Table 2.
Evidence-rating results based on the results of statistical analyses of the 20 associations on cancer prognosis.
| Study | Association between metformin use and | Summary relative risk estimate (random-effect P)* |
Cases >1000 |
Largest study relative risk estimate P < 0.05 |
I2 < 50% |
Small study effects |
95% prediction interval exclude the null value | Excess significance |
10% credibility ceiling survival |
|---|---|---|---|---|---|---|---|---|---|
| Associations supported by highly suggestive evidence (1) | |||||||||
| Du et al. (28) | Colorectal cancer overall survival | +++ | + | + | – | + | – | – | + |
| Associations supported by suggestive evidence (4) | |||||||||
| Lega et al. (22) | All cancer overall survival | ++ | + | – | – | + | – | + | + |
| Tang et al. (24) | Breast cancer overall survival | +++ | + | – | – | + | – | + | + |
| Xin et al. (37) | Lung cancer overall survival | ++ | + | + | – | – | – | – | + |
| Zhou et al. (29) | Pancreatic cancer overall survival | ++ | + | – | – | – | – | – | + |
| Associations supported by weak evidence (11) | |||||||||
| Franciosi et al. (21) | All cancer cancer-specific survival | ++ | – | + | + | – | – | – | + |
| Hu et al. (23) | Bladder cancer cancer-specific survival | + | – | + | – | # | # | + | – |
| Hu et al. (23) | Bladder cancer recurrence-free survival | + | – | – | – | – | – | + | – |
| Du et al. (28) | Colorectal cancer cancer-specific survival | + | + | + | – | – | – | – | – |
| Guo et al. (31) | Endometrial cancer overall survival | ++ | – | + | + | – | – | – | – |
| Chu et al. (30) | Endometrial cancer recurrence-free survival | + | – | – | + | – | – | – | + |
| Li et al. (33) | Kidney cancer overall survival | + | – | – | + | – | – | – | + |
| Li et al. (33) | Kidney cancer cancer-specific survival | + | – | – | + | – | – | – | + |
| Ma et al. (35) | Liver cancer overall survival | + | + | – | – | – | – | + | + |
| Xiao et al. (40) | Prostate cancer overall survival | + | + | + | – | – | – | – | – |
| Xiao et al. (40) | Prostate cancer recurrence-free survival | + | – | – | + | – | – | – | – |
| Associations supported by not suggestive evidence (4) | |||||||||
| Hu et al. (23) | Bladder cancer overall survival | – | – | – | – | – | – | + | – |
| Xu et al. (25) | Breast cancer cancer-specific survival | – | – | – | + | – | – | – | – |
| Mansourian et al. (26) | Colorectal adenoma recurrence-free survival | – | – | – | – | – | – | – | – |
| Xiao et al. (40) | Prostate cancer cancer-specific survival | – | + | + | – | – | – | – | – |
*
P-value calculated using random-effect model: +++, P < 10−6; ++, P < 10−3; +, P < 0.05; –, P > 0.05. For other items, + = yes, – = no.
#Hu et al. (23) included only two primary studies, thus Egger's test and prediction interval are not available.