons per outcome for each of the 3 prespecified
substitutions. We were also unable to conduct sensitivity analyses in the network as several of
the RCTs contained multiple comparison arms.
Balancing these strengths and limitations, the certainty of the available evidence was assessed as
moderate for the primary outcome of body weight in the substitution of NSBs for SSBs, and low
for the primary outcome of body weight in the substitution of water for SSBs and in the
substitution of NSBs for water. For all other secondary adiposity and cardiometabolic outcomes
the certainty of the evidence ranged from low to high for the 3 prespecified beverage
substitutions. On average, a rating of moderate was given for the substitution of NSBs for SSBs,
and low for the substitution of water for SSBs and NSBs for water. Downgrades were made for
evidence of serious indirectness and serious imprecision in the substitution of NSBs for SSBs
and the substitution of water for SSBs, while downgrades were made for evidence of serious
inconsistency, serious indirectness and serious imprecision in the substitution of NSBs for water.
46
4.5.3 Implications
The findings of this study are highly relevant for informing guidance on the role of NSBs as a
caloric displacement for SSBs in the context of excess sugar intake which has been implicated in
the dual epidemics of obesity and diabetes [172, 173]. Particular focus has been directed to SSBs
as they are the most important source of added/free sugars in several countries worldwide [49,
69, 174] and their overconsumption has been associated with weight gain, diabetes, and
downstream complications of hypertension and coronary heart disease (CHD) [6-9]. These
conditions not only place an enormous impact on the healthcare system but can significantly
reduce an individual’s quality of life [175]. Although the findings from this network meta-
analysis are limited to NSBs, they are of relevance to NNSs as they are added to several foods
and beverages by consumers and product developers in an attempt to reduce calories [85]. Most
importantly, of all product categories containing NNSs as a single food matrix, NSBs are
globally the most consumed source [18, 85].
Water is considered the “standard of care” substitution for SSBs by authoritative bodies [11-13,
15-18]. Yet, many of these health organizations warn against replacing SSBs with NSBs out of
concern that they contribute to an increased risk of obesity and diabetes. A concern that is
echoed in several media headlines[176-184]. Contrary to these concerns, the available evidence
shows that when NSBs are substituted for SSBs, benefits are seen with no evidence of harm on
body weight and cardiometabolic outcomes over the moderate term. For heavy SSB consumers
who are unable to switch to water, NSBs are a viable alternative as they provide the intended
benefit of calorie displacement. This finding is particularly important considering that
individuals struggle to maintain a substantial change to dietary behavior, especially if the change
is associated with cravings or hunger [185]. NSBs, due to their sweet taste, offer an alternative
that may provide the least cognitive bias towards adherence.
4.5.4 Conclusions
When used as the intended substitution for SSBs, NSBs improves body weight and several
markers of cardiometabolic risk factors including adiposity, IHCL and TGs without evidence of
harm. NSBs were shown to be similar to water, the “standard of care” substitution for SSBs, in
their effect on body weight and cardiometabolic risk factors. The available evidence supports the
use of NSBs as an alternative replacement strategy for SSBs in overweight/obese adults at risk
47
for or with diabetes over the moderate term. Due to uncertainty in the estimate, across the 3
prespecified substitutions, further research is needed to improve our estimates. There is a need
for more high quality RCTs of longer duration comparing these beverages to clarify if the
substitution of NSBs for SSBs results in further risk reduction. These RCTs should focus on
quantifying the types of NNS blends in NSBs, while comparing their effectiveness as a caloric
displacement for SSBs against the “standard of care” matched substitution without caloric
displacement of water, to more accurately assess their long-term effects. In this regard, we await
the results of the STOP Sugars NOW trial (ClinicalTrials.gov, NCT03543644); the Effect of
Non-nutritive Sweeteners of High Sugar Sweetened Beverages on Metabolic Health and Gut
Microbiome trial (ClinicalTrials.gov, NCT03259685); the Study of Drinks With Artificial
Sweeteners in People With Type 2 Diabetes trial (ClinicalTrials.gov, NCT03944616); and the
SWITCH trial (ClinicalTrials.gov, NCT02591134). Furthermore, contributions from research
using a range of study designs will likely be required to validate if the intended benefits and lack
of harm with NSBs for SSBs are durable for cardiometabolic risk outcomes. Additional research
should also assess the health impact of our 3 prespecified substitutions among groups not
included, or inadequately captured in this evidence syntheses, such as younger individuals and
higher risk ethnic groups (ex. Aboriginal peoples), to confirm certainty of directionality.
4.5.5 Funding Statement
The Diabetes and Nutrition Study Group (DNSG) of the European Association of the Study of
Diabetes (EASD) commissioned this systematic review and meta-analysis and provided funding
and logistical support for meetings as part of the development of the EASD Clinical Practice
Guidelines for Nutrition Therapy. This work was also supported by the Canadian Institutes of
Health Research (funding reference number, 129920) through the Canada-wide Human Nutrition
Trialists’ Network (NTN). The Diet, Digestive tract, and Disease (3-D) Centre, funded through
the Canada Foundation for Innovation (CFI) and the Ministry of Research and Innovation’s
Ontario Research Fund (ORF), provided the infrastructure for the conduct of this project. NDM
was supported by a CIHR-Masters Award and a Research Training Centre scholarship through
St. Michel’s Hospital. JLS was funded by a PSI Graham Farquharson Knowledge Translation
Fellowship, Diabetes Canada Clinician Scientist award, CIHR INMD/CNS New Investigator
Partnership Prize, and Banting & Best Diabetes Centre Sun Life Financial New Investigator
Award. JSS, gratefully acknowledges the financial support by ICREA under the ICREA
48
Academia program. EMC holds the Lawson Family Chair in Microbiome Nutrition Research at
the University of Toronto. With the exception of the Clinical Practice Guidelines Committee of
the DNSG of the EASD, none of the sponsors had a role in any aspect of the present study,
including design and conduct of the study; collection, management, analysis, and interpretation
of the data; and preparation, review, approval of the manuscript or decision to publish.
4.5.6 Conflict of Interest
NDM was a full-time employee with Loblaws Companies Limited from November 2011 to
December 2017 and in 2019 received the CIHR-Masters Award. TAK has received research
support from the CIHR and an unrestricted travel donation from Bee Maid Honey Ltd. He has
also spoken as an invited speaker at a Calorie Control Council annual general meeting for which
he received an honorarium. LW and RZ report no relevant competing interests. LC is a Mitacs-
Elevate post-doctoral fellow jointly funded by the Government of Canada and the Canadian
Sugar Institute. FAY reports no relevant competing interests. EMC has received research support
from Lallemand Health Solutions and Ocean Spray; and has received consultant fees or speaker
or travel support from Danone, Nestlé and Lallemand Health Solutions. VSM has been on a pro
bono retainer for expert support for the Center for Science in the Public Interest in litigation
related to sugar sweetened beverages and has served as a consultant for the City of San Francisco
for a case related to health warning labels on soda. JOH receives research funding from NIH and
from the National Cattlemen's Beef Association. He is a member of the Scientific Advisory
Committee for General Mills, McCormick Science Institute, and Milk Producers Educational
Program. He is a member of the Board of Trustees for the International Life Sciences Institute
(ILSI). He has equity in Gelesis and Shakabuku, LLC. LAL reports no relevant competing
interests. AA reports no relevant competing interests. PBJ is running a nonprofit public funded
research project entitle: Innosweet- Integrated perception, psychology, and physiology for
maintaining sweetness perception via sugar replacement and reduction for value added healthy
beverage applications (6150-00037A). He is a honorary member of the EUSTAS and was until
2018 Board Member of the Diabetes and Nutrition Study Group (DNSG) of the EASD. DR is
director of Vuk Vrhovac University Clinic for Diabetes, Endocrinology and Metabolic Diseases
at Merkur University Hospital, Zagreb, Croatia. He is the president of Croatian Society for
Diabetes and Metabolic Disorders of Croatian Medical Association. He serves as a Executive
Committee member of Croatian Endocrine Society, Croatian Sociaty for Obesity and Croatian
49
Society for Endocrine Oncology. He was a board member and secretary of IDF Europe and
currently he is the chair of IDF YLD Programme. He has served as a Executive Committee
member of Diabetes and Nutrition Study Group of EASD and currently he serves as a a
Executive Committee member of Diabetes and Cardiovascular Disease Study Group of EASD.
He has served as principal investigator or co-investigator in clinical trials of AstraZeneca, Eli
Lilly, MSD, Novo Nordisk, Sanofi Aventis, Solvay and Trophos. He has received travel support,
speaker fees and honoraria from advisory board engagements and/or consulting fees from
Abbott, Amgen, AstraZeneca, Bayer, Belupo, Boehringer Ingelheim, Eli Lilly, Lifescan –
Johnson & Johnson, International Sweeteners Association, Krka, Medtronic, Mediligo, Novartis,
Novo Nordisk, MSD, Merck Sharp & Dohme, Pfizer, Pliva, Roche, Salvus, Sandoz, Sanofi
Aventis and Takeda. HK is Director of Clinical Research at the Physicians Committee for
Responsible Medicine, a nonprofit organization providing nutrition education and research. JSS
reports serving on the board of and receiving grant support through his institution from the
International Nut and Dried Fruit Council, and Eroski Foundation. Reports serving in the
Executive Committee of the Instituto Danone Spain. Has received research support from the
Instituto de Salud Carlos III, Spain; Ministerio de Educación y Ciencia, Spain; Departament de
Salut Pública de la Generalitat de Catalunya, Catalonia, Spain; European Commission and USA
National Institutes of Health. Has received research support from California Walnut
Commission, Sacramento CA, USA; Almond Board of California, USA; Patrimonio Comunal
Olivarero, Spain; La Morella Nuts, Spain; and Borges S.A., Spain. Reports receiving consulting
fees or travel expenses from Danone; California Walnut Commission, Eroski Foundation,
Instituto Danone - Spain, Nuts for Life, Australian Nut Industry Council, Nestlé, Abbot
Laboratories, and Font Vella Lanjarón. He is on the Clinical Practice Guidelines Expert
Committee of the European Association for the study of Diabetes (EASD), and served in the
Scientific Committee of the Spanisch Food and Safety Agency, and the Spanish Federation of
the Scientific Societies of Food, Nutrition and Dietetics. He is a member of the International
Carbohydrate Quality Consortium (ICQC), and Executive Board Member of the Diabetes and
Nutrition Study Group (DNSG) of the EASD. CWCK has received grants or research support
from the Advanced Food Materials Network, Agriculture and Agri-Foods Canada (AAFC),
Almond Board of California, American Peanut Council, Barilla, Canadian Institutes of Health
Research (CIHR), Canola Council of Canada, International Nut and Dried Fruit Council,
50
International Tree Nut Council Research and Education Foundation, Loblaw Brands Ltd, Pulse
Canada and Unilever. He has received in-kind research support from the Almond Board of
California, American Peanut Council, Barilla, California Walnut Commission, Kellogg Canada,
Loblaw Companies, Quaker (PepsiCo), Primo, Unico, Unilever, WhiteWave Foods/Danone. He
has received travel support and/or honoraria from the American Peanut Council, Barilla,
California Walnut Commission, Canola Council of Canada, General Mills, International Nut and
Dried Fruit Council, International Pasta Organization, Loblaw Brands Ltd, Nutrition Foundation
of Italy, Oldways Preservation Trust, Paramount Farms, Peanut Institute, Pulse Canada, Sun-
Maid, Tate & Lyle, Unilever and White Wave Foods/Danone. He has served on the scientific
advisory board for the International Tree Nut Council, International Pasta Organization,
McCormick Science Institute and Oldways Preservation Trust. He is a member of the
International Carbohydrate Quality Consortium (ICQC), Executive Board Member of the
Diabetes and Nutrition Study Group (DNSG) of the European Association for the Study of
Diabetes (EASD), is on the Clinical Practice Guidelines Expert Committee for Nutrition Therapy
of the EASD and is a Director of the Toronto 3D Knowledge Synthesis and Clinical Trials
foundation. JLS has received research support from the Canadian Foundation for Innovation,
Ontario Research Fund, Province of Ontario Ministry of Research and Innovation and Science,
Canadian Institutes of health Research (CIHR), Diabetes Canada, PSI Foundation, Banting and
Best Diabetes Centre (BBDC), American Society for Nutrition (ASN), INC International Nut
and Dried Fruit Council Foundation, National Dried Fruit Trade Association, The Tate and Lyle
Nutritional Research Fund at the University of Toronto, The Glycemic Control and
Cardiovascular Disease in Type 2 Diabetes Fund at the University of Toronto (a fund established
by the Alberta Pulse Growers), and the Nutrition Trialists Fund at the University of Toronto (a
fund established by the Calorie Control Council). He has received in-kind food donations to
support a randomized controlled trial from the Almond Board of California, California Walnut
Commission, American Peanut Council, Barilla, Unilever, Unico/Primo, Loblaw Companies,
Quaker (Pepsico), Kellogg Canada, and WhiteWave Foods. He has received travel support,
speaker fees and/or honoraria from Diabetes Canada, Mott’s LLP, Dairy Farmers of Canada,
FoodMinds LLC, PepsiCo, The Ginger Network LLC, International Sweeteners Association,
Nestlé, Pulse Canada, Canadian Society for Endocrinology and Metabolism (CSEM), GI
Foundation, Abbott, Biofortis, ASN, Health Sciences North, INC Nutrition Research &
51
Education Foundation, and Physicians Committee for Responsible Medicine. He has or has had
ad hoc consulting arrangements with Perkins Coie LLP, Tate & Lyle, and Wirtschaftliche
Vereinigung Zucker e.V. He is a member of the European Fruit Juice Association Scientific
Expert Panel. He is on the Clinical Practice Guidelines Expert Committees of Diabetes Canada,
European Association for the study of Diabetes (EASD), Canadian Cardiovascular Society
(CCS), and Obesity Canada. He serves as an unpaid scientific advisor for the Food, Nutrition,
and Safety Program (FNSP) and the Technical Committee on Carbohydrates of the International
Life Science Institute (ILSI) North America. He is a member of the International Carbohydrate
Quality Consortium (ICQC), Executive Board Member of the Diabetes and Nutrition Study
Group (DNSG) of the EASD, and Director of the Toronto 3D Knowledge Synthesis and Clinical
Trials foundation. His wife is a former employee of Unilever Canada.
4.5.7 Ethics Approval
Not required
4.5.8 Data Sharing
No additional data are available.
The lead author affirms that this manuscript is an honest, accurate, and transparent account of the
study being reported; that no important aspects of the study have been omitted; and that any
discrepancies from the study as planned (and, if relevant, registered) have been explained.
This is an Open Access article distributed in accordance with the Creative Commons Attribution
Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt,
build upon this work non-commercially, and license their derivative works on different terms,
provided the original work is properly cited and the use is non-commercial. See:
http://creativecommons.org/licenses/by-nc/4.0/.
52
1Table 4.1: Trial Characteristics
Study Subjects Population Age
Mean
(SD)
Sex No.
(%) M:F
NNS Beverage Dose (mL/d)
Setting Design Duration
(Weeks)
Funding
Source
NSB Water SSB
Bonnet et
al. 2018
50 OW/NW, otherwise healthy,
non/low-NNS consumers
31.1
(10.3)
22:28 Asp/
Ace-K
660 660 NA France C 12 A/I
Bruun et
al. 2015
35 OW/OB, otherwise healthy 39
(1.1)
(44:56) Asp 1000 1000 1000 Denmark P 26 A/I
Campos
et al.
2015
27 OW/OB, otherwise healthy,
regular SSB consumers
NR 14:21 NR 1300 NA 1300 Switzerland P 12 A
Engel et
al. 2017
45
OW/OB, otherwise healthy 38.6
(7.6)
(40:60) Asp 1000 1000 1000 Denmark P 26 A/I
Hernánd
ez-
Cordero
et al.
2014
240
OW/OB, otherwise healthy,
regular SSB consumers
33.3
(6.7)
14:13 NR NA 250+ 250+ Mexico P 39 I
Higgins
et al.
2018
93 NW, healthy, non/low-NNS
consumers
22.9
(1.0)
(52:48) Asp 500 500 NA USA P 12 I
Higgins
& Mattes
2019
154
OW/OB, otherwise healthy,
non/low-NNS consumers
27.3
(9.6)
16:29 Sac/ Asp/
RebA/
Suc
1250-
1750
NA 1250-
1750
USA P 12 A
Madjd et
al. 2015
62
OW/OB, otherwise healthy,
regular NSB consumers
32
(6.9)
(36:64) NR 250+ 250+ NA Iran P 24 A
Madjd et
al. 2017
81
OB, T2DM (only on
metformin to control
diabetes), regular NSB
consumers
34.8
(7.2)
0:240 NR 250+ 250+ NA Iran P 24 A
Maersk
et al.
2012
35
OW/OB, otherwise healthy 39
(26)
(0:100) Asp 1000 1000 1000 Denmark P 26 A/I
Peters et
al. 2016
308 OW/OB, otherwise healthy,
weight stable, regular NSB
consumers
47.8
(10.5)
43:50 NR 710 710 NA USA P 52 I
53
A, agency; Asp, aspartame; Ace-K, acesulfame potassium; F, females; I, industry; M, males; NA, not applicable; No., number; NNS, non-nutritive sweetener;
NSB, non-nutritive sweetened beverage; NW, normal weight; NR, not reported; OB, obese; OW, overweight; RebA, rebaudioside A; Sac, saccharin; SD, standard
deviation; Suc, sucralose; SSB, sugar-sweetened beverage; T2DM, type 2 diabetes mellitus.
*Secondary analyses to Maersk et al. 2012. As more outcomes were reported in the Engel et al. 2017 analysis, data from that trial was used for the majority of
outcomes, except for uric acid (Bruun et al.) and IHCL (Maersk et al).
Reid et
al. 2007
133
NW, weight watchers and
non-weight watchers
31.8
(9.1)
(46:54) Asp 1000 NA 1000 England P 4 A
Reid et
al. 2010
53
OW, otherwise healthy 33.7
(9.9)
67:87 Asp 1000 NA 1000 Scotland P 4 A
Reid et
al. 2014
41
OB, otherwise healthy 35
(9.1)
(44:56) Asp 1000 NA 1000 Scotland P 4 A/I
Tate et
al. 2012
213 OW/OB, otherwise healthy,
regular SSB consumers
42
(10.7)
0:62 NR 1420-
2000
1420-
2000
NA USA P 26 I
Tordoff
et al.
1990
30
NW, healthy 25.6
(5.3)
(0:100) Asp 1135 NA 1135 USA C 3 A
54
All reports identified through database searching: 4541
(through 28 March, 2019)
MEDLINE: 1052
EMBASE: 1475
The Cochrane Library: 2014
Manual Searches: 5
Total reports after duplicates removed:
3142
Reports excluded by title and abstract: 2778
Non-Human: 137
Cross-Sectional: 19
Case Study: 3
Case-Control: 8
Prospective Cohort: 13
Observational: 81
Commentaries/Editorials/Letters: 5
Review Papers/Conference
Highlights/Guidelines/Protocol Papers: 301
Meta-Analysis: 44
Systematic Review: 48
Drug Study: 282
Inadequate Intervention: 1774
Unsuitable End Points: 63
Reports assessed for full review:
364
Reports excluded by full review: 355
Cross-Sectional: 8
Prospective Cohort: 1
Observational: 11
Review Papers/Conference
Highlights/Guidelines/Protocol Papers: 60
Systematic Review: 3
Drug Study: 2
Inadequate Intervention: 174
Inadequate Comparator: 22
Unsuitable End Points: 41
RCT Acute (<3 weeks): 31
Irretrievable: 2
Total reports included: 16 (14 RCTs)
Trials comparing NSBs vs water: 7
Trials comparing NSBs vs SSBs: 7
Trials comparing SSBs vs water: 2
Adiposity: 14
Glycemic control: 10
Blood lipids: 8
Blood pressure: 5
NAFLD: 3
Uric acid: 2
1Figure 4.1 Literature Search for RCTs of NSBs reporting on Adiposity, Glycemic Control, Blood Lipids,
Blood Pressure, NAFLD and Uric Acid.
