Once elderly patients are diagnosed with Alzheimer's disease, the presence of other illnesses and the patients' ethnic background appear to affect their length of survival, investigators report in the medical journal Neurology.

Led by Dr. Yaakov Stern, the research team at Columbia University Medical Center identified 323 cases of Alzheimer's disease from a group of approximately 4300 Medicare recipients who were enrolled in the Washington Heights Inwood Columbia Aging Project in New York.

More than half of the subjects were Hispanic, one third were African American and about 10 percent were white. Stern's group followed these patients for an average of 4 years, to a maximum of about 13 years.

Although overall the overall average lifespan (92 years) and age at diagnosis (83 years) did not differ by racial or ethnic group, the average length of survival after Alzheimer's disease was diagnosed was 8 years among Hispanics, significantly longer than among whites (4 years) and African Americans (5 years).

A history of diabetes and high blood pressure both independently shortened survival. Closer analysis of the data revealed that the mortality risk was increased by 2.6-fold among Alzheimer's disease patients with high blood pressure, and by 2-fold among patients who also had diabetes, both statistically significant differences.

Age was also a significant predictor of shortened survival time, with an average post-diagnosis survival of 10 years among those between 67 and 74 years old; 7 years among those between 75 and 84 years; and 4 years among those between 85 and 100 years old, all statistically significant differences.

In contrast to most other studies, the team did not detect any association between survival duration and gender, history of heart disease or malignancy, or the presence of APOE-epsilon-4 alleles - the gene associated with Alzheimer's disease that is detected in some but not all who develop this type of dementia.

The lack of any apparent association with the APOE gene, Stern's group suggests, "may reflect a differential effect of epsilon-4 in the earlier stages of disease that is eclipsed by other factors (medical, social, or disease-related factors) later in the disease course."