Huntington’s Disease and Employment: The Relative Contributions of Cognitive and Motor Decline to the Decision to Leave Work

Abstract

Background:

Huntington’s disease (HD) presents with motor, cognitive, and behavioral symptoms that impair functional capacity and the ability to maintain employment. The relative contribution of cognitive decline to work disability remains controversial.

Objective:

To evaluate the association of cognitive decline, compared with motor decline, with the decision to leave work.

Methods:

Data from the Enroll-HD observational study were analyzed. The correlation of age of cognitive symptom onset and age of motor symptom onset with age at leaving work was assessed. The association of the Stroop Color Naming Test (SCNT) cognitive assessment and the Total Motor Score (TMS) assessment (reverse scored) with the Total Functional Capacity (TFC) assessment was also assessed.

Results:

For every year delay in cognitive symptom onset, there was a 0.806 year increase in age at leaving work (SE = 0.030, p < 0.001, adj-R² = 0.628). For every year delay of motor symptom onset, there was a 0.814 year increase in age at leaving work (SE = 0.031, p < 0.001, adj-R² = 0.603). For every additional correct SCNT response given and for every unit increase in TMS, there was a 0.105 unit increase (SE = 0.006, p < 0.001, adj-R² = 0.315) and a 0.104 unit decrease in TFC (SE = 0.003, p < 0.001, adj-R² = 0.640), respectively.

Conclusions:

Cognitive symptoms have a significant association, comparable to that of motor symptoms, with occupational functioning and the decision to leave work, suggesting that development of therapies for both cognitive and motor decline would be important for allowing people with HD to remain in the workforce longer.

Keywords

Cognition disability evaluation drug development Huntington’s disease symptoms unemployment

INTRODUCTION

Huntington’s disease (HD) is an autosomal dominant neurodegenerative disorder that results in the progressive loss of nerve cells in the brain. HD has a multifactorial impact, progressively affecting motor, cognitive, and behavioral function. The symptoms associated with HD affect functional capacity: the ability to perform daily activities including completing chores, managing finances, and engaging in employment. The average age of symptom onset is 30–50 years of age [1]. A challenging consequence of HD is the inability to maintain gainful employment. Needing to leave one’s job can have a substantial impact on quality of life. A recent review of the financial, physical, and mental health status following involuntary job loss in the general population lists adverse effects not only on individuals, but also on family members [2]. A study at the University of Iowa investigating functional impairment in individuals with pre-manifest HD (n = 265) found occupational decline was the most prevalent functional impairment reported. Sixty-five percent of participants reported at least some reduction in ability to engage in their typical work and 37% reported being unable to engage in any gainful employment [3].

Recent research aims to clarify the relationship between the multifaceted symptoms of HD and the decision or need to leave work; motor, cognitive, and behavioral symptoms all appear to contribute. A longitudinal cohort study (N = 208) by Ross et al. at Johns Hopkins University consisting of clinic patients at any stage in disease progression found that the participants who discontinued working during the time between annual visits (n = 35) had a significant worsening of both cognitive symptoms and motor symptoms in the time between visits, as determined by measuring change in cognitive and motor assessment scores from the last visit that occurred while still working to the first unemployed visit [4]. Ross et al. suggested that, despite small sample size, these data imply that cognitive and motor impairment have “closely intertwined” associations with functional and occupational disability [4]. A prospective study by Marder et al. (N = 960) followed subjects with HD for 18 months and demonstrated that individuals with higher baseline cognitive scores had slower decline in functional capacity, assessed via the TFC, further demonstrating a potential influence of cognitive impairment on functional decline [5]. Different conclusions, however, were drawn from a recent study by Goh et al., which used the Enroll-HD observational study database to explore work patterns among a cross-sectional cohort of 656 pre-manifest individuals with HD. This study found that there were increased odds ratios for work impairment among those with increased motor symptoms, although this seemed to be mediated by anxiety. This same study found cognitive symptoms were not significantly associated with work disability [6]. By contrast, a recent study of Social Security Disability Insurance recipients (not specific to HD) participating in work support interventions found mild cognitive deficits to significantly predict work ability. Participants in this study had psychiatric disorders but relatively normal cognitive functioning [7]. Overall, these analyses suggest that both cognitive and motor decline may be associated with functional capacity and the ability to work.

