Table 1. Characteristics of the included studies for the risk of prostate cancer analysis.
| Author publication year | Country/region | Study design | Study period(years) | Male subjects | PCa cases | Treatment comparison | Measure of outcome | Adjustments* | Study quality score |
| Currie et al., 2009 | UK | retrospective cohort study | 2.4 (2000–2003) | 32261 | 301 | metformin vs Sulfonylureas; metformin vs Insulin-based therapies | aHR | 1–3 | 7 |
| Ferrara et al., 2011 | USA | retrospective cohort study | 8 (1997–2005) | 134864 | 2105 | metformin vs pioglitazone | aHR | 1,2,5,9,13,15 | 5 |
| Azoulay et al., 2011 | Canada | nested case–control study | 21(1988–2009) | 8098 | 739 | Metformin vs other hypoglycemic agent | aRR | 2,3,5,6,9,12,16 | 7 |
| Murtola et al., 2008 | Finland | Case-Control Study | 7(1995–2002) | 49446 | 24723 | Metformin user vs nonuser | aOR | 1,4,16 | 6 |
| Wright et al.,2009 | USA | Case-Control Study | 3(2002–2005) | 1943 | 1001 | Metformin user vs nonuser | aOR | 1,5–8,16 | 6 |
| Hsieh et al.,2012 | Taiwan | retrospective cohort study | 8(2000–2008) | 5680 | 84 | metformin vs Insulin; metformin vs Sulfonylureas | aOR | 1 | 6 |
| Onitilo et al.,2013 | USA | retrospective cohort study | 14(1995–2009) | 4956 | 237 | Metformin vs other hypoglycemic agent | aHR | 1,2,4,6,10,14,15 | 6 |
| Ruiter et al.,2012 | Netherlands | retrospective cohort study | 10(1998–2008) | 40131 | 236 | metfromin vs Sulfonylurea derivatives | aHR | 1,5,10 | 8 |
| Morden et al.,2011 | USA | Retrospective cohort study | 5(2003–2008) | 25660 | 2072 | metformin vs insulin | aHR | 1,2,6,10,13,15 | 5 |
| Margel et al.,2013 | Canada | nested case–control study | 15(1994–2009) | 31836 | 5306 | Metformin vs other hypoglycemic agent | aOR | 4–6,10,15,16 | 7 |
| Tseng,2011 | Taiwan | retrospective cohort study | 2(2003–2005) | 494630 | 889 | Metformin user vs nonuser | aHR | 1,4–6,10,15,16 | 5 |
| Preston et al.,2014 | Denmark | nested case–control study | 22(1989–2011) | 134486 | 12226 | Metformin user vs nonuser | aOR | 10,16 | 8 |
*Adjustments: 1.Age, 2.smoking status, 3.prior cancer diagnosis, 4.place of residency, 5.use of other diabetes medications, 6.BMI, 7.PSA value, 8.family history of prostate cancer, 9.HbA1c, 10.Comorbidities, 11.Townsend index of deprivation, 12.excessive alcohol use, 13.race/ethnicity, 14.year of diagnosis, 15.socioeconomic status, 16.other drug use, 17.ADT treatment, 18.T stage, 19.Gleason score, 20.primary treatment with radiation or surgery, 21.surgical margin, 22.lymph node metastasis, 23. extracapsular extension, 24. seminal vesicle invasion.