{"paper_id":"dd6e5fa0-e1ca-420e-9d61-c94817ab32c3","body_text":"Obesity is an important risk factor for health problems\nand is deemed to be 1 of the 10 global diseases that contributes\nto an increased health burden. There is a rapidly\nincreasing incidence of this complication in many industrialized\ncountries, particularly the United States, and in\ndeveloping Asian countries ( 1 ).\nIn numerous studies, researchers evaluated the effects\nof obesity on assisted reproductive technology (ART) cycle\noutcomes in women ( 2 - 13 ) and reported inconsistent\nresults. Koning et al. ( 14 ) in a review article, reported that\nthere were limited data despite 14 available studies in this\narea and concluded that further studies were needed to\nachieve an accurate insight.\nCurrently, there is no evidence to indicate that obesity\nincreases the risk for ART complications; however, some\nresearchers have reported the negative effects of obesity\non pregnancy rates ( 14 ). In contrast, a review article\npublished by Rittenberg et al.( 15 ) reported an association\nbetween obesity and excess weight in women with\npoor pregnancy outcomes. This finding included reduced\nrates for clinical pregnancy and live births. Luke et al. ( 7 )\nconcluded that obesity had a negative impact on clinical\npregnancy and live birth rates along with ART cycles with\nautologous oocytes. They emphasized that this risk could\nbe brought under control by the use of donor oocytes.\nThe mechanism of the effects of female obesity on ART\noutcomes is controversial. The impact of obesity on ART\noutcomes in men is less studied ( 1 ,  16 ,  17 ) with conflicting\nresults. A systemic review and meta-analysis by MacDonald\net al. ( 18 ) published in 2010, has found no evidence of\na relationship between increased body mass index (BMI)\nand semen parameters. Thus, further studies would be \nwarranted in this field. Petersen et al. ( 19 ) reported that \nmaternal and paternal BMI, both independently and combined, \nexerted negative effects on live birth rates after \nin vitro fertilization (IVF) cycles, but this association in \nintracytoplasmic sperm injection (ICSI) cycles was less \nobvious. In light of the current evidence, we designed the \npresent study to assess the impacts of obesity in a couple \non ART outcomes. This study sought to answer the question \nof whether obesity simultaneously in a couple has a \nnegative effect on ICSI cycle outcomes in comparison to \ncouples who have normal BMIs.\n\nThis was a cross-sectional study performed at Royan \nInstitute between January 2013 and January 2014. The \nReview Board and Ethics Committees of Royan Institute \napproved the study protocol. All participating couples \nprovided ethical permission at their initial visit for the use \nof their treatment outcomes. Participant confidentiality \nfor all participants was assured during the research and \nwritten informed consent was obtained from them.\nWe evaluated the data recorded during the study period \nfrom all of the study participants. The study population \nwas limited to patients who underwent ICSI or IVF/ICSI \ncycles that resulted in the transfer of 2 or 3 fresh embryos. \nHeight and weight were recorded for all couples. Couples \nwhose female partner was =39 years of age and the male \npartner was <55 years of age ( 17 ) at the time of the treatment \ncycle onset were enrolled to minimize the effect of \nage as a confounding factor. We excluded all cases with \nuterine factor, severe male factor, severe endometriosis, \nand gamete or embryo donor recipients ( Fig .1 ).\nSampling procedure and the distribution of the couples according to their BMI. IVF; In vitro fertilization, ICSI; Intra-cytoplasmic sperm injection, BMI; \nBody-mass index, IUI; Intrauterine insemination, PGD; Pre-gestational diagnosis, TESE; Testicular sperm extraction, PESE; Percutaneous epididymal sperm \nextraction, PESA; Percutaneous epididymal sperm aspiration, and TESA; Testicular sperm aspiration.