HEALTH-EFFECTS OF AN AIR-POLLUTION EPISODE IN LONDON, DECEMBER 1991

Background - In December 1991 London experienced a unique air pollution episode during which concentrations of nitrogen dioxide rose to record levels, associated with moderate increases in black smoke. The aim of this study was to investigate whether this episode was associated with adverse health effects and whether any such effects could be attributed to air pollution. Methods - The numbers of deaths and hospital admissions occurring in Greater London during the week of the episode were compared with those predicted using data from the week before the episode and from equivalent periods from the previous four years. Relative risks (RR) (episode week versus predicted) for adverse health events were estimated using log linear modelling and these were compared with estimates from control areas which had similar cold weather but without increased air pollution. Results - In all age groups mortality was increased for all causes (excluding accidents) (relative risk = 1.10) and cardiovascular diseases (1.14); non-significant increases were observed for all respiratory diseases (1.22), obstructive lung diseases (1.23), and respiratory infections (1.23). In the elderly (65 + years) the relative risk of hospital admission was increased for all respiratory diseases (1.19) and for obstructive lung diseases (1.43), and a nonsignificant increase was observed for ischaemic heart disease (1.04). In children (0-14 years) there was no increase in admissions for all respiratory diseases and only a small non-significant increase for asthma. When compared with control areas the relative risks became non-significant but remained increased. Conclusions - The air pollution episode was associated with an increase in mortality and morbidity which was unlikely to be explained by the prevailing weather, a coincidental respiratory epidemic, or psychological factors due to publicity. Air pollution is a plausible explanation but the relative roles of nitrogen dioxide and particulates cannot be distinguished.