Table 2.
a: Leave-one-out sensitivity analysis for GMIs (ertugliflozin vs. control). b: Sensitivity analysis by excluding two studies that were not placebo-controlled.
| Study excluded | RR (95% CI) | Z-test p value | Heterogeneity (I2) |
|---|---|---|---|
| a | |||
| 15 mg vs. control 26 weeks | |||
| Dagogo-Jack (2018) [19] | 4.42 (1.96, 10.00) | p = 0.0004 | p = 0.33; I2 = 12% |
| Ji (2019) [14] | 5.98 (2.84, 12.56) | p < 0.00001 | p = 0.53; I2 = 0% |
| Pratley (2018) [17] | 4.34 (1.99, 9.47) | p = 0.0002 | p = 0.36; I2 = 8% |
| Rosenstock (2018) [6] | 4.62 (1.93, 11.03) | p = 0.0006 | p = 0.29; I2 = 20% |
| Terra (2017) [21] | 6.39 (2.12, 19.27) | p = 0.0010 | p = 0.28; I2 = 21% |
| 5 mg vs. control 26 weeks | |||
| Dagogo-Jack (2018) [19] | 3.67 (1.73, 7.77) | p = 0.0007 | p = 0.38; I2 = 2% |
| Ji (2019) [14] | 4.74 (2.21, 10.15) | p < 0.0001 | p = 0.40; I2 = 0% |
| Pratley (2018) [17] | 3.55 (1.70, 7.42) | p = 0.0008 | p = 0.46; I2 = 0% |
| Rosenstock (2018) [6] | 3.77 (1.73, 8.26) | p = 0.0009 | p = 0.36; I2 = 7% |
| Terra (2017) [21] | 5.93 (2.25, 15.58) | p = 0.0003 | p = 0.46; I2 = 0% |
| 15 mg vs. control 52 weeks | |||
| Aronson (2018) [20] | 9.32 (4.03, 21.52) | p < 0.00001 | p = 0.91; I2 = 0% |
| Dagogo-Jack (2018) [19] | 5.24 (2.44, 11.27) | p < 0.0001 | p = 0.23; I2 = 33% |
| Hollander (2018) [18] | 4.73 (2.14, 10.44) | p = 0.0001 | p = 0.29; I2 = 19% |
| Pratley (2018) [17] | 5.82 (2.27, 14.93) | p = 0.0002 | p = 0.16; I2 = 45% |
| 5 mg vs. control 52 weeks | |||
| Aronson (2018) [20] | 8.66 (3.74, 20.06) | p < 0.00001 | p = 0.89; I2 = 0% |
| Dagogo-Jack (2018) [19] | 4.73 (2.00, 11.16) | p = 0.0004 | p = 0.17; I2 = 44% |
| Hollander (2018) [18] | 4.51 (1.74, 11.70) | p = 0.002 | p = 0.21; I2 = 37% |
| Pratley (2018) [17] | 5.35 (1.85, 15.42) | p = 0.002 | p = 0.11; I2 = 55% |
| 15 mg vs. control 104 weeks | |||
| Gallo (2019) [16] | 11.67 (3.62, 37.65) | p < 0.0001 | NA |
| Hollander (2019) [15] | 5.44 (1.61, 18.38) | p = 0.006 | NA |
| 5 mg vs. control 104 weeks | |||
| Gallo (2019) [16] | 10.43 (3.22, 33.80) | p < 0.0001 | NA |
| Hollander (2019) [15] | 4.38 (1.27, 15.13) | p = 0.02 | NA |
|
| |||
| b | |||
| 15 mg vs. control 52 weeks | |||
| Hollander (2018) [18]; Pratley (2018) [17] | 4.73 (1.29, 17.39) | p = 0.02 | p = 0.18; I2 = 43% |
| 5 mg vs. control 52 weeks | |||
| Hollander (2018) [18]; Pratley (2018) [17] | 4.41 (0.94, 20.64) | p = 0.06 | p = 0.13; I2 = 56% |
RR: risk ratio; CI: confidence interval; NA: not available.