\nThe patients’ age (years) was recorded at the beginning \nof treatment. At the onset of treatment, we classified \nparticipants as smokers or non-smokers according \nto the number of cigarettes smoked per day. The \ndiagnosis of infertility was determined according to \nthe 10th revision of the International Classification of \nDiseases ( 11 ). Accordingly, women participants were \ncategorized as ovulatory or an ovulatory. Standard \novarian stimulation protocols were performed according \nto routine clinical practice. In brief, suppression \nof the endogenous luteinizing hormone surge was performed \nwith either gonadotropin-releasing hormone \nagonists or antagonists. Controlled ovarian stimulation \nwas performed with recombinant follicle-stimulating \nhormone (rFSH) and/or human menopausal gonadotropin \n(hMG); trans-vaginal ultrasound guided ovum \npickup was performed 34-36 hours after administration \nof human chorionic gonadotropin (hCG). ICSI for retrieved \nMII oocytes, with or without insemination, was \nperformed in accordance with standard general recommendations.\nWe defined normal fertilization as the appearance of \nthe 2 nd  polar body at 16-19 hours after insemination \nor microinjection. In our institute, embryo quality is \ngraded as A, B, C, and D, with \"A\" defined as the best \nquality and \"D\", the worst, according to cell numbers, \npercentage of fragmentation, and cell symmetry. All \nembryo transfers were performed with a Labotect catheter \n(Labotect, Germany) by experienced gynaecologists \nand embryologists on day 3 after IVF/ICSI. Luteal \nphase support was provided by administration of 400 \nmg of vaginal progesterone twice a day until the day of \nthe ß-hCG test. Luteal support with progesterone was \nprescribed until the observation of foetal heart activity \nand subsequently tapered until week 8 of gestation. The \nmain outcomes were fertilization, implantation, clinical \npregnancy, and live birth rates. The implantation \nrate was denoted as the number of visualized intrauterine \ngestational sacs divided by the number of transferred \nembryos. A clinical pregnancy was documented \nby ultrasound observation of an intrauterine gestational \nsac with foetal cardiac activity. We defined spontaneous \nabortion as the loss of clinical pregnancy prior to \n20 weeks gestation. Trained nurses routinely measured \nheight and weight in participants of both genders prior \nto the onset of the treatment cycle. The balance scale \nfor the measurement of weight was calibrated daily \nand verified by a one kg counterweight. We used the \nWorld Health Organization’s definition of BMI (kg/m 2 ) \nto classify male and female participants as underweight \n(<18.5 kg/m 2 ), normal (18.5-24.9 kg/m 2 ), overweight \n(25-29.9 kg/m 2 ), or obese (=30 kg/m 2 ) ( 20 ). The small \nnumber of underweight couples precluded their inclusion \nin the couples’ analysis. We divided the couples \ninto 3 groups based on male and female BMI results: \ngroup 1 (normal weight), group 2 (overweight), and \ngroup 3 (obese). The main outcomes were compared \namong the three groups.\nStatistical analysis was carried out using the Statistical \nPackage for the Social Sciences (SPSS), version \n20.0 (SPSS Inc., Chicago, IL, USA). The study population’s \ncharacteristics were compared according to the \ncouples’ BMI (normal, overweight, and obese) using \none-way analysis of variance (ANOVA), Kruskal-\nWallis nonparametric analysis of variance, and the \nchi-square test when appropriate. Multilevel logistic \nregression analysis was applied to determine the \nodds of live births following ICSI cycles. The analysis \nwas conducted according to the female and male \nBMI groups. Normal-weight patients were considered \nto be the reference group. Analysis of female BMI was \nadjusted for age and duration of infertility. Likewise, \nanalysis of the male BMI was adjusted for age, duration \nof infertility, and smoking status.\nA multilevel logistic regression analysis was used to \ndetect the predictive factors for live births after ICSI \ncycles. All possible factors that affected the live birth \nrate, which included female and male ages, couples’ \nBMI (<25 kg/m 2  and =25 kg/m 2 ), male smoking status, \ncause and duration of infertility, ovarian stimulation \nprotocol [long gonadotropin-releasing hormone \n(GnRH) agonist and GnRH antagonist protocols], and \nnumber and quality of transferred embryos were incorporated \ninto the model. The results of the multilevel \nlogistic regression analysis have been presented as adjusted \nodds ratios (ORs) with 95% confidence intervals \n(CIs). P values <0.05 were considered statistically \nsignificant.