The correlation between raised body mass index and assisted reproductive treatment outcomes: a systematic review and meta-analysis of the evidence

Background Public funding for fertility services within the United Kingdom is limited, and therefore, strict guidance exists regarding who can be offered treatment under the National Health Service (NHS). Body mass index (BMI) is a universal criteria adopted by both the public and private sector. This study addresses an important aspect of the impact of a raised BMI on fertility treatment outcomes. We standardise the analysis of the data by only including studies incorporating the WHO BMI criteria; the current reference point for clinicians and clinical commissioning groups in ascertaining which group of patients should receive treatment. This study is an update of the previous systematic review performed in 2010, with the inclusion of a larger number of cycles from central databases such as the Society for Assisted Reproductive Technology (SART). Methods An electronic literature search was conducted through the Cochrane, Medline and Embase libraries. Data extraction for each outcome measure was pooled and expressed as an odds ratio with 95% confidence intervals. Where clinical heterogeneity was evident, the random effects model was used to calculate the risk ratio and a fixed effects model was used for the remaining studies. A p value < 0.05 was considered statistically significant. Results A total of 49 studies have been identified and included in this systematic review. Overweight and obese (BMI ≥ 25 kg/m2) women have a statistically significant lower live birth rate (OR 0.81, 95% CI 0.74–0.89, p < 0.00001) following Assisted Reproductive Technology (ART) when comparisons are drawn to women with a normal BMI. An increase is also demonstrated in the number of miscarriages experienced by women with a BMI ≥ 30 kg/m2 (OR 1.52, 95% CI 1.28–1.81, p < 0.00001). Conclusion Although this review concludes that a clear impact of BMI on ART outcomes is demonstrated, there remains questions as to the pathophysiology underlying these differences. This review supports the government’s stringent criteria regarding BMI categories under which NHS funding is made available for ART, through a clear description of poor reproductive outcomes in women with a BMI ≥ 30 kg/m2.


Plain English summary
This study highlights the impact of an overweight or obese female partner on fertility treatment outcomes, in particular focusing on IVF. Women who are overweight or obese have been shown to be less likely to have a life birth outcome from an IVF cycle. They are also more likely to suffer from early miscarriages whilst undergoing fertility treatments.

Main manuscript
The correlation between raised body mass index and assisted reproductive treatment outcomes: A systematic review and meta-analysis of the evidence.

Background
Obesity is a major challenge for today's clinicians. In 2016, the World Health Organisation (WHO) [1] stated that a staggering 39% of adults aged > 18 years fell into the overweight category, of which 40% were accounted for by women. Furthermore, 13% of the adult population were documented to be obese, with women accounting for 15% (WHO Global Health Observatory Data 2016). A raised body mass index (BMI) has been linked to a number of medical comorbidities, as well as being implicated in having a detrimental impact on the reproductive capacity of women in particular. Women who fall into high BMI categories can present with hypothalamicpituitary ovarian dysfunction and thus, low fecundity rates. In 2011, Rittenberg et al., [2] concluded that women with a BMI ≥ 25.0 kg/m 2 had a lower live birth rate through assisted reproductive treatments (ART) compared with women of a normal BMI. This has been further supported by multiple large studies evaluating the impact of BMI on ART outcomes.
The WHO classification of BMI is widely referred to, and provides standardisation for comparison of research outcomes. A documented BMI of 18.5-24.9 kg/m 2 is considered normal and healthy and the preferred range. A BMI of 25-29.9 kg/m 2 refers to overweight and a BMI ≥ 30 kg/m 2 is considered obese. The latter range is further subdivided into Class 1 (30.0-34.9 kg/m 2 ), Class 2 (35.0-39.9 kg/m 2 ) and Class 3 (≥ 40.0 kg/m 2 ). This paper, considers the current evidence regarding the impact of raised BMI on outcomes following ART treatment. A systematic review and meta-analysis of the available evidence will help provide or refute the current recommendations from the government regarding the allocation of resources for fertility treatment.

Search strategy
Literature searches were conducted through the Cochrane, Embase and Medline libraries . The medical subject headings (MeSH) were generated for two categories: 1. Body mass index (BMI, overweight, obesity); 2. in vitro fertilisation (IVF)/ intracytoplasmic sperm injection (ICSI) (embryo, embryo transfer, ART). All identified papers were reviewed by two authors (PRS and MM) independently. All discrepancies, regarding inclusion or exclusion of the data were discussed with a final decision mutually agreed upon.

