Conversion of mild cognitive impairment to dementia in elderly subjects: A preliminary study in a memory and cognitive disorder unit

Prevalence and incidence of predementia syndromes vary as a result of different diagnostic criteria as well as different sampling and assessment procedures. Mild cognitive impairment (MCI) is thought to be a prodromal phase of dementia and therefore highly predictive of subsequent conversion. The aim of our study was to investigate the risk of conversion to dementia for different MCI subtypes diagnosed according to standardized and recently revised criteria (amnestic; impairment of memory plus other cognitive domains; nonamnestic). Participants were recruited among the 2,866 patients referring to the Memory and Cognitive Disorders Unit of the Local Health Unit of Bologna, Maggiore Hospital, between October 2000 and February 2006. In this preliminary study we analyzed data from 52 elderly outpatients with a diagnosis of MCI and a mean follow-up of 1.21 +/- 0.61 years (range 0.23-3.10 years). Mean age was 72.8 +/- 6.6 years, males were 61.5%. Mean baseline mini mental state examination (MMSE) score was 27.1 +/- 1.5. There were 15 incident cases of dementia (28.8%), with Alzheimer's disease (AD) accounting for 53.3% of all cases AD with cerebrovascular disease for 33.4% and fronto-temporal dementia for 13.3%. Overall rate of conversion was 23.8 per 100 person-years. During the same follow-up period, 53.8% of participants remained stable and 17.3% reverted to normal. Rates of conversion for the specific MCI subtypes were 38 per 100 person-years for amnestic MCI, 20 per 100 person-years for non-amnestic MCI, and 16 per 100 person-years for memory plus other cognitive domains MCI. With respect to non-converters, converters were generally older (76.1 +/- 4.2 vs. 71.5 +/- 7.0 years, p = 0.021), had a lower MMSE score (26.4 +/- 1.66 vs. 27.4 +/- 1.4, p = 0.035) and a higher prevalence of atrophy at neuroimaging (73.7% vs. 42.4%, p = 0.047). Moreover, with respect to non-converter, converters tended to have higher serum high density lipoprotein (HDL) levels, and lower serum folate levels. No difference was observed for the other study variables, included MCI subtype. Our findings suggest that the current definitions for MCI subtypes, particulary those referring to individuals with multiple or non-amnestic cognitive impairment, include a substantial number of individuals who may not progress to dementia. The possible role of cortical atrophy and low folate in the conversion from MCI to dementia could have important implications, because both conditions are easily identifiable. Moreover, low folate status is potentially amenable to therapeutic options. Although discouraging with respect to the clinical usefulness of currently available MCI criteria, our results raise the possibility that defining a protocol of multiple clinical risk factors may be useful in identifying MCI individuals at increased risk of conversion.