Current issues of ACP Journal Club are published in Annals of Internal Medicine


Dementia severity, wandering and falling, and behavioral problems were associated with shorter survival in Alzheimer disease

ACP J Club. 1991 Jan-Feb;114:21. doi:10.7326/ACPJC-1991-114-1-021

Source Citation

Walsh JS, Welch G, Larson EB. Survival of outpatients with Alzheimer-type dementia. Ann Intern Med. 1990;113:429-34.



To examine the clinical course of Alzheimer-type dementia and those factors that might predict or influence the length of survival.


Inception cohort assembled at the time of initial diagnosis and followed prospectively until death or for at least 6 years.


Outpatients at a university hospital. 98% were referred from the Geriatrics and Family Services Program, Department of Psychiatry, University of Washington.


126 patients diagnosed with Alzheimer-type dementia were selected from among 200 consecutive outpatients evaluated for suspected dementia from 1980 to 1982. The diagnosis was based on published guidelines and the DSM-Ill criteria. All 126 patients were followed successfully.

Assessment of prognostic factors

All patients had a standardized diagnostic evaluation that included initial medical, psychiatric, diagnostic laboratory, and neuropsychologic assessments. Historical data were collected, with particular attention paid to symptom duration; age at symptom onset; and concurrent medical, neurologic, psychologic, and behavioral problems. The Mini-Mental State Examination (MMSE) and the modified Blessed Dementia Rating Scale were used to rate cognitive impairment as well as habits and activities of daily living.

Main outcome measure

Length of survival.

Main results

By the end of the follow-up period, 77 (61%) of the 126 patients had died. The median survival from symptom onset was 9.3 years (range 1.8 y to 16+ y; interquartile range 6.1 y to 10.3 y), whereas the median survival from study entry (initial diagnosis) was 5.3 years (range 0.2 y to 7.2+ y; interquartile range 2.9 y to 6.2 y). Patients who were older at symptom onset had a shorter survival (P < 0.01). Dementia severity, as measured by the MMSE, was strongly associated with survival (P < 0.001); the median survival of patients with scores of 18 or below was 2.9 years less than that of patients with scores above 18 (relative risk [RR] for death in patients with low MMSE scores 2.7, 95% CI 1.6 to 4.4). Comorbid conditions and symptom duration were not related to survival. A multivariate analysis of age at symptom onset and of historical features showed that the combination of wandering and falling (RR 2.1, CI 0.9 to 5.2) and behavioral problems (RR 1.4, CI 0.7 to 2.9) at the time of evaluation appeared to be associated with shorter survival.


The survival of patients with Alzheimer-type dementia was highly variable. Dementia severity, the combination of wandering and falling, and behavioral problems were associated with shorter survival.

Sources of funding: National Institute on Aging; Alzheimer Disease Research Center; Alzheimer Disease Patient Registry.

Address for article reprint: Dr. E. B. Larson, Medical Director's Office, RD-30, University of Washington, Seattle, WA 98195, USA.


In Alzheimer disease, the clinician's role is supportive, helping to manage behavioral and functional complications and educating about the disease. Families usually want to learn about prognosis. Previous research about prognosis produced conflicting results, in part because of methodologic flaws. Walsh and colleagues have correctly applied survival analysis, usefully distinguishing between time of symptom onset and time of diagnosis. Their use of a multivariate analysis to control for variable interactions importantly controls for age as an independent predictor. The Cox model does not show explained variance, so the importance of their predictors cannot be determined. Their finding that dementia severity was related to mortality is supported by another recently published cohort study (1). In their own multivariate analysis of prognostic indicators, Drachman and co-workers (1) found that depression, anxiety, and psychotic symptoms were not associated with earlier deterioration. Also, as Walsh and co-workers note, treating behavioral problems with psychotropic medications may cause complications.

The marked variability in disease duration makes it difficult to apply these predictors to clinical practice. However, the observation that patients with wandering and falling have a 50% 3-year mortality, compared with 20% among those without, is useful. This association, however, should not be seen as a justification to restrain wanderers. The association with behavioral problems should also be viewed with caution. Additional research is needed to develop accurate patient-specific prognoses.

Calvin H. Hirsch, MD
University of California, DavisDavis, California, USA


1. Drachman DA, O'Donnell BF, Lew RA, Swearer JM. The prognosis in Alzheimer's disease: "How far" rather than "how fast" best predicts the course. Arch Neurol. 1990;47:851-6.