Original Article Systematic Review The Effect of Lifestyle Intervention on Pregnancy and Birth Outcomes on Obese Infertile Women: A Systematic Review and Meta-Analysis Juan J Espinós, M.D.1*, Ivan Solà, M.D., Ph.D.2, 3, 4, Claudia Valli, M.Sc.2, Ana Polo, M.D.5, Lucja Ziolkowska, M.D.2, 6, M José Martínez-Zapata, M.D.2, 3, 4 1. Department of Obstetrics and Gynaecology, Hospital de la Santa Creu i Sant Pau, Barcelona, Universitat Autònoma de Barcelona (UAB), Bellaterra, Spain 2. IbCC Iberoamerican Cochrane Center, Barcelona, Spain 3. CIBERESP CIBER Epidemiología y Salud Pública (CIBERESP), Barcelona, Spain 4. IIB Sant Pau Biomedical Research Institute Sant Pau (IIB Sant Pau), Barcelona, Spain 5. Department of Reproduction Puigvert Foundation, Barcelona, Spain 6. Medical University of Silesia, Katowice, Poland Abstract Obesity has been associated with negative effects on natural fertility and poor prognosis when assisted reproduc- tive technologies (ART) are performed. Patients attending for fertility treatments are often advised to optimize their weights to improve the outcomes. There is lack of enough information on how weight-loss would be effective for improving fertility in women who are overweight or obese. We conducted a systematic review to evaluate whether weight-loss achieved by lifestyle program improves natural or assisted reproduction in obese infertile women. We searched CENTRAL, MEDLINE, and EMBASE up to March 2018. Two reviews were selected as randomised tri- als assessing a lifestyle intervention in women with obesity before receiving treatments for infertility and appraised their risk of bias. We extracted data on pregnancy, birth, and miscarriage rates as the primary outcomes and pooled effect estimates using a random effects model. The primary outcome was the live birth rate. We reported summary measures as the relative risk (RR), 95% confidence interval (CI), and percentage of heterogeneity (I2). We included eight randomised trials with 1175 women. Lifestyle programmes, improved pregnancy rates (RR: 1.43, CI: 95% 1.02 to 2.01; I2=60%; 8 RCTs; N=1098) but had no impact on live births (RR: 1.39, CI: 95% 0.90 to 2.14; I2=64%; 7RCTs; N=1034). Our findings suggest that women participating in lifestyle interventions had an increased risk of miscarriage (RR: 1.50, CI: 95% 1.04 to 2.16; I2=0; 6RCTs; N=543). We rated the quality of evidence for these outcomes as the moderate-to-low. Lifestyle interventions slightly increased the pregnancy rate, while it would be uncertain whether it can improve the live birth. Lifestyle interventions can increase the risk of miscarriage. More research is needed to further explore lifestyle interventions on reproductive outcomes in obese infertile women. Keywords: Diet, Infertility, Live Birth Rate, Obesity, Physical Exercise Citation: Espinós JJ, Solà I, Valli C, Polo A, Ziolkowska L, Martínez-Zapata MJ. The effect of lifestyle intervention on pregnancy and birth outcomes on obese infertile women: a systematic review and meta-analysis. Int J Fertil Steril. 2020; 14(1): 1-9. doi: 10.22074/ijfs.2020.5921. This open-access article has been published under the terms of the Creative Commons Attribution Non-Commercial 3.0 (CC BY-NC 3.0). Introduction implantation, as well as reducing the fertility rate in The prevalence of overweight and obesity among women with a normal menstrual cycle (4-7). The extent of women have increased more than three times in the impact of obesity on in vitro fertilization (IVF) outcomes last years, creating a global pandemic affecting both is unknown due to the heterogeneity of studies conducted industrialized and developing countries (1, 2). Obesity in this area, the retrospective nature of most investigations, has been associated with negative effects on both general and lack of standardized criteria (8-10). Obesity has been and reproductive health. Natural fertility is compromised associated with an increase in gonadotropin need, more in both, men and women (3). In the last, polycystic days of treatment, higher cancellation rates of cycles ovarian disease (which is typically associated with due to the inadequate response, decreased numbers of central obesity, insulin resistance, and hyperinsulinism) total and mature eggs, reduced rates of fertilization, and and alterations affecting obesity-related hormones (e.g., consequently fewer high-quality embryos. Obesity has leptin, adipokines, ghrelin, and endorphins) can affect also been associated with endometrial abnormalities and oocyte quality, fertilization, embryo development, and lower implantation rates (11-14). Received: 1/ April/2019, Accepted: 5/October/2019 *Corresponding Address: Department of Obstetrics and Gynecology, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona (UAB), Bellaterra, Spain Email: jespinos@santpau.cat Royan Institute International Journal of Fertility and Sterility Vol 14, No 1, April-June 2020, Pages: 1-9 1 Lifestyle Intervention in Obese Infertile Women Weight-loss has been appreciated as one of the most Outcomes effective means of increasing