a Section of Social Medicine, Department of Public Health, and b Fertility Clinic, Juliane Marie Center, Rigshospitalet,University of Copenhagen, Copenhagen, Denmark
Objective: To investigate the independent and combined associations between female and male body mass index (BMI) on the prob-ability of achieving a live birth after treatments with in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) under adjust-ment for relevant covariates.Design: Population-based cohort study.Setting: Danish national registers.Patient(s): Patients with permanent residence in Denmark receiving IVF or ICSI treatment with use of autologous oocytes from Jan-uary 1, 2006, to September 30, 2010.Intervention(s): None.Main OutcomeMeasure(s): Live birth. Analyses were adjusted for age and smoking at treatment initiation and results stratified by BMIgroups and presented by IVF/ICSI treatment.Result(s): In total, 12,566 women and their partners went through 25,191 IVF/ICSI cycles with 23.7% ending in a live birth. Overweightand obese women with regular ovulation had reduced odds of live birth (adjusted OR 0.88, 95% CI 0.79–0.99 and adjusted OR 0.75, 95%CI 0.63–0.90, respectively) compared with normal-weight women. IVF-treated couples with both partners having BMIR25 kg/m2 hadthe lowest odds of live birth (adjusted OR 0.73, 95% CI 0.48–1.11) compared with couples with BMI <25 kg/m2. BMI showed nosignificant effect on chance of live birth after ICSI.
Use your smartphone
Conclusion(s): Increased female and male BMI, both independently and combined, negativelyinfluenced live birth after IVF treatments. With ICSI, the association with BMI was less clear.(Fertil Steril� 2013;-:-–-. �2013 by American Society for Reproductive Medicine.)KeyWords: Body mass index, in vitro fertilization, intracytoplasmic sperm injection, live birth,multilevel analysis
Discuss: You can discuss this article with its authors and with other ASRM members at http://fertstertforum.com/petersengl-body-mass-index-live-birth-art/
to scan this QR codeand connect to thediscussion forum forthis article now.*
* Download a free QR code scanner by searching for “QRscanner” in your smartphone’s app store or app marketplace.
he prevalence of obesity is in- overweight and obese individuals reproductive age in both sexes (3).
T creasing and constitutes a majorworldwide epidemic affecting
more than one billion people worldwide(1, 2). In Denmark, the proportion of
Received October 23, 2012; revised December 21, 20G.L.P. has nothing to disclose. L.S. reports a grant fro
unrelated to this work). A.P. has nothing to disReprint requests: Gitte Lindved Petersen, M.Sc. (Public
constituted 32.5%–39.8% amongwomen and 46.2%–57.6% among menaged 25–44 years in 2010. Theprevalence tends to increase during
12; accepted January 10, 2013.m and travel expenses paid by Merck Serono (allclose. M.K.-J. has nothing to disclose.Health), Section of Social Medicine, University ofpenhagen K, Denmark (E-mail: [email protected]).
0282/$36.00Medicine, Published by Elsevier Inc.
Overweight and obesity is known tobe associated with a number ofcomorbidities, such as type 2 diabetes,hypertension, certain cancers, andstroke (4). Over the past decades,attention has been directed toward theeffect of obesity on fertility.
In women, adipose tissue affectsthe gonad hormonal balance, leadingto increased levels of leptin and de-creased levels of adiponectin, which isnegatively associated with plasma
insulin levels. The altered hormone levels may result in hyper-androgenism (1), and in adipose women contribute to anovu-lation. However, even in ovulating women, increased bodymass index (BMI) reduces conception rates. Increased BMImay thus detrimentally affect the oocytes and/or embryoquality (5).
In men, increased scrotal temperatures with increasedBMI can adversely influence spermatogenesis (6). Moreover,leptin and E2 levels correlate with the amount of fat tissue,and higher levels of E2 may cause shifts in levels of gonado-tropins and in the LH/FSH ratio, disturbing the spermatogen-esis (7). However, a prospective study of 1,466 men did notfind any negative effect of increased BMI on semen quality(8). Nor did a systematic review with meta-analysis find anyoverall relationship between levels of E2 and BMI in men (9).
