Table 2.
Sensitivity analyses of mortality associated with incretin based treatment in meta-analysis of randomised controlled trials in patients with type 2 diabetes. Figures are Mantel-Hanzel point estimates (relative risk or odds ratio) unless stated otherwise
| Comparison | Point estimate (95% CI) |
|---|---|
| Primary analysis | Peto OR 0.96 (0.90 to 1.02) |
| Sensitivity analyses | |
| Alternative effect measure: RR (fixed model) | RR 0.95 (0.90 to 1.01) |
| GLP-1 agonists v control | RR 0.89 (0.81 to 0.98) |
| DPP-4 inhibitors v control | RR 0.99 (0.92 to 1.07) |
| Alternative pooling method: Mantel-Hanszel’s method (fixed model) | OR 0.95 (0.89 to 1.01) |
| GLP-1 agonists v control | OR 0.89 (0.80 to 0.98) |
| DPP-4 inhibitors v control | OR 0.99 (0.92 to 1.08) |
| Statistical models: random effects | OR 0.95 (0.89 to 1.02) |
| GLP-1 agonists v control | OR 0.88 (0.80 to 0.98) |
| DPP-4 inhibitors v control | OR 1.00 (0.92 to 1.08) |
| A continuity correction of 0.5 | OR 0.95 (0.89 to 1.01) |
| Excluding trials with zero or one event across arms | Peto OR 0.96 (0.90 to 1.02) |