Assessments have been developed and validated to measure the cognitive, motor, and functional impairments associated with HD. The dataset used in this study included data from multiple different cognitive assessments; however, there is no established way to create a composite cognitive assessment score from the multiple different assessments, and so we selected the number of correct Stroop Color Naming Test (SCNT) responses as the measure of cognitive capacity. The SCNT is used in many HD studies and measures processing speed [8, 9]. Additionally, the SCNT is highly sensitive to pre-symptomatic decline [8]. In one study, Stout et al. compared 29 cognitive assessments and found the SCNT to be one of the two most sensitive assessments, in addition to the Symbol-Digit Modality Test [8]. We thus chose the SCNT to be able to capture cognitive impairment in a range of subjects, including those who were moving from pre-manifest to early manifest HD. During the SCNT, subjects report verbally, in order, the colors on a page during a 45 second period, and the number of correct responses is recorded. Assessment of motor symptoms is most commonly performed with the Unified Huntington Disease Rating Scale (UHDRS) Total Motor Score (TMS). The TMS has undergone extensive validity testing to confirm internal and inter-rater consistency [10]. The TMS is a categorical scale from 0 to 124, with higher scores representing greater impairment. The TMS measures oculomotor impairment, bradykinesia and fine motor impairment, rigidity, dystonia, and chorea [5, 10]. Functional capacity is measured by the UHDRS Total Functional Capacity (TFC). The TFC has been extensively validated in clinical trials and is the most commonly used measure of function in HD. The TFC is a categorical scale from 0–13, with higher scores representing less impairment, and measures ability to carry out gainful employment, manage finances, complete domestic chores, perform activities of daily living, and live independently [10].

Policy and drug development has allotted disproportionally more attention to motor symptoms than to cognitive symptoms, despite mounting evidence of cognitive symptoms’ contribution to quality of life and functional decline. This may be because motor symptoms are an easier target and motor outcomes can be more easily measured. Similarly, a clinical diagnosis of HD is currently made based solely on the presence of motor symptoms [11]. Before 2016, this focus on motor symptoms was associated with delays for individuals who were trying to receive disability insurance but who needed to stop or reduce work due to cognitive symptoms prior to motor symptom onset. In response to these and other related difficulties, the HD Parity Act was proposed. One tenet of this act called for revisions of the Social Security Administration’s (SSA) medical guidelines for determining disability. While the act itself has not passed, pressure from advocates ultimately led to these revisions being updated in the guidelines in June 2016 and put into practice in September 2016 [12, 13]. Similar success has not occurred in drug development [14]. Currently, there are only two FDA approved drugs indicated for HD, tetrabenazine and deutetrabenzine, both of which target motor symptoms [15]. Reports from the FDA’s 2015 patient-focused drug development initiative survey (N = 110) revealed that over two thirds of participants felt that cognitive impairment significantly affected daily life and found that participants believed existing treatments “do not adequately manage their most disabling symptoms” [16]. If a significant association of cognitive impairment with occupational decline can be demonstrated, this may continue to encourage research efforts to target cognitive decline.

This study’s aims were 1) to evaluate the association of cognitive symptoms with the decision to leave work and to a decline in occupational functioning; and 2) to compare the strength of this association to the strength of the association of motor symptoms with the decision to leave work and decline in occupational functioning. The primary objective was to assess if age of cognitive symptom onset and age of motor symptom onset were significantly associated with age at leaving work, as determined by assessing the associations of age of cognitive symptom onset and age of motor symptom onset with age at leaving work. We also explored the association of cognitive and motor assessment scores with functional capacity assessment scores, as determined by assessing the associations of the SCNT scores and TMS scores with TFC scores. It is important to note that, although this study focuses on HD’s cognitive and motor symptoms, people with HD also have behavioral symptoms, even when they are pre-manifest. These behavioral symptoms, which include but are not limited to depression and apathy, are not part of this investigation as they can be more difficult to characterize than HD’s cognitive and motor symptoms: they do not follow a linear pattern of progression and, as noted by Goh et al., are difficult to differentiate from the same behavioral conditions commonly found in the general public [6].

MATERIALS AND METHODS

Study design

This descriptive, correlational study used data from the 2015 periodic data cut (N = 4,146) of the Enroll-HD observational study (https://www.enroll-hd.org). Enroll-HD is an international longitudinal study that recruits individuals with and at risk for HD and caregivers and controls without HD. Enroll-HD follows these subjects for up to ten years, with the goal of a better understanding of symptom progression over time. Subjects attend annual visits during which they complete a variety of assessments and questionnaires to clarify understanding of the natural history and progression of HD [17, 18].