Female obesity has been associated with lower chance ofconception, regardless of menstrual cycle pattern (2, 10), andRamlau-Hansen et al. found a dose-response relationship be-tween increasing female BMI and prolonged time to preg-nancy (TTP) in a Danish cohort study (11). Reportedly, TTPincreased by 2.84 days per 1 kg increase in body weight,and increased risk of TTP >12 months was observed amongcouples with both partners being obese (11).
Among men, the odds of taking >12 months to conceivehas been detected to increase by 12%with each 3-unit increasein BMI, compared with BMI 20–22 kg/m2 (6). Another studyfound 20% and 36% higher risks of TTP R12 months amongmen with BMIs 25.0–29.9 kg/m2 and 30.0–34.9 kg/m2, respec-tively, compared with men with BMI 20.0–22.4 kg/m2. Con-trasts may even be underestimated, because couples withoutconception were not included (12).
Also, studies of outcome from assisted reproductive tech-nology (ART) treatments have detected negative effects of in-creased female BMI (13–15). In recent reviews, the probabilityof live birth after ART was found to be 9%–10% lower amongoverweight/obese women compared with normal-weightwomen (16, 17). Obesity affects the outcome even morethan overweight (16, 18, 19), with suggestions of a dose-response relationship (16). However, other studies found noeffect of increased female BMI on pregnancy rate after IVF(20) or on deliveries after IVF/ICSI (21).
Knowledge about effects of male BMI on ART outcome isscarce, but increased male BMI has been found to reduce thechance of live birth (22). However, the effect may be possibleto overcome by the use of ICSI (intracytoplasmic sperm injec-tion) rather than IVF (23).
In Denmark, the cumulative proportion of infertility is16%–26% in women who tried conceiving for R12 months(24, 25). Approximately 10% of all Danish children were bornafter medically assisted reproduction (MAR) treatment in2010, and >15,000 ART treatments are initiated each year(26). The Danish health care system offers infertile couplesand single women up to three fully reimbursed IVF or ICSItreatment cycles with fresh embryos and an unlimitednumber of frozen embryo transfer and insemination cycles(in practice, 3–6 cycles) if the woman's age does not exceeds40 years. According to Danish law, women can be treated ina private setting until their 46th birthday. In Denmark,information on all MAR treatments, including characteristics
2
of both partners, is recorded in publicly administratedregistries. The objective of the present study was toinvestigate the independent and combined associationsbetween female and male BMI and the probability ofachieving a live birth following treatment with IVF or ICSI.
MATERIALS AND METHODSStudy Population
This study was based on nationwide Danish register data. In-formation on ART treatments was collected from the IVF reg-istry and linked to information on all births and abortionsobtained from the Danish Medical Birth Registry (MBR) andthe Danish National Patient Registry (NPR). Linkage wasdone by use of the central personal registry (CPR) number,which is a unique identity allocated to all citizens with perma-nent residence in Denmark. The CPR number allows linkage ofDanish registries with little or no error.
Women with permanent residence in Denmark were eligi-ble for inclusion if they had at least one IVF or ICSI treatmentcycle with the use of autologous oocytes from January 1,2006, to September 30, 2010. Information on treatment-specific conditions was registered at first visit to the fertilityclinic. Height and weight information was routinely regis-tered from 2006 for women and from 2009 for men. Forwomen with one or more registered live births from January1, 2006, to June 30, 2011, we calculated the number of daysbetween treatment cycle and live birth. A live birth was con-sidered to be the result of the particular ART treatment if thedifference was 140–308 days (20–44 weeks). The calculateddifference was compared with the reported gestational age(in days) from the MBR by subtracting the gestational agefrom the calculated difference. Births with differences of 0–29 days were included. If the difference was >29 days, thebirth was considered to be the outcome of another cycle. Be-cause women and men could have several treatments, the ob-servation unit was treatment cycle.
Figure 1 shows exclusions from the baseline dataset andnumbers in the final cohort. The majority of the exclusionswere due tomissing female BMI information.We excluded cou-ples if their first treatment cycle was a frozen embryo transfercycle, because this indicates a treatment cycle before the inclu-sion period, during which information was not available.
This study was approved by the Danish Data ProtectingAgency (2011-41-6052). According to Danish law, no ap-proval is required from the National Committee for Health Re-search Ethics for registry-based research.
