Variability of age at onset (AO) of Alzheimer disease (AD) among members of the same family is important as a biological clue and because of its clinical effects.
To evaluate which clinical variables influence the discrepancy in AO among affected relatives with familial AD.
Clinical genetic project of Spanish kindred with AD conducted by 4 academic hospitals in Madrid, Spain.
Age at onset of AD in 162 families and discrepancy in AO in intragenerational and intergenerational affected pairs were analyzed in relation to age, sex, maternal or paternal transmission, pattern of inheritance, and apolipoprotein E genotype.
Maternal transmission of AD was significantly more frequent than paternal transmission (P < .001). In 27% of the affected individuals, AO occurred before the patient was 65 years old. Discrepancy in AO among siblings was within 5 years in 44% of the families, 6 to 10 years in 29%, and more than 10 years in 27% (range, 0-22). This discrepancy was independent of the sex of the sibling pairs and was significantly lower with maternal transmission of AD (P = .02). Segregation analysis showed no differences in the inheritance pattern between families with low (≤5 years) or high (>5 years) AO discrepancy. Age at onset in carriers of the apolipoprotein E ε4 allele was slightly younger. However, among siblings, an extra apolipoprotein E ε4 allele was not consistently associated with earlier onset of AD. Eighty percent of patients, independent of sex or mode of transmission, were already affected at their parents' reported AO.
There is a wide discrepancy in AO in affected siblings that is not clearly explained by a single clinical variable or apolipoprotein E genotype. The interaction of many factors probably determines AO in each affected individual. However, maternal transmission of AD seems to result in a similar AO in offspring, and the risk of developing dementia after the parent's reported AO decreases significantly.
Alzheimer disease (AD) is the most common degenerative dementia. Age at onset (AO) varies between 40 and 90 years and in most patients is after age 65 years. Earlier onset, or presenile AD, has provided the first genetic insights into the disease via the study of patients with autosomal-dominant transmission. It is well known that genetic factors such as the presence of presenilins or amyloid precursor protein mutations are associated with early AO in kindreds.1 However, even in these families, AO may vary considerably. The presence of the ε4 allele within the apolipoprotein E (APOE) genotype also decreases AO in a dose-dependent way in late-onset familial and sporadic cases.1 There are probably other genetic AO modifiers that are still unknown, and nongenetic factors such as educational achievement level, environmental factors, or trigger events may also contribute to the variability in AO.
The possibility of anticipating or understanding the mechanisms that influence AO of AD is of great importance in clinical settings because it enables more accurate genetic counseling, follow-up of presymptomatic cases, and exclusion of individuals at risk after a certain age and aids in the decision of when initiation of therapeutic measures is appropriate. We analyzed data from the GENODEM Project (a research program that collects clinical genetic information and DNA samples of Spanish kindreds with familial AD) to assess the discrepancy in AO among affected members with familial AD and the clinical variables that may influence this variability.
The GENODEM Project involves 4 academic hospitals in Madrid (Fundación Jiménez Díaz, Hospital Clíinico San Carlos, Hospital Ramón y Cajal, and Fundación Hospital Alcorcón) that use common standardized dementia protocols.
A total of 162 probands with a clinical diagnosis of probable AD according to standardized criteria,2 AO before 85 years, and at least 1 affected first-degree relative known to have a diagnosis of AD or having a similar clinical dementia (insidious onset, evolving in years, and unrelated to trauma or alcohol abuse) have been studied. For each proband, detailed data on family history of dementia were collected. For all affected individuals in each family, data on sex, AO of symptoms, and age at death were obtained. Special effort was made to ensure that AO data in siblings were as accurate as possible by reviewing medical records from other centers and by telephone interviews with the closest relative. For AO in parents, we relied on information provided by their offspring and double-checked it with several family members when available.
