POLYCYSTIC OVARY SYNDROME (PCOS) is the most common endocrinopathy in women of reproductive age, with a prevalence of up to 10%. It is a heterogeneous syndrome with the characteristics of hirsutism, acne, anovulation, hyperandrogenemia, polycystic ovaries, and infertility [1]. It is associated with an adverse metabolic profile including insulin resistance. There is a clear association between obesity, the development of PCOS, and the severity of its phenotypic, biochemical and metabolic features [2]. About 50% of PCOS patients are obese [3]. It is currently assumed that adipose tissue is not only the main energy reservoir but also a pivotal endocrine organ secreting a variety of bioactive cytokines named adipokines, such as leptin, resis- Abstract. Studies have examined the association between retinol-binding protein 4 (RBP4) and polycystic ovary syndrome (PCOS). However, the results have been inconsistent. To investigate the association between RBP4 and PCOS, we performed a meta-analysis. The Cochrane Library, MEDLINE, the ISI Web of Science, and EMBASE were searched to identify all of the studies that examined the relationship between circulating RBP4 levels and PCOS. Standard mean difference (SMD) values and 95% confidence interval (CI) were estimated and pooled using meta-analysis methodology. A total of seven studies were involved in the meta-analysis, which included a total of 636 subjects (260 controls and 376 patients with PCOS). The RBP4 level was higher in PCOS patients than in non-PCOS patients (random effects MD (95% CI)=0.69, [0.20, 1.18], P=0.006). However, the RBP4 level was not higher in nonobese PCOS patients than in nonobese controls (random effects MD (95% CI)=0.38, [-0.21, 0.98], P=0.20). The effect size revealed that the RBP4 level was higher in overweight or obese PCOS patients than weight-matched controls (fixed effects MD (95% CI) =7.95, [5.96, 9.93], P<0.05). In the subgroup analysis by region, the RBP4 level was higher in PCOS patients in Asia than controls (random effects MD (95% CI)=0.85, [0.54, 1.15], P<0.05), but not in European PCOS patients compared with controls (random effects MD (95% CI)=0.34, [-1.12, 1.80], P=0.65). This subgroup analysis also showed that nonobese PCOS patients have higher RBP4 levels than controls in Asia. Our meta-analysis results indicated that RBP4 might be a useful tool for identifying PCOS women.