A retrospective study of the impact of lifestyle on age at onset of Huntington disease

A retrospective study of the impact of lifestyle on age at onset of Huntington disease
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A Retrospective Study of the Impact of Lifestyle on Age atOnset of Huntington Disease M. Kaye Trembath, BSc(Hons), PhD, Grad Dip (Genetic Counselling), 1 Zoe¨ A. Horton, BBNSc, Grad Dip Sci, Hons Genetics, Grad Dip (Genetic Counselling), 1 Lynette Tippett, PhD, 2 Virginia Hogg, MA(hons), Dip Clin Psych, 2 Veronica R. Collins, BSc, Grad Dip Ed, MSc(Epidemiol), PhD, 1 Andrew Churchyard, BSc(hons), MBBS, FRACP, PhD, 3 Dennis Velakoulis, MBBS, FRANZCP, 4 Richard Roxburgh, FRACP PhD, 5 and Martin B. Delatycki, MBBS, FRACP, PhD, 1,6,7 * 1  Bruce Lefroy Centre for Genetic Health Research, Murdoch Childrens Research Institute, Royal Children’s Hospital, Parkville, Victoria, Australia  2  Department of Psychology, The University of Auckland, Auckland, New Zealand   3  Huntington Service, Calvary Healthcare Bethlehem, Caulfield, Victoria, Australia  4  Neuropsychiatry Unit, Royal Melbourne Hospital, Parkville, Victoria, Australia  5  Department of Neurology, Auckland City Hospital, Auckland, New Zealand  6   Department of Clinical Genetics, Austin Health, Heidelberg, Victoria, Australia 7   Department of Paediatrics, University of Melbourne, Parkville, Victoria, Australia Abstract:  In transgenic mouse models of Huntington disease(HD) environmental enrichment significantly delays diseaseonset. A questionnaire-based survey of 154 adults with diag-nosed HD (mean 4.2 years postdiagnosis) and a known  IT15 CAG repeat length, explored whether premorbid lifestylemay relate to age-at-onset (AO). Participants were drawnfrom HD outpatient clinics in Australia and New Zealand.Premorbid physical, intellectual, and passive activity levelswere used to generate scores in the categories of leisure, non-leisure (education, occupation and domestic duties) and totallifestyle. AO was associated with increased CAG repeatlength as expected (r   5 2 0.72,  P  <  0.001), but also with alifestyle that included higher levels of passive activity (r   52 0.38,  P  <  0.001). Multiple linear regression modelingshowed lifestyle passivity to be a variable independent of CAG repeat length in predicting AO (R 2 5  0.54, b  5 2 0.22,  P  5  0.005). Comparison of the mean AO across tertiles of lifestyle passivity scores showed onset 4.6 years (95% CI  5 1.3–7.9) later in the least compared with the most passivetertile. CAG repeat length was also shown to predict lifestylepassivity (R 2 5  0.12,  b  5  1.08,  P  <  0.0005). Neither intel-lectual nor physical activity showed significant relationshipsto AO or CAG repeat length in this cohort. Our study leadsto two conclusions: that a passive lifestyle may be a preclini-cal expression of HD, and that it actually contributes to theearlier onset of symptoms. Overcoming the tendency to bepassive may substantially delay onset of HD.    2010 Move-ment Disorder Society Key words:  age-at-onset; Huntington disease; environ-mental modifiers; lifestyle; passivity Huntington disease (HD) is an autosomal dominantneurodegenerative disease characterized by variablephenotypic expression in motor, cognitive and psychi-atric symptoms, age-at-onset (AO), and severity andduration of disease. 1,2 AO is commonly in the 30s or 40s and strongly dependent upon the length of anabnormal CAG repeat expansion within the  IT15  genethat accounts for 44 to 73% of the varianceobserved. 1,3,4 There is evidence that about 40% of theremaining variance (not accounted for by CAG repeatlength) is attributable to other genetic factors, while Potential conflict of interest: The authors report no conflict of interest.*Correspondence to: Prof. Martin B. Delatycki, Bruce LefroyCentre for Genetic Health Research, Murdoch Childrens ResearchInstitute, Royal Children’s Hospital, Parkville, Victoria 3052,Australia. E-mail: martin.delatycki@ghsv.org.auReceived 16 October 2009; Revised 10 December 2009; Accepted27 February 2010 Published online 29 March 2010 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/mds.23108 1444  Movement Disorders Vol. 25, No. 10, 2010, pp. 1444–1450   2010 Movement Disorder Society  60% is due to undefined environmental factors. 