Table 1.
Characteristics of included studies
| Included studies | Location | Sample size (hyperandrogenemia/nonhyperandrogenemia) | Mean age (range, year) | PCOS diagnostic criteria | Type of study | Extracted indexj |
|---|---|---|---|---|---|---|
| Hosseinpanah 2014 [6] | Iran | 136 (109/27) | 33.6 (18 ~ 45) | 2003 Rotterdam criteria | Cross-sectional | b, d, f |
| Kim 2014 [7] | Korea | 700 (432/268) | 27.9 (15 ~ 40) | 2003 Rotterdam criteria | Cross-sectional | a |
| Lerchbaum 2014 [8] | Austria | 706 (352/354) | 27h (16 ~ 45) | 2003 Rotterdam criteria | Cross-sectional | a, b |
| Livadas 2014 [9] | Greece | 1218 (716/502) | 23h | 2003 Rotterdam criteria | Cross-sectional | – |
| Sung 2014 [10] | Korea | 1062 (645/417) | 24 | 2003 Rotterdam criteria | Cross-sectional | a, c, d, e, f |
| Tehrani 2014 [11] | Iran | 85 (72/13) | 29.07 (18 ~ 45) | 2003 Rotterdam criteria | Cross-sectional | a, c, d, e, f, g |
| Ates 2013 [12] | Turkey | 410 (334/76) | 24.55 | 2003 Rotterdam criteria | Cross-sectional | a, c, e, f, g |
| Di Sarra 2013 [13] | Italy | 89 (65/24) | 23.6 (18 ~ 40) | 2003 Rotterdam criteria | Cross-sectional | d, e, f, g |
| Zhu 2013 [14] | Shanghai, China | 53 (28/25) | 22.82 | 2003 Rotterdam criteria | Cross-sectional | d, e, f, g |
| Gluszak 2012 [15] | Poland | 93 (88/5) | 23.95 | 2003 Rotterdam criteria | Cross-sectional | c, d, e, f, g |
| Jones 2012 [16] | United Kingdom | 29 (19/10) | 28 | 2003 Rotterdam criteria | Cross-sectional | – |
| Li 2012 [17] | Guangdong, China | 131 (62/69) | 29.57 | 2003 Rotterdam criteria | Cross-sectional | c, d, e, f, g |
| Ozkaya 2012 [18] | Turkey | 132 (100/32) | 24.21 | 2003 Rotterdam criteria | Cross-sectional | c, d, e, f, g |
| Cupisti 2011i [19] | Germany | 309 (293/16) | 27.16 | 2006AES criteriai | Cross-sectional | c, d, e, f, g |
| Mehrabian 2011 [20] | Iran | 539 (287/252) | 29.3 (18 ~ 42) | 2003 Rotterdam criteria | Cross-sectional | a, b, c, f |
| Melo 2011 [21] | Brazil | 226 (175/51) | 26.45 | 2003 Rotterdam criteria | Cross-sectional | a, c, d, e, f, g |
| Wijeyaratne 2011 [22] | Sri Lanka | 469 (374/95) | 25 | 2003 Rotterdam criteria | Cross-sectional | a |
| Yilmaz 2011 [23] | Turkey | 127 (103/24) | 25.36 (18 ~ 35) | 2003 Rotterdam criteria | Cross-sectional | a, c, d, e, f, g |
| Castelo-Branco 2010 [24] | Spain | 197 (152/45) | 28.4 | 2003 Rotterdam criteria | Cross-sectional | e, f, g |
| Guo 2010 [25] | Shandong, China | 615 (571/44) | 28.3 (20 ~ 41) | 2003 Rotterdam criteria | Cross-sectional | a, c, d, e, f, g |
| Goverde 2009 [26] | Netherlands | 157 (101/56) | 29 (17 ~ 43) | 2003 Rotterdam criteria | Cross-sectional | a, b, c, f |
| Barber 2007 [27] | United Kingdom | 309 (267/42) | 33.26 | 2003 Rotterdam criteria | Cross-sectional | a |
| Shroff 2007 [28] | United States | 258 (224/34) | 27.86 (18 ~ 45) | 2003 Rotterdam criteria | Cross-sectional | a, c, d, e, f, g |
| Chen H 2014 [29] | Shanghai, China | 126 (34/92) | 27 | 2003 Rotterdam criteria | Cross-sectional | c, d, e, f, g |
| Li YC 2014 [30] | Guangxi, China | 68 (42/26) | 25.51 (18 ~ 37) | 2003 Rotterdam criteria | Cross-sectional | d, e, f, g |
| Ha LX 2013 [31] | Ningxia, China | 267 (127/140) | 25.21 | 2003 Rotterdam criteria | Cross-sectional | c, d, e, f, g |
| Tao T 2013 [32] | Shanghai, China | 305 (248/57) | 26.44 (18 ~ 45) | 2003 Rotterdam criteria | Cross-sectional | a |
| Li J 2011 [33] | Shanghai, China | 95 (84/11) | Unknown | 2003 Rotterdam criteria | Cross-sectional | c, d, e, f, g |
| Liu L 2011 [34] | Zhejiang, China | 48 (34/14) | 27.15 (23 ~ 33) | 2003 Rotterdam criteria | Cross-sectional | d, e, f, g |
| Qu ZY 2011 [35] | Shandong, China | 306 (177/129) | Unknown | 2003 Rotterdam criteria | Cross-sectional | b |
| Xu LS 2010 [36] | Tianjin, China | 256 (152/104) | 23.8 (14 ~ 39) | 2003 Rotterdam criteria | Cross-sectional | b, c |
| Zhang L 2010 [37] | Jiangsu, China | 35 (15/20) | 29.43 (21 ~ 35) | 2003 Rotterdam criteria | Cross-sectional | b |
aNumber of cases with MetS; bNumber of cases with IR; cHOMA-IR value; dTC value; eTG value; fHDL value; gLDL value; hMedian; iPCOS typing had10 subtypes, and the rest had four subtypes; jMeant that the corresponding outcome data were not exactable if they were data of median or quartiles that could not be converted into mean ± standard deviation