55
2Figure 4.2 Network Results: Substitution of NSBs for SSBs
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for SSBs on
outcomes of body weight, other measures of adiposity, glycemic control, blood lipids, blood pressure, non-alcoholic
fatty liver disease (NAFLD) and uric acid. Data was pooled using network random effects models and expressed as
mean differences (MD) and 95% confidence intervals (CIs). To display the results for outcomes on the same plot,
standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the MDs) were
calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. 2h-PG, two-hour post
prandial glucose; ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass index; DBP, Diastolic
Blood Pressure; FPG; Fasting Plasma Glucose; FPI, Fasting Plasma Insulin; HbA1c; hemoglobin A1c; HDL-C,
high-density lipoprotein-cholesterol; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; IHCL,
intra-hepatocellular lipid; LDL-C, low-density lipoprotein-cholesterol; Non-HDL-C, non-high-density lipoprotein-
cholesterol; NSBs, non-nutritive sweetened beverages; SBP, Systolic Blood Pressure; SSBs, sugar-sweetened
beverages; TC, total cholesterol; TGs, triglycerides; WC, Waist Circumference.
56
3Figure 4.3: Network Results: Substitution of Water for SSBs
Network analyses of randomized controlled trials investigating the effect of the substitution of water for SSBs on
outcomes of body weight, other measures of adiposity, glycemic control, blood lipids, blood pressure, non-alcoholic
fatty liver disease (NAFLD) and uric acid. Data was pooled using network random effects models and expressed as
mean differences (MD) and 95% confidence intervals (CIs). To display the results for outcomes on the same plot,
standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the MDs) were
calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. 2h-PG, two-hour post
prandial glucose; ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass index; DBP, Diastolic
Blood Pressure; FPG; Fasting Plasma Glucose; FPI, Fasting Plasma Insulin; HbA1c; hemoglobin A1c; HDL-C,
high-density lipoprotein-cholesterol; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; IHCL,
intra-hepatocellular lipid; LDL-C, low-density lipoprotein-cholesterol; Non-HDL-C, non-high-density lipoprotein-
cholesterol; SBP, Systolic Blood Pressure; SSBs, sugar-sweetened beverages; TC, total cholesterol; TGs,
triglycerides; WC, Waist Circumference.
57
4Figure 4.4 Network Results: Substitution of NSBs for Water
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for water
on outcomes of body weight, other measures of adiposity, glycemic control, blood lipids, blood pressure, non-
alcoholic fatty liver disease (NAFLD) and uric acid. Data was pooled using network random effects models
and expressed as mean differences (MD) and 95% confidence intervals (CIs). To display the results for
outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally
scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by
the black horizontal line. 2h-PG, two-hour post prandial glucose; ALT, alanine transaminase; AST, aspartate
transaminase; BMI, body mass index; DBP, Diastolic Blood Pressure; FPG; Fasting Plasma Glucose; FPI,
Fasting Plasma Insulin; HbA1c; hemoglobin A1c; HDL-C, high-density lipoprotein-cholesterol; HOMA-IR,
Homeostatic Model Assessment of Insulin Resistance; IHCL, intra-hepatocellular lipid; LDL-C, low-density
lipoprotein-cholesterol; Non-HDL-C, non-high-density lipoprotein-cholesterol; NSBs, non-nutritive sweetened
beverages; SBP, Systolic Blood Pressure; TC, total cholesterol; TGs, triglycerides; WC, Waist Circumference.
58
2Appendix Table 4.1: PRISMA-NMAa Checklist
Section/topic Item
# b* Checklist item† Reported
on page #
TITLE
Title 1 Identify the report as a systematic review incorporating a network
meta-analysis (or related form of meta-analysis).
29
ABSTRACT
Structured
summary
2 Provide a structured summary including, as applicable:
Background: main objectives
Methods: data sources; study eligibility criteria, participants, and
interventions; study appraisal; and synthesis methods, such as network
meta-analysis.
Results: number of studies and participants identified; summary
estimates with corresponding confidence/credible intervals; treatment
rankings may also be discussed. Authors may choose to summarize
pairwise comparisons against a chosen treatment included in their
analyses for brevity.
Discussion/Conclusions: limitations; conclusions and implications of
findings.
Other: primary source of funding; systematic review registration
number with registry name.
30
INTRODUCTION
Rationale 3 Describe the rationale for the review in the context of what is already
known, including mention of why a network meta-analysis has been
conducted.
31-32
Objectives 4 Provide an explicit statement of questions being addressed, with
reference to participants, interventions, comparisons, outcomes, and
study design (PICOS).
31-32
METHODS
Protocol and
registration
5 Indicate whether a review protocol exists and if and where it can be
accessed (e.g., Web address); and, if available, provide registration
information, including registration number.
32
Eligibility
criteria
6 Specify study characteristics (e.g., PICOS, length of follow-up) and
report characteristics (e.g., years considered, language, publication
status) used as criteria for eligibility, giving rationale. Clearly describe
eligible treatments included in the treatment networt and note whether
any have been clustered or merged into the same node (with
justification).
32-33, 63
Information
sources
7 Describe all information sources (e.g., databases with dates of
coverage, contact with study authors to identify additional studies) in
the search and date last searched.
34
Search 8 Present full electronic search strategy for at least one database,
including any limits used, such that it could be repeated.
62
59
Study selection 9 State the process for selecting studies (i.e., screening, eligibility,
included in systematic review, and, if applicable, included in the meta-
analysis).
33, 63
Data collection
process
10 Describe method of data extraction from reports (e.g., piloted forms,
independently, in duplicate) and any processes for obtaining and
confirming data from investigators.
33-34
Data items 11 List and define all variables for which data were sought (e.g., PICOS,
funding sources) and any assumptions and simplifications made.
33
Geometry of
the network
S1 Describe methods used to explore the geometry of the treatment
network under study and potential biases related to it. This should
include how the evidence base has been graphically summarized for
presentation, and what characteristics were compiled and used to
describe the evidence base to readers.
34-35
Risk of bias in
individual
studies
12 Describe methods used for assessing risk of bias of individual studies
(including specification of whether this was done at the study or
outcome level), and how this information is to be used in any data
synthesis.
34
Summary
measures
13 State the principal summary measures (e.g., risk ratio, difference in
means). Also describe the use of additional summary measures
assessed, such as treatment rankings and surface under the cumulative
ranking curve (SUCRA) values, as well as modified approaches used to
present summary findings from meta-analyses.
34
Planned methods
of results
14 Describe the methods of handling data and combining results of studies
for each network meta-analysis. This should include, but not be limited
to:
• Handling of multi-arm trials;
• Selection of variance structure;
• Selection of prior distributions in Bayesian analyses; and
• Assessment of model fit.
34-35
Assessment of
inconsistency
S2 Describe the statistical methods used to evaluate the agreement of
direct and indirect evidence in the treatment network(s) studied.
Describe efforts taken to address its presence when found.
35
Risk of bias
across studies
15 Specify any assessment of risk of bias that may affect the cumulative
evidence (e.g., publication bias, selective reporting within studies).
34
Additional
analyses
16 Describe methods of additional analyses if done, indicating which were
pre-specified. This may include, but not be limited to, the following:
• Sensitivity or subgroup analyses;
• Meta-regression analyses;
• Alternative formulations of the treatment network; and
Use of alternative prior distributions for Bayesian analyses (if
applicable).
35-36
60
RESULTS
Study selection 17 Give numbers of studies screened, assessed for eligibility, and included
in the review, with reasons for exclusions at each stage, ideally with a
flow diagram.
37-38, 52-
53
Presentation of
network
structure
S3 Provide a network graph of the included studies to enable visualization
of the geometry of the treatment network.
88-97
Summary of
network
geometry
S4 Provide a brief overview of characteristics of the treatment network.
This may include commentary on the abundance of trials and
randomized patients for the different interventions and pairwise
comparisons in the network, gaps of evidence in the treatment network,
and potential biases reflected by the network structure.
40, 68-71
Study
characteristics
18 For each study, present characteristics for which data were extracted
(e.g., study size, PICOS, follow-up period) and provide the citations.
52-53
Risk of bias
within studies
19 Present data on risk of bias of each study and, if available, any outcome
level assessment.
38
Results of
individual
studies
20 For all outcomes considered (benefits or harms), present, for each study:
(a) simple summary data for each intervention group (b) effect estimates
and confidence intervals, ideally with a forest plot. Modified approaches
may be needed to deal with information from larger networks.
38-39
Synthesis of
results
21 Present results of each meta-analysis done, including
confidence/credible intervals. In larger networks, authors may focus on
comparisons versus a particular comparator (e.g. placebo or standard
care), with full findings presented in an appendix. League tables and
forest plots may be considered to summarize pairwise comparisons. If
additional summary measures were explored (such as treatment
rankings), these should also be presented.
55-57, 76-
87
Exploration for
inconsistency
S5 Describe results from investigations of inconsistency. This may include
such information as measures of model fit to compare consistency and
inconsistency models, P values from statistical tests, or summary of
inconsistency estimates from different parts of the treatment network.
39, 64-69
Risk of bias
across studies
22 Present results of any assessment of risk of bias across studies for the
evidence base being studied.
70-71
Results of
additional
analyses
23 Give results of additional analyses, if done (e.g., sensitivity or subgroup
analyses, meta-regression analyses, alternative network geometries
studied, alternative choice of prior distributions for Bayesian analyses,
and so forth).
39
DISCUSSION
Summary of
evidence
24 Summarize the main findings including the strength of evidence for each
main outcome; consider their relevance to key groups (e.g., healthcare
providers, users, and policy makers).
40-41
Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at
review-level (e.g., incomplete retrieval of identified research, reporting
bias). Comment on the validity of the assumptions, such as transitivity
and consistency. Comment on any concerns regarding network geometry
(e.g., avoidance of certain comparisons).
44-45
Conclusions 26 Provide a general interpretation of the results in the context of other
evidence, and implications for future research.
41-43, 46-
47
FUNDING
61
Funding 27 Describe sources of funding for the systematic review and other support
(e.g., supply of data); role of funders for the systematic review. This
should also include information regarding whether funding has been
received from manufacturers of treatments in the network and/or
whether some of the authors are content experts with professional
conflicts of interest that could affect use of treatments in the network.
47
a Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for
Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097.
doi:10.1371/journal.pmed1000097. For more information, visit: www.prisma-statement.org. b Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, Ioannidis JP, Straus S, Thorlund K,
Jansen JP, Mulrow C, Catalá-López F, Gøtzsche PC, Dickersin K, Boutron I, Altman DG, Moher D. The PRISMA
extension statement for reporting of systematic reviews incorporating network meta-analyses of health care
interventions: checklist and explanations. Ann Intern Med. 2015 Jun 2;162(11):777-84.
* Boldface indicates new items to this checklist according to PRISMA extension for reporting network meta-
analyses. † Text in italics indicates wording specific to reporting of network meta-analyses that has been added to guidance
from the PRISMA statement.
PICOS = population, intervention, comparators, outcomes, study design.
62
3Appendix Table 4.2: Search Strategy
1 Identical search terms were used in all three databases. For all databases, the original search date was 28 March 2019.
Database MEDLINE, EMBASE, Cochrane1
Search
Terms
1. aspartame.mp. 2. exp Aspartame/
3. neotame.mp.
4. saccharin.mp. 5. exp Saccharin/
6. sucralose.mp.
7. stevia.mp. 8. exp Stevia/
9. acesulfame.mp.
10. exp Sweetening Agents/ 11. sugar substitute*.mp.
12. noncaloric.mp.
13. non-caloric.mp. 14. nonnutritive.mp.
15. non-nutritive.mp.
16. no calorie*.mp. 17. low calorie sweeten*.mp.
18. sugar-free.mp.
19. artificial sweet*.mp. 20. diet beverage.mp.
21. diet soda.mp.
22. exp Carbonated Beverages/ 23. ssb.mp.
24. sugar*sweetened beverage*.mp.
25. Fructose/ 26. fructose.mp.
27. dietary sucrose/
28. sucrose.mp. 29. High Fructose Corn Syrup/
30. High Fructose Corn Syrup.mp.
31. Soft drink*.mp. 32. cola.mp.
33. sugar*.mp.
34. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 11 or 12 or
13 or 14 or 15 or 16 or 17 or 18 or 20 or 21 or 22 or
23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or
32 or 33
35. Blood Pressure/
36. systolic blood pressure.mp. 37. SBP.mp.
38. diastolic blood pressure.mp.
39. DBP.mp. 40. 35 or 36 or 37 or 38 or 39
41. glyc*m*.mp. 42. Hemoglobin A, Glycosylated/
43. glyc*mia.mp. 44. insulin*.mp.
45. gly*albumin.mp.
46. OGTT.mp.
47. hba1c.mp.
48. HOMA*.mp.
49. fructosamine*.mp. 50. Insulin/
51. exp Glucose/
52. Glucose Tolerance Test/ 53. 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or
51 or 52
54. triglyceride.mp.
55. triacylglycerol.mp.
56. VLDL.mp. 57. very low density lipoprotein.mp.
58. lipid*.mp.
59. lipids/ 60. cholesterol/
61. cholesterol.mp.
62. lipoprotein.mp. 63. lipoproteins/
64. (hdl or high density lipoprotein).mp.
65. (ldl or low density lipoprotein).mp. 66. exp hyperlipidemias/
67. apolipoprotein*.mp.
68. non-HDL.mp. 69. 54 or 55 or 56 or 57 or 58 or 59 or 60 or 61 or 62 or 63 or
64 or 65 or 66 or 67 or 68
70. fatty liver.mp.
71. non-alcoholic fatty liver disease/
72. NAFLD.mp. 73. transaminases/
74. alanine transaminase/
75. alt.mp. 76. aspartate aminotransferase/
77. ast.mp.
78. IHCL.mp.
79. Intrahepatocellular lipid.mp.
80. transamin*.mp.
81. 70 or 71 or 72 or 73 or 74 or 75 or 76 or 77 or 78 or 79 or
80
82. exp uric acid/ 83. uric acid.mp.
84. urate.mp.
85. hyperuricemia/ 86. hyperuricemia.mp.
87. hyperuricaemia.mp.
88. uric.mp.
89. 82 or 83 or 84 or 85 or 86 or 87 or 88
90. randomized controlled trial.mp,pt.
91. randomized.mp. 92. placebo.mp.
93. 90 or 91 or 92
94. 40 or 53 or 69 or 81 or 89 95. 34 and 93 and 94
Database Total
Medline 1052
Embase 1475
Cochrane 2014
Manual Searches 5
TOTAL 4546
63
4Appendix Table 4.3: PICOTSb Framework
Participants Intervention Comparison Outcome Time Study Design
Adult men and
women
excluding
pregnant or
breastfeeding
women and
children. All
health
backgrounds
NSBs or
SSBs or
water
NSBs or
SSBs or water
Adiposity,
glycemic
control,
established
blood lipid
targets, blood
pressure, non-
alcoholic fatty
liver and uric
acid
≥ 3
weeks
Human
randomized
controlled
trials
bHutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, Ioannidis JP, Straus S, Thorlund K,
Jansen JP, Mulrow C, Catalá-López F, Gøtzsche PC, Dickersin K, Boutron I, Altman DG, Moher D. The PRISMA
extension statement for reporting of systematic reviews incorporating network meta-analyses of health care
interventions: checklist and explanations. Ann Intern Med. 2015 Jun 2;162(11):777-84.
64
5Appendix Table 4.4: Side-Splitting Approach for Inconsistency for Body Weight
Side Direct Indirect Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water
for SSBs
0.4057214 1.177256 0.042085 0.839443 -0.44781 1.443613 0.756
NSBs
for SSBs
1.066648 0.4535466 1.519124 1.39197 -0.45248 1.464935 0.757
NSBs
for
Water
1.100907 0.7015187 0.668898 1.277997 0.432009 1.456246 0.767
6Appendix Table 4.5: Side-Splitting Approach for Inconsistency for BMI
Side Direct
Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.1778188 0.282524 -0.7871 0.450022 0.609281 0.529387 0.25
NSBs for
SSBs
0.3601485 0.138045 -0.26236 0.519513 0.622504 0.537329 0.247
NSBs for
Water
0.4060461 0.41773 0.189944 0.319592 -0.59599 0.525259 0.257
BMI, body mass index; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
7Appendix Table 4.6: Side-Splitting Approach for Inconsistency for Body Fat %
Side Direct
Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
-0.192686 0.7165101 -1.147723 1.942673 1.128455 2.067054 0.585
NSBs for
SSBs
0.6185242 0.2192872 0.6197683 2.074374 1.238293 2.085881 0.553
NSBs for
Water
0.5359688 1.898476 0.6054734 0.7514356 -1.141442 2.039145 0.576
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
8Appendix Table 4.7: Side-Splitting Approach for Inconsistency for Waist Circumference
Side Direct
Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.30000
01
2.03127
8
0.820217
2
141.4286 -
0.5202172
141.4432 0.99
7
NSBs for
Water
0.82023
07
1.02721
7
-
0.220309
1
282.9657 1.04054 282.9689 0.99
7
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
65
9Appendix Table 4.8: Side-Splitting Approach for Inconsistency for HbA1c
Side Direct Direct Indirect
Differenc
e
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for SSBs
-0.05
0.219310
6
-
0.11329
0.17317
4 0.063291
0.27943
9
0.82
1
NSBs for SSBs
-0.11 0.125894
-
0.17329
0.24947
3 0.063291
0.27943
9
0.82
1
NSBs for
Water -
0.223292
0.118911
1 -0.16
0.25287
7 -0.06329 0.27944
0.82
1 HbA1c, hemoglobin A1c; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
10Appendix Table 4.9: Side-Splitting Approach for Inconsistency for FPG
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
-0.03654 0.13924 -0.18162 0.107769 0.145082 0.175492 0.408
NSBs for
SSBs
0.168507 0.092779 0.004328 0.168865 0.164179 0.191957 0.392
NSBs for
Water
-0.00184 0.040021 0.135048 0.200939 -0.13689 0.204885 0.504
FPG, fasting plasma glucose; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
11Appendix Table 4.10: Side-Splitting Approach for Inconsistency for 2h-PG
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
NSBs for
SSBs
0.29122 0.378731 1.71E-03 142.5415 -0.29293 142.542 0.998
NSBs for
Water
0.19005 0.099326 -0.7558 284.7283 0.565751 284.7282 0.998
2h-PG, two-hour post prandial glucose; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
12Appendix Table 4.11: Side-Splitting Approach for Inconsistency for FPI
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
12.09007 21.76729 -32.2253 15.84592 44.31534 29.52399 0.133
NSBs for
SSBs
12.1009 10.80856 -58.2259 46.91402 70.32684 48.63387 0.148
NSBs for
Water
-7.99187 5.604813 31.18152 42.82504 -39.1734 42.73854 0.359
FPI, fasting plasma insulin; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
66
13Appendix Table 4.12: Side-Splitting Approach for Inconsistency for HOMA-IR
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.32139 0.496521 -0.86131 0.670215 1.182697 0.915703 0.197
NSBs for
SSBs
0.241025 0.3648 -1.02249 1.101164 1.26352 1.182708 0.285
NSBs for
Water
-0.04597 0.198863 1.566582 1.878476 -1.61255 1.882096 0.392
HOMA-IR, homeostatic model assessment of insulin resistance; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
14Appendix Table 4.13: Side-Splitting Approach for Inconsistency for LDL-C
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.08227
0.13972 0.03716 0.15838 0.07855 0.20937 0.70800 NSBs for
SSBs 0.06182 0.14030 -0.12531 0.16127 0.18713 0.21375 0.38100 NSBs for
Water 0.02700 0.04922 0.07387 0.22275 -0.04687 0.22814 0.83700 LDL-C, low-density lipoprotein cholesterol, NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
15Appendix Table 4.14: Side-Splitting Approach for Inconsistency for Non-HDL-C
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
-0.04854 0.105695 -0.09359 0.150071 0.04505 0.183311 0.806
NSBs for
SSBs
0.158931 0.124875 0.00913 0.124738 0.149801 0.176104 0.395
NSBs for
Water
0.021362 0.067712 0.026765 0.182856 -0.0054 0.196445 0.978
Non-HDL-C, Non-high-density lipoprotein cholesterol, NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
16Appendix Table 4.15: Side-Splitting Approach for Inconsistency for TGs
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
-0.28485 0.1587012 -0.0957 0.157552 -0.18915 0.222016 0.394
NSBs for
SSBs
0.217221 0.1340943 0.279237 0.186124 -0.06202 0.226371 0.784
NSBs for
Water
0.056847 0.0561295 -0.10666 0.241273 0.163505 0.248033 0.51
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; TGs, triglycerides.