3 Several specific genes or loci that act as independentmodifiers of AO have been identified, 5–11 plus parame-ters with a genetic, or at least familial componentincluding parental age-at-onset, paternal vs maternaltransmission, 12,13 and an interaction between theexpanded and normal alleles of   IT15 . 14 Environmentalfactors influencing onset of human HD remain largelyunidentified.In transgenic mouse models of HD, environmentalenrichment significantly delays the onset and slows theprogression of disease 15,16 and dietary restriction 17 or supplementation with essential fatty acids 18 amelioratestheir HD phenotype. In each instance, associated mo-lecular changes are evident in the developing and adultbrain. 19 There is evidence that environmental factors includ-ing cognitive stimulation, education, leisure activities,occupation, diet, and smoking influence the onset andprogression of other neurodegenerative disorders suchas Alzheimer disease, Parkinson disease and dementia,either directly or by interacting with genetic risk fac-tors. 19–24 Our study explored the premorbid leisure activitiesand nonleisure activities of individuals with HD toestablish whether lifestyle may have influencedtheir AO. METHODSParticipants Participants were recruited through HD outpatientclinics coordinated by Huntington’s Victoria in Aus-tralia and by Auckland, Wellington and Waikato HDAssociations and the Auckland City Hospital in NewZealand. Inclusion criteria included a formal clinicaldiagnosis of HD, an expanded CAG repeat in the  IT15 gene of known length, being 18 years or older andadjudged by the attending clinician (neurologist, neuro-psychiatrist or psychologist) as being capable of responding meaningfully to a questionnaire. AO wasdefined as the age when the first enduring change infunctioning consistent with HD symptomatology wasnoted, and established by consensus between the par-ticipant and attending family members. A single inter-viewer, with no knowledge of the participant’s  IT15 CAG repeat length at the time of interview, wasinvolved in each country to maintain consistency.Local ethics committees approved the study, andwritten informed consent was obtained from allparticipants. Measures A questionnaire was developed based on Friedlandand colleagues’ 25 retrospective study of the relation-ship between lifetime activity and Alzheimer disease.A single interview with each individual (usually with afamily member), provided data pertaining to the‘‘nominal premorbid lifetime’’ defined as the periodfrom 13 years of age to AO. A range of ‘‘leisure’’activities, classified as predominantly physical, intellec-tual, or passive, were covered for three life-stages,namely the teen years (13–19), the 20s and 30s, andthe 40s and 50s, or for as long as the individualremained asymptomatic. Leisure activity scores werecalculated from the frequency with which each activitywas undertaken. ‘‘Nonleisure’’ activity scores weregenerated incorporating education, occupation, anddomestic activity data. From these two scores, compos-ite ‘‘lifestyle’’ scores for each of the three activitytypes were derived (summarised in Fig. 1). Thirtyleisure activities were recorded on a 0–7 scale as per Verghese et al. 26 Only activities performed regularlycontributed: 0  5  never or occasional, 1  5  about onceper week on average, 2 to 6  5  several times per week,7  5  significant activity most days. For televisionwatching a score of 7 required  > 30 hours per week,with each point representing about 4 hours. Physicalactivities recorded included walking for exercise or transport, running, gardening, and a variety of teamsports. Intellectual activities included reading, drawingand painting, craftwork, playing musical instruments,board games or cards. Passive activities (defined asthose requiring little or no initiative or lacking in phys-ical or intellectual challenge) included listening tomusic, watching movies or television, conversing bytelephone or other passive social interaction. Withineach applicable life-stage, scores were totalled for eachcategory of activity. Lifetime average scores (i.e.leisure passivity, leisure intellectuality, and leisure FIG. 1.  