Premium
Predictors of mortality in presenile and senile dementia
Author(s) -
Kaszniak Alfred W.,
Fox Jacob,
Gandell David L.,
Garron David C.,
Huckman Michael S.,
Ramsey Ruth G.
Publication year - 1978
Publication title -
annals of neurology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 4.764
H-Index - 296
eISSN - 1531-8249
pISSN - 0364-5134
DOI - 10.1002/ana.410030311
Subject(s) - dementia , medicine , neuropsychology , abnormality , atrophy , pediatrics , cerebral atrophy , discriminant function analysis , neuropsychological test , disease , cognition , psychiatry , machine learning , computer science
Forty‐seven hospitalized patients with a diagnosis of presenile or senile dementia and without focal neurological disease or major systemic illness were given complete neurological, neuroradiological, and neuropsychological examinations. Mortality at one year after hospitalization was determined, 12 patients being lost to follow‐up. Of the remaining 35 patients, 19 were alive and 16 had died. These two groups did not differ in age, education, length of dementia history, sex, race, or degree of cerebral atrophy (by computerized tomography). Significant differences were found in degree of electroencephalographic abnormality and in 8 of 14 cognitive measures, the largest single difference being on a test of expressive language. Discriminant function analysis made with the cognitive measures as dependent variables yielded a correct prediction (classification analysis) of mortality in all but 1 case (97% accuracy). These results suggest that (1) degree of functional brain impairment rather than degree of cerebral atrophy may be the more important influence on mortality in dementia patients without focal lesions, (2) short‐term survival may be accurately predicted in such patients by tests of cognitive functioning, and (3) expressive language deficit in such patients may indicate a particularly poor prognosis for survival.