67
17Appendix Table 4.16: Side-Splitting Approach for Inconsistency for HDL-C
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.040828 0.052674 -0.01915 0.071828 0.059977 0.089072 0.501
NSBs for
SSBs
0.0106076 0.059408 -0.06617 0.063332 0.076774 0.086835 0.377
NSBs for
Water
0.0089119 0.024449 0.043649 0.092413 -0.05256 0.095592 0.582
HDL-C, high-density lipoprotein cholesterol, NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
18Appendix Table 4.17: Side-Splitting Approach for Inconsistency for TC
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
-0.0606 0.2349 -0.0937 0.1625 -0.0331 0.2825 0.907
NSBs for
SSBs
0.1603 0.1392 -0.1716 0.2913 0.3319 0.3196 0.299
NSBs for
Water
0.0212 0.0635 -0.0829 0.3428 0.1041 0.3488 0.765
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; TC, total cholesterol.
19Appendix Table 4.18: Side-Splitting Approach for Inconsistency for SBP
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.460598 2.588089 -3.12636 5.446124 3.586955 6.020967 0.551
NSBs for
SSBs
5.694315 3.635833 0.220911 3.470725 5.473405 5.045351 0.278
NSBs for
Water
2.319843 1.840343 7.140843 6.00167 -4.821 6.25626 0.441
NSBs, non-nutritive sweetened beverages; SBP, systolic blood pressure; SSBs, sugar-sweetened beverages.
20Appendix Table 4.19: Side-Splitting Approach for Inconsistency for DBP
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
-0.17673 1.634782 -5.322171 3.461488 5.145445 3.813087 0.177
NSBs for
SSBs
5.158686 2.408139 -0.8695577 2.334642 6.028243 3.349012 0.072
NSBs for
Water
0.989036 2.240188 4.606612 4.569729 -3.61758 5.252017 0.491
DBP, diastolic blood pressure; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
68
21Appendix Table 4.20: Side-Splitting Approach for Inconsistency for IHCL
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.46000 0.21000 0.34051 0.24464 0.11949 0.32241
NSBs for
SSBs
0.42051 0.14089 0.54000 0.29000 -0.11949 0.32241
NSBs for
Water
-0.08000 0.20000 0.03949 0.25288 -0.11949 0.32241
IHCL, intra-hepatocellular lipid; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
22Appendix Table 4.21: Side-Splitting Approach for Inconsistency for ALT
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
NSBs for
SSBs
3.5000 7.1000 0.0000 223.6259 3.5000 223.7386 0.988
NSBs for
Water
-1.1998 1.4881 5.8002 447.2031 -7.0000 447.2007 0.988
ALT, alanine transaminase; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
23Appendix Table 4.22: Side-Splitting Approach for Inconsistency for AST
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
NSBs for
SSBs
1.5 3.25 2.36E-04 223.6165 1.499764 223.6402 0.995
NSBs for
Water
-0.2 1.781909 2.800086 447.2568 -3.00009 447.2532 0.995
AST, aspartate transaminase; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
24Appendix Table 4.23: Side-Splitting Approach for Inconsistency for Uric Acid
Side Direct Direct Indirect
Difference
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. P>z
Water for
SSBs
0.05 0.023463 0.036245 0.038575 0.013756 0.045151
NSBs for
SSBs
0.026253 0.0201564 0.0399999 0.039979 -0.01375 0.044773
NSBs for
Water
0.01 0.0324097 0.0237565 0.031373 -0.01376 0.045107
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
69
25Appendix Table 4.24: Loop-Specific Approach for Inconsistency
Outcome P-value Inconsistency
Body weight P=0.074 Not present
BMI P=0.8 Not present
Body fat P=0.98 Not present
Waist circumference No triangular loop found
HbA1c Unable to calculate*
FPG P=0.25 Not present
2h-PG No triangular loop found
FPI P=0.75 Not present
HOMA-IR P=0.76 Not present
LDL-C P=0.95 Not present
Non-HDL-C P=0.66 Not present
TGs P=0.04 Present
HDL-C P=0.49 Not present
TC P=0.63 Not present
SBP P=0.15 Not present
DBP P=0.57 Not present
IHCL Unable to calculate*
ALT No triangular loop found
AST No triangular loop found
Uric acid Unable to calculate* 2h-PG, two-hour post prandial glucose; ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass
index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; FPI, fasting plasma insulin; HbA1c, hemoglobin
A1c; HDL-C, high-density lipoprotein-cholesterol; HOMA-IR, Homeostatic Model Assessment of Insulin
Resistance; IHCL, intra-hepatocellular lipid; LDL-C, low-density lipoprotein-cholesterol; Non-HDL-C, non-high-
density lipoprotein-cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TGs, triglycerides.
* Calculation issues with loop-specific approach in Stata with multiple-arm trials as model did not converge,
therefore, “unable to calculate” for all missing ones.
70
Figures
5Appendix Figure 4.1: Cochrane Risk of Bias Summary for all Included Trials
71
6Appendix Figure 4.2: Risk of Bias Proportion for all Included Trials
72
7Appendix Figure 4.3: Transitivity Analysis_Box Plots Showing the Distribution of the Mean
Age (Years) of the Trials Across the Available Direct Comparisons
73
8Appendix Figure 4.4: Transitivity Analysis_Box Plots Showing the Distribution of the study
Length (Weeks) of the Trials Across the Available Direct Comparisons
74
9Appendix Figure 4.5: Transitivity Analysis_Box Plots Showing the Distribution of the Sample
Size of the Trials Across the Available Direct Comparisons
75
10Appendix Figure 4.6: Transitivity Analysis_Box Plots Showing the Distribution of the % Males
of the Trials Across the Available Direct Comparisons
76
11Appendix Figure 4.7: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Adiposity Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for SSBs on outcomes of adiposity. Data
was pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise
differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To
display the results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent
which categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for
the network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network estimate.
BMI, body mass index; GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; MID, Minimally Important
Difference; NE, not estimable; No., number; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; WC, waist circumference. 1The certainty of evidence for the indirect estimates is based off the lowest rating in the first order loop. 2We rated down the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%, P<0.10). 3We rated down the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We rated down the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was available
for assessment. 5We rated down the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify treatment effect in the direct comparisons (intransitivity). 6We rated down the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.5kg for body weight, 0.2kg/m2 for BMI, 2% for body fat and 2cm for waist circumference. 7The certainty of the evidence for the network was based on the certainty of the evidence for the indirect estimate as no direct evidence was
available. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
77
12Appendix Figure 4.8: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Glycemic Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for SSBs on glycemic outcomes. Data was
pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct
pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the
results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the
MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the
network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network
estimate. 2h-PG, two-hour post prandial glucose; FPG, fasting plasma glucose; FPI, fasting plasma insulin; GRADE, Grading of Recommendations,
Assessment, Development, and Evaluation; HbA1c, hemoglobin A1c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; MID,
Minimally Important Difference; NE, not estimable; No., number; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages. 1The certainty of evidence for the indirect estimates is based off the lowest rating in the first order loop. 2We rated down the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We rated down the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We rated down the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was available for assessment. 5We rated down the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We rated down the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.3% for HbA1c, 0.5mmol/L for FPG, 10% for 2h-PG, 5pmol for FPI and 1 for
HOMA-IR. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
78
13Appendix Figure 4.9: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Blood Lipid Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for SSBs on blood lipid outcomes. Data was
pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct
pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the
results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the
MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the
network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network
estimate. GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-
density lipoprotein-cholesterol; MID, Minimally Important Difference; No., number; Non-HDL-C, non-high-density lipoprotein-cholesterol;
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; TC, total cholesterol; TGs, triglycerides. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.1mmol/L. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
79
14Appendix Figure 4.10: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for SSBs
on Blood pressure, NAFLD and Uric Acid Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for SSBs on blood pressure, NAFLD and
uric acid outcomes. Data was pooled using network random effects models and expressed as mean differences (MD) and 95% confidence
intervals (CIs). MDs for the direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a
common comparator. To display the results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs
(proportionally scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which categories downgrades were made among the direct, indirect and network estimates. The starting point
for the certainty of the evidence for the network estimate was based on the certainty of the evidence for the direct or indirect estimate that
contributed the most weight to the network estimate. ALT, alanine transaminase; AST, aspartate transaminase; DBP, diastolic blood pressure; GRADE, Grading of Recommendations, Assessment,
Development, and Evaluation; IHCL, intra-hepatocellular lipid; MID, Minimally Important Difference; NE, not estimable; No., number; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; SBP, systolic blood pressure. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 2mmHg for SBP, 2mmHg for DBP, 10% for IHCL, 10% for ALT, 10% for AST,
and 0.22 mg/dl (0.013mmol/L) for uric acid. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005d Brignardello-Petersen, R., et al.,
The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates from a Network
Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005
80
15Appendix Figure 4.11: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Adiposity Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of water for SSBs on outcomes of adiposity. Data was pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the
direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise
differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95%
CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent
which categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network
estimate.
BMI, body mass index; GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; MID, Minimally Important Difference; NE, not estimable; No., number; SSBs, sugar-sweetened beverages; WC, waist circumference. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.5kg for body weight, 0.2kg/m2 for BMI, 2% for body fat and 2cm for waist
circumference. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
81
16Appendix Figure 4.12: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Glycemic Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of water for SSBs on glycemic outcomes. Data was
pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct
pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the
results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the
MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the
network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network
estimate. 2h-PG, two-hour post prandial glucose; FPG, fasting plasma glucose; FPI, fasting plasma insulin; GRADE, Grading of Recommendations,
Assessment, Development, and Evaluation; HbA1c, hemoglobin A1c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; MID,
Minimally Important Difference; NE, not estimable; No., number; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.3% for HbA1c, 0.5mmol/L for FPG, 10% for 2h-PG, 5pmol for FPI and 1 for
HOMA-IR. 7The certainty of the evidence for the network was based on the certainty of the evidence for the indirect estimate as no direct evidence was
available. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
82
17Appendix Figure 4.13: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Blood Lipid Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of water for SSBs on blood lipid outcomes. Data was
pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences
were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the
results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which
categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the
network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network estimate.
GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-
density lipoprotein-cholesterol; MID, Minimally Important Difference; No., number; Non-HDL-C, non-high-density lipoprotein-cholesterol; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; TC, total cholesterol; TGs, triglycerides. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%, P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.1mmol/L. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
83
18Appendix Figure 4.14: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of Water for SSBs
on Blood pressure, NAFLD and Uric Acid Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of water for SSBs on blood pressure, NAFLD and
uric acid outcomes. Data was pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs
for the indirect pair-wise differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a
common comparator. To display the results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal
line. Filled, black boxes represent which categories downgrades were made among the direct, indirect and network estimates. The starting point
for the certainty of the evidence for the network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network estimate.
ALT, alanine transaminase; AST, aspartate transaminase; DBP, diastolic blood pressure; GRADE, Grading of Recommendations, Assessment,
Development, and Evaluation; IHCL, intra-hepatocellular lipid; MID, Minimally Important Difference; NE, not estimable; No., number; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; SBP, systolic blood pressure. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%, P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 2mmHg for SBP, 2mmHg for DBP, 10% for IHCL, 10% for ALT, 10% for AST, and 0.22 mg/dl (0.013mmol/L) for uric acid. 7The certainty of the evidence for the network was based on the certainty of the evidence for the indirect estimate as no direct evidence was
available. c Puhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. d Brignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
84
19Appendix Figure 4.15: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
on Adiposity Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for water on outcomes of adiposity. Data was pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the
direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise
differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95%
CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent
which categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network
estimate.
BMI, body mass index; GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; MID, Minimally Important Difference; NE, not estimable; No., number; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; WC waist
circumference. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified MIDs for benefit and/or harm. MIDs for each outcome were: 0.5kg for body weight, 0.2kg/m2 for BMI, 2% for body fat and 2cm for waist
circumference. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
85
20Appendix Figure 4.16: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for water on glycemic outcomes. Data was
pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct
pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences
were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the
results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the
MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which
categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the
network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network
estimate.
2h-PG, two-hour post prandial glucose; FPG, fasting plasma glucose; FPI, fasting plasma insulin; GRADE, Grading of Recommendations,
Assessment, Development, and Evaluation; HbA1c, hemoglobin A1c; HOMA-IR, Homeostatic Model Assessment of Insulin Resistance; MID,
Minimally Important Difference; NE, not estimable; No., number; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.3% for HbA1c, 0.5mmol/L for FPG, 10% for 2h-PG, 5pmol for FPI and 1 for
HOMA-IR. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
86
21Appendix Figure 4.17: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
on Blood Lipid Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for water on blood lipid outcomes. Data was
pooled using network random effects models and expressed as mean differences (MD) and 95% confidence intervals (CIs). MDs for the direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs for the indirect pair-wise differences
were synthesized with the use of generic inverse variance random effects models and 95% CIs using a common comparator. To display the
results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs (proportionally scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal line. Filled, black boxes represent which
categories downgrades were made among the direct, indirect and network estimates. The starting point for the certainty of the evidence for the
network estimate was based on the certainty of the evidence for the direct or indirect estimate that contributed the most weight to the network estimate.
GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-
density lipoprotein-cholesterol; MID, Minimally Important Difference; No., number; Non-HDL-C, non-high-density lipoprotein-cholesterol; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; TC, total cholesterol; TGs, triglycerides. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%, P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 0.1mmol/L. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
87
22Appendix Figure 4.18: Superplot of the Direct, Indirect and Network Analysis with GRADEc,d
Assessment of the Certainty of the Evidence on the Effect of the Substitution of NSBs for Water
on Blood pressure, NAFLD and Uric Acid Outcomes
Network analyses of randomized controlled trials investigating the effect of the substitution of NSBs for water on blood pressure, NAFLD and
uric acid outcomes. Data was pooled using network random effects models and expressed as mean differences (MD) and 95% confidence
intervals (CIs). MDs for the direct pair-wise differences were synthesized with the use of generic inverse variance random effects models. MDs
for the indirect pair-wise differences were synthesized with the use of generic inverse variance random effects models and 95% CIs using a
common comparator. To display the results for outcomes on the same plot, standardized mean differences (SMDs) and pseudo 95% CIs
(proportionally scaled to the 95% CIs of the MDs) were calculated. SMDs are represented by a red diamond and 95% CIs by the black horizontal
line. Filled, black boxes represent which categories downgrades were made among the direct, indirect and network estimates. The starting point
for the certainty of the evidence for the network estimate was based on the certainty of the evidence for the direct or indirect estimate that
contributed the most weight to the network estimate.
ALT, alanine transaminase; AST, aspartate transaminase; DBP, diastolic blood pressure; GRADE, Grading of Recommendations, Assessment,
Development, and Evaluation; IHCL, intra-hepatocellular lipid; MID, Minimally Important Difference; NE, not estimable; No., number; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages; SBP, systolic blood pressure. 1The certainty of evidence for the indirect estimates is based off the lowest certainty of evidence in the first order loop. 2We downgraded the evidence for serious inconsistency in the direct estimates as there was substantial unexplained heterogeneity (I2≥50%,
P<0.10). 3We downgraded the evidence for serious inconsistency in the network estimates as there was evidence of incoherence (the 95% CIs of the direct
and indirect estimates were significantly different, P<0.05). 4We downgraded the evidence for serious indirectness in the direct estimates as only one RCT that was not sufficiently generalizable was
available for assessment. 5We downgraded the evidence for serious indirectness in the indirect estimates as there were differences in study characteristics that may modify
treatment effect in the direct comparisons (intransitivity). 6We downgraded the evidence for serious imprecision in the direct and/or network estimates as the 95% CIs were wide and crossed prespecified
MIDs for benefit and/or harm. MIDs for each outcome were: 2mmHg for SBP, 2mmHg for DBP, 10% for IHCL, 10% for ALT, 10% for AST,
and 0.22 mg/dl (0.013mmol/L) for uric acid. cPuhan, M.A., et al., The GRADE Working Group (2014). A GRADE Working Group Approach for Rating the Quality of Treatment Effect
Estimates from Network Meta-Analysis. BMJ 349. doi:10.1136/bmj.g5630. dBrignardello-Petersen, R., et al., The GRADE Working Group (2018). Advances in the GRADE Approach to Rate the Certainty in Estimates
from a Network Meta-Analysis. Journal of Clinical Epidemiology 93. doi:10.1016/j.jclinepi.2017.10.005.
88
23Appendix Figure 4.19: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Body
Weight
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SSBs, sugar-sweetened beverages.
24Appendix Figure 4.20: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on BMI
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. BMI; body mass index, NSBs, non-
nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
89
25Appendix Figure 4.21: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Body Fat
%
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SSBs, sugar-sweetened beverages.
26Appendix Figure 4.22: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Waist
Circumference
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SSBs, sugar-sweetened beverages.
90
27Appendix Figure 4.23: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on HbA1c
Blue nodes represent the study size for each beverage under investigation. The thickness of the
black lines represents the number of studies directly comparing one beverage to another. HbA1c,
hemoglobin A1c; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
28Appendix Figure 4.24: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on FPG
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. FPG, fasting plasma glucose; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
91
29Appendix Figure 4.25: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on 2h-PG
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. 2h-PG, two-hour post prandial
glucose; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
30Appendix Figure 4.26: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on FPI
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. FPI, fasting plasma insulin; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
92
31Appendix Figure 4.27: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on HOMA-IR
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. HOMA-IR; homeostatic model
assessment of insulin resistance; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
32Appendix Figure 4.28: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on LDL-C
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. LDL-C, low-density lipoprotein
cholesterol; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
93
33Appendix Figure 4.29: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on Non-HDL-
C
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. Non-HDL-C, non-high-density
lipoprotein cholesterol, NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
34Appendix Figure 4.30: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on TGs
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SSBs, sugar-sweetened beverages; TGs, triglycerides.
94
35Appendix Figure 4.31: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on HDL-C
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. HDL-C, high-density lipoprotein
cholesterol, NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
36Appendix l Figure 4.32: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on TC
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SSBs, sugar-sweetened beverages; TC, total cholesterol.
95
37Appendix Figure 4.33: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on SBP
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SBP, systolic blood pressure; SSBs, sugar-sweetened beverages.
38Appendix Figure 4.34: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on DBP
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. DBP, diastolic blood pressure; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
96
39Appendix Figure 4.35 Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on IHCL
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. IHCL, intra-hepatocellular lipid;
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
40Appendix Figure 4.36: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on ALT
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. ALT, alanine transaminase; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
97
41Appendix Figure 4.37: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on AST
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. AST, aspartate transaminase; NSBs,
non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages.
42Appendix Figure 4.38: Network Diagram for Randomized Controlled Trials Investigating the
Effect of the Substitution of NSBs for SSBs, Water for SSBs, and NSBs for Water on uric acid
Blue nodes represent the study size for each beverage under investigation. The thickness of the black lines
represents the number of studies directly comparing one beverage to another. NSBs, non-nutritive sweetened
beverages; SSBs, sugar-sweetened beverages.
