First author [ref.] | Country | Study period and type | DR/DS | Method of patient selection | Index patients n | Age years mean±sd | Sex F/M % | HIV status | DR pattern n (%) | Contacts n (mean) | Outcome measures for M. tuberculosis infection | TB infection events n/N (%) | Latent TB therapy | TB disease events n/N (%) | Timing of diagnosis | Risk of infection RR (95% CI)# | Risk of disease RR (95% CI)# | Overall risk of bias |
Snider [10] | USA | 1975–1977, cohort (prospective), NR (study period 32 months) | DR | Recruited from CDC laboratory | 398 | NR | NR | NR | INH-resistant: 178 (44.5); SM-resistant: 136 (34.0); INH/SM-resistant: 86 (21.5) | Paediatric contacts: 627 (1.6) | TST ≥5mm (12 clinics), TST ≥10mm (3 clinics), unknown (2 clinics) | DR: 239/601 (39.8) | NR | DR: 4/601 (0.6 of total, 1.7 of infected) | NR | DR-non-MDR versus DS: 1.19 (1.03–1.36) | DR-non-MDR versus DS: 0.84 (0.24–2.94) | High-moderate (selection bias likely, comparability and outcome ascertainment likely) |
DS | Matched to study patients for age, race, sex and location | 398 | NR | NR | NR | Fully susceptible | 778 (2.0) | TST ≥5 mm (12 clinics), TST ≥10 mm (3 clinics), unknown (2 clinics) | DS: 252/751 (33.6) | NR | DS: 6/753 (0.8 of total, 2.4 of infected) | NR | ||||||
Barroso [4] | Brazil | 1990–1999, cohort (retrospective), 2 years | DR | Based on the results of DST at medical facilities | 126 | 39±25 | 37.3/62.7 | 78 of 126 tested, all results negative | MDR | 557 (4.4) | NR | NR | NR | MDR: 25/557 (4.5) | NR | NR | MDR versus DS: 0.84 (0.52–1.37) | Low-moderate (selection bias likely, comparability and outcome ascertainment likely) |
DS | Matched to study patients for sex, age, and year of first treatment | 176 | 41±14 | 37.5/62.5 | 97 of 176 tested, all results negative | Fully susceptible | 752 (4.3) | NR | NR | NR | DS: 41/752 (5.5) | NR | ||||||
Johnston [8] | Canada | 1990–2008, cohort (retrospective), 123 (IQR 19–239) months | DR | Recruited from national TB registry | 124 | NR | NR | NR | INH-mono- resistant (HMR): 96 | HMR: 249 (3.0) | Non-MDR (HMR)¶: 121/249 (49) | Non-MDR (HMR): 8/249 (3.0) | DR-non-MDR versus DS: 1.53 (1.34 – 1.75) | DR-non-MDR versus DS: 1.43 (0.71–2.87) | Low-moderate (selection bias likely, comparability and outcome ascertainment likely) | |||
INH/RMP-resistant: 28 (MDR) | MDR: 89 (3.0) | TST ≥5 mm (3 months to <1 year after source diagnosis) | MDR: 42/89 (47) | 12/89 treated | MDR: 5/89 (6.0), all susceptible | All diagnosed within 3 months of index patients | MDR versus DS: 1.49 (1.19–1.86) | MDR versus DS: 2.5 (1.05–5.93) | ||||||||||
DS | Recruited from national TB registry | 2895 | NR | Fully susceptible | 7309 (3.0) | TST ≥5 mm (3 months to <1 year after source diagnosis) | DS: 2321/7309 (32) | NR | DS: 168/7472 (2.0) | NR | ||||||||
Teixeira [5] | Brazil | 1994–1998, cohort (prospective), NR (study period 54 months) | DR | Recruited from TB referral centre | 26 | 39.5±12 | 23/77 | HIV+: 5 (20%) HIV-: 21 (80%) | INH/RMP-resistant: 6 (23); INH/RMP/PZA-resistant: 11 (43); INH/RMP/PZA/SM-resistant: 5 (19) INH/RMP/PZA/SM-resistant: 1 INH/RMP/PZA/EMB-resistant: 1 INH/RMP/SM/ PZA-resistant: 1 INH/RMP/PZA/ ethionamide: 1 | 157 (6.0) | TST ≥10 mm | MDR: 59/133 (44), no therapy | MDR 6/157 (4.0) | 3 a median 10 (range 2–34) months after initial evaluation, | MDR versus DS: 1.19 (0.92–1.54) | MDR versus DS: 0.7 (0.19–2.59) | Good (comparability and outcome ascertainment likely) | |
