Methods-271 participants of a community based longitudinal study of aging in north Manhattan without evidence of significant cognitive impairment were interrogated for a history of head injury on two occasions at entry into the study. The examining physician sought a history of head injury with loss of conciousness. Independently, a risk factor interviewer inquired about a history of head injury with loss of consiousness or amnesia, the duration of any loss of consiousness, and the date of the head injury. Patients were followed up with standardised annual evaluations for up to five years to determine the first occurrence of dementia.

Results-Over the course of the study incident dementia due to probable or possible Alzheimer's disease was diagnosed in 39 patients. Cox proportional hazards modelling showed that a history of head injury with loss of consiousness reported to the physician was associated with earlier onset of dementia due to Alzheimer's disease (relative risk (RR) = 4.1, 95% confidence interval (95% CI) 1.3-12.7). head injury with loss of consiousness or amnesia reported to the risk factor interviewer was not significantly associated with earlier onset of Alzheimer's disease overall (RR 2.0, 95% CI 0.7-6.2), but those who reported loss of consiousness exceeding five minutes were at significantly increased risk (RR 11.2, 95% CI 2.3-59.8). Incident Alzheimer's disease was significantly associated with head injury which occurred within the preceding 30 years (RR 5.4, 95% CI 1.5-19.5).

Conclusion-The results of this cohort study are consistent with the findings of several case-control studies suggesting that head injury may be a risk factor for Alzheimer's disease.