RESEARCH ARTICLE
A Systematic Review and Meta-Analysis of Clinical Variables
Used in Huntington Disease Research
Sonia Franciosi, PhD, Yaein Shim, BSc, Margaret Lau, BCom, Michael R. Hayden, MB, ChB, PhD, and Blair R. Leavitt, MD, CM*
Center for Molecular Medicine and Therapeutics and Department of Medical Genetics, Child and Family Research Institute,
University of British Columbia, Vancouver, British Columbia, Canada
ABSTRACT: Treatment effect in Huntington dis-
ease (HD) clinical trials has relied on primary outcome
measures such as total motor score or functional rating
scales. However, these measures have limited sensitiv-
ity, particularly in pre- to early stages of the disease.
We performed a systematic review of HD clinical stud-
ies to identify endpoints that correlate with disease
severity. Using standard HD keywords and terms, we
identified 749 published studies from 1993 to 2011
based on the availability of demographic, biochemical,
and clinical measures. The average and variability of
each measure was abstracted and stratified according
to pre-far, pre-close, early, mild, moderate, and severe
HD stages. A fixed-effect meta-analysis on selected
variables was conducted at various disease stages. A
total of 1,801 different clinical variables and treatment
outcomes were identified. Unified Huntington Disease
Rating Scale (UHDRS) Motor, UHDRS Independence,
and Trail B showed a trend toward separation between
Huntington disease (HD) is a neurodegenerative dis-
order characterized by a myriad of symptoms, includ-
ing cognitive decline, psychiatric disturbances, and
motor abnormalities such as chorea. Current therapies
for HD predominantly target symptom management
and do not alter the underlying disease. A subject with
HD is typically clinically diagnosed with the disorder
------------------------------------------------------------
*Correspondence to: Dr. Blair R. Leavitt, Center for Molecular Medicine
and Therapeutics, 950 West 28th Avenue, Room 2020, Vancouver, BC,
V5Z 4H4, Canada; bleavitt@cmmt.ubc.ca
Funding agencies: This study was supported by CHDI, The Canadian
Institutes for Health Research, and the Huntington Society of Canada.
M.R.H. is a Killam University Professor and holds a Canada Research
Chair in Human Genetics.
Relevant conflicts of interest/financial disclosures: Nothing to report.
Full financial disclosures and author roles may be found in the online ver-
sion of this article.
Received: 14 March 2013; Revised: 24 June 2013; Accepted: 11
August 2013
Published online 18 October 2013 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/mds.25663
HD stages. Other measures, such as UHDRS Apathy,
Verbal Fluency, and Symbol Digit, could only distinguish
between pre- and early stages of disease and later
stages, whereas other measures showed little correla-
tion with increasing HD stages. Using cross-sectional
data from published HD clinical trials, we have identified
potential endpoints that could be used to track HD dis-
ease progression and treatment effect. Longitudinal
studies, such as TRACK-HD, are critical for assessing
the value of potential markers of disease progression
for use in future HD therapeutic trials. A list of variables,
references used in this meta-analysis, and database is
available at http://www.cmmt.ubc.ca/research/investiga-
tors/leavitt/publications. V
C 2013 International Parkinson
and Movement Disorder Society
K e y W o r d s : meta-analysis; literature; Huntington
disease; clinical endpoint; disease severity
when choreic or other extrapyramidal motor abnor-
malities occur; however, data suggest that HD-related
changes, such as striatal atrophy, and behavioral
symptoms, including depression, can occur as early as
10 to 15 years before motor onset.1 One of the chal-
lenges that the HD research community faces is the
lack of sensitive disease outcome measures to track
disease progression. This makes monitoring and symp-
tom management difficult and also impedes the devel-
opment of potential treatments for HD. Furthermore,
because HD progresses slowly and is heterogenous in
symptomatology, a person presenting with HD may
be misdiagnosed if the motor signs are not present.
Subjects with HD are currently monitored over time
using a battery of tests and scales, including the Uni-
fied Huntington’s Disease Rating Scale (UHDRS).2
Outcome measures commonly used in HD clinical tri-
als typically include total motor score (TMS) and total
functional capacity (TFC), two subscales of the
Movement Disorders, Vol. 28, No. 14, 2013 1987
F R A N C I O S I E T A L .
UHDRS. These scales are designed to evaluate motor
and cognitive deficits, behavioral disturbances, and
functional daily activities in HD. These rating scales
have their limitations, however, because they are often
subjective, biased, and inconsistent between raters and
are not sensitive to detect subtle changes over pro-
longed periods of time.3
