Predictability of Age at Onset in Huntington Disease in the Dutch Population

Introduction Huntington disease (HD) is an autosomal dominant, progressive neuropsychiatric disorder. The main clinical symptoms are chorea, dementia, and changes in personality, mood, and behavior. The disease is incurable and leads to death usually within 17 years after onset, with a range of 2–45 years (18,53). HD serves often as a paradigm for other adult-onset hereditary diseases, especially repeat expansion disorders. Because of the great variability and gradual onset of the first symptoms, the precise age at onset is difficult to determine. Gradual behavioral changes may only become recognized as abnormal over a longer period of time and choreic movements may go unnoticed for a long period. Individuals who are familiar with the disease and/or are members of HD families recognize symptoms earlier than patients without a family history known to them. Most studies have taken as the point of onset the first definite abnormality, whatever its nature, recorded by a reliable witness. There is, usually, an 8-year delay between age at onset and age at diagnosis (32). Early studies (7,66) show a mean age at onset between 35 and 44 years. These studies show bias because later-onset gene carriers who will develop the disease are omitted. A life-table analysis applies a correction for carriers who have not yet developed the disorder and shows a later mean age at onset, between 46.0 and 48.9 years (60). The use of family data that are sufficiently remote from the time of the study to be certain that all those who are going to develop HD will have done so gives also a correction and a later age at onset of 43.4 years (60). Those cases with an age at onset less than or equal to 20 years are defined as juvenile HD (3%–10%) (18), and onset more than or equal to 60 years can be considered as HD in the elderly (8%–23%) (45,66). These types form the extremes of the distribution curve of the age at onset (2–80 yr) (18). The mean age at onset was found to be higher in female patients than in males (52). Paternal transmission has preferentially been found in juvenile patients, with an age at onset less than or equal to 20 years (41,43,51,67). In late-onset HD patients (more than or equal to 60 yr), the transmission was mostly confined to the mother (44,45) although this was not found in other studies (15,67). Age at onset and sex of the affected parent and grandparent seem to influence the age at onset in offspring. A lower age at onset and paternal and grandmaternal inheritance result in lower age at onset in offspring and anticipation (earlier age at onset in successive generations) (13,28,51,52). The HD gene has been localized on chromosome 4p16.3 (16), and the mutation was identified in 1993 (22). HD, like other neurodegenerative disorders such as spinocerebellar ataxia (SCA) type 1, 2, 3, 6, and 7; dentato-rubro-pallido-luysian atrophy (DRPLA); and spinobulbar muscular atrophy (SBMA), is caused by an expanded CAG repeat, coding for an expanded polyglutamine stretch in the cognate protein. The wildtype allele has 6–35 copies but affected subjects have repeats between 36 and 250 (30,47,56). As in SCA 1 and 2—and as opposed to SCA 3, 6, and 7, SBMA, and DRPLA—there is some overlap between normal and disease ranges in HD, and reduced penetrance is found in the lower range of the disease repeat. The normal repeat typically segregates simply and stably as a polymorphic locus, although occasionally changes are seen (11), but the higher-normal range (27–35 repeats) has increased instability and occasionally yields a repeat size in the affected range. The repeats in the HD range are even more unstable, and besides expanding, they also occasionally contract in successive generations (6). Due to instability of the high-normal repeats, the boundaries of the repeat size have been refined: the normal CAG repeat is considered to contain fewer than 27 repeats; the intermediate repeat, 27–35 repeats; the disease repeats with reduced penetrance, 36–39 repeats; and the regular disease repeat more than 39 repeats (3). The genetic status of an individual can now be ascertained with >99% accuracy (22,30). Like other CAG repeat expansion disorders, HD shows an inverse relationship between repeat size and age at onset. The CAG repeat accounts for approximately 50%–77% of the variation in the age at onset (4,8,10,12,23,24,26,27,33,35,38,40,46,49,57,58,59,61,64,65). In a 1997 Canadian study (8), age-dependent likelihood of onset curves were proposed, which differed significantly for each CAG repeat length in the 39–50 range. No studies of probability of onset for the other CAG repeat expansion disorders have been performed