The following selection criteria were adopted: (a) studies based on animal models were excluded. (b) lean was defined as individuals with BMI ≤ 25 kg/m2, and for Asian populations, BMI ≤ 23 kg/m², (c) overweight/obese individuals were those with a BMI of ≥25.0 kg/m², and for the Asian population, a BMI ≥23 kg/m², (d) studies comparing lean PCOS (LP) with overweight/obese PCOS.

Studies that did not meet the pre-set eligibility criteria were excluded during the preliminary screening phase of relevant studies. Duplicated studies were excluded using the EndNote reference library program. The selected articles were further reviewed to identify the existence of different versions of the articles and those with accessible full texts. To enhance consistency and accuracy in the extracted data, four authors were involved in data extraction, which included all outcomes of interest and baseline characteristics. The main outcome was differences in metabolic and hormonal indicators between the lean PCOS and obese PCOS individuals. Moreover, we investigated the development of systemic diseases and the prevalence of metabolic syndrome (MS) in lean versus obese PCOS (OP) patients. The secondary outcome of the study was the differences in critical clinical parameters, including low-density lipoprotein (LDL) cholesterol levels, triglycerides, systolic and diastolic blood pressures (BPs), and fasting blood sugar (FBS) levels between lean and OP patients. To enhance the integrity of the extracted data, any uncertainties arising during the data extraction process were addressed through collaborative discussions among the investigators involved in data extraction.

All statistical comparisons between lean PCOS and obese PCOS were performed using the Revman software (Version 5.4.1). Standard mean difference (SMD) and their respective 95% confidence intervals (CIs) were calculated for the two groups and the I² statistic was employed, with values less than 50% indicating mild heterogeneity. In cases where there was significant heterogeneity, sensitivity analyses were conducted using the leave-one-out method. A p value below 0.05 indicated significant difference between the groups.

Relevant studies were selected following the PRISMA 2020 guidelines. A total of 64,045 records were identified from the electronic databases (10,518 from PubMed; 20,885 from ISI Web of Science; 14,443 from Embase; 18,199 from Cochrane Library) and 23 via manual search. After excluding 49,255 duplicates, 14,813 records were further screened by reading the titles/abstracts, leading to the exclusion of 12,921 articles (7,975 as reviews/abstracts/editorials; 1,583 as animal studies; 3,363 involving irrelevant populations). Subsequently, the remaining 1,892 full-text articles were analyzed, among which 1,819 were excluded (837 as pooled/subgroup/post-hoc analyses of included trials; 532 without BMI in eligibility criteria; 450 lacking reported outcomes of interest). Finally, 73 clinical trials were included in the systematic review and meta-analysis ( Figure 1 ).

The 73 studies were from 18 countries grouped as follows: Turkey (n=14), China (n=12), India (n=11), and Italy (n=5). The sample sizes ranged from 22 to 458, with all studies stratifying participants into lean PCOS and obese PCOS groups. Lean PCOS counts per study ranged 10–352, obese PCOS 10–352. Regarding the age distribution of the participants (61 studies reported age) with those of lean PCOS group having the mean age of 20.25 (1.45)–34.4 (4.8) years; obese PCOS with a mean age of 20.43 (1.53)–34.7 (4.4) years (12 studies NA). Analysis of the BMI data (59 studies) revealed that the mean BMI of the lean PCOS group was 19.92 (3.51)–23.23 (1.65) kg/m², while that of obese PCOS was 26.36 (1.86)–36.8 (4.8) kg/m² (14 studies NA). Age and BMI were reported as mean (SD), mean (min-max), mean (SE), median (5-95% CI), mean (5-95% CI), or medians (interquartile ranges) ( Supplementary Table 2 ).

The methodological quality of the included studies was determined using the Newcastle-Ottawa Scale, covering Selection, Comparability, and Outcome domains; the results are shown in Supplementary Table 3 . Most studies exhibited a minimal risk of bias, indicating they were highly reliable.

Fasting blood sugar levels were lower in the lean PCOS compared with levels in the obese PCOS group (SMD −0.48, 95% CI −0.58 to –0.39, p <0.01) ( Figure 2 ). Considering the BP measurements, lean PCOS participants had slightly lower levels of diastolic (SMD −0.56, 95% CI −0.79 to −0.33, p < 0.01) and systolic (SMD −0.58, 95% CI −0.80 to −0.36, p < 0.0003) BP relative to individuals with obese PCOS ( Figure 3 ).

Analysis of the lipid profiles revealed that lean PCOS individuals presented with favorable outcomes compared with their obese PCOS patients. They had lower levels of LDL (SMD −0.49, 95% CI −0.60 to −0.38, p < 0.01) and triglycerides (SMD −0.72, 95% CI −0.85 to −0.59, p < 0.01). These differences in lipid levels indicate that lean PCOS may have a lower cardiovascular risk than obese PCOS ( Figure 4 ).

