{"paper_id":"d25c6643-5d10-4640-b851-9220edfd6d19","body_text":"Search strategy : All international databases including MEDLINE, Web of Science, Google scholar, Scopus and Cumulative Index to Nursing and Allied Health Literature were searched for the original articles without language and time limitation, written from January 1990 to December 2018. Keywords were searched electronically by two independent Boolean operators using a specified search strategy. The protocol of this study was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42017075275) 8 .\nInclusion/exclusion criteria : Two reviewers independently carried out the literature search and evaluation of the searched articles based on the inclusion and exclusion criteria. The structures of the searched articles were appraised by the reconstructed PRISMA checklist 9 . All the articles with full text in English conducted as a cross-sectional design reporting the prevalence of endometriosis in any stage of the disease in women aged 15-60 yr old, were include.\nData extraction : Using a uniform excel sheet, two reviewers independently extracted the required data from the data contained in the identified articles. Discrepancies in the extracted data were resolved through consensus, and if agreement could not be reached, they resolved it by referral to a third investigator. The STROBE checklist 10  was used as a standard checklist for reporting the results of the included studies.\nQuality assessment and risk of bias : The quality of each study was assessed according to the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies 11  which included the quality of research question, study population, sampling strategy, groups recruited from the same population and uniform eligibility criteria, sample size justification, exposure assessed before outcome measurement, sufficient timeframe to observe an effect, different levels of the exposure of interest, exposure measurement and assessment, repeated exposure assessment, outcome measurement, blinding of outcome assessors, follow up rate and statistical analyses. This tool measures 14 different criteria which are used to give each study an overall quality rating of good, fair or poor.\nStatistical analysis : Random effects models were used based on the presence of heterogeneity. Heterogeneity was assessed using Q Cochran’s test and  I 2  index 12 . The forest plot was implemented for showing the results of the individual and pooled effects of all the studies. The Egger’s test was also used to evaluate the presence of publication bias 12 . In addition, a subgroup analysis was done (by sample size, stages of endometriosis, diagnostic method, continent and clinical symptoms) to identify different sources of heterogeneity. A  P <0.05 was considered significant for all statistical tests, except for Q Cochran’s, meta-regression and Egger’s test (<0.1). All statistical analyses were performed through STATA version 12.0 (STATA Corp., College Station, TX, USA).\n\nA total of 2433 articles were extracted for this study. Finally, 17 studies 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29  with 127,476 women suffering from endometriosis were included for estimating the prevalence of endometriosis ( Fig. 1 ). The characteristics of the included studies are described in  Table I . The quality assessment details for the included studies are shown in  Table II .\nFlow diagram showing literature search and study selection.\nCharacteristics of the included studies\nMRI, magnetic resonance imaging\nQuality assessment of included articles based on Quality Assesment Tool for observational cohort & cross-sectional studies 11 .\nCD, cannot determine; NA, not applicable; NR, not reported\nA meta-analysis was conducted regarding the stage of the disease, diagnostic method, continent and clinical symptoms to estimate the prevalence of endometriosis. The pooled prevalence estimate of endometriosis regardless of the stage of disease, diagnostic method, continent and clinical symptoms was 18 per cent [95% confidence interval (CI): 16-20;  Fig. 2 ]. The prevalence of endometriosis based on the stage of the disease ranged from two per cent (95% CI: 1-4) for stage 4 to 20 per cent (95% CI: 11-28) for stage 1 ( Table III ).\nPooled prevalence estimate of endometriosis.\nPooled prevalence estimates along with 95% confidence intervals of endometriosis by stage of disease, diagnostic method and continent\n* P <0.05. MRI, magnetic resonance imaging; CI, confidence interval\nThe prevalence of endometriosis by the diagnostic method ranged from 12 per cent (95% CI: 3-21) for endometriosis diagnosed with other diagnostic methods to 20 per cent (95% CI: 17-22) for endometriosis diagnosed with laparoscopic method ( Table III ). The CI of Egger’s test did not include zero, showing significant publication bias ( Fig. 3 ). Furthermore, high statistical heterogeneity ( I 2 >99%,  P <0.001) was identified in total analyses ( Table III ).\nFunnel plot for pooled prevalence estimate by different stages of endometriosis.\nThe prevalence of endometriosis by the continent ranged from 17 per cent (95% CI: 12-21) for Europe to 36 per cent (95% CI: 5-69) for Asia, and by the sample size ranged from 28 per cent (95% CI: 19-37) for studies with less than of 1000 individuals to seven per cent (95% CI: 3-11) for studies with more than of 1000 individuals ( Table III ). The results of subgroup analysis indicated that the prevalence rates of endometriosis in women with infertility, chronic pelvic pain and asymptomatic women were 31 (95% CI: 15-48), 42 (95% CI: 25-58) and 23 per cent (95% CI: 19-26), respectively ( Table III ).\nMeta-regression was used to explore the sources of between-study heterogeneity including age and diagnostic method. According to the results, the prevalence of endometriosis did not show a relationship with age ( P >0.10) and diagnostic method ( P >0.10).\n\nRegarding the stage of endometriosis, the results of this study showed that the prevalence of minimal endometriosis (stage 1) was higher than other stages of endometriosis. Considering the diagnostic method, the prevalence of endometriosis diagnosed with laparoscopy, ultrasound and magnetic resonance imaging (MRI) methods was higher than endometriosis diagnosed with other diagnostic methods in the world. Another study showed the same results 30 .\nThe precise prevalence of endometriosis in female adult population is not known. The prevalence in fertile women undergoing sterilization is four per cent (1.5-5%), whereas it is 13.5 per cent (2-68%) in infertile women 5 . The reported prevalence of this disease in women undergoing laparoscopy for chronic pelvic pain ranged from 15 to 75 per cent 6 . A study conducted in north India in 2015 showed that the precise prevalence of endometriosis was not known; however, it was estimated to be 2-10 per cent in the general population, but up to 50 per cent in infertile women 31 .\nEndometriosis affects approximately 70 per cent of women with dysmenorrhoea and dyspareunia 32 . Adolescents may have more severe symptoms. Reports from Germany showed that 0.05, 1.93 and 6.1 per cent of the patients were in the age groups of 10-14, 15-19 and 20-24 yr, respectively 33 34 35 . A review of previous studies indicated that global estimates varied significantly and ranged from approximately 2-45 per cent based on the diagnostic criteria and the study population 35 . The world statistics suggest that 10-15 per cent of women in the world are affected by endometriosis, which is consistent with the results of our study, as the overall prevalence was estimated to be 18 per cent 35 . According to our analysis, the prevalence of endometriosis in developing countries was higher than in developed countries as also shown by another study 36 .\nThe present study had several limitations. First, the number of studies in some subgroups was small and did not provide sufficient statistical power to assess the source of heterogeneity. Second, there were some other factors (such as diagnostic accuracy, quality of detective equipment and physicians’ skills in detecting endometriosis) that might be important sources of heterogeneity, but we could not evaluate their role in heterogeneity due to the lack of information. Third, some of the included studies did not measure the variables such as age or prevalence of endometriosis by the stage of the disease.\nDespite the high heterogeneity of the studies, this systematic review and meta-analysis showed a high prevalence of endometriosis in developing countries. The prevalence of endometriosis in women with chronic pelvic pain was higher than those with infertility. Future studies are needed to explore factors affecting endometriosis prevalence worldwide, which may help develop future prevention programmes.","source_license":"CC0","license_restricted":false}