as yet. Information concerning the age-dependent likelihood of being affected, with a particular CAG repeat size, may be useful in predictive testing programs and for the design of clinical trials for persons at increased risk for HD (8). We retrospectively analyzed a Dutch cohort to compare our data with the Canadian data. In addition, we analyzed the relation of CAG repeat size and age at onset in sibs and in parent-child pairs of this cohort. We also analyzed the influence of the sex of the at-risk person and the sex of the parent and grandparent on the age at onset, and family factors relative to the repeat size. Subjects and Methods Subjects In the Netherlands the estimated prevalence of HD is 6.5:100,000, based on the number of living affected individuals recorded at the Leiden Roster for HD. In the roster, 3,115 individuals at 50% risk (children of known affected individuals) are registered. For this study we included those individuals with a CAG repeat size of at least 36, tested between 1993 and 2000. This amounts to 755 individuals, from 344 different HD families, of which 614 are affected and 141 are asymptomatic gene carriers. In all individuals we ascertained the age at onset or oldest age while still asymptomatic as obtained by retrospective studies of patient charts and records of predictive test applicants. The age at onset was determined as the youngest age of the subjective or objective start of neurologic (and psychiatric) complaints due to HD. The subjective start was usually several years before a neurologist or psychiatrist examined the patient. The oldest age while still asymptomatic was the age when the patient subjectively did not have neurologic and/or psychiatric complaints. Methods DNA analysis: DNA analysis for HD in the Netherlands is centralized in the DNA diagnostic section of the Department of Clinical Genetics in Leiden. DNA was extracted from leukocytes by standard procedures (42). Analysis of the CAG repeat was performed according to standard procedures, using end-labeling PCR primers that flank the CAG repeat alone. The CAG repeat size was assessed by electrophoresis on the Automated Laser Fluorescent DNA Sequencer (Pharmacia Biotech Benelux, Roosendaal, The Netherlands), using internal as well as external standards consisting of DNA samples with known CAG size. Data analysis: The cumulative probability of having onset of HD by a certain age was calculated for each year of age and CAG size including all individuals, both those asymptomatic at-risk and those affected, by use of Kaplan-Meier survival analysis. Differences between ages at onset of different patient groups were illustrated with Kaplan-Meier curves, and analyzed with a Cox proportional regression model extended with a frailty parameter, which was assumed to be gamma-distributed (21). This frailty parameter represents characteristics shared by the individuals from the same family pedigree. Such sharing, like the shared genetic makeup, leads to the possibility that the age at onset of individuals from the same pedigree is correlated. As such correlation violates the basic Cox regression model (which assumes that all patients are independent), the Cox model was extended with a frailty parameter, which takes such correlation into consideration. Differences between groups were analyzed using chi-square tests, and p = 0.05 was set as the criterion for statistical significance. Results Since there were no affected individuals with a CAG repeat length of less than 36 repeats, as described earlier (30,56) individuals with a CAG repeat length of 36 or more repeats were considered to be a cohort at risk, from birth to either neurologic or psychiatric onset or last contact or until death. Repeat sizes, age at onset/while still asymptomatic, numbers of affected and unaffected individuals, and numbers of individuals with juvenile (≤20 years) or late (≥60 years) age at onset of the Dutch HD cohort of 755 individuals ascertained are shown in Table 1. In the data set, most individuals (n = 614, 81%) were already affected, and 141 (19%) were not affected at the time of assessment (1993–2000). HD patients with juvenile onset show significantly larger repeat sizes than those with an age at onset over 60 years. In patients with juvenile onset the father was the affected parent in 88% of cases and the mother, in 13% of cases, while in patients with onset over 60 years the mother was the affected parent in 59% of cases and the father, in 41% of cases.TABLE 1: Characteristics of the Dutch Huntington disease cohortIn the total sample there was no significant difference between the median age at onset of 341 men and of 406 women (8 patients, sex unknown). In contrast the sex of the transmitting