Prevalence of Polycystic Ovarian Syndrome in India: A Systematic Review and Meta-Analysis

Stein-Leventhal syndrome, often known as polycystic ovarian syndrome (PCOS), is a syndrome that affects women's reproductive health. PCOS is one of the most common endocrine and metabolic disorders in women of reproductive age. The etiology of PCOS remains unknown mainly, and the estimation of PCOS burden in a specific geographical location will impact disease control strategies. Hence, this study estimated the pooled prevalence of PCOS in Indian women. Databases such as PubMed, CINHAL, Scopus, and Google Scholar were thoroughly searched. Only those published Indian studies that reported the prevalence of PCOS from 2010 to 2021 and had at least one of the following diagnostic PCOS criteria were included in the systematic review: the National Institutes of Health (NIH), Rotterdam's criteria, or/and Androgen Excess Society (AES). MetaXL version 5.3 software was used for data analysis. The risk of bias was assessed using modified Joanna Briggs Institute criteria for cross-sectional studies. Out of 17132 articles, 11 articles were selected for systematic review and meta-analysis. The pooled prevalence of PCOS was 11.33(7.69-15.59) using the random effect. The proportion of Hirsute using the Ferriman-Gallwey score was highly variable, ranging from 1.6% to 37.9% (n=6). The prevalence rate of PCOS is high among Indian women. The pooled prevalence of PCOS was close to 10% using Rotterdam's criteria and AES criteria, while it was 5.8% using NIH criteria. The study's overall finding emphasizes the need for more acceptable and uniform diagnostic criteria for screening PCOS. At the same time, policy-makers should consider giving more importance to PCOS in their effort to control non-communicable diseases.


Introduction And Background
In the mid-1900s, Stein and Leventhal (Chicago, IL, USA) investigated the mechanisms of female sterility. According to Stein and Leventhal, women with sterility, equated with infertility, had abundant body hair and disturbed menstrual cycles. Irving Freiler Stein Sr. wrote "The Stein-Leventhal Syndrome: A Curable Form of Sterility" in 1958, detailing his findings on Stein-Leventhal syndrome diagnosis and surgical therapy. Stein-Leventhal syndrome, often known as polycystic ovarian syndrome (PCOS), is a syndrome that affects women's reproductive health. Excess hair in the body, absence of menstrual cycle (amenorrhea), and infertility are all common symptoms of PCOS [1]. In the 21st century, reproductive health remains a top public health priority issue that needs a holistic approach to address it.
PCOS is one of the most commonly reported endocrine and metabolic disorders among women of reproductive age. It is a heterogeneous condition characterized by features of androgen excess and ovarian dysfunction symptoms in the absence of another diagnosis. Although the etiopathology of PCOS is not so well proven, accumulating evidence suggests that it is a multi-gene condition with substantial epigenetic and environmental impacts, including nutrition and lifestyle variables. Menstrual abnormalities and reproductive dysfunction are the most commonly reported signs of PCOS, leading to female infertility [2,3]. Cardiovascular disease, hypertension, lipid metabolic problems, and endometrial cancer are all two to six times more common in PCOS patients than in the general population [4]. PCOS is easy to diagnose and treat; it just takes judicious utilization of already available standardized diagnostic tests and the application of appropriate approaches to address hyperandrogenism, the consequences of ovarian dysfunction, and the metabolic abnormalities that arise with it [5].
In the last few years, several attempts have been made to standardize the diagnostic criteria for PCOS [6]. But still, the diagnostic criteria for PCOS are debatable. First, in 1990, the National Institutes of Health (NIH) established criteria for PCOS [7], followed by Rotterdam criteria in 2003 [8]. This criterion involves the presence of any two of the three conditions: (a) oligomenorrhea/anovulation, (b) clinical/biochemical hyperandrogenism, and (c) polycystic ovaries (each ovary containing ≥12 follicles measuring 2-9 mm). In 2006, AES criteria were given by the Androgen Excess Society (AES), featuring clinical/biochemical hyperandrogenism with either oligo/anovulation or polycystic ovaries [9].
As indicated by the NIH diagnostic criteria, the revealed predominance of PCOS went from 6% to 9% in the United States, the United Kingdom, Spain, Greece, Australia, Asia, and Mexico [10]. Related to variances in research populations, limitations because of types of recruitment and sampling, and an absence of standardized definitions for the phenotypes, there is substantial disparity in reported prevalence even when using the same diagnostic criteria. The impact of race and nationality on the clinical presentation of androgen excess [11], as well as the gradual improvement in the presence of antral follicles by ultrasonography [12], may potentially impact the differences in reported prevalence. The ambiguity surrounding PCOS findings must be addressed promptly to give doctors and their patients more diagnostic accuracy, minimizing incorrect classification and the possible psychological distress that misdiagnosis can be caused by it [13].
The prevalence of a disease in a particular region is always a necessary tool for any control measures. However, there are no full-fledged published data on PCOS prevalence and distribution patterns in India because of an absence of well-designed studies with a robust methodology. As a result, a systematic review that provides a suitable pooled prevalence is highly required. With this goal, the present study was planned to measure the pooled prevalence of PCOS among Indian women from 2010 to 2021. regarding the prevalence, or were not published as peer-reviewed original research publications, they were eliminated. Two blinded investigators (M.D.B. and V.R.) conducted the initial searching and screening of titles and abstracts. After a full-text review regarding the inclusion of the particular study, the third investigator (JG) was consulted for the final decision. The initial search from PubMed, CINHAL, Scopus, and Google Scholar yielded a total of 17,132 articles ( Figure 1). After the initial removal of duplicates, screening from abstracts and titles, only 30 relevant articles were undertaken for full-text review for eligibility. Furthermore, on the exclusion of 19 articles for various reasons (Figure 1), 11 articles were included in the quantitative synthesis.

