Longitudinal Cognitive and Motor Changes Among Presymptomatic Huntington Disease Gene Carriers FREE

HUNTINGTON disease (HD) is an autosomal-dominant, neurodegenerative disorder characterized neuropathologically by the loss of medium-sized spiny neurons in the neostriatum and clinically by a triad of progressive motor, cognitive, and emotional symptoms. The HD phenotype is caused by an increased number of triplet (CAG) repeats in the 5‘-translated region of the HD gene on chromosome 4.¹ Typically, individuals with HD have 38 or more CAG repeats. A negative correlation between the number of CAG repeats and age at onset of HD has been found, with individuals with juvenile HD having the largest number of repeats (>55).^{1 - 2} The clinical implication of an allele with 32 to 37 repeats is ambiguous. These repeat sizes are often considered intermediate alleles with variable penetrance.

Great variation is seen in the age at onset of clinical symptoms, the initial symptom presentation, the clinical course of HD, and the duration of illness.^{3 - 5} The development of manifest HD occurs as early as childhood and as late as 80 years of age. The average age at onset is late in the fourth decade,^{6 - 7} and, as a result, the individual who has a parent with HD has many years of at-risk status.

Results of previous studies^{8 - 9} show that motor changes, including extraocular movements, mild chorea, brisk muscle stretch reflexes, and diminished rapid alternating movements, are the most consistent early findings of manifest HD. Cognitive decline, with progression from early changes in information processing, cognitive inflexibility, and memory retrieval to more severe and widespread abnormalities later in the disease course, has been demonstrated as part of the HD phenotype.¹⁰ Early changes in executive functions,^{11 - 13} visuospatial abilities,^{14 - 15} cognitive speed or efficiency,^{16 - 17} sensorimotor function,^{16 ,18} and olfactory functioning¹⁹ and recognition²⁰ have been reported.

Determination of the exact age at symptom onset is frequently difficult, and the categorization of an individual as affected or unaffected can be equivocal. Although the actual sequence of symptom appearance can be variable, the diagnosis of HD in at-risk individuals is usually made after the onset of involuntary, choreiform movements.^{6 ,21} Conflicting results have been found regarding neuropsychologic impairment and cognitive decline in otherwise asymptomatic individuals. Deficits reported in presymptomatic individuals at risk for HD include attentional learning and planning functions,²² memory,^{23 - 24} visuospatial tasks,^{23 ,25 - 26} performance of visuomotor skills,²⁷ language tasks,²⁸ tactile-motor coordination,²⁹ and stimulus perception and encoding.^{30 - 31} Others have reported no difference between presymptomatic gene carriers (PSGCs) and non–gene carriers (NGCs) for sustained attention and mental processing speed,^{32 - 33} verbal memory and learning,^{23 ,32 ,34} visuospatial and motor task performance,^{32 - 33 ,35} and intelligence subtest patterns.^{32 ,36 - 38}

In previous studies,³⁹ cross-sectional differences in cognitive function for PSGCs as measured by the Wechsler Adult Intelligence Scale–Revised (WAIS-R) were observed. Participants who were PSGCs showed a tendency to score lower on all subscales of the WAIS-R test, with the digit symbol and picture arrangement subscales reaching statistical significance. In addition, PSGCs also had abnormal neurologic findings for saccades velocity and muscle stretch reflexes.⁴⁰ In contrast, PSGCs did not have significantly more chorea than NGCs, confirming that this sample was indeed clinically presymptomatic. The objective of the present study was to reexamine a subset of the original study population to determine whether longitudinal decline in cognitive or motor function could be detected among a sample of PSGCs who maintained presymptomatic status throughout follow-up and to estimate whether longer follow-up using the entire sample would be useful.

In the original study, a sample of individuals at risk for HD was recruited through the National Huntington Disease Research Roster for Patients and Families at the Indiana University School of Medicine, Indianapolis. At-risk individuals reporting themselves as asymptomatic and without previous diagnosis of HD were invited to participate. Details regarding patient recruitment can be found in the study by Foroud et al.³⁹ A total of 45 participants from this original study, regardless of HD status, agreed to participate and were examined an average of 3.7 years after their initial study visit. Participants were informed of the risks and benefits involved in the study, and informed consent was obtained.

