PERFORMANCE ON THE DIGIT SYMBOL SUBSTITUTION TEST AND 5-YEAR MORTALITY IN THE WESTERN COLLABORATIVE GROUP-STUDY

The relation between performance on the digit symbol substitution test and subsequent 5-year mortality was examined in 1,118 male participants in the Western Collaborative Group Study (mean age at time of examination = 70.6 years). Cox regression analyses revealed that the unadjusted relative risk from all-cause mortality was 1.79 (95% confidence interval 1.45-2.21) for subjects who scored 10 points (1 standard deviation) lower on the digit symbol substitution test. The relative risk remained significant at 1.44 (95% confidence interval 1.12-1.86) after adjustment for age, education, baseline serum cholesterol, and systolic blood pressure levels and for the prevalence of cancer, cardiovascular/cerebrovascular disease, and ever smoking. Tree-structured survival analysis revealed the presence of a subgroup of subjects (n = 130) with low digit symbol substitution performance (mean = 26.3) who experienced a mortality rate of 34.9 deaths per 1,000 person-years. This mortality rate was not significantly different from that encountered by another subgroup of subjects with a history of cancer (n = 128; 45.7 deaths per 1,000 person-years). If age-related slowing of information processing is a marker for aging of the central nervous system, then people with lower digit symbol scores may be ''older'' physiologically than similarly aged persons with higher digit symbol scores. Aging of the central nervous system might be expected to reduce adaptive capacity in a general way, thereby increasing susceptibility to death from a variety of causes.