Derivation of activity scores over premorbid lifetime. Allscores are an average over the individuals’ premorbid lifetime(defined as the period between 13 years of age and AO). 1445 LIFESTYLE AND AGE AT ONSET OF HD  Movement Disorders, Vol. 25, No. 10, 2010  physicality scores) were then generated by averagingthe life-stage scores in each category over the nominalpremorbid lifetime of each individual.We defined an effective full-time year of study(EFTYS) as 1 year full-time or 4 years part-time (e.g.,the formal educational portion of apprenticeships,where the remainder contributed to the occupationscore). The education component was scored asEFTYS from age 13  3  40 (hours per week), averagedover the (AO-13) years of the nominal premorbid life-time. Education was defined as two-thirds intellectualand one-third passive. School sport was consideredpart of leisure activity. Details of all major periods of employment until AO were collected. Individuals werequestioned closely about the nature and duration of their work. Each separate job was classified into one of six groups on the basis of being physical, intellectual,or passive (all in a single category of activity or 50/50in two categories); namely physical (e.g., builder or cleaner), intellectual (e.g., computer programmer or teacher), passive (e.g., telephonist or some processworkers), physical and passive (e.g., truck driver or hairdresser), intellectual and physical (e.g., nurse or mechanic), intellectual and passive (e.g., secretary or taxi driver). Occupation scores were averaged over (AO–13) years as hours per week in each category.Data were collected concerning a range of domesticduties including cooking, cleaning, household repairs/ maintenance, yard maintenance, and record keeping,covering the three life-stages (as appropriate) up toAO. As with leisure activities, the 0–7 frequencyscores were totalled in categories for each life-stageand pre-morbid lifetime averages computed. The scoresdid not strictly reflect hours per week as domesticduties were not simple to quantify nor was the rangecomprehensive, however it was the relativity of thescores that was important not their absolute value. Togenerate lifetime average nonleisure scores the domes-tic, education, and occupation scores were summedwithin categories. Lifestyle scores comprised the sumof lifetime average leisure and nonleisure scores(within categories). Data Analysis Analyses were conducted using STATA 10. AO,CAG repeat length and all activity measures weretreated as continuous variables. As the distributions of AO, CAG repeat length and the various activity param-eters did not deviate from normal, relationships wereexamined using Pearson correlations. Linear regressionanalyses were used to assess the association betweenactivity measures and AO. Where appropriate, differentactivity scores were also separately entered as variablesinto fully adjusted multiple linear regression modelsthat all included CAG as a predictor variable becauseof the strength of association between CAG repeatlength and AO. The criterion for significance of associ-ation was  P  <  0.05.Activity scores were analysed across tertiles andmean AO (with 95% CIs) calculated for each. RESULTSCohort Characteristics A total of 154 participants were interviewed, 129 inAustralia and 25 in New Zealand, with a refusal rateof 3/157. At least one family member accompanied theparticipant at 73% of interviews, a nonfamily carer at3%, and the remainder were interviewed alone.Accompanied participants were a mean of 4.3  6  3.8years postdiagnosis compared with 4.1  6  3.6 years for the group who were unaccompanied. AO precededclinical diagnosis by a mean of 2.8  6  2.9 years. Thecohort was largely unrelated with only 12 pairs of sib-lings and one parent-child pair.Mean AO was 45.7  6  11.5 years with a median of 45 years (range 21–76 years). There was no significantdifference in AO between males and females, thosewho had predictive