Dementia in Newfoundland: identification of a geographical isolate?

STUDY OBJECTIVE

The aims were (1) to identify from death certificates regions with an increased incidence of dementia mortality; and (2) to determine whether a previously observed excess of patients with Alzheimer disease originating from a small area could be confirmed in a survey of death certificates.

DESIGN

The study identified all individuals dying with dementia, recorded on death certificates as an immediate, antecedent, underlying, or contributing cause of death. Rather than the usual residence, the birthplace of these individuals was used to determine regional differences in dementia mortality. A comparison was made of two areas to test the significance of a geographical isolate of persons. To test for a possible genetic component of the excess, an analysis was made of the frequencies of family names. To test for a possible environmental component an analysis was made of standard measurements of drinking water quality.

SETTING

The survey data were derived from all 1985 and 1986 deaths in the province of Newfoundland.

MEASUREMENTS AND MAIN RESULTS

Based on the current census population, the prevalence of dementia at death for 1985 and 1986 was 34 and 37/100,000. For both years there was a significant excess of persons originating from a small area (95% CI, 1.1-20.7%, and 2.5-20.4%). This excess could not be explained by differences in age, sex, ethnic origin, or by variation in mobility patterns. The study area has a high concentration of aluminium in the drinking water. An analysis of the family names gave inconclusive evidence of a clustering among the dementia cases.

CONCLUSIONS

If all contributing causes of death are recorded and the birthplace of individuals is noted, mortality statistics can reveal regional differences in dementia rates. This shows the need to examine areas smaller than census districts to identify subpopulation variation in the prevalence of dementia. Environmental influences can vary substantially in areas relatively close together, as evidenced in measurements of drinking water chemistry. Genetic influences are more likely to be revealed from the birthplace of individuals, which may indicate a common ancestry.