Urine LAM as a predictor of mortality among participants receiving anti-TB therapy
| Predictor | Number of deaths/number at risk (%) | Univariate HR (95% CI) | p Value | Multivariate HR* (95% CI) | p Value |
|---|---|---|---|---|---|
| Baseline visit | |||||
| Laboratory-based LAM negative | 19/70 (27.1) | 1.00 (reference) | – | 1.00 (reference) | – |
| Laboratory-based LAM positive | 6/20 (30.0) | 1.13 (0.45–2.83) | 0.79 | 1.19 (0.40–3.55) | 0.75 |
| Rapid LAM negative | 16/61 (31.0) | 1.00 (reference) | – | 1.00 (reference) | – |
| Rapid LAM positive | 9/29 (31.0) | 1.22 (0.54–2.76) | 0.63 | 1.41 (0.54–3.70) | 0.49 |
| Rapid LAM score <2 | 21/76 (27.6) | 1.00 (reference) | – | 1.00 (reference) | – |
| Rapid LAM score ≥2 | 4/14 (28.6) | 1.04 (0.36–3.04) | 0.66 | 0.99 (0.28–3.58) | 0.99 |
| 2-month visit | |||||
| Laboratory-based LAM negative | 12/64 (18.8) | 1.00 (reference) | – | 1.00 (reference) | – |
| Labratory-based LAM positive | 4/9 (44.4) | 3.05 (0.98–9.48) | 0.05 | 4.13 (0.88–19.4) | 0.07 |
| Rapid LAM negative | 12/61 (19.7) | 1.00 (reference) | – | 1.00 (reference) | – |
| Rapid LAM positive | 4/12 (33.3) | 2.00 (0.65–6.21) | 0.23 | 1.99 (0.52–7.65) | 0.32 |
| Rapid LAM grade <2 | 12/65 (18.5) | 1.00 (reference) | – | 1.00 (reference) | – |
| Rapid LAM grade ≥2 | 4/8 (50.0) | 3.63 (1.17–11.3) | 0.03 | 5.58 (1.24–25.2) | 0.03 |
| 6-month visit | |||||
| Rapid LAM negative | 4/45 (8.9) | 1.00 (reference) | – | 1.00 (reference) | – |
| Rapid LAM positive | 2/5 (40.0) | 5.56 (1.00–30.7) | 0.05 | 42.1 (1.87–952) | 0.02 |
*Models adjusted by age, gender, Karnofsky performance score and HIV status.
Rapid LAM result modelled as negative (grade <1) versus positive (grade ≥1), and grade <2 versus grade ≥2.
LAM, lipoarabinomannan; TB, tuberculosis.