the probability of fertility We set the following primary outcomes: live birth in infertile overweight or obese women (15, 16). Few (including spontaneous live birth, IVF live birth and studies have analyzed the actual effects of a lifestyle cumulative live birth per initial cycle), cumulative intervention, including diet and exercise on obese women pregnancy rate and miscarriage (pregnancy ending within wishing to become pregnant. Additionally, the findings the first 20 weeks of gestation). Secondary outcomes were of these studies have been inconsistent, probably owing pregnancy (including multiple pregnancies), ongoing to methodological shortcomings (17). A prior systematic pregnancy, and implantation rates. review, including randomized and non-randomized controlled trials and studies using weight reduction Data extraction and risk of bias assessment drugs showed an increase in the feasibility of becoming pregnant, with no significant adverse effect on live birth Two authors extracted independently the relevant rates (18). data from chosen trials using a predefined extraction form and an additional author revised the process for In this systematic review, we aimed to evaluate whether accuracy. We registered the characteristics of included weight-loss achieved by a lifestyle intervention improved studies in descriptive tables. We contacted authors from the pregnancy outcomes in obese infertile women, with a included studies to request missing data in published specific focus on the live birth rate. papers. Materials and Methods We assessed independently the risk of bias from included trials using the Cochrane tool for that purpose We conducted this systematic review according to the (21). We assessed the trial randomisation sequence methodological guidance of Cochrane (19). We reported generation and its concealment, the concealment of the the findings from the review according the PRISMA intervention to participants, researchers, and outcomes statement (20). assessors, attrition, and incomplete outcome data and selective outcome reporting. Search strategies We searched MEDLINE (via PubMed), EMBASE Data analysis and findings description (via Ovid), and CENTRAL (via The Cochrane Library) We analysed the effect measures for dichotomous from the databases inception up to March 2018. We variables using risk ratios (RR) and mean differences designed a search strategy combining text words and (MD) for continuous variables calculating their 95% controlled vocabulary adapted to the requirements confidence intervals (CI). We considered statistic of each database. We included the complete search significant difference between compared groups when strings in the Materials S1 (See Supplementary Online 95% CI was not included. The unit of the analysis of Information at www.ijfs.ir). Additionally, we searched interest was the participants in included trials and we the reference list of all eligible studies and contacted used the available-case analysis approach to calculate the authors of the included trials to request additional effect estimates. information. When appropriate, we calculated pooled effect estimates for each outcome using a fixed-effect model Study selection or a random effect model when there was statistical We included randomised controlled trials assessing a heterogeneity (22). We assessed heterogeneity lifestyle intervention in obese women before receiving comparing characteristics from included studies and treatments for infertility. The lifestyle interventions that through the I square statistics (23) considering a we considered in this study consisted of any type of substantial statistical heterogeneity for values greater structured physical exercise and/or any low calorie intake than 50% and considerable heterogeneity for values diet refered by the primary included studies. Eligible greater than 75% scenario in which we did not perform trials included women with a body mass index of 29 or the pool effect estimates. We performed sub-group higher who were candidates for IVF. The selected trials analyses according to the lifestyle programme assessed assessed the structured health promotion programmes in the included trials (diet alone or combined with consisting of dietary intake reduction alone or combined physical activity). We planned sensitivity analyses with physical activity compared with an inactive control excluding trials with the highest risk of bias or those group (e. g. women on a waiting list) or women receiving that were a suspected source of heterogeneity. As any weight loss advice. Three authors independently evaluated pooled analyses included more than 10 trials, we were whether the references retrieved from the searches met not able to conduct formal tests to assess the impact of the inclusion criteria and resolved disagreements by publication bias (24). We used the statistical package discussion or through adjudication by an additional author. in the open access software Review Manager (v 5.3.5) We obtained full copies of eligible references for a final to conduct