98
Chapter 5: Rationale, Design and Baseline Characteristics Assessing the
Effect of Substituting NSBs versus Water for SSBs on Gut Microbiome,
Glucose Tolerance, and Cardiometabolic Risk Factors: Strategies To
OPpose SUGARS with Non-nutritive sweeteners Or Water trial (STOP
Sugars NOW)
Néma McGlynn1,2,3, Sabrina Ayoub-Charette1,2,3, Tauseef A Khan1,2,3, Sonia Blanco-Mejia1,2,3, Laura Chiavaroli1,2,3,
Meaghan Kavanagh1,2,3, Danielle Lee1,2,3, Elena M Comelli3,4, Cyril WC Kendall1,2,3,5, Lawrence A Leiter1,2,3,6,7,8,
John L Sievenpiper1,2,3,6,7,8
1Toronto 3D Knowledge Synthesis and Clinical Trials Unit, Toronto, ON, Canada; 2Clinical Nutrition and Risk Factor Modification Centre, St. Michael's Hospital, Toronto, ON, Canada; 3Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, ON, Canada 4Joannah and Brian Lawson Centre for Child Nutrition, Toronto, ON, Canada. 5College of Pharmacy and Nutrition, University of Saskatchewan, Saskatoon, SK, Canada; 6Division of Endocrinology and Metabolism, Department of Medicine, St. Michael’s Hospital, Toronto, ON,
Canada; 7Department of Medicine, Faculty of Medicine, University of Toronto, Toronto, ON, Canada; 8Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada;
Corresponding Author: John L Sievenpiper MD, PhD, FRCPC, St. Michael's Hospital, #6138-61 Queen Street East,
Toronto, ON, M5C 2T2, CANADA, Tel: 416 867 7475, Fax: 416 867 7495, email: [email protected]
99
5.1 Abstract
Background: Concerns exist that non-nutritive sweetened beverages (NSBs) do not have
established benefits and adversely affect glucose tolerance through compositional changes of the
gut microbiome. Whether the intended benefits of NSBs improve glucose control similar to
water in their substitution for sugar-sweetened beverages (SSBs) is unclear. The STOP Sugars
NOW study is a randomized crossover trial on the effect of 3 prespecified substitutions (NSBs
for SSBs, water for SSBs and NSBs for water) on gut microbiome and glucose tolerance. This
report provides the rationale, design and baseline characteristics of the trial.
Methods: A randomized crossover trial with three 4-week intervention phases (SSBs, NSBs and
water) in 81 overweight and obese adult participants, who consume at least one 355mL can of
SSBs daily. Each intervention phase is separated by a ≥4-week washout phase where participants
revert back to their regular SSB intake. The protocol includes 6 study visits. Two primary
outcomes, gut microbiome beta-diversity and glucose tolerance, are analyzed through 16S rRNA
sequencing of fecal sample collections and 2-hour 75g oral glucose tolerance tests, respectively.
Adherence to study beverages is assessed through biomarker analysis of 24-hour urine, collected
at each study visit.
Results: Participants were mainly recruited through Toronto Transit Commission subway
advertisements and a third-party participant recruitment service (Trialfacts). At baseline,
participants had a mean age of 41.8 ± 12.9 years, weight of 94.2 ± 18.9kg, BMI of 33.8 ±
6.7kg/m2, and waist circumference of 108.7 ± 13.4cm. Mean baseline levels of blood pressure
were normal. STOP Sugars NOW participants consumed an average of two 355mL cans of SSBs
daily, 43.2% consumed Coke, and 25.9% consumed Canada Dry Ginger Ale. Projected NSB
intake indicates 95% of participants will be consuming a blend of aspartame and acesulfame
potassium. Sixty-one percent of participants reported no consumption of any foods or beverages
with non-nutritive sweeteners (NNSs) over the last 6 months at baseline, whereas 24.7% of those
who consumed NNSs did so from NSBs.
Conclusion: The STOP Sugars NOW trial will be the first randomized controlled trial to
investigate the effect of 3 prespecified substitutions (NSBs for SSBs, water for SSBs and NSBs
for water) on changes in the gut microbiota and glucose tolerance.
Trial registration number: ClinicalTrials.gov identifier, NCT03543644
100
5.2 Introduction
Health authorities discourage the consumption of sugar-sweetened beverages (SSBs) [11, 73,
186, 187] as their excess intakes have been associated with weight gain, type 2 diabetes mellitus
(T2DM), and their downstream complications of hypertension and coronary heart disease (CHD)
[6-9, 188].
Non-nutritive sweetened beverages (NSBs) provide a viable alternative for SSBs, yet concerns
exist that they do not have established benefits, with mixed recommendations for NSBs in
dietary and clinical practice guidelines [11, 12, 15, 16, 18], the majority recommending that
water replace SSBs. Whether the intended benefits of NSBs improve glucose control similar to
water in their substitution for SSBs is unclear.
Systematic reviews and meta-analyses of prospective cohort studies, including a WHO-
commissioned review [21], indicate that NSBs are associated with an increased risk of the
conditions they are intended to prevent including weight gain, type 2 diabetes mellitus (T2DM)
and coronary heart disease (CHD) [7, 19], even though reverse causality and confounding by
indication are the most likely explanations [134, 143].
Syntheses of randomized controlled trials (RCTs), have failed to fully account for the calories
available to be displaced by NSBs with caloric (e.g. SSBs) and noncaloric (e.g. water, placebo,
weight loss diets) comparators pooled together, or noncaloric comparators used as the sole
comparator leading to an underestimation of the effect of NSBs in their intended substitution for
SSBs [19-21].
Although biological mechanisms involving impaired sensory and endocrine signaling mediated
by the sweet taste receptor [79, 80] have been offered in support of these observations, there is
particular concern that NSBs may induce and promote glucose intolerance through changes to
the microbiome [130, 131]. Much of this concern derives from a single intervention study that
showed a decrease in glucose tolerance after 6 days of saccharin in an NSB format with
subsequent transfer of the phenotype to germ free mice by fecal transplant [81]. This study,
despite its concerning methodological weaknesses, including a small sample size of 7 healthy
participants, daily continuous 75g Oral Glucose Tolerance Tests (OGTT) and no control group,
101
resulted in an overall negative view of non-nutritive sweeteners (NNSs) in the media [182, 183,
189-194].
These prevailing uncertainties signal a need to address the ongoing concerns related to NSBs.
Health Canada, in particular, has indicated that studies of sugar reduction strategies that use
NNSs and target the microbiome are an important research priority [195]. We have therefore,
designed and implemented a Canadian Institutes for Health Research (CIHR)-funded RCT to
assess the effect of a “real world” strategy of 3 prespecified substitutions: NSBs for SSBs
(intended substitution with caloric displacement), water for SSBs (“standard of care” substitution
with caloric displacement) and NSBs for water (matched substitution without caloric
displacement) on gut microbiome and glucose tolerance in overweight/obese participants: The
STOP Sugars NOW trial.
5.3 Objective and Hypotheses
There are 3 primary objectives of this study.
1. To assess the effect of our 3 prespecified substitutions (NSBs for SSBs, water for SSBs
and NSBs for water) on the first primary outcome of diversity of gut microbiome over 4-
weeks in overweight/obese participants who are regular SSB drinkers.
2. To assess the effect of our 3 prespecified substitutions (NSBs for SSBs, water for SSBs
and NSBs for water) on the second primary outcome of glucose tolerance over 4-weeks
in overweight/obese participants who are regular SSB drinkers.
3. To assess the effect our 3 prespecified substitutions (NSBs for SSBs, water for SSBs and
NSBs for water) on the secondary outcomes of body weight, body mass index (BMI),
waist circumference and glucose and insulin regulation over 4-weeks in overweight/obese
participants who are regular SSB drinkers.
The hypothesis for each objective is listed below.
1. Substituting NSBs for SSBs and water for SSBs will improve the diversity of gut
microbiome after 4-weeks in overweight/obese participants who are regular SSB
drinkers, with no difference between the substitution of NSBs for water.
102
2. Substituting NSBs for SSBs and water for SSBs will improve glucose tolerance after 4-
weeks in overweight/obese participants who are regular SSB drinkers, with no difference
between the substitution of NSBs for water.
3. Substituting NSBs for SSBs and water for SSBs will decrease measures of adiposity and
improve glucose and insulin regulation after 4-weeks in overweight/obese participants
who are regular SSB drinkers, with no difference between the substitution of NSBs for
water.
5.4 Methods
5.4.1 Study Design
The STOP Sugars NOW trial is a single-center, open label, randomized controlled crossover trial
among 81 overweight or obese participants. This study was approved by The Unity Health
Toronto Research Ethics Board in Jan 2018 and is registered at www.clinicaltrials.gov
(NCT03543644).
In order to compare the effect of our 3 prespecified substitutions on gut microbiome diversity
and glucose tolerance, participants consume SSBs, NSBs or water over a period of 4-weeks each.
Due to the high inter-individual heterogeneity in the gut microbiome, a crossover design was
selected [196]. A schematic representation of the study design can be seen in Figure 5.1.
Participants start the study as their own control receiving the 3 interventions in random order,
with each separated by a ≥4-week washout phase.
103
5.4.2 Blinding
Blinding of the participants and investigators is not possible due to the packaging, taste and look
of the interventions. However, outcome assessors (laboratory, microbiome analysis) and the
statistician will be blinded to the identity of the treatments.
5.4.3 Participants
We recruited 81 participants consuming ≥1 can (355ml) serving of SSBs per day (for which
there is an NSB equivalent that contains the NNSs acesulfame potassium (Ace-K) or sucralose),
who were overweight or obese based on ethnic-specific cut-offs (BMI ≥23kg/m2 for Asian
individuals and ≥ 25kg/m2 other individuals [37]), with a high waist circumference (≥94cm in
men, ≥80cm in women of Europid, Sub-Saharan African, Eastern Mediterranean, and Middle
Eastern origin; ≥90cm in men and ≥ 80cm in women of South Asian, Chinese, Japanese, and
South and Central American origin [35]) between the ages of 18-75 years, non-smokers, not
taking medications that have a clinically relevant effect on the primary outcomes, no intentions
of making major alterations to their diet or physical activity regime during the study duration,
had a primary care physician, and were proficient in English were eligible to participate in the
trial. The full set of exclusion criteria can be found on Table 5.1. Participants had to agree to
consume only one type of SSB and brand-matched NSB for each month of intervention. If
participants were consuming NSBs in addition to SSBs they were asked to discontinue
consuming them for the run-in and duration of the study. All participants are expected to
participate in the study for approximately six months (one run-in phase of ≥2-weeks, three
intervention phases of 4-weeks each, and two washout phases of ≥4-weeks each) and be able to
attend all study visits.
104
5.4.4 Sample Size (Power) Calculation
The study is being performed in 81 participants and powered to show a difference between the
NSB, water and SSB arms in 60 participants in the two primary outcomes. Assuming a drop-out
rate of 20%, we recruited 75 participants. Our final recruitment strategy resulted in a larger
number of participants who completed the run-in phase. Consequently, we increased our
recruitment to replace drop-outs, resulting in the recruitment 81 participants. Changes in beta
diversity of gut microbiome and changes in glucose tolerance are the two primary outcomes of
the trial. It is hypothesized that substituting NSBs for SSBs and water for SSBs will increase the
diversity of the gut microbiome and improve glucose tolerance after 4-weeks in regular SSB
consumers, with no difference between the substitution of NSBs for water. We are powered to
detect a difference between all three interventions (based on a power calculation difference
between the NSB and water arm), and this difference is expected to be smaller than the NSB for
SSB and water for SSB comparisons.
The first primary outcome of gut microbiome beta diversity is being analyzed through 16S
ribosomal rRNA gene sequencing using pairwise weighted UniFrac distances [197]. UniFrac is a
distance metric based upon the unique fraction of branch length in a phylogenetic tree built from
two sets of taxa. Comparison of microbiome samples will be performed via weighted UniFrac,
which considers the relative abundance of taxa. We simulated the within-group distance as 0.2,
and the standard deviation (SD) of within-group distances as 0.07. Previous studies have shown
that diversity of the gut microbiome taxa can change significantly with small dietary alterations
over a short period (5 to 7 days) in small groups of people (10 to 25) [196, 198]. Therefore, a
weighted UniFrac distance of 0.04 was set based on the effect of 0.5 observed in the study by
Suez et al. [81]. Due to the crossover design of this trial with a within-person correlation of 0.7,
it was calculated to show a difference between the water and NSB arms in both primary
outcomes with 98% power, 75 participants were required, assuming a drop-out rate of 20%.
The second primary outcome is change in glucose tolerance, as measured by incremental Area
Under the Curve (iAUC) from a 2-hour 75g OGTT. Assuming absolute numbers for mean and
SD from a recent unpublished randomized trial [199] and a within-person correlation of 0.7, 60
participants provide 89% power to detect a 20% change in mean iAUC between the three arms,
if the direction of change is similar to Suez et al. [81]. The 20% difference for glucose iAUC was
105
based on the minimally important difference proposed by Health Canada for postprandial
glycaemic response health claims [200] (Table 5.2).
To account for the adjustment required for multiple testing of two primary outcomes, this power
calculation uses the Benjamini-Hochberg procedure as suggested by the Food and Drug
Administration’s draft document “Multiple Endpoints in Clinical Trials Guidance for Industry”
[201]. This procedure is a step-down method that controls for false discovery rate, while
maintaining high power [202, 203]. To show a difference between SSBs, NSBs and water for all
primary and secondary outcomes, a truncated Benjamini-Hochberg method with parallel
gatekeeping is being implemented, whereby an unused portion of the alpha (α) is passed onto the
secondary family of outcomes if either of the primary outcomes is significant. An α of 0.0375 for
both primary outcomes were calculated based on a p-value of 0.05. For all secondary outcomes,
an α of 0.0125 was used as it’s the lowest possible starting α for secondary outcomes as per the
truncated Benjamini-Hochberg procedure.
5.4.5 Recruitment, Consent and Screening
Participants were recruited through two main advertising campaigns: The Toronto Transit
Commission (TTC) subway advertisements and Trialfacts, a third-party participant recruitment
service for clinical trials. Some participants completed an online Survey Monkey to assess
general eligibility. Other forms of recruitment included online advertising through Craigslist and
Kijiji and postcards on the University of Toronto, St George, campus and St Michael’s
Hospital’s (SMH) elevators in Toronto, Ontario, Canada. Interested individuals completed a
telephone-screen using a questionnaire and were briefed on the details of the study, treatments,
and procedures.
Eligible and interested individuals were asked to come to the Clinical Nutrition and Risk Factor
Modification Centre (CNRFMC) at SMH for an in-person consent review. Once consent was
obtained, individuals received a screening identification (ID) number and an in-person screening
was scheduled to obtain anthropometric measures such as height, weight and waist
circumference, completing a questionnaire regarding SSB intake, and blood pressure
measurements to confirm study qualification. Blood pressure was taken by Automated Office
Blood Pressure (AOBP) measurements on both arms. The arm with the highest mean systolic
measurements was used for all study visits. If the average of 3 AOBP measurements displayed a
106
systolic blood pressure ≥160mmHg or diastolic blood pressure ≥100mmHg, the individual was
excluded for uncontrolled hypertension [204].
Significant results were provided to participant’s family physician for further follow-up. Those
who met all eligibility criteria were enrolled in the run-in phase of the study and assigned a study
ID number.
5.4.6 Run-In Phase
Prior to randomization, participants underwent a ≥2-week run-in phase to determine their usual
SSB intake and their beverage preference options for SSBs, NSBs and water (flat or carbonated).
Before starting the run-in phase, participants were scheduled for their first study visit. They were
instructed on how to collect a 24-hour urine sample and fecal sample and provided kits to do so
for their first study visit.
5.4.7 Randomization
Randomization, with no stratification, was done using the Research Electronic Data Capture
(REDCap) program by the Applied Health Research Centre (AHRC) after successful completion
of the run-in phase and first study visit. Participants were randomly allocated to six possible
SSB, NSB and water groups using blocked (Latin squares) randomization with a similar number
of participants allocated to each treatment sequence (Table 5.3). The randomization schedule
was created by AHRC through REDCap. Participants were randomized and given their study
beverages once all measures from the first study visit were collected. Allocating participants
through REDCap kept study staff unaware of each participant’s group assignment until revealed
to them at the end of their first visit.
5.4.8 Intervention
There are three interventions: SSBs (355mL, 140kcal, 39g sugars per can); NSBs (355mL, 0kcal,
0g sugars per can); and water (355mL, 0kcal, 0g sugars per can or bottle of still or carbonated
water). The complete list of all available study beverages is listed in Table 5.4. SSBs were
defined as any SSB (sodas and soft drinks) that contains at least 50kcal per 8-oz serving. For the
purpose of the study, SSBs do NOT include fruit drinks, sports/energy drinks, sweetened iced
tea, coffee, or homemade SSBs such as frescas or 100% fruit juice.
107
For each intervention phase, participants are receiving relevant study beverages based on their
group assignment, with instructions to consume only one type of SSB, its brand-matched NSB
and carbonated or flat water for 4-weeks each in random order, while freely consuming their
usual background diet. To allow for “real-world” substitutions using products available on the
market, the calories of the intervention groups are not matched and the number of cans each
participant consumes each day are based off their ≥2-week run-in beverage logs.
5.4.9 Study Visits
All study visits are being conducted at the CFI-funded Toronto 3D Clinical Research Centre and
the CNRFMC at SMH. Participants are completing the three interventions in random order, each
separated by a ≥4-week washout. Motivational phone calls are being made every 2-weeks
between visits to provide participants with reminders and ensure fidelity to the study protocol.
The complete participant visit schedule is outlined in Table 5.5.
After successful completion of the run-in phase, participants came to the study site for their first
visit where their ≥2-week run-in beverage logs were checked for completion and any changes
required were communicated. Participants were notified of their randomization assignments after
completing all baseline tests.
Prior to each study visit, participants are given standard stool and urine collection kits and
instructions on how to complete a weighted three-day diet record (3DDR) which is being
completed and collected for visits 1, 2, 4 and 6 in order to assess adherence to background diet.
Participants are being instructed to collect their fecal sample and 24-hour urine as close to their
visit dates as possible. The fecal kits are required to assess diversity of gut microbiome and the
urine kits for urinary biomarker analysis of adherence. Participants are being informed to collect
their stool samples in the specimen containers provided. If collected prior to a visit date, they are
to store the fecal sample in their personal freezer using the containers and bags provided. All
fecal samples are to arrive at the centre in the bags provided, surrounded with dry ice packs. If
24-hour urine collections are taken ≥1 day before the participants’ study visits, the sample is
being kept in the participant’s fridge until their study visit date. At each visit, study staff retrieve
the stool sample, the 24-hour urine collection, and beverage logs from the participants. Once
samples are received from participants, they are labeled with the participant’s visit number, date
108
and given a fecal randomization code in order to keep study staff blinded for this outcome.
Samples are immediately placed in the lab’s freezer at –20°C until all samples are provided for
that visit. They are then transported with dry ice to the University of Toronto and frozen at –
80°C.
For all six study visits, participants are coming to the study site in a 10 to 12-hour fasted state.
Prior to their visits, participants are being reminded to maintain the same dietary and exercise
patterns the evening before each test and to consume at least 150g of carbohydrate for three non-
consecutive days before their scheduled visit date. Participants are being provided with concrete
examples of what 150g of carbohydrate looks like to ensure instructions are followed. At each
visit, weight, waist circumference and blood pressure measurements are being taken;
questionnaires assessing physical activity levels, NNS intake in foods and beverages and
medication and supplement changes are being administered and a fasting blood sample for
primary and secondary outcomes (including glucose, insulin and biomarkers of adherence) is
taken by an intravenous (IV) nurse. Administration of a 2-hour 75g OGTT is being performed at
each study visit, following standard protocol, to assess glucose tolerance. Between OGTT blood
draws, participants’ weighted 3DDRs and 24-hour urine and fecal sample forms are verified.
After the OGTT participants are given breakfast.
Once all measures are collected, the participant is provided the intervention. Beverage logs are
provided for each of the intervention and washout phases with reminders to log beverage intake
daily. For the washout phases, participants are purchasing their own SSBs. Subjects take home
one week’s worth of their beverages at the first visit of each phase. The rest of their study drinks
for that phase are picked-up by the participant or shipped to their home addresses by Starhawk, a
third-party delivery company. Consent for the shipment or pick-up of study drinks was obtained
in the consent review.
