METHODS: This study analysed cross-sectional data from the Research with East London Adolescents Community Health Survey, a prospective population survey of adolescents attending state schools in East London, England. The history of NSSI was obtained using two items from the Lifestyle and Coping questionnaire (whether they have ever engaged with self-harm and the last time they engaged in such behaviours). The presence of TDIs, increased overjet and inadequate lip coverage were determined through clinical assessments by two trained dentists. Survey logistic regression was fitted to test the association of NSSI with TDIs. Odds ratios (ORs) were adjusted for socio-demographic and clinical characteristics as potential confounders.

RESULTS: The lifetime and last-year prevalence of NSSI were 11.9% and 6.7%, respectively, whereas the prevalence of TDIs was 16.5%. Neither the lifetime prevalence of NSSI (OR: 1.02, 95% confidence interval: 0.56-1.85) nor the last-year prevalence of NSSI (OR: 0.76, 95% CI: 0.36-1.61) were associated with TDIs in regression models adjusted for confounders.

METHODS: We did a sequential annual cross-sectional study of 2164 children aged 8-9 years attending primary schools between 2009-10 and 2013-14 in central London, UK, following the introduction of London's LEZ in February, 2008. We examined the association between modelled pollutant exposures of nitrogen oxides (including nitrogen dioxide [NO2]) and particulate matter with a diameter of less than 2·5 μm (PM2·5) and less than 10 μm (PM10) and lung function: postbronchodilator forced expiratory volume in 1 s (FEV1, primary outcome), forced vital capacity (FVC), and respiratory or allergic symptoms. We assigned annual exposures by each child's home and school address, as well as spatially resolved estimates for the 3 h (0600-0900 h), 24 h, and 7 days before each child's assessment, to isolate long-term from short-term effects.

FINDINGS: The percentage of children living at addresses exceeding the EU limit value for annual NO2 (40 μg/m3) fell from 99% (444/450) in 2009 to 34% (150/441) in 2013. Over this period, we identified a reduction in NO2 at both roadside (median -1·35 μg/m3 per year; 95% CI -2·09 to -0·61; p=0·0004) and background locations (-0·97; -1·56 to -0·38; p=0·0013), but not for PM10. The effect on PM2·5 was equivocal. We found no association between postbronchodilator FEV1 and annual residential pollutant attributions. By contrast, FVC was inversely correlated with annual NO2 (-0·0023 L/μg per m3; -0·0044 to -0·0002; p=0·033) and PM10 (-0·0090 L/μg per m3; -0·0175 to -0·0005; p=0·038).

INTERPRETATION: Within London's LEZ, a smaller lung volume in children was associated with higher annual air pollutant exposures. We found no evidence of a reduction in the proportion of children with small lungs over this period, despite small improvements in air quality in highly polluted urban areas during the implementation of London's LEZ. Interventions that deliver larger reductions in emissions might yield improvements in children's health.