Table 3. Individual patient factors associated with the three study outcomes in the first 6 months after starting ART.
| Patient factors | No follow-up |
Initially followed-up, then lost |
Death |
|||||
|---|---|---|---|---|---|---|---|---|
| Odds ratioa (95% CI) | P-valueb | Hazard ratioa (95% CI) | P-valueb | Hazard ratioa (95% CI) | P-valueb | |||
| Age (years) | < 0.0001 | 0.0040 | 0.061 | |||||
| 16–29 | 0.83 (0.61–1.14) | 1.38 (1.05–1.83) | 0.84 (0.38–1.83) | |||||
| 30–39 | 0.72 (0.56–0.94) | 1.05 (0.82–1.36) | 1.36 (0.67–2.78) | |||||
| 40–49 | 0.73 (0.64–0.83) | 1.28 (0.98–1.66) | 0.84 (0.38–1.84) | |||||
| ≥ 50 | 1 | 1 | 1 | |||||
| Sex | 0.25 | 0.93 | 0.31 | |||||
| Male | 1 | 1 | 1 | |||||
| Female | 0.76 (0.49–1.20) | 0.99 (0.86–1.15) | 0.83 (0.58–1.18) | |||||
| Year of starting ART | 0.055 | < 0.0001 | 0.35 | |||||
| ≤ 2000 | 1 | 1 | 1 | |||||
| 2001–2002 | 4.54 (1.28–16.1) | 2.70 (1.64–4.46) | 1.38 (0.69–2.78) | |||||
| 2003–2004 | 5.06 (1.28–20.0) | 7.62 (4.55–12.8) | 1.02 (0.44–2.36) | |||||
| Initial ART regimen | 0.31 | 0.35 | 0.58 | |||||
| NNRTI-based | 1 | 1 | 1 | |||||
| PI-based | 0.30 (0.06–1.45) | 0.79 (0.50–1.23) | 1.38 (0.63–3.02) | |||||
| Unknown or other combination | 1.24 (0.44–3.44) | 1.21 (0.77–1.92) | 1.75 (0.40–7.72) | |||||
| Baseline CD4-cell count (cells/µl)c | < 0.0001 | < 0.001 | < 0.0001 | |||||
| ≥ 50 | 1 | 1 | 1 | |||||
| 25–50 | 2.76 (1.69–4.51) | 1.03 (0.81–1.32) | 1.52 (0.79–2.93) | |||||
| < 25 | 2.49 (1.43–4.33) | 1.48 (1.23–1.77) | 3.34 (2.10–5.30) | |||||
| Not measured | 2.88 (1.43–5.77) | 1.16 (0.96–1.40) | 1.81 (0.97–3.40) | |||||
| Clinical staged | 0.26 | 0.83 | 0.036 | |||||
| CDC stage A/B, WHO stage I/II | 1 | 1 | 1 | |||||
| CDC stage C, WHO stage III/IV | 0.85 (0.43–1.69) | 0.96 (0.69–1.35) | 5.35 (1.50–19.1) | |||||
| Not assessed | 3.73 (0.77–18.1) | 1.07 (0.76–1.51) | 4.36 (0.93–20.5) | |||||
ART, antiretroviral therapy; CDC, Centers for Disease Control and Prevention; CI, confidence interval; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor. a Odds ratios were calculated using multivariable random-effects logistic regression, and hazard ratios were calculated using multivariable Weibull proportional hazard models incorporating the competing risks of death and loss to follow-up. Estimates were adjusted for all variables listed in Table 3. b P-values were calculated using the Wald test. c Not measured in 1404 patients. d Not assessed in 3004 patients.