Early-life exposure to indoor air pollution or tobacco smoke and lower respiratory tract illness and wheezing in African infants: a longitudinal birth cohort study

Background Indoor air pollution (IAP) and environmental tobacco smoke (ETS) are associated with lower respiratory tract illness (LRTI) or wheezing in children. However, the effect of the timing of these exposures, specifically antenatal versus postnatal, and of alternate fuel sources such as the increasingly used volatile organic compounds have not been well studied. We longitudinally investigated the effect of antenatal or postnatal IAP and ETS on LRTI or wheezing prevalence and severity in African infants. Methods Mother and infant pairs enrolled over a 3-year period in a birth cohort study in two centres in Paarl, South Africa, were followed for the first year of life for LRTI or wheezing illness. We measured exposure to IAP (particulate matter, nitrogen dioxide, sulphur dioxide, carbon monoxide, and volatile organic compounds benzene and toluene) using devices placed in homes, antenatally and postnatally. We measured ETS longitudinally by maternal self-report and by urine cotinine measures. Study staff trained in recognition of LRTI or wheeze documented all episodes, which were categorised according to WHO case definition criteria. We used multivariate logistic and Poisson regressions to explore associations. Findings Between March 1, 2012, and March 31, 2015, we enrolled 1137 mothers with 1143 livebirths. Of 1065 infants who attended at least one study visit, 524 episodes of LRTI occurred after discharge with a wheezing prevalence of 0·23 (95% CI 0·21–0·26) episodes per child year. Exposures associated with LRTI were antenatal maternal smoking (incidence rate ratio 1·62, 95% CI 1·14–2·30; p=0·004) or particulate matter (1·43, 1·06–1·95; p=0·008). Subanalyses of LRTI requiring hospitalisation (n=137) and supplemental oxygen (n=69) found antenatal toluene significantly increased the risk of LRTI-associated hospitalisation (odds ratio 5·13, 95% CI 1·43–18·36; p=0·012) and need for supplemental oxygen (13·21, 1·96–89·16; p=0·008). Wheezing illness was associated with both antenatal (incidence rate ratio 2·09, 95% CI 1·54–2·84; p<0·0001) and postnatal (1·27, 95% CI 1·03–1·56; p=0·024) maternal smoking. Antenatally, wheezing was associated with maternal passive smoke exposure (1·70, 1·25–2·31; p=0·001) and, postnatally, with any household member smoking (1·55, 1·17 −2·06; p=0·002). Interpretation Antenatal exposures were the predominant risk factors associated with LRTI or wheezing illness. Toluene was a novel exposure associated with severe LRTI. Urgent and effective interventions focusing on antenatal environmental factors are required, including smoking cessation programmes targeting women of childbearing age pre-conception and pregnant women. Funding Bill & Melinda Gates Foundation, Discovery Foundation, South African Thoracic Society AstraZeneca Respiratory Fellowship, Medical Research Council South Africa, National Research Foundation South Africa, and CIDRI Clinical Fellowship. We did a longitudinal study of children enrolled in the Drakenstein Child Health Study (DCHS),16 a birth cohort study in a peri-urban area of South Africa that included follow-up through the first year of life. Consecutive consenting pregnant women were enrolled at 20–28 weeks' gestation at two public primary health clinics serving different populations: Mbekweni (serving a predominantly black African population) and Newman (serving a predominantly mixed-race population)16 from March 1, 2012, to March 31, 2015. We chose a 3-year period for the DCHS study so as to ensure constant enrolment over different seasons and time periods, with more than 90% of the DCHS population attending the public health service (appendix p 2).16 We excluded participants who were younger than 18 years, who did not attend study clinics for postnatal care (and thus could not be readily followed up), or who were intending to move out of the district within 2 years after the infant's birth.16 All children were born at Paarl Hospital (Paarl, South Africa). Mother and infant pairs were followed at 6–10 weeks, 14 weeks, and 6, 9, and 12 months after birth. Study questionnaires and clinical data were collected at enrolment and at each follow-up visit. We applied a composite socioeconomic status score to each participant and categorised them into quartiles as lowest, low-to-moderate, moderate-to-high, or highest socioeconomic status (appendix p 2).12, 16, 17 The study was approved by the Faculty of Health Sciences Human Research Ethics Committees of the University of Cape Town and of Stellenbosch University, and by the Western Cape Provincial Health Research committee. An antenatal (within 4 weeks of enrolment) and postnatal (between 4 and 6 months of the infant's life) home visit was undertaken to assess the home environment and