Biomarker research in HD includes both small-
scale–focused studies for specific outcome measures
and larger, longitudinal observational studies, such as
PREDICT-HD4 and TRACK-HD,5 that include inves-
tigation of potential clinical, cognitive, neuroimaging,
and biochemical outcome measures.6 The focus of
PREDICT-HD is to prospectively characterize clinical,
neurobiological, and behavioral markers of HD before
clinical diagnosis in subjects known to carry the HD
CAG expansion. The aim of this study was to identify
outcome measures that change before manifestation of
motor abmormalities. The findings from PREDICT-
HD have been expanded in TRACK-HD, a study
aimed at determining specific clinical and biological
markers of HD progression from baseline data over a
prolonged period in early HD, premanifest HD, and
control subjects.7 The long-term goal of TRACK-HD
is to determine which outcome measures with a
related functional outcome would allow efficacy test-
ing of potential therapeutics over a short period with
a practical number of participants.
Severity of HD is currently staged as early, mild,
moderate, and severe using the TFC subscale of the
UHDRS.8 Subjects without motor symptoms are con-
sidered premanifest or prodromal and can be further
categorized according to estimated years to onset as
either far or close using a CAG- and age-based predic-
tive model.9 Identification of objective and quantita-
tive biomarkers that change significantly both
preceding disease onset and with each stage of a dis-
ease will facilitate the identification of new treatments
that slow or even prevent the disease. A biomarker
should be quantifiable in a consistent manner and
therefore reproducible. We reviewed the medical liter-
ature and determined the statistical distributions of
clinical and laboratory measures obtained from sub-
jects with HD according to disease stage.
Materials and Methods
Study Identification and Selection
Using HD keywords such as “Huntington Disease”
or “HD” and “clinical” in PubMed, ISI, and Cochrane
Review databases, we searched the HD literature for
clinical studies conducted between January 1, 1993
and December 31, 2011. Study selection was based on
the availability of demographic, biochemical, and clin-
ical measures, particularly in the absence of life-
1988 Movement Disorders, Vol. 28, No. 14, 2013
threatening comorbidities. We excluded reviews and
case reports.
Objective of the Literature Review
Our objective was to determine the characteristics
of published study population and understand how
baseline variables and potential markers changed with
HD stages. Study subjects included premanifest, early,
mild, mild-moderate, moderate, and severe HD sub-
jects as well as control subjects. Other data retrieved
concerned the study question, experimental design,
and the presence or absence of individual subject data.
Data Abstraction and Stratification
Baseline data were abstracted and sorted according to
premanifest (far and close to onset) and each of the five
clinical stages according to the UHDRS TFC score, as
has been previously described.8 Premanifest HD subjects
consisted of individuals who were gene positive for the
HD mutation and had no clinical symptoms or signs of
the disease. Premanifest subjects were considered far
(greater than 12 years) or close (less than 12 years) from
HD, similar to the TRACK-HD study,10 based on esti-
mated years to diagnosis calculated using a CAG- and
age-based predictive model derived by Langbehn et al.9
Subjects manifesting HD were subdivided further into
four groups: early (stage 1; TFC score of 11–13); mild
(stage 2; TFC score of 7–10.9); moderate (stage 3; TFC
score of 3–6.9); and severe (stage 4/5; TFC score of 0–
2.9). A significant number of studies also reported age
and/or CAG size of the larger allele for a mild-moderate
stage, which encompassed subjects in stages 1 to 2. Con-
trol data from healthy subjects were also abstracted.
Data were further organized according to sample size.
The complete list of variables, references used in the
meta-analysis, and database is available at http://
www.cmmt.ubc.ca/research/investigators/leavitt/
publications.
Statistical Analysis
The central tendency and variability of each measure
(mean, median, standard error, standard deviation
[SD], 95% confidence interval [CI], and interquartile
ranges) were tabulated and calculated using Microsoft
Excel. A fixed-effect meta-analysis on each variable was
used to determine the point estimates, 95% CIs, and
study data distributions (i.e., 2 SDs).11 Because of the
large number of variables reported in the HD literature
from 1993 to 2011, a meta-analysis was performed on
only the more frequently reported variables.
Results
Description of Studies
At this time, we identified 749 suitable studies from
1993 to 2011 from the medical literature and
 M A R K E R S O F H U N T I N G T O N D I S E A S E S E V E R I T Y