Sample selection

For all linear regressions, the sample was limited to only gene positive subjects (subjects with a CAG expansion of≥36 repeats) who had a recorded age of both cognitive and motor symptom onset, a recorded age at leaving work, and recorded TMS, SCNT, and TFC scores. Data from subjects who qualified as having juvenile HD (age of motor symptom onset as less than 20 years of age) were excluded, because juvenile HD generally has faster disease progression and presents differently than adult onset HD [19]. Data from individuals who had a recorded age of less than 18 were also excluded, because these subjects are likely to be in school and therefore unlikely to be employed. Similarly, data from subjects older than 64 years old were excluded, because qualification for retirement at age 65 may serve as a confounder when analyzing unemployment trends (Table 1). For the attempted survival analyses, data was limited to all gene positive individuals (as defined above) who had a recorded value for the age of motor and cognitive symptom onset, and who met the adult onset and age requirements described above.

Table 1

Sample set narrowing process

Variable	Number excluded	Sample size after exclusion
Original data set	–––	4,147
Gene positive only 1	972	3,175
Remove age “<18”	4	3,171
Remove age 65+	449	2,722
Remove juvenile HD	25	2,697
Age of unemployment present	1,623	1,074
Age of cognitive symptom onset present	380	694
Age of motor symptom onset present	20	674
SCNT score present	26	648
TMS score present	4	644
TFC score present	2	642

¹ Refers to subjects with a CAG expansion of≥36 repeats.

Determining age at leaving work

During the Enroll-HD visits, the “Variable Items” form collects data on work status (see Supplementary Material). The subject is asked to report whether they are full-time employed, part-time employed, self employed, or not employed. If not employed, they are asked to indicate the reason. Age at leaving work data is collected on those that indicate the reason as either “retirement” or “unemployed.”

Cognitive, motor, and functional capacity assessment descriptions

Cognitive impairment: SCNT

Cognitive capacity was assessed via number of correct SCNT responses. During this assessment, the subject was provided with a sheet containing eleven rows, one of which was a practice row. Each row contained ten colored boxes, and each box was one of three colors: blue, red, and green. The subject verbally reported in order from left to right and top to bottom the colors observed. The number of correct, incorrect, and self-corrected responses given over 45 seconds was recorded [9].

Motor impairment: TMS

Motor impairment was assessed via the TMS. The TMS ranges from 0 – 124 points, with a higher score indicating more motor symptoms. It was developed to measure HD’s motor symptoms, including oculomotor impairment (0 – 24 points), bradykinesia and fine motor impairment (0 – 44 points), rigidity (0 – 8 points), dystonia (0 – 20 points), and chorea (0 – 28 points) [5].

Functional capacity: TFC

The TFC assessment, a component of the UHDRS, was used as an indicator of ability to work. Question number 1 on the TFC is whether the subject is able to engage in gainful employment; possible responses are: “unable” (0 points), “marginal work only” (1 point), “reduced capacity for usual job” (2 points), and “normal” (3 points) [5]. In addition to assessing ability to engage in gainful employment, the TFC measured ability to manage finances, complete domestic chores, and carry out activities of daily living. TFC scores were categorical and ranged from 0 to 13, with higher scores indicating greater functional ability. [5].

Statistical methods

Relationship between age of symptom onset and age at leaving work

Regression analyses were performed using the White’s sandwich estimator in R [20]. Two univariate models (Fig. 1) and one multivariate model (Table 2) were fit to evaluate whether self-reported age of cognitive symptom onset and self-reported age of motor symptom onset were significant predictors of age at stopping work. The p-value, adjusted R-squared value, and robust standard errors were calculated. Figure 1 presents scatter plots of self-reported age of cognitive symptom onset (Fig. 1A) and self-reported age of motor symptom onset (Fig. 1B) on the x-axis and age at leaving work on the y-axis. The multivariate regression controlled for sex, marital status, educational attainment, self-identified race or ethnicity, region of residence, type of residence, and apathy. Presence of apathy was defined as having a severity score of two or greater on the Problem Behavior Assessment (PBA) apathy subscale. Per convention, the reference values for these confounding factors were male, single, primary school education, Caucasian, North America, rural, and non-apathetic respectively.

Table 2

Description of the effects of the variables held constant during the multivariate regression on age at leaving work

