Page 1 of 5 Compeng interests: none declared. Conflict of interests: none declared. All authors contributed to the concepon, design, and preparaon of the manuscript, as well as read and approved the final manuscript. All authors abide by the Associaon for Medical Ethics (AME) ethical rules of disclosure. For citation purposes: Christensen R, Bartels EM, Bliddal H. Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: Overview of reviews and indirect comparison with Rosa canina (a novel nutraceutical). OA Arthritis 2013 Feb 01;1(1):1. Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY) Systematic Review therapy as long as they perceive a benefit and cover the costs of the treatment themselves. It may be used as an ethically ineffective active comparator in osteoarthritis trials. Introduction Osteoarthritis (OA) is a common joint disorder and may occur in any synovial joint in the body, although the condition is most common in hands, knees, hips and spine 1 . The clinical problems, along with the pathological and radiographic changes, include joint pain, stiffness, movement with a restricted range and cracking of joints (crepitus) and decreased function with loss of muscle strength 2 . Drug therapy in OA mainly consists of analgesics and non steroidal anti-inflammatory drugs (NSAIDs). Pharmacological therapy for pain relief is widely perceived to be the backbone of effective manage- ment 3 . Unfortunately, many of the most effective pain relievers, espe- cially NSAIDs, have well-known side effects that limit their use 4,5 . In OA, specific food substances such as nutraceuticals are expected to affect the course of the disease in the degenerating joints. Nutraceuticals are functional ingredients commer- cially available as powders, pills and other medicinal forms and are gener- ally not associated with food 6 . Glucosamine, which is classified as a ‘dietary supplement’ in the United States and as a pharmaceu- tical in some European countries, is available over the counter, appears to be safe and is widely marketed for pain relief in OA 7 . However, its efficacy is uncertain; heterogeneity Abstract Introduction In the absence of randomised trials making head-to-head comparisons, we wanted to evaluate the pain- reducing effect of glucosamine hydro- chloride (GH) and the use of the Rosa canina (RP) therapy in patients with osteoarthritis by using an indirect comparison meta-analysis. Materials and methods Randomised controlled trials includ- ed in published meta-analyses were eligible for inclusion. The standard- ised mean difference (SMD) was used as an effective measure, and the difference between GH and RP was calculated using the Bucher approach. Results There was no heterogeneity for any nutraceuticals. GH studies were consist- ently located around an effect size of zero [SMD = –0.01 (–0.14; –0.12)], whereas RP studies consistently pointed towards an effect size of 0.4 [SMD = –0.37 (–0.60; –0.13)]. Testing whether there was any significant difference between GH and RP (both in contrast to placebo), we applied the Bucher approach, which resulted in an effect size of 0.36 in favour of RP [SMD = 0.36 (0.09 ; 0.62), Z = 2.65; P = 0.0082]. Conclusion GH should not be used in clinical practice. However, we see no harm in having patients continuing with GH among trials of glucosamine is larger than that expected by chance 8 . Among trials with involvement of pharma- ceutical industries, effect sizes are consistently higher. Potential expla- nations include different glucosamine preparations, inadequate allocation concealment and industry bias 9 . On the basis of three randomized controlled trials (RCTs), Vlad et al. concluded that glucosamine hydro- chloride (GH) has consistently no effect on pain, and the absence of heterogeneity (i.e. homogeneity) among these trials suggests that this summary effect is robust and valid; i.e., future studies of this preparation are unlikely to yield useful results 9 . In contrast, the hip powder Rosa canina [or rosehip powder (RP)] has shown promising results as a pain-killing nutraceutical—based on three RCTs with homogeneous efficacy results— without any contradictory results 10 . As guidelines such as those from the Osteoarthritis Research Society International are intended to provide concise, patient-focused, up-to-date, evidence-based, expert consensus recommendations for the manage- ment of hip and knee OA, an empirical comparison of the observed efficacy from the assumed well-known GH and any new nutraceutical compound seems relevant 11 . Scientifically, efficacy is most convinc- ingly established by demonstrating superiority to placebo in a placebo- controlled trial, by showing superiority to an active control treatment or by demonstrating a dose-response rela- tionship. This type of trial is referred to as a superiority trial: a trial with the primary objective of showing that the Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: an overview of reviews and indirect comparison with Rosa canina R Christensen 1,2* , EM Bartels 1 , H Bliddal 1 * Corresponding author Email: [email protected]1 Musculoskeletal Statistics Unit, The Parker Institute, Copenhagen University Hospital, Frederiksberg, Denmark 2 Institute of Sports Science and Clinical Biomechanics, University of Southern, Denmark Musculoskeletal Conditions
