Intended for healthcare professionals


Heterogeneity of air pollution effects is related to average temperature

BMJ 1997; 315 doi: (Published 01 November 1997) Cite this as: BMJ 1997;315:1161
  1. Martin Bobak (martinb{at}, Clinical lecturer in epidemiologya,
  2. Anthony Roberts, Consultant neurologistb
  1. a Department of Epidemiology and Public Health, University College London, London WC1E 6BT
  2. b 74 Harley Street, London W1N 1AE

    Editor—The APHEA project (air pollution and health: a European approach) found that the effects of daily variation in air pollution on mortality were significantly stronger in western Europe than in eastern Europe.1 The authors have put forward several explanations for this inconsistency, all of which, generally speaking, suggest that the small effects found in eastern Europe are an artefact. We propose a less dismissive explanation.

    Inspired by the Eurowinter study,2 we plotted the relative risk of death in cities in the APHEA project1 against average winter temperatures (taken from a previous report3). Spearman correlation coefficients were 0.85 for sulphur dioxide (12 cities, P<0.001; 1) and 0.72 for black smoke (eight cities, P=0.045). Correlations with average annual temperatures were 0.77 and 0.61, respectively. This observation is the more convincing, because the relation is continuous: estimates from western European cities lie between those from southern and eastern Europe.


    We can only speculate about the mechanisms underlying this observation. There may be an interaction between air pollution and climate: an identical increase in pollution may not be equally harmful at different climatic conditions. In the APHEA project, effects of air pollution were stronger in the warm season.1 Ambient concentrations in warmer cities or seasons may also reflect more closely personal exposures.

    A residual confounding by climatic factors may also have a role. Daily temperatures are inversely correlated with daily concentrations of sulphur dioxide and particulates and with mortality. The slope of the association between low temperature and mortality is steeper in places with warmer winters, probably because of inadequate protection against cold in the latter.2 In consequence, any residual confounding by temperature would be stronger in warmer cities, leading to higher estimates of pollution effects. The statistical adjustment may not be sufficient to remove entirely all the effects of temperature or climate; other aspects (for example, daily minima or variations) may be important as well. This explanation is also biologically plausible. Effects of temperature on mortality are most pronounced for cardiovascular and respiratory causes, and more so among elderly people4—remarkably similar to the reported effects of air pollution.5

    The heterogeneity in air pollution effects in the APHEA project seems genuine and related to temperature. It is unlikely that all the short term effects of air pollution on mortality are spurious. Some of them, however, and part of the inconsistencies between studies, may be due to the role of temperature or climate. Existing data should provide the answer to this important question.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.