The study, led by Xiaoju Wan and colleagues including corresponding author Jun Tan at Jiangxi Maternal and Child Health Hospital in Nanchang, was published in the Journal of Ovarian Research. The team began with 11,524 eligible women divided into three cohorts: 2,881 with polycystic ovary syndrome, commonly abbreviated PCOS; 4,354 with diminished ovarian reserve, known as DOR; and 4,289 with tubal factor infertility, in which damaged or blocked fallopian tubes prevent natural conception. After restricting the analysis to women with complete data suitable for mediation analysis, the final sample comprised 2,507 women with PCOS, 2,425 with DOR, and 3,991 with tubal factor infertility. The outcome analyses drew on somewhat different subsets: 6,708 women for clinical pregnancy, 4,182 for pregnancy loss, and 6,523 for live birth.

The central methodological innovation of the study lies in its use of mediation analysis, a statistical framework designed to decompose an observed association into direct and indirect components. In this case, the researchers asked whether body mass index exerts part of its influence on oocyte yield, the number of eggs retrieved during an IVF cycle, through a specific intermediate variable: basal follicle-stimulating hormone, or FSH. FSH is the pituitary hormone that drives follicular growth in the ovaries, and its baseline level in the blood is a well-established marker of ovarian reserve. The logic of the mediation model is straightforward. If higher body mass index alters basal FSH levels, and if those altered FSH levels in turn change how many eggs are retrieved, then FSH functions as a mediator, carrying an indirect effect of body weight on ovarian response.

The results of this mediation analysis were nuanced. Body mass index was indeed associated with a small indirect effect on oocyte yield operating through FSH, with the proportion of the total effect mediated by FSH ranging from 5.31 to 8.21 percent across the cohorts. In other words, only a modest fraction of the relationship between body weight and egg retrieval numbers runs through the hormonal pathway the researchers measured. The direction of this indirect association was similar across all three diagnostic groups, which is precisely the kind of consistency the study was designed to detect. However, the magnitude of the effect was small and the confidence intervals were wide, particularly within the diminished ovarian reserve cohort, where fewer eggs are retrieved overall and statistical precision suffers accordingly. The authors are careful to note that the clinical relevance of this FSH-mediated pathway remains uncertain.

Where the study becomes considerably more consequential is in its findings on pregnancy loss. Obesity was significantly associated with a higher risk of miscarriage after IVF, with an odds ratio of 1.60 and a 95 percent confidence interval of 1.02 to 2.53, yielding a statistically significant result with a P value of 0.041. The researchers translated this into a clinically meaningful metric as well: a number needed to harm of 19, meaning that roughly one additional pregnancy loss could be expected for every 19 women with obesity compared with women of lower body weight. Strikingly, this elevated risk was consistent across the PCOS, diminished ovarian reserve, and tubal factor cohorts, suggesting that the miscarriage risk associated with obesity is not confined to any single infertility diagnosis but operates as a general phenomenon.

The picture for other reproductive outcomes was less definitive. Obesity showed a non-significant trend toward lower live birth rates, with an odds ratio of 0.77 and a P value of 0.078, but the authors emphasize that this analysis had only 5.3 percent statistical power, rendering it inconclusive. Power refers to a study’s ability to detect a true effect if one exists, and a power of 5.3 percent is extraordinarily low, meaning the analysis could almost certainly not have detected even a substantial difference in live birth rates. The researchers are explicit on this point: the live birth finding should be considered exploratory and hypothesis-generating, and no clinical recommendation regarding live birth can be made on the basis of this study. Meanwhile, no association was observed between obesity and clinical pregnancy, the earlier milestone at which a fetal heartbeat is typically confirmed.

These findings carry particular weight because of how they were structured. Rather than treating all IVF patients as a homogeneous group, the researchers stratified their analysis by infertility etiology, allowing them to test whether the effects of body mass index differ between women whose ovaries are polycystic, women whose ovaries are running low on eggs, and women whose reproductive tracts are mechanically obstructed. The consistency of the associations across these three biologically distinct groups is the study’s headline contribution. It suggests that whatever mechanisms link body weight to ovarian response and miscarriage, those mechanisms are not specific to the hormonal chaos of PCOS or the depleted follicular pool of DOR, but instead operate through pathways shared by all women.

Several biological explanations for such shared pathways are plausible, though the study itself was not designed to test them directly. Adipose tissue is not merely passive storage; it is an endocrine organ that secretes inflammatory cytokines, adipokines such as leptin, and enzymes that convert androgens to estrogens. Chronic low-grade inflammation associated with obesity has been implicated in impaired endometrial receptivity and compromised placental development, both of which could plausibly raise miscarriage risk independently of the cause of infertility. Similarly, the insulin resistance that frequently accompanies elevated body mass index could affect oocyte quality and early embryonic development across diagnostic categories. The modest FSH-mediated pathway identified in the study may represent one thread of this larger web, but the small proportion mediated indicates that other mechanisms, unmeasured in this analysis, likely account for most of the effect.

The study also illustrates the strengths and limits of retrospective cohort research. Because the data were collected for clinical rather than research purposes, and because the analysis was restricted to complete cases, the possibility of selection bias cannot be fully excluded; women excluded for missing data may have differed systematically from those included. The ethics committee of Jiangxi Maternal and Child Health Hospital approved the study under approval number SZYY-202604 and waived the need for informed consent given its retrospective nature, and the work was conducted in accordance with the Declaration of Helsinki. Funding came from the JXHC Key Laboratory of Fertility Preservation, and the authors declared no competing interests. As an observational study, it can establish associations but cannot prove causation, and unmeasured confounders could influence the results.

For patients and clinicians, the practical message is a measured one. Body mass index appears to matter for IVF in broadly similar ways regardless of infertility diagnosis, with the clearest and most actionable signal being the increased risk of pregnancy loss among women with obesity. The effect on egg retrieval, while real in a statistical sense, is small and partly channeled through basal FSH in a way that may not justify changes to stimulation protocols on its own. The live birth question remains open, awaiting adequately powered studies. What this research adds is a degree of generality: the conversation about weight and fertility treatment need not be segmented by diagnosis, because the evidence suggests the underlying biology is shared.

Subject of Research: The association of body mass index with IVF ovarian response and pregnancy loss across different infertility etiologies

Article Title: Consistent associations of BMI with ovarian response and pregnancy loss risk across PCOS, DOR, and tubal factor infertility: a complete-case mediation analysis of 8,923 IVF cycles

Article References: Wan, X., Yu, M., Wu, X., Huang, Z., & Tan, J. (2026). Consistent associations of BMI with ovarian response and pregnancy loss risk across PCOS, DOR, and tubal factor infertility: a complete-case mediation analysis of 8,923 IVF cycles. Journal of Ovarian Research. https://doi.org/10.1186/s13048-026-02299-0

Image Credits: AI Generated

DOI: 10.1186/s13048-026-02299-0

Keywords: body mass index, IVF, PCOS, diminished ovarian reserve, tubal factor infertility, pregnancy loss, oocyte yield, follicle-stimulating hormone, mediation analysis, obesity, reproductive medicine, miscarriage