Skip to main content
. 2017 Apr 8;17:251. doi: 10.1186/s12879-017-2309-9

Table 2.

Univariable logistic regression analysis of potential predictors of tuberculosis death

Predictors Overall
N = 2250
Dead patients
N = 213 (%)
Treatment success
N = 2037 (%)
Crude OR (95% CI) p-value
Gender 0.914
 Male 1286 (57.2) 121 (56.8) 1165 (57.2) 0.94 (0.74–1.31)
 Female 964 (42.8) 92 (43.2) 872 (42.8) 1
Age, years 35.8 (12.7) 42.4 (13.5) 35.1 (12.4) 1.04 (1.03–1.05) <0.0001
Adjusted BMI, kg/m2 21.4 (4.0) 19.2 (4.9) 21.6 (3.9) 0.83 (0.80–0.87) <0.0001
Initial hospitalization
 Yes 1363 (60.6) 146 (68.5) 1217 (59.7) 1.47 (1.08–1.99) 0.012
 No 887 (39.4) 67 (31.7) 820 (40.3) 1
Clinical form <0.0001
 PTB+ 1527 (67.9) 107 (50.2) 1420 (69.7) 1
 PTB− 210 (9.3) 42 (19.7) 168 (8.2) 3.32 (2.24–4.91)
 ETB 513 (22.8) 64 (30.0) 449 (22.0) 1.89 (1.36–2.62)
Type of patient 0.393
 New cases 2119 (94.2) 199 (93.4) 1920 (94.3) 1
 Retreatment 106 (4.7) 13 (6.1) 93 (4.6) 1.35 (0.74–2.4)
 Other 25 (1.1) 1 (0.5) 24 (1.2) 0.40 (0.05–2.99)
HIV serology
 Positive 788 (35.0) 149 (69.9) 639 (31.4) 5.09 (3.74–6.93) <0.0001
 Negative 1462 (65.0) 64 (30.0) 1398 (68.6) 1

OR odds ratio, BMI  body mass index adjusted by divided weight by the square of 1.70m  for men and 1.60m for women, SD standard deviation, PTB+ smear positive pulmonary tuberculosis, PTB- smear negative pulmonary tuberculosis, ETB extra-pulmonary tuberculosis