Stroke in a defined elderly population, 1967-1985. A less lethal and disabling but no less common disease

Decline in stroke mortality in recent decades has been well documented in the United States and other countries. This study, based on a well-defined population with comprehensive medical records available for research purposes, seeks to explain decline in stroke mortality among older persons between 1967 and 1985. The study specifically explores the competing explanatory mechanisms of decreased incidence of stroke versus decreased case-fatality rate.

We conducted a retrospective analysis of three successive period cohorts (1967 through 1971, 1974 through 1978, and 1981 through 1985) of persons > or = 65 years of age enrolled in a large group model HMO in a metropolitan community. All new hospitalized and a sample of nonhospitalized strokes were ascertained, and samples of first-ever strokes were studied. Incidence, case-fatality rates, survival times, and comorbidities were compared across cohorts.

There was no significant change in stroke incidence over time; however, 1-month case fatality declined dramatically from 33% in 1967 through 1971 to 18% in 1981 through 1985 (P < .01); median survival increased from 213 to 1092 days. Indices of reduced severity included declines in coma from 27% to 12% (P < .01) and in wheelchair- or bed-bound status from 40% to 30% (P = .067). Cases with and without CT scan in 1981 to 1985, when this procedure became widely available in the health plan, were similar in severity, thereby reducing the possibility of ascertainment bias.

In this well-defined older population, stroke has become a less lethal and disabling though no less common disease. This finding fails to support the "compression of morbidity" hypothesis while supporting a model of delayed progression for stroke in this age group.