all of the analyses (25). We assessed the decision with respective to their inclusion and reported the quality of evidence to judge the confidence in the effect reason that led to exclusion of studies. estimates obtained from each primary outcome. We Int J Fertil Steril, Vol 14, No 1, April-June 2020 2 Espinós et al. rated the quality of evidence as high, moderate, low reporting weight-loss due to the presence of high or very low according to the impact of each outcome heterogeneity (95%). on the risk of bias, indirectness, and effect estimates inconsistency, and imprecision (26). We summarized Risk of bias the effect estimates for primary outcomes and their Most trials implemented random sequences quality of evidence in a summary of the Table of generated adequately using lists of computer findings (27). generated numbers (16, 29-34) and had proper Results allocation concealment, using opaque envelopes in most of the cases (16, 31-34). With the exception of Study selection and characteristics one trial (30), the rest was open or did not provide Our search strategy yielded 726 records of which 48 details on blinding of researchers or participants, but were potentially eligible to be included. The flowchart four implemented a blinded outcome assessment (29- (Fig.1) describes the complete eligibility process, and 32). We considered three trials having high risk of we describe the reasons for excluding 40 studies and bias because the data available for the analysis were the main characteristics of eight included trials (28- partially complete (16, 28, 32). Finally, two trials 34) in the Materials S2 (See Supplementary Online had high risk of selective reporting bias because Information at www.ijfs.ir) and the Table 1, respectively. some outcomes included in their protocols did not Table 2 shows the summary of findings of the review coincide with those reported in the published reports Fig.1 with a judgement on their quality of evidence. of their findings (Fig.2) (28, 34). 726 records identified from bibliographic databases 39 CENTRAL 129 duplicates removed 264 MEDLINE 423 EMBASE 597 records screened 546 records excluded according inclusion criteria by title or abstract 43 articles excluded (See Materials S2) 20 cohort studies 51 full text articles 9 ineligible study designs assessed for eligibility 7 secondary publications from included studies 5 did not include eligible outcomes 2 ineligible interventions 8 trials included Fig.1: Flowchart for study eligibility. In total, we included 1175 infertile women. The mean age ranged from 29 to 34 years old, and the body mass index (BMI) from 24 to 38. The included trials compared lifestyle-structured programmes with the usual care. The assessed programmes consisted of dietary intake reduction (28-30) or combined with physical activity interventions (6, 30-33). Women in control groups immediately received infertility treatment with no history of interventions or were included in a waiting list for IVF (28-30, 32) or received standard advice for weight-loss (16, 31, 34). All lifestyle interventions significantly reduced the weight of infertile women compared with control group in the Materials S3 (See Supplementary Online Information at www.ijfs.ir). The mean weight loos values ranged between 3 and 10 kg at the end of the intervention. We did not pooled the results of studies Fig.2: Risk of bias. 3 Int J Fertil Steril, Vol 14, No 1, April-June 2020 Lifestyle Intervention in Obese Infertile Women Table 1: Characteristics of included studies Study ID, Women Age (Y) Mean BMI at baseline Experimental Control Outcomes Follow-up Funding Setting, years (SD) Mean (SD) intervention intervention (months) country Experimental/ Experimental/ control group control group Becker et al. (28), 35 31.36 (SE 0.89)/ 28.67 (SE 0.60)/ Hypocaloric diet Maintenance of Live birth (spontane- 12 Not reported 2015 Obstetrics 31.25 (SE 0.78) 28.82 (SE 0.98) with a low gly- the body weights ous) ndesirable and Gynaecol- cemic index and and usual diets effects (miscarriage) ogy Service of low glycemic Pregnancy rate (clin- the Hospital de load ical) BMI change Clinicas de Porto Weight change Alegre, Brazil Einarsson et al. 317 31.5 (4.3)/ 33.1 (1.3)/ A low calorie IVF with no previ- Live birth 12 Sahlgren- (29), 2017 Infer- 31.7 (4.1) 33.0 (1.5) liquid formula ous interventions (spontaneous ska University tility clinics Swe- diet of 880 kcal/ IVF) Undesirable Hospital (ALF- den, Denmark day effects (miscarriage, GBG-70 940), and Iceland ectopic pregnancy) Merck AB, Pregnancy rate Solna, Sweden (clinical, multiple) (an affiliate of BMI change Weight Merck KGaA, change Darmstadt, Germany), Impolin AB, Hjalmar Svensson Foun- dation and Jane and Dan Olsson Foundation Espinós et al. 41 32.0 (3.2)/ 34.6 (3.0)/ Diet and exercise IVF/ICSI with no Live birth (IVF, 12 Grant from the (30), 2017 32.9 (3.9) 34.0 (4.1) previous interven- cumulative) Undesir- Instituto de Fertility Unit of tions able effects (miscar- Salud Carlos III Hospital de la riage) Pregnancy rate (PI11/02816) Santa Creu i