Study inclusion and exclusion criteria
All relevant published studies reporting on the effects of BMI on IVF and ICSI pregnancy outcomes were included. Studies that reported donor cycles, conception by natural cycles, intrauterine insemination, waist hip ratio, and non-WHO classification of BMI were excluded. In addition, studies reporting on the effects of paternal body mass index on IVF/ ICSI outcomes were also excluded.

Outcome measures
The primary outcome measure assessed was live birth rate following an IVF/ ICSI cycle. Secondary outcome measures included: clinical pregnancy rate; and, miscarriage rate. The presence of a gestational sac on an ultrasound scan at least four weeks following on from an embryo transfer was used as confirmation for a clinical pregnancy. The clinical pregnancy rate was calculated per IVF/ ICSI cycle. For the purpose of this review, miscarriage was defined as pregnancy loss ≤ 20 weeks gestation. The miscarriage rate was calculated per clinical pregnancy.

Statistical analysis
Data extraction for each outcome measure was pooled and expressed as an odds ratio (OR) with a 95% confidence interval (CI). Clinical heterogeneity (I 2 ) [3] was considered significant when the I 2 value was < 50%. Where clinical heterogeneity was evident, the random effects model (DerSimonian and Laird, 1986) was used to calculate the risk ratio, and clinical heterogeneity was explored by comparing the variation in studies, such as, study design, study quality and interventions. Particular care was taken to further evaluate studies with similar first authors to avoid heterogeneity in the study population. For the remaining pooled data, the fixed effect model [4] was used to calculate the risk ratio. Statistical analysis was performed using the RevMan 5.3 software. A p value < 0.05 was considered statistically significant.

Results
The search strategy yielded 7458 electronic citations (Fig. 1). Of this, 2830 were removed secondary to duplications. Titles and abstracts were reviewed for the remaining 4628 publications. After screening of the titles and abstracts, 4508 publications were further excluded.
Full manuscripts were obtained for the remaining 120 articles. A further 16 articles were excluded as they did not use the WHO classification for BMI categories. A further 55 articles were excluded as per the inclusion exclusion criteria. The remaining 49 articles met all requirements and were included in this systematic review and meta-analysis (Table 1).

Primary outcome measure
Life birth rate per IVF/ ICSI cycle In women with a BMI ≥25 kg/m 2 versus BMI < 25 kg/ m 2 , a total of 14 studies were pooled and a statistically significant reduction in the live birth rate (OR 0.81, 95% CI 0.74-0.89, p < 0.00001; Fig. 2a) was seen. There was significant heterogeneity between the included studies (I 2 = 65%).
A total of 11 studies compared women with a normal BMI against those who were overweight (BMI 25-29.9 kg/m 2 ). An analysis of the pooled data showed a statistically significant reduction in the live birth rate in women with a BMI 25-29.9 (OR 0.92, 95% CI 0.86-0.97, p = 0.005; Fig. 2b). No significant heterogeneity (I 2 = 18%) was documented.
Data for women with a normal BMI versus BMI ≥ 30 kg/m 2 came from the pooling of 10 studies. The live birth rate for women with a BMI ≥30 kg/m 2 was statistically significantly lower than for women with a normal BMI (OR 0.81, 95% CI 0.79-0.82, p < 0.00001; Fig. 2c). No significant heterogeneity (I 2 = 0%) was detected in the data source.

Secondary outcome measures Clinical pregnancy rate
A total of 37 studies were pooled for BMI < 25 kg/m 2 versus BMI ≥25 kg/m 2 . A statistically significant reduction in the clinical pregnancy rate was demonstrated for women with a BMI ≥25 kg/m 2 (OR 0.82, 95% CI 0.77-0.88, p < 0.00001; Fig. 3a). However, there was significant heterogeneity (I 2 = 58%, p < 0.00001) between the studies analysed.
A statistically significant reduction in the clinical pregnancy rate was demonstrated for women with a BMI between 25 and 29.9 kg/m 2 when compared to women with a normal BMI (19 studies pooled, OR 0.89, 95% CI 0.84-0.94, p < 0.00001; Fig. 3b). No significant heterogeneity (I 2 = 31%) was seen between the studies.      Pooled analysis from 18 studies demonstrated a statistically significant reduction in the clinical pregnancy rate for women with a BMI ≥30 kg/m 2 when compared to women with a normal BMI (OR 0.80, 95% CI 0.74-0.87, p < 0.00001; Fig. 3c). There was no significant heterogeneity (I 2 = 32%) present between the studies.
The risk of miscarriage is further increased in women with a BMI ≥30 kg/m 2 when compared to women who fall into a normal BMI category (17 studies pooled, OR 1.52, 95% CI 1.28-1.81, p < 0.00001; Fig. 4c). No significant heterogeneity (I 2 = 46%) was demonstrated between the studies.