Transcript
Page 1 of 5
Com
petin
g in
tere
sts:
non
e de
clar
ed. C
onfli
ct o
f int
eres
ts: n
one
decl
ared
. A
ll au
thor
s co
ntrib
uted
to th
e co
ncep
tion,
des
ign,
and
pre
para
tion
of th
e m
anus
crip
t, a
s w
ell a
s re
ad a
nd a
ppro
ved
the
final
man
uscr
ipt.
A
ll au
thor
s ab
ide
by th
e A
ssoc
iatio
n fo
r Med
ical
Eth
ics
(AM
E) e
thic
al ru
les
of d
iscl
osur
e.
For citation purposes: Christensen R, Bartels EM, Bliddal H. Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: Overview of reviews and indirect comparison with Rosa canina (a novel nutraceutical). OA Arthritis 2013 Feb 01;1(1):1.
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)
Systematic Review
therapy as long as they perceive a benefit and cover the costs of the treatment themselves. It may be used as an ethically ineffective active comparator in osteoarthritis trials.
IntroductionOsteoarthritis (OA) is a common joint disorder and may occur in any synovial joint in the body, although the condition is most common in hands, knees, hips and spine1. The clinical problems, along with the pathological and radiographic changes, include joint pain, stiffness, movement with a restricted range and cracking of joints (crepitus) and decreased function with loss of muscle strength2. Drug therapy in OA mainly consists of analgesics and non steroidal anti-inflammatory drugs (NSAIDs). Pharmacological therapy for pain relief is widely perceived to be the backbone of effective manage-ment3. Unfortunately, many of the most effective pain relievers, espe-cially NSAIDs, have well-known side effects that limit their use4,5. In OA, specific food substances such as nutraceuticals are expected to affect the course of the disease in the degenerating joints. Nutraceuticals are functional ingredients commer-cially available as powders, pills and other medicinal forms and are gener-ally not associated with food6.
Glucosamine, which is classified as a ‘dietary supplement’ in the United States and as a pharmaceu-tical in some European countries, is available over the counter, appears to be safe and is widely marketed for pain relief in OA7. However, its efficacy is uncertain; heterogeneity
Abstract IntroductionIn the absence of randomised trials making head-to-head comparisons, we wanted to evaluate the pain-reducing effect of glucosamine hydro-chloride (GH) and the use of the Rosa canina (RP) therapy in patients with osteoarthritis by using an indirect comparison meta-analysis. Materials and methodsRandomised controlled trials includ- ed in published meta-analyses were eligible for inclusion. The standard-ised mean difference (SMD) was used as an effective measure, and the difference between GH and RP was calculated using the Bucher approach. ResultsThere was no heterogeneity for any nutraceuticals. GH studies were consist-ently located around an effect size of zero [SMD = –0.01 (–0.14; –0.12)], whereas RP studies consistently pointed towards an effect size of 0.4 [SMD = –0.37 (–0.60; –0.13)]. Testing whether there was any significant difference between GH and RP (both in contrast to placebo), we applied the Bucher approach, which resulted in an effect size of 0.36 in favour of RP [SMD = 0.36 (0.09 ; 0.62), Z = 2.65; P = 0.0082]. ConclusionGH should not be used in clinical practice. However, we see no harm in having patients continuing with GH
among trials of glucosamine is larger than that expected by chance8. Among trials with involvement of pharma-ceutical industries, effect sizes are consistently higher. Potential expla-nations include different glucosamine preparations, inadequate allocation concealment and industry bias9. On the basis of three randomized controlled trials (RCTs), Vlad et al. concluded that glucosamine hydro-chloride (GH) has consistently no effect on pain, and the absence of heterogeneity (i.e. homogeneity) among these trials suggests that this summary effect is robust and valid; i.e., future studies of this preparationare unlikely to yield useful results9. In contrast, the hip powder Rosa canina [or rosehip powder (RP)] has shown promising results as a pain-killing nutraceutical—based on three RCTs with homogeneous efficacy results—without any contradictory results10.
