Traditionally, schizophrenia has been studied in early adulthood. Its course and outcome during senescence are largely unknown and subject to controversy. We reviewed the consecutive neuropathologic records of 1,046 patients who were chronically hospitalized in New York State mental institutions, and we selected for analysis all 802 patients who died after age 50 with a clinical antemortem diagnosis, as recorded in the autopsy notes, of schizophrenia (n = 544) or dementia (n = 258). The prevalence of neuropathologic diagnoses consistent with Alzheimer's disease (AD) was 51 percent in the dementia group and 28 percent in the schizophrenia sample. This prevalence rate in the schizophrenia sample (mean age = 77) was considerably higher than that estimated for the general population. When evaluated against age of death, AD findings in demented patients were age invariant, whereas the rate of such findings in schizophrenia patients rose monotonously from under 5 percent below age 60 to 50 percent at age 90 and over. The age-relative rate of AD diagnosis in schizophrenia patients was similar to a curve postulated for first-degree relatives of familial AD patients and was markedly higher than population estimates. These findings as a basis for comparison with other retrospective studies of pathological records are presented. Our own study, as well as others, suffers from three intrinsic limitations. The clinical diagnoses are taken from death notes and have no formal verification. Likewise, neuropathologic diagnoses were based upon informal criteria in use at the time; since that time, formal diagnostic criteria have been evolving, and new staining methods have become available. Finally, it is not possible to determine from this material whether these patients are representative of all elderly schizophrenia patients or even of those who are institutionalized. Therefore, despite the large sample size on which our current findings are based, a new study has begun to address these weaknesses by complete review and rediagnosis of medical records and neuropathological material, using current methods, standardized criteria, and quantitative measures of degenerative changes. Specifically, the new ongoing study examines whether autopsied patients are representative by performing detailed diagnostic reviews of a control sample of nonautopsied patients from the same institutions. These results, if confirmed in the new study, demonstrate substantially greater vulnerability of chronic schizophrenia patients to the development of AD (or, at least, to histological changes typical of this disease). Possible association with chronic neuroleptic treatment and pathophysiological mechanisms remains to be elucidated.