testing and those who had not, or between Australians and New Zealanders (Table 1). CAG Repeat Length as a Predictor of AO Linear regression analysis showed that AO becameearlier as the CAG repeat length of the expanded alleleincreased, with the length of the CAG repeat account-ing for 51% of the variance in AO in this cohort(R 2 5  0.51, b 5 2 2.65,  P  <  0.001). Relationships Between Activity Scores and AO Inspection of the correlation coefficients for AO,CAG repeat length and the various measures of activ-ity (Table 2) revealed a number of associations thatwarranted further investigation.AO became earlier as the average passive activityscores increased, regardless of whether the activitytook place in leisure, nonleisure, or was reflected inthe composite lifestyle score. Perhaps surprisingly,intellectual and physical activity scores showed no evi-dence of significant correlation with AO in any facet.As a consequence, linear regression modeling wasundertaken to test whether lifetime average activity 1446 M.K. TREMBATH ET AL.  Movement Disorders, Vol. 25, No. 10, 2010  scores significantly predicted AO (see Fig. 2 for life-style passivity score data). Separate multiple regres-sion models were computed for the composite life-style activity scores, leisure activity scores and non-leisure scores, including CAG repeat length as apredictor (Table 3). The model comprising the threecomposite lifestyle activity scores (passive, intellec-tual, and physical) and CAG repeat length was signif-icant (R 2 5 0.54, F(4,149)  5  43.5,  P  <  0.001). Passiv-ity was a predictor of AO independent of CAG repeatlength and showed that the more passive the lifestyle,the earlier the onset of disease regardless of whether passivity occurred in leisure, nonleisure or the com-posite lifestyle measure. TABLE 1.  Demographic profile of study participants Age at onset inyears mean 6  SD (range)CAG repeatlength mean  6 SD (range) nTotal cohort 45.7 6  11.5 (21–76) 43.8 6  31 (36–54) 154Gender Male 44.6 6  11.7 (21–76) 44.3  6 3.4 (36–54) 86Female 47.1 6  11.2 (29–73) 43.3  6 2.7 (39–51) 68Mean difference  2 2.5 1.0  P  value 0.18 0.05Parental srcinPaternal 42.4 6  10.5 (24–66) 44.3  6 2.9 (40–52) 51Maternal 44.2 6  10.2 (21–76) 44.1  6 3.2 (36–54) 80Mean difference –1.8 0.2  P  value 0.33 0.7Unknown 58.4 6  9.7 (32–73) 41.9  6 2.5 (39–50) 23Mean difference a 14.9 2.3  P  value  < 0.001  < 0.001Predictive testingYes 44.6 6  10.8 (21–66) 47No 46.2 6  11.8 (24–76) 107Mean difference  2 1.6  P  value 0.4NationalityAustralian 45.8  6 11.7 (21–76) 43.9  6 3.2 (36–54) 129New Zealand 45.3 6  10.7 (26–67) 43.3  6 2.8 (38–52) 25Mean difference 0.5 0.6  P  value 0.84 0.38Two-way comparisons of mean AO or mean CAG repeat length between groups were performed using t   tests. a Difference between mean known (paternal or maternal) and mean unknown parental srcin. TABLE 2.  Correlations between AO, CAG repeat lengthand the various activity scores Variable Age at onset inyears  r   (  P  value)CAG repeatlength  r   (  P  value) CAG repeat length  2 0.72 ( < 0.001)Passive activityLifestyle  2 0.38 ( < 0.001) 0.34 ( < 0.001)Leisure  2 0.22 (0.005) 0.15 (0.07)Nonleisure  2 0.34 ( < 0.001) 0.34 ( < 0.001)Intellectual activityLifestyle  2 0.02 (0.8)  2 0.02 (0.77)Leisure 0.04 (0.64)  2 0.1 (0.23)Nonleisure  2 0.05 (0.57) 0.03 (0.75)Physical activityLifestyle 0.07 (0.39)  2 0.09 (0.27)Leisure  2 0.03 (0.74) 0.07 (0.38)Nonleisure 0.09 (0.28)  2 0.13 (0.11)All variables showed normally distributed histograms, thus associa-tions were examined using Pearson coefficients (r) FIG. 2.  Linear regression analysis using lifestyle passivity score as apredictor of AO. The line is fitted and the 95% confidence interval isshaded. R 2 5 0.14, b 52 0.45,  P 5  < 0.001. 