TABLE 1. Numbers of HD clinical variables segregated by
category reported in the literature from 1993 to 2011
Clinical Variables by Category No. of Variables
Demographic
Sociodemographic
Neurological
Behavioral
Executive function
Neuropsychological
Neuroimaging
Biochemical
Cardiovascular
Miscellaneous
Total no. of variables
compiled a list of 1,801 variables for data abstraction
(Table 1). A complete list of these studies, the
abstracted variables, and database are available at
http://www.cmmt.ubc.ca/research/investigators/leavitt/
publications. The categorical breakdown of the clini-
cal variables and treatment outcomes for data abstrac-
tion are shown in Table 1. The total number of
subjects in these studies was 109,568, of which 22%
were control, 12% pre-HD, and 66% HD. Of the pre-
HD subjects, 31% were designated as far from onset,
11% close to onset, and 58% were pre-HD as a result
of lack of data for further subcategorization into “far”
or “close.” Of the HD group, 8% were early, 24%
109 mild, 4% moderate, 2% severe, and 62% as HD simi-
28 larly as a result of lack of data enabling them to be
293
150
further subcategorized according to disease stage.
247
348
Clinical Variables and Relationship With HD
355
213 Stage
10 The meta-analysis of typical subject demographics is
48 shown in Table 2. We found that age increased with
1,801
advancing HD stage and plateaued from mild to
severe. The CAG size of the larger allele between HD
subjects in the different disease stages did not change,
but was lower in control subjects, compared to HD
subjects, as would be expected. Body mass index
(BMI; kg/m2) did not change in the early stages of
HD; however, the number of studies used in the analy-
sis was low. There was a trend for baseline IQ to
decrease with advancing disease stage. We also found
that 72% of studies reported gender information; in
these studies with control subjects, pre-HD far and
close to onset, mild, and severe HD groups had more