Variable	n (%)	Delay in leaving work (y) 1	Standard Error 2	p-value
Age of Cognitive Symptom Onset		0.476	0.075	<0.001
Age of Motor Symptom Onset		0.395	0.079	<0.001
Sex = Male	298 (46%)	2.280	0.596	<0.001
Region 3 = Latin America	2 (0.3%)	7.682	3.005	0.011
Region = Europe	349 (54%)	–0.737	0.688	0.284
Region = Australasia	17 (3%)	–0.078	1.144	0.945
Marital Status = Married	329 (51%)	1.392	0.735	0.058
Marital Status = Partnership	38 (6%)	1.901	1.156	0.100
Marital Status = Divorced	111 (17%)	1.118	0.892	0.210
Marital Status = Widowed	12 (2%)	1.825	2.278	0.423
Marital Status = Legally Separated	18 (3%)	1.592	2.332	0.495
Race or ethnicity 4 = Other	8 (1%)	3.057	1.567	0.051
Race or ethnicity = Asian	4 (1%)	0.543	1.280	0.672
Race or ethnicity = American Indian	4 (1%)	0.034	0.849	0.968
Race or ethnicity = Mixed	3 (0.5%)	–5.303	1.495	<0.001
Race or ethnicity = Hispanic or Latino	12 (2%)	3.066	2.824	0.278
Race or ethnicity = African American	11 (2%)	–4.582	3.400	0.178
Education = Doctoral Degree	9 (1%)	4.964	3.154	0.116
Education = First Stages of Tertiary Education	136 (21%)	1.045	1.630	0.521
Education = Lower Secondary	125 (19%)	0.720	1.487	0.628
Education = Upper Secondary	224 (35%)	0.237	1.595	0.882
Education = Specific Occupational Training	130 (20%)	–0.491	1.741	0.778
Residence = Village	113 (18%)	–1.776	1.681	0.291
Residence = Town	274 (43%)	–0.830	1.546	0.592
Residence = City	233 (36%)	–0.715	1.549	0.644
Apathy = present	79 (12%)	0.696	0.645	0.281

¹ Positive scores represent an association with increased age at leaving work.

²The standard errors reported are robust standard errors, found using the White’s sandwich estimator [20].

³ Refers to the region of the world in which the subject resides.

⁴ Refers to the self-reported response a subject provided when asked the race the subject identified with.

Fig. 1.

Age of cognitive symptom onset and age of motor symptom onset have similar directly proportional linear relationships with age at leaving work. (A) For every year delay in age of cognitive symptom onset, there was a 0.806 year increase in age at leaving work (SE = 0.030, p < 0.001, adj-R² = 0.628), without controlling for possible confounding variables. (B) For every year delay in age of motor symptom onset, there was a 0.814 year increase in age at leaving work (SE = 0.031, p < 0.001, adj-R² = 0.603), without controlling for possible confounding variables.

A survival analysis was attempted using the semi-parametric Cox-proportion hazard method to further evaluate the relationship between symptom age of onset and age at leaving work. This analysis would have minimized the bias associated with a linear regression, which necessitated the exclusion of individuals who were still employed. The analysis was attempted by right censoring with current age for subjects who had not yet left work, and with t = 0 set to time at which cognitive or motor symptom onset began [21, 22]. As with the linear regression model, this model controlled for sex, marital status, education level, self-identified race or ethnicity, region of residence, type of residence, and presence of apathy. This survival analysis could not be performed, however, because, of the subjects included in the survival analysis, 31% left work prior to cognitive symptom onset and 23% left work prior to motor symptom onset.

Relationship between cognitive or motor scores and functional scores

Two univariate models (Fig. 2) and one multivariate model (Table 3) were fit to determine whether SCNT score and TMS score were significant predictors of TFC score. The p-value, adjusted R-squared value, and robust standard errors were reported. Figure 2 presents scatter plots of SCNT score (Fig. 2A) and TMS score (Fig. 2B) on the x-axis and TFC score on the y-axis.

Fig. 2.

The Total Functional Capacity (TFC) assessment has a direct relationship with the Stroop Color Naming Test (SCNT) and an inverse relationship with the Total Motor Score (TMS). (A) For every additional correct response given durnig the SCNT, there was a corresponding 0.105 unit increase in TFC (SE = 0.006, p < 0.001, adj-R² = 0.315), without controlling for possible confounding variables. (B) For every unit increase in TMS there was a coresponding 0.104 unit decrease in TFC (SE = 0.003, p < 0.001, adj-R² = 0.640), without controlling for possible confounding variables.

Table 3

Multivariate regression of the relationship between TFC and both SCNT and TMS

	Increase in TFC (points) 1	Standard Error 2	p-value
SCNT (number correct)	0.037	0.005	<0.001
TMS	–0.085	0.005	<0.001

¹ Positive scores represent an association with increased (and thus improved) TFC score.

² The standard errors reported are robust standard errors.