5.4.10 Washout phase
To control for any carry-over effects of one beverage type over another, each of the three
intervention phases is separated by a ≥4-week wash out phase where participants revert back to
their regular SSB intake. Participants are given beverage logs to complete over this ≥4-week
phase. No beverages are provided by the study site during the washout phase. Participants are
109
contacted via phone and email by study staff near the middle and end of each phase to remind
them of their group assignment and the next study visit.
5.4.11 Outcomes
There are two primary outcomes: change in gut microbiome beta diversity as measured by 16S
rRNA gene sequencing, and plasma glucose iAUC measured by a 2-hour 75g OGTT. Secondary
outcomes are changes in waist circumference, body weight, BMI, fasting plasma glucose (FPG),
2h plasma glucose (2h-PG), and the Matsuda whole body insulin sensitivity index (Matsuda
ISIOGTT) [205]. Exploratory outcomes consisting of a range of cardiometabolic/metabolomic
outcomes is also being considered. All outcome assessments are being performed by qualified
study staff and follow standard operating procedures.
5.4.12 Adherence Assessment
Adherence to study drinks is being assessed through completed beverage logs, returned unused
beverage containers, and objective biomarkers of SSBs (increased 13C/12C ratios in serum fatty
acids, increased urinary fructose), NSBs (urinary Ace-K, sucralose, decreased 13C/12C ratios in
serum fatty acids, decreased urinary fructose) and water (decreased 13C/12C ratios in serum
fatty acids, decreased urinary fructose, decreased urinary Ace-K and sucralose) intake.
(see biomarkers of adherence).
To determine adherence to background diet and evaluate changes in diet quality, participants are
being asked to complete a weighted 3DDR over two non-consecutive weekdays and one
weekend day during the run-in phase and last week of each intervention phases. To increase
accuracy of reporting, participants are provided with a food scale during the in-person screen.
Detailed instructions are given, using food models, on how to complete the record by study
dietitians who also review each 3DDR record with participants at their study visits. The 3DDRs
is being analyzed by Nutritics (Dublin, Ireland), a nutrition software tool.
5.4.13 Analytical Techniques
5.4.13.1 Anthropometric Analyses
All anthropometric measurements are being taken while the participant is wearing light clothes
and no shoes. Body weight is being assessed by electronic beam scale (Rice Lake® Weighing
System, Israel, model 240-10). A wall-mounted stadiometer (Perspective Enterprises, Portage,
110
MI, USA) is used to measure height. Waist circumference is being assessed using the Heart and
Stroke Foundation methodology [206] where participants are standing upright with feet
shoulder-width apart. The iliac crest is being determined using the sides of hands and index
fingers. As locating the iliac crest is often challenging among overweight and obese individuals,
to reduce measuring errors, a measuring tape (AMG Medical Inc, vinyl tape measure, 150cm,
AZGAMG116870.) is being used to determine length from the floor up to the top of the iliac
crest for each participant, during the in-person screen. This value is recorded in the participants’
charts and used at each study visit to ensure consistency in waist measurements. Measurements
from the floor to the top of the iliac crest are being done on both the right and left sides of the
body. Waist circumference is measured around the abdomen at the point of the iliac crest.
Participants are instructed to relax by taking in two normal breaths. After the second breath the
tape is to be tightened around the waist, the measurement collected and then recorded in the
participant’s chart. To obtain blood pressure and resting heart rate, participants are first seated in
a quiet, temperature-controlled room for at least 10-15 minutes. Blood pressure is then measured
oscillometrically by AOBP using the OMRON Intellisense HEM-907 according to JNC VII
criteria [207], where an average of three measurements, each separated by one minute, is taken.
5.4.13.2 Biochemical Analyses
The microbiome is being sequenced by next-generation sequencing (Illumina MiSeq) with
primers targeting the appropriate variable V3-V4 region of 16S rRNA [208] for pair-end
sequencing [209]. Sequencing data is being analyzed to assign Operational Taxonomic Units
(OTUs) to determine their relative abundance. Data for selected taxa is being confirmed by
qPCR [208]. Alpha- and beta-diversity indexes is being calculated. The metagenome is being
inferred from compositional data in silico [210]. Microbiome analysis is being performed at the
Department of Nutritional Sciences, University of Toronto.
Plasma samples for glucose and insulin is being separated by centrifuge in the wet laboratory at
the CNRFMC at SMH, within 4 hours. After each study visit, collected blood samples are
transported to the University of Toronto and immediately frozen at –80°C. Analyses of the
plasma glucose is being done at the Banting and Best Diabetes Centre (BBDC) using the glucose
oxidase method [211] at the University of Toronto. Plasma insulin analysis is being done using
enzyme-linked immunosorbent assay (ELISA), a solid phase two-site enzyme immunoassay
111
method using Mercodia AB Insulin ELISA (Uppsala, Sweden) kits [212]. The Thermo
ScientificTM MultiskcanTM FC Microplate Photometer plate with SkanItTM software (Ratastie,
Finland) is being used to process insulin data. Insulin analysis is taking place at the Guelph
Research and Development Centre, Ontario, Canada. Plasma glucose and insulin curves are
being plotted as the incremental change over time and iAUC is being calculated geometrically
for each participant using the trapezoid method, ignoring areas below the fasting value [213].
The Matsuda ISI (Matsuda ISIOGTT) is being calculated using the 75g OGTT derived plasma
glucose (PG) and insulin (PI) values, according to the formula by Matsuda et al. [205]: 10,000
divided by the square root of ([FPG x FPI] x [mean PG x PI insulin during OGTT]), where PG
is expressed in mg/dl (0.0551mmol/L) and PI in U/ml (6pmol/L). The early insulin secretion
index (∆PI30-0/∆PG30-0) is being calculated as the change in PI from 0 minutes to 30 minutes
divided by the change in PG over the same period [214].
5.4.13.3 Biomarkers of Adherence
The detection of SSB and water intake through 13C/12C ratios in serum fatty acids (e.g.
palmitoleic acid) derived from the de novo lipogenesis (DNL) of added sugars from sugar cane
(sucrose) or corn (high fructose corn syrup [HFCS]) is being assessed using a novel gas
chromatography, compound specific isotope ratio mass spectrometry (GC-CSIRMS) method at
the Department of Nutritional Sciences, University of Toronto. SSB and water intake adherence
is also being verified through an increase or decrease, respectively, of urinary fructose as
measured by gas chromatography mass spectrometry (GS-MS) [215]. NSB adherence is being
determined in urinary sucralose or Ace-K excretion through liquid chromatography with mass
spectrometry using electrospray ionization (LC-ESI-MS/MS) [216].
5.4.14 Antibiotic Use
Antibiotic use during an intervention phase is going to require a 30-day washout period at the
completion of the antibiotic course and a restart of that intervention phase. If the antibiotic
course is being taken during the washout phase, a 30-day additional washout is going to be
instituted.
112
5.4.15 Compensation
Participants are being provided monetary compensation for each study visit at the end of each
intervention phase. If a participant withdraws from the study, compensation is only being
provided for each visit that was completed. Participants are also being reimbursed for any study-
related transportation costs.
5.4.16 Statistical Analyses
An intention to treat principle using repeated measures mixed effects models in STATA 14
(StataCorp, Texas, USA) is being used to analyze the data. Sensitivity analysis is being
performed on the basis of complete data availability for primary endpoints. A
separate sensitivity analysis is being performed on the basis of antibiotic use during the trial.
For all primary and secondary outcomes effect modification by sex is being explored. The
truncated Benjamini-Hochberg false discovery rate controlling method with parallel gatekeeping
procedure is being used to correct for multiple comparisons for all primary and secondary
outcomes if at least one primary outcome reaches significance. If none of the primary outcomes
reach significance, the secondary outcomes are going to be analyzed as exploratory variables
with no adjustment for false discovery rate [201].
5.4.16.1 Subgroup Analysis
A priori subgroup analysis is being conducted by age, sex, ethnicity, baseline BMI, baseline
waist circumference, baseline FPG, baseline 2h-PG (75g OGTT), baseline iAUC, medication
use, sweeteners consumed from study beverages in the NSB arm, and background sweetener use.
5.4.17 Adverse Effects
Participants are being requested to report any adverse effects of the intervention (study
beverages) or visit examinations (75g OGTT solution, blood draw, stool sample collection)
immediately to study staff. No adverse effects are anticipated from the study drinks. Participants
may experience nausea after consuming the glucose drink, anxiety over having their blood drawn
and possible contamination of skin, clothing or food from the stool sample collection. To
mitigate these risks, participants are being given detailed instructions on the OGTT and blood
draw and proper collection of all bodily samples.
113
5.5 Results
5.5.1 Participant Flow
During the 18-month recruitment period, a total of 1086 individuals contacted the site to inquire
about the study. Of these individuals, 260 passed the telephone screen and were scheduled to
come in for the consent review. Main reasons for ineligibility from the phone screen included
consuming <1 355mL can SSB per day (n=54); low BMI (n=50); self-reported disease or illness
(n=48); did not have a primary care physician (n=28); or were smokers, recreational drug users
or heavy alcohol users (>3 drinks/day) (n=24). A total of 155 individuals consented and
completed the in-person screen, of which 141 were eligible and scheduled for their first visit.
Due to study withdrawal (n=36) and lost-to-follow up (n=22) a total of 81 participants were
randomized after completing visit one. A diagram of the flow of participants from enrollment to
randomization, following the Consolidated Standards of Reporting Trials (CONSORT) [217] is
shown in Figure 5.2.
5.5.2 Baseline Characteristics
Baseline characteristics of randomized participants in the STOP Sugars NOW trial are displayed
in Table 5.6. Mean age is 42 years. Forty-eight percent of participants are male and 52% are
female. Participants are predominately mixed (23.5%) and European (19.8%) decent with over
half born in Canada (59.3%). Among those who immigrated, the average year of immigration
was 1996. Many participants work ≥32 hours per week (50.6%), have an undergraduate degree
(33.3%) and report themselves as non-alcohol consumers (25.9%). Although subjects are mainly
obese (BMI 33.8kg/m2) with an elevated waist circumference (108.7cm), the mean blood
pressure level is normal (116.1/75.9), with only 6 participants (7%) on blood pressure lowering
medications. SSB intake (collected during the run-in phase), revealed Coke (43.2%) as the most
popular beverage consumed, followed by Canada Dry Ginger Ale (25.9%), and an overall mean
intake of two 355mL cans per day. Projected NSB intake, therefore, indicates that an
overwhelming majority of participants are going to be consuming a blend of aspartame and Ace-
K (95%). Most subjects indicated they did not consume any NNSs (60.5%), over the past 6
months, at their first visit. Of those who did, the most common source of NNS intake was from
NSBs (24.7%).
114
5.6 Discussion
The STOP Sugars NOW trial successfully recruited and enrolled 81 participants who are
receiving SSBs, NSBs and water in randomized order over the course of 4-weeks each. This
study is the first RCT to investigate how NSBs compare to water as a substitution for SSBs in
their effect on diversity of the gut microbiome and glucose tolerance.
This study overcomes several limitations of previous studies in the substitution of NSBs for
SSBs. It includes a “real-world” approach by assessing intake of NSBs based on products
available in the market as opposed to the addition of NNSs to beverages. Most NSBs are sold as
blends of NNSs to enhance the palatability of the product [91]. Therefore, adding them to
beverages as single-type sweeteners, often in amounts higher than those found on store shelfs,
prevents conclusions to be made on their efficacy as SSB replacements based on “real world”
intakes. Several studies have also failed to track the type and quantity of NNSs consumed in
beverages or investigate NNS intake through foods and beverages [156, 161-164, 168]. By
limiting NNS intake to NSBs and tracking intake through beverage logs, returned cans and
urinary biomarkers, possible deleterious (or beneficial) effects of NNS types or quantities can be
determined. Comparing NSBs for SSBs allows for the displacement of energy by the
comparator, which is often overlooked in RCT syntheses, including a recent WHO-
commissioned review [21]. Comparing NSBs for water (the “standard of care” replacement
beverage for SSBs), is going to clarify if NSBs are like water in their effect on gut bacteria and
diabetes risk.
Despite the strengths of this trial, there are challenges and limitations. Study length may decrease
adherence as the use of a crossover design requires participant involvement for a minimum of 5.5
months. Issues around antibiotic use; scheduling due to vacations; religious customs (fasting);
and changes in employment, school and family dynamics have arisen. These conflicts often lead
to an increase in study length through the extension of a washout phase to help participants
adhere to consuming their study beverages. Although our study population includes individuals
at risk for T2DM, we do not know if those living with the disease are going to see improvement
in glucose control when replacing their usual SSB intake with NSBs at “real-world” intakes.
Furthermore, even though we are assessing dietary intake through one of the best available
115
methods for measuring “free-living” food intake (weighted 3DDRs), it is possible that dietary
compensation resulting from displaced energy intake through NSBs or water, may not be
captured by these records, as dietary intake is not being assessed at the beginning of the 2nd and
3rd intervention phases.
In addition to this trial, a few others are currently investigating the effect of NNSs on outcomes
of cardiometabolic risk and/or gut microbiome. These trials include the “Microbiome and Non-
caloric Artificial Sweeteners in Humans” trial (ClinicalTrials.gov, NCT03708939), in which
participants are receiving NNSs compared to glucose on outcomes of glycemic control and gut
microbiome; the “Effects of Non-nutritive Sweeteners on the Composition of the Gut
Microbiome” trial (ClinicalTrials.gov, NCT02877186) in which participants are receiving NSBs
compared to water on outcomes of gut microbiome; and the “Effect of Non-nutritive Sweeteners
of High Sugar Sweetened Beverages on Metabolic Health and Gut Microbiome” trial
(ClinicalTrials.gov, NCT03259685) in which participants are receiving NSBs compared to SSBs
on outcomes of metabolic syndrome parameters and gut microbiome. Although these trials are
going to help strengthen the evidence base on the effect of NNSs on cardiometabolic risk and gut
microbiome, they do not assess the effect of NNSs through all 3 of our prespecified substitutions:
intended substitution with caloric displacement, “standard of care” substitution with caloric
displacement, and as a matched substitution without caloric displacement.
5.7 Conclusion
Although intakes of added sugars are declining in North America, they are still being consumed
above targets with SSBs being the most important source of added sugars [66, 218]. As a result,
strategies targeting the substitution of SSBs is an essential recommendation by authoritative
bodies. NSBs provide a viable alternative, yet their recommendation in dietary and clinical
practice guidelines are inconsistent [11-13, 15-18, 142], and concerns exists that they adversely
affect glucose tolerance by altering the composition of the gut microbiome [80, 81]. This unique
trial is going to inform guidelines on the usefulness of NSBs as SSB replacements; aid in
knowledge translation related to the health effects of NSBs for SSBs; and improve health
outcomes by educating healthcare providers and patients, stimulating industry innovation, and
guiding future research design.
116
5.8 Acknowledgements and Funding
A sincere thanks to all study participants; student volunteers; CNRFMC colleagues, including
the SMH Core Lab and MRS Research Centre, BBDC and the Department of Nutritional
Sciences (University of Toronto) who are making this study possible.
This work is being supported by the Canadian Institutes of Health Research (funding reference
number, 129920) through the Canada-wide Human Nutrition Trialists’ Network (NTN). The
Diet, Digestive tract, and Disease (3-D) Centre, funded through the Canada Foundation for
Innovation (CFI) and the Ministry of Research and Innovation’s Ontario Research Fund (ORF),
is providing the infrastructure for the conduct of this project. NM and SAC are being supported
by the Toronto 3D Internship and CIHR Canada Graduate Scholarship - Master's (CGS-M)
awards. NM is also receiving funding through the St. Michael’s Hospital Research Training
Centre Scholarship. JLS was funded by a PSI Graham Farquharson Knowledge Translation
Fellowship, Diabetes Canada Clinician Scientist award, CIHR INMD/CNS New Investigator
Partnership Prize, and Banting & Best Diabetes Centre Sun Life Financial New Investigator
Award. EMC holds the Lawson Family Chair in Microbiome Nutrition Research at the
University of Toronto. None of the sponsors have a role in any aspect of the present study,
including the design and conduct of the study; collection, management, analysis, and
interpretation of the data; and the preparation, review and approval of the manuscript or decision
to publish
5.9 Statement of Contribution
Néma McGlynn contributed to ethic submissions; provided input on how to build the trial
database in REDCap and participant recruitment; secured and trained study volunteers;
performed telephone and in-person screens; assisted in securing study materials; organized the
process for the purchase and delivery of study beverages; was a key player in the management
and execution of participant study visits; organized and managed the collection and storage of
participant data; helped set up a system for participant reminders; and performed all of the data
cleaning and analysis of baseline characteristics.
117
Visits
Visits
Visits
Visits
1Figure 5.1: Study Design1
NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages 1All possible groups not shown.
Anal4
1
5
3
S2
B
6
4
Group 1
Group 2
NSB
Group 3
SSB
NSB
Water Water
Washout
Washout
Washout
SSB
NSB
Water
Washout
Washout
Washout
SSB
NSB
Water
Randomization
4 weeks
≥4 weeks
4 weeks
≥4 weeks
4 weeks
164 Days (5.5 months)
164 Days (5.5 months)
Analysis
Analysis
1
1
5
5
3
3
2
2
6
6
s)
≥2 weeks
≥2 weeks
≥2 weeks
≥2 weeks
Run-in phase
Run-in phase
Run-in phase
Run-in phase
118
1Table 5.1: Study Exclusion Criteria1
Age between 18 and 75 years
BMI ≥23kg/m2 for Asian individuals and ≥25kg/m2 other individuals
Waist circumference ≥94cm in men, ≥80cm in women in Europid, Sub-Saharan African, Eastern Mediterranean, and
Middle Eastern individuals; ≥90cm in men and ≥80cm in women for South Asian, Chinese, Japanese, and South and
Central American individuals
Not regularly drinking SSBs (≥1 355mL can serving per day)2
Self-reported pregnant or breast-feeding females, or women planning on becoming pregnant throughout study duration
Regular medication use that have a clinically relevant effect on the primary outcomes (except birth control and PRN meds
such as Advil, Tylenol, etc.)
Antibiotic use in the last 3 months
Use of CAM deemed inappropriate by investigators
Self-reported diabetes
Self-reported hypertension (or systolic blood pressure ≥160mmHg or diastolic blood pressure ≥100mmHg [219])
Self-reported polycystic ovarian syndrome
Self-reported cardiovascular disease
Self-reported gastrointestinal disease
Previous bariatric surgery
Self-reported liver disease
Self-reported hyperthyroidism or hypothyroidism
Self-reported kidney disease
Self-reported chronic infection
Self-reported lung disease
Self-reported cancer/malignancy
Self-reported schizophrenia spectrum and other psychotic disorders, bipolar and related disorders, and dissociative disorders
Major surgery in the last 6 months
Other major illness or health-related incidence within the last 6 months
Smokes cigarettes
Regular recreational drug users
Heavy alcohol use (>3 drinks/day)
Does not have a primary care physician
Participation in any trials within the last 6 months or planning on participating in other trials for the duration of this study
Individuals planning on making dietary or physical activity changes throughout study duration
BMI, body mass index; CAM, complementary or alternative medicine; SSBs, sugar-sweetened beverages 1Disease exclusions will be based upon self-reported diagnosis. 2SSBs include sodas and soft drinks that contain at least 50 kcal per 8-oz serving. They do not include fruit drinks or
100% fruit juice, sports or energy drinks, sweetened iced tea or coffee, or homemade SSBs such as frescas or juices.
119
2Table 5.2 Power Calculation for Primary Outcomes Between Water and NSB Arms
Outcome
Type Outcome Mean Change SD Correlation N Alpha*
Power
(Beta)
N (corrected for
20% loss to follow
up)
Primary Outcomes Family
Primary
Microbiome
UniFrac
Distance
0.04 0.07 0.7 60 0.0375 98% 75
Primary Glucose iAUC
44.81
mmol/L/min
(20%)
113
mmol/L
/min
0.7 60 0.0375 89% 75
120
3Table 5.3: Latin Square Randomization for Three Group Crossover Study1
Intervention Phase 1 Intervention Phase 2 Intervention Phase 3
Sequence Group 1 A B C
Sequence Group 2 B C A
Sequence Group 3 C A B
Sequence Group 4 B A C
Sequence Group 5 A C B
Sequence Group 6 C B A 1Rows represent all possible sequences and columns represent different intervention phase.