Sant (clinical, multiple) Pau-Fundacio Pu- Weight change igvert, Barcelona Implantation rate Spain Fertilization rate Moran et al. (31), 46 33.8 (3.5)/ 34.0 (4.5)/ A nutritionally A standard advice Live birth Undesir- Not NHMRC Pro- 2016 Repromed, 32.5 (3.3) 33.9 (4.4) adequate reduced on appropriate able effects (miscar- reported gram Grant to Adelaide Aus- energy diet and diet and lifestyle riage) Pregnancy rate RJN, a Brails- tralia exercise interven- factors influencing BMI change Weight ford Robertson tion andcontact fertility provided change Grant and The with investigators face-to face at one University of session with no Adelaide in active follow-up Adelaide, Australia, and sponsored with a product (Opti- fast VLCD) by Novartis USA Mutsaerts et 577 29.7 (4.5)/ 27.7 Motivational Prompt infertility Live birth Undesir- 24 Grant (50- al. (32), 2016 29.8 (4.6) (range 24.4-31.0)/ counselling: treatment with no able effects (miscar- 50110-96-518) University medi- outpatient previous interven- riage) Pregnancy rate from the cal centres and visits, telephone tions (clinical, multiple) Netherlands general hospitals consultations, BMI change Weight Organization Netherlands assistance of an change for Health online diet diary, Research and advise to engage Development in moderate intensity physical activity Palomba et al. 96 28.43 (8.31)/ 31.05 (2.98)/ Structured 2 weeks of obser- BMI change Weight Not Not reported (33), 2010* 26.50 (4.26) 32.3 (3.73) exercise training vation followed change Ovulation reported Setting Units of plus hypocaloric by one cycle of rate Reproductive Reproductive diet for 6 weeks, CC therapy outcomes Changes Medicine and with one cycle of in anthropometric Surgery Italy CC after the first and hormonal and 2 weeks metabolic param- eters Compliance with the interven- tions Int J Fertil Steril, Vol 14, No 1, April-June 2020 4 Espinós et al. Table 1: Continued Study ID, Women Age (Y) Mean BMI at baseline Experimental Control Outcomes Follow-up Funding Setting, years (SD) Mean (SD) intervention intervention (months) country Experimental/ Experimental/ control group control group Rothberg et 14 33 (5.0)/ 41 (4)/ Intensive weight Standard-of-care Live birth Pregnancy 12 Grant from the al. (34), 2016 30 (4.0) 41 (4) loss interventions nutritional coun- rate BMI change Michigan Insti- University of consisted of 12 selling consisted Weight change tute for Clinical Michigan (UM) weeks of very- of 16 weeks of Research (grant Health System, low-energy diet conventional U040012 PI Ann Arbor, (800 kcal/day) food-based diet to A.R.); the Michigan USA plus 4 weeks of core services of a low-calorie the Michigan conventional Nutrition Obe- food-based diet sity Research Centre (grant DK089503); and the Michi- gan Centre for Diabetes Research (grant P30DK020572) Sim et al. (16), 49 32,9 (3.3)/ 35.1 (3.8)/ A very-low-en- Recommenda- Live birth Undesir- 12 National Health 2014, Royal 32,8 (3.1) 38.0 (5.2) ergy diet for the tions for weight able effects (miscar- and Medical Prince Alfred initial 6 weeks loss and the same riage) Pregnancy rate Research Coun- Hospital (RPAH) followed by a printed material as (clinical, assisted, cil of Australia Fertility Unit, hypocaloric diet, the intervention. natural) BMI change and from the Sydney, Australia combined with Weight change Sydney Univer- a weekly group sity Nutrition multidisciplinary Research programme Foundation to KAS. Prima Health Solu- tions provided the VLED (KicStart) SD; Standard deviation, SE; Standard error, CC; Clomiphene citrate, BMI; Body mass index, IVF; In vitro fertilization, ICSI; Intracytoplasmic sperm injection, *; Palomba et al. study had 3 groups, but we include only group B and C described in the Table. The group A received structured exercise training plus hypocaloric diet for 6 weeks without CC. Table 2: Summary of review findings Anticipated absolute effects (95% CI) Relative effect (95% CI) Number of participants Quality of the evidence Outcomes Risk with usual care Risk with lifestyle interventions (*) Live births 242 per 1.000 346 per 1.000 RR 1.43 433 (4 RCTs) IVF live births (181 to 655) (0.75 to 2.71) Low1, 2 Live births 405 per 1.000 563 per 1.000 RR 1.39 1034 (7 RCTs) All live births (365 to 867) (0.90 to 2.14) Low2, 3 All pregnancies 502 per 1.000 718 per 1.000 RR 1.43 1034 (7 RCTs) (507 to 1.000) (1.01 to 2.02) Moderate3 Miscarriage 142 per 1.000 213 per 1.000 RR 1.50 543 (6 RCTs) (148 to 307) (1.04 to 2.16) Moderate4 *; The risk in the intervention group [and its 95% confidence interval (CI)] is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI), 1; Two studies had high risk of performance bias (open trials), and additional one high risk of attrition bias, 2; The confidence interval of effect estimate includes both an effect for the intervention and the control condition, 3; Five studies had high risk of performance bias or detection bias (open trials), and two reported selectively their outcomes, and 4; Four studies had high risk of performance bias or detection bias (open trials), three had high risk of