Wilkinson, P; Kovats, S; Pattenden, S; Armstrong, B; Mangtani, P; Fletcher, A; McMichael, T

Objective: To examine the determinants of vulnerability to winter/cold-related death in the elderly British population in the context of the large, but inadequately explained, winter excess of deaths in the UK. DESIGN Cohort study of people aged 75+ years recruited between 1995 and 1998.

Setting: One hundred and six General Practices from the Medical Research Council Trial of Assessment and Management of Older People in Great Britain. METHODS The extent to which the winter (December to March) excess of deaths was modified by individual characteristics was examined by Poisson models with interaction terms for the winter:non-winter ratio.

Results: There were 4221 deaths in 42,162 person-years of follow-up in winter months (rate = 100.1 deaths per 1000 person-years (95% Cl 97.1 to 103.1)) and 5902 in 77,227 person years of follow-up in other months of the year (rate = 76.4 deaths per 1000 person-years (95% CI 74.5 to 78.4)). The overall winter: non-winter rate ratio was 1.31 (95% CI 1.26 to 1.36). Month to month variation accounted for 17% of annual all-cause mortality, but only 7.8% after adjustment for temperature, and 5.2% after further adjustment for influenza A. There was little evidence that the winter:non-winter ratio of deaths varied by geographical region or age within this elderly population. The winter:non-winter ratio was 1.11 (95% CI 1.00-1.23) times greater in women than men after adjustment for region, age-group, indicators of clinical status and socio-economic position. However, there was no evidence that socio-economic deprivation or difficulty in making ends meet were predictive of winter death, but a self-reported history of respiratory illness was: those with such a history had a winter:non-winter ratio 1.20 (95% CI 1.08-1.34) times that of people without a history of respiratory illness.

Conclusions: The lack of socio-economic gradient in vulnerability to winter death is consistent with the findings of other studies of the UK population, but appears counter-intuitive. Female gender and a history of pre-existing respiratory illness were the only individual factors for which we found clear evidence of vulnerability to winter death in this 75+ population. Although we cannot exclude modest increases in risk associated with the other medical, behavioural and social factors, the patterns and determinants of winter death appear complex and require further elucidation.

(C) 2003 Lippincott Williams & Wilkins, Inc.