Body Mass Index
BMI was calculated as body weight (kg) divided by heightsquared (m2) and categorized according to the World HealthOrganization classifications as underweight (<18.5 kg/m2),normal weight (18.5–24.9 kg/m2), overweight (25.0–29.9kg/m2), and obese (R30 kg/m2).
Live Births
We obtained information on live births from the MBR, whichcontains information on all births in Denmark.
VOL. - NO. - / - 2013
FIGURE 1
Flow diagram of exclusions from the study population. BMI, bodymass index; FET, frozen embryo transfer; PGD, preimplantation genetic diagnosis.a Number of women with a male partner.Petersen. Body mass index and IVF/ICSI outcome. Fertil Steril 2013.
Fertility and Sterility®
Covariates
Age was recorded in years at treatment initiation and groupedinto 5-year intervals. Smoking was reported as number ofcigarettes smoked per day at treatment initiation and groupedinto smokers versus nonsmokers. Infertility diagnoses werecoded according to the 10th revision of the InternationalClassification of Diseases, and, based on female infertility diag-nosis, women were categorized as ovulatory or anovulatory.
Statistical Analyses
Descriptive results are presented as frequencies with row per-centages or medians with interquartile ranges (IQRs). BMIgroups were compared by use of c2 tests (frequencies) orKruskal-Wallis tests (medians).
Multilevel logistic regression analyses were performedwith the use of Proc Glimmix to estimate the odds of live birthfollowing IVF or ICSI treatments in consecutive treatmentcycles. Analyses were performed according to female andmale BMI groups, respectively, with normal-weight patientsas reference group. Analyses of female BMI were adjustedfor female age and female smoking (both recorded at treat-ment initiation). Analyses were performed separately forovulatory and anovulatory women. Analyses of male BMI
VOL. - NO. - / - 2013
were adjusted for male age and male smoking. To examinepotential linear trends, we also included BMI as a continuousvariable in the analyses of female and male BMI, respectively,and adjusted for the same covariates as in the categoricanalyses.
Multilevel logistic regression was repeated as coupleanalyses. To increase statistical power, BMI groups for womenand men were analyzed as BMI <25 and BMI R25. The ref-erence group was couples with both partners having BMI<25 kg/m2. Estimates were adjusted for female and maleage and smoking. Women were included regardless of ovula-tory status.
Results from the multilevel logistic regression analysesare presented as adjusted odds ratios (ORs) with 95% confi-dence intervals (CIs) and P values from Wald c2 tests.
All statistical analyses were performed with the use ofSAS version 9.2.
RESULTSIn total, 12,566 unique women and their partners wentthrough 25,191 treatment cycles with IVF or ICSI, of which23.7% ended in live birth (23.4% of IVF and 24.0% of ICSI).Median BMI was lower in women (23.2 kg/m2, IQR
3
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY
21.1–26.2) than men (25.2 kg/m2, IQR 24.1–27.8). Womenhad a median age of 34.0 years (IQR 30.0–37.0) and men35.0 years (IQR 32.0–39.0), and the median number of treat-ment cycles during the study period was 2 (IQR 1–3).Table 1 shows characteristics of studied treatment cycles ac-cording to sex and BMI groups. The majority of the treatmentcycles were performed in normal-weight and overweight pa-tients (respectively, 63.9% and 23.4% of treatment cyclesamong women and 47.5% and 38.9% of treatment cyclesamong men). Except for fertility treatment among men, allvariables were significantly unevenly distributed amongBMI groups. Stratified by sex, a larger proportion of the cycleswere performed in the lowest BMI and younger female agegroups compared with the male groups. Despite this unequaldistribution of BMI and age groups in women andmen, cross-tabulation showed that couples tended to resemble each otherin BMI, age, and smoking (data not shown).
In Table 2, results from the multilevel analyses are pre-sented according to female and male BMI. In general, both fe-male and male overweight and obesity were associated withdecreased odds of live birth following IVF. Among ovulatorywomen, a significant difference between BMI groups wasseen, with 12% (95% CI 0.79–0.99) and 25% (95% CI 0.63–0.90) reduction in odds of live birth among the overweightand obese, respectively. From the trend analysis, a statisticallysignificant decrease of 2% (95% CI 0.97–0.99) was observedwith every 1-unit increase in BMI (P¼ .003). Similar tenden-cies were seen among anovulatory women and men. ForICSI treatments, the association between BMI and live birthshowed no clear tendencies. Even though ORs were suggestiveof a lower probability of live birth among underweight andobese anovulatory women and men, 95% CIs were nonsignif-icant. Analyses of IVF and ICSI cycles together revealed ten-dencies similar to IVF cycles only, though with a less strongeffect of BMI (Table 2).