Post hoc analysis of sample characteristics

Post hoc analysis was limited to all gene positive subjects (as defined above) between the ages of 18 and 64 (inclusive) who were either still employed or had a recorded age at leaving work and who were not classified as having juvenile HD. Subjects who did not have a recorded age of symptom onset despite reporting the presence of cognitive or motor symptoms were also excluded. Chi-squared tests were performed using the IBM SPSS Statistics software package [23] (i) to determine if order of symptom onset was significantly different between unemployed and employed subjects, (ii) to determine if point in disease progression at which subjects left work was significantly different between subjects who left work due to age and subjects who left work due to health, and (iii) to examine if point in disease progression at which subjects left work was significantly different between subjects who reported experiencing motor symptoms first, cognitive symptoms first, and simultaneous symptom onset (defined as onset of both cognitive and motor symptoms occurring during the same year).

RESULTS

Sample selection

Data from 4,146 individuals, 3,175 of whom were gene positive, were available in the Enroll-HD 2015 periodic data cut. Data from 642 of these individuals met the selection criteria as defined in the Materials and Methods section (Table 1) and were incorporated into the linear regression analyses.

Relationship between age of symptom onset and age at leaving work

Both age of cognitive symptom onset and age of motor symptom onset were significantly associated with age at leaving work. For every year that cognitive symptom onset or motor symptom onset was delayed, there was a corresponding 0.806 (SE = 0.030, p < 0.001, adj-R² = 0.628) and 0.814 (SE = 0.031, p < 0.001, adj-R² = 0.603) year increase in age of leaving work, respectively (Fig. 1). Upon controlling for sex, marital status, education attainment, self identified race or ethnicity, region of residence, type of residence, and apathy, every year cognitive or motor symptom onset was delayed corresponded to a 0.476 (SE = 0.075, p < 0.001) and 0.395 (SE = 0.040, p < 0.001) year increase in age at leaving work, respectively (Table 2).

Relationship between cognitive or motor scores and functional scores

Both SCNT and TMS assessment scores were significantly associated with TFC assessment score. For every additional correct answer recorded during the SCNT, there was a corresponding increase in the TFC of 0.105 points (SE = 0.0006, p < 0.001, adj-R² = 0.315) (Fig. 2). If the TMS is held constant, this corresponded to an increase in the TFC of 0.037 points (SE = 0.007, p < 0.001) (Table 3). For every additional point on the TMS, indicating worsening motor symptoms, there was a corresponding decrease in the TFC of 0.104 points (SE = 0.003, p < 0.001, adj-R² = 0.640) (Fig. 2). If SCNT score was held constant, this corresponded to a decrease in the TFC of 0.085 points (SE = 0.005, p < 0.001) (Table 3).

Post hoc analysis of sample characteristics

The chi-squared test revealed that distribution of order of symptom onset was significantly different between the employed and unemployed subjects (p < 0.001). A greater percent of employed subjects (55%) than unemployed subjects (4%) reported no symptom onset, whereas experiencing either cognitive symptom onset first, motor symptom onset first, or simultaneous symptom onset was more prevalent in the unemployed subjects (17%, 66%, and 14% respectively) than in the employed subjects (5%, 38%, and 3% respectively) (Table 4). The chi-squared test also revealed that, when assessing the point in disease progression at which a subject left work (e.g., prior to any symptom onset, before cognitive symptom onset but after motor symptom onset, etcetera), the distribution was significantly different between subjects who left work due to health and subjects who left work due to age (p < 0.001) (Table 5). Distribution of the point in disease progression at which a subject left work was also significantly different between subjects who reported different orders of symptom onset (p < 0.001) (Table 6).

Table 4

Comparison of symptom progression in subjects employed and unemployed at baseline 1

Order in which subject experienced symptoms	Unemployed subjects [n (%)]	Employed subjects [n (%)]
Cognitive symptoms first	108 (17%)	34 (5%)
Motor symptoms first	419 (65%)	253 (38%)
Simultaneous symptom onset	89 (14%)	18 (3%)
No symptoms yet	24 (4%)	368 (55%)
Total	640	673

¹Different sample selection criteria, as described in the Methods section, were used in post hoc analysis than in the original analyses to include subjects with onset of only one or neither symptom, to compare employed subjects to unemployed subjects, and to control for reason for leaving work.

Table 5

Point in disease progression during which subject left work

	Reason for stopping work
Symptoms 1	Either health or age [n (%)]	Health [n (%)]	Age [n (%)]
No symptoms	116 (18%)	84 (15%)	32 (54%)
Only cognitive symptoms present	36 (6%)	35 (6%)	1 (2%)
Only motor symptoms present	206 (32%)	188 (32%)	18 (31%)
Both symptoms present	282 (44%)	274 (47%)	8, (14%)
Total	640	581	59

¹Refers to which symptoms were present at time that subject left work.

