This cross-sectional study investigates the relation between a broad category of cognitive impairment and depression in a sample of 654 subjects aged 65 or over. This sample represents 74% of all subjects of that age living in a defined electoral district in North London, UK. The presence of depression and cognitive impairment was ascertained by interviewing all subjects with Short-CARE. Information was collected also for a variety of socto-demographic factors, level of social support and variables of functional limitation (i.e., impairment, disability and handicap). We found a cross-sectional association between depression and cognitive impairment (OR = 3.3; 95% CI: 2.1-3.1). However, when the analysis was adjusted for potential confounders using stratified analysis and logistic regression, we found that variables of functional limitation (especially disability and handicap) acted as confounders of the above association. This confounding effect did not differ significantly across sexes in our study.