attrition bias and one reported selectively its outcomes. Grade working group grades of evidence: High quality; Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality; Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality; Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality; We are very uncertain about the estimate. Effect of lifestyle interventions in primary outcomes women (RR: 1.43, CI: 95% 1.02 to 2.01; I2=60%; Fig.4) (16, Seven studies reported the live birth with a total 28-31, 33). Twenty-one more women out 100 participating number of 1034 patients (28-34), and showed that in a lifestyle intervention became pregnant in comparison to lifestyle interventions had no effect on live birth rates women receiving usual care (CI: 95% 0.5 to 38 more). (RR: 1.39, CI: 95% 0.90 to 2.14; I2=65%; Fig.3). A subgroup analysis of studies assessing interventions We rated this outcome as low-quality due to limitations based on dietary restriction (28, 29) or in combination with in study designs and imprecision in the effect estimate. On physical activity (16, 30-32, 34) did not show changes any the other hand, the intervention led to higher pregnancy rates the effect estimates magnitude or direction (in the Materials according the pooled results of seven trials including 1098 S4, See Supplementary Online Information at www.ijfs.ir). 5 Int J Fertil Steril, Vol 14, No 1, April-June 2020 Fig.3 Lifestyle Intervention in Obese Infertile Women Lifestyle interventions Usual care Risk ratio Risk ratio Study or subgroup Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI Becker et al. (28), 2015 3 14 0 12 2,1% 6,07 [0,34, 106,85] Einarsson et al. (29), 2017 45 152 42 153 26,3% 1,08 [0,76, 1,54] Espinos et al. (30), 2017 13 21 6 20 16,3% 2,06 [0,98, 4,36] Moran et al. (31), 2016 7 18 5 20 12,5% 1,56 [0,60, 4,04] Mutsaerts et al. (32), 2016 123 280 153 284 30,5% 0,82 [0,69, 0,97] Rothberg et al. (33), 2016 3 6 0 5 2,3% 6,00 [0,38, 94,35] Sim et al. (16), 2014 12 27 3 22 10,0% 3,26 [1,05, 10,12] Subtotal (95% CI) 518 516 100,0% 1,39 [0,90, 2,14] Total events 206 209 Heterogeneity: Tau²=0,15; Chi²=17,11, df=6 (P=0,009); I² = 65% Test for overall effect: Z=1,49 (P=0,14) 0.01 0.1 1 10 100 Usual care Lifestyle intervention Fig.4 Fig.3: Live Birth-pooled analysis. Lifestyle interventions Usual care Risk ratio Risk ratio Study Events Total Events Total Weight M-H, Random, 95% CI M-H, Random, 95% CI Becker et al. (28), 20151 3 14 0 12 1.3% 6.07 [0.34, 106.85] Einarsson et al. (29), 2017 66 152 56 153 27.8% 1.19 [0.90, 1.56] Espinos et al. (30), 2017 13 21 6 20 12.8% 2.06 [0.98, 4.36] Moran et al. (31), 2016 12 18 8 20 15.7% 1.67 [0.89, 3.12] Mutsaerts et al. (32), 2016 175 280 186 284 32.6% 0.95 [0.84, 1.08] Palomba et al. (33), 2010 1 32 0 32 1.1% 3.00 [0.13, 71.00] Rothberg et al. (34), 2016 3 6 0 5 1.5% 6.00 [0.38, 94.35] Sim et al. (16), 2014 13 27 3 22 7.2% 3.53 [1.15, 10.84] Total (95% CI) 550 548 100.0% 1.43 [1.02, 2.01] Total events 286 259 Heterogeneity: Tau²=0.09; Chi²=17.40, df=7 (P=0.01); I²=60% 0.005 0.1 1 10 200 Test for overall effect: Z=2.07 (P=0.04) Usual care Lifestyle interventions 1; Three women in the intervention group experienced 3 spontaneous pregnancies. Fig.5 Fig.4: Pregnancy rate-pooled analysis. Lifestyle interventions Usual care Risk ratio Risk ratio Study or subgroup Events Total Events Total Weight M-H, Fixed, 95% CI M-H, Fixed, 95% CI Becker et al. (28), 2015 0 3 0 0 Not estimable Einarsson et al. (29), 2017 8 66 5 56 14,2% 1,36 [0,47, 3,91] Espinos et al. (30), 2017 1 14 1 7 3,5% 0,50 [0,04, 6,86] Moran et al. (31), 2016 5 12 3 8 9,4% 1,11 [0,36, 3,40] Mutsaerts et al. (32), 2016 42 175 27 186 68,6% 1,65 [1,07, 2,56] Sim et al. (16), 2014 5 13 1 3 4,3% 1,15 [0,20, 6,59] Total (95% CI) 283 260 100,0% 1,50 [1,04, 2,16] Total events 61 37 Heterogeneity: Chi²=1,26, df=4 (P=0,87); I²=0% 0.005 0.1 1 10 200 Test for overall effect: Z=2,16 (P=0,03) Usual care Lifestyle intervention Fig.5: Miscarriage-pooled analysis. Notably, the results from six studies with a total number miscarriage in comparison to women receiving usual of 543 participants (16, 28-32) showed a statistically care (CI: 95% 0.6 to 9 more). We rated pregnancy rates significant increase in the risk of miscarriage in women and miscarriage as moderate quality due to limitations allocated to lifestyle interventions (RR: 1.50, CI: 95% in studies design. This increase in the risk of miscarriage 1.04 to 2.16; I2=0; Fig.5), resulting in seven women more disappeared in a subgroup analysis of interventions that out of 100 allocated to lifestyle interventions having a were exclusively based on a dietary restriction according Int J Fertil Steril, Vol 14, No 1, April-June 2020 6 Espinós et al. the pooled results from two trials (125 participants; RR: reports. This fact did not allow us to undertake reliable 1.36; 95% CI: 0.47 to 3.91; I2=0) (Materials S4, See analysis to explore the effects of the intervention in Supplementary Online Information at www.ijfs.ir). terms of different characteristics of women participating in other studies, the interventions assessed or the control After exploring possible sources of heterogeneity, conditions. Also, we limited inclusion to randomised we performed a sensitivity analysis excluding from trials that allowed us to obtain reliable effect estimates the pooled analyses one trial that could have an impact but omitted the results from a body of controlled on the consistency of effect estimates (32). The results observational studies (see excluded studies at Materials of these analyses resulted in a statistically significant S2, See Supplementary Online Information at www.ijfs. increase in live birth rates that favoured the intervention ir) that could bring light to the findings of our review. (6 trials, 470 participants; RR: 1.69; 95% CI: 1.05 to We also found some high heterogeneity related with the 2.70; I2=34%), while the impact on miscarriage switched different types of interventions for reducing weight and to a non-significant difference (5 trials, 182 participants; the discrepancies in women’s characteristics, such as RR: 1.16; 95% CI: 0.59 to 2.30; I2=0%) (Materials S5) age and the baseline values of women’s weight between (See Supplementary Online Information at www.ijfs.ir). studies. We rated the quality of evidence for primary outcomes as moderate-to-low due to the limitations in Cumulative pregnancy rate was not reported in the the included studies design and the imprecision in effect included studies. estimates. Effect of lifestyle interventions in secondary outcomes The increase in the miscarriage rate is an unexpected The participation in a lifestyle intervention did not finding since obesity has been related to a lower oocyte show differences, compared to the usual care, in the quality and endometrial receptivity increasing the risk rate of ongoing pregnancies (32) (317 participants; RR: of pregnancy loss. However, the study by Mutsaerts et 0.91, CI: 95% 0.79 to 1.05) or implantation rates (30) al. (32) in comparison with the other studies introduced (65 participants; RR: 1.32, CI: 95% 0.72 to 1.69). We clinical heterogeneity because women had lower rated these outcomes as low due to the imprecision in BMI and the control group received a higher number effect estimates. of infertility treatments; furthermore, the assessed intervention lasted for a longer period and the study Discussion presented attrition bias (22% of losses). For these reasons, we excluded Mutsaerts et al. (32) study in the We included eight trials, providing a total of 1175 sensitivity analysis. In consequence, results changed to infertile obese women randomised to receive a type of lifestyle interventions increased of live birth and there diet and/or exercise structured program versus usual care was not difference in the risk of miscarriage compared before undergoing an assisted reproduction program. with the control group. These results are more consistent In all included studies, experimental interventions with recent data that show an association of weight gain significantly lowered the women’s weight; however, ≥5% with a higher risk of pregnancy loss compared with there were some variations in the measure effects between maintaining a constant weight. The weight loss ≥5% did the studies. The main findings of our systematic review not associate with the increased risk of pregnancy loss suggests that lifestyle interventions may have little or no (35). Other systematic reviews have reported the effect impact on the live birth rates of obese infertile women of diet and/or exercise on obese fertile women. One who wish pregnancy. review (36) assessed the effect of low carbohydrate diet On the other hand, our results showed an increase on fertility hormones and pregnancy in overweight and in the risk of miscarriage rate in seven more pregnant obese women with a methodology that differed from our women out of 100 receiving the intervention instead review and with inconclusive results regarding the impact of the usual care. The sub-group analyses according to of intervention on the pregnancy rate. Another review the components of the intervention of interest (dietary also focused on assessed weight-loss interventions in restriction alone or in combination with physical overweight and obese women with broader inclusion activity) did not have major impact on our findings. No criteria (the review included non-randomized studies studies reported the cumulative pregnancy rate. and also assessed weight reduction drugs) (18). Pooled analysis from randomized trials showed similar results Our review surveyed rigorous methodological for the pregnancy rate and live birth, but did not show standards, and we set the methods used in our review in any increase in the rate of miscarriage, as shown by our a protocol prospectively registered. Most of the review findings. steps were conducted independently by pairs of reviews to ensure the accuracy of judgements and data. We made Lifestyle intervention programmes targeted to people an effort to identify all the relevant trials eligible for our with overweight or obesity