DS | Two DS index patients were matched to each new MDR-TB patient | 52 | 38.4±13 | 23/77 | HIV+: 5 (10%) HIV-: 47 (90%) | Fully susceptible | 251 (5.0) | TST ≥10 mm | DS: 86/231 (37) | NR | DS: 11/251 (4.0) | Median 10 (range 2–34) months after initial evaluation | ||||||
Palmero [7] | Argentina | 1998–2000, cohort (retrospective), 3 years | DR | Recruited from TB registry | 37 | 31.3±9.3 | 33/68 | HIV+: 21 (57%) HIV-: 16 (43%) | MDR | 97 (2.6) | TST ≥10 mm | MDR: 17/97 (17.5) | NR | MDR: 2/97 (2.1) | NR | MDR versus DS: 1.45 (0.87–2.43) | MDR versus DS: 0.92 (0.2–4.25) | Low-moderate (selection bias likely, comparability and outcome ascertainment likely) |
DS | Recruited from TB registry | 100 | 29.6±8.6 | 24/76 | HIV+: 38 (38%) HIV-: 62 (62%) | Fully susceptible | 356 (3.5) | TST ≥10 mm | DS: 43/356( 12.1) | NR | DS: 8/356 (2.2%) | NR | ||||||
Grandjean [6] | Peru | 2010–2013, cohort (prospective), DR 1425 person years (mean 494 days) | DR | Recruited at diagnosis from reference laboratories | 213 | 32 | 61/39 | HIV+: 18 (8%) HIV-: 195 (92%) | MDR | 1055 (4.0) | NR | NR | 12.5% | MDR: 35/1055 (3.3%), 28 DST performed of which 24 MDR and 4 DS | Day 1 of follow-up to day 600 of follow-up | NR | MDR versus DS: 0.71 (0.49–1.03) | Good (outcome ascertainment likely) |
2010–2013, cohort (prospective), DS 2620 person years (mean 406 days) | DS | Matched to study patients for age, race, sex and geographic location | 487 | 33 | 61/39 | HIV+: 20 (4%) HIV-: 467 (96%) | Fully susceptible | 2362 (4.0) | NR | NR | 17.2% | DS: 114/2441 (4.8%) | Day 1 of follow-up to day 600 of follow-up | |||||
Laniado-Laborin [9] | Mexico | 2011–2013, cross-sectional, no follow-up | DR | Recruited from TB clinic based on culture and DST performed at the clinic | 33 (20 MDR)¶ | NR | NR | MDR: 96/41 (4.0), paediatric contacts | TST ≥5 mm, IGRA ≥0.35 IU·m-1 | MDR TST-positive: 31/41 (75.6); MDR IGRA-positive: 24/41 (58) | Not treated | NR | MDR versus DS (TST): 0.91 (0.74–1.11) | NR | Low-moderate (selection bias likely, comparability and outcome ascertainment likely) | |||
DS | Recruited from TB clinic based on culture and DST performed at the clinic | 37 | NR | NR | 77 (2.3) | TST ≥5 mm, IGRA ≥0.35 IU·m-1 | DS TST-positive: 64/77 (83); DS IGRA-positive: 32/77 (42) | Treated with INH or RMP, unclear proportion | NR | |||||||||
Meta-analysis | Events in DR-non-MDR contacts versus DS contacts: 360/850 versus 2573/8060; events in MDR contacts versus DS contacts: 149/360 versus 2514/7973 | Events in DR-non-MDR contacts versus DS contacts: 12/850 versus 174/8225; events in MDR contacts versus DS contacts: 73/1931 versus 342/11279 | DR-non-MDR versus DS: 1.33 (1.2–1.46); MDR versus DS: 1.24 (1.08–1.42) | DR-non-MDR versus DS: 1.23 (0.67–2.27); MDR versus DS: 0.81 (0.64–1.06) |
DR: drug-resistant; DS: drug-susceptible; F: female; M: male; M. tuberculosis: Mycobacterium tuberculosis; TB: tuberculosis; NR: not recorded; CDC: Centers for Disease Control and Prevention; INH: isonicotinylhydrazide (isoniazid); SM: streptomycin; TST: tuberculin skin test; MDR: multidrug-resistant; DR-non-MDR: other resistance or not specified, not INH/ rifampicin (RMP); DST: drug-susceptibility testing; IQR: interquartile range; HMR: isoniazid mono-resistant; PZA: pyrazinamide; EMB: ethambutol; IGRA: interferon-γ release assay. #: fixed effects meta-analysis; ¶: additional information provided by authors, used in the meta-analysis.