parent and grandparent does show a significant effect on the mean age at onset (Table 2). Transmission of the disease through the paternal and/or grandmaternal line decreases the mean age at onset.TABLE 2: Influence of the sex of the transmitting parent and grandparent on the mean age at onsetThe CAG repeat size is significantly longer if the gene is inherited through the paternal line (Table 3). No influences of the sex of the grandparents on the CAG repeat size can be seen, however.TABLE 3: Influence of the sex of the transmitting parent and grandparent on the mean CAG repeat sizeThe overall cumulative incidence of HD in our cohort is illustrated in Figure 1. At 40 years the probability of onset is 30%, and at 60 years, 80%.Fig. 1: Overall cumulative probability of the age at onset of Huntington disease, with associated 95% confidence bands, in 755 persons, 614 symptomatic and 141 asymptomatic, with at least 36 CAG repeats.The cumulative incidence of HD is given in Figure 2 for each number of CAG repeats studied. There was a significant increase in the probability of onset (p < 0.0001) with increasing CAG repeat length. Clearly, the median age at onset decreases (p < 0.0001) with increasing number of CAG repeats: the median ages are given in Table 4 and are illustrated in Figure 3. Only 50% of the HD carriers with ≤40 repeats will be affected at age 68 years, but this decreases to age 50 with 45 repeats, to age 34 with 50 repeats, and to age 23 with ≥53 repeats.TABLE 4: Median age at onset for each number of CAG repeatsFig. 2: Cumulative probability of the age at onset with confidence bands for each number of CAG repeats in 755 persons with at least 36 CAG repeats. There is a significant increase in the probability of onset (p < 0.0001) with increasing repeat length from ≤40 to ≥53. (Horizontal axis: age at onset in years; vertical axis: cumulative probability. For the number of cases per CAG repeat, see Table 4).Fig. 3: Median age at onset with associated 95% confidence bands of 755 persons with at least 36 CAG repeats decreases significantly (p < 0.0001) with increasing number of CAG repeats (≤40 to ≥53).The univariate and multivariate relative risks for number of CAG repeats, sex, and family are summarized in Table 5. There were 344 different families in the current data set: the number of cases ascertained per family varied from 1 to 51 with a median of 1, a mean of 2.2, and a standard deviation (SD) of 3.3. There were significant differences between the families with respect to age at onset (p < 0.0001): we assumed that the families had a random frailty, and the SD of the gamma-frailty distribution was estimated to be 0.71. This indicated that families differed with respect to the risk of age at onset.TABLE 5: Univariate and multivariate relative risk for number of CAG repeats, sex, and familyThe posterior relative risks per family compared to the average family are given in Figure 4. These relative risks suggest major differences between families with respect to age at onset; some families have almost 2.5 times more risk of disease onset at a certain age than the average family. This is illustrated by the following example: the members of family A and B have comparable mean numbers of CAG repeats (43.0 and 43.1, respectively), but the average age at onset/while still asymptomatic in family A is 61.8 years (5 affected of 7 members, as yet), whereas the average age at onset/while still asymptomatic in family B is 42.8 years (5 affected of 9 members, as yet).Fig. 4: Posterior relative risks (RR) of age at onset with associated 95% confidence bands (grey vertical bars) per family compared to the average family (posterior RR = 1, black horizontal line), suggesting major differences between families with respect to age at onset.Multivariately, the effect of the number of CAG repeats on age at onset remained highly significant (p < 0.0001). The differences between families also remained significant (p < 0.0001), although the differences were smaller when the CAG repeats were taken into consideration (SD of the frailty distribution decreased to 0.36). The correlation between the age at onset and the CAG repeat in this cohort is −0.74. There is a stronger correlation if HD is inherited through the paternal (−0.84) and grandpaternal (−0.77) than through the maternal (−0.64) and grandmaternal (−0.60) line (Table 6). The correlation between the CAG repeat and the age at onset in sibs and parent-child pairs were respectively −0.62 and −0.61.TABLE 6: Influence of the sex of the parent and grandparent on the correlation between CAG repeat size and age at onsetDiscussion Different studies of the correlation between the CAG repeat length and the age at onset have