\n\nIn total, there were 4203 ART cycles during the study \nperiod. A total of 990 eligible women and their husbands \nunderwent 927 ICSI and 63 ICSI with insemination \n(IVF/ICSI) cycles according to the inclusion \ncriteria. The sampling procedure and distribution of \nthe couples according to their BMI has been illustrated \nin ( Fig .1 ). According to BMI, of the 990 women participants, \nthere were 59 (6%) underweight, 357 (36%) \nwith normal weight, 412 (41.6%) overweight, and 162 \n(16.4%) obese participants. Of the 990 men evaluated, \nthere were 18 (1.8%) underweight, 325 (32.7%) normal \nweight, 425 (43%) overweight, and 223 (22.5%) obese \nparticipants.\nThe characteristics of the study population according to \ngender and BMI have been presented in ( Table 1 ). The \ndistribution of smoking in the males significantly differed \namong the BMI groups (P=0.006). The majority of females \n(n=786, 79.3%) had normal menses and ovulation. \nAs expected, there were more anovulatory cases in the \nobese group than in the other groups (P=0.003).\nAnovulatory cases in the present study consisted of \nparticipants with PCOS (n=153, 75%) and age factor \n(over 37 to 39 years, n=51, 25%).\nBasic characteristics of the studied population according to gender and body mass index\nData are presented as mean ± SD or n (%).\nWe separately evaluated the impact of female and male \nBMI on the live birth rate in ICSI. The results of the multilevel \nregression analysis according to female and male \nBMI has been shown in ( Table 2 ). Among the ovulatory \nwomen, there was a significant difference between the \nBMI groups, with a 60% (95% CI: 0.11-0.83) decrease in \nthe odds for live birth among overweight individuals and \n84% (95% CI: 0.02-0.99) decrease in the odds for live \nbirth among obese individuals. Trend analysis showed a \nsignificant reduction of 9% (95% CI: 0.83-0.99) with each \none unit increase in BMI (P=0.04). Among anovulatory \nwomen, the association between BMI and live births presented \nno clear tendencies, even though the ORs indicated \nlower probabilities for live births among overweight and \nobese anovulatory women. The 95% CIs were not significant. \nAmong anovulatory women, the trend analysis \nshowed a significant reduction of 15% (95% CI: 0.72-\n0.98) with every one unit increase in BMI (P=0.02). In \nboth ovulatory and anovulatory underweight women, we \nobserved a significant elevation in the odds of live births \nof 6.5 times (95% CI: 2.1-20.65) and 7.3 times (95% CI: \n0.99-55.1), but the CIs were too wide because of the low \nsample size. The results for men participants presented no \nsignificant relationship between BMI and live births.\nThe comparison of the three groups of couples according \nto BMI has been demonstrated in ( Table 3 ). \nThe overweight (P=0.01) and obese (P<0.001) couples \nwere significantly older than normal weight couples. \nThe results indicated that the three groups were comparable \nin terms of type, cause of infertility, number \nof previous ART cycles, and stimulation outcomes. \nThere were no significant differences between the three \ngroups in terms of fertilization, implantation, clinical \npregnancy, multiple pregnancy, miscarriage, and live \nbirth rates.\nMultilevel logistic regression analyses of the probability for live births following all ICSI or IVF/ICSI cycles according to gender and stratified by BMI\nOR; Odds ratio, CI: Confidence interval, BMI; body mass index, IVF; In vitro fertilization, ICSI; Intra-cytoplasmic sperm injection,  a ; ORs with 95% CIs and P values from\nWald tests, and  b ; Female analyses adjusted for age and duration of infertility. Male analyses adjusted for age, duration of infertility, and smoking status.\nComparison of study population characteristics and cycle outcomes among three groups of couples according to BMI\nData are presented as mean ± SD or n (%). BMI; Body mass index, FSH; Follicle stimulating hormone, LH; Luteinizing hormone, AMH; Anti-Müllerian hormone, TSH; Thyroid stimulating hormone, ART; Assisted reproductive technology, rFSH; Recombinant follicle-stimulating hormone, ANOVA: One-way analysis of variance,  a ; Obese couples vs. overweight couples (P=0.015), normal BMI vs. obese couples (P=0.040) according to Tukey’s test,  b ; Normal BMI vs. overweight couples (P=0.013), normal BMI vs. obese couples (P<0.001), overweight vs. obese couples (P=0.086) according to Tukey’s test,  c ; Normal BMI vs. overweight couples (P=0.032), normal BMI vs. obese couples (P=0.021) according to Tukey’s test, *; Good quality embryos-all ET were A, B, or AB, Fair-half of ET were good quality (AC, BC), Poor quality-all ET were C, D, or CD.