The authors' purpose in this study was to determine whether changes in weather conditions were associated with daily mortality among people aged 65 years and older diagnosed as having congestive heart failure in Montreal, Canada, and who died in the urban area between 1984 and 1993. The authors used a time-stratified case-crossover design and adjusted the models for nitrogen dioxide and ozone. They found a strong nonlinear association with maximum temperature in the warmer months of the year, with a threshold at about 25°C. The authors observed no associations after lag 3 days. In the cold period, they found that risks increased linearly with increasingly colder temperatures, but only after lag 2 days. The authors found no associations with relative humidity. For change in barometric pressure from the previous day, they found no associations in the cold period, but an increase in pressure from the previous day increased risk for lags 0 or 1 days. The authors found some differences between men and women. Copyright © 2008 Heldref Publications.