Background. Features that may allow early identification of patients at risk of prolonged delirium, and therefore of poorer outcomes, are not well understood. The aim of this study was to determine if preoperative delirium risk factors and delirium symptoms (at onset and clinical symptomatology during the course of delirium) are associated with delirium duration. Methods. This study was conducted in prospectively identified cases of incident delirium. We compared patients experiencing delirium of short duration (1 or 2 days) with patients who had more prolonged delirium (≥3 days) with regard to DRS-R-98 (Delirium Rating Scale Revised-98) symptoms on the first delirious day. Delirium symptom profile was evaluated daily during the delirium course. Results. In a homogenous population of 51 elderly hip-surgery patients, we found that the severity of individual delirium symptoms on the first day of delirium was not associated with duration of delirium. Preexisting cognitive decline was associated with prolonged delirium. Longitudinal analysis using the generalised estimating equations method (GEE) identified that more severe impairment of long-term memory across the whole delirium episode was associated with longer duration of delirium. Conclusion. Preexisting cognitive decline rather than severity of individual delirium symptoms at onset is strongly associated with delirium duration.