As guidelines such as those from the Osteoarthritis Research Society International are intended to provide concise, patient-focused, up-to-date, evidence-based, expert consensus recommendations for the manage-ment of hip and knee OA, an empirical comparison of the observed efficacy from the assumed well-known GH and any new nutraceutical compound seems relevant11.
Scientifically, efficacy is most convinc-ingly established by demonstrating superiority to placebo in a placebo-controlled trial, by showing superiority to an active control treatment or by demonstrating a dose-response rela-tionship. This type of trial is referred to as a superiority trial: a trial with the primary objective of showing that the
Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: an overview of reviews and
indirect comparison with Rosa caninaR Christensen1,2*, EM Bartels1, H Bliddal1
*Corresponding author Email: [email protected] Musculoskeletal Statistics Unit, The Parker
Institute, Copenhagen University Hospital, Frederiksberg, Denmark
2 Institute of Sports Science and Clinical Biomechanics, University of Southern, Denmark
For citation purposes: Christensen R, Bartels EM, Bliddal H. Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: Overview of reviews and indirect comparison with Rosa canina (a novel nutraceutical). OA Arthritis 2013 Feb 01;1(1):1.
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)
Systematic Review
calculated based on the individuals SMD and number of patients included10,18. As the unadjusted (Cohen’s) SMD in principle does not treat the variance (SE2) as an estimate, we applied (i.e. via multiplication) the Hedges’ bias-correction by default, adjusting for small sample bias19.
We used a standard inverse-variance random effects meta-analysis tech-nique to combine effect sizes across trials and calculated the variance estimate τ2 as a measure of heteroge-neity20, for GH and RP, respectively. The I2 index for evaluation of the amount of heterogeneity was also calculated21, which describes the inconsistency via the percentage of total variation across trials that is attributable to heterogeneity rather than chance22.
In the absence of randomised trials making head-to-head comparisons, an indirect comparison is possible using a common comparator23,24. The indirect method may produce less biased results than direct comparisons when evaluating new pharmaceutical interventions25. As described above, we performed two separate meta-analyses, one combining trials of GH versus placebo to obtain estimated SMD(GH) and the second comparing RP versus placebo yielding an esti-mated SMD(RP).
The estimated difference in effect of GH and RP, SMD(GH-RP) was calcu-lated using the Bucher approach23, comparing the two estimated SMDs as follows:
SMD[GH-RP] = SMD(GH) - SMD(RP)and
Var[SMD(GH-RP)] = [SE(GH)]2 + [SE(RP)]2
SE[SMD(GH-RP)] = Ö{Var[SMD(GH-RP)]}
where ‘Var’ (as in variance) indi-cates the square of the standard error. From these values we calculated a 95% confidence interval for SMD(GH-RP). This analysis can be seen as the simplest form of meta-regression with a single binary trial factor or, equivalently, as an examination of the interaction between treatment effect and type of nutraceutical26.
text words related to OA, which were combined with validated filters for controlled clinical trials and meta-analyses. Details of the search strategy are described in previous papers10,14. A specific search methodology for locating published systematic reviews with meta-analyses on these OA topics is described elsewhere15.