1447  LIFESTYLE AND AGE AT ONSET OF HD  Movement Disorders, Vol. 25, No. 10, 2010  Leisure Activities During Different Life-Stages Regression models were also computed involvingleisure activity levels during different life periods. Lei-sure passivity in the teen years was the best predictor of AO (R 2 5  0.54,  F (4,149)  5  44.4,  P  <  0.001), (b  52 0.35,  P  5  0.002), despite the fact that generally theteen years had lower levels of passive activity than theother life-stages and the range in scores was muchwider. Leisure passivity reported for the 20s and 30speriod (or the part thereof while the participantremained asymptomatic) also independently predictedAO, but less strongly (R 2 5  0.52, F(4,148)  5  40.0,  P  <  0.001) (b  5 2 0.26,  P  5  0.03). Levels of leisurepassivity in the 40s and 50s for individuals with later onset HD did not predict AO. Levels of intellectualand physical leisure activities showed no associationwith AO in any life-stage (data not shown). Impact of Lifestyle Passivity on AO An estimate of the impact that lifestyle passivitymay have on AO was made in two ways. First, a hier-archical regression model was computed and thechange in R 2 value resulting from the inclusion of thelifestyle passivity variable into the model containingCAG repeat length as the sole independent variablewas noted. R 2 increased from 0.51 to 0.54, indicatingthat   6% of the residual variance in AO could beattributed to the lifestyle passivity score.The second method involved comparing the meansof AO across tertiles of the lifestyle passivity score,which were 51.3 years for the group with the lowestlevels of passive activity compared with 40.5 years for the third with the highest levels. This represented aCAG adjusted difference of 4.6 years (95% CI  5  1.3to 7.9,  P  5  0.006), and reflected a substantially later onset for the group with the least passive lifestyle. Relationships Between Activity Scores and CAGRepeat Length Table 2 also shows a substantial correlation betweenCAG repeat length and average passive activity scores(e.g. lifestyle scores r   5  0.34,  P  <  0.001). Regressionmodels designed to test CAG repeat length as a predic-tor of various behaviors as measured through activityscores, showed that the larger the expansion in the dis-ease allele the more passive the premorbid lifestyletended to be on average (R 2 5  0.12, b  5  1.08,  P  < 0.001). There was no evidence of CAG repeat lengthpredicting intellectual or physical activity in the life-style of any group. There was no interaction betweenCAG repeat length and lifestyle passivity in determin-ing AO. DISCUSSION The focus of this study was to identify features of the premorbid lifestyle of individuals with HD thatinfluenced their AO. This was important because cur-rently those at risk of HD have no strategies availableto them that may delay onset of the symptoms of thisdebilitating disease. Lifestyle and AO The major contribution of this study is to show theexistence of a meaningful and significant association inhuman HD between passivity in lifestyle and AO. Themore that passive activities are pursued the earlier symptoms of HD become apparent, regardless of whether the passivity derives from leisure or nonleisureactivities or is a feature of the combined measure of lifestyle. This finding implies an environmental influ-ence on AO in human HD and indicates that behav-ioral endeavours can affect the disease course. CAG Repeat Length and a Passive Lifestyle A second important finding is the associationbetween CAG repeat length and passivity in this sam-ple. The relationship suggests that a gene positive sta-tus, and in particular the length of the CAG repeat,may be driving a tendency to be passive for manyyears, while individuals are still considered to be pre-morbid, and therefore, passivity may be an early fea-ture of the disease itself. There are numerous prece- TABLE 3.  Multiple linear regression modeling testing for predictors of age at onset  Variable Lifestyle  b  (  P  value) Leisure  b  (  P  value) Nonleisure  b  (  P  value)Passive activity  2 0.22 (0.005)  2 0.28 (0.04)  2 0.27 (0.02)Intellectual activity  2 0.06 (0.22)  2 0.03 (0.8)  2 0.11 (0.12)Physical activity  2 0.07 (0.24) 0.08 (0.55)  2 0.11 (0.12)CAG repeat length  2 2.45 ( < 0.001)  2 2.6 ( < 0.001)  2 2.47 ( < 0.001)R  2 0.54 0.53 0.53F (4, 149) 43.5 ( < 0.001) 41.6 ( < 0.001) 42.3 ( < 0.001) 1448 M.K. TREMBATH ET AL.  Movement Disorders, Vol. 25, No. 10, 2010
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks