Summary characteristics for participants at the enrolment study visit
| Total cohort | Males | Females | p-value# | |
| Subjects n | 734 | 223 | 511 | |
| Age years | 34 (28–40) | 37 (32–43) | 32 (27–38) | <0.001 |
| Smoking status¶ | <0.001 | |||
| Current | 90 (12) | 62 (28) | 28 (5) | |
| Former | 82 (11) | 44 (20) | 37 (7) | |
| Never | 562 (77) | 116 (52) | 446 (87) | |
| Household asset ownership index+ | ns | |||
| Lowest quartile | 235 (32) | 62 (28) | 173 (34) | |
| Second quartile | 173 (24) | 57 (26) | 116 (23) | |
| Third quartile | 177 (24) | 58 (26) | 119 (23) | |
| Highest quartile | 140 (19) | 42 (19) | 98 (19) | |
| High-risk occupation§ | 237 (32) | 117 (52) | 120 (23) | <0.001 |
| Cooking fuel type | 0.001 | |||
| Charcoal | 297 (40) | 70 (31) | 227 (44) | |
| Firewood | 428 (58) | 149 (67) | 279 (55) | |
| Self-reported pulmonary disease history | ||||
| Pulmonary tuberculosis | 60 (8) | 26 (12) | 34 (7) | 0.02 |
| Pneumocystis pneumonia | 38 (5) | 14 (6) | 24 (5) | ns |
| Cryptococcal pneumonia | 4 (1) | 2 (1) | 2 (<1) | ns |
| CD4 count | ns | |||
| <100 cells·μL−1 | 198 (27) | 73 (33) | 125 (24) | |
| 100–349 cells·μL−1 | 441 (60) | 124 (56) | 317 (62) | |
| 350–499 cells·μL−1 | 59 (8) | 18 (8) | 41 (8) | |
| ≥500 cells·μL−1 | 34 (5) | 7 (3) | 27 (5) | |
| Viral load | 0.01 | |||
| ≤10 000 copies·mL−1 | 79 (11) | 14 (6) | 65 (13) | |
| >10 000 copies·mL−1 | 649 (88) | 206 (92) | 443 (87) | |
| Study visits per participant n | 20 (7–28) | 16 (7–28) | 21 (8–28) | ns |
| Observation time years | 5 (2–7) | 4 (2–7) | 5 (2–7) | ns |
Data are presented as n (%) or median (interquartile range), unless otherwise stated. #: p-values for comparisons between sexes, calculated using t-test for normally distributed variables, Wilcoxon rank sum testing for non-normally distributed variables and Chi-squared testing for dichotomous and categorical variables; ¶: never-smokers defined as those who did not smoke prior to cohort enrolment and continued not to smoke; former smokers defined as those who smoked prior to cohort enrolment, but did not smoke throughout the duration of the study; current smokers defined as those who smoked at cohort enrolment and continued to smoke throughout the duration of the study; +: constructed following the method of Filmer and Pritchett [9], using the following binary indicator variables: home ownership; land ownership; livestock ownership; presence of the following items (yes/no) in the household: iron, stove, fridge, phone, motorcycle, clock, bed, sofa, bike, television, lantern, cupboard, shoes, car, radio, mattress or electricity; characteristics of the home: flushing toilet, cement wall, cement floor or piped water; §: includes farmer, cook, motorcyclist, brick worker, builder, carpenter, cattle keeper, dry cleaner and sand miner. ns: nonsignificant.