Data extraction
Two authors (M.D.B. and V.R.) created a data table form for the data extraction process, which was pilottested to ensure author unanimity. Data extraction was done by all three Investigators (M.D.B., V.R., and J.G.) independently in pretested and piloted format in a Microsoft Excel sheet; regarding any disagreement on the extracted data, final consensus was made after discussion with the fourth investigator (R.R.). Data were extracted using Microsoft Excel sheet for the following variables: author, title, journal name, publication year, region, sampling frame, study setting, sample size, study population, risk-of-bias appraisal, and the criteria used to measure the prevalence of PCOS. The primary outcome was the pooled prevalence of PCOS.

Quality assessment of studies/risk of bias
The quality assessment of the included studies was done using the modified Joanna Briggs Institute (JBI) criteria by the University of Adelaide [16]. The bias risk was appraised by all three investigators by giving a response of "yes," "no," "unclear," and "not applicable." All three reviewers independently assessed the bias risk using the modified JBI criteria. In case of a mismatch of results, the common opinion of any two reviewers was the final decision. The evaluated articles were divided into three categories: high risk of bias (JBI score < 50%), moderate risk of bias (JBI score between 50% and 69%), and low risk of bias (JBI score ≥ 70%) [17].

Data analysis
MetaXL version 5.3 software was used for data analysis. Cochrane's Q test evaluated the probable sources of heterogeneity to identify the presence of heterogeneity, and I2 statistics were used to measure the amount of heterogeneity within and between studies using each of the three diagnostic criteria. Q test with p < 0.10 was considered statistically significant heterogeneity, and I2 > 75% was regarded as high heterogeneity [18]. The pooled prevalence of PCOS has been estimated using the random-effects model (DerSimonian-Laird method) [19]. Transformed double arcsine transformation has been used for stabilizing the variance of each study's proportions. Publication bias was evaluated using the Doi plot and Luis Furuya-Kanamori (LFK) asymmetry index [20]. Sensitivity analysis has been done to indicate the major determinant for the pooled prevalence of PCOS and to identify the main source of heterogeneity.

Results
A total of 30 articles were reviewed for a full text, and 11 articles were included in the present study [21][22][23][24][25][26][27][28][29][30][31]. Most of the selected studies were from Southern India, and none of the studies selected were from Eastern India. Table 1 shows the details of the selected study.  While performing the risk of bias assessment using modified JBI criteria ( Table 2), most of the studies were based on community settings except for Singh et al.'s study [28] and Laddad et al's study [29], which were carried out in the outpatient departments of hospitals. All the selected papers reported have used Rotterdam's criteria in addition to those three papers that used NIH and AES criteria. Most selected papers give details about oligo/amenorrhea except for one paper, Bhuvanashree et al. [23], where no detailed information was available for the study's diagnostic criteria.   [31], reported the proportion of females presenting with biochemical hyperandrogenism, and most of them took more than two standard deviation of serum testosterone level in comparison to average women in their reproductive age groups as the cut-off for biochemical hyperandrogenism. Only five studies reported the prevalence of polycystic ovaries, and most of them took the total number of cysts per ovary (n>10-12) and ovarian volume > 10 ml as diagnostic criteria; in addition to, one study, Bhuvanashree et al. [23], also took bilateral presence of multiple sub-cortical ovarian cysts arranged in a necklace pattern as diagnostic criteria for polycystic ovaries.

PCOS, polycystic ovarian syndrome
Similarly, the pooled prevalence of PCOS using the AES and NIH criteria are shown in Figure 3 and Figure 4, respectively.

Heterogeneity and publication bias
The 11 included studies were analyzed for heterogeneity and publication bias. High heterogeneity was found in the analysis with the Q test (p <0.001) and I2 statistics (I2 = 96%). For publication bias, the Doi plot showed asymmetry confirming the presence of bias, and minor asymmetry was seen in the LFK index (LFK index = 1.87) ( Figure 5).