Two of the 45 participants were eliminated before data analysis. One of these participants exhibited sufficient neurologic symptoms at the follow-up examination to warrant a clinical diagnosis of HD. Therefore, this individual could not be considered clinically presymptomatic and was removed from further data analysis. A second participant was eliminated from the analysis because his DNA did not amplify properly, thus preventing gene status determination. Demographic information from the 43 remaining study participants is summarized in Table 1. No significant differences between PSGCs and NGCs were found except for the expected expanded allele size among PSGCs (P<.001).

DNA samples were analyzed with a polymerase chain reaction–based diagnostic screen as previously described.³⁹ Briefly, the CAG repeat implicated in the disease phenotype was amplified, and the resulting alleles were compared with previously sequenced cosmids having 18 and 48 CAG repeats. Each gel was interpreted independently by 2 individuals who were unaware of the participants' neurologic and neuropsychologic test results.

There is uncertainty regarding disease outcome for allele sizes occurring in the intermediate size range. However, we chose a conservative allele classification scheme, with individuals with 2 alleles having less than 32 CAG repeats classified as NGCs (n=31) and those with 1 allele having 38 or more CAG repeats categorized as HD gene carriers (n=12). There were no participants with an allele in the range between 32 and 37 CAG repeats. Gene testing was performed anonymously, and the results were not provided to the participants.

To measure cognitive function, each participant was examined by standard administration procedures on several verbal and performance subscales of the WAIS-R. Verbal subscales included digit span, vocabulary, and arithmetic. Performance subscales included picture arrangement, block design, and digit symbol.

A quantified neurologic examination adapted from previously described protocols^{6 ,37} was administered by a board-certified neurologist with specialization in movement disorders (E.S.). These tests measured extraocular movements, gait and stability, chorea, dystonia, parkinsonism, tremor, muscle stretch reflexes, and cerebellar function. Ratings were assigned as 0 (normal), 1 (possibly abnormal), or 2 (abnormal) for 29 neurologic tests.⁴⁰ Consistent with findings of previous studies,^{39 - 40} participants were classified as having manifest HD using criteria designed to approximate those used in clinical practice. Individuals with definite chorea, in the absence of other possible causes (eg, hyperthyroidism), were classified as having manifest HD. Individuals with an expanded CAG repeat (>37), without sufficient abnormalities to be considered manifest HD, were termed PSGCs. The neurologic examination was designed to detect early motor features of HD; thus, an attempt was made to maximize the sensitivity of the ratings, even at the expense of some specificity.⁴⁰ Because neuropsychologic testing was performed separately, the neurologist did not ascertain specific information for evaluation of the cognitive or emotional status of the participant. In addition, the neurologist was unaware of the previous and current neuropsychologic and gene testing results.

One-way repeated-measures analysis of variance was used to test for differences between PSGCs and NGCs in longitudinal cognitive measures (repeated measures) evaluated at the initial and follow-up examinations. Raw scores for the WAIS-R subscales were used in this analysis to prevent the possible loss of sensitivity that may result from decade-related adjustments in published tables.

Although data for 29 variables were collected as part of the quantified neurologic examination, only those variables for which at least 1 PSGC had a score greater than 0 at the follow-up examination were included in the statistical analysis (17 of 29 variables). No PSGC had abnormalities in any of the following variables: extraocular range, saccades performance and latency, eye closure motor impersistence, dystonia of the trunk, bradykinesia of the body, gait and station stability, plantar reflex, resting tremor of the head and neck, and resting tremor of the extremities bilaterally and unilaterally. For the remaining 17 quantified neurologic variables, the Fisher exact test was used to compare PSGC and NGC groups for the longitudinal measures.

Confirmatory factor analysis, using the principle components method, was used separately in the PSGC and NGC samples to reduce the number of relevant items to a subset of factors. Factor analysis was used to test for evidence of a single underlying determinant associated with HD resulting in the observed longitudinal changes. The psychomotor, oculomotor, and motor function measures sensitive to progression were loaded into the factor analysis. To establish which factors were significant, only those with eigenvalues greater than 1.00 were considered.

In addition to the longitudinal analyses, the follow-up examination data were evaluated cross sectionally. For this analysis, the individual raw scores on each of the subscales of the WAIS-R were adjusted for age from published tables.⁴¹ One-tailed Student t tests were performed on the neuropsychologic scales to determine whether the PSGC group demonstrated significantly greater abnormality than the NGC group at the follow-up examination. The Fisher exact test was used to compare quantified neurologic variables between the 2 groups.