TABLE 2. Meta-analysis of typical subject demographics

Typical Subject Demographics

Study Data
Variable/HD Stage Total Studies Total Subjects Point Estimate (95% CI) Distributions (2 SDs)

Age, years 660 41,412

Control 364 19,104 47.0 (46.4–47.6) 37.4–56.5
Pre-HD-far 35 3,148 31.9 (30.1–33.7) 22.3–41.5
Pre-HD-close 30 1,198 39.4 (37.5–41.2) 30.2–48.5
Early 56 4,217 42.0 (40.4–43.5) 32.1–51.9
Mild 106 10,587 47.0 (45.6–48.4) 32.5–61.5
Mild-moderate 13 198 46.4 (41.9–50.9) 31.1–61.8
Moderate 39 2,198 56.1 (54.2–58.1) 46.7–65.5
Severe 17 960 46.6 (41.6–51.6) 25.4–67.8
CAG (larger allele) 219 6,759
Control 43 2,996 23.6 (22.9–24.3) 19.2–28.0
Pre-HD-far 29 2,643 41.9 (41.2–42.5) 38.4–45.4
Pre-HD-close 28 1,194 42.2 (41.6–42.8) 39.5–44.9
Early 32 1,120 44.9 (44.1–45.7) 40.8–49.0
Mild 59 4,928 44.2 (43.5–44.8) 39.3–49.1
Mild-moderate 4 58 44.5 (41.2–47.8) 37.9–51.0
Moderate 19 395 42.2 (41.2–43.2) 38.6–45.9
Severe 5 67 45.8 (42.4–49.3) 38.9–52.7
BMI (kg/m2) 41 9,636
Control 16 6,692 24.6 (23.8–25.4) 22.2–26.9
Pre-HD 2 12 22.5 (17.4–27.7) 15.3–29.7
Early 9 740 21.1 (20.4–21.7) 19.5–22.6
Mild 12 2,168 24.3 (22.1–26.6) 16.2–32.4
Moderate 1 24 23.3 (20.6–25.9) 20.0–26.5
IQ 49 2,104
Control 31 1,027 111.6 (108.8–114.5) 96.8–128.0
Far 3 306 107.4 (98.0–116.7) 87.6–131.6
Close 6 259 110.2 (101.7–118.7) 89.8–130.6
Early 5 262 99.5 (91.8–107.2) 82.6–116.4
Mild 5 168 93.9 (84.8–103.1) 73.7–114.1