measure IAP. Dwellings were categorised12 and the most common pollutants and by-products of combustion measured. Particulate matter of diameter 10 μm or less (PM10) was measured using a personal air sampling pump (AirChek 52; SKC, Eighty Four, PA, USA) and carbon monoxide with an Altair (Troy, MI, USA) carbon monoxide single gas detection unit, left in homes for 24 h. Diffusion tubes placed in homes for 2 weeks measured nitrogen dioxide, sulphur dioxide (Radiello absorbent filters in polyethylene diffusive body; Sigma-Aldrich, St Louis, MO, USA), and the volatile organic compounds benzene and toluene (Markes thermal desorption tubes; Llantrisant, UK). As described previously,12 an average concentration based on the 2-week duration in the home was obtained for nitrogen dioxide, sulphur dioxide, and volatile organic compounds; 24-h averages were obtained for PM10. Carbon monoxide data were downloaded to a computer and the frequency of exceedance above the hourly ambient standard was calculated (appendix p 2).12 The South African National Ambient Air Quality Standards18 were used to define expected exposure levels for each pollutant based on an averaging period of 1 year for each measure: PM10 40 μg/m3; nitrogen dioxide 40 μg/m3; benzene 5 μg/m3; toluene 240 μg/m3; and carbon monoxide 30 mg/m3 (based on an averaging period of 1 h; no more than 88 h of exceedence per year; appendix p 2).18 During the postnatal home visit, these same measurements were repeated. To measure exposure to ETS, questionnaires of maternal and paternal smoking and household exposure to tobacco smoke were administered at enrolment, at the antenatal visit, and at each follow-up visit during the postnatal follow-up period.19 Maternal exposure to ETS was also measured using urine cotinine at the second antenatal visit (28–32 weeks' gestation) and at birth, with the highest result used to assign the mother's smoking status (appendix p 2).19 Urine cotinine levels were classified as less than 10 ng/mL (non-smoker), 10–499 ng/mL, (passive smoker or exposed), or 500 ng/mL or more (active smoker).19 We categorised respiratory disease as an episode of LRTI or wheeze. Study staff trained in the recognition of LRTI or wheezing illness documented all episodes, either ambulatory or hospitalised. We defined LRTI and severe LRTI using WHO case definition criteria (appendix p 2).13, 20 Active surveillance for LRTI in the cohort was established (appendix p 2).13 LRTI which occurred at or shortly after birth prior to discharge was defined seperately. Episodes of wheeze were self-reported by a caregiver at a study visit or diagnosed on auscultation by trained study staff at a study visit or intercurrent illness. Study staff were trained in the recognition and auscultation of wheezing; caregivers were also trained in clinical recognition (appendix p 2). Recurrent wheezing was defined as two or more episodes of wheezing. We used simple descriptive statistics to characterise the study population, summarising continuous data as median (IQR) and categorical data as proportions (95% CI). We used Wilcoxon rank-sum test to compare medians and the χ2 test to compare proportions. We used mixed-effects Poisson regression clustered around the infant for multivariate analysis of LRTI incidence and multivariable Poisson regression for wheezing; results are presented as incidence rate ratios (IRRs) and 95% CIs. We used univariate mixed effects logistic regression clustered around the infant to explore associations between demographic, household, and socioeconomic characteristics, indoor air pollutants, and smoke exposure between severe versus non-severe LRTI, hospitalised versus ambulatory, LRTI requiring oxygen versus not requiring oxygen, and wheeze at LRTI versus no wheeze in the subset of infants that had an LRTI; results are presented as odds ratios and 95% CIs. Univariate analysis tested the association between environmental and socioeconomic factors and respiratory disease. (appendix p 2). We included variables that were associated with these outcomes and those of clinical relevance in multivariate (mixed effects) logistic regression models to determine the effect of severity of disease. We used the Wilcoxon signed-rank test to compare differences in the median pollutants measured antenatally to postnatally. We included confounding variables (birthweight, sex, ethnicity [site], socioeconomic status, weight-for-age Z score [WAZ],21 maternal HIV status, crowding, household characteristics, fossil fuel usage, vaccination status, nutritional status, and feeding in the first 6 months status) that showed an effect in the final analysis models (appendix p 2). All statistical tests were two-sided at α=0·05. We used STATA (version 13.0) for all data analysis. The sponsors of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to all the data and had final responsibility for the decision to submit for publication.