\nThe results of the multilevel logistic regression model for the detection of the predictive factors for the live birth rate showed that none of the included variables remained in the final model as significant factors. The results also revealed no significant association between the couples’ BMI and live births ( Table 4 ).\nMultilevel logistic regression analysis for detection of predictive factors for live birth after ICSI or IVF/ICSI cycles in the studied population\nICSI; Intracytoplasmic sperm injection, IVF; In vitro fertilization, BMI; Body mass index, OR; Odds ratio, and CI; Confidence interval.\n\nPrevious studies separately evaluated the effects of both \ngenders’ BMI on ART outcomes. The synergistic effects \nof obesity in couples were reported in limited studies ( 8 , \n 19 ). We have excluded the main confounding factors that \naffect live birth rates in order to accurately assess the independent \neffects of a couple’s obesity on ART outcomes. \nOur results revealed that a couple’s BMI had no effect on \nthe outcomes of ICSI with fresh cleavage-stage embryo \ntransfer cycles.\nOur results supported those published in 2013 by Petersen \net al. ( 19 ), who reported that the combined increased \nmaternal and paternal BMI had no significant effect \non live birth rate in ICSI cycles. However, the authors \nhave presented the negative impacts of increased female \nand male BMI, both individually and combined, on live \nbirths in IVF cycles. In our institute, treatment cycles with \nonly IVF are uncommon and the majority of treatment cycles \ninclude ICSI or IVF/ICSI procedures. Therefore, we \ncould not evaluate these subjects according to IVF cycles.\nThe effects of female BMI on ART outcomes were evaluated \nin several studies. Our findings showed that among \novulatory women, BMI had a negative impact on live \nbirths. In anovulatory women, we observed a tendency \nfor less odds of live births in the obese group, which was \nnot statistically significant. Therefore, our results agreed \nwith some recent studies where female BMI negatively \nimpacted ART outcomes ( 8 ,  11 ,  13 ,  15 ,  20 ). On the other \nhand, previous studies indicated no negative effect of female \nBMI on ART outcomes ( 2 ,  12 ,  21 - 23 ). Petersen et \nal. ( 19 ) demonstrated that the female BMI had a negative \nimpact on live birth rates in IVF cycles, but this was \nless clear in ICSI cycles. A prospective study conducted \nby Chavarro et al. ( 24 ) evaluated 170 women who underwent \n233 ART cycles and found an association between \noverweight and obese women with decreased live birth \nrates. Moragianni et al. ( 25 ), in a retrospective research \nof 4609 patients, found that obesity had significant negative \neffects on ART outcomes, with up to 68% lower odds \nof live births following the first ART cycle. Rittenberg et \nal. ( 15 ), in a meta-analysis of 47967 IVF/ICSI cycles, reported \nthat an increased female BMI was aligned with adverse \npregnancy outcomes in IVF/ICSI treatment cycles \nand this effect was observed in both overweight and obese \nwomen. Since the earlier investigations did not categorize \ntheir findings according to type of treatment cycle (IVF \nor ICSI), a more adverse influence of increased BMI in \nIVF compared to ICSI might have been concealed and \nthe intensity of the BMI impact on IVF/ICSI possibly relied \non the IVF and ICSI cycle distributions in the sample \nsize ( 19 ). Although a number of multiparous women are \nobese, a negative association of obesity with women’s reproductive \nhealth has been reported ( 26 ). Because of the \nconflicting results reported by studies, the mechanism action \nof maternal obesity on IVF or IVF/ICSI outcomes \nremains unclear ( 27 ). Although a number of oocyte donation \nstudies have suggested negative effects of obesity \non the endometrium ( 6 ,  28 ), others have not ( 2 ,  21 ,  29 ). \nEndocrine changes related to obesity such as hyperandrogenism \nand insulin resistance as well as alterations in the \nlocal insulin-like growth factors (IGFs), cytokines, and \nleptin levels may play a major role in the adverse effects \nof an increased BMI on ART outcomes ( 4 ). According to \nprevious studies ( 13 ), the mechanism of action of obesity \nin anovulatory cases, especially PCOS women, is different \nand depends on the intensity of the endocrine changes.