Prevalence, incidence, and associated risk factors of tuberculosis in children with HIV living in the UK and Ireland (CHIPS): a cohort study.

Background: Tuberculosis is the most common serious co-infection in people living with HIV worldwide, but little is known about its incidence in HIV-infected children living in high-resource settings with low tuberculosis prevalence. We aimed to assess the incidence and prevalence of tuberculosis in children with HIV living in the UK and Ireland to understand rates, risk factors, and outcomes of the disease in this group.

Methods: We did an analysis of children enrolled in CHIPS, an observational multicentre cohort of children receiving HIV care in the UK and Ireland. We assessed characteristics and prevalence of tuberculosis at baseline, measured incidence of disease through the follow-up period using the CHIPS database, and calculated associated risk factors in these children with multivariable logistic and Cox regression models.

Findings: Between Jan 1, 1996, to Sept 18, 2014, data for 1848 children with 14 761 years of follow-up were reported to CHIPS. 57 (3%) children were diagnosed with tuberculosis: 29 children had tuberculosis at presentation (prevalent tuberculosis) and 29 had the disease diagnosed during follow-up (incident tuberculosis), including one child with recurrent tuberculosis events. Median age at diagnosis was 9 years (IQR 5-12). 25 (43%) children had pulmonary tuberculosis, 24 (41%) had extrapulmonary tuberculosis with or without pulmonary involvement, and the remainder (n=9; 16%) had unspecified-site tuberculosis. The overall incidence rate for the follow-up period was 196 cases per 100 000 person-years (95% CI 137-283). In our multivariable model, tuberculosis at presentation was associated with more severe WHO immunological stage at baseline (odds ratio 0·25, 95% CI 0·08-0·74; p=0·0331; for none vs severe) and being born abroad (odds ratio 0·28, 0·10-0·73; p=0·0036; for UK and Ireland vs abroad). Incident tuberculosis was associated with time-updated more severe WHO immunological stage (hazard ratio 0·15, 95% CI 0·06-0·41; p=0·0056; for none vs severe) and older age at baseline (1·11, 0·47-2·63; p=0·0027; for age >10 years vs 5-9 years).

Interpretation: Tuberculosis rates in HIV-infected children in the UK and Ireland were higher than those reported in the general paediatric population. Further study is warranted of tuberculosis screening and preventive treatment for children at high-risk of this disease to avoid morbidity and mortality in this population.

Funding: NHS England, PENTA Foundation.