Table 4. Prognostic factors associated with treatment success in a time-dependent Cox regression model.
Covariate | Classification (n) | Multivariate a aHR (95% CI) | p value | Multivariate aHR b (95% CI) | p value |
TMTC era | 24 months (290) | 2.3 (1.9–2.8) | <0.001 | 2.4 (1.9–2.9) | <0.001 |
36 months (290) | 6.0 (3.9–9.1) | <0.001 | 6.3 (4.2–9.5) | <0.001 | |
Sex | Male (494) | 0.9 (0.8–1.2) | 0.579 | 0.9 (0.8–1.1) | 0.477 |
Female (157) | Reference | Reference | |||
Aboriginal | Yes (124) | 1.2 (1.0–1.6) | 0.080 | 1.2 (1.0–1.5) | 0.092 |
No (527) | Reference | Reference | |||
BMI | <22 (397) | Reference | — | ||
22∼26 (191) | 1.0 (0.8–1.2) | 0.989 | — | ||
>26 (63) | 1.2 (0.9–1.7) | 0.206 | — | ||
Age | <35 (128) | 0.9 (0.7–1.2) | 0.334 | 0.9 (0.7–1.1) | 0.237 |
35∼60 (346) | 0.8 (0.6–1.0) | 0.083 | 0.8 (0.6–1.0) | 0.021 | |
>60 (177) | Reference | Reference | |||
Alcohol | Yes (125) | 0.9 (0.7–1.1) | 0.368 | — | |
No (526) | Reference | — | |||
Diabetics | Yes (234) | 1.0 (0.8–1.2) | 0.664 | — | |
No (417) | Reference | — | |||
Hypertension | Yes (104) | 1.1 (0.9–1.4) | 0.375 | — | |
No (547) | Reference | — | |||
Hepatitis B | Yes (51) | 1.0 (0.7–1.4) | 0.863 | — | |
No (600) | Reference | — | |||
Hepatitis C | Yes (56) | 0.7 (0.5–0.9) | 0.022 | 0.7 (0.5–1.0) | 0.030 |
No (595) | Reference | Reference | |||
Cavitary lesion on CXR | Yes (282) | 0.9 (0.7–1.1) | 0.240 | — | |
No (369) | Reference | — | |||
Smear-negative at the time of MDR-TB diagnosis | Negative (232) | 1.1 (0.9–1.3) | 0.378 | — | |
Positive (419) | Reference | — | |||
Number of first-line drugs to which isolate is resistant | ≥3 (308) | 0.9 (0.8–1.1) | 0.322 | — | |
<3 (343) | Reference | — | |||
Treatment delay c | No (457) | 0.8 (0.6–0.9) | 0.012 | 0.8 (0.6–1.0) | 0.018 |
Yes (194) | Reference | Reference | |||
Patient classification | New (245) | 3.5 (2.3–5.1) | <0.001 | 3.6 (2.5–5.4) | <0.001 |
Relapse+ treatment after default+ treatment after failure of the first treatment (350) | 2.5 (1.7–3.6) | <0.001 | 2.5 (1.7–3.7) | <0.001 | |
Treatment after failure of re-treatment (56) | Reference | Reference |
Adjusted with all covariates and stratified with culture converted before second-line drug.
Adjusted with covariates chosen by Akaike information criterion (AIC) and stratified with culture converted before second-line drug.
Treatment delay: the lag between sputum collection of MDR-TB and start of second-line drug > 120 days.
Abbreviations: aHR: adjusted hazard ratio; BMI: body mass index; CI: confidence interval; CXR: chest X-ray; TMTC: Taiwan Multi-drug Resistance Tuberculosis Consortiums.