4Table 5.4: Complete List of All Available Study Beverages
Ace-K, acesulfame potassium; NSBs, non-nutritive sweetened beverages; SSBs, sugar-sweetened beverages
SSBs group
355 ml can
42 grams of sugars
NSBs group
355 ml can
0 grams of sugar
Water group
355 ml bottle or
can
0 grams of sugar
Coca-Cola
Pepsi
Canada Dry Ginger
Ale
Schweppes Ginger
Ale
Mountain Dew
Sprite
7UP
Orange Crush
Diet Coke (Aspartame, Ace-K)
Coca Cola Zero (Aspartame, Ace-K)
Diet Pepsi (Aspartame, Ace-K)
Diet Canada Dry Ginger Ale (Aspartame, Ace-
K)
Diet Schweppes Ginger Ale (Ace-K, Sucralose)
Diet Mountain Dew (Aspartame, Ace-K,
Sucralose)
Sprite Zero (Aspartame, Ace-K)
Diet 7UP (Aspartame, Ace-K)
Diet Orange Crush (Sucralose)
Carbonated or
still water
121
5Table 5.5: Participant Visit Schedule
Phase1 Screening
(0 wks)
Run-in
(2 weeks)
Intervention
phase 1 (4 weeks)
Washout
(4 weeks)
Intervention
phase 2 (4 weeks)
Washout
(4 weeks)
Intervention
phase 3 (4 weeks)
Study Visit 0 No visit Week 0
(Start)
Week 4
(End)
No visit Week 0
(Start)
Week 4
(End)
No visit Week 0
(Start)
Week 4
(End)
Written
Informed
Consent
Provided
x
Anthropometric measures
Height x
Weight x x x x x x x
Waist
circumference
x x x x x x x
Blood pressure x x x x x x x
Biochemical measures
Fecal sample x x x x x x
24-hour urine
collection
(biomarkers
for adherence)
x x x x x x
Blood sample
drawn
x x x x x x
75g OGTT x x x x x x
Questionnaires
Personal
information
x
Demographic
Data
x
Medical
History
x
Case Report
Form (e.g.
medications,
physical
activity)
x x x x x x x
Beverage log x x x x x x x x x x
3DDR
(adherence to
background
diet)
x x x x
3DDR, 3-day diet record; OGTT, oral glucose tolerance test 1Length in weeks
122
2Figure 5.2: Trial Flow: Screening and Randomization
123
6Table 5.6 Baseline Characteristics of Randomized Participants1
Variable Value
Anthropometry Mean ± SD
Age (years) 41.8 ± 12.9
Females, n (%) 42 (51.9)
Height, cm 167.3 ± 10.6
Weight, kg 94.2 ± 18.9
BMI, kg/m2 33.8 ± 6.7
Waist circumference, cm 108.7 ± 13.4
Systolic blood pressure 116.1 ± 12.6
Diastolic blood pressure 75.9 ± 9.1
Resting pulse 72.9 ± 10.3
SSB preference n (mean intake/day)
7-UP 3 (2)
Coke 35 (2)
CD Ginger Ale 21 (1.5)
Orange Crush 3 (1)
Pepsi 11 (2)
Sprite 7 (2)
Unknown 1 (0)
Projected NSB intake (NNS blends) n (%)
Diet 7-UP (Asp & Ace-K) 3 (3.7)
Diet Coke (Asp & Ace-K) 16 (19.8)
Coke Zero (Asp & Ace-K) 19 (23.5)
CD Diet Ginger Ale (Asp & Ace-K) 21 (25.9)
Diet Orange Crush (Sucralose) 3 (3.7)
Diet Pepsi (Asp & Ace-K) 11 (13.6)
Sprite Zero (Asp & Ace-K) 7 (8.6)
Unknown (Unknown) 1 (1.2)
Non-nutritive sweetener intake n (%)
NSBs 20 (24.7)
NNS containing foods 3 (3.7)
Table-top NNS 9 (11.1)
None 49 (60.5)
Multiple categories selected 11 (13.6)
Ethnicity n (%)
Aboriginal 2 (2.5)
European 16 (19.8)
African/Caribbean 5 (6.2)
Scottish/Irish 7 (8.6)
English 5 (6.2)
Latin American 5 (6.2)
Indian 4 (4.9)
East Asian 6 (7.4)
South East Asian 6 (7.4)
124
Other 6 (7.4)
Mixed 19 (23.5)
Education level n (%)
High School 18 (22.2)
College or Diploma 18 (22.2)
Undergraduate Degree 27 (33.3)
Graduate Degree 16 (19.8)
Other 2 (2.5)
Work status n (%)
FT (≥ 32 hrs/wk) 41 (50.6)
PT (≤ 32 hrs/wk) 13 (16)
Casual 6 (7.4)
Stay at home parent 6 (7.4)
FT student 3 (3.7)
Disability 3 (3.7)
Other 9 (11.1)
Alcohol intake n (%)
None 21 (25.9)
1-2 times per year 9 (11.1)
Every 2-3 months 16 (19.8)
1-2 times per month 16 (19.8)
1-2 times per week 16 (19.8)
Daily 3 (3.7) Ace-K, acesulfame potassium; CD, Canada Dry; FT, full-time; NNS, non-nutritive sweetener; NSBs, non-
nutritive sweetened beverage; PT, part-time; SSB, sugar-sweetened beverage. 1N=81
125
Chapter 6: General Discussion
6.1 Summary
To determine whether non-nutritive sweetened beverages (NSBs) improve body weight and
cardiometabolic risk factors similar to water in their substitution for sugar-sweetened beverages
(SSBs), we conducted a systematic review and network meta-analysis of randomized controlled
trials (RCTs) assessing the effect of NSBs, as a replacement for SSBs, through 3 prespecified
substitutions: NSBs for SSBs (intended substitution with caloric displacement), water for SSBs
(“standard of care” substitution with caloric displacement) and NSBs for water (matched
substitution without caloric displacement) on measures of body weight and cardiometabolic risk
factors. Fourteen RCTs (n=1530), predominantly in people with overweight/obesity at risk for or
with diabetes, comparing NSBs, SSBs and/or water on the primary measure of body weight and
secondary outcomes of other measures of adiposity, glycemic control, blood lipids, blood
pressure, non-alcoholic fatty liver disease (NAFLD) and uric acid were identified. The
substitution of NSBs for SSBs improved body weight, body mass index (BMI), body fat,
triglycerides (TGs) and intra-hepatocellular lipid (IHCL). The substitution of water for SSBs
improved only uric acid and there was no effect of the substitution of NSBs for water except for
a small increase in hemoglobin A1c (HbA1c). The certainty of the evidence was moderate
(NSBs for SSBs) and low (water for SSBs and NSBs for water) for the primary outcome of body
weight and ranged from low to high for all other outcomes across all substitutions. Downgrades
were mainly made for imprecision. The available evidence supports the use of NSBs as an
alternative to water when used as a replacement for SSBs in overweight/obese people at risk for
or with diabetes. Due to imprecision in the estimate and concerns that NSBs adversely affect
glucose tolerance through compositional changes of the gut microbiome, there is a need for more
high-quality trials. To address this concern, we provided a design, rationale and baseline
characteristics report for a RCT investigating the effect of NSBs, as a replacement for SSBs,
through our 3 prespecified substitutions (NSBs for SSBs, water for SSBs and NSBs for water) on
gut microbiome and glucose tolerance over 4-weeks in overweight/obese participants who are
regular SSB drinkers: Strategies To OPpose SUGARS with Non-nutritive sweeteners Or Water
trial (STOP Sugars NOW). Baseline characteristics of study participants revealed an average
intake of two 355mL cans of SSBs per day, the majority of which were consuming Coke (n=37,
126
45.7%), followed by Canada Dry Ginger Ale (n=27, 33.3%). Projected NSB intake, based on the
brand-matched NSB options, indicates a NNS-blend of primarily aspartame and acesulfame
potassium (Ace-K) (n=77, 95%). This trial will be the first to determine if substituting NSBs for
SSBs, water for SSBs and NSBs for water impacts glucose control through changes in the gut
microbiota.
6.2 Strengths and Limitations
Strengths of the network meta-analysis are:
1. Accounting for the nature of the comparator to allow for the displacement of energy from
SSBs by NSBs in comparison to the preferred replacement beverage of water.
2. Use of network meta-analysis allowed for the quantitative comparison of our 3
prespecified beverage interventions that were not directly compared in trials. The
network meta-analysis then strengthened the evidence base by analyzing both direct and
indirect evidence collectively, thereby providing more precise estimates of the effect of
each beverage comparison on our outcomes.
3. A rigorous search and selection process of the available literature examining the effect of
NSBs, SSBs and water on outcomes of body weight and cardiometabolic risk factors.
4. Inclusion of only RCTs, a design that provides the greatest protection against bias.
5. Utilization of the GRADE approach to assess the certainty of our estimates.
Strengths of the STOP Sugars NOW trial are:
1. Implementation of the Benjamini-Hochberg procedure for the power calculation
which accounts for the adjustment required for multiple testing of two primary
outcomes (glucose iAUC and Microbiome UniFrac Distance).
2. Use of a “real-world” approach by assessing intake of NSBs based on products
available in the market, and through the equal displacement of SSBs with NSBs and
water based on each participant’s average intake at baseline.
3. Limiting NNS intake to NSBs and tracking adherence through urinary biomarkers to
determine possible deleterious (or beneficial) effects of NNS types or quantities.
127
4. Comparing NSBs to water in their substitution for SSBs which allows for the
displacement of energy by the comparator. This comparison will clarify if NSBs are
like water in their effect on gut bacteria and diabetes risk.
5. Although we will not be able to determine the effect of single NNSs on gut
microbiota changes and glucose tolerance, we are taking a pragmatic approach by
looking at “real-world” intakes of NNSs from NSBs, of which the majority exist as
blends of different NNSs.
Limitations of the network meta-analysis are:
1. Downgrades for evidence of serious inconsistency on the secondary adiposity outcome of
waist circumference in the substitution of NSBs for SSBs; on the secondary
cardiometabolic outcomes of HbA1c and FPI in the substitution of water for SSBs; and in
the primary outcome of body weight and several secondary adiposity and
cardiometabolic outcomes in the substitution of NSBs for water (waist circumference,
HbA1c, FPI, HOMA-IR, Non-HDL-C and TGs).
2. Downgrades for evidence of serious indirectness, as only one RCT or less of direct
comparisons was available for several secondary adiposity and cardiometabolic outcomes
in the analyses for the substitution of NSBs for SSBs (waist circumference, HbA1c, 2h-
PG, ALT and AST), water for SSBs (waist circumference, HbA1c, 2h-PG, FPI, HOMA-
IR, IHCL, ALT, AST and uric acid) and NSBs for water (IHCL, ALT, AST and uric
acid).
3. Downgrades for evidence of serious imprecision in the primary outcome of body weight
and several secondary adiposity and cardiometabolic outcomes in the substitution of
NSBs for SSBs (BMI, waist circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-HDL-C,
TGs, HDL-C, TC, SBP, DBP, IHCL, ALT, AST and uric acid); water for SSBs (body
weight, BMI, waist circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-HDL-C, TGs,
HDL-C, TC, SBP, DBP, IHCL, ALT, AST and uric acid); and NSBs for water (body
weight, BMI, waist circumference, HbA1c, 2h-PG, FPI, LDL-C, Non-HDL-C, TGs, TC,
SBP, DBP, IHCL, ALT, AST and uric acid) crossed our prespecified minimally
important differences. Accordingly, we downgraded the evidence in all cases for serious
imprecision.
128
4. The inability to assess for publication bias or conduct subgroup analyses as there were
less than 10 direct RCT comparisons per outcome for each of the 3 prespecified
substitutions. We were also unable to conduct sensitivity analyses in the network as
several of the RCTs contained multiple comparison arms.
Limitations/challenges of the STOP Sugars NOW trial are:
1. The length of the trial may decrease adherence as the use of a crossover design requires
participant involvement for a minimum of 5.5 months, which may further increase due to
each participants life’s circumstances (vacations; religious customs (fasting); and changes
in employment, school and family dynamic) and possible antibiotic use.
2. Our study population only includes individuals at risk for T2DM. Therefore, we won’t
know if those living with the disease will see improvement in glucose control or
microbiome diversity when replacing their usual SSB intake with NSBs at “real-world”
intakes.
3. Dietary compensation resulting from displaced energy intake through NSBs or water may
not be captured by weighted 3-day diet records (3DDRs), as dietary intake is not being
assessed at every study visit (only visits 1, 2, 4 and 6).
6.3 Clinical Implications
Current dietary and clinical practice guidelines are inconsistent in their recommendations for
NSBs [11-13, 15, 16, 18], even though results from SRMAs of RCTs consistently show non-
significant or beneficial effects of NSBs on outcomes of cardiometabolic risk in individuals with,
and without diabetes [19-21, 23-26, 138]. In our network meta-analysis, it was demonstrated that
when NSBs are substituted for SSBs, improvements are seen on body weight and
cardiometabolic outcomes over the moderate term with no evidence of harm. Although we did
not see benefits with the substitution of NSBs for SSBs on measures of glucose control, the
benefits observed on the primary outcome of body weight and other secondary outcomes of
adiposity, lipids and IHCL suggest that longer-term substitutions of NSBs for SSBs lead to
improvements in glycemic response with subsequent reduction in, and management of, diabetes
risk. Even though the findings from this NMA are limited to NSBs, they are of relevance to
NNSs in general as NNSs are added to several foods and beverages by consumers and product
129
developers in an attempt to reduce calories [85]. Most importantly, of all product categories
containing NNSs as a single food matrix, NSBs are globally the most consumed source [85].
Given the current concern with dietary sugars in the obesity and diabetes epidemic, the clinical
implications of our findings indicate that NSBs are a viable alternative to SSBs and are
comparable to water in their effect on body weight, glycemic control and cardiometabolic risk
factors. The use of NSBs as a displacement of energy from SSBs, however, should be part of a
broader strategy where guidance is given on replacements for all sources of excess caloric intake
from sugars. Emphasis should be placed on the overall quality of the diet by the incorporation of
a dietary pattern that best aligns with an individual’s values and preferences to help strengthen
adherence over the long-term [12].
6.4 Future Directions
The findings of this network meta-analysis are highly relevant for informing guidance on the role
of NSBs in the context of excess sugar intake from SSBs. While our results demonstrate that
NSBs have no adverse effect on blood glucose and insulin regulation in adults living with and
without diabetes, we were unable to establish any benefit on these outcomes when NSBs
substituted SSBs. Furthermore, as imprecision was detected in the network estimate for HbA1c,
2h-PG, and fasting insulin among all three beverage comparisons, there is a need for larger and
longer RCTs of a year or more to elucidate the potential benefits of NSBs on glycemic control.
Our inability to assess for publication bias or do subgroup analyses as there were less than 10
direct RCT comparisons per outcome for each of the 3 prespecified substitutions, and the sparse
network meta-analysis for outcomes of IHCL, ALT, AST and uric acid, resulting in downgrades
for serious imprecision among all 3 beverage comparisons further emphasize the need for more
high-quality RCTs. With additional RCTs we would be able to determine if the effect sizes
observed were overestimated and with false-positive results; if effect modifiers such as weight,
age, health status, background diet or type of NNS in the beverages could have influenced the
results; and if the results for outcomes with limited studies in the network were spurious.
The STOP Sugars NOW trial will not only add to the evidence-base on the effect of our 3
prespecified beverages on metabolic outcomes but will add to the limited evidence concerning
NNSs on glucose tolerance through possible changes in the gut microbiome based on “real-
world” intakes. If changes in the gut microbiome are observed with subsequent glucose
130
intolerance, additional well-designed RCTs will be required to elucidate a causal biological
plausibility of this relationship, while considering the nature of the comparator.
Considering 95% of participants are projected to consume a blend of aspartame and Ace-K,
determinations could also be made on the potential benefits or harms of this sweetener blend in
the NSBs consumed. Future trials could investigate the impact of these NNSs in certain foods or
other sweeteners that are gaining in popularity such as stevia. As the glucose units of steviol
glycosides, the sweet tasting compounds of the stevia plant, are cleaved off by bacteria in the
colon [91], a future study could mimic our trial design to investigate any potential effects this
NNS may have on glucose tolerance through changes in the gut microbiome, thereby adding to
the limited evidence of this sweetener on gut health [130].
Chapter 7: Conclusions
Conclusions:
The findings from this thesis demonstrated the following:
1. In a systematic review and network meta-analysis of 14 randomized controlled trials of
21 trial comparisons involving 1530 predominantly overweight or obese adult
participants, at risk for or with diabetes, over the moderate term, showed that substituting
non-nutritive sweetened beverages (NSBs) for sugar-sweetened beverages (SSBs), water
for SSBs and NSBs for water did not adversely affect cardiometabolic risk. Benefits were
observed for outcomes of body weight, BMI, body fat, triglycerides, and
intrahepatocellular lipids with the substitution of NSBs for SSBs. The overall certainty of
the evidence was rated as low for the substitutions of water for SSBs and NSBs for water,
and moderate for the substitution of NSBs for SSBs, with downgrades mainly for
imprecision, indicating a need for more high-quality trials to improve our confidence in
the estimates of effect.
2. The STOP Sugars NOW trial successful recruited 81 participants to receive SSBs, NSBs
and water in randomized order over the course of 4-weeks each. Baseline characteristics
indicated that participants consumed an average of two 355mL cans of SSBs per day with
131
a projected NNS intake of a blend of aspartame and Ace-K. This trial will address the
limitations of the trials used in our network meta-analysis while adding the necessary
evidence to improve our confidence in the estimates of effect of our 3 prespecified
substitutions (NSBs for SSBs, water for SSBs and NSBs for water) on body weight and
cardiometabolic risk. It will also help determine if NSBs for SSBs, water for SSBs and
NSBs for water, negatively affect the gut microbiome with associated developments of
adverse glucose tolerance.
This is the first network meta-analysis to investigate the role of NSBs compared to water as a
displacement of energy for SSBs on outcomes of body weight and cardiometabolic risk factors.
Our results are consistent with the previous meta-analyses of RCTs assessing the effect of NNSs
on body weight and cardiometabolic risk. The STOP Sugars NOW trial is the first RCT to
investigate how NSBs compare to water as a substitution for SSBs in their effect on diversity of
gut microbiome and glucose tolerance.
132
References
1. World Health Organization. Obesity and overweight. 2018; Available from: https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight.
2. Health Canada. Tackling obesity in canada: obesity and excess weight rates in canadian adults. 2018; Available from: https://www.canada.ca/en/public-health/services/publications/healthy-living/obesity-excess-weight-rates-canadian-adults.html.
3. O’Neill, M., K. Kornas, and L. Rosella, The future burden of obesity in Canada: a modelling study. Canadian Journal of Public Health, 2019. 110(6): p. 768-778.
4. The GBD 2015 Obesity Collaborators, Health effects of overweight and obesity in 195 countries over 25 years. New England Journal of Medicine, 2017. 377(1): p. 13-27.
5. World Health Organization. Global report on diabetes. 2016; Available from: https://apps.who.int/iris/bitstream/handle/10665/204871/9789241565257_eng.pdf;jsessionid=BD0EB38572D0002E58B5706B4FCF2E67?sequence=1#targetText=GlOBAl%20BurDEN,8.5%25%20in%20the%20adult%20population.
6. Malik, V.S., et al., Sugar-sweetened beverages and weight gain in children and adults: a systematic review and meta-analysis. Am J Clin Nutr, 2013. 98(4): p. 1084-102.
7. Imamura, F., et al., Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: systematic review, meta-analysis, and estimation of population attributable fraction. BMJ, 2015: p. h3576.
8. Jayalath, V.H., et al., Sugar-sweetened beverage consumption and incident hypertension: a systematic review and meta-analysis of prospective cohorts. Am J Clin Nutr, 2015. 102(4): p. 914-21.
