OBJECTIVE

To determine incidence rates of non-dementia cognitive impairment, to examine the impact of attrition due to death on the observed incidence estimates, and to compare the observed and corrected estimates of non-dementia cognitive impairment with dementia incidence rates.

METHODS

A total of 1,435 persons without dementia aged 75+ from the Kungsholmen Project were evaluated for occurrence of dementia over 9 years. A total of 1,070 cognitively unimpaired subjects were also followed using amnestic mild cognitive impairment (aMCI) and other cognitive impairment, no dementia (OCIND) definitions. To correct the observed incidence rates for attrition due to death, cognitive status for subjects lost due to death was imputed using information on previous cognitive and health status. Observed and corrected incidence rates (IR) and 95% CIs were calculated with the person-years method, using Poisson distribution.

RESULTS

Incidence rates per 1,000 person-years were as follows: dementia IR = 70.4 (64.0 to 77.4); aMCI observed IR = 11.4 (8.6 to 15.1), corrected IR = 13.7 (10.3 to 18.2); OCIND observed IR = 33.8 (28.7 to 39.8), corrected IR = 42.1 (36.5 to 48.6). Both aMCI and OCIND incidence increased with advancing age. Observed incidence of aMCI and OCIND together was similar to that of dementia at age 75 to 79 but lower at more advanced ages. However, the cognitive impairment incidence after age 79 increased substantially when the estimates were corrected for attrition due to death during follow-up.

CONCLUSIONS

Non-dementia cognitive impairment is common and often underestimated in population studies that do not adjust for attrition.