Table 3 shows the results of the multilevel analyses ofcouple as the unit of analysis. The majority of treatments(31.5%) were performed in the reference group (coupleswith BMI <25 kg/m2). The lowest OR of live birth was foundin IVF-treated couples with BMIR25 kg/m2 (OR 0.73, 95% CI0.48–1.11). Reduced ORs of live birth were observed as well inall other combined BMI groups, but the reductions in OR didnot reach statistical significance. The analysis of ICSI-treatedcouples showed only a small insignificant reduction in oddsof live birth across groups of couples with one or both part-ners having BMIR25 kg/m2. In accordance with the findingsfrom the separate analyses on women and men in Table 2, theanalysis of IVF and ICSI treatments in the couple analysisshowed a tendency of lower odds of live birth with higherBMI groups similar to that in couples treated with IVF butwith a less strong impact of BMI (Table 3).
In a sensitivity analysis, we excluded anovulatory womenfrom the couple analyses. However, no considerable changesin results were found (data not shown).
Subsample Analyses
Information on treatments predating January 1, 2006, wasnot available. To identify cycle number including first treat-
4
ment cycle with reasonable accuracy, we conducted analysessimilar to those presented in Table 2 of a subsample includingonly women initiating their first treatment cycle after Decem-ber 31, 2007. Results showed results similar to those of the fullcohort. It was not possible to analyze anovulatory ICSI-treated women owing to small group sizes.
Finally, we examined if any statistical significant differ-ence between odds of live birth following IVF/ICSI treatmentby treatment cycle number could be identified. In a logistic re-gression model we compared treatment cycle 1, 2, and 3 iden-tified in the subsample (January 1, 2008–June 30, 2010). Theprobability of live birth in treatment cycles 2 and 3 did notvary significantly from that of the first treatment cycle.
DISCUSSIONIn summary, the major finding of the present study was thatboth female and male increased BMIs influence on the odds oflive birth after IVF treatment negatively. After ICSI, the find-ings were less clear. In the couple analyses, similar associa-tions between BMI and live birth were found.
The negative effect of increased female BMI on live birthafter ART treatments found in this study is supported by mostearlier findings. Recently, a retrospective study of 4,609 pa-tients found BMI R30 kg/m2 to be associated with 37%–
68% lower probability of live birth after IVF/ICSI comparedwith patients with BMI 18.50–24.99 kg/m2 (27). In a cohortstudy of 383 women, Fedorcs�ak et al. (28) found significantlylower live birth rates among obese women compared withlean. In 2004, Fedorcs�ak et al. (29) published another studyof 2,660 IVF/ICSI-treated couples, showing female BMI R30kg/m2 to be associated with lower probability of live birthcompared with female BMI 18.5–24.9 kg/m2. Contradictingresults also have been published. A study of 573 patients goingthrough IVF/ICSI showed no negative effect of obesity on em-bryo implantation rate, pregnancy rate, or miscarriage rate(20), and from a retrospective study of 278 couples, no differ-ence in clinical pregnancy rates was seen between womenwith BMI >24 kg/m2 and BMI <25 kg/m2 (21).
Systematic reviews suggest that female overweight andobesity negatively affect outcomes from ART treatment, butwhen using live birth as outcome measure, the results areless conclusive. Maheshwari et al. (30) found insufficient ev-idence on the effect of BMI on chance of live birth followingART treatment, whereas Rittenberg et al. (16) detected signif-icantly lower live birth rates (relative risk 0.84, 95% CI0.76–0.92) in women with BMI R25 kg/m2. From anothersystematic review, an OR of live birth of 0.90 (95% CI 0.82–1.0) was found among women with BMI >25 kg/m2, and itwas concluded that there was no evidence of increased riskof complications after ART from overweight or obesity (17).Results from the present study suggest that female and maleBMI are not associated similarly with probability of live birthafter IVF and ICSI treatments. Our results indicate that femaleBMI has greater influence on live birth after IVF comparedwith ICSI cycles. Because most earlier studies do not stratifyresults according to IVF or ICSI treatment, a more negativeeffect of increased BMI in IVF than in ICSI may have beenmasked. The degree to which BMI affects outcomes of IVF/
VOL. - NO. - / - 2013
TABLE 1
Characteristics of the study population according to gender and body mass index.