Table 6

Point in disease progression during which subjects left work, given order of symptom onset and reason for leaving work

Order of symptom onset	Point in disease progression during which subject left work (indicated by which symptoms were present)	Reason for leaving work: either health or age [n (% 1 )]	Reason for leaving work: health [n (%)]	Reason for leaving work: age [n (%)]
Cognitive symptom onset prior to motor symptom onset	No symptoms	7 (6% 2 )	5 (5%)	2 (40%)
	Only cognitive symptoms	36 (33%)	35 (34%)	1 (20%)
	Both cognitive and Motor symptoms	65 (60%)	63 (61%)	2 (40%)
	Total	108 (100%)	103 (100%)	5 (100%)
Motor symptom onset prior to cognitive symptom onset	No symptoms	64 (15%)	47 (12%)	17 (44%)
	Only motor symptoms	206 (49%)	188 (49%)	18 (46%)
	Both cognitive and Motor symptoms	149 (36%)	145 (38%)	4 (10%)
	Total	419 (100%)	380 (100%)	39 (100%)
Cognitive symptom onset and motor symptom onset during same year	No symptoms	21 (24%)	20 (23%)	1 (33%)
	Both cognitive and Motor symptoms	68 (76%)	66 (77%)	2 (67%)
	Total	89 (100%)	86 (100%)	(100%)

¹Percent given based on only those subjects with same order of symptom onset and reason for leaving work.

²e.g.) Of the subjects who both experienced cognitive symptom onset first and reported leaving work due to either health or age, 7 subjects (6%) left work at the point in disease progression during which they had neither cognitive nor motor symptom onset yet.

DISCUSSION

The primary purpose of this project was to determine the association of cognitive symptoms with the decision to leave work, and to compare this to the association of motor symptoms with the decision to leave work. The results suggest that cognitive and motor symptoms have similar, significant associations with ability to maintain employment; cognitive and motor symptoms may both influence the decision to leave work.

The significant and similarly strong positive correlation of age of both motor and cognitive symptom onset with age at leaving work suggests that cognitive symptoms contribute to a similar extent as motor symptoms to the decline in occupational functioning. There are many ways by which symptom onset can affect ability to work. Motor, cognitive, and psychiatric symptoms may all reduce the speed or accuracy with which a subject can perform both mental and physical tasks. These symptoms may also be noticed by colleagues or clients and deemed odd or unprofessional, reducing occupational functioning by impairing social interaction. The decision to leave work after symptom onset may also be driven by reasons other than occupational decline.

Additionally, as the post hoc analysis revealed, 18% of subjects reported having neither motor symptoms nor cognitive symptoms prior to leaving work. This suggest that the strong correlations between age of symptom onset and age at leaving work may be due to not only the influence of HD symptoms on occupational functioning, but also the effect of leaving work on symptom severity. Literature suggests that the reduction in stimulation, disruption of routine, and loss of social networks that follow both voluntary and involuntary job leaving, for example, might worsen cognitive decline, restlessness, depression, and/or anxiety [24, 25].

The significant positive correlation found between the SCNT and the TFC and between the TMS and the TFC also suggests that both cognitive and motor symptoms affect functional ability. These results are corroborated by Tabrizi et al.’s study, which used the TRACK-HD longitudinal observational study to measure change in assessment scores over a 36-month period. Tabrizi et al. concluded that both TMS and Stroop scores significantly predict changes in TFC scores [26]. The stronger correlation between the TMS and the TFC (adj-R² = 0.640) than between the SCNT and the TFC (adj-R² = 0.315) may indicate either that motor symptoms have a stronger effect on functional ability or that the TMS assessment is a better indicator than the SCNT assessment of functional decline. Clinical significance of these results for any one individual should be considered in the context of his or her employment situation. These conclusions are in contrast to the Goh study mentioned above, which used the same observational study database and found motor symptoms, but not cognitive symptoms, to be significantly associated with occupational capacity [6]. Goh et al.’s study, however, aimed to assess the influence of early cognitive and motor impairment in working individuals classified as asymptomatic. As such, it used a very different sample and relied on the assessments’ sensitivity to early deficits.

Post-hoc analyses during the present study revealed interesting findings about the sample characteristics. After controlling for order of symptom onset, significantly more subjects reported leaving work with only motor symptoms (49%) than with only cognitive symptoms (33%). Further investigation should be performed to understand what this discrepancy reveals about the potential contributions of cognitive and motor decline to the decision to leave work. This may indicate that motor symptoms have a more detrimental effect than cognitive symptoms on occupational functioning, perhaps because it is more difficult to hide motor symptoms than cognitive symptoms. Given that many of these subjects left work prior to the September 2016 SSA guidelines revisions, which provided more weight to cognitive symptoms as a contributor to disability, it is also possible that it was easier to obtain disability insurance and thus leave work with only motor symptoms than to do so with only cognitive symptoms [12, 13]. Further, since a clinical diagnosis of HD is traditionally based on motor symptoms alone, receiving the HD diagnosis due to motor symptoms may prompt individuals to cease working. Post hoc analysis also revealed that 50% of subjects did not leave work until cognitive symptoms were present. The development of treatment that addresses cognitive decline in HD may substantially benefit these individuals by extending the duration that they are able to engage in gainful employment.

