12 Med Getiet 1993; 30: 1028-1035

Ethical and social issues in presymptomatictesting for Huntington's disease: a EuropeanCommunity collaborative study

European Community Huntington's Disease Collaborative Study Group

AbstractAn analysis of social and ethical aspectsof presymptomatic testing for Hunt-ington's disease has been carried out,based on data on linked DNA markers,from four major testing centres in dif-ferent European Community countries(Belgium, Italy, Netherlands, and UnitedKingdom). Information was available on603 applicants, with 213 final resultsgiven, of which 32% gave an increasedrisk. A series of specific issues and prob-lems were documented systematicallyfor all applicants, results being given onfrequency of occurrence and illustratedby individual case histories. The princi-pal issues could be grouped as problemsof inappropriate referral, problemsinvolving relatives, and problems relat-ing to disclosure of results. At least oneimportant problem was encountered in46% of applicants, emphasising the im-portance of expert counselling, prepara-tion, and support of applicants, and ofclose liaison between clinical, counsell-ing, and laboratory staff. The extensiveand detailed information available forHuntington's disease from this and otherstudies will be of considerable value inrelation to genetic testing for other lateonset genetic disorders and will be evenmore relevant to Huntington's diseasenow that specific mutation analysis ispossible for this disorder.(] Med Getiet 1993;30:1028-35)

Huntington's disease (HD) is one of the mostserious and frequent neurogenetic disorders oflater life. The principal features include pro-gressive motor deterioration with involuntarymovements, mental involvement with demen-tia, and general neurodegeneration.' Onset canbe at any age, but is most commonly in adultlife, with severe disability and frequently pro-found psychological effects on family mem-bers.The gene for HD has recently been isolated

and the underlying mutation identified as atrinucleotide (CTG) repeat sequence that isexpanded in HD patients.2 The molecular gen-etic aspects are beginning to be explored indetail,3-5 but the need for caution in applyingthe new information until greater experience isavailable has been stressed.67The existence of close DNA markers for the

HD gene for almost 10 years has allowedmodification of risks of family members by

analysing the pattern of linked markers in thefamily.8-"' Such risk alteration has now been inprogress in programmes throughout the worldfor over seven years and a number of largeseries has been reported."-'4 The seriousconsequences of presymptomatic testing forsuch a serious late onset disorder have alsobeen appreciated and documented.'5-"Although future presymptomatic testing for

HD is likely to be based largely on mutationanalysis rather than linked markers, the ethicaland social issues involved will remain largelyunchanged; indeed they may well becomemore important as the laboratory aspects aresimplified.The four centres involved in our study have

all been concerned with major programmes ofHD prediction in their own countries, and ourcombined experience has provided a consider-able proportion of all such testing undertakenin Europe so far.Some results of testing have already been

reported by the individual centres,'3 1418 1 but itwas considered that a combined analysis of thesocial and ethical aspects of testing and, inparticular, the related problems encounteredduring testing would help to identify thoseissues likely to arise more generally as genetictesting becomes more widespread.

Study methodsThe foundation of the present study was pro-vided by the individual series of predictionsundertaken in the four centres. No additionaltesting was undertaken specifically for thestudy. All applications for HD presymptoma-tic testing in each centre up to the end of 1991were analysed, together with the results of allcompleted tests.The data were extracted from the original

records by the individual centre and trans-ferred to a standard form designed for thestudy. All data used were anonymous andwithout identifying detail, this being knownonly to the individual centre that had per-formed the test. Completed forms were sent tothe Cardiff centre, where data were enteredonto a microcomputer for further analysis.The core of the form consisted of (1) a sheet

giving general relevant details, (2) a summarysheet of the specific problems to be analysed,(3) individual pages to record details of eachspecific problem, and (4) details of test resultfor those applicants reaching this stage.The general data were collected on all sub-

jects who had applied for predictive testing,

Participating centres:Institute of MedicalGenetics, Cardiff, UK.D BallP S HarperA Tyler

Department ofClinical Genetics,Leiden, TheNetherlands.A TibbenM Vegter-van der Vlis

Centre for HumanGenetics, Leuven,Belgium.J J CassimanG Evers-Kiebooms

Institute ofExperimentalMedicine and Instituteof Pyschology,National ResearchCouncil, Rome, Italy.M FrontaliA G Jacopini

Correspondence to ProfessorPeter S Harper, Institute ofMedical Genetics,University Hospital ofWales, Heath Park, CardiffCF4 4XN, UK.

Received 26 August 1993.Accepted 9 September 1993.

