Effect of female body mass index on intrauterine insemination outcomes: a systematic review and meta-analysis

Abstract

The prevalence of women with a raised body mass index (BMI) seeking assisted conception treatment is increasing. Findings of existing studies evaluating the effect of female BMI on intrauterine insemination (IUI) treatment outcomes remain inconsistent. This systematic review and meta-analysis evaluate the effect of female BMI on IUI treatment outcomes. Two authors independently conducted data extraction and assessed study quality. Risk ratios (RR) and 95% confidence intervals were calculated using the Mantel-Haenszel approach for dichotomous outcomes. 11 studies involving 23,145 IUI treatment events, comprising 21,211 cycles from 8 studies, and 1,934 participants in three studies, met the inclusion criteria for the meta-analysis. Two cohorts of women undergoing IUI treatment were compared - women with normal BMI < 25 kg/m² were compared with a second cohort of women with a BMI category ≥ 25 kg/m². There was no statistically significant difference in live birth rate (LBR) (RR 1.06, 95% CI 0.86–1.307); clinical pregnancy rate (CPR) (RR 0.94, 95% CI 0.78–1.13); miscarriage (RR 0.92, 95% CI 0.31–2.74) or ectopic pregnancy rate (RR 2.20, 95% CI 0.78–6.23). Our meta-analysis showed that a raised female BMI did not affect IUI treatment outcomes. Nevertheless, weight loss counselling should be offered to women with a raised BMI undergoing IUI, to reduce the associated obstetric morbidity.

TWEETABLE ABSTRACT

A meta-analysis of 11 studies found that having a raised female BMI did not change IUI treatment outcomes.

Keywords:

• BMI
• iUI
• clinical pregnancy
• livebirth
• systematic review
• meta-analysis

Introduction

Being overweight is defined as having a Body Mass Index (BMI) of 25 kg/m² or more and is a global epidemic with over 1.8 billion people overweight worldwide (World Health Organization, 2021). Furthermore, it is estimated that up to a half of all women in Europe within the reproductive age group are overweight (Gallus et al., 2015; Koning et al., 2010).

Being overweight is associated with major health consequences including risks of developing diabetes or hypertensive disorders but can also be a significant risk factor for female infertility. Reasons for this include hypothalamic-pituitary-ovarian dysregulation, menstrual irregularity, and anovulation (Broughton & Moley, 2017; Silvestris et al., 2018). Furthermore, being overweight significantly increases the risk of pregnancy complications including miscarriage, stillbirth and maternal disease such as diabetes and pre-eclampsia (Catalano & Shankar, 2017; Lashen et al., 2004). Assisted conception treatment in these women is also challenging with the potential deleterious effect of a raised BMI on ART outcomes. Sermondade et al.’s recent systematic review and meta-analysis suggested that female obesity is negatively associated with LBR following IVF (Sermondade et al., 2019). However, the evidence as to whether a raised BMI affects treatment outcomes following IUI appears conflicting. Some studies suggest there is a negative effect on outcomes as well as an increase in adverse events such as ectopic pregnancy in women with a raised BMI (Na et al., 2018; Yavuz et al., 2013). However, other studies suggest there is no effect of a raised female BMI on outcomes following IUI (Isa et al., 2014; L. T. Wang et al., 2020). One study has even reported an increase in success rates in women with a raised BMI compared to normal BMI (J. X. Wang et al., 2004).

In the year 2016, data from the European Society for Human Reproduction and Embryology, European IVF- consortium monitoring showed that 213,415 IUI cycles took place within Europe (European IVF-monitoring Consortium (EIM)‡ for the European Society of Human Reproduction and Embryology (ESHRE), Wyns et al., 2020). Given how prevalent a raised BMI is likely to be within this population, it is important to counsel women on any effect their BMI may have on IUI treatment outcomes.

This systematic review and meta-analysis evaluated whether raised female BMI affected IUI treatment outcomes.

Methods

The conduct and reporting of this review was guided by PRISMA guidelines and prospectively registered (PROSPERO: CRD42021239666). The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist was used while writing this review (Moher et al., 2009).