In a previous comparison of individuals at risk for HD, significant differences in psychomotor, oculomotor, and functional motor performance were found between PSGCs and NGCs. When a local subset of the original sample (12 PSGCs and 31 NGCs) was compared longitudinally, significantly more rapid longitudinal decline among PSGCs was found in psychomotor speed (digit symbol subscale, P<.05) (Table 2), oculomotor function (optokinetic nystagmus, P<.01) (Table 3), and cerebellar motor function (rapid alternating movements, P<.005) (Table 3) than for NGCs. In addition to the significant decline in digit symbol subscale performance, longitudinal change in the arithmetic subscale (P<.10) tended to be greater in PSGCs than in NGCs.

In a secondary set of analyses, confirmatory factor analysis was used to test for evidence of a single underlying factor that resulted in the significant longitudinal differences in the 3 measures sensitive to progression: the digit symbol subscale of the WAIS-R and the 2 neurologic variables—optokinetic nystagmus and rapid alternating movements. If these observed longitudinal changes are caused by 1 process associated with HD, they would be expected to contribute to 1 factor for PSGCs and have a different pattern in NGCs. To test this hypothesis, factor analysis was run separately for PSGCs and NGCs. One factor (eigenvalue=1.79) was extracted for PSGCs. This supported the idea that 1 process is responsible for the longitudinal change in PSGCs. All 3 measures sensitive to progression—the digit symbol subscale, optokinetic nystagmus, and rapid alternating movements—contributed to this factor (Table 4). The NGCs had 2 factors with eigenvalues greater than 1.00 (eigenvalues=1.17 and 1.06). This supported the hypothesis that different processes are responsible for their longitudinal changes. For NGCs, the first factor is composed of similar loadings of optokinetic nystagmus and rapid alternating movements, whereas the digit symbol subscale was weighted in the opposite direction. In the second factor, rapid alternating movements loads strongly, whereas optokinetic nystagmus loads in the opposite direction (Table 4).

Similar to their first evaluation 3.7 years earlier, PSGCs demonstrated significantly worse performance on the digit symbol subscale than NGCs (P<.05) (Table 2). Of the 17 neurologic variables used in the analysis, PSGCs demonstrated a significantly higher rate of possible or definite abnormalities than did NGCs for 4 of the neurologic tests: optokinetic nystagmus (P<.001), rapid alternating movements (P<.005), overall ocular saccades (P<.05), and chorea of the trunk (P<.05) (Table 3).

This study, currently the largest published longitudinal study of clinically presymptomatic individuals at risk for HD, to our knowledge,^{32 ,37} provides significant evidence for longitudinal decline in psychomotor, oculomotor, and motor function before the onset of manifest HD. The PSGC group demonstrated greater decline in psychomotor speed, optokinetic nystagmus, and rapid alternating movements an average of 3.7 years after their first evaluation. Although PSGCs demonstrated a tendency toward a higher rate of abnormalities than did NGCs, after elimination of the 1 patient noted previously, none of the gene carriers included in this study exhibited sufficient neurologic symptoms at the follow-up examination to warrant a clinical diagnosis of HD. Our studies, and those of others, reveal subtle motor, cognitive, and behavioral changes before the onset of unequivocal chorea, suggesting that decline occurs earlier than the first clinical diagnosis.

Results of the factor analysis support the hypothesis that a single underlying mechanism may be responsible for the decline in the psychomotor speed, optokinetic nystagmus, and rapid alternating movements seen in PSGCs. For PSGCs, all 3 variables contributed with similar weights into 1 factor. Conversely, this same relationship was not observed for NGCs. Two factors are needed to explain the variance in these measures in NGCs, suggesting that different mechanisms may be responsible for the longitudinal change in these variables among NGCs compared with PSGCs.