Meta-analyses of randomized or quasi-randomized trials in patients with OA of the knee or hip were eligible if they evaluated patient-reported pain16 in patients allocated to GH9 or RP10 compared with patients allocated to placebo.
Data collection and quality assessment A standardised form was used to extract data from the original reports of individual trials on authors of the study, year of publication, trial design, study length, number of patients randomised (i.e. the ITT population, Ntotal), number of patients for whom detailed outcome data were available for meta-analysis in each group [E(exposed) and C(control)] included in the individual-study statistical tests (NE and NC, respectively), average patient age, sex and site of OA. When necessary, we approximated means and measures of dispersion from figures and P-values. For crossover trials, we extracted data from the first period only17. Disagreements were resolved by discussion with a third reviewer and subsequent consensus.
The risk of bias assessment according to the quality of included studies is described in detail else-where8,10, including assessment of appropriateness of randomisation, masking and handling of withdrawals.
Data synthesisWe expressed treatment effects as SMDs by dividing the difference in mean (MD) values by pooled standard deviation (SD)15. Negative SMD values indicated a beneficial (pain-reducing) effect of the experimental intervention. The corresponding variance (SE2) was
response to the investigational product is superior to a comparative agent (active or placebo control)12. In some cases, an investigational product is compared with a reference treatment without the objective of showing supe-riority. This type of trial is referred to as a non-inferiority trial: a trial with the primary objective of showing that the response to the investigational product is not clinically inferior to a comparative agent (active or placebo control)12.
Using overview of reviews meth-odology, our objectives were firstly, to evaluate the analgesic effect of GH products opposed to that of RP compounds used in OA treatment via an indirect comparison meta-analysis, and secondly, to include these estimates in a concise meta-epidemiological study13 applicable for researchers designing a superiority trial with the objective of using GH as an ineffective ‘active comparator’ in a head-to-head comparison with a new nutraceutical in OA treatment.
Materials and methods
This overview of reviews was de -signed to compile and compare the evidence from published systematic reviews on GH9 and RP for pain reduction in OA10.
Searches and selection of meta-analysesThe Cochrane Library, Medline, EMBASE and CINAHL were searched using a combination of keywords and
This work conforms to the values laid down in the Declaration of Helsinki (1964). The protocol of this study has been approved by the relevant ethical committee related to our institution in which it was performed. All subjects gave full informed consent to participate in this study.
Active comparators should be chosen with care. We sepculate that as GH is widely used and broadly displays lack of efficacy, we are confident that this result will be consistent across all trials.
Page 3 of 5
Com
petin
g in
tere
sts:
non
e de
clar
ed. C
onfli
ct o
f int
eres
ts: n
one
decl
ared
. A
ll au
thor
s co
ntrib
uted
to th
e co
ncep
tion,
des
ign,
and
pre
para
tion
of th
e m
anus
crip
t, a
s w
ell a
s re
ad a
nd a
ppro
ved
the
final
man
uscr
ipt.
A
ll au
thor
s ab
ide
by th
e A
ssoc
iatio
n fo
r Med
ical
Eth
ics
(AM
E) e
thic
al ru
les
of d
iscl
osur
e.
For citation purposes: Christensen R, Bartels EM, Bliddal H. Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: Overview of reviews and indirect comparison with Rosa canina (a novel nutraceutical). OA Arthritis 2013 Feb 01;1(1):1.
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)
Systematic Review
worry about heterogeneity for any nutraceuticals. As indicated by the I2 value (0%), there was no inconsist-ency: GH studies are consistently located around an effect size of zero [SMD= -0.01 (-0.14; 0.12)], whereas RP studies consistently point towards an effect size of 0.4 [SMD = -0.37 (-0.60; -0.13)]. As illustrated by the dotted line indicating all studies combined, there seems to be a clear difference between the empirical data derived from these nutraceuticals.