9. Xi, B., et al., Sugar-sweetened beverages and risk of hypertension and CVD: a dose-response meta-analysis. Br J Nutr, 2015. 113(5): p. 709-17.
10. World Health Organization. Guideline: Sugars intake for adults and children. 2015; Available from: https://apps.who.int/iris/bitstream/handle/10665/149782/9789241549028_eng.pdf?sequence=1.
11. Health Canada. Canada's dietary guidelines for health professionals and policy makers. 2018; Available from: https://food-guide.canada.ca/static/assets/pdf/CDG-EN-2018.pdf.
12. Diabetes Canada, Clinical practice guidelines committees. Canadian Journal of Diabetes, 2018. 42.
13. United States Department of Agriculture. Scientific report of the 2015 dietary guidelines advisory committee. 2015.
14. Australian Government, Australian dietary guidelines, D.o.H.a. Ageing, Editor. 2013. 15. American Diabetes Association. Standards of medical care in diabetes - 2020. The journal of
Clinical and Applied Research and Education 2020; Available from: https://care.diabetesjournals.org/content/diacare/suppl/2019/12/20/43.Supplement_1.DC1/Standards_of_Care_2020.pdf.
16. Diabetes UK. Evidence-based nutrition guidelines for the prevention and management of diabetes. 2018; Available from: https://diabetes-resources-production.s3.eu-west-1.amazonaws.com/resources-s3/2018-03/1373_Nutrition%20guidelines_0.pdf.
17. Mann, J.I., et al., Evidence-based nutritional approaches to the treatment and prevention of diabetes mellitus. Nutr Metab Cardiovasc Dis, 2004. 14(6): p. 373-394.
18. Johnson, R.K., et al., Low-calorie sweetened beverages and cardiometabolic health: A science advisory from the american heart association. Circulation, 2018. 138(9).
133
19. Azad, M.B., et al., Nonnutritive sweeteners and cardiometabolic health: a systematic review and meta-analysis of randomized controlled trials and prospective cohort studies. Canadian Medical Association Journal, 2017. 189(28): p. E929-E939.
20. Rogers, P.J., et al., Does low-energy sweetener consumption affect energy intake and body weight? A systematic review, including meta-analyses, of the evidence from human and animal studies. International Journal of Obesity, 2016. 40(3): p. 381-394.
21. Toews, I., et al., Association between intake of non-sugar sweeteners and health outcomes: systematic review and meta-analyses of randomised and non-randomised controlled trials and observational studies. BMJ, 2019. 364: p. k4718.
22. Greenwood, D.C., et al., Association between sugar-sweetened and artificially sweetened soft drinks and type 2 diabetes: systematic review and dose–response meta-analysis of prospective studies. British Journal of Nutrition, 2014. 112(5): p. 725-734.
23. Miller, P.E. and V. Perez, Low-calorie sweeteners and body weight and composition: a meta-analysis of randomized controlled trials and prospective cohort studies. The American Journal of Clinical Nutrition, 2014. 100(3): p. 765-777.
24. Santos, N.C., et al., Metabolic effects of aspartame in adulthood: A systematic review and meta-analysis of randomized clinical trials. Crit Rev Food Sci Nutr, 2018. 58(12): p. 2068-2081.
25. Laviada-Molina, H., et al., Effects of nonnutritive sweeteners on body weight and BMI in diverse clinical contexts: Systematic review and meta-analysis. Obesity Reviews, 2020.
26. Onakpoya, I.J. and C.J. Heneghan, Effect of the natural sweetener, steviol glycoside, on cardiovascular risk factors: A systematic review and meta-analysis of randomised clinical trials. European Journal of Preventive Cardiology, 2015. 22(12): p. 1575-1587.
27. Obesity Canada. Report Card on Access to Obesity Treatment for Adults in Canada. . 2019; Available from: http://obesitycanada.ca/wp-content/uploads/2019/04/OC-Report-Card-2019-Eng-F-web.pdf.
28. Brauer, P., et al., Recommendations for prevention of weight gain and use of behavioural and pharmacologic interventions to manage overweight and obesity in adults in primary care. Canadian Medical Association Journal, 2015. 187(3): p. 184-195.
29. International Diabetes Federation. IDF diabetes atlas: Eighth edition. 2017; Available from: https://diabetesatlas.org/IDF_Diabetes_Atlas_8e_interactive_EN/.
30. Basu, S., et al., The relationship of sugar to population-level diabetes prevalence: an econometric analysis of repeated cross-sectional data. PLoS ONE, 2013. 8(2): p. e57873.
31. Tremmel, M., et al., Economic Burden of Obesity: A Systematic Literature Review. International Journal of Environmental Research and Public Health, 2017. 14(4): p. 435.
32. Bommer, C., et al., Global Economic Burden of Diabetes in Adults: Projections From 2015 to 2030. Diabetes Care, 2018. 41(5): p. 963-970.
33. Maluly, H.D.B., et al., Low- and No- Calorie Sweeteners (LNCS): critical evaluation of their safety and health risks. Food Science and Technology, 2020.
34. Leiter, L.A., et al., Cardiometabolic Risk in Canada: A Detailed Analysis and Position Paper by the Cardiometabolic Risk Working Group. Canadian Journal of Cardiology, 2011. 27(2): p. e1-e33.
35. Alberti, K.G.M.M., et al., Harmonizing the metabolic syndrome. Circulation, 2009. 120(16): p. 1640-1645.
36. Lau, D.C.W., et al., 2006 Canadian clinical practice guidelines on the management and prevention of obesity in adults and children Canadian Medical Association Journal, 2007. 176(8): p. 1-117.
37. Hsu, W.C., et al., BMI cut points to identify at-risk Asian Americans for type 2 diabetes screening. Diabetes Care, 2015. 38(1): p. 150-8.
134
38. WHO Expert Consultation, Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet, 2004. 363(9403): p. 157-63.
39. International Diabetes Federation. The IDF consensus worldwide definition of the metabolic syndrome. 2006; Available from: https://www.pitt.edu/~super1/Metabolic/IDF1.pdf.
40. Tan, C.E., et al., Can We Apply the National Cholesterol Education Program Adult Treatment Panel Definition of the Metabolic Syndrome to Asians? 2004. 27(5): p. 1182-1186.
41. World Health Organization, Obesity: Preventing and managing the global epidemic. WHO Technical Report Series, 2000. 894(i–xii): p. 1-253.
42. Adab, P., M. Pallan, and P.H. Whincup, Is BMI the best measure of obesity? BMJ, 2018: p. k1274. 43. Yusuf, S., et al., Obesity and the risk of myocardial infarction in 27,000 participants from 52
countries: a case-control study. Lancet, 2005. 366(9497): p. 1640-9. 44. World Health Organization, Waist circumference and waist–hip ratio: report of a WHO expert
consultation. 2008: Geneva, Switzerland. 45. Malik, V.S., et al., Sugar-sweetened beverages and risk of metabolic syndrome and type 2
diabetes: A meta-analysis. Diabetes Care, 2010. 33(11): p. 2477-2483. 46. Meng Wang, M.Y., Le Fang, Ru-Ying Hu, Association between sugar-sweetened beverages and
type 2 diabetes: A meta-analysis. Journal of Diabetes Investigation, 2015. 6(3): p. 360-366. 47. Malik, V.S., et al., Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and
cardiovascular disease risk. Circulation, 2010. 121(11): p. 1356-1364. 48. Te Morenga, L.A., et al., Dietary sugars and cardiometabolic risk: systematic review and meta-
analyses of randomized controlled trials of the effects on blood pressure and lipids. The American Journal of Clinical Nutrition, 2014. 100(1): p. 65-79.
49. Malik, V.S. and F.B. Hu, Sugar-sweetened beverages and cardiometabolic health: An update of the evidence. Nutrients, 2019. 11(8): p. 1840.
50. Dimeglio, D. and R. Mattes, Liquid versus solid carbohydrate: effects on food intake and body weight. International Journal of Obesity, 2000. 24(6): p. 794-800.
51. Cassady, B.A., R.V. Considine, and R.D. Mattes, Beverage consumption, appetite, and energy intake: what did you expect? The American Journal of Clinical Nutrition, 2012. 95(3): p. 587-593.
52. Mourao, D.M., et al., Effects of food form on appetite and energy intake in lean and obese young adults. International Journal of Obesity, 2007. 31(11): p. 1688-1695.
53. Tieken, S.M., et al., Effects of solid versus liquid meal-replacement products of similar energy content on hunger, satiety, and appetite-regulating hormones in older adults. Horm Metab Res, 2007. 39(5): p. 389-94.
54. Atkinson, F.S., K. Foster-Powell, and J.C. Brand-Miller, International tables of glycemic index and glycemic load values: 2008. Diabetes Care, 2008. 31(12): p. 2281-2283.
55. Tey, S.L., et al., Effects of aspartame-, monk fruit-, stevia- and sucrose-sweetened beverages on postprandial glucose, insulin and energy intake. International Journal of Obesity, 2017. 41(3): p. 450-457.
56. J Bloomer, R., et al., Blood glucose and insulin response to artificially- and sugar-sweetened sodas in healthy men. Integrative Food, Nutrition and Metabolism, 2016. 3(1).
57. Ludwig, D.S., The glycemic index. JAMA, 2002. 287(18): p. 2414. 58. Liu, S., et al., Relation between a diet with a high glycemic load and plasma concentrations of
high-sensitivity C-reactive protein in middle-aged women. The American Journal of Clinical Nutrition, 2002. 75(3): p. 492-498.
59. Chiavaroli, L., et al., Effect of fructose on established lipid targets: a systematic review and meta-analysis of controlled feeding trials. J Am Heart Assoc, 2015. 4(9): p. e001700.
60. David Wang, D., et al., Effect of fructose on postprandial triglycerides: a systematic review and meta-analysis of controlled feeding trials. Atherosclerosis, 2014. 232(1): p. 125-33.
135
61. Choo, V.L., et al., Food sources of fructose-containing sugars and glycaemic control: systematic review and meta-analysis of controlled intervention studies. BMJ, 2018: p. k4644.
62. Wang, D.D., et al., The effects of fructose intake on serum uric acid vary among controlled dietary trials. J Nutr, 2012. 142(5): p. 916-23.
63. Chiu, S., et al., Effect of fructose on markers of non-alcoholic fatty liver disease (NAFLD): a systematic review and meta-analysis of controlled feeding trials. European Journal of Clinical Nutrition, 2014. 68(4): p. 416-423.
64. Langlois, K., et al., Change in total sugars consumption among Canadian children and adults, Statistics Canada, Editor. January 16, 2019. p. 10-19.
65. Welsh, J.A., et al., Consumption of added sugars is decreasing in the United States. Am J Clin Nutr, 2011. 94(3): p. 726-34.
66. Bailey, R., et al., Sources of added sugars in young children, adolescents, and adults with low and high intakes of added sugars. Nutrients, 2018. 10(1): p. 102.
67. Rosinger, A., et al., Sugar-sweetened beverage consumption among U.S. adults, 2011–2014. NCHS Data Brief, 2017. 270: p. 1-8.
68. Singh, G.M., et al., Global, regional, and national consumption of sugar-sweetened beverages, fruit juices, and milk: A systematic assessment of Beverage intake in 187 countries. PLoS One, 2015. 10(8): p. e0124845.
69. Popkin, B.M. and C. Hawkes, Sweetening of the global diet, particularly beverages: patterns, trends, and policy responses. The Lancet Diabetes & Endocrinology, 2016. 4(2): p. 174-186.
70. Scientific Advisory Committee on Nutrition. Carbohydrates and Health. 2015; Available from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/445503/SACN_Carbohydrates_and_Health.pdf.
71. Rippe, J.M., et al., What is the appropriate upper limit for added sugars consumption? Nutrition Reviews, 2017. 75(1): p. 18-36.
72. Colchero A, R.-D.J., Popkin BM, and Ng SW, In Mexico, evidence of sustained consumer response two years after implementing a sugar-sweetened beverage tax. Health Affairs, 2017. 36.
73. Diabetes Canada. Sugar & diabetes. 2019; Available from: https://www.diabetes.ca/en-CA/advocacy---policies/our-policy-positions/sugar---diabetes.
74. Public Health England. Sugar reduction: The evidence for action. 2015; Available from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/470179/Sugar_reduction_The_evidence_for_action.pdf.
75. Mexican Academy of Medicine, Dietary and physical activity guidelines in the context of overweight and obesity in the Mexican population, Consejo Nacional de Ciencia y Tecnologia, Editor. 2015.
76. The Supreme Council of Health Health Promotion and Non-Communicable Diseases. Qatar dietary guidelines. 2015; Available from: http://www.fao.org/3/a-az908e.pdf.
77. Ministry of Health of Brazil. Dietary guidelines for the Brazil population. 2014; Available from: https://www.paho.org/hq/dmdocuments/2015/dietary-guides-brazil-eng.pdf.
78. Istituto Nazionale di Ricerca per gli Alimenti e la Nutrizione. Linee guida per una sana alimentazione italiana. 2003; Available from: http://www.fao.org/3/a-as686o.pdf.
79. Sylvetsky, A.C. and K.I. Rother, Nonnutritive sweeteners in weight management and chronic disease: A review. Obesity (Silver Spring), 2018. 26(4): p. 635-640.
80. Hunter, S.R., et al., Low Calorie Sweeteners Differ in Their Physiological Effects in Humans. Nutrients, 2019. 11(11): p. 2717.
81. Suez, J., et al., Artificial sweeteners induce glucose intolerance by altering the gut microbiota. Nature, 2014. 514(7521): p. 181-186.
136
82. Diabetes Canada. Sugars and sweeteners. 2018; Available from: https://guidelines.diabetes.ca/docs/patient-resources/sugars-and-sweeteners.pdf.
83. Health Canada. Sugar alcohols (polyols) and polydextrose used as sweeteners in foods 2018; Available from: https://www.canada.ca/en/health-canada/services/food-nutrition/food-safety/food-additives/sugar-substitutes/sugar-alcohols-polyols-polydextrose-used-sweeteners-foods-food-safety.html.
84. Health Canada. Sweeteners definitions. 2018; Available from: http://www.inspection.gc.ca/food/requirements-and-guidance/labelling/industry/sweeteners/eng/1387749708758/1387750396304?chap=5.
85. Sylvetsky, A.C. and K.I. Rother, Trends in the consumption of low-calorie sweeteners. Physiology & Behavior, 2016. 164: p. 446-450.
86. European Food Safety Authority. Sweeteners. Available from: http://www.efsa.europa.eu/en/topics/topic/sweeteners.
87. U.S. Food and Drug Administration. High-intensity sweeteners. 2017; Available from: https://www.fda.gov/food/food-additives-petitions/high-intensity-sweeteners.
88. Health Canada. The safety of sugar substitutes. 2008; Available from: https://www.canada.ca/en/health-canada/services/healthy-living/your-health/food-nutrition/safety-sugar-substitutes.html.
89. Serra-Majem, L., et al., Ibero–american consensus on low- and no-calorie sweeteners: Safety, nutritional aspects and benefits in food and beverages. Nutrients, 2018. 10(7): p. 818.
90. World Health Organization. Alphabetical list of WHO member states. [cited 2019; Available from: https://www.who.int/choice/demography/by_country/en/.
91. Magnuson, B.A., et al., Biological fate of low-calorie sweeteners. Nutrition Reviews, 2016. 74(11): p. 670-689.
92. Health Canada. List of permitted sweeteners (lists of permitted food additives). 2019; Available from: https://www.canada.ca/en/health-canada/services/food-nutrition/food-safety/food-additives/lists-permitted/9-sweeteners.html.
93. Ahmad, S.Y., J.K. Friel, and D.S. MacKay, The effect of the artificial sweeteners on glucose metabolism in healthy adults: a randomized double-blinded crossover clinical trial. Appl Physiol Nutr Metab, 2019.
94. Nettleton, J.E., R.A. Reimer, and J. Shearer, Reshaping the gut microbiota: Impact of low calorie sweeteners and the link to insulin resistance? Physiol Behav, 2016. 164(Pt B): p. 488-493.
95. Franz, M. Amounts of sweeteners in popular diet sodas. Diabetes Self-Management 2010; Available from: https://static.diabetesselfmanagement.com/pdfs/DSM0310_012.pdf
96. Calorie Control Council. Cyclamate information center. 2018; Available from: https://www.cyclamate.org/facts.html.
97. Markets, M.a. Sugar substitutes market by type, composition, application, & by region: Forecast to 2020. 2015; Available from: https://www.marketsandmarkets.com/Market-Reports/sugar-substitute-market-1134.html.
98. Magnuson, B.A., A. Roberts, and E.R. Nestmann, Critical review of the current literature on the safety of sucralose. Food and Chemical Toxicology, 2017. 106: p. 324-355.
99. Canadian Food Inspection Agency. Labelling requirements for sweeteners and foods that contain sweeteners. 2020; Available from: https://www.inspection.gc.ca/food-label-requirements/labelling/industry/sweeteners/eng/1387749708758/1387750396304?chap=3.
100. Burke, M.V. and D.M. Small, Physiological mechanisms by which non-nutritive sweeteners may impact body weight and metabolism. Physiology & Behavior, 2015. 152: p. 381-388.
101. Dhillon, J., J.Y. Lee, and R.D. Mattes, The cephalic phase insulin response to nutritive and low-calorie sweeteners in solid and beverage form. Physiology & Behavior, 2017. 181: p. 100-109.
137
102. Just, T., et al., Cephalic phase insulin release in healthy humans after taste stimulation? 2008. 51(3): p. 622-627.
103. Swithers, S.E., A.A. Martin, and T.L. Davidson, High-intensity sweeteners and energy balance. Physiology & Behavior, 2010. 100(1): p. 55-62.
104. Tucker, R.M. and S.Y. Tan, Do non-nutritive sweeteners influence acute glucose homeostasis in humans? A systematic review. Physiol Behav, 2017. 182: p. 17-26.
105. Kojima, I., et al., Glucose-sensing receptor T1R3: A new signaling receptor activated by glucose in pancreatic β-cells. 2015. 38(5): p. 674-679.
106. Kyriazis, G.A., M.M. Soundarapandian, and B. Tyrberg, Sweet taste receptor signaling in beta cells mediates fructose-induced potentiation of glucose-stimulated insulin secretion. 2012. 109(8): p. E524-E532.
107. Nakagawa, Y., et al., Sweet taste receptor expressed in pancreatic β-cells activates the calcium and cyclic AMP signaling systems and stimulates insulin secretion. PLoS ONE, 2009. 4(4): p. e5106.
108. Margolskee, R.F., et al., T1R3 and gustducin in gut sense sugars to regulate expression of Na+-glucose cotransporter 1. Proceedings of the National Academy of Sciences, 2007. 104(38): p. 15075-15080.
109. Moran, A.W., et al., Expression of Na+/glucose co-transporter 1 (SGLT1) is enhanced by supplementation of the diet of weaning piglets with artificial sweeteners. British Journal of Nutrition, 2010. 104(05): p. 637-646.
110. Stearns, A.T., et al., Rapid upregulation of sodium-glucose transporter SGLT1 in response to intestinal sweet taste stimulation. Ann Surg, 2010. 251(5): p. 865-71.
111. Mace, O.J., et al., Sweet taste receptors in rat small intestine stimulate glucose absorption through apical GLUT2. The Journal of Physiology, 2007. 582(1): p. 379-392.
112. Jang, H.J., et al., Gut-expressed gustducin and taste receptors regulate secretion of glucagon-like peptide-1. Proceedings of the National Academy of Sciences, 2007. 104(38): p. 15069-15074.
113. Brown, R.J. and K.I. Rother, Non-nutritive sweeteners and their role in the gastrointestinal tract. The Journal of Clinical Endocrinology & Metabolism, 2012. 97(8): p. 2597-2605.
114. Brown, R.J., M. Walter, and K.I. Rother, Effects of diet soda on gut hormones in youths with diabetes. Diabetes Care, 2012. 35(5): p. 959-964.
115. Sylvetsky, A.C., et al., Hormonal responses to non-nutritive sweeteners in water and diet soda. Nutr Metab (Lond), 2016. 13(1): p. 71.
