SIR—In their article published recently, Freter et al. [1] assessed the feasibility of incorporating the Delirium Elderly At-Risk (DEAR) instrument into routine nursing care of elective orthopaedic patients and evaluated its usefulness to predict post-operative delirium. We would like to contribute to this topic with personal data. We assessed predictors of delirium during in-hospital rehabilitation in elderly patients who underwent an elective hip arthroplasty. All patients (n= 244) consecutively admitted to our unit for this reason from 1 January 2002 to 28 February 2004 underwent a multidimensional assessment including sociodemographic (age, gender, living condition), Geriatric Depression Scale (GDS), Charlson Index, body mass index and Barthel Index (BI) referring to 1 month before surgical intervention (BI pre-surgery) and to admission (BI admission). Delirium was ascertained on admission and during in-hospital rehabilitation according to the DSM IV criteria using the Confusion Assessment Method (CAM) [2]. The Mini-Mental State Examination (MMSE) was administered within 72 hours of admission or, in the case of delirium, after 3 consecutive days of negative CAM. Appendix 1 (available as supplementary data on the journal’s website, www.ageing.oxfordjournals.org) shows the characteristics of all patients stratified into two groups (delirium and no delirium). In comparison with the others, patients with delirium were significantly older, predominantly male, more impaired in cognitive and functional status, and had a higher comorbidity detected with the Charlson Index. When the effect of all variables that were significantly associated in the univariate model was tested in a multiple logistic regression, male gender (OR= 13.7, 95% CI 2.8–65.5, P= 0.001), greater comorbidity, as measured by a Charlson Index score ≥ 3 (OR= 9.4, 95% CI 1.3–66.9, P= 0.02), a MMSE score