Comparing the two separate meta-analysesAs presented in Figure 1, two clear distinctions between GH and RP are obvious: one being that the number of included patients in the RP trials (287 patients) is far less than that included in the GH trials (933 patients)—with the majority of clin-ical observations (76%) being from
in the present meta-epidemiological study.
Effects of interventionsFigure 1 illustrates a typical meta-analysis plot—a so-called forest plot35—where the individual study effect sizes are illustrated by squares, and the statistical uncertainty is presented with 95% confidence inter-vals; i.e. if the 95% confidence intervals are not overlapping the null-line, the statistical test (for the null hypothesis) would correspond to a P-value of <0.05. The area of the square is proportional to the preci-sion assigned to that study in a fixed-effect meta-analysis (1/SE2). These forest plots include the result of the overall effect from the GH and RP meta-analyses at the bottom of the graph; a diamond is used to distin-guish it from the individual studies. As anticipated, there is no need to
Modelling the empirical data distributionTo explore potential model distribu-tional differences between the observed efficacy data for GH and RP, we used Monte Carlo simulation to generate and model the assumed normally distributed data for the SMD values, and subsequently non-parametric descriptive statistics to describe the empirical 95% confi-dence intervals (using the 2.5 and 97.5 percentiles in the overall distri-bution of mean values) for each of the products, GH and RP, respectively.
ResultsDescription of included reviewsAs described by Vlad et al.,9 there were three RCTs eligible for inclu-sion, all applying GH27–29. The exten-sion data from the GAIT study30 were considered ineligible for our meta-epidemiological study, as the data available on pain were considered secondary to those in the Clegg et al. paper29. This is in agreement with the Cochrane review recently updated by Towheed et al.8 One minor discrep-ancy between Vlad et al. and Towheed et al. is that Towheed (first author on the Cochrane review) considers the McAlindon trial28 to be a glucosa-mine sulphate trial, in contrast to Vlad et al. who describes it as a hydrochloride trial. As TE McAlindon is co-authoring the paper published by Vlad et al., the judgement by Vlad and co-authors on the labelling of the product used is preferred9,28. However, the actual data (means and SDs) used in this meta-epidemiological study for GH were adapted from the published tables of Towheed et al8. In terms of comparing RP with placebo—or any other active compound—no new RCTs were available since the meta-analysis by Christensen et al.10; thus, three studies were eligible for inclusion31–33. A recent RCT34 assessed the symptomatic efficacy in rheuma-toid arthritis patients; thus, it was not considered to be eligible for inclusion
Figure 1: Two meta-analyses (of 3 trials) testing a nutraceutical against placebo for pain reduction in OA patients [SMDs (95% CI)]. The area of the block is proportional to the weight assigned to that study in the meta-analysis. The dotted line indicates the overall random-effects analysis if one (falsely) were to combine both GH and RP in the same meta-analysis (Q = 7.662, I2 = 35%), ignoring the clinical difference between nutraceuticals.
Page 4 of 5
Com
petin
g in
tere
sts:
non
e de
clar
ed. C
onfli
ct o
f int
eres
ts: n
one
decl
ared
. A
ll au
thor
s co
ntrib
uted
to th
e co
ncep
tion,
des
ign,
and
pre
para
tion
of th
e m
anus
crip
t, a
s w
ell a
s re
ad a
nd a
ppro
ved
the
final
man
uscr
ipt.
A
ll au
thor
s ab
ide
by th
e A
ssoc
iatio
n fo
r Med
ical
Eth
ics
(AM
E) e
thic
al ru
les
of d
iscl
osur
e.
For citation purposes: Christensen R, Bartels EM, Bliddal H. Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: Overview of reviews and indirect comparison with Rosa canina (a novel nutraceutical). OA Arthritis 2013 Feb 01;1(1):1.