confirmed that not all genetic variation in HD can be attributed to the HD locus itself. The sex of the affected parent and grandparent has previously been described as a modifier of the age at onset (13,28,50,51,52). Our data confirm these observations, that is, that paternal inheritance leads to a lower age at onset than maternal inheritance and that the same applies to grandmaternal and grandpaternal inheritance when studied independently. However, when 2-generation inheritance is taken into account, the lowest age at onset is seen when HD is inherited through the father from the grandmother (39.8 yr), and the highest age at onset is seen in patients who inherited HD through the mother from the grandfather (46.1 yr) (see Table 2). While the paternal effect on the CAG repeat size (see Table 3) was described before (57), no grandpaternal effect has thus far been reported. The fact that this effect is opposite from the expectation suggests that mechanisms other than the CAG repeat size must influence the age at onset. Mechanisms such as meiotic instability of the CAG repeat in transmission through the male germline and multiple modifying genes or epigenetic changes in methylation of nucleic acid of the genome (imprinting) are potential causes of the grandparental effect on the age at onset (51,57). Although it has been reported (52) that the sex of the patients themselves also influenced the age at onset, we could not conclude this from our cohort, which comprised only CAG-tested individuals, recently diagnosed and therefore perhaps with a more accurately determined age at onset. Other modifiers of the age at onset have been proposed, for example, the normal repeat (1,14,63), a commonly occurring HD haplotype (37), the adjacent CCG repeat (4,5,55,65), a Δ2642 deletion codon (1,2,36,65), and apolipoprotein E with genotype ε2ε3 (25). Their influence on the age at onset is minor, however, and cannot fully explain the existing differences, therefore there must be other, yet unknown, modifiers of the age at onset. Our study confirms, for a large cohort ascertained uniformly and analyzed by standardized, accurate CAG repeat size measurement (8,34,39,56) that the CAG repeat size is the major determinant of age at onset in HD. The correlation between the age at onset and the CAG repeat in this cohort is −0.74, comparable to the −0.73 described earlier (8). Consistent with decreased age at onset (or increased repeat size), this correlation is even stronger if HD is inherited through the paternal (−0.84) than through the maternal line (−0.64) (see Table 6). Previous reports have shown that family factors influence the HD phenotype: the age at onset and other phenotypic factors such as the symptoms of persons within families (4,29,62). Dutch families differed with respect to the risk of age at onset (see Table 5): the age at onset of persons within the families was more similar than can be explained by the number of CAG repeats. An aggregation of the age at onset amongst sibs has been observed (54), especially with juvenile onset (19,20). We hypothesized that if environmental factors were principally responsible (17), one would expect a greater correlation for sibs than between parent and child, whereas, conversely, this would not be the case if genetic factors (other than the repeat length) were the predominant factor. Our analysis turns out to favor the latter: the correlation between the age at onset and the CAG repeat of parent-child pairs and sibs in our series was quite comparable, respectively −0.61 and −0.62. This result seems to point to greater influence of a shared genetic background than of environmental factors. Clearly, the development of accurate onset and survival curves has great value for genetic counseling. The number of years lived before the appearance of HD symptoms is an important variable in the decision-making processes of individuals who undergo the testing protocol (48). On the other hand, it might in some cases have a negative bearing on the hope of an asymptomatic carrier for a late age at onset of HD. These issues should therefore be discussed in pretest counseling sessions to learn if the asymptomatic test applicant would be informed of this additional information, especially given that our policy with regard to disclosing results is not to communicate the repeat length unless applicants ask for it (9). Furthermore, onset and survival curves might ultimately be helpful for the design of clinical trials (8). A significant extension of the age at onset beyond the median age might be a useful indicator of a therapeutic effect. We note that there is a significant difference between the median age at onset in the present Dutch study and the Canadian study (8) (Figure 5). For each repeat size the Dutch median age at onset is about 10 years later than the Canadian. This difference must be accounted for by a systematic measurement difference of 3 CAG repeats. However, this was not confirmed after exchanging samples between the 2 laboratories: test results showed minor differences in CAG repeat size in the DNA samples exchanged (differences, n = 36: mean: 0.47; median: 0; SD: 1.18; range: −2 to +3).Fig. 5: Median age at onset with associated 95% confidence bands (stippled lines) in the present Dutch cohort (black solid line) (n = 755) and the Canadian cohort (grey solid line) (n = 866) (reference 8, Brinkman et al, 1997) of persons with at least 36 CAG repeats, showing a significant difference of 10 years.The mean age at onset of the Dutch cohort is 44 years, which is lower than the range of 46.0–48.9 years described before (61). Because the group of asymptomatic carriers (19%) is underrepresented compared to symptomatic carriers, not due to systematic bias in the ascertainment of data, the age at onset could have been underestimated. This underrepresentation of asymptomatic carriers is also found in the Canadian study (23%), however (mean age at onset not described). Differences in the number of asymptomatic carriers cannot explain the existing differences, because the asymptomatic carriers in both cohorts are younger than the symptomatic carriers and therefore the median age at onset will almost not be influenced. In the Canadian study, telephone interviews were made to obtain the age at onset or the youngest age while still asymptomatic. Possibly there is a perceptual difference between consults and telephone interviews, which may lead to a difference in the reported age at onset for symptomatic individuals. A more likely explanation might be a difference in ascertainment of the age at onset. Ascertainment of the age at onset has been proved to be difficult. There is a known delay between the age at onset and age at diagnosis of 8 years (32). It is worth noting however, especially considering the statistical evidence for greater impact of secondary genetic factors than environmental factors, the different genetic background of the relatively homogenous Dutch cohort and the Canadian cohort, which was reported to consist of European, Asian, black South African, Arabian, and South American families. These secondary genetic factors may significantly influence the age at onset. In conclusion, for the proper comparison and clinical use of epidemiologic studies in HD, it is vital to cross-validate repeat size estimates and to take into account as yet undefined genetic factors in the background population, which contribute to the age at onset. When applied in this perspective, knowledge about the 95% confidence interval of the median age at onset given the CAG repeat size in a specific population might be quite helpful by providing estimates of symptom-free survival to individuals participating in a predictive testing program and helping them plan for the future. Summary As previously described, the age at onset of Huntington disease (HD) ranges from 2 to 80 years, with a mean between 46.0 and 48.9 years. The number of repeats, in the causal CAG repeat expansion, is inversely related to the age at onset and accounts for 50%–77% of the variation in age at onset. We analyzed a Dutch cohort of 755 individuals retrospectively to assess the probability of onset for any given CAG repeat. The repeat size is the major determinant of age at onset, with a correlation of −0.74, stronger (−0.84) in paternal than in maternal inheritance (−0.64), consistent with increased repeat expansion and stronger anticipation in the paternal line. The age at onset within families was more similar than could be explained by the resemblance of the repeat size of persons in the same family. We hypothesized that if environmental factors were principally responsible for this familial aggregation, one would expect a greater correlation for sibs than for parents and children. This was not found to be the case. These observations suggest that genetic factors may play a greater role in the onset of HD than a shared environment. Finally, we discuss several explanations for the fact that the Dutch median age at onset for all expanded repeat sizes studied is significantly later, by about 10 years, than that found in a Canadian study. Acknowledgments Dr MR Hayden and Dr RR Brinkman from the Department of Medical Genetics, Centre for Molecular Medicine and Therapeutics, University of British Columbia, Vancouver, Canada, are acknowledged for their kind cooperation in exchanging samples for CAG repeat sizing.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.