Table 4. The pooled estimates of isoniazid preventive therapy effect on all-types of Tuberculosis.
| Exposure category | Number of studies | Sample size | Pooled RR* (95% CI) | 95% PI | |
|---|---|---|---|---|---|
| Overall estimate | 10 | 7619 | 0.65 (0.51, 0.84) | (0.37, 1.17) | |
| TST | Positive | 5 | 1703 | 0.48 (0.29, 0.82) | (0.13, 1.81) |
| Negative | 9 | 3140 | 0.79 (0.58, 1.08) | (0.54, 1.16) | |
| Unknown | 4 | 2776 | 0.68 (0.42, 1.10) | (0.11, 4.23) | |
| ART | Treated | 2 | 2226 | 0.67(0.47, 0.96) | Not estimable** |
| Not treated | 8 | 4234 | 0.73 (0.53, 1.02) | (0.33, 1.60) | |
| IPT dose | 300mg | 8 | 6819 | 0.62 (0.47, 0.82) | (0.34, 1.12) |
| 900mg | 2 | 800 | 0.89 (0.36, 2.18) | Not estimable** | |
| IPT duration | 6 months | 6 | 5837 | 0.61 (0.45, 0.82) | (0.30, 1.22) |
| 12 months | 4 | 1782 | 0.79 (0.45, 1.37) | (0.11, 5.73) | |
* Random effect model
**Prediction interval can only be estimated for more than two studies
ART: Antiretro viral therapy; CI: Confidence interval; IPT: Isoniazid preventive therapy; PI: Prediction interval; RR: Relative Risk; TST: Tuberculin Skin Test