Body mass indexa
P valueb TotalaUnderweight Normal weight Overweight Obese
Women 760 (3.0) 16,100 (63.9) 5,892 (23.4) 2,439 (9.7) – 25,191 (100.0)Age at treatment initiation, y <25 30 (4.0) 297 (1.8) 127 (2.2) 60 (2.5) < .001 514 (2.0)
Note: ICSI ¼ intracytoplasmic sperm injection; IQR ¼ interquartile range.a Presented as numbers with column percentages for categoric variables and medians with IQRs for continuous variables. The unit of analysis was treatment cycle.b Comparison of BMI groups by use of c2 tests for the categoric variables (age, smoking, fertility treatment, and number of treatments during study period) or Kruskal-Wallis tests for the continuous variables (years of infertility).
Petersen. Body mass index and IVF/ICSI outcome. Fertil Steril 2013.
VOL.-
NO.-
/-
20135
Fertilityand
Sterility®
TABLE 2
Results from multilevel logistic regression analyses of probability of live birth following all IVF or ICSI cycles according to sex and stratified on body mass index (BMI).
n (%)
Live births per IVF cyclea Live births per ICSI cyclea Live births per IVF/ICSI cyclea
OR 95% CI P valueb OR 95% CI P valueb OR 95% CI P valueb
Note: CI ¼ confidence interval; ICSI ¼ intracytoplasmic sperm injection; OR ¼ odds ratio.a ORs with 95% CIs and P values from Wald c2 tests.b Female analyses adjusted for female age and smoking status at treatment initiation. Male analyses adjusted for male age and smoking status at treatment initiation. The unit of analysis was treatment cycle.c Represents inclusion of the BMI variable in its continuous form to analyze a possible linear association between BMI and probability of live birth following IVF or ICSI treatment.d Missing information was caused by male BMI not being routinely collected until 2009.
Petersen. Body mass index and IVF/ICSI outcome. Fertil Steril 2013.
6VOL.-
NO.-
/-
2013
ORIG
INALARTIC
LE:ENVIRO
NMEN
TANDEPID
EMIOLO
GY
TABLE 3
Results from multilevel logistic regression analyses of probability of live birth following IVF or ICSI treatment of couples.
Combination of BMI (kg/m2) Live birth per IVF cyclea Live birth per ICSI cyclea Live birth per IVF/ICSI cyclea
Women Men n (%) OR 95% CI P value OR 95% CI P value OR 95% CI P value
<25 <25 601 (31.5) 1 – .455 1 – .974 1 – .488<25 R25 508 (26.7) 0.78 (0.52–1.16) 0.91 (0.59–1.41) 0.83 (0.62–1.12)R25 <25 313 (16.4) 0.83 (0.50–1.37) 0.94 (0.60–1.47) 0.87 (0.62–1.21)R25 R25 484 (25.4) 0.73 (0.48–1.11) 0.92 (0.59–1.42) 0.81 (0.60–1.09)Note: Abbreviations as in Table 2.a Adjusted for female age, female smoking, male age, and male smoking. The unit of analysis was treatment cycle, and results are given as ORs with 95% CIs and P values from Wald c2 tests.
Petersen. Body mass index and IVF/ICSI outcome. Fertil Steril 2013.
Fertility and Sterility®
ICSI may therefore depend on the distribution of IVF and ICSItreatments in the sample.
The reason for BMI having less influence on ICSI cyclesmay be that the ICSI technology overcomes the possible directnegative influence of fat on the oocytes or the fact that severemale infertility is the major reason for infertility in the ICSIgroup, although in the IVF group the fertility problem maybe caused by factors related to overweight or obesity. It hasbeen shown that female obesity is associated with increasedrisk of spontaneous abortion after oocyte donation (31).This suggests that female obesity is an independent risk factorfor not achieving a live birth after fertility treatment, and thatthe reason may not be related solely to the quality of the oo-cytes. A later study by Bellver et al. (32) also reported poorerongoing pregnancy rates per oocyte donation cycle in over-weight and obese women (n ¼ 450 and 122, respectively).That finding supports the hypothesis that excess body weightnegatively affects the endometrium or its environment.
