Volume 19, Number 3—March 2013
Research
Treatment Outcomes for Extensively Drug-Resistant Tuberculosis and HIV Co-infection
Max R. O’Donnell
, Nesri Padayatchi, Charlotte Kvasnovsky, Lise Werner, Iqbal Master, and C. Robert Horsburgh
, Nesri Padayatchi, Charlotte Kvasnovsky, Lise Werner, Iqbal Master, and C. Robert HorsburghTable 4
Predictors of 36 deaths at 24 months of treatment for 82 HIV-positive XDR TB patients, KwaZulu-Natal Province, South Africa*
| Predictor | No. died/total (%) | Univariate analysis |
Multivariate analysis† |
|||
|---|---|---|---|---|---|---|
| Hazard ratio (95% CI) | p value | Hazard ratio (95% CI) | p value | |||
| Sex | ||||||
| F | 26/52 (50.0) | 1.55 (0.75–3.21) | 0.2405 | 1.82 (0.83–4.01) | 0.1349 | |
| M |
10/30 (33.3) |
1.0 (referent) |
NA |
1.0 (referent) |
NA |
|
| Age, y | ||||||
| <36 | 20/45 (44.4) | 1.05 (0.55–2.03) | 0.8765 | NA | NA | |
| >36 |
16/37 (43.2) |
1.0 (referent) |
NA |
NA |
NA |
|
| Previous TB treatment‡ | ||||||
| Yes | 29/66 (43.9) | 1.68 (0.51–5.52) | 0.3913 | 1.70 (0.51–5.65) | 0.3865 | |
| No | 3/12 (25.0) | 1.0 (referent) | NA | 1.0 (referent) | NA | |
| Unknown |
4/4 (100.0) |
NA |
NA |
NA |
NA |
|
| Initial sputum smear result | ||||||
| Positive | 22/48 (45.8) | 1.03 (0.53–2.01) | 0.9354 | NA | NA | |
| Negative |
14/34 (41.2) |
1.0 (referent) |
NA |
NA |
NA |
|
| Initial CD4 cell count/mm3 | ||||||
| <200 | 13/29 (44.8) | 1.02 (0.52–2.02) | 0.9495 | NA | NA | |
| >200 |
23/53 (43.4) |
1.0 (referent) |
NA |
NA |
NA |
|
| Receiving ART | ||||||
| Yes | 18/50 (36.0) | 0.54 (0.28–1.03) | 0.0633 | 0.46 (0.22–0.94) | 0.0333 | |
| No |
18/32 (56.3) |
1.0 (referent) |
NA |
1.0 (referent) |
NA |
|
| Adverse event | ||||||
| Yes | 18/35 (51.4) | 1.43 (0.74–2.76) | 0.2832 | 1.89 (0.92–3.86) | 0.0812 | |
| No | 18/47 (38.3) | 1.0 (referent) | NA | 1.0 (referent) | NA | |
*XDR TB, extensively drug-resistant tuberculosis; NA, not applicable; ART, antiretroviral therapy.
†Significant variables or variables that caused >10% change in the bivariate hazard ratio were included in the multivariate model.
‡Patients whose previous TB treatment history was unknown were excluded from analyses.