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)
Systematic Review
ConclusionGH should not be used in clinical practice. However, we echo the conclusion from Wandel et al., according to which we see no harm in having patients continuing with GH therapy as long as they perceive a benefit and cover the costs of the treatment themselves. GH may be used as an inactive comparator in trials. Further evidence of the effect of nutraceuticals should be gathered in studies independent of the industry manufacturing products.
AcknowledgementsThis study was supported by grants from the Oak Foundation, The Danish Rheumatism Association, Frederiks-berg Hospital and Copenhagen Hospital Corporation. The authors acknowledge the personal and scientific support of Professor Bente Danneskiold-Samsøe, MD, Head of the Parker Institute.
References1. Dieppe PA, Lohmander LS. Pathogenesis and management of pain in osteoarthritis. Lancet 2005 Mar;365(9463):965–73.2. Altman RD. Criteria for classification of clinical osteoarthritis. J Rheumatol Suppl. 1991 Feb;27:10–2.3. Bliddal H, Christensen R. The manage-ment of osteoarthritis in the obese patient: practical considerations and guidelines for therapy. Obes Rev. 2006 Nov;7(4):323–31.4. Deeks JJ, Smith LA, Bradley MD. Efficacy,tolerability, and upper gastrointestinal safety of celecoxib for treatment of osteo-arthritis and rheumatoid arthritis: system-atic review of randomised controlled trials. BMJ 2002 Sept;325(7365):619.5. Bjordal JM, Ljunggren AE, Klovning A, Slordal L. Non-steroidal anti-inflamma-tory drugs, including cyclo-oxygenase- 2 inhibitors, in osteoarthritic knee pain: meta-analysis of randomised placebo
DiscussionThis paper clearly illustrates the inef-fectiveness of GH in the treatment of OA. As this type of glucosamine is still widely used, it is obvious that it could be easily applied as an ineffective active comparator in a (pseudo-) superiority trial, where a novel nutra-ceutical may want to benchmark against a widely used nutraceutical.
We provide such an example based on a previous meta-analysis on RP for OA—apparently a promising nutra-ceutical—although it still remains to show efficacy in a promising phase-3-like trial. As evident from these data, manufacturers of a new nutraceutical with an anticipated pain-reducing capacity over GH (ES ≥ 0.36) would need to include 163 patients in each group—for a two-group compar-ison—to be able to show that RP is superior to GH (i.e. given a power of 90% and P < 0.05). In comparison, if the authors want to lower power a priori to around 80%, then they would require 122 patients in each group36.
The quality of the evidence of inef-ficiency of GH is very consistent and robust and should exclude potential biases in the process. The data from three studies of RP are equally consistent, while studies from inde-pendent sources are lacking.
GH trials. The other major difference is the clear distinction between where the observed effect sizes are situated; eye-balling the data clearly indicates a systematic difference between GH and RP. To formally test whether there is any difference between GH and RP (both in contrast to placebo), we applied the Bucher approach23, resulting in an effect size of 0.36 in favour of RP [SMD = 0.36 (0.09; 0.62), Z = 2.65; P = 0.0082].
Empirical data distributionWhen combining the six individual data sets, we were able to analyse the data set in a more classically statistical way. Figure 2 illustrates the empirical data distributions revealing (again) that trials testing GH compared with those testing placebo cannot be expected to add value to the patients (empirical mean = -0.03, SE = 0.16, Z = 0.18, P = 0.86); whereas, RP still seems a promising therapeutic strategy to be explored in more detail with an expected effect size of -0.37 (empirical mean = -0.37, SE = 0.21). In contrast to the basic statistical meta-analysis model20, this simulated empirical model indicates that there is some evidence that RP could be more efficacious than placebo; however, till date, no truly convincing data have been reported (Z = 1.77, P = 0.076).
Figure 2: Empirical data distributions comparing GH with placebo (empirical mean = -0.03, SE = 0.16, Z = 0.18, P = 0.86); RP seems to be a promising therapeutic strategy to explore in more detail (empirical mean = -0.37, SE = 0.21).