The following data were analyzed: general demographic data about the entire sample of families with AD including distribution of affected and unaffected members according to sex, maternal or paternal transmission, number of affected members per family, distribution of affected members according to AO, and influence of sex and parental transmission in AO. Both intragenerational and intergenerational discrepancy in AO within families were analyzed. Intragenerational discrepancy was analyzed with consideration of differences across all affected siblings to establish the range of variability and within sibling pairs to consider the influence of sex, AO, parental transmission, and pattern of inheritance. Genetic data were analyzed using the Genetic Analysis Package program (GAP; Epicenter Software, University of Southern California, Pasadena), which enables construction of a complete schema of a family tree including previously defined data such as sex, AO, age at death, and maternal or paternal transmission. Pattern of inheritance was analyzed by complex segregation analysis to compare the best-fit model of transmission between families with low (≤5 years) or high (>5 years) AO discrepancy. This analysis was carried out according to the unified model of complex segregation analysis implemented with POINTER software (Division of Biostatistics, Washington University Medical School, St Louis, Missouri).3 The model partitions the total variation of the underlying liability for AD into 3 independent components: a biallelic single major locus, a polygenic background, and a random environmental component.
Intergenerational AO discrepancy was examined comparing pairs of affected parent and offspring. We calculated cumulative survival rates with Kaplan-Meier curves, which expressed the probability of the descendents still being asymptomatic at a certain age in comparison with the parent's reported AO. The effect of sex was also examined.
Age at onset in all probands and discrepancy in AO in 42 intragenerational sibling pairs were also examined according to APOE genotype. Blood samples were collected after the patient or surrogate gave informed consent. The study was approved by the Research Ethics Committee of Fundación Jiménez Díaz. APOE genotype was determined by polymerase chain reaction digestion according to the method of Hixson and Vernier.4 Differences in AO comparing APOE-concordant with APOE-discordant sibling pairs were tested using the t test. For discordant sibling pairs, differences in AO were considered positive when dementia developed earlier in the sibling with more ε4 alleles and negative when it developed later.
Demographic and general ao data
A total of 162 probands or families were recruited. There was a mean of 3 affected members per family, with 504 affected cases of 2024 individuals in the family trees. The number of affected female members was significantly higher (30%) than affected male members (17%) (P = .001). Of all affected individuals, 66% were women and 34% were men. Maternal transmission was more frequent than paternal transmission (120 vs 47; P < .001).
Mean ± SD AO in the probands was 69 ± 9 years (age range, 29-85 years). Reliable AO data were also available for 90 siblings of a total of 252 affected individuals. The distribution of affected individuals according to AO is shown in Figure 1. For 27% of the affected individuals AO occurred before these persons were 65 years old; in 73% AO occurred when these persons were 65 years old or older. There were no differences in AO between men and women (P = .34) or between maternal and paternal transmission (P = .36; Table 1).
Discrepancy in ao within affected pairs
Information about AO in several affected siblings was available in 79 families and demonstrated a wide discrepancy (Figure 2). The value chosen was the highest discrepancy in AO among all affected siblings in a family, that is, 1 value per family. This discrepancy was within 5 years in 44% of the families, 6 to 10 years in 29%, and more than 10 years in 27%. There was a tendency toward a smaller difference when the AO in the proband was higher (P = .06).
The discrepancy in AO between sibling pairs was independent of sex (P = .31) (Figure 3A). Mean ± SD discrepancy was 4.9 ± 3.6 years between sisters, 3.2 ± 3.5 years between brothers, and 4.2 ± 4.0 years when the sibling pair was a brother and a sister. The mean ± SD discrepancy in AO within sibling pairs was significantly lower with maternal transmission of AD (3.9 ± 3.7 years vs 6.3 ± 5.2 years with paternal transmission; P = .02; Figure 3B). Complex segregation analysis showed that both families with a high (>5 years) or low (≤5 years) discrepancy in AO among siblings better fit in the same model of inheritance (Table 2).
