Menu
Menu
Menu
Menu
Background Tuberculosis is a leading cause of global childhood mortality. However, the epidemiology and burden of tuberculosis in infancy is not well understood. We aimed to investigate tuberculin skin test conversion and tuberculosis in the Drakenstein Child Health study, a South African birth cohort in a community in which tuberculosis incidence is hyperendemic. Methods In this prospective birth cohort study, we enrolled pregnant women older than 18 years who were between 20 and 28 weeks’ gestation and who were attending antenatal care in a peri-urban, impoverished South African setting. We followed up their children for tuberculosis from birth until April 1, 2017, or age 5 years. All children received BCG vaccination at birth. Tuberculin skin tests were administered to children at 6, 12, 24, 36, 48, and 60 months of age, and at the time of a lower respiratory tract infection. An induration reaction of 10 mm or more was considered to be a tuberculin skin test conversion. To prevent boosting, we censored children with a reactive, negative tuberculin skin test. Findings Among 915 mother–child pairs (201 [22%] HIV-positive mothers and two [<1%] HIV-positive children), 147 (16%) children had tuberculin skin test conversion, with increasing cumulative hazard with age (0·08 at 6 months, 0·17 at 12 months, 0·22 at 24 months, and 0·37 at age 36 months). For every 100 child-years, the incidence was 11·8 (95% CI 10·0–13·8) for tuberculin skin test conversion, 2·9 (2·4–3·7) for all diagnosed tuberculosis, and 0·7 (0·4–1·0) for microbiologically confirmed tuberculosis. Isoniazid preventive therapy was effective in averting disease progression (adjusted hazard ratio 0·22, 95% CI 0·08–0·63; p<0·0001). Children with a lower respiratory tract infection were significantly more likely to also have tuberculosis than were those without one (2·27, 1·42–3·62; p<0·0001). Interpretation Greater focus should be placed on the first years of life as a period of high burden of transmission and clinical expression of tuberculosis infection and disease. Multifaceted interventions, such as isoniazid preventive therapy and tuberculosis screening of infants with LRTIs, beginning early in life, are needed in high-burden settings. Funding Bill & Melinda Gates Foundation, Medical Research Council South Africa, and National Research Foundation South Africa.
In this prospective birth cohort study, we enrolled pregnant women who were between 20 and 28 weeks' gestation and attending antenatal care in Paarl, a peri-urban setting outside of Cape Town, South Africa.14 In 2015, tuberculosis incidence in this area was estimated to be 880 new cases per 100 000 population.13 Participants were recruited from two clinics, TC Newman and Mbekweni, which are a few miles apart. Both clinics serve impoverished, heterogeneous communities. People attending TC Newman are of admixed ancestry, whereas the Mbekweni clinic serves mostly a black, Xhosa population. All infants were given BCG vaccination at birth, per national policy. All mothers accessed care in the public sector, which has a strong primary health-care programme, including an effective mother-to-child HIV prevention and antiretroviral therapy programme. Women were followed up through pregnancy and childbirth, and newborn infants were followed up into early childhood, up to age 5 years. Exclusion criteria for pregnant women were being younger than 18 years and intending to leave the area within 1 year. We obtained ethics approval from the University of Cape Town Faculty of Health Sciences Human Research Ethics Committee (reference numbers 401/2009 and 651/2013) and the Provincial Child Health Research Committee. Mothers provided written informed consent at enrolment and verbal assent for infants, which was renewed annually. Comprehensive questionnaires about maternal health were administered at enrolment and antenatal data were collected concurrently. Detailed birth information was obtained at delivery. Obstetric care and all births took place at the regional hospital in Paarl. Follow-up visits, including clinical examinations, were done at 6, 12, 24, 36, 48, and 60 months of age. Data for environmental exposures, household characteristics, respiratory risk factors, anthropometry, and child symptoms were obtained at scheduled visits. Missed visits were rebooked with a study mobile phone network system or by study community-based fieldworkers. Mothers were counselled about respiratory symptoms at every visit and advised to attend the study site or contact study staff between scheduled study visits whenever the child developed cough or difficulty breathing. Socioeconomic status comprised a comprehensive composite of asset ownership, household income, employment, and education.14 HIV tests were given to all mothers during pregnancy. Adults were tested with Abbott Determine HIV 1/2 rapid HIV antibody test (Abbott Laboratories, North Chicago, IL, USA). If positive, a confirmatory enzyme-linked immunosorbent assay was done. Infants of HIV-positive mothers were tested with DNA PCR (Cobas Ampliprep system, Roche Molecular Systems, Branchburg, NJ, USA) at age 6 weeks, and 6 weeks after the end of breastfeeding. Children were re-tested at 18 months with the rapid antibody test. As part of the Drakenstein Child Health Study, we established surveillance systems for the detection of lower respiratory tract infections in infants; children with such infections were seen and had specimens (induced sputum, nasopharyngeal samples, blood, urine [all participants], and blood culture [hospitalised infants with lower respiratory tract infections]) taken by trained study nurses and staff.14, 15 Briefly, study nurses were trained to diagnose lower respiratory tract infection or severe lower respiratory tract infection