1028

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

Ethical and social issues in presymptomatic testing for Huntington's disease

Table 1 Principal characteristics ofHD testing centres.

Cardiff Leiden Leuven Rome

Population South Wales, Netherlands Belgium; majority Italysouthern England from Flanders

Contact with Longstanding and extensive in south Longstanding and extensive Longstanding and Longstanding and extensive forHuntington's disease Wales; not so for southern England extensive Latium region; not for other regionsPrevious register Yes Yes No Yes

Laboratory Integral. Previous molecular research Integral. Previous molecular Integral Integral. Previous molecular researchon HD research on HD on HD

Medical support for Family doctor involved Family doctor not directly Family doctor not Family doctor not directly involvedthose tested involved. Occasional psychiatric directly involved

or neurological referral

Payment for test Nil Nil Nil Nil

though the definition of an applicant variedslightly between centres. In the Cardiff series,data were recorded on all those who had madea serious enquiry about testing, though notnecessarily reaching the stage of initial inter-view; data on this subgroup are recordedseparately from those in all four centres whoreached this stage of the testing procedure.The standardised list of problems allowed

each problem to be identified if it had occurredin the testing of a person. The detailed recordswere reviewed in order to complete this sheetand the subsequent specific sheets. Data oneach problem were recorded in a way thatallowed quantitative analysis, while a spacewas left to give an individual description of theproblem; the illustrative case historiesrecorded later in this report are not taken fromthese sheets, but were supplied by the indi-vidual centres after the preliminary analysis ofthe data.For the subset of applicants actually given a

final test result, details of prior and final riskwere recorded. This permitted a comparisonof the problems arising and the composition ofthe series between this group and the largernumber of overall applicants.

ResultsTHE CHARACTERISTICS OF THE INDIVIDUALCENTRESThe four participating centres, three in north-ern and one in southern Europe, all had exten-sive experience in predictive testing andgenetic counselling for HD, as mentioned pre-viously. All were based in genetics centreshaving a wider involvement in genetic coun-selling for other disorders, and all had closelinks with the molecular genetics laboratoriesinvolved in the HD testing, which formed partof the same institution in all centres. Table 1summarises the principal features of the fourcentres.The background to HD presymptomatic

Table 2 Presymptomatic testing for Huntington's disease. Applications and finalresults from the different centres

Applicants % male Final results % male

Cardiff 284 (216)* 48 42 26Leiden 183 41 127 43Leuven 93 48 37 46Rome 43 56 7 57Total 603 213

*Figure excluding those applicants not proceeding to full interview (see text).

testing varied between the centres. In three(Cardiff, Leiden, and Rome), testing had de-veloped from a programme of HD geneticsresearch, longstanding in Cardiff and Leiden,while in Rome the establishment of the testingprogramme originated in the context of morerecent research activity on HD. In Leuven,testing had developed from a longstandingtradition of genetic counselling for HD andfrom a long and close collaboration with thepatient association.

OVERALL RESULTS OF THE SERIESThe total numbers in the different series aresummarised in table 2. Data on 603 applicantswere available, with 213 subjects having beengiven a final result. The distribution betweenseries varied considerably, with the largestseries of final results coming from Leiden,while the Cardiff series of initial applicantsexceeded the others. As noted before, a pro-portion of these did not proceed to -the firstformal interview, the remainder (216 out of284) giving a figure more closely in line withother centres. The great majority of applicantsand those tested had an affected parent; sub-jects with a healthy parent but affected grand-parent accounting for only 69 of 603 applicants(11%) and 14 of 213 actually tested (7%).The age distribution of applicants and those

tested was similar in the four centres. Onlythree applicants were aged under 20, whilefour were over 60, leaving the great majority ofthose tested (97%) between 20 and 60 years,with a mean age of 32-7 years (male 32-2,female 33 2) for applicants, and of 32-6 years(male 30 9, female 33 8) for those given a finalresult.The sex ratio of applicants (table 2) showed

47% of applicants to be male, a proportionsimilar to that in the general population of thisage range. However, the ratio for thoseactually tested is lower (40%), though thisdifference is contributed mainly by the Cardiffseries (26% males).Table 3 gives data on education levels and

employment in the different series.The distribution of final risks, where an

informative result was obtained, is shown inthe figure, while table 4 shows the results bycentre, including uninformative outcomes,which comprised 11% of the total.The proportion of informative results giving

a raised risk is 32%, a value comparable to that

1029

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

European Community Huntington's Disease Collaborative Study Group

Table 3 Educational levels and employment in the different series.

