Polycystic ovary syndrome (PCOS) is a common cause of female infertility, with a global prevalence ranging from 4% to 21% (1). In China, the prevalence increased from 5.6% in 2013 to 7.8% in 2022 (2, 3). PCOS is more prevalent among women of reproductive age. Patients with PCOS constitute a substantial proportion of those undergoing assisted reproductive technology (ART), and their endocrine and metabolic abnormalities may influence ART outcomes. PCOS is a metabolic disorder characterized by endocrine abnormalities. Compared to age-matched individuals without PCOS, those with PCOS are more likely to be overweight or obese (4). Previous studies have shown that, although the clinical pregnancy rate in obese women undergoing in vitro fertilization (IVF) is similar to that in women of normal weight, women with both PCOS and obesity have a relatively higher miscarriage rate (5, 6), which contributes to a reduced live birth rate (7). In recent years, increasing attention has been paid to the impact of body mass index (BMI) on assisted reproduction outcomes. However, BMI classification is inconsistent across studies, and most investigations examining the impact of BMI on pregnancy outcomes in patients with PCOS are based on the World Health Organization (WHO) criteria, which are primarily derived from European populations. Given that the standard BMI cut-off may underestimate obesity-related risks in Asian populations, the WHO redefined the classification for the Asia–Pacific region in 2000: overweight as BMI ≥ 23 kg/m² and obesity as BMI ≥ 25 kg/m² (8). Previous studies have also shown that the prevalence of metabolic abnormalities increases significantly in individuals with a BMI ≥ 23 kg/m², supporting this threshold as a valid cut-off for identifying those at high risk of metabolic complications, including Chinese women with PCOS (9, 10). However, the Chinese Obesity Working Group classifies BMI as follows: underweight, < 18.5 kg/m²; normal, 18.5 ≤ BMI < 24 kg/m²; overweight, 24 ≤ BMI < 28 kg/m²; and obesity, ≥ 28 kg/m² (11). Liang H et al. also demonstrated that this stratification system is appropriate for the Chinese PCOS population (12). Given the potential for BMI abnormalities in many patients with PCOS, further investigation is needed to establish a more accurate BMI cut-off point for optimizing assisted reproduction outcomes in different populations.

Women with PCOS have high rates of obesity, and most exhibit abdominal obesity, which is more likely to result in metabolic abnormalities, particularly insulin resistance (13). These metabolic disturbances affect not only overweight or obese PCOS patients but also those with normal weight (14). An WHO survey reported that Asians tend to have lower BMI but higher body fat percentages, along with increased risks of diabetes and cardiovascular disease (15). A study conducted by the Obstetrics and Gynecology Hospital of Fudan University in China also demonstrated that some non-obese women with PCOS exhibited high body fat and low muscle mass. Although their BMI was within or below the normal range, they still presented with insulin resistance (16). Previous studies have emphasized the adverse effects of obesity on IVF outcomes, including reduced live birth rates and increased risk of miscarriage (17). However, there remains a significant lack of research on assisted reproductive outcomes specifically in non-obese patients with PCOS.

Generally, if patients with PCOS present with excessively high BMI values, most ART centers recommend BMI adjustment before initiating treatment, which may lead to delays in conception and increased patient anxiety. Therefore, identifying an optimal BMI cut-off is essential for guiding clinical decision-making. In this study, we compared the pregnancy outcomes of PCOS patients with varying BMI levels following IVF–ET, including clinical pregnancy rate, early miscarriage rate, ongoing pregnancy rate, and live birth rate, in order to determine the most appropriate BMI cut-off—indicating the target BMI value during the preconception management phase.

In this retrospective cohort study, 649 women with PCOS underwent fresh embryo transfer following IVF/ICSI. The live birth rate was lowest in the low BMI group, and this difference was statistically significant compared with the normal BMI group. Although overweight and obese patients with PCOS did not show statistically significant differences from those with normal BMI, a trend toward lower live birth rates was observed.

Nonlinear curve analysis identified an inflection point at approximately 22 kg/m². When BMI was below this threshold, live birth rate increased progressively with increasing BMI. In contrast, when BMI exceeded 22 kg/m², the live birth rate showed a declining trend. Additionally, a significant interaction between age and BMI was observed, with the adverse effect of low BMI on live birth rate more pronounced in younger patients. These findings suggest that increasing BMI may improve pregnancy outcomes in individuals with low BMI. Conversely, in patients aged > 32 years who were overweight or obese, live birth rates were lower.