Movement Disorders, Vol. 28, No. 14, 2013 1989

F R A N C I O S I E T A L .

female participants than male. More than 93% of sub-
jects in each of the groups were right-handed. There
was no difference in average years of education
between each of the groups (12–13 years). Further-
more, there was an increase in percentage of subjects
taking medications with advancing disease stage, com-
pared to control (21%) and pre-HD (14%), with 43%
of early HD subjects, 46% of mild, 68% of moderate,
and 76% of severe HD subjects on medications.
Neurological Variables
As expected, the point estimates and 95% CIs for
TFC (Fig. 1A) showed a clear separation between the
control subjects and disease and also between each of
the different HD stages. This reflects the fact that the
HD stages are based on TFC scores. Other neurologi-
cal variables, such as UHDRS Motor Score (Fig. 1B),
also showed a separation between control or premani-
fest and disease and also between the different HD
stages, with some overlap between early and mild
stages. Furthermore, UHDRS Independence (Fig. 1C)
showed a clear difference between control or premani-
fest subjects and disease states and also between con-
trol and the different HD stages, with some overlap
between early and mild stages. Other neurological var-
iables, such as UHDRS Chorea (Fig. 1D), showed a
difference between control and premanifest, early, and
mild stages of the disease, with moderate-to-severe
scores returning to control/premanifest levels. In the
case of the Mattis Dementia Rating Scale (MDRS), a
biphasic pattern was observed in that differences were
only noted between the presence of disease irrespective
of stage and the healthy or premanifest state (data not
shown). For other outcome measures, such as finger-
tapping frequency, for example, there were insufficient
data to perform a meta-analysis (data not shown).

FIG. 1. Fixed-effect meta-analysis results of selected neurological variables. For each clinical variable, the point estimates (point), 95% CIs (box),
and 2 SDs (bars) are presented for (A) UHDRS TFC, (B) UHDRS Motor, (C) UHDRS Independence, and (D) UHDRS Chorea in control, HD far from
onset, HD close to onset, and early, mild, moderate, and severe HD. “N” indicates the total studies (without duplicates), and “n” the total subjects
for the specific stage. Data indicate that there is a trend for disease stage to worsen with UHDRS TFC (total studies 5 173), UHDRS motor (total
studies 5 191), and UHDRS Independence (total studies 5 71) and also between control and disease states. UHDRS Chorea (total studies 5 40), on
the other hand, showed worsening in early-to-mild stages of the disease, with subjects returning to premanifest or control levels in moderate-to-
severe stages. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

1990 Movement Disorders, Vol. 28, No. 14, 2013

 M A R K E R S O F H U N T I N G T O N D I S E A S E S E V E R I T Y

FIG. 2. Fixed-effect meta-analysis results of selected behavioral and executive function outcome measures. For each clinical variable, the point esti-
mates (point), 95% CIs (box), and 2 SDs (bars) are presented for (A) UHDRS Apathy, (B) Verbal Fluency, and (C) Symbol Digit in control, HD far from
onset, HD close to onset, and early, mild, moderate, and severe HD. “N” indicates the total studies (without duplicates), and “n” the total subjects
for the specific stage. Data indicate a slight trend for UHDRS Apathy to change from early-to-mild stages of the disease (total studies 5 24) and a
slight trend for disease stage to worsen with both Verbal Fluency (total studies 5 59) and Symbol Digit (n 5 80) in early-to-moderate stages of the
disease. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Behavioral Variables stages, whereas Trail A (Fig. 3A) showed difference