Comparative analysis demonstrated that endocrine hormone levels were significantly different between lean PCOS and obese PCOS patients.

Data shown in Figure 5 indicate that the level of LH was higher in lean PCOS patients than in obese PCOS patients (SMD 0.23, 95% CI 0.11 to 0.36, p < 0.01). The level of FSH in lean PCOS patients was also higher than that in obese ones (SMD 0.10, 95% CI 0.04 to 0.17, p = 0.02). Moreover, the LH/FSH ratio in lean PCOS patients exceeded that in obese PCOS patients (SMD 0.23, 95% CI 0.07 to 0.40, p < 0.01). Notably, lean PCOS exhibited higher levels of SHBG compared to obese PCOS(SMD 0.81, 95% CI 0.61 to 1.02, p < 0.01). The levels of DHEA-S were significantly higher in the lean PCOS than in the obese PCOS patients(SMD 0.19, 95% CI 0.03 to 0.35, p < 0.01) ( Figure 6 ).

A higher Ferriman-Gallwey (F-G) score was observed in obese PCOS individuals compared to lean PCOS patients (SMD -0.36, 95% CI -0.58 to -0.14, p = 0.87) ( Figure 7 ). The levels of F-T were significantly different between the two groups, being higher in obese PCOS patients than in lean PCOS patients (SMD -0.4, 95% CI -0.6 to -0.2, p < 0.01) ( Figure 8 ). Moreover, the homeostasis model assessment of insulin resistance (HOMA-IR) was higher in obese PCOS patients (SMD -0.88, 95% CI -1.03 to -0.72, p < 0.01), and their ovarian volume was also larger compared with that in the lean PCOS patients (SMD -0.17, 95% CI -0.32 to -0.01, p = 0.23) ( Figure 9 ).

The analysis showed that there was no significant difference in AMH levels and androstenedione between the two groups.

These findings of this study indicate that obese PCOS patients have higher BP, blood lipid, and blood glucose levels compared to their lean counterparts, which is consistent with earlier reports (10). Notably, insulin resistance is a key feature of PCOS, and obesity has been implicated in the pathogenesis of this condition (11). In obese individuals, the expansion of adipose tissue increases the secretion of various adipokines and inflammatory substances. These compounds can disrupt insulin signaling pathways, thereby decreasing insulin sensitivity (12). Therefore, the body produces more insulin to maintain normal blood glucose levels. The increased insulin production stimulates lipid synthesis and storage in the liver and adipose tissue, thereby elevating blood lipid levels (13).

Our meta-analysis shows that, compared with obese patients with Polycystic Ovary Syndrome (PCOS), lean PCOS patients often present with a favorable metabolic profile. However, studies also indicate that when compared to the general population, lean PCOS patients are at a higher risk of developing metabolic disorders. Even if their body weight is within the normal range, they may have underlying metabolic problems such as mild insulin resistance or abnormal lipid metabolism (14).

The differences in metabolic and endocrine characteristics between lean and obese PCOS patients may have important implications for clinical practice. For obese PCOS patients, weight management should be incorporated in the treatment plan. Lifestyle modifications, such as a balanced diet and regular exercise, can potentially alleviate insulin sensitivity, reduce androgen levels, and restore ovulatory function (29). Moreover, it has been shown that metformin is more effective as an ovulation stimulation agent when administered to non-obese women with PCOS (30). For lean PCOS patients, although their metabolic status may be relatively better, they still require monitoring to track their metabolic parameters. In some cases, appropriate treatments for ovulatory dysfunction and hyperandrogenism should be administered. The specific treatment approach should be customized to each patient’s symptoms, fertility needs, and overall health.

The results of this study are valuable in terms of several strengths. We utilized a comprehensive literature search, stringent inclusion and exclusion criteria to identify eligible studies. First, to the best of our knowledge, this is the first meta-analysis to evaluate the metabolic and endocrine characteristics between lean and obese patients with polycystic ovary syndrome (PCOS). Second, the present study comprehensively compared the metabolic and endocrine characteristics of lean and obese PCOS, providing important data to guide the clinical classification of patients. Moreover, it clarifies the differences in key indicators between the two types of patients, which is essential to the implementation of targeted treatment. However, the included studies exhibited some degree of heterogeneity, and differences in diagnostic criteria and detection methods, which may affect the stability of the results. In addition, there was selection bias since the samples were mainly selected from specialized outpatient clinics, implying lack of population representativeness.

In conclusion, this meta-analysis provides valuable information regarding the profile of metabolic and endocrine characteristics between lean and obese PCOS patients. Further research is advocated to investigate the underlying mechanisms and to develop more targeted treatment strategies for different subgroups of PCOS patients.