Table 2

HIV cohorts

First author, yearCountryYears of studyStudy descriptionNumber of TB casesNumber in cohortPatient-yearsTB incidence, per cent per year (95% CI)ART at enrolment (%)Number maleAge rangeQuality, TB incidence (selection/outcomes)*Quality, immunosuppression (selection/comparability/outcomes)*Quality, CD4% (selection/comparability/outcomes)*Quality, time-on-ART (selection/comparability/outcomes)*Quality, ART protection (selection/comparability/outcomes)*
Abuogi, 201360Kenya2009–2010Prospective cohort study of HIV-infected children (6 weeks to 14 years, some on ART and some not) followed-up for the incidence of clinically diagnosed TB. Study at two urban clinics with children recruited over a 2-year period106867201.39 (0.75 to 2.58)743220–14B/AB/C/Ananana
Alarcón, 201261Brazil2002–2007Prospective cohort study of HIV-infected children (<21 years) at 17 sites in Latin America; children recruited over a 5-year period and followed for the incidence of a number of opportunistic infections, including TB773125230.28 (0.07 to 0.48)673250–22B/Bnananana
Auld, 201462Côte d'Ivoire2004–2008A nationally representative retrospective cohort study of children (<15 years) starting ART over a 5-year period at 29 facilities56196041901.34 (1.03 to 1.74)10010580–14C/CnaC/A/CB/C/Cna
Bakeera-Kitaka, 201163Uganda2003–2006Retrospective cohort study of children and adolescents starting ART over a 3½-year period at one specialist clinic3111806169018.4 (17.1 to 20.3)100932nsC/CnanaB/C/CB/A/C
Braitstein, 200964Kenya2001–2007Retrospective observational study of all children (<13 years) enrolled in a network of HIV clinics over a 5-year period7656301436817.6 (16.3 to 18.8)1931440–13C/CnanaB/C/CB/A/C
Brennan, 201465South Africa2004–2011Retrospective cohort of all children (<19 years) on ART enrolled over a 7-year period at 12 urban clinics113332928284.0 (3.3 to 4.8)10016080–18C/CnaC/A/CB/A/Cna
Ciaranello, 201466East Africa2002–2008All HIV-infected infants (<12 months) enrolled over a 6-year period at 7 sites across 3 countries12884773817.4 (14.5 to 20.6)0418<1C/CnaC/C/Cnana
Crook, 201614Uganda, Zimbabwe2007–2012Randomised controlled trial assessing different monitoring strategies for providing ART. All HIV-infected children (3 months to 17 years) enrolled from 4 centres in 2 countries over an 18-month period6996936321.9 (1.5 to 2.4)1004880–17B/BnaB/A/BA/C/Bna
Curtis, 201267Eight countries2002–2010Data from 25 HIV treatment centres in 8 countries collected prospectively over an 8-year period. Adults and children included if starting ART at one of the centres. Data from children (<15 years) disaggregated from adults140394616878.3 (8.1 to 8.6)100ns0–15C/CnanaB/C/Cna
Dankne, 200168USA1988–1998Data from 13 studies over a period of 10 years with children (<21 years) included. From pre-ART period. Wide range of opportunistic infections assessed27333167500.4 (0.3 to 0.6)018520–21C/CnaC/C/Cnana
De Beaudrap, 201369Burkina Faso/Côte d'Ivoire2006–2007/2000–2004Pooled data from a clinical trial (recruited over 1 year) and observational cohort (recruited over 4 years). Children (15 months to 15 years) enrolled at initiation of ART and followed for 2 years101883552.9 (1.4 to 5.1)10066nsC/AnaC/C/AB/C/Ana
Edmonds, 200970Democratic Republic of Congo2004–2008Prospective study of children (<18 years) enrolled from one hospital clinic over a 3½-year period. All children ART-naïve at baseline with some starting ART during follow-up8136459613.6 (10.8 to 16.9)01720–18B/BB/C/BnaA/C/BA/A/B
Gray, 201471South Africa2005–2011Control (placebo) arm of a randomised controlled trial of isoniazid preventive therapy. Children (over 8 weeks) recruited over a 4-year period from three sites7822402.9 (1.2 to 6.0)10042nsA/Bnananana
Kouakoussui, 200472Côte d'Ivoire2000–2003All children recruited from one clinic and followed-up for incident TB. All children in the study initiated ART at some point in the study to provide pre-ART and post-ART TB incidences72822063.4 (1.5 to 6.7)651531–16C/BnaC/C/Cnana
Li, 201373Tanzania2004–2011Prospective study of all children (<15 years) from 28 clinics recruited over a 7-year period. ART started based on local guidelines over follow-up376504072215.2 (4.7 to 5.8)5124600–14B/CB/A/CnaA/A/CA/A/C
Martinson, 200974South Africa1994–2006Retrospective cohort of children (<15 years) recruited from four clinics over a 12-year period. Incidence determined for the period pre-ART and post-ART2811132171312.3 (9.8 to 12.6)25630–14B/CnananaA/A/C
Prasitsuebsai, 201475SE Asia1993–2009Retrospective and prospective cohort of HIV-infected children (≤18 years) from 14 clinics in 5 countries in SE Asia. Recruited over a 16-year period. Wide range of OIs assessed. Period covered before ART, during monotherapy or dual-therapy and on HAART4772280182621.5 (19.4 to 23.7)711360–18C/Cnananana
Thomas, 200076USA (New York City)1989–1995Retrospective study in HIV-exposed and HIV-infected children enrolled in a longitudinal study. TB cases determined by review of clinical records and by linkage to local TB registry. Registry data assessed for a 6-year period45142673840.61 (0.44 to 0.82)04750–12B/Bnananana
Walters, 200877South Africa2003–2005Chart review of all children (<13 years) starting HAART at one hospital over a 3-year period and followed-up until the end of the study period for incident TB. Cases defined as pre-HAART (if in the 9 months prior to HAART initiation) or post-HAART1029054625.1 (21.1 to 29.6)01420–14C/Anananana
Walters, 201478South Africa2003–2010Retrospective cohort of all children (<2 years) starting HAART at one hospital over an 8-year period. Children divided into those developing TB-IRIS (developing TB with 3 months of starting HAART) and those with incident TB (developing TB after 3 months)5534185525.3 (22.0 to 28.9)0161<2B/CnanaA/C/Cna
Yirdaw, 201479Ethiopia2007–2010Retrospective cohort study of the electronic records of all patients (adults and children) starting ART over a 3-year period in five hospitals in one region of Ethiopia. Patients followed until the end of the study period and the interventions of ART and IPT evaluated. Incidence stratified by age enabling extraction children (<15 years)2347510992.1 (1.3 to 3.1)nsns0–15B/CnananaA/C/C
Zar, 200780South Africa2003–2004Double-blind randomised controlled trial evaluating the impact of IPT on TB incidence in HIV-infected children (aged older than 8 weeks). Children recruited from two sites with the trial stopped after 17 months. Data on TB incidence taken from placebo arm131315623.4 (12.4 to 50.0)8740–14C/Cnananana
  • *Adapted Newcastle-Ottawa tool for cohort studies: A=high quality; B=moderate quality; C=low quality; see online supplementary material pp. 19–21, 26–27.

  • ART, antiretroviral therapy; HAART, highly-active antiretroviral therapy; IPT, isoniazid preventive therapy; IRIS, immune reconstitution inflammatory syndrome; OIs, opportunistic infections; na, not applicable; ns, not stated.