usually result in poor inclusion criteria and asked missing data in published compliance rates and gender have been identified as one reports to the authors to avoid selective reporting bias. The of the critical predictors for adherence, which is lower review has also some limitations, and we obtained few in women (37). On the other hand, a great majority of missing data from trials and the data extracted from trial obese women facing an infertility treatment with interest 7 Int J Fertil Steril, Vol 14, No 1, April-June 2020 Lifestyle Intervention in Obese Infertile Women in a supervised medical weight-loss programme would adipose hormones in female reproduction. Hum Reprod Update. not be willing to delay the fertility treatment more 2006; 12(5): 585-601. 8. Maheshwari A, Stofberg L, Bhattacharya S. Effect of overweight than three months to attempt weight-loss (38). These and obesity on assisted reproductive technology--a systematic re- considerations are relevant in the light of the review view. Hum Reprod Update. 2007; 13(5): 433-444. findings when making a decision to initiate a programme 9. Metwally M, Ong KJ, Ledger WL, Li TC. 2008. Does high body such those described but facing low expectations from mass index increase the risk of miscarriage after spontaneous and it in terms of the fertility treatment success. In that assisted conception? A meta-analysis of the evidence. Fertil Steril. 2008; 90(3): 714-726. context, an individualized and shared decision should be 10. Legge A, Bouzayen R, Hamilton L. Young D. The impact of ma- made exploring patient motivation and other compliance ternal body mass index on in vitro fertilization outcomes. J Obstet predictors, such as age, baseline BMI, and mood (37). Gynaecol Can. 2014; 36(7): 613-619. 11. Setti AS, Braga DP, Figueira Rde C, Vingris L, Iaconelli A, Borges E Jr. Body mass index is negatively correlated with the response to Conclusion controlled ovarian stimulation but does not influence oocyte mor- phology in ICSI cycles. Eur J Obstet Gynecol Reprod Biol. 2012; Lifestyle interventions in obese infertile women 163(2): 175-179. based on dietary restrictions and physical activity 12. Robker RL. Evidence that obesity alters the quality of oocytes and probably lead to a slightly increase in the pregnancy rate embryos. Pathophysiology. 2008; 15(2): 115-121. compared with the usual care and make little difference 13. Mioni R, Chiarelli S, Xamin N, Zuliani L, Granzotto M, Mozzanega B, et al. Evidence for the presence of glucose transporter 4 in the in the improvement of live birth. Furthermore, our endometrium and its regulation in polycystic ovary syndrome pa- findings suggested a link between these interventions tients. J Clin Endocrinol Metab. 2004; 89(8): 4089-4096. and a slightly increase of the risk of miscarriage. More 14. Rittenberg V, Seshadri S, Sunkara SK, Sobaleva S, Oteng-Ntim E, research is needed in obese women undergoing infertility El-Toukhy T. Effect of body mass index on IVF treatment outcome: programs to further confirm or refute our findings. an updated systematic review and meta-analysis. Reprod Biomed Online. 2011; 23(4): 421-439. 15. Moran L, Tsagareli V, Norman R, Noakes M. Diet and IVF pilot study: short-term weight loss improves pregnancy rates in over- Acknowledgements weight/obese women undertaking IVF. Aust N Z J Obstet Gynae- Dr. Mª José Martinez Zapata is funded by a Miguel col. 2011; 51(5): 455-459. 16. Sim KA, Dezarnaulds GM, Denyer GS, Skilton MR, Caterson ID. Servet research contract from the Instituto de Salud Weight loss improves reproductive outcomes for obese women Carlos III and European Social Fund (investing in Your undergoing assisted reproductive technology: a randomised con- Future) (CP15/00116). There is no conflict of interest in trolled trial. Clin Obes. 2014; 14: 792-805. this study. 17. Sim KA, Partridge SR, Sainsbury A. Does weight loss in overweight or obese women improve fertility outcomes? A systematic review. Obes Rev. 2014; 15(10): 839-850. Authors’ Contributions 18. Best D, Avenell A, Bhattsacharya S. How effective are weight-loss interventions for improvin fertility in womenand men who are over- J.J.E.; Conceived the study. I.S.; Designed and weight or obese? A systematic review and meta-analysis of the evi- conducted the search. C.V., L.Z., M.J.M.-Z.; Screened dence. Hum Reprod Update. 2017; 23(6): 681-705. search results for eligibility and extracted data from 19. Higgins JPT, Green S. Cochrane Handbook for Systematic Re- views of Interventions Version 5.1.0. The Cochrane Collabora- relevant studies. I.S., C.V., L.Z., M.J.M.-Z.; Assessed tion, 2011. Available from www.handbook.cochrane.org. [updated the risk of bias. M.J.M.-Z., J.J.E., I.S., C.V.; Drafted March 2011]. the manuscript and the rest of authors contributed to 20. Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group. preparation of the manuscript. J.J.E., A.P.; Designed Preferred reporting items for systematic reviews and meta-analy- and conducted one of the included trials. All authors ses: the PRISMA statement. PLoS Med. 2009; 6(7): e1000097. 