Irregular menstrual cycle patterns have been found tonegatively affect TTP (10). The present study also suggests dif-ferences in odds of live birth by BMI from IVF and ICSI treat-ment by female ovulation status. However, the group ofanovulatory women was small and risk estimates too uncer-tain to draw any firm conclusions.
The effect of male BMI on ART outcomes is less studied,but earlier results do suggest, in linewith this study, a negativeimpact of increasing male BMI on ART outcome. In a retro-spective analysis of 305 couples undergoing ART, increasedmale BMI was associated with significantly reduced live birthrates (22). Another retrospective study of 290 ART treatments,showed that male BMI>25 kg/m2 was associated with signif-icantly lower clinical pregnancy rates compared with maleBMI <25 kg/m2. The authors hypothesized that ICSI maypartly overcome obesity-related impairment of the sperm-egg interaction (23). We found higher odds of live birth afterICSI treatment compared with IVF among overweight andobese men. From trend analyses, a 1% lower decrease in prob-ability by each 1-unit increase in BMI was found among ICSI-treated men compared with IVF-treated. Our results therebysupport the proposed hypothesis that ICSI may overcomeobesity-related impairment of the sperm-egg interaction.
The negative impact of increased male BMI might be dueto altered hormonal profile, with increased leptin and E2 levelscausing reduced semen quality due to disturbance of the sper-matogenesis (7). A pilot weight-reduction study of 43 men
VOL. - NO. - / - 2013
with BMI >33 kg/m2 found obesity to be associated withpoor semen quality and altered reproductive hormonal profile(33), and increased male BMI to be associated with decreasedblastocyst development (22). A literature review recently con-cluded that the effects of male obesity-associated infertilityare modest but that it should be of great concern to healthprofessionals owing to the increasing rates of obesity in thegeneral population (34). Results from our study support thisnotion, because we found associations between male BMIand ART similar to those of female BMI.
To the best of our knowledge, no studies on the effect ofcouples' BMI on ART treatment have been published before,so the findings of the present study need to be further ex-plored before being accepted. In general, most existing studiesof BMI and ART are concerned with female BMI only, but re-sults from the present study call for increased focus on thecouple's BMI.
Groups of couples with combined high and low BMI weresmall, resulting in uncertain estimates from the statisticalanalyses. People tend to select partners resembling them-selves in BMI (35), which places some couples in a veryunfavorable position regarding weight-related infertility.
Multilevel logistic regression analyses were selected forthe analyses, because we found a minor decrease in chanceof live birth by increasing number of treatment cycles. Thisstatistical model allowed for inclusion of several treatmentsin each couple under consideration of the difference in chanceof live birth by treatment number. Increased number of ana-lyzed treatments thus allowed for inclusion of covariates inthe model without considerable loss of statistical power.
Because live birth is a frequent outcome after ART treat-ment, ORs do not reflect RRs very well. Interpretation of theORs as RRs would cause an overestimation of the RRs.
The major strength of the present study is its nationwideunselected population representing all ART-treated couplesliving in Denmark. Inclusion of information on men andwomen makes this study unique. In addition, inclusion ofall treatments by use of multilevel analyses ensured an un-precedented large sample size under adjustment for the inde-pendence of ART treatment cycles. However, certainlimitations also apply.
Owing to missing BMI information for most men, someanalyzed groups were small. However, we found no sign of se-lection bias, because the missing information was caused bymale BMI not being routinely collected before 2009. For
7
ORIGINAL ARTICLE: ENVIRONMENT AND EPIDEMIOLOGY
women, we retained BMI information between treatmentcycles if some were missing. Unfortunately, this was not pos-sible for men, because male CPR numbers were not availablein the data set.
Adjusting the association between BMI and live birth forduration of infertility was feasible only in a subset of data,because it had not been registered in 86% of the treatmentcycles. However, in a sensitivity analysis on the remaining14% we found tendencies similar to those not adjusted foryears of infertility (data not shown).
