A study has been carried out in our geriatric rehabilitation wards, to know better the role of cognitive dysfunction in treatment outcome and processing, mainly functional recovery and comorbidity. We analyzed 478 frail inpatients, 2/3 of them were females, aged 78.7 + 9.2 years, consecutively admitted to the unit. Assessment of cognitive function was performed by the mini mental state examination (MMSE), of functional status by Barthel index (BI), of co-morbidity by cumulative illness rating scale (CIRS) both at the admission and discharge for each patient, together with the usual clinical parameters and social outcome. More than one people in three lived alone and showed some communication problems; one in seven needed modifications in food preparation or nutritional system;more than one in five had pressure sore >/= 2 (EPUAP = European Pressure Ulcers Advisory Panel classification). The mean (+/- SD) levels of the admitted were: MMSE = 20 +/- 7.2; BI total score = 45.5 +/- 28.9; CIRS class = 4 +/- 2, CIRS severity score = 1.9 +/- 0.4. Of the patients, 74.5% were discharged to home. Mean functional gain was 20 points at BI: 65.6 +/-30.7; MMSE improved to 21.6 +/- 7.1 (p = 0.00005 for both comparisons, by Wilcoxon test).MMSE was positively correlated to BI (r = 0.6, p = 0.0005) and negatively correlated (p =0.0005) to CIRS comorbidity (r = -0.33) and severity (r = -0.26), and to age (r = -0.38, p =0.0005) both at admission and at discharge. MMSE at admission was correlated neither to functional gain, nor to improvement of clinical indicators, nor to measures of functional and clinical efficiency. MMSE at discharge showed similar results. In a stepwise multiple correlation analysis, taking the gain in BI as the dependent variable, while BI, serum albumin level, MMSE, CIRS severity and comorbidity indexes at admission as independent variables, MMSE together with BI admission total score and CIRS severity index retained a strong association with functional gain, whilst admission serum albumin levels and CIRSindex lost it. Consistently with some literary data, we showed the actual possibility of functional and clinical gains for people cared in a geriatric rehabilitation ward, all through a wide range of MMSE score, including subnormal scores. Poor cognitive status bears heavily on frailty, but does not hamper the outcomes of genuine rehabilitative efforts in geriatrics.