Increased risk of dementia in mothers of Alzheimer's disease cases

Several studies of families with AD have reported a higher frequency of progressive dementia in mothers of subjects with AD than in fathers, ^[1-3] suggesting maternal transmission. The relative frequencies of dementia reported in these previous studies are difficult to interpret, however, because they do not take into account the greater life expectancy of women and the resulting greater maternal lifetime risk for AD. In the present study, we test the hypothesis of increased relative risk of progressive dementia in mothers of AD cases using survival analysis to adjust for gender differences in life expectancy.

Methods.

The diagnostic criteria and procedures for obtaining family histories from the 118 patients with probable AD in the present study have been described by Silverman et al. ^[4] These probands were selected from 11 university medical centers participating in the Consortium to Establish a Registry for Alzheimer's Disease. ^[5] For each of these 118 probands, a living nondemented spouse served as informant and as control subject. To date, 25 of 118 probands have been autopsied; the diagnosis of AD was neuropathologically confirmed in all 25.

Results.

The mean age of the mothers of probands was significantly greater than that of the fathers at time of death (77.0 +/- 13.4 versus 71.8 +/- 13.2 years; p = 0.004). At the time of death, 19 mothers had manifestations of primary progressive dementia (PPD) consistent with AD, as compared with only five fathers. Using the Cox proportional hazards model, the age-adjusted risk of PPD was found to be 2.8 times higher in mothers than in fathers (95% CI 1.1 to 7.7, log rank chi squared = 4.86, p = 0.028). Figure 1 shows age-specific cumulative risk estimates of PPD in parents of AD probands.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Cumulative risk of primary progressive dementia (PPD) in mothers and fathers of AD probands. By age 90, the cumulative risk of PPD is estimated to be 27.7% for mothers and 12.7% for fathers.

Family history data were also available for the parents of the spouse control subjects and for the siblings of the probands with AD. In the event of maternal transmission, one would expect the female-to-male relative risk to be lower in these groups compared with the 2.8:1 ratio observed in the proband parents. Among the parents of the spouse control subjects, 10 mothers and 4 fathers were affected. The age-adjusted female-to-male relative risk of PPD was 1.6, (95% CI 0.5 to 5.0). For the siblings of probands, the age-adjusted relative risk was 2.1 (95% CI 0.9 to 4.7). The relative risk in both groups was lower than that observed in the parental generation, although the differences did not reach statistical significance. Larger study samples are needed to resolve this question.

To pursue this further, we made a separate study of 10 families in this cohort with a strong familial pattern, that is, having one affected parent and at least one affected sibling beyond the proband case. The observed ratio of affected mothers versus fathers was 9:1, consistent with a subtype of AD with maternal transmission.

Discussion.

We found the age-adjusted risk of PPD consistent with AD to be significantly higher in the mothers than in the fathers. Furthermore, the age-adjusted female-to-male relative risk is higher in the parents of AD cases than in the parents of nondemented spouse control subjects or in the sibling generation. These latter findings are important because female gender may in itself be a risk factor for AD. ^[2-4,6] If female gender was a risk factor, the apparent risk of dementia in the mothers of proband cases would be increased regardless of a maternal transmission effect. Comparisons of risk ratios in proband parents versus risk ratios in other groups, such as the spouse control parent and the proband sibling groups reported here, should be free of this potential confound. These comparisons found that there was a risk to mothers of AD cases that was above and beyond the female gender-associated risk observed in these control groups.

In this regard, other family history studies of AD reporting generation- and gender-specific data are consistent with our findings Table 1. The observed proportion of female-to-male cases varies considerably across these studies, perhaps because of the differences in diagnostic criteria. However, comparisons within these studies control for these factors and should remain valid. Their data consistently suggest a higher relative risk for women in the parental generation than in the sibling generation, with the female-to-male case ratio being about twice as high in the parental than the sibling generation.

Several artifacts may have contributed to these findings. Falsely attributed paternity would lead to an understatement of paternal transmission rates and bias relative risk estimates toward the maternal lineage. Longer survival time within female AD cases ^[7] may increase the likelihood of a discernable progression of dementia before death in women and bias female-to-male relative risk estimates; this bias, however, would also effect relative risk estimates for spouse control parents and siblings, so that conclusions based on comparisons against these groups should remain valid. Finally, the proband case series and the research series summarized in Table 1 may be more representative of familial or early-onset cases. Hence, inference to the general AD case population from these studies is not warranted.

One possible avenue of maternal transmission of AD is the maternally inherited mitochondrial genome. There are reports of several point mutations (variants) in mitochondrial DNA in AD cases, ^[8,9] with speculative but plausible mechanisms of involvement of the mitochondrion and by implication mitochondrial DNA. ^[8-10] Genomic imprinting mechanisms may also explain a maternal inheritance pattern in AD.

In conclusion, we demonstrated evidence of an increased number of affected mothers in probands with AD, consistent with maternal transmission patterns in AD. Confirmation of our findings will require additional study, using a larger number of well-characterized probands and their families.