This systematic review aims to compare two cohorts – a raised BMI ⩾25 kg/m² cohort with a normal BMI < 25 kg/m² cohort on treatment outcomes following IUI. Primary outcome was Live Birth Rate (LBR) and secondary treatment outcomes were Clinical Pregnancy Rate (CPR), miscarriage rate and ectopic pregnancy rate. Subgroups analyses were also performed comparing normal BMI (BMI < 25 kg/m²) with overweight (BMI ⩾25 kg/m² – BMI < 30 kg/m²) and obese (BMI ⩾30 kg/m) on IUI outcomes. Data was only be combined for analysis if they fell within these BMI categorisations.

Literature search strategy and eligibility criteria

All study designs investigating the association of BMI and IUI were included. The pre-established inclusion criteria included all IUI cycles for both female or male factor infertility with an exposure to raised BMI. Observational cohort studies assessing solely other factors such as perinatal outcomes or other parameters such as waist to hip ratio were excluded. The search strategy, selection criteria, data extraction, quality assessment and statistical analyses described below were defined a priori.

Ovid Medline, Ovid Embase and the Cochrane Library were searched for relevant literature. The search strategy was limited to 1990 and 16^th May 2021 for all electronic databases, corresponding with the widespread use of IUI as an assisted conception treatment since this time (Tucker et al., 1990). Articles were also restricted to being published in English language. Further efforts were made to identify all available studies, including searching grey literature and conference abstracts from ESHRE, and the American Society for Reproductive Medicine. The search strategy for electronic databases used the following combined search terms: (‘Overweight’ [MeSH Terms] OR ‘Obesity’ [MeSH Terms] OR ‘Body Weight’ [MeSH Terms] OR ‘BMI’[Title/Abstract] OR ‘body mass index’[Title/Abstract] AND ‘Insemination, Artificial’[MeSH Terms] OR ‘IUI’[Title/Abstract] OR ‘Intrauterine Insemination’[Title/Abstract]).

Study selection and data extraction

Two reviewers (VS and HK) independently performed a screening of titles and abstracts of all articles, clinical studies and abstracts of congresses to exclude citations deemed irrelevant by both observers. Based on the pre-established inclusion criteria, full texts of potentially relevant articles were retrieved and assessed for inclusion by two reviewers (VS and HK). Any disagreement or uncertainty was resolved by discussion with a third reviewer (SKS). The methodological quality of the selected studies was assessed using the Newcastle-Ottawa Quality Assessment Scale for cohort studies (Stang, 2010). Data were extracted from included articles by two independent reviewers (VS and HK) using a data extraction template developed for the present review. All qualifying articles with quantitative data for either LBR or CPR were included in the meta-analysis. Authors were contacted for missing data and given a time of 3 weeks to respond; no authors replied. When only percentages were available, and when possible, the number of events was derived from total number of cycles/patients in each population group.

Data synthesis and meta-analysis

The following study details were extracted to characterize the included studies: study authors, publication year, country, study design, eligibility criteria, number of IUI cycles, stimulated vs natural cycle, cause of infertility and average age group of participants. For each cohort –the sample size, age, percentage and/or number of live births or clinical pregnancies were noted. When data were dispatched by subgroups in the article (e.g. natural vs stimulated IUI cycles), extracted data were pooled for overall meta-analysis. Dichotomous variables were expressed as relative risk (RR) and the precision of the estimates was evaluated by the 95% CI. The clinical relevance of all comparisons was assessed based on the precision of the estimates. Fixed and random effects models of analysis were used based on the heterogeneity of the studies.

The software Review Manager 5.3.5 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to combine and analyse the results for the meta-analysis. The meta-analysis was performed using a random effect model with the Mantel–Haenszel (M-H) method. Pooled effect sizes were deemed statistically significant at P < 0.05. In addition to computed estimates of between-study variance (Tau2), the statistic heterogeneity across the studies was calculated by chi-square statistic, and inconsistency was judged by the value of I² statistic. I² ≥ 50% indicates substantial heterogeneity (Higgins et al., 2003). For each study included in the meta-analysis, risk of bias was assessed by two independent reviewers (VS and HK) using ROBINS-1 tools (Sterne et al., 2016): confounding, selection of participants, intervention classification, intervention deviations, missing data, outcome measurement and selection of reported results. Each study was assigned a ‘low’, ‘high’ or ‘unclear’ risk of bias. A funnel plot was used to assess the presence of small-study effects suggestive of publication bias.