Page 5 of 5
Com
petin
g in
tere
sts:
non
e de
clar
ed. C
onfli
ct o
f int
eres
ts: n
one
decl
ared
. A
ll au
thor
s co
ntrib
uted
to th
e co
ncep
tion,
des
ign,
and
pre
para
tion
of th
e m
anus
crip
t, a
s w
ell a
s re
ad a
nd a
ppro
ved
the
final
man
uscr
ipt.
A
ll au
thor
s ab
ide
by th
e A
ssoc
iatio
n fo
r Med
ical
Eth
ics
(AM
E) e
thic
al ru
les
of d
iscl
osur
e.
For citation purposes: Christensen R, Bartels EM, Bliddal H. Superiority trials in osteoarthritis using glucosamine hydrochloride as comparator: Overview of reviews and indirect comparison with Rosa canina (a novel nutraceutical). OA Arthritis 2013 Feb 01;1(1):1.
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)
Systematic Review
30. Sawitzke AD, Shi H, Finco MF,Dunlop DD, Bingham CO, III, Harris CL, et al. The effect of glucosamine and/or chondroitin sulfate on the progression of knee osteoarthritis: a report from the glucosamine/chondroitin arthritis intervention trial. Arthritis Rheum. 2008 Oct;58(10):3183–91.31. Warholm O, Skaar S, Hedman E, Mølmen HM, Eik L. The effects of a stand-ardized herbal remedy made from a subtype of Rosa canina in patients with osteoarthritis: a double-blind, rand-omized, placebo-controlled clinical trial. Curr Ther Res. 2003;64(1):21–31.32. Rein E, Kharazmi A, Winther K. A herbal remedy, Hyben Vital (stand. powder of a subspecies of Rosa canina fruits), reduces pain and improves general wellbeing in patients with osteoarthritis—a double-blind, placebo-controlled, randomised trial. Phytomedi-cine 2004 Jul;11(5):383–91.33. Winther K, Apel K, Thamsborg G. A powder made from seeds and shells of a rose-hip subspecies (Rosa canina) reduces symptoms of knee and hip osteo-arthritis: a randomized, double-blind, placebo-controlled clinical trial. Scand J Rheumatol. 2005 Jul;34(4):302–8.34. Willich SN, Rossnagel K, Roll S,Wagner A, Mune O, Erlendson J, et al. Rose hip herbal remedy in patients with rheumatoid arthritis - a randomised controlled trial. Phytomedicine 2010 Feb;17(2):87–93.35. Lewis S, Clarke M. Forest plots: trying to see the wood and the trees. BMJ 2001 Jun;322(7300):1479–80.36. Campbell MJ, Julious SA, Altman DG. Estimating sample sizes for binary, ordered categorical, and continuous outcomes in two group comparisons. BMJ 1995 Oct;311(7013):1145–8.