Information on treatment cycles predating January 1,2006, was not available, and the assumed first treatment cyclemay thus not be the actual first. Sensitivity analyses wereconducted on patients with no identified treatments beforeJanuary 1, 2008, with the intention to identify cycle numberwith reasonable accuracy. Results were similar to or amplifiedthe suggested tendencies found from analyses of the fullcohort. Any effect of unidentified previous treatment cyclesamong analyzed patients is expected to have been negligiblebut may imply a minor risk of type 2 errors.
Official recommendations and introduction of specific re-quirements of weight loss before ART treatment initiation willneed support and guidance programs for patients tomeet theirtarget BMI. Future studies should investigate how weight re-duction programs can be designed and implemented, and itmay also be relevant to identify possible age cutoffs as previ-ously discussed by Pinborg et al. (19). Weight loss amongoverweight and obese patients seeking ART treatment may,besides increasing the chance of achieving a live birth, implypositive health-related side effects and economic cost savingson a societal level. Finally, we anticipate that focusing onjoint action within couples will make lifestyle interventionsmore successful and easier to implement.
Acknowledgments: The authors thank Steen Rasmussen,Danish National Board of Health, for his assistance in extract-ing the data from the registries.
REFERENCES1. Metwally M, Li TC, LedgerWL. The impact of obesity on female reproductive
function. Obes Rev 2007;8:515–23.2. van der Steeg JW, Steures P, Eijkemans MJ, Habbema JD, Hompes PG,
3. Christensen AI, Ekholm O, Davidsen M, Juel K. Sundhed og sygelighedi Danmark 2010 & udviklingen siden 1987. The National Health Inter-view Survey 2010. Copenhagen: National Institute of Public Health;2012.
4. Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham CL, Anis AH. The in-cidence of co-morbidities related to obesity and overweight: a systematicreview and meta-analysis. BMC Public Health 2009;9:88.
5. Robker RL. Evidence that obesity alters the quality of oocytes and embryos.Pathophysiology 2008;15:115–21.
6. Sallmen M, Sandler DP, Hoppin JA, Blair A, Baird DD. Reduced fertilityamong overweight and obese men. Epidemiology 2006;17:520–3.
7. Fejes I, Koloszar S, Zavaczki Z, Daru J, Szollosi J, Pal A. Effect of body weighton testosterone/estradiol ratio in oligozoospermic patients. Arch Androl2006;52:97–102.
8. Duits FH, van WM, van d, V, Gianotten J. Healthy overweight male partnersof subfertile couples should not worry about their semen quality. Fertil Steril2010;94:1356–9.
8
9. MacDonald AA, Herbison GP, Showell M, Farquhar CM. The impact of bodymass index on semen parameters and reproductive hormones in humanmales: a systematic review with meta-analysis. Hum Reprod Update 2010;16:293–311.
10. Jensen TK, Scheike T, Keiding N, Schaumburg I, Grandjean P. Fecundabilityin relation to body mass and menstrual cycle patterns. Epidemiology 1999;10:422–8.
11. Ramlau-Hansen CH, Thulstrup AM, Nohr EA, Bonde JP, Sørensen TIA,Olsen J. Subfecundity in overweight and obese couples. Hum Reprod2007;22:1634–7.
12. Nguyen RH, Wilcox AJ, Skjaerven R, Baird DD. Men's body mass index andinfertility. Hum Reprod 2007;22:2488–93.
13. Ferlitsch K, Sator MO, Gruber DM, R€ucklinger E, Gruber CJ, Huber JC. Bodymass index, follicle-stimulating hormone and their predictive value in in vitrofertilization. J Assist Reprod Genet 2004;21:431–6.
14. Lintsen AME, Pasker-de Jong PCM, de Boer EJ, Burger CW,Jansen CAM, Braat DDM, et al. Effects of subfertility cause, smokingand body weight on the success rate of IVF. Hum Reprod 2005;20:1867–75.
15. Wang JX, Davies M, Norman RJ. Body mass and probability of pregnancyduring assisted reproduction treatment: retrospective study. Br Med J2000;321:1320–1.
16. Rittenberg V, Seshadri S, Sunkara SK, Sobaleva S, Oteng-Ntim E, El-Toukhy T. Effect of body mass index on IVF treatment outcome: an up-dated systematic review and meta-analysis. Reprod Biomed Online 2011;23:421–39.
