BACKGROUND Elders undergo approximately 40% of more than 1 million major abdominal operations annually. Yet evidence about recovery to preoperative levels of functional independence is limited. This study details course and predictors of functional recovery after elective major abdominal operations in the elderly. METHODS This was a prospective cohort of 372 consecutive patients, 60 years old or more, enrolled from surgeons in private practice and two university-affiliated hospitals, assessed preoperatively and postoperatively at 1, 3, and 6 weeks, 3 and 6 months, using self-report and performance-based measures (Activities of Daily Living [ADL], Instrumental Activities of Daily Living [IADL], Medical Outcomes Study Short Form-36 Physical Component and Mental Component Scales [PCS, MCS], Geriatric Depression Scale [GDS], Folstein Mini-Mental State Exam [MMSE], timed walk, functional reach, hand grip strength). RESULTS Mean age was 69 +/- 6 years with 56% men, 47% nonHispanic Caucasian, and 42% Mexican American; hospital distribution was 49% private, 51% university-affiliated. Maximum functional declines (95% CI) occurred 1 week postoperatively: ADL, 2.8 points (2.4 to 3.2); IADL, 7.6 points (7 to 8.3); SF-36 PCS, 6.5 points (5.4 to 7.6); Mini-Mental State Exam, 0.5 points (0.2 to 0.7); timed walk, 6.8 seconds (5.2 to 8.4); functional reach, 1.7 inches (1.2 to 2.2); grip strength, 2 kilograms (1.3 to 2.7) (p < 0.001 for all). SF-36 mental component scale and Geriatric Depression Scale scores did not worsen. Mean recovery times were: Mini-Mental State Exam, 3 weeks; timed walk, 6 weeks; ADL, SF-36 PCS, and functional reach, 3 months; and IADL, 6 months. Mean grip strength did not return to preoperative status by 6 months. The incidence of persistent disability at 6 months, compared with preoperative status, was: ADL, 9%; IADL, 19%; PCS, 16%; mental component scale, 17%; timed walk, 39%; functional reach, 58%; and grip strength, 52%. Potentially modifiable independent predictors of ADL and IADL recovery were preoperative physical conditioning and depression plus serious postoperative complications. CONCLUSIONS The clinical course of functional recovery varied across different measures. Protracted disability at 6 months after operation was substantial. Several potentially modifiable factors consistently predicted recovery.