One limitation of the Enroll-HD database was that the presence and age of onset of symptoms was self-reported. Four times more subjects reported experiencing motor symptoms first than experiencing cognitive symptoms first. This contradicted Duff et al.’s study, which assessed cognitive scores of all gene-positive subjects without motor symptoms enrolled in the PREDICT-HD observational study and found that 54% of subjects near clinical diagnosis met the criteria for mild cognitive impairment [27]. This, in concert with lack of insight as a common symptom of HD, suggests that some subjects who reported leaving work prior to cognitive symptom onset may have been, unknowingly, cognitively impaired [1]. The possibility of incorrect self-reported data may contribute to the finding that, even after controlling for order of symptom onset, subjects were more likely to report leaving with only motor symptom onset than with only cognitive symptom onset. This could also have played a role in the finding that 18% of subjects reported neither cognitive nor motor symptoms prior to leaving work. Self-selection bias was another limitation of using data from the Enroll-HD observational study. For example, subjects who were having difficulty at work may have been more likely to participate because of the opportunity to discuss their employment issues with physicians and possibly social workers during visits, and then to apply for disability. Additionally, unemployed subjects generally have more free time to devote to study participation. Lastly, it is important to note that the SCNT cognitive assessment does have an oral motor component, and thus is influenced by and may reflect not only cognitive capacity but also motor capacity.

Overall, the results of this investigation suggest that cognitive symptoms have a significant association with occupational functioning and the decision to leave work, comparable to the association of motor symptoms. This suggests that the development of therapies for both cognitive and motor decline would be important for allowing individuals with HD to remain in the workforce for longer [14]. Further studies should be performed that assess which components of the cognitive and motor impairment associated with HD affect work functioning to the greatest extent. Additionally, behavioral symptoms were not assessed in the present study. This is because currently there is no validated way to differentiate behavioral symptoms caused by the CAG expansion from conditions that would have developed regardless of the CAG expansion or from reactive psychiatric conditions that developed in response to the knowledge of one’s risk for HD. Behavioral symptoms are likely important predictors of job loss. A prospective study of subjects with Parkinson’s disease, a neurodegenerative motor disorder that like HD is characterized by dopamine dysregulation, compared baseline behavioral and motor assessment scores of employed subjects that remained employed (n = 195) and became unemployed (n = 224) and found mental health symptoms, but not motor symptoms, to be associated with leaving work [28]. Additionally, the Enroll-HD database does not report job complexity or level of cognitive and physical demand, nor does it report whether a subject has moved to a lower complexity job due to impairment [4]. It is recommended that future studies be performed that investigate these topics to more fully understand the effect of HD on occupational functioning.

CONFLICTS OF INTEREST

The authors have no conflict of interest to report.

Footnotes

ACKNOWLEDGMENTS

The researchers would like to thank the Griffin Foundation for their ongoing support of the Georgetown MedStar Huntington’s Disease Care, Education, and Research Center. The researchers would also like to thank the Georgetown Undergraduate Research Opportunities Program for their financial support of this project, as well as Jan LaRocque, PhD, and Joan B. Riley, MS, MSN, FNP-BC, FAAN, for their invaluable feedback and contributions to this manuscript.

The table provided in the supplementary material outlines the questions asked as part of the Variable Items form during the Enroll-HD study visits to collect information regarding employment status.

The supplementary material is available in the electronic version of this article: .

References

Bates

, Dorsey

, Gusella

, Hayden

, Kay

, Leavitt

, et al.Huntington disease. Nat Rev Dis Primers. 2015;(1):1–21.

Brand

. The far-reaching impact of job loss and unemployment. Annu Rev Sociol. 2015;41:359–75.

Beglinger

, O’Rourke

, Wang

, Langbehn

, Duff

, Paulsen

. Earliest functional decline in Huntington disease. Psychiatry Res. 2010;178(2):414–8.

Ross

, Pantelyat

, Kogan

, Brandt

. Determinants of functional disability in Huntington’s Disease: Role of cognition and motor dysfunction. Mov Disord. 2014;29(11):1351–8.