Sensitivity analysis
Each study's effect (i.e. eleven studies) on the pooled prevalence of PCOS has been analyzed by excluding each study step by step using sensitivity analysis ( Table 3). It showed that three studies (Joshi et al. [24], Deswal et al. [25], and Nanjaiah [27]) were comparatively the prime determinants of the pooled prevalence of PCOS, and the higher source of heterogeneity comes from the study by Nidhi et al. [21].   [32,33]. The current paper is the first systematic review and meta-analysis to estimate the overall prevalence of PCOS in India as per three diagnostic criteria. This paper demonstrates that the pooled PCOS prevalence estimates according to Rotterdam's criteria is 11.34% in India. These findings are slightly higher when compared to the metaanalysis conducted by Wu et al., where the overall prevalence of PCOS was 10.01% among Chinese women [34]. The PCOS prevalence rates among Chinese females varied by region; the prevalence rates of PCOS in eastern regions (7.82%) are much lower than those in central (14.24%) and western regions (13.35%) [34]. Since only a few published articles were found in India on the prevalence of PCOS, we could not perform a subgroup analysis based on regions in India. The prevalence of PCOS diagnosed using Rotterdam's criteria (2003) is reportedly higher than the NIH criteria (1990) and AES Criteria (2006) [35].
According to a few studies, Rotterdam's criteria may include some individuals with mild phenotypes of PCOS in the diagnosis, thereby raising PCOS prevalence [36].
The prevalence of PCOS had significant differences among regions, occupation, age, time of publication, diagnostic criteria, and survey populations [34]. The current systematic review shows a difference in the PCOS diagnostic criteria across the included studies. Regarding hyperandrogenism, the cut-off score of modified Ferriman-Gallwey criteria used for hirsutism and the biochemical parameters measuring hormones for hyperandrogenemia was uneven throughout the studies. For example, Nidhi et al.'s study stated that they reported prevalence according to Rotterdam's criteria, which included women with an F-G score cut-off of ≥6 to diagnose hirsutism [21]. However, five studies have stated that they have used an F-G score cut-off of ≥8 to classify it as hirsutism [24,[28][29][30][31]. Similarly, Skiba et al.'s study found a lack of adherence to the recognized PCOS diagnostic criteria across various studies. It further stated that consistent use of Rotterdam's criteria in the research context is complex, and it might raise further issues about its utility as a diagnostic framework [13].
Since the threshold used to measure PCOS by ultrasonography is not mentioned in all the studies included in the current systematic review, it could have led to discrepancies among studies. However, it is unclear if the variation in PCOS prevalence is linked to different thresholds used for measuring the antral follicle count (AFC) and ovarian volume, necessitating more study in this area. The frequency of the transducer used to define PCOs morphology may also have a role in the disparities in prevalence rates [6,37,38].
Including school and college-going adolescents in this current review may have inflated the pooled prevalence estimate. Similarly, Joshi et al.'s study included adolescents and young girls in Mumbai, revealing the highest prevalence of PCOS estimates using Rotterdam's criteria [24]. Various studies had stated that the inclusion of adolescents in their samples might amplify the prevalence estimate when Rotterdam's criteria were used, as both oligo-anovulation and PCOS are common in adolescent girls [39,40]. Furthermore, students have long mental work hours and may be under long-term stress, resulting in increased catecholamine secretion, endocrine function disorder, sympathetic nerve excitability, and secretion of hypothalamus-pituitary-adrenal cortex hormone, all of which reduce immune function. Students frequently make poor lifestyle choices, such as inconsistent eating and little exercise. These variables might hasten the onset of PCOS [34,41].
Though the quality of the study was appraised using the JBI criteria, this paper failed to assess the standard of individual diagnostic methods used to evaluate each diagnostic criterion for PCOS. Furthermore, the age group of all the studies is not uniform. Regional variations were not found as most of the studies are from the southern region of India, and none were from the Eastern part of India; therefore, the result may not reflect India as a whole.

Data availability statement
Data available within the article or its supplementary materials (Appendix Table 5).

Conclusions
The pooled prevalence of PCOS was close to 10% using Rotterdam's criteria and AES criteria, while it was 5.8% using the NIH criteria. The study's overall finding emphasizes the need for more acceptable and uniform diagnostic criteria for screening PCOS. Although physicians are crucial in identifying PCOS and educating the public about this condition, the extra cost and amount of time it takes for a diagnosis and treatment may deter some young women from seeking assistance. Additionally, it is critical for healthcare professionals to communicate this information with cultural sensitivity. The guidelines for the management and awareness of PCOS in India need to be established with the assistance of this evidence by policy-makers, government organizations, and healthcare professionals.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.