Cross-sectional analyses of the follow-up data for at-risk individuals confirmed the results of a previous study,^{39 - 40} which suggested that individuals who inherited the HD gene have oculomotor, functional motor, and psychomotor deficits that precede the onset of definite abnormal neurologic functioning. Results of other studies demonstrate that these changes are the earliest features of manifest HD. These include changes in extraocular movements,^{8 - 9} motor function (including diminished rapid alternating movements, mild chorea, and brisk muscle stretch reflexes),^{8 - 9} and psychomotor speed.^{42 - 43} Our results are consistent with those previously reported, indicating that gene carriers who can be considered presymptomatic on clinical examination demonstrate greater abnormality for the oculomotor variables of optokinetic nystagmus and overall ocular movements and for the motor function variables of rapid alternating movement and chorea of the trunk. Although the changes we found in PSGCs represent a statistical increase in possible abnormalities on examination compared with NGCs, no PSGC had sufficient abnormalities to be classified as having manifest HD. Chorea of the extremities would be expected to be an earlier sign of HD than chorea of the trunk; however, chorea of the extremities was not significantly different between PSGCs and NGCs. In our study, a large proportion of PSGCs (42%) and NGCs (23%) were rated as having possible chorea of the extremities, thus limiting the power to discriminate between them. This was most likely the result of the neurologist's attempt to maximize the sensitivity of the ratings to avoid classifying symptomatic individuals as presymptomatic.

We show further that longitudinal changes are greater in the digit symbol subscale than in other measures of the WAIS-R. This suggests that presymptomatic individuals carrying the HD allele may have the greatest difficulties with the ability to learn new tasks, visual motor dexterity, and rapidity of task completion⁴⁴ early in disease progression. Asymptomatic HD gene carriers have been shown to have changes in caudate metabolism⁴⁵ and volume.⁴⁶ In addition, animals and humans have demonstrated cognitive impairment after caudate lesions.⁴⁷ Therefore, it is likely that the observed early cognitive changes in HD are related to changes in caudate function.

Our study avoided some of the difficulties that have plagued other studies of at-risk individuals. The psychosocial issues related to at-risk status that may affect cognitive performance, even among NGCs, have been experienced by all participants in our sample because we examined only individuals at risk for HD by virtue of having an affected parent. Therefore, the at-risk NGCs are an excellent control sample. Our gene carriers were all clinically presymptomatic at the time of examination, with an average age of 43.1 years, near the typical age at onset. Assuming that the rate of presymptomatic, longitudinal change is likely to be greatest just before onset of clinically manifest HD, the power to detect change in this sample of PSGCs should be improved compared with that in a younger sample of PSGCs. Furthermore, because all individuals participating in the initial study were adult, presymptomatic individuals, we would not have expected, and we did not have, any participants with a very large number of CAG repeats (>50). Rather, our restricted range of expanded CAG repeats (38-48) suggests that this is a representative sample of patients with adult-onset HD, and the longitudinal decline that we detected before disease onset may be the typical HD prodrome.

In this study, criteria for manifest HD were used that would approximate those used in clinical practice. Using these criteria, subtle changes have been found among at-risk individuals who do not fulfill clinical criteria for manifest HD. The present study expands on these findings and suggests that these abnormalities progress before a clinical diagnosis of HD can be established. These results demonstrate the importance of including neuropsychologic tests in the study of therapies designed to prevent or slow the clinical progression of disease. Given the subtlety of the neuropsychologic and motor findings, a longitudinal design using serial assessment of a large, well-characterized sample of at-risk individuals, such as ours, will be necessary to provide important data regarding neuropsychologic and motor symptom variability and progression. Further longitudinal analyses and follow-up of a larger subset of our original at-risk sample will allow us to evaluate the course of the longitudinal changes reported herein. A more complete understanding of the physiologic and cognitive changes that occur early in the course of HD will help better define the disease process and will be important for future diagnostic and therapeutic strategies. Such an understanding will assist in assessing the value of motor and cognitive assessments as relevant outcome measures in controlled clinical trials aimed at slowing the underlying neurodegenerative changes in HD.

Accepted for publication August 25, 1998.

This study was supported in part by grants R01-AG-08918, M01-RR-750, and PHS N01-NS-2326 from the Public Health Service, Washington, DC, and by Medical Genetics Training Grant Fellowship NICHD-T3HD07373 from the National Institute of Child Health and Human Development, Bethesda, Md (Ms Kirkwood).

We thank all participants and members of the National Huntington Disease Research Roster for Patients and Families for their efforts, and Dawn Kleindorfer and Jacqueline Gray for their assistance, Indiana University School of Medicine, Indianapolis.

Reprints: Tatiana Foroud, PhD, Department of Medical and Molecular Genetics, Indiana University School of Medicine, 975 W Walnut St, Indianapolis, IN 46202 (e-mail: tforoud@medgen.iupui.edu).