Discussion
Public funding for fertility services within the United Kingdom is limited, and therefore, strict guidance exists regarding who can be offered treatment under the National Health Service (NHS). Body mass index (BMI) is a universal criteria adopted by both the public and private sector. This study addresses an important aspect of the impact of a raised BMI on fertility treatment outcomes.
We standardise the analysis of the data by only including studies incorporating the WHO BMI criteria; the current reference point for clinicians and clinical commissioning groups in ascertaining which group of patients should receive treatment. This study is an update of the previous systematic review performed in 2010, with the inclusion of a larger number of cycles from central databases such as the Society for Assisted Reproductive Technology (SART).
This systematic review and meta-analysis has clearly highlighted the negative impact of a raised BMI on the outcomes following ART treatment, with documented lower success rates and higher rates of miscarriages as well as higher total dosage of gonadotrophin usage with Normal BMI versus BMI ≥30 kg/m 2 no effect on the duration of stimulation. The latter may have been balanced by higher dosages of treatment which can also have a cost implication. However, as most studies have included a BMI category of < 25 kg/ m 2 , which would also include underweight women with a BMI < 18 kg/m 2 , the detrimental effects of which have been addressed in a number of previous studies, a risk of bias cannot be confidently excluded. This has been addressed through the inclusion of studies allowing for a sub-group analysis of women with a normal BMI with overweight and obese women.
The presented data is able to demonstrate statistical significance with low clinical heterogeneity for a number of factors reflective of success through ART treatment. Despite this, caution is advised for interpretation of the presented information as only a few of the included studies controlled for confounding factors such as age, smoking and duration of infertility. In order to reduce further clinical heterogeneity, studies not incorporating the WHO classification for BMI and paternal BMI were excluded.
The included studies were considered relevant if they conformed to the WHO classification of BMI, despite this, a considerable amount of methodological and clinical heterogeneity existed. The level of statistical heterogeneity for the primary outcome measure live birth rate and secondary outcome measures clinical pregnancy rate and miscarriage rate were limited. However, despite a significant increase in total gonadotrophin dosage requirements with increasing BMI categories, the studies demonstrated significant statistical heterogeneity, limiting their value.
The presented data can act as an aid in the counselling of subjects secondary to a clear impact on ART outcomes being demonstrated across all BMI categories. The evidence supports the government's stringent allocation of funding when resources are significantly limited.
A raised BMI impacts reproductive health at the pre and post embryological stage of development, affecting oocyte quality and the endometrial environment [2].
A recent meta-analysis and systematic review by Best et al., [5] has demonstrated that weight loss can improve (c) Normal BMI versus BMI ≥30 kg/m 2 pregnancy rate and ovulatory status with a trend favouring spontaneous conception. However, these effects have not been seen through ART. Of note, miscarriage rates were unaltered with a change in weight.
Besides the reproductive health effects of a raised BMI, clinicians should also be aware of the increased rate of pregnancy complications such as pregnancy induced hypertension, pre-eclampsia and gestational diabetes in women with a raised BMI. Women are also at an increased risk of an emergency caesarean section with increasing BMI [6,7].
A holistic approach should be used when counselling patients seeking ART treatments using an open discussion method to inform patients of the effects of raised BMI on ART and obstetric care. This will allow couples to make an informed decision and to take ownership of their well-being.

Conclusion
This systematic review and meta-analysis further emphasises the negative impact of a raised BMI on ART outcomes. However, the underlying pathophysiology is beyond the scope of this systematic review and will need to be evaluated in future studies. The quality of this systematic review would be further improved if future study designs included the WHO classification of BMI and controlled for confounding variables.