\nThe influence of male BMI on ART outcomes has been \nless studied. The existing literature contains only 7 studies \non this subject ( 1 ,  8 ,  11 ,  16 ,  17 ,  19 ,  30 ). The first study, \npublished in 2011 by Bakos et al. ( 1 ), reported an association \nbetween high paternal BMI with significantly reduced \nclinical pregnancy and live birth rates after ART. \nTwo recent studies presented that male BMI was associated \nwith a negative impact on clinical pregnancy and \nlive birth rates after IVF, but not after ICSI. Additionally, \nthe previous studies in this field reported that ICSI might \novercome the negative impact of obesity on sperm-oocyte \ninteraction ( 16 ,  19 ). On the other hand, a prospective \nstudy conducted by Colaci et al. ( 17 ) evaluated 114 \ncouples who underwent 172 ICSI cycles and concluded \nthat male obesity was associated with decreased odds \nfor live births after ICSI. Our results indicated that the \nmale BMI had no effect on live birth rates after ICSI. The \ndeleterious effects of male obesity could be due to an altered \nhormonal profile and decreased semen quality related \nto increased leptin and E2 levels, and disturbance in \nspermatogenesis ( 19 ,  31 ). However, a systematic review \nwith meta-analysis found no evidence of an association \nbetween an increased BMI and semen parameters ( 32 ). \nA systematic review by Campbell et al. ( 33 ) in 2015 reported \nthat the rate of births per ART cycle was reduced \nby 35% in obese men. The salient weak point in the previous \nstudies and our study was the use of BMI as a marker\nof body fat in men. In view of these conflicting results, \nwe suggest that prospective studies should evaluate the \neffects of male and female abdominal obesity on reproductive \nparameters and ART outcomes via other anthropometric \nmeasurements (waist and hip circumferences). \nCurrently, the role of the male BMI in ART processes and \noutcomes is partly understood. Further investigations are \nneeded to arrive at reliable conclusions ( 13 ).\nWe analysed the couples and found no synergistic negative \nimpact of increased female and male BMI on live \nbirths after ICSI cycles. This finding agreed with studies \nby Petersen et al. ( 19 ) and Schliep et al. ( 34 ). Some \nstudies assessed the effects of combined male and female \nBMI on ART outcomes ( 4 ,  8 ,  10 ,  19 ). Petersen et al. ( 19 ) \nevaluated the effects of parental BMI on live birth rates \nafter ART cycles. They reported that increased combined \nfemale and male BMI had a negative impact on live birth \nrates after IVF cycles; however, its effects in terms of ICSI \nwere less clear. Schliep et al. performed a prospective assessment \nof 721 couples and found no influence of the \ncouples’ weight status on IVF outcomes ( 34 ). In contrast, \na recent study by Wang and colleagues retrospectively investigated \n12061 first fresh IVF/ICSI cycles and reported \nthat female obesity exerted negative effects on live births \nafter IVF; nonetheless, there was no evidence of a negative \nimpact by the parental BMI on ICSI outcomes ( 4 ). \nIn contrast, Umul et al. ( 10 ) found that couples’ obesity \nhad a negative impact on clinical pregnancy rates and live \nbirth rates following ICSI cycles. In the present study we \nmeticulously analysed the characteristics of the couples in \nICSI cycles and adjusted the impact of confounding factors \non our results. Recent data have confirmed the findings \nof those previous studies that reported no significant \ninfluence of the parental BMI on ICSI success. In view of \nthe conflicting results, we suggest that more research be \nundertaken to shed sufficient light on this issue.\nThe present study has some limitations. There was no \ndata about the specific hormonal profile and android or \ngynoid distribution of fat in anovulatory and ovulatory \nwomen, and no data about semen analysis parameters to \ncompare among different BMI groups. We propose that \nthese parameters should be considered in future studies.\n\nBased on the current findings, an increased maternal \nBMI independently influenced negatively live birth rate \nafter ICSI cycles, whereas increased paternal BMI separately \nand in combination with maternal BMI did not \nshow this affect.","source_license":"CC-BY-4.0","license_restricted":false}