21. Higgins JPT, Altman DG, Sterne JAC. Assessing risk of bias in read and approved the final manuscript. included studies. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. The Cochrane Collaboration, 2011. Available from www.handbook. References cochrane.org. (updated March 2011). 1. Mitchell S, Shaw D. The worldwide epidemic of female obesity. 22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Best Pract Res Clin Obstet Gynaecol. 2015; 29(3): 289-299. Clin Trials. 1986; 7(3): 177-188. 2. Apovian CM. Obesity: definition, comorbidities, causes, and bur- 23. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring in- den. Am J Manag Care. 2016; 22(7 Suppl): S176-S185. consistency in meta-analyses. BMJ. 2003; 327(7414): 557-560. 3. Practice Committee of the American Society for Reproductive Med- 24. Sterne JAC, Egger M, Moher D. Addressing reporting biases. In: icine. Obesity and reproduction: a committee opinion. Fertil Steril. Higgins JPT, Green S, editors. Cochrane Handbook for System- 2015; 104(5): 1116-1126. atic Reviews of Interventions Version 5.1.0. Available from www. 4. Talmor A, Dunphy B. Female obesity and infertility. Best Pract Res cochrane-handbook.org: The Cochrane Collaboration, 2011. [up- Clin Obstet Gynaecol. 2015; 29(4): 498-506. dated March 2011]. 5. Grodstein F, Goldman MB, Cramer DW. Body mass index and ovu- 25. Review Manager (RevMan) [Computer program]. Version 5.3. Co- latory infertility. Epidemiology. 1994; 5(2): 247-250. penhagen: The Nordic Cochrane Centre, The Cochrane Collabora- 6. Kuchenbecker WK, Groen H, Zijlstra TM, Bolster JH, Slart RH, van tion, 2014. der Jagt EJ, et al. The subcutaneous abdominal fat and not the 26. Schünemann H, Brozek J, Guyatt G, Oxman A, editores. GRADE intraabdominal fat compartment is associated with anovulation in handbook for grading quality of evidence and strength of recom- women with obesity and infertility. J Clin Endocrinol Metab. 2010; mendations [Internet]. The GRADE Working Group, 2013. Avail- 95(5): 2107-2112. able in: http://gdt.guidelinedevelopment.org/app/handbook/hand- 7. Gosman GG, Katcher HI, Legro RS. Obesity and the role of gut and book.html. [updtaed 10/2013; acessed February 2018]. Int J Fertil Steril, Vol 14, No 1, April-June 2020 8 Espinós et al. 27. Guyatt GH, Oxman AD, Santesso N, Helfand M, Vist G, Kunz R, et 33. Palomba S, Falbo A, Giallauria F, Russo T, Rocca M, Tolino A, et al. GRADE guidelines: 12. Preparing summary of findings tables- al. Six weeks of structured exercise training and hypocaloric diet binary outcomes. J Clin Epidemiol. 2013; 66(2): 158-172. increases the probability of ovulation after clomiphene citrate in 28. Becker GF, Passos EP, Moulin CC. Short-term effects of a hypoca- overweight and obese patients with polycystic ovary syndrome: loric diet with low glycemic index and low glycemic load on body a randomized controlled trial. Hum Reprod. 2010; 25(11): 2783- adiposity, metabolic variables, ghrelin, leptin, and pregnancy rate 2791. in overweight and obese infertile women: a randomized controlled 34. Rothberg A, Lanham M, Randolph J, Fowler C, Miller M, Smith Y. trial. Am J Clin Nutr. 2015; 102(6): 1365-1372. The Feasibility of a brief, intensive weight loss intervention to im- 29. Einarsson S, Bergh C, Friberg B, Pinborg A, Klajnbard A, Karlström prove reproductive outcomes in obese, subfertile women: a pilot PO, et al. Weight reduction intervention for obese infertile women study. Fertil Steril. 2016; 106(5): 1212-1220. prior to IVF: a randomized controlled trial. Hum Reprod. 2017; 35. Radin RG, Mumford SL, Sjaarda LA, Silver RM, Wactawski-Wende 32(8): 1621-1630. J, Lynch AM, et al. Recent attempted and actual weight change 30. Espinós JJ, Polo A, Sánchez-Hernández J, Bordas R, Pares P, in relation to pregnancy loss: a prospective cohort study. BJOG. Martínez O, et al. Weight decrease improves live birth rates in 2018; 125(6): 676-684. obese women undergoing IVF: a pilot study. Reprod Biomed On- 36. McGrice M, Porter J. The effect of low carbohydrate diets on fertil- line. 2017; 35(4): 417-424. ity hormones and outcomes in overweight and obese women: a 31. Moran LJ, Tsagareli V, Noakes M, Norman R. Altered preconcep- systematic review nutrients. Nutrients. 2017; 9(3). pii: E204. tion fatty acid intake is associated with improved pregnancy rates 37. Burgess E, Hassmén P, Pumpa KL. Determinants of adherence in overweight and obese women undertaking in vitro fertilisation. to lifestyle intervention in adults with obesity: a systematic review. Nutrients. 2016; 8(1). pii: E10. Clin Obes. 2017; 7(3): 123-135. 32. Mutsaerts MA, van Oers AM, Groen H, Burggraaff JM, Kuchen- 38. Sacha CR, Page CM, Goldman RH, Ginsburg ES, Zera CA. Are becker WK, Perquin DA, et al. Randomized trial of a lifestyle pro- women with obesity and infertility willing to attempt weight loss pri- gram in obese infertile women. N Engl J Med. 2016; 374(20): 1942- or to fertility treatment? Obes Res Clin Pract. 2018; 12(1): 125-128. 1953. 9 Int J Fertil Steril, Vol 14, No 1, April-June 2020