Use this URL to cite or link to this record in EThOS:
Title: Air pollution and mortality : an investigation into the lag structure between exposure to air pollution, temperature and mortality from pneumonia, chronic obstructive pulmonary disease, & ischaemic heart disease
Author: Gittins, Matthew
ISNI:       0000 0004 7655 9783
Awarding Body: University of Manchester
Current Institution: University of Manchester
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
Introduction: The association between daily air pollution exposure and risk of mortality is well established. Few studies have investigated in detail the associations beyond a seven day lag. The aim of this thesis was to investigate the change in risk across longer (30 day) periods post exposure for three specific causes of death: pneumonia, chronic obstructive pulmonary disease (COPD), and ischaemic heart disease (IHD). Methods: Daily Scottish mortality data (1980-2011) was matched to measurements from local fixed site pollution (Black smoke, PM10, PM2.5, SO2, & NO2) and temperature monitors. Exposure on subjects' 'day of death' was compared with control days in a time-stratified case-crossover analysis. Exposure effects on 30 days prior to day of death were modelled using distributed lag non-linear, lag stratified, and cubic distributed lag models. Matching hospital admissions data inferred subject location during exposure, further analyses investigated extreme outliers and missing data using multiple imputation techniques. The analysis accounted for several confounders including accurately modelling temperature relationships unique for each cause of death. Results: Of the 919,301 deaths, 20% were classified as being caused by pneumonia, 9.5% as COPD, and 30% as IHD in the 'any' cause of death field. Non-linear effects for temperature and linear effects for the pollutants were present across all 30 days. Temperature-mortality was observed to be U-shaped at shorter lags. Consistently increased risk occurred for longer in cold temperatures with 1oC increase (30 days lag) = %RR -0.35% Pneumonia, -0.62% COPD, and -0.26% IHD. PM2.5 on all three outcomes, and all pollutants on COPD showed the greatest effect sizes. In general, COPD risk only occurred after a delay, peaking between 12-18 days. COPD risk due to PM2.5 was immediate (%RR (95% C.I.) = 1.05% (0.14%,2.01%)) and lasted the full 30 days. Pneumonia risk often reported the shortest lag of 10-15 days, whereas IHD risk occurred 2 days after exposure but lasted the remaining 30 days. There was some evidence especially for pneumonia of a smaller association between air pollution on mortality when subjects included were present in hospital. A simulation study indicated slight improvement in accuracy when 'multiple imputation' was performed compared to 'complete cases' analysis; though both techniques reported similarly underestimated effect estimates. Extreme outliers in the main analysis of pollution exposure did not appear to have a strong influence on the risk. However, large variability between monitor measurements of pollution exposure was present and appeared to be influencing the results. Conclusion: This study provides additional evidence on the link between air pollution, and temperature, and acute mortality. Particular focus was on three causes of death (pneumonia, COPD, and IHD) that are shown to be influenced by air pollution in subtly different ways. Results also indicated that the 'true' effect of air pollution on mortality might be greater than shown by mortality studies which do not use hospital admission location during exposure into account.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Distributed Lag ; Case-crossover design ; Epidemiology ; Ischaemic Heart Disease ; Lag stratified ; Pneumonia ; Acute mortality ; Air pollution ; Chronic Obstructive Pulmonary Disease