Table 8.
Outcome and data sources | Events/total population | Effect estimate (95% CI) | |
---|---|---|---|
Fixed effect | Random study specific effects* | ||
Composite outcome | |||
IPD | RSG: 274/11 837; control: 219/9319 | 1.39 (1.15 to 1.68) | 1.39 (1.15 to 1.83) |
Myocardial infarction | |||
IPD | RSG: 147/11 837; control: 133/9319 | 1.27 (1.00 to 1.64) | 1.27 (1.00 to 1.92) |
IPD+summary | RSG: 190/24 467; control: 173/20 372 | 1.20 (0.97 to 1.49) | 1.20 (0.96 to 1.51) |
Heart failure | |||
IPD | RSG: 122/11 837; control: 80/9319 | 1.64 (1.22 to 2.21) | 1.64 (1.22 to 2.24) |
Cardiovascular related deaths | |||
IPD | RSG: 15/11 837; control: 10/9319 | 1.25 (0.56 to 2.92) | 1.25 (0.46 to 2.92) |
IPD+summary | RSG: 41/24 020; control: 30/19 908 | 1.22 (0.76 to 1.98) | 1.22 (0.74 to 1.98) |
Non-cardiovascular related deaths | |||
IPD | RSG: 22/11 837; control: 13/9319 | 1.36 (0.69 to 2.84) | 1.36 (0.61 to 2.83) |
IPD=individual patient level data; RSG=rosiglitazone.
Simmonds and Higgin’s model with random study specific effects.