Data on AO in parents were available for 119 families. The mean ± SD differences in AO were similar for paternal transmission to son (n = 12; 1.3 ± 5.3 years) or daughter (n = 22; 3.3 ± 6.5 years) compared with maternal transmission to son (n = 18; 3.7 ± 8.6 years) or daughter (n = 67; 2.9 ± 6.6 years). Figure 4 shows cumulative survival rates (ie, percentage of still asymptomatic offspring) at a certain age compared with reported AO in the parent. Only about 20% of affected offspring had no symptoms at the AO in their parent (cumulative survival rate, 0.23), and this percentage was the same for maternal or paternal transmission. At 5 years after AO in the parent, only 6% of offspring affected were still without symptoms.
Sixty-six percent of 106 individuals with AD carried the APOE 3/4 (53%) or APOE 4/4 (13%) genotype, and 34% carried the APOE 3/3 genotype. Distribution of female and male carriers was similar for the 3 APOE genotypes. There were no carriers with APOEε2 alleles. Mean ± SD AO in APOE 4/4 carriers was younger (66 ± 5 years) compared with APOE 3/4 (69 ± 7 years) and APOE 3/3 carriers (71 ± 6 years) (P = .11) (Figure 5A).
APOE genotype was available for 42 sibling pairs (Figure 5B). The distribution of 29 APOE-concordant sibling pairs and mean ± SD difference in AO was APOE 3/3 (n = 6; 6.3 ± 5.2 years), APOE 3/4 (n = 16; 3.5 ± 3.0 years), and APOE 4/4 (n = 7; 6.5 ± 3.4 years), and for 13 APOE-discordant sibling pairs was APOE 3/4 vs APOE 4/4 (n = 8; 4.2 ± 2.4 years) and APOE 3/3 vs APOE 3/4 (n = 5; 6.6 ± 2.9 years). There were no APOE 3/3 vs APOE 4/4 sibling pairs. The mean ± SD difference in AO in APOE-concordant sibling pairs (5.1 ± 3.9 years) was not significantly different from that in discordant sibling pairs (5.1 ± 2.8 years; P = .16). When sibling pairs were APOE discordant (1 of the pair had 1 APOE ε4 allele more), only in 50% of this occurrence the sibling with more ε4 alleles had an earlier AO than its pair (positive value in Figure 5).
Our study findings show a wide variability in AO of symptoms in siblings with familial AD that is not dependent on a single clinical variable or APOE genotype, although it points to maternal transmission as a clinical feature that may decrease this difference. They also show that most affected individuals already have symptomatic disease at their parents’ reported AO.
Several studies suggest that AO is a clinical feature strongly determined by genetic factors. This is supported in that it is more homogeneous within families than between families,5 40% of the variance of this trait can be explained by familial effects,6 amyloid precursor protein and presenilin-1 and presenilin-2 mutations are associated with specific AO ranges,7,8 and some genes have been identified as influencing AO in AD, including the APOE genotype.9 In this series, in almost half of the families (44%), the difference in AO was within 5 years among siblings, but more than half of the families had a much wider range of difference (6-22 years). The discrepancy in AO was not influenced by sex of the affected siblings, and there did not seem to be a different pattern of inheritance, as shown by segregation analysis, between families with high or low discrepancy.
However, the difference in AO within sibling pairs was significantly lower when the disease was maternally transmitted. This is important because maternal transmission, in this study and in others,10 is much more frequent than paternal transmission. In these families, there were overwhelming data for a female preponderance in most aspects: substantially more women were affected, and there were more cases of maternal transmission and more mother-daughter transmission. Many studies have analyzed why AD samples always exhibit more affected women than men. Some argue that the explanation may be longer life expectancy in women and that there is not a sex-specific risk for AD11; others find female sex to be an independent risk factor alone12,13 or when associated with the presence of the ε4 allele.14,15 Insofar as transmission, the offspring of affected mothers do not seem to be at greater risk of dementia than the offspring of affected fathers.16 It is possible that a percentage of paternally transmitted AD cases would be included within sporadic AD, given the lack of clinical information owing to premature deaths. In this respect, the risk of developing AD in a longitudinal study suggests that a genetic component is also important in apparently sporadic cases.17
Possible explanations for a maternal inheritance pattern of AD have been suggested, such as genomic imprinting mechanisms or transmission through mitochondrial DNA alterations.13 Bassett et al18 reported several genomic regions linked to late-onset AD specific to disease transmission from the mother. Maternal transmission of AD may have distinct characteristics, such as a more homogeneous AO in offspring.