17. Koning AM, Mutsaerts MA, Kuchenbecher WK, Broekmans FJ, Land JA,Mol BW, et al. Complications and outcome of assisted reproductiontechnologies in overweight and obese women. Hum Reprod 2012;27:457–67.
18. Luke B, BrownMB, Stern JE, Missmer SA, Fujimoto VY, Leach R. Female obe-sity adversely affects assisted reproductive technology (ART) pregnancy andlive birth rates. Hum Reprod 2011;26:245–52.
19. Pinborg A, Gaarslev C, Hougaard CO, Nyboe AA, Andersen PK, Boivin J,et al. Influence of female bodyweight on IVF outcome: a longitudinal multi-centre cohort study of 487 infertile couples. Reprod Biomed Online 2011;23:490–9.
20. Dechaud H, Anahory T, Reyftmann L, Loup V, Hamamah S, Hedon B. Obesitydoes not adversely affect results in patients who are undergoing in vitro fer-tilization and embryo transfer. Eur J Obstet Gynecol Reprod Biol 2006;127:88–93.
21. Matalliotakis I, Cakmak H, Sakkas D, Mahutte N, Koumantakis G, Arici A.Impact of body mass index on IVF and ICSI outcome: a retrospective study.Reprod Biomed Online 2008;16:778–83.
22. Bakos HW, Henshaw RC, Mitchell M, Lane M. Paternal body mass index isassociated with decreased blastocyst development and reduced live birthrates following assisted reproductive technology. Fertil Steril 2011;95:1700–4.
23. Keltz J, Zapantis A, Jindal SK, Lieman HJ, Santoro N, Polotsky AJ. Overweightmen: clinical pregnancy after ART is decreased in IVF but not in ICSI cycles. JAssist Reprod Genet 2010;27:539–44.
24. Rachootin P, Olsen J. Prevalence and socioeconomic correlates of subfe-cundity and spontaneous abortion in Denmark. Int J Epidemiol 1982;11:245–9.
25. Schmidt L. Infertility and assisted reproduction in Denmark. Epidemiologyand psychosocial consequences. Dan Med Bull 2006;53:390–417.
26. Danish Fertility Society. Annual report 2011. Available at: http://www.fertilitetsselskab.dk/images/stories/dfs2011.pdf. Accessed October 4, 2012.
27. Moragianni VA, Jones SM, Ryley DA. The effect of body mass index on theoutcomes of first assisted reproductive technology cycles. Fertil Steril 2012;98:102–8.
28. Fedorcsak P, Storeng R, Dale PO, Tanbo T, Abyholm T. Obesity is a risk factorfor early pregnancy loss after IVF or ICSI. Acta Obstet Gynecol Scand 2000;79:43–8.
29. Fedorcsak P, Dale PO, Storeng R, Ertzeid G, Bjercke S, Oldereid N, et al. Im-pact of overweight and underweight on assisted reproduction treatment.Hum Reprod 2004;19:2523–8.
30. Maheshwari A, Stofberg L, Bhattacharya S. Effect of overweight and obesityon assisted reproductive technology—a systematic review. Hum Reprod Up-date 2007;13:433–44.
31. Bellver J, Rossal LP, Bosch E, Zuniga A, Corona JT, Melendez F, et al. Obesityand the risk of spontaneous abortion after oocyte donation. Fertil Steril2003;79:1136–40.
32. Bellver J, Melo MA, Bosch E, Serra V, Remohi J, Pellicer A. Obesity and poorreproductive outcome: the potential role of the endometrium. Fertil Steril2007;88:446–51.
VOL. - NO. - / - 2013
33. Hakonsen LB, Thulstrup AM, Aggerholm AS, Olsen J, Bonde JP,Andersen CY, et al. Does weight loss improve semen quality and reproduc-tive hormones? Results from a cohort of severely obese men. Reprod Health2011;8:24.
34. Reis LO, Dias FG. Male fertility, obesity, and bariatric surgery. Reprod Sci2012;19:778–85.
35. Jokela M, Elovainio M, Kivim€aki M. Lower fertility associated with obesityand underweight: The US National Longitudinal Survey of Youth. Am JClin Nutr 2008;88:886–93.
