To identify and to measure the association between socio-demographic factors and the prevalence of adverse drug reaction at hospital admission in an elderly population, we carried out a cross-sectional study on a sample of 19,070 patients of the Gruppo Italiano di Farmacoepidemiologia nell'Anziano data base (Italy). The following socio-demographic variables were considered as independent variables: gender, age, education, living arrangement, widowhood/widowerhood, and alcohol and tobacco habit. We estimated the association of these variables with diagnosis of any adverse drug reaction at hospital admission. We used a logistic regression model to obtain estimates of this association, taking into account hepatic and renal diseases, number of diseases, number of drugs used, albumin index (indicator of nutritional condition) and cognitive function. Out of the 19,070 patients included in our study, 878 (4.3%) were diagnosed with an adverse drug reaction at admission. The multivariate analysis shows that none of the socio-demographic variables and health habits considered are associated with adverse drug reactions at admission. A worse nutritional condition of the subject (OR = 1.39; 95% CI: 1.17-1.64), the number of drugs consumed in the month previous to the admission (OR = 1.74; 95% CI: 1.47-2.08) and the presence of renal disease (OR = 1.76; 95% CI: 1.41-2.55) are associated with ADR at hospital admission. Our results suggest that special attention should be paid to avoid adverse drug reactions in elderly people with nutritional problems, renal failure or those who are using two drugs or more. Additional precautions should be taken with people with cognitive impairment, because of their reduced capacity to report their symptoms.