The ε4 allele was clearly overrepresented in this series with familial AD (66% had 1 or 2 ε4 alleles), inasmuch as its presence in the Spanish population has been reported in 33% of individuals with sporadic AD and 12% of control individuals.19 The aggregation of ε4 carriers in familial AD has also been reported by other authors.10 It is confirmed as a contributor to earlier AO, although in this group it did not reach statistical significance. However, the effect of the APOE genotype was not strong within families, first, because the same variability in AO was found in APOE-concordant or APOE-discordant sibling pairs, as also reported by Axelman et al,20 and second, AO in almost half of the ε4 allele–positive sibling pairs was later than in their ε4 allele–minus siblings.
Other environmental factors can trigger or influence the onset of dementia. Sometimes the death of a spouse or depression are mentioned by relatives as the beginning of cognitive decline. Alternatively, a high level of educational achievement has been considered a preventive or delaying factor for dementia.21 This variable was not informative in our sample because there was a narrow range of education, either primary or secondary school. Within the same family and in individuals who were reared in the 1920s to 1950s, women had a lower level of educational achievement than men; thus, when comparing sibling pairs with different sex, we could have indirectly found some effect of this factor. However, the difference in AO was similar for sibling pairs of the same or different sex.
In these families, 80% of the affected offspring already had symptoms at the reported AO in the parent and only 6% developed the disease 5 years after that AO. These data may be biased by the different attention to symptoms paid in previous generations, when dementia could have been neglected until clearly established, whereas today early cognitive symptoms are detected and diagnosed; thus, no strong conclusions can be drawn. However, our data are consistent with those of other studies that show that genetic risk of developing AD in relatives peaks at a certain age and then decreases significantly.22 That is, it is probable that genetic factors causing familial AD interact during a certain period, whereas risk of developing AD in individuals in their late 80s has more to do with aging itself than with familial features. On the other hand, our data may be clinically valuable because they can support giving positive expectations when an individual has no symptoms at the age when his or her parent was recognized as having cognitive impairment. Prospective studies should examine further whether the risk of developing dementia decreases significantly once an individual is older than the AO in the parent and whether there may be a tendency to anticipate development of AD.
In summary, AO of AD is a clinical feature probably conditioned by the complex interaction of several genetic factors and a slight influence of extragenetic variables. Although it is difficult to predict an approximate AO in a certain individual with familial AD, maternal inheritance suggests a closer AO among siblings and that children who have no symptoms by the reported AO of dementia in their parents are less prone to develop AD.
Correspondence: Estrella Gómez-Tortosa, MD, PhD, Department of Neurology, Fundación Jiménez Díaz, Avda Reyes Católicos 2, 28040 Madrid, Spain (email@example.com).
Accepted for Publication: January 29, 2007.
Author Contributions: Dr Gómez-Tortosa had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Gómez-Tortosa, Barquero, and Jiménez-Escrig. Acquisition of data: Gómez-Tortosa, Barquero, Barón, Sainz, Manzano, Payno, Ros, Almaraz, and Gómez-Garré. Analysis and interpretation of data: Gómez-Tortosa and Barquero. Drafting of the manuscript: Gómez-Tortosa and Barón. Critical revision of the manuscript for important intellectual content: Gómez-Tortosa, Barquero, Sainz, Manzano, Payno, Ros, Almaraz, Gómez-Garré, and Jiménez-Escrig. Statistical analysis: Gómez-Tortosa, Barón, and Jiménez-Escrig. Obtained funding: Gómez-Tortosa and Barquero. Administrative, technical, and material support: Gómez-Tortosa, Barquero, Sainz, Manzano, and Payno. Study supervision: Gómez-Tortosa and Barquero.