116. Pepino, M.Y., et al., Sucralose affects glycemic and hormonal responses to an oral glucose load. Diabetes Care, 2013. 36(9): p. 2530-2535.
117. Wittekind, A., et al., A workshop on "dietary sweetness-is it an issue?'. International Journal of Obesity, 2018. 42: p. 934-938.
118. Zhang, G.H., et al., Effects of mother's dietary exposure to acesulfame-K in Pregnancy or lactation on the adult offspring's sweet preference. Chem Senses, 2011. 36(9): p. 763-70.
119. Drewnowski, A. and C. Rehm, Consumption of low-calorie sweeteners among U.S. adults is associated with higher healthy eating index (HEI 2005) scores and more physical activity. 2014. 6(10): p. 4389-4403.
120. Gibson, S., et al., Low calorie beverage consumption is associated with energy and nutrient intakes and diet quality in british adults. Nutrients, 2016. 8(1): p. 9.
121. Patel, L., G. Alicandro, and C. La Vecchia, Low-calorie beverage consumption, diet quality and cardiometabolic risk factors in british adults. Nutrients, 2018. 10(9): p. 1261.
122. Leahy, M., et al., Consumption of low-calorie sweetened beverages compared to water is associated with reduced intake of carbohydrates and sugar, with no adverse relationships to
138
glycemic responses: Results from the 2001–2012 national health and nutrition examination survey. Nutrients, 2017. 9(9): p. 928.
123. Ashwell, M., et al., Expert consensus on low-calorie sweeteners: facts, research gaps and suggested actions. Nutrition Research Reviews, 2020: p. 1-10.
124. Abou-Donia, M.B., et al., Splenda alters gut microflora and increases intestinal p-glycoprotein and cytochrome p-450 in male rats. Journal of Toxicology and Environmental Health, Part A, 2008. 71(21): p. 1415-1429.
125. Uebanso, T., et al., Effects of low-dose non-caloric sweetener consumption on gut microbiota in mice. Nutrients, 2017. 9(6): p. 560.
126. Wang, Q.P., et al., Non-nutritive sweeteners possess a bacteriostatic effect and alter gut microbiota in mice. PLoS One, 2018. 13(7): p. e0199080.
127. Bian, X., et al., Saccharin induced liver inflammation in mice by altering the gut microbiota and its metabolic functions. Food and Chemical Toxicology, 2017. 107: p. 530-539.
128. Anderson, R.L. and J.J. Kirkland, The effect of sodium saccharin in the diet on caecal microflora. 1980. 18(4): p. 353-355.
129. Naim, M., et al., Effects of sodium saccharin on the activity of trypsin, chymotrypsin, and amylase and upon bacteria in small intestinal contents of rats. 1985. 178(3): p. 392-401.
130. Mela, D.J., J. McLaughlin, and P.J. Rogers, Perspective: Standards for research and reporting on low-energy (“artificial”) sweeteners. Advances in Nutrition, 2020.
131. Thomson, P., et al., Short-term impact of sucralose consumption on the metabolic response and gut microbiome of healthy adults. Br J Nutr, 2019. 122(8): p. 856-862.
132. Ruanpeng, D., et al., Sugar and artificially sweetened beverages linked to obesity: a systematic review and meta-analysis. QJM: An International Journal of Medicine, 2017. 110(8): p. 513-520.
133. Cheungpasitporn, W., et al., Associations of sugar-sweetened and artificially sweetened soda with chronic kidney disease: A systematic review and meta-analysis. Nephrology, 2014. 19(12): p. 791-797.
134. Malik, V.S., Non-sugar sweeteners and health. BMJ, 2019: p. k5005. 135. Pan, A., et al., Changes in water and beverage intake and long-term weight changes: results
from three prospective cohort studies. International Journal of Obesity, 2013. 37(10): p. 1378-1385.
136. Mozaffarian, D., et al., Changes in diet and lifestyle and long-term weight gain in women and men. New England Journal of Medicine, 2011. 364(25): p. 2392-2404.
137. Field-Fote, E., Mediators and Moderators, Confounders and Covariates. Journal of Neurologic Physical Therapy, 2019. 43(2): p. 83-84.
138. Wiebe, N., et al., A systematic review on the effect of sweeteners on glycemic response and clinically relevant outcomes. BMC Medicine, 2011. 9(1): p. 123.
139. Riley, R.D., et al., Multivariate and network meta-analysis of multiple outcomes and multiple treatments: rationale, concepts, and examples. BMJ, 2017: p. j3932.
140. Dias, S. and D.M. Caldwell, Network meta-analysis explained. Archives of Disease in Childhood - Fetal and Neonatal Edition, 2019. 104(1): p. F8-F12.
141. Tonin, F.S., et al., Network meta-analysis: a technique to gather evidence from direct and indirect comparisons. Pharmacy Practice, 2017. 15(1): p. 943-943.
142. Evert, A.B., et al., Nutrition therapy for adults with diabetes or prediabetes: a consensus report. Diabetes Care, 2019. 42(5): p. 731-754.
143. Sievenpiper, J.L., et al., The importance of study design in the assessment of nonnutritive sweeteners and cardiometabolic health. Canadian Medical Association Journal, 2017. 189(46): p. E1424-E1425.
139
144. Puhan, M.A., et al., A GRADE Working Group approach for rating the quality of treatment effect estimates from network meta-analysis. 2014. 349(sep24 5): p. g5630-g5630.
145. Higgins, J., Thomas, J. Cochrane handbook for systematic reviews of interventions. Ed 6. 2019; Available from: https://training.cochrane.org/handbook/current.
146. Hutton, B., et al., The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: Checklist and explanations. Annals of Internal Medicine, 2015. 162(11): p. 777.
147. Chaimani, A., et al., Graphical tools for network meta-analysis in stata. PLoS ONE, 2013. 8(10): p. e76654.
148. Dias, S., et al., Checking consistency in mixed treatment comparison meta-analysis. Statistics in Medicine, 2010. 29(7-8): p. 932-944.
149. Jackson, D., et al., A design-by-treatment interaction model for network meta-analysis with random inconsistency effects. 2014. 33(21): p. 3639-3654.
150. Guyatt, G., et al., GRADE guidelines: 1. Introduction—GRADE evidence profiles and summary of findings tables. Journal of Clinical Epidemiology, 2011. 64(4): p. 383-394.
151. Brignardello-Petersen, R., et al., Advances in the GRADE approach to rate the certainty in estimates from a network meta-analysis. Journal of Clinical Epidemiology, 2018. 93: p. 36-44.
152. Brignardello-Petersen, R., et al., GRADE approach to rate the certainty from a network meta-analysis: avoiding spurious judgments of imprecision in sparse networks. Journal of Clinical Epidemiology, 2019. 105: p. 60-67.
153. Brignardello-Petersen, R., et al., GRADE approach to rate the certainty from a network meta-analysis: addressing incoherence. Journal of Clinical Epidemiology, 2019. 108: p. 77-85.
154. Bonnet, F., et al., Consumption of a carbonated beverage with high-intensity sweeteners has no effect on insulin sensitivity and secretion in nondiabetic adults. The Journal of Nutrition, 2018. 148(8): p. 1293-1299.
155. Bruun, J.M., et al., Consumption of sucrose-sweetened soft drinks increases plasma levels of uric acid in overweight and obese subjects: a 6-month randomised controlled trial. European Journal of Clinical Nutrition, 2015. 69(8): p. 949-953.
156. Campos, V., et al., Sugar- and artificially sweetened beverages and intrahepatic fat: A randomized controlled trial. Obesity, 2015. 23(12): p. 2335-2339.
157. Engel, S., et al., Effect of high milk and sugar-sweetened and non-caloric soft drink intake on insulin sensitivity after 6 months in overweight and obese adults: a randomized controlled trial. European Journal of Clinical Nutrition, 2018. 72(3): p. 358-366.
158. Hernández-Cordero, S., et al., Substituting Water for Sugar-Sweetened Beverages Reduces Circulating Triglycerides and the Prevalence of Metabolic Syndrome in Obese but Not in Overweight Mexican Women in a Randomized Controlled Trial. The Journal of Nutrition, 2014. 144(11): p. 1742-1752.
159. Higgins, K.A., R.V. Considine, and R.D. Mattes, Aspartame consumption for 12 weeks does not affect glycemia, appetite, or body weight of healthy, lean adults in a randomized controlled trial. The Journal of Nutrition, 2018. 148(4): p. 650-657.
160. Higgins, K.A. and R.D. Mattes, A randomized controlled trial contrasting the effects of 4 low-calorie sweeteners and sucrose on body weight in adults with overweight or obesity. The American Journal of Clinical Nutrition, 2019. 109(5): p. 1288-1301.
161. Madjd, A., et al., Effects on weight loss in adults of replacing diet beverages with water during a hypoenergetic diet: a randomized, 24-wk clinical trial. The American Journal of Clinical Nutrition, 2015. 102(6): p. 1305-1312.
140
162. Madjd, A., et al., Beneficial effects of replacing diet beverages with water on type 2 diabetic obese women following a hypo-energetic diet: A randomized, 24-week clinical trial. Diabetes, Obesity and Metabolism, 2017. 19(1): p. 125-132.
163. Maersk, M., et al., Sucrose-sweetened beverages increase fat storage in the liver, muscle, and visceral fat depot: a 6-mo randomized intervention study. The American Journal of Clinical Nutrition, 2012. 95(2): p. 283-289.
164. Peters, J.C., et al., The effects of water and non-nutritive sweetened beverages on weight loss and weight maintenance: A randomized clinical trial. Obesity, 2016. 24(2): p. 297-304.
165. Reid, M., et al., Long-term dietary compensation for added sugar: effects of supplementary sucrose drinks over a 4-week period. British Journal of Nutrition, 2007. 97(1): p. 193-203.
166. Reid, M., R. Hammersley, and M. Duffy, Effects of sucrose drinks on macronutrient intake, body weight, and mood state in overweight women over 4 weeks. Appetite, 2010. 55(1): p. 130-136.
167. Reid, M., et al., Effects on obese women of the sugar sucrose added to the diet over 28 d: a quasi-randomised, single-blind, controlled trial. British Journal of Nutrition, 2014. 111(3): p. 563-570.
168. Tate, D.F., et al., Replacing caloric beverages with water or diet beverages for weight loss in adults: main results of the Choose Healthy Options Consciously Everyday (CHOICE) randomized clinical trial. Am J Clin Nutr, 2012. 95(3): p. 555-63.
169. Tordoff, M.G. and A.M. Alleva, Effect of drinking soda sweetened with aspartame or high-fructose corn syrup on food intake and body weight. Am J Clin Nutr, 1990. 51(6): p. 963-9.
170. Sievenpiper, J.L., et al., Effect of fructose on body weight in controlled feeding trials. 2012. 156(4): p. 291.
171. De Ruyter, J.C., et al., A trial of sugar-free or sugar-sweetened beverages and body weight in children. New England Journal of Medicine, 2012. 367(15): p. 1397-1406.
172. Hu, F.B., Globalization of Diabetes: The role of diet, lifestyle, and genes. Diabetes Care, 2011. 34(6): p. 1249-1257.
173. Lustig, R.H., L.A. Schmidt, and C.D. Brindis, The toxic truth about sugar. Nature, 2012. 482(7383): p. 27-29.
174. Kirkpatrick, S.I., et al., Top dietary sources of energy, sodium, sugars, and saturated fats among Canadians: insights from the 2015 Canadian Community Health Survey. Applied Physiology, Nutrition, and Metabolism, 2019. 44(6): p. 650-658.
175. Canadian Diabetes Association. Diabetes: Canada at the tipping point. Available from: file:///C:/Users/Student/Downloads/WEB-Eng-CDA-Report.pdf.
176. Canadian Medical Association Journal. Artificial sweeteners linked to risk of weight gain, heart disease and other health issues. 2017; Available from: https://www.sciencedaily.com/releases/2017/07/170717091043.htm.
177. Azad, M. The effects of nonnutritive sweeteners on cardiometabolic health. Physician's Weekly 2018; Available from: https://www.physiciansweekly.com/the-effects-of-nonnutritive-sweeteners-on-cardiometabolic-health/.
178. LaMotte, S. Artificial sweeteners: Where do we stand? CNN Health 2018; Available from: https://www.cnn.com/2016/01/18/health/where-do-we-stand-artificial-sweeteners/index.html.
179. Shreeves, R. What artificial sweeteners do to your body: They've been linked to diabetes, dementia and decreased gut health. MNN.com 2019; Available from: https://www.mnn.com/food/healthy-eating/blogs/scientists-may-have-found-link-between-artificial-sweeteners-and-diabetes.
180. Beck, L. Evidence suggesting artificial sweeteners may be harmful should give us pause. The Globe and Mail 2018; Available from: https://www.theglobeandmail.com/life/health-and-fitness/article-evidence-suggesting-artificial-sweeteners-may-be-harmful-should-
141
give/?utm_medium=Newsletter&utm_source=Globe%20Life&utm_type=text&utm_content=GlobeLife&utm_campaign=2018-10-15_16.
181. Brueck, H. More evidence just landed that artificial sweeteners really are bad for you. Science Alert 2018; Available from: https://www.sciencealert.com/artificial-sweeteners-like-aspartame-can-be-toxic-to-our-gut-microbes.
182. Kirkey, S. Fake sugars linked to obesity, heart disease. National Post 2017; Available from: http://nationalpost.com/news/0717-na-sugar.
183. Leung, W. Sugar substitutes associated with weight gain and health problems, study says. Globe and Mail 2017; Available from: https://www.theglobeandmail.com/news/national/sugar-substitutes-linked-to-weight-gain-and-health-problems-study-says/article35704562/.
184. Gardener, H. and M.S.V. Elkind, Artificial sweeteners, real risks. Stroke, 2019. 50(3): p. 549-551. 185. Pagoto, S.L. and B.M. Appelhans, A Call for an end to the diet debates. JAMA, 2013. 310(7): p.
687-688. 186. Heart and Stroke Foundation of Canada. Sugar, heart disease and stroke. 2014; Available from:
https://www.heartandstroke.ca/-/media/pdf-files/canada/2017-position-statements/sugar-ps-eng.ashx.
187. US Department of Health and Human Services. Cut down on added sugars. 2016; Available from: https://health.gov/dietaryguidelines/2015/resources/DGA_Cut-Down-On-Added-Sugars.pdf.
188. Malik, V.S., et al., Long-term consumption of sugar-sweetened and artificially sweetened beverages and risk of mortality in US adults. Circulation, 2019. 139(18): p. 2113-2125.
189. Chang, K. Artificial sweeteners may disrupt body’s blood sugar controls. The New York Times 2014; Available from: https://well.blogs.nytimes.com/2014/09/17/artificial-sweeteners-may-disrupt-bodys-blood-sugar-controls/
190. Faherty, S. Artificial sweeteners may have despicable impacts on gut microbes. Scientific American 2014; Available from: https://blogs.scientificamerican.com/guest-blog/artificial-sweeteners-may-have-despicable-impacts-on-gut-microbes/
191. Sifferlin, A. Artificial sweeteners are linked to weight gain - not weight loss. Time 2017; Available from: http://time.com/4859012/artificial-sweeteners-weight-loss/.
192. Vergano, D. Artificial sweeteners may trigger blood sugar risks. National Geographic 2014; Available from: http://news.nationalgeographic.com/news/2014/09/140917-sweeteners-artificial-blood-sugar-diabetes-health-ngfood/
193. Brait, E. More research needed into the effects sugar substitutes have on health. The Toronto Star 2017; Available from: https://www.thestar.com/life/2017/07/17/more-research-needed-into-the-effects-sugar-substitutes-have-on-health.html
194. Thompson, D. Do artificial sweeteners raise odds for obesity? . WebMD 2017; Available from: http://www.webmd.com/food-recipes/news/20170717/do-artificial-sweeteners-raise-odds-for-obesity#1
195. Canadian Institutes of Health Research in partnership with Health Canada. Operating grant : Sugar and health ARCHIVED. 2016; Available from: https://www.researchnet-recherchenet.ca/rnr16/vwOpprtntyDtls.do?prog=2554&printfriendly=true.
196. Zeevi, D., et al., Personalized nutrition by prediction of glycemic responses. Cell, 2015. 163(5): p. 1079-1094.
197. Kelly, B.J., et al., Power and sample-size estimation for microbiome studies using pairwise distances and PERMANOVA. Bioinformatics, 2015. 31(15): p. 2461-2468.
198. David, L.A., et al., Diet rapidly and reproducibly alters the human gut microbiome. Nature, 2014. 505(7484): p. 559-563.
199. Braunstein, C., et al., A double-blind, randomized controlled, acute feeding equivalence trial of small, catalytic doses of fructose and allulose on postprandial blood glucose metabolism in
142
healthy participants: the fructose and allulose catalytic effects (FACE) trial. Nutrients, 2018. 10(6): p. 750.
200. Health Canada, Draft guidance document on food health claims related to the reduction in post-prandial glycaemic response, Bureau of Nutritional Sciences, Editor. 2013, Health Products and Food Branch Editor.
201. Food and Drug Administration, Multiple endpoints in clinical trials: guidance for industry [draft guidance]. 2017.
202. Benjamini, Y.a.D.Y., The control of the false discovery rate in multiple testing under dependency. Annals of Statistics, 2001. 29(4): p. 1165-1188.
203. Benjamini, Y.a.Y.H., Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the royal statistical society, 1995: p. 289-300.
204. Leung, A.A., et al., Hypertension Canada's 2017 Guidelines for Diagnosis, Risk Assessment, Prevention, and Treatment of Hypertension in Adults. Can J Cardiol, 2017. 33(5): p. 557-576.
205. Matsuda, M.a.R.A.D., Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care, 1999. 22(9): p. 1462-1470.
206. Heart and Stroke Foundation of Canada. Healthy weight and waist. 2018; Available from: https://www.heartandstroke.ca/get-healthy/healthy-weight/healthy-weight-and-waist.
207. Nerenberg, K.A., et al., Hypertension Canada’s 2018 guidelines for diagnosis, risk assessment, prevention, and treatment of hypertension in adults and children. Canadian Journal of Cardiology, 2018. 34(5): p. 506-525.
208. Mouzaki, M., et al., Intestinal microbiota in patients with nonalcoholic fatty liver disease. Hepatology, 2013. 58(1): p. 120-7.
209. Caporaso, J.G., et al., Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J, 2012. 6(8): p. 1621-4.
210. Langille, M.G., et al., Predictive functional profiling of microbial communities using 16S rRNA marker gene sequences. Nat Biotechnol, 2013. 31(9): p. 814-21.
211. Kadish, A.H.a.H., D, A new method for the continuous monitoring of blood glucose by measurement of dissolved oxygen. Clinical Chemistry, 1965. 11(9): p. 869-875.
212. Mercodia, Mercodia insulin ELISA. 2008-2017: Uppsala, Sweden. 213. Wolever, T.M., et al., The glycemic index: methodology and clinical implications. Am J Clin Nutr,
1991. 54(5): p. 846-54. 214. Phillips, D.I., et al., Understanding oral glucose tolerance: comparison of glucose or insulin
measurements during the oral glucose tolerance test with specific measurements of insulin resistance and insulin secretion. Diabet Med, 1994. 11(3): p. 286-92.
215. Theytaz, F., et al., Metabolic fate of fructose ingested with and without glucose in a mixed meal. Nutrients, 2014. 6(7): p. 2632-49.
216. Logue, C., et al., Application of liquid chromatography–tandem mass spectrometry to determine urinary concentrations of five commonly used low-Calorie sweeteners: A novel biomarker approach for assessing recent intakes? Journal of Agricultural and Food Chemistry, 2017.
217. Schulz, K.F., D.G. Altman, and D. Moher, CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ, 2010. 340(mar23 1): p. c332-c332.
218. Government of Canada. Sugars. 2018; Available from: https://www.canada.ca/en/health-canada/services/nutrients/sugars.html.
219. Leung, A., et al., Hypertension Canada's 2017 guidelines for diagnosis, risk assessment, prevention, and treatment of hypertension in adults. Canadian Journal of Cardiology, 2017.