Background: The negative interactions between Alzheimer's disease (AD) patients and their caregivers may provoke criticism, hostility, and emotional overinvolvement that characterise highly expressed emotion (EE) attitudes. In this study, we hypothesised that affective temperament traits of AD caregivers are related to their high EE levels independent from other patient and caregiver characteristics. Methods: Eighty AD patients were assessed through Clinical Dementia Rating Scale (CDR), and Mini-Mental State Examination. Expressed Emotion Scale (EES), Temperament Evaluation of Memphis, Pisa, Paris and San Diego auto-questionnaire, and Hamilton Depression Rating Scale were applied to the caregivers. The high (n = 41) and low EE caregivers (n = 39) were compared with respect to some sociodemographic and clinical characteristics of the patients and caregivers, and affective temperament traits of caregivers. The associations of caregiver EES scores with multiple variables related to patients and caregivers were examined by Pearson correlation tests. We performed multiple linear regression analysis to determine the possible predictors of total EES scores. Results: High EE caregivers had significantly higher depressive, cyclothymic, and anxious temperament traits than in low EE subjects. A weak positive correlation was found between the total EES scores and Personal Care scores of CDR. Home and hobbies subscale scores of CDR had a moderately significant positive correlation with total EES scores. There was also moderate significant positive correlations between total EE scores and depressive, cyclothymic, anxious, and irritable temperament traits. Linear regression analysis indicated that depressive temperament significantly predicted the high levels of caregiver EE. Conclusion: These findings suggest that caregivers' depressive temperament is predominantly related to their EE levels even after controlling for the severity of AD, and lower educational level of caregivers. Our results may provide evidence that high EE might be a reflection of caregivers' depressive temperament traits, in accordance with the trait hypothesis.