Most behavioral outcomes analyzed did not show a
change between the different HD stages. Of the behav-
ioral outcome measures analyzed, UHDRS Apathy
(Fig. 2A) did not show a change until mild stages of
the disease. Other measures, such as UHDRS Behavior
and Beck Depression (data not shown), showed no
change between control, premanifest, or HD stages.
For other behavioral variables, such as UHDRS Sad-
ness or UHDRS Aggression, there were insufficient
studies to perform a meta-analysis.
Executive Function Variables
None of the executive function outcomes analyzed
showed a change between the different HD stages. Of
the executive function measures analyzed, Verbal Flu-
ency (Fig. 2B) and Symbol Digit (Fig. 2C) did not show
a difference until early to mild stages of the disease.
Other measures, such as Stroop Color and Stroop Inter-
ference, showed a biphasic pattern in that differences
were noted between control or premanifest and disease
stages (data not shown). For other variables, such as
Symptom Checklist 90 or Tower of London, too few
studies were reported to perform a meta- analysis.
Neuropsychological Variables
Some sample neuropsychological variables are
shown in Figure 3. Of the neuropsychological varia-
bles, Trail B (Fig. 3B) showed a difference between
control or premanifest and each of the different HD
only between healthy and disease stage. The higher
variability in the moderate-to-severe stages could be
the result of the low study number used in the analy-
sis. Other neuropsychological variables, such as the
Mini–Mental State Examination (MMSE) or MDRS,
showed differences between healthy or premanifest
and disease states (data not shown). Several other neu-
ropsychological variables were reported on, such as
the Wisconsin Card Sorting Test and the Weschsler
Adult Intelligence Scale, but not enough reports to
perform a meta-analysis.
Neuroimaging and Metabolite Variables
No difference in either striatal volume (Fig. 3C) or
bicaudate ratio (data not shown) was noted between
control, premanifest, and disease stages. For other
morphometric outcomes, such as total brain volume,
or for brain metabolites, such as N-acetylaspartate or
lactate, the numbers of studies reporting on these vari-
ables were insufficient for a meta-analysis. Overall,
studies reporting on brain imaging measures were few
and subject numbers used in these studies were low,
increasing the variability noted between studies using
these outcome measures.
Biochemical Variables and Other Outcome
Measures
Several biochemical outcome measures were
reported, from plasma, serum, and cerebrospinal fluid

Movement Disorders, Vol. 28, No. 14, 2013 1991

F R A N C I O S I E T A L .

FIG. 3. Fixed-effect meta-analysis results of selected neuropsychological variables and striatal volume. For each clinical variable, the point estimates
(point), 95% CIs (box), and 2 SDs (bars) are presented for (A) Trail A, (B) Trail B, and (C) striatal volume in control, HD far from onset, HD close to
onset, and early, mild, moderate, and severe HD. “N” indicates the total studies (without duplicates), and “n” the total subjects for the specific
stage. Data indicate that although there is no trend for disease stage to worsen with Trail A (total studies 5 36), there is a trend for Trail B scores to
worsen with disease stage (n 539) and no trend for striatal volume to change from early-to-moderate stages (n 5 16) of the disease. [Color figure
can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

(CSF), and included pro- and anti-inflammatory cyto-
kines, such as interleukin (IL)26, IL-8, and IL-10,
cholesterol-related endpoints, such as low- and high-
density lipoprotein, glucose, insulin, brain-derived neu-
rotrophic factor, and insulin-like growth factor.
Reports for each of these variables were too few to
perform a meta-analysis.
Discussion
Sensitive, robust, and reliable trial measures that
have relevance to clinical outcome are critical aspects
of studies designed to determine potential treatment
benefits in HD. To date, endpoints such as TMS and
TFC have been common in HD clinical trials. The
focus of recent well-designed studies, such as TRACK-
HD, are to determine prospectively which clinical
measures might be robust endpoints and/or sensitive
biomarkers in longitudinal HD studies. However, few
studies have confirmed the effectiveness of these end-
points. Therefore, we conducted a review of the HD
literature to assess the clinical trial variables used in
previous HD studies and extracted baseline data on
HD subjects as a way of determining which markers
correlate with disease severity based on TFC score.
We determined that several of the variables ana-
lyzed—UHDRS motor (Fig. 1B), UHDRS Independ-
ence (Fig. 1C), and Trail B (Fig. 3B) correlated with
TFC staging criteria. Several outcomes measures, such
as UHDRS Apathy (Fig. 2A), Verbal Fluency (Fig.
2B), and Symbol Digit (Fig. 2C), showed a change
only after mild stages of the disease. Other outcome
measures, such as UHDRS Chorea (Fig. 1D), showed
a difference between control and premanifest to mild
stages of the disease, but not at later stages of disease.
Outcome measures such as Trail A (Fig. 3A), for
example, showed a biphasic pattern in that changes
were apparent between healthy, or premanifest, and
manifest states only. Similarly, MDRS, MMSE, and
UHDRS Stroop Word and Interference showed a
biphasic pattern. Several other outcome measures,
such as UHDRS Behavior and Beck Depression (data
not shown), showed no change between control, pre-
manifest, and manifest stages. Because of the low
number of studies and subject numbers, brain imaging
measures, such as bicaudate ratio and metabolite out-
comes, showed little to no change either in premani-
fest or manifest stages, compared to control (data not
shown). Striatal volume (Fig. 3C) also showed little
change between controls, premanifest, or manifest
stages. Reports of outcome measures from CSF,
plasma, or serum reported in the literature were too
low to even perform a meta-analysis.
Results from this study are generally consistent with
cross-sectional findings from PREDICT and TRACK-
HD studies. Results from PREDICT-HD indicate that
UHDRS TMS, chorea, bradykinesia, and oculomotor
endpoints are sensitive measures closer to diagnosis and
in HD cases versus controls.12 Our data suggest that
UHDRS TMS (Fig. 1B) and chorea (Fig. 1D) show cross-
sectional differences in HD cases versus controls.

