BACKGROUND: Few studies have identified specific factors that increase mortality during heat waves. This study investigated socio-demographic characteristics and pre-existing medical conditions as effect modifiers of the risk of dying during heat waves in a cohort of elderly residents in Rome. METHODS: A cohort of 651,195 residents aged 65 yrs or older was followed from 2005 to 2007. During summer, heat wave days were defined according to month-specific thresholds of maximum apparent temperature. The adjusted relative risk of dying during heat waves was estimated using a Poisson regression model including all the considered covariates. Risk differences were also calculated. All analyses were run separately for the 65-74 and 75+ age groups. RESULTS: In the 65-74 age group the risk of dying during heat waves was higher among unmarried subjects and those with a previous hospitalization for chronic pulmonary disease or psychiatric disorders. In the 75+ age group, women, and unmarried subjects were more susceptible to heat. Furthermore, a higher susceptibility to heat among those with previous hospitalization for diabetes, diseases of the central nervous system (CNS), psychiatric disorders and cerebrovascular diseases resulted from risk differences. DISCUSSION: Results showed a higher susceptibility to heat among those older than seventy-five years, females and unmarried. Pre-existing health conditions play a different role among the two considered age groups. Moreover, compared with previous studies the pattern of susceptibility factors have slightly changed over time. For the purposes of public health programmes, susceptibility should be considered as time, space and population specific.