Marder

, Zhao

, Myers

, Cudkowicz

, Kayson

, Kieburtz

, et al.Rate of functional decline in Hunt-ington’s disease. Huntington Study Group. Neurology. 2000;54(2):452–8.

Goh

, You

, Perin

, Clay

, Loi

, Ellis

, et al.Predictors of workplace disability in a premanifest Hunt-ington’s disease cohort. J Neuropsychiatry Clin Neurosci. 2018;30(2):115–21.

McGurk

, Drake

, Xie

, Riley

, Milfort

, Hale

, et al.Cognitive predictors of work among social security disability insurance beneficiaries with psychiatric disorders enrolled in IPS supported employment. Schizophr Bull. 2018;44(1):32–7.

Stout

. Neurocognitive signs in prodromal Huntington’s disease. Neuropsychology. 2011;25(1):1–14.

Stroop

. Studies of interference in serial verbal reactions. J Exp Psychol Gen. 1935;18(6):643–62.

10.

Huntington Study Group. Unified Huntington’s disease rating scale: Reliability and consistency. Mov Dis. 1996;11:136–42.

11.

Loy

, Elizabeth

. Is a motor criterion essential for the diagnosis of clinical Huntington disease?PLoS Curr. 2013Apr 11 [revised 2013 Apr 11];5. pii: ecur-rents.hd.f4c66bd51e8db11f55e1701af937a419.

12.

Huntington’s Disease Society of America. Huntington’s disease parity act. Available from: http://hdsa.org/about-hdsa/advocacy/huntingtons-disease-parity-act-2/

13.

Caring Voice Coalition. Social security’s new Huntington’s disease listing. 2016.

14.

Simpson

, Lovecky

, Kogan

, Vetter

, Yohrling

. Survey of the Huntington’s disease patient and caregiver community reveals most impactful symptoms and treatment needs. J Huntingtons Dis. 2016;5:395–403.

15.

Teva Pharmaceutical Industries Ltd.Teva Announces FDA Approval of AUSTEDO™ (deutetrabenazine) Tablets for the Treatment of Chorea Associated with Huntington’s Disease. Businesswire. 2017.

16.

Center for Drug Evaluation and Research (CDER), US Food and Drug Administration (FDA). Huntington’s disease: The voice of the patient. A series of reports from the US FDA’s patient-focused drug development initiative. 2016.

17.

CHDI Foundation. Enroll-HD: About this study. Available from: https://www.enroll-hd.org/learn/about-this-study/

18.

Landwehrmeyer

, Fitzer-Attas

, Guiliano

, Goncalves

, Anderson

, Cardoso

, et al.Data analytics from Enroll-HD, a global clinical research platform for Huntington’s disease. Mov Disord Clin Pract. 2016;4:212–24.

19.

Young

, Shoulson

, Penney

, Starosta-Rubinstein

, Gomez

, Travers

, et al.Huntington’s disease in Venezuela: Neurologic features and functional decline. Neurology. 1986;36(2):244–9.

20.

Weisstein WW. Bonferroni correction. Mathworld-a Wolfram web resource. Retrieved from http://mathworld.wolfram.com/BonferroniCorrection.html.

21.

Fox

, Weisberg

Cox proportional-hazards regression for survival data in R. An appendix to an R companion to applied regression, second edition.2011.

22.

Hosmer

, Lemeshow

, May

. Introduction to regression modeling of survival data. Applied Survival Analysis: Regression Modeling of Time-to-Event Data. 2008;1–15. Hoboken, NJ:Wiley-InterScience.

23.

IBM Corp. (Released 2013) IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp.

24.

Bonsang

, Adam

, Perelman

Does retirement affect cognitive functioning? Paper presented at SHARE Users Conference. 2009. doi: 10.1016/j.jhealeco.2012.03.005

25.

Rohwedder

, Willis

. Mental retirement. J Econ Perspect. 2010;24(1):119–38.

26.

Tabrizi

, Scahill

, Durr

. Predictors of phenotypic progression and disease onset in premanifest and early stage Huntington’s disease in the TRACK-HD study: Analysis of 36-month observational data. Lancet Neurol. 2013;12:637–49.

27.

Duff

, Paulsen

, Mills

, Beglinger

, Moser

, Smith

, et al.Mild cognitive impairment in pre-diagnosed Huntington disease. Neurology. 2010;75(6): 500–7.

28.

Armstrong

, Gruber-Baldini

, Reich

, Fishman

, Lachner

, Shulman

. Which features of Parkinson’s disease predict earlier exit from the workforce?Parkinsonism Relat Disord. 2014;20:1257–9.