1992 Movement Disorders, Vol. 28, No. 14, 2013

 M A R K E R S
Similarly, results from PREDICT-HD also indicate that
using the Functional Assessment Scale, work and finan-
cial capacity appear to be first areas of decline in early
HD and that the fatigue or low-energy subscale of the
Beck Depression Inventory, rather than depression, are
highly associated with functional impairment in early
HD.13 Our results are consistent with these results,
because the UHDRS Independence Scale showed discrete
changes that correlated with higher disease stage (Fig.
1C). Results from TRACK-HD indicate that motor
changes occur in premanifest subjects. In particular,
speeded tapping interval and frequency were sensitive
outcome measures in premanifest stages, compared to
manifest subjects and controls.14 Also, antisaccade error
rates and tongue protrusion force coordination deficits
showed stepwise increases between control, premanifest,
and manifest stages, and self-paced tap precision showed
a stepwise decrease between control, premanifest, and
manifest stages.5 Furthermore, cognitive measures, such
as Symbol Digit, Stroop Word, Indirect Circle, and
speeded tapping tap duration, decline in early HD.7 Our
results would indicate that Trail B (Fig. 3B), a cognitive
measure of executive functioning, correlates with
advancing disease stage in agreement with TRACK-HD.
Similarly, results from TRACK-HD indicate no changes
in neuropsychiatric measures with the exception of prob-
lem behaviours assesment (PBA) Apathy.7 Similarly, our
results indicate that behavioral endpoints, such as
UHDRS Behavior and Beck Depression, do not correlate
with advancing disease stage, with the exception of
UHDRS Apathy (Fig. 2A), which showed a change only
between mild and later stages of the disease, compared
to early, pre-HD, and control stages.
Imaging endpoints have gained attention as possibly
sensitive markers, because changes as early as 15 years
before HD onset have been reported.1 Interestingly,
our results did not show a clear correlation in any
imaging endpoints with advancing disease stage. Other
imaging endpoints, such as caudate volume and puta-
men volume, showed differences in premanifest or
manifest versus control subjects or no change between
groups. Differences between our results and those of
TRACK-HD could be the result of limitations of our
study design, because few studies have reported on
these measures, increasing the variability of our
results, less rigorous characterization of subject groups
in our analysis, differences in measurement techniques
in different studies, and differences in the specific
imaging parameters used in the TRACK-HD versus
parameters used in the studies included in this meta-
analysis. Furthermore, subjects could be on different
medications and other environmental influences can
affect data, making interpretation of meta-analysis
results across multiple studies difficult.
Interpretation of the results from this meta-analysis
study should be made with caution. Retrospective anal-
ysis of cross-sectional data from multiple sources has
O F H U N T I N G T O N D I S E A S E S E V E R I T Y
inherent limitations,15 which includes the lack of causal
inferences, because biomarker-disease measurements
occur simultaneously and those mentioned above. Cur-
rent HD clinical trial outcomes measures, such as TFC,8
are not sensitive in the premanifest or early stages of the
disease and require a large number of participants and
lengthy study duration.2,5,13 Early results from TRACK-
HD indicate that none of the new clinical measures
seemed to outperform (on the basis of effect sizes) the
standard UHDRS measures.16 A caveat of TRACK-HD
is that investigators are not blinded to clinical status,
which may have biased TRACK-HD investigators.
Imaging markers were most sensitive in TRACK-HD,7
but do not always correlate with functional changes and
are costly to perform in large trials.17 Our results sug-
gest that UHDRS TMS, UHDRS Independence and pos-
sibly Trail B are reasonable clinical measures that could
be used as endpoints in HD clinical trials.
We have performed a systematic review of clinical
measures using cross-sectional data from existing HD
clinical trials and provide access to all of the accumu-
lated data in an online database. Although our results
suggest that certain measures show good cross-
sectional correlations with disease stages in HD (con-
sistent with previously published individual studies),
definitive statements about the value of any specific
measures reviewed as potential endpoints for clinical
trials are beyond the scope of this work. Longitudinal
studies, such as TRACK-HD, are critical for assessing
the value of potential markers of disease progression
for use in future HD therapeutic trials. Ultimately, the
performance of any of the measures we have reviewed
will require replication in well-controlled and blinded
therapeutic phase III clinical trials.
References
1. Paulsen JS, Langbehn DR, Stout JC, et al. Detection of Hunting-
ton’s disease decades before diagnosis: the Predict-HD study. J
Neurol Neurosurg Psychiatry 2008;79:874–880.
2. Huntington Study Group. Unified Huntington’s Disease Rating
Scale: reliability and consistency. Mov Disord 1996;11:136–142.
3. de Boo G, Tibben A, Hermans J, et al. Subtle involuntary move-
ments are not reliable indicators of incipient Huntington’s disease.
Mov Disord 1998;13:96–99.
4. Paulsen JS, Hayden M, Stout JC, et al. Preparing for preventive
clinical trials: the Predict-HD study. Arch Neurol 2006;63:883–
890.
5. Tabrizi SJ, Langbehn DR, Leavitt BR, et al. Biological and clinical
manifestations of Huntington’s disease in the longitudinal TRACK-
HD study: cross-sectional analysis of baseline data. Lancet Neurol
2009;8:791–801.
6. Weir DW, Sturrock A, Leavitt BR. Development of biomarkers for
Huntington’s disease. Lancet Neurol 2011;10:573–590.
7. Tabrizi SJ, Reilmann R, Roos RA, et al. Potential endpoints for
clinical trials in pre-manifest and early Huntington’s disease in the
TRACK-HD study: analysis of 24 month observational data. Lan-
cet Neurol 2012;11:42–53.
8. Shoulson I, Fahn S. Huntington’s disease: clinical care and evalua-
tion. Neurology 1979;29:1–3.
9. Langbehn DR, Brinkman RR, Falush D, et al. A new model for
prediction of the age of onset and penetrance for Huntington’s dis-
ease based on CAG length. Clin Genet 2004;65:267–277.
Movement Disorders, Vol. 28, No. 14, 2013 1993
F R A N C I O S I E T A L .
10. Tabrizi SJ, Scahil R, Durr A, et al. Biological and clinical changes
in premanifest and early stage Huntington’s disease in the
TRACK-HD study: the 12-month longitudinal analysis. Lancet
Neurol 2011;10:31–42.
11. Hedges LV, Vevea, JL. Fixed- and random-effects models in meta-
analysis. Psychol Methods 1998;3:486–504.
12. Biglan KM, Ross CA, Langbehn DR, et al. Motor abnormalities in
pre-manifest persons with Huntington’s disaese: The Predict-HD
Study. Mov Disord 2009;24:1763–1772.
13. Paulsen JS, Wang C, Duff K, et al. Challenges assessing clinical
endpoints in early Huntington disease. Mov Disord 2010;25:2595–
2603.
1994 Movement Disorders, Vol. 28, No. 14, 2013
14. Bechtel N, Scahill RI, Rosas HD, et al. Tapping linked to function
and structure in premanifest and symptomatic Huntington disease.
Neurology 2010;75:2150–2160.
15. Micheel CM, Ball JR, eds. Evaluation of Biomarkers and Surrogate
Endpoints in Chronic Disease. Institute of Medicine Consensus
Report. Washington, DC: National Academies Press; 2010:108–
109.
16. Kieburtz K, Venuto C. TRACK-HD: both promise and disappoint-
ment. Lancet Neurol 2012;11:24–25.
17. Henley SM, Ridgway GR, Scahill RI, et al. Pitfalls in the use of
voxel-based morphometry as a biomarker: examples from Hun-
tington disease. Am J Neuroradiol 2010;31:711–719.