according to WHO clinical case definitions.16 A lower respiratory tract infection was diagnosed in children with cough or difficulty breathing and age-specific tachypnoea, or if the child had lower-chest-wall indrawing. Severe lower respiratory tract infections were diagnosed in children younger than 2 months with tachypnoea or lower-chest-wall indrawing, or in children of any age if the child had a general danger sign. We derived Z scores from WHO child growth standards at birth and at every follow-up visit, and we used the median of all the weight-for-age Z scores for each child to summarise nutrition status during the follow-up period. Children were considered to be severely underweight or stunted if weight-for-age or length-for-age Z scores were less than −2. Normal weight was −2 to 2, and overweight was a score greater than 2. Tuberculin skin tests were done at the 6-month visit and then at 12, 24, 36, 48, and 60 months of age, and at the time of a lower respiratory tract infection. Tuberculin skin test conversion was defined as an induration reaction greater than or equal to 10 mm, to minimise the risk of misclassification due to BCG vaccination or exposure to environmental mycobacteria, as recommended by WHO and South Africa's Department of Health.10, 17 To prevent misinterpretation of boosted skin-test reactions due to recurrent tuberculin skin testing as tuberculosis infection, children with a reactive but negative skin test (1–9 mm) were not given another test, and were censored for the tuberculin skin test conversion analysis at that point in time. Because of the high number of censored skin tests before age 48 months, we excluded tuberculin skin tests taken after 36 months of age. Children with positive skin tests were referred to local tuberculosis clinics for isoniazid preventive therapy; however, the study investigators could not enforce that this was prescribed. Children were followed up for tuberculosis from birth until April 1, 2017, or age 5 years. Trained study staff collected induced sputum specimens in duplicate for tuberculosis culture and mycobacterial PCR investigation (Xpert MTB/RIF; Cepheid, Sunnyvale, CA, USA) from all children with a tuberculin skin test induration of at least 10 mm, and from children who were suspected to have or had been diagnosed with tuberculosis by local health services. A chest radiograph was taken in all children with suspected pulmonary tuberculosis. Tuberculosis was diagnosed by experienced physicians and nurses in local tuberculosis community clinics, and chest radiographs were read and reported by an experienced tuberculosis clinician. We compared results using three different definitions of tuberculosis: all tuberculosis cases (clinically, radiographically suggestive, or microbiologically confirmed cases); cases that were microbiologically confirmed or radiographically suggestive only; and microbiologically confirmed cases only (positive Xpert MTB/RIF or sputum culture). Mother–child pairs were included in this analysis if they had at least one tuberculin skin test. For exploratory data analysis, we summarised continuous variables as medians with IQRs, and categorical variables using proportions. For tuberculin skin test conversion, time-to-event was determined by the date on which the child was administered a skin test and had a positive result; a child was determined to have no conversion on the date of the last negative skin test. Follow-up was censored at a reactive skin test, death, or age 3 years. For tuberculosis, time-to-event was determined when a child was diagnosed with tuberculosis. Follow-up was censored at death, 5 years of age, or April 1, 2017. Children with reactive but negative tuberculin skin tests were not censored for the tuberculosis survival analysis. We analysed factors associated with tuberculin skin test conversion and tuberculosis for the whole cohort and adjusted for the enrolment site for each analysis. Because the first tuberculin skin test was administered at different ages (because of missed study visits or lower respiratory tract infections), we also adjusted for the age of first tuberculin skin test administration in the multivariable model of skin test conversion. We used Cox proportional hazard models for multivariable modelling, including a random intercept for each child using a gamma distribution, and results are presented as hazard ratios (HRs). For each independent variable, we present univariable models (indirect effects) and multivariable models (direct effects, after adjustment for the effects of confounding variables). We explored the effect of the censoring approach used in our tuberculin skin test conversion analysis by adjusting our definition of conversion to infants with induration reactions more than 5 mm (rather than the 10 mm cutoff used in our primary conversion analysis). In this analysis, all children with reactive but negative tuberculin induration reactions less than 5 mm were censored. We investigated the timing of lower respiratory tract infections and tuberculosis diagnoses to determine whether case ascertainment bias could possibly explain an association seen between the two diseases. We assumed case ascertainment was likely if diagnosis of both diseases occurred simultaneously in children with multiple diagnoses. To investigate this, among children diagnosed with both diseases during the study period, we calculated the proportion diagnosed with a lower respiratory tract infection and tuberculosis less than 2 weeks apart. We used two-sided p values and 95% CIs to assess statistical significance in all models. The likelihood ratio test was used to derive all p values from Cox regression models. We did all analyses with Stata (version 14.1). The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.