Financial Disclosure: None reported.
Funding/Support: This study was supported in part by Red Centro de Investigación de Enfermedades Neurológica.
Additional Contributions: V. Sanchez, RN, assisted in obtaining blood samples.
L Causative and susceptibility genes for Alzheimer's disease: a review. Brain Res Bull
1- 24PubMedGoogle ScholarCrossref
EM The clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's disease. Neurology
939- 944PubMedGoogle ScholarCrossref
R A unified model for complex segregation analysis. Am J Med Genet
1983;35816- 826PubMedGoogle Scholar
DT Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with Hha
I. J Lipid Res
545- 548PubMedGoogle Scholar
et al. Interfamilial and intrafamilial phenotypic heterogeneity in familial Alzheimer's disease. J Geriatr Psychiatry Neurol
1- 6PubMedGoogle ScholarCrossref
et al. Familial influence on variation in age of onset and behavioural phenotype in Alzheimer's disease. Br J Psychiatry
2000;176156- 159PubMedGoogle ScholarCrossref
M AO in familial early onset Alzheimer's disease correlates with genetic etiology. Am J Med Genet
129- 130PubMedGoogle ScholarCrossref
et al. Familial Alzheimer's disease: site of mutation influences clinical phenotype. Ann Neurol
376- 379PubMedGoogle ScholarCrossref
et al. Age at onset in two common neurodegenerative diseases is genetically controlled. Am J Hum Genet
985- 993PubMedGoogle ScholarCrossref
et al. Alzheimer's disease: interaction of apolipoprotein E genotype, family history of dementia, gender, education, ethnicity, and age of onset. Neurology
1575- 1579PubMedGoogle ScholarCrossref
DA Is the risk of developing Alzheimer's disease greater for women than for men? Am J Epidemiol
132- 133PubMedGoogle ScholarCrossref
J Increased risk of familial late-onset Alzheimer's disease in women. Neurology
126- 129PubMedGoogle ScholarCrossref
JC Increased risk of dementia in mothers of Alzheimer's disease cases: evidence for maternal inheritance. Neurology
254- 256PubMedGoogle ScholarCrossref
et al. Gender difference in apolipoprotein E–associated risk for familial Alzheimer disease: a possible clue to the higher incidence of Alzheimer disease in women. Am J Hum Genet
803- 811PubMedGoogle Scholar
et al APOE-epsilon4 count predicts age when prevalence of AD increases, then declines: the Cache County Study. Neurology
[published correction appears in Neurology
et al. Genetic epidemiological study of maternal and paternal transmission of Alzheimer's disease. Am J Med Genet
378- 382PubMedGoogle ScholarCrossref
KL Familial aggregation in Alzheimer's disease: comparison of risk among relatives of early- and late-onset cases, and among male and female relatives in successive generations. Neurology
1988;38207- 221PubMedGoogle ScholarCrossref
D Evidence for parent of origin effect in late-onset Alzheimer disease. Am J Med Genet
679- 686PubMedGoogle ScholarCrossref
et al. Genotype and phenotype of apolipoprotein E in patients with Alzheimer's disease in Castille and Leon [in Spanish]. Neurologia
384- 388PubMedGoogle Scholar
L Apolipoprotein E and alpha1-antichymotrypsin genotypes and age of onset of familial Alzheimer's disease. Dement Geriatr Cogn Disord
1- 5PubMedGoogle ScholarCrossref
et al. Education modifies the relation of AD pathology to level of cognitive function in older persons. Neurology
1909- 1915PubMedGoogle ScholarCrossref
M Variability of familial risk of Alzheimer disease across the late life span. Arch Gen Psychiatry
565- 573PubMedGoogle ScholarCrossref