17. Elbourne DR, Altman DG, Higgins JP, Curtin F, Worthington HV, Vail A. Meta-analyses involving cross-over trials: methodological issues. Int J Epidemiol. 2002 Feb;31(1):140–9.18. Normand SL. Meta-analysis: formu-lating, evaluating, combining, and reporting. Stat Med. 1999 Feb;18(3):321–59.19. Hedges LV. Distribution theory for glass’s estimator of effect size and related estimators. J Educational Statistics 1981; 6(2):107–28.20. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials 1986 Sept;88(3):177–88.21. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002 Jun;21(11):1539–58.22. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003 Sept;327 (7414):557–60.23. Bucher HC, Guyatt GH, Griffith LE,Walter SD. The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials. J Clin Epidemiol. 1997 Jun;50(6):683–91.24. Song F, Altman DG, Glenny AM, Deeks JJ. Validity of indirect comparison for estimating efficacy of competing inter-ventions: empirical evidence from published meta-analyses. BMJ 2003 Mar; 326(7387):472.25. Song F, Harvey I, Lilford R. Adjusted indirect comparison may be less biased than direct comparison for evaluating new pharmaceutical interventions. J Clin Epidemiol. 2008 May;61(5):455–63.26. Davey SG, Egger M, Phillips AN. Meta-analysis. Beyond the grand mean? BMJ 1997 Dec;315(7122):1610–4.27. Houpt JB, McMillan R, Wein C,Paget-Dellio SD. Effect of glucosamine hydrochloride in the treatment of pain of osteoarthritis of the knee. J Rheu-matol. 1999 Nov;26(11):2423–30.28. Mcalindon T, Formica M, LaValley M, Lehmer M, Kabbara K. Effectiveness of glucosamine for symptoms of knee osteo-arthritis: results from an internet-based randomized double-blind controlled trial. Am J Med. 2004 Nov;117(9):643–9.29. Clegg DO, Reda DJ, Harris CL, Klein MA, O’Dell JR, Hooper MM, et al. Glucosa-mine, chondroitin sulfate, and the two in combination for painful knee osteoar-thritis. N Engl J Med. 2006 Feb;354 (8):795–808.
controlled trials. BMJ 2004 Dec; 329(7478):1317.6. Ameye LG, Chee WS. Osteoarthritis and nutrition. From nutraceuticals to func-tional foods: a systematic review of the scientific evidence. Arthritis Res Ther. 2006 Jul;8(4):R127.7. Richy F, Bruyere O, Ethgen O, Cucherat M, Henrotin Y, Reginster JY. Structural and symptomatic efficacy of glucosamine and chondroitin in knee osteoarthritis: a comprehensive meta-analysis. Arch Intern Med. 2003 Jul;163(13):1514–22.8. Towheed TE, Maxwell L, Anastassiades TP, Shea B, Houpt J, Robinson V, et al. Glucosamine therapy for treating osteo-arthritis. Cochrane Database Syst Rev. 2005;(2):CD002946.9. Vlad SC, LaValley MP, McAlindon TE, Felson DT. Glucosamine for pain in osteo-arthritis: why do trial results differ? Arthritis Rheum. 2007 Jul;56(7):2267–77.10. Christensen R, Bartels EM, Altman RD, Astrup A, Bliddal H. Does the hip powder of Rosa canina (rosehip) reduce pain in osteoarthritis patients?—a meta-analysis of randomized controlled trials. Osteoar-thritis Cartilage 2008 Apr;16(9):965–72.11. Zhang W, Moskowitz RW, Nuki G,Abramson S, Altman RD, Arden N, et al. OARSI recommendations for the manage-ment of hip and knee osteoarthritis, Part II: OARSI evidence-based, expert consensus guidelines. Osteoarthritis Cartilage 2008 Feb;16(2):137–62.12. Lewis JA. Statistical principles for clin-ical trials (ICH E9): an introductory note on an international guideline. Stat Med. 1999 Aug;18(15):1903–42.13. Sterne JA, Juni P, Schulz KF, Altman DG, Bartlett C, Egger M. Statistical methods for assessing the influence of study char-acteristics on treatment effects in ‘meta-epidemiological’ research. Stat Med. 2002 Jun;21(11):1513–24.14. Bliddal H, Christensen RD, Kristensen PK, Astrup AV. [Glucosamine effectiveness in the treatment of knee osteoarthritis. Presentation of a Cochrane analysis with the perspective on the GAIT trial]. Ugeskr Laeger. 2006 Dec;168(50):4405–9.15. Bliddal H, Christensen R. The treatmentand prevention of knee osteoarthritis: a tool for clinical decision-making. Expert Opin Pharmacother. 2009 Aug;10(11):1793–804.16. Dougados M. Monitoring osteoarthritisprogression and therapy. Osteoarthritis Cartilage 2004;12 Suppl A:S55–S60.