The most unexpected finding in our study was that the low BMI group exhibited both the highest miscarriage and lowest live birth rate, which contrasts with prior studies (6, 10, 12, 17). A prospective study involving 3,604 women also reported that underweight, overweight, and obese individuals had higher rates of subfertility compared with normal-weight women (20). BMI may influence IVF outcomes through several mechanisms. While healthy lifestyle modifications and weight loss are established first-line treatments for overweight or obese patients with infertility, low BMI has received comparatively less clinical attention. Notably, Asian individuals tend to have higher body fat percentages than Europeans at equivalent BMI values (21), which may lead to underestimation of the risk for metabolic complications such as diabetes and cardiovascular disease (15). A proportion of non-obese patients with PCOS exhibit high body fat, low muscle mass, abdominal obesity, and insulin resistance, despite having a BMI within the normal or even low range (16, 22). Two clinical studies have reported that lifestyle interventions—particularly those aimed at increasing muscle mass—can significantly improve insulin resistance and enhance rates of spontaneous ovulation. Several other studies have also suggested that increasing skeletal muscle mass can substantially improve insulin sensitivity (23–25). Additionally, a large retrospective cohort study conducted at Xiamen University, which included 13,745 intrauterine insemination cycles, found that women with low body weight had lower cumulative pregnancy and live birth rates compared with normal-weight women. The study concluded that low body weight may impair endometrial receptivity (26). Souter et al. also reported a positive correlation between BMI and endometrial thickness, suggesting that lower BMI may be a detrimental factor for endometrial receptivity (27). Furthermore, an animal study demonstrated that decreased oocyte quality in a lean mouse model of PCOS was associated with impaired mitochondrial ultrastructure and function (28), which may offer insights for future mechanistic investigations. In summary, considering the BMI inflection point identified in our study, increasing preconception BMI—particularly through muscle-building interventions—may improve pregnancy outcomes in PCOS patients with BMI < 22 kg/m², especially in younger individuals.

According to the Chinese Obesity Working Group classification system (11), patients with PCOS were categorized into underweight, normal weight, overweight, and obese groups. Although live birth rates in the overweight and obese groups were not significantly different from those in the normal weight group, a decreasing trend was observed, consistent with previous research findings (5–7, 12, 17). This supports the concept that obese patients with PCOS tend to have lower live birth rates compared with those of normal weight. Previous studies have suggested that individuals with obesity and PCOS are more likely to exhibit insulin resistance, which is an independent risk factor for early miscarriage (10, 29). Moreover, obesity is widely recognized for its adverse effects on both the quantity and quality of oocytes, as well as for reducing endometrial receptivity—ultimately impairing embryo implantation and lowering clinical pregnancy rates (30). Although our study did not demonstrate a statistically significant reduction in live birth rates among individuals with high BMI, the presence of obesity-related metabolic disturbances still warrants BMI optimization prior to conception. This may help reduce the risk of maternal and neonatal complications, as well as adverse effects on offspring (31, 32).

This study has several strengths. First, it exclusively included patients with PCOS who underwent fresh embryo transfers, thereby reducing the influence of certain confounding variables. Second, the association between BMI and live birth rate was further stratified by age, allowing for subgroup analysis. However, several limitations should be acknowledged. First, the retrospective study design limits causal inference, even though the temporal sequence between BMI and pregnancy outcomes was established. Second, the sample size was relatively small, with only 6.8% of participants having a BMI < 18.5 kg/m² and 9.5% having a BMI ≥ 28 kg/m², which may have influenced the robustness of subgroup analyses. Nevertheless, the median BMI was 22.5 kg/m², and 260 patients (40%) had a BMI < 22 kg/m², which provided adequate power for evaluating associations in this subgroup. Third, body composition analysis was not performed for all participants, resulting in a lack of data on fat percentage and muscle mass. Moreover, BMI does not distinguish between subcutaneous and visceral adipose tissue. Future studies should incorporate direct assessments of body composition. Fourth, this study only included fresh embryo transfer cycles. A parallel investigation that includes frozen embryo transfer cycles is currently underway, which will improve the generalizability of the findings.

This study was the first to propose a lower threshold for BMI management (BMI = 22 kg/m²) in Chinese patients with PCOS, highlighting the need for clinicians to consider weight-related risks at both ends of the BMI spectrum. Specifically: (1) Young PCOS patients with a BMI < 22 kg/m² should undergo nutritional assessment and receive individualized weight gain recommendations; (2) For individuals with obesity aged over 32 years, BMI should be optimized prior to conception. In summary, our study demonstrated for the first time that low BMI in PCOS patients is associated with adverse pregnancy outcomes. These findings underscore the importance of recognizing low BMI as a risk factor in reproductive planning and offer a new clinical perspective for individualized weight management and assisted reproductive strategies in patients with PCOS.