Sensitivity analysis

For the meta-analysis, a fixed-effect model was performed to compare the estimates of the intervention compared to a random-effect models in order to assess the robustness of findings. Sensitivity analysis was performed by excluding the outliers identified in the Funnel plot. To verify whether the conclusion would have been different if eligibility was restricted to studies with low risk of bias, another sensitivity analysis was performed by omitting all studies with at least one high risk of bias. Subgroup analyses were performed to separate the categories of raised BMI into overweight (BMI ⩾25 kg/m – BMI < 30 kg/m²) and obesity (BMI ⩾30 kg/m²).

Results

The search strategy identified a total of 1,463 articles, including duplicates and articles irrelevant to the primary research questions. After removing 674 duplicates, 892 records were reviewed, and 12 full-text articles were assessed for eligibility for qualitative analysis. Among this 11 articles seemed potentially appropriate to be included in the meta-analysis and were all cohort studies.

The size of the study population ranged from 301 to 2,040 participants in each study and a total of 23,145 events were analysed. This is comprised of 21,211 cycles from eight studies, and 1,934 participants in three studies.

Quality of studies

The methodological quality of the studies was assessed using the Newcastle – Ottawa score and the ROBINS-1 tool. Overall the quality assessment of these studies showed a low risk of bias. Amongst the nine applicable stars assessing the three main categories of selection, comparability and outcomes, the eligible studies received between eight and nine stars. Only 2 out of the 11 studies demonstrated a high risk of bias using the ROBINS-1 tool. Our meta-analysis included four prospective cohort studies and six retrospective cohort studies; one study was unclear as to whether this was prospective or retrospective. Six studies stated that BMI was measured before the treatment or insemination. Five studies did not mention when this was recorded; all but one of these studies were retrospective studies.

Results of Primary and secondary outcomes

There was no statistically significant difference detected for LBR for women undergoing IUI with a BMI ⩾25 kg/m² vs BMI < 25 kg/m², (RR 1.06, 95% CI 0.86–1.30 1 study with 3217 cycles) (Whynott et al., 2021). There was also no statistically significant difference detected for CPR for women undergoing IUI (RR 1.07, 95% CI 0.99–1.15); (RR 0.73, 95% CI 0.58–0.92); 11 studies with 23,145 events: divided into 21,211 cycles and 1,934 patients. Statistical heterogeneity was high at I² = 87%. It was not possible to conduct a meta-analysis of the CPR outcomes across all 11 studies, as 8 studies investigated the outcome per cycle, and 3 studies showed the outcome per patient.

There was no statistically significant difference detected for the miscarriage rate (RR 0.92, 95% CI 0.31–2.74, 3 studies with 5301 cycles) or ectopic pregnancy rate (RR 2.20, 95% CI 0.78–6.23, 1 study with 3217 cycles) for women undergoing IUI with a BMI ⩾25 kg/m² vs BMI < 25 kg/m².

Figure 1. Flow diagram showing the study selection process for systematic review on effect of body mass index (BMI) on IUI treatment outcome, in adherence to the PRISMA flow chart.

Sensitivity analyses

A funnel plot revealed four outlier studies (Koloszár et al., 2002; Souter et al., 2011; J. X. Wang et al., 2004; Yavuz et al., 2013) and a sensitivity analysis was performed. Excluding data from these studies did not change the outcome of CPR (RR 1.07, 95% CI 0.98–1.18 for CPR outcomes per cycle; RR 0.87, 95% CI 0.56–1.35 for CPR outcomes per patient). Furthermore, using odds ratios instead of relative risk did not change any of the outcomes. The use of the fixed effects model did not modify the result for any of the outcomes.

Subgroup analysis

Subgroup analyses were performed according to BMI categories comparing overweight vs normal BMI and obese vs normal BMI. However data was only available for CPR. Only 3 out of the 11 studies had data that could be pooled for this subgroup analysis due to heterogeneity in the BMI catergorisation; 2 studies were analysed due to the discrepancy in one study investigating the outcome per patient, as opposed to per cycle (Isa et al., 2014). There was no statistically significant difference detected for CPR for overweight vs normal BMI in women undergoing IUI; 2 studies with 2224 cycles (RR 0.87, 95% CI 0.52–1.44). There was also no statistically significant difference identified for CPR for obese vs normal BMI in women undergoing IUI; 3 studies with 2182 cycles (RR 0.86, 95% CI 0.48–1.53).

Figure 2. Random effect meta-analysis of female BMI ≤ 25 kg/m² vs BMI < 25 kg/m² on clinical pregnancy rates following IUI treatment, separated to 1)outcomes per cycle; 2) outcomes per patient.

Figure 3. Random effect meta-analysis of female BMI ⩾25 kg/m² vs BMI < 25 kg/m² on miscarriage rates following IUI treatment.

Discussion

Main findings

This meta-analysis based on 23,145 cycles and women from 11 studies showed no significant difference in either LBR or CPR in women with a raised BMI undergoing IUI treatment compared to women with a normal BMI. Furthermore, women who had a raised BMI were no more likely to have a miscarriage or ectopic pregnancy following IUI. Only a single study reported on livebirth rate (Whynott et al., 2021). Subgroup analysis for overweight vs normal BMI and obese vs normal BMI did not show any difference on CPR. However, this review has not been designed to detect a dose effect relationship of BMI on IUI outcomes.

Strengths and limitations

This review has some limitations. Firstly the paucity of studies reporting on the primary outcome of livebirth rate is a limitation of this review, especially given that LBR has been defined as a core outcome for infertility research (Duffy et al., 2020). Secondly, the multitude of potential confounding factors such as age (Isa et al., 2014), smoking (Huyghe et al., 2017), ethnicity (Craig et al., 2018), indication for IUI (Starosta et al., 2020) or insemination technique (Smith et al., 2002) could not be adjusted for in this study. Thirdly, the WHO standardized classification for BMI was chosen, studies which did not fit this categorization therefore had to be excluded. Lastly although studies were analysed for methodological quality, there was still significant heterogeineity in terms of study populations and reported outcomes.

Figure 4. Random effect meta-analysis of female (BMI ≤ 25 kg/m – BMI < 30 kg/m²) with normal BMI (<25 kg/m²) and obese (BMI ≥ 30 kg/m²) with normal BMI on clinical pregnancy rates following IUI treatment.

Strengths of this meta-analysis included that this review had the largest sample size published to date as well as strategies to minimise risk of bias in the review process including two researchers independently conducting the search, data extraction and undertaking the methodological quality assessment. However, it is conceivable that some studies may not have been identified during the search because of language restrictions. We attempted to contact all authors for clarification or for unpublished data, only one responded but did not subsequently release the unpublished data (Whynott et al., 2021).

Interpretation

Possible explanations for these findings include some suggestion that obesity may enhance endometrial proliferation and subsequently therefore implantation (Wolff et al., 2013). A more physiologically plausable explanation is that anovulation, which is widely prevalent within the overweight and obese population, may be overcome through ovulation induction agents or superovulation (McKnight et al., 2011; Starosta et al., 2020). Finally as the number of failed cycles of IUI prior to moving to IVF is likely to vary significantly between studies, an association between BMI and IUI outcomes may not clearly be seen. However, it is unclear why the findings for women with a raised BMI differ for IUI compared to IVF and further research is needed in this area (Sermondade et al., 2019). Furthermore, it is also conceivable that study findings will be prone to type II errors with small effect sizes not being detected.

Conclusion

In conclusion this current review did not find an association for women with a raised body mass index and IUI treatment outcomes. The same conclusions were drawn for the subgroup analysis comparing overweight and obese women with IUI treatment outcomes. Further research is needed to understand these findings, with the need for studies to categorise body mass index according to WHO standards and to report outcomes as per the recommended minmum core outcome set reporting for infertility research (Duffy et al., 2020). Nevertheless weight loss counselling should be offered to overweight and obese patients to reduce the associated obstetric morbidity.

Ethical approval

None required

Authors’ contribution

SKS conceived the idea. VS and HK performed data collection, tabulated data and analysis. VS and MSK performed statistical analysis. VS, HK, FB, YB, IS, MSK and SKS contributed to writing the manuscript and approved of the final version.

Acknowledgements

The authors would like to express their gratitude to Leorita Henry, Clinical Support Librarian at King’s College Hospital for her help in conducting the search.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

None received