Educational levels

Pre-school Less than high school High school plus Degree N/K

Cardiff 11 (9*) 151 (140*) 48 (42*) 23 (21*) 51 (0*)Leiden 0 89 81 9 2Leuven 0 31 37 9 16Rome 0 22 14 7 0

Employment

Manual Non-manual N/K Total

Cardiff 147 (123*) 99 (88*) 38 (6*) 284 (217*)Leiden 104 76 1 181Leuven 36 36 21 93Rome 22 21 0 43

*Represents group (216) which received counselling.

of other reported series; the basis of this dif-ference from the 50% and 25% prior risk is atleast in part related to the age of applicants. Itcan be clearly seen from the figure that the ageadjusted risk of many subjects is considerablyless than their prior risk, having already passedthrough a significant portion of their riskperiod.A significant proportion of those requesting

testing were unable to be given a final result.The most frequent reason (57 cases) was aninadequate family structure, with key familymembers dead or unavailable for testing.However, difficulties in confirming the dia-gnosis of HD were also prominent (31 cases),while no fewer than 47 applicants provedalready to be clinically affected.

100 r

80

60k

40

20

L . .vI

0 0V (NC

I I n n n0 LO 0 LO 0 LO Lo 0 Lo 0 Lo 0 Lo 0X 'T LO L t c0 co 0 0 C m 0D

% risk

_ Pre-test LZ Post-test

Age adjusted risk before and after testing.

Table 4 Risk alteration after presymptomatic testing(see figure for detailed risk distribution) .

Risk raised Risk lowered Uninformative

No % No % No %

Cardiff 12 29 25 60 5 12Leiden 45 35 74 58 8 6Leuven 10 27 18 49 9 24Rome 1 14 5 71 1 14

Total 68 32 122 57 23 1 1

Table 5 Specific problems encountered in HD presymptomatic testing.

Cardiff Leiden Leuven Rome Total

Problems of referralReferral without permission 6 ( 3) 0 3 0 9Referral of minor 29 (23) 0 5 3 37Referral under third party pressure 11 (11) 5 6 2 24

Problems involving relativesThreat to privacy of information 27 (23) 1 0 1 29Refusal to supply blood sample 16 (15) 10 4 1 31Use of stored research sample 1 ( 1) 4 7 0 12Unsolicited risk alteration 2 ( 1) 4 0 1 7

Problems of disclosureDisclosure to physician refused 3 ( 3) 4 0 0 7Disclosure to relative/spouse refused 2 ( 2) 11 1 0 14Request for disclosure by relative 0 ( 0) 0 0 0 0Request for disclosure by third party 10 ( 5) 0 2 0 12Request for disclosure to self 1 ( 1) 11 0 1 13Non-compliance with protocol 20 (11) 4 5 8 37Other problem 20 (17) 18 9 1 48

( ) indicated date restricted to applicants proceeding to full interview in Cardiff series.

ANALYSIS OF PROBLEMSThe proportion of applicants in whom at leastone important problem arose was high, 46%(280 out of 603). The specific problems onwhich data were collected on individual sheetsare listed in table 5. Those encountered in theCardiff series are subdivided according towhether they were encountered in the series ofpreliminary enquiries, since this category wasabsent in the other centres, but analyses arebased on the whole series. Details on eachproblem are given below, with specific caseexamples where relevant. Three principalgroups of problems could be identified: thefirst relates to referrals or applications thatmight be considered inappropriate; the secondto the requirement for involving other familymembers in the testing process; while the thirdinvolves issues of disclosure and confidentia-lity.

Problems of referralREFERRAL WITHOUT PERMISSIONThis indicated that an applicant had beenreferred without that person being aware ofthis, or without having given permission fortesting. Of the nine instances, five werereferred from psychiatrists.

Example (Cardiff)A psychiatrist referred a 50 year old man fortesting. He had been under his care for manyyears, suffering from severe depression, buthad never showed any neurological signs. Itwas suspected that the depression was relatedto the family history of HD. The psychiatristasked for the patient to be tested without hisknowledge and offered to send a blood sample.He also asked the genetic centre to send theresult to him so that he could decide whetheror not to inform his patient. The genetic centreexplained why this was not possible andobtained permission to see the patient, whowas quite uninterested in testing; therefore, itwas not carried out. (This case also illustrates arequest for information from a third party.)

REFERRAL OF A MINORThis was defined as any referral under 18 yearsof age (the usual age of majority in the coun-

U)a)

0a).0Ez

1030

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

Ethical and social issues in presymptomatic testing for Huntington's disease

tries concerned). The mean age of such chil-dren was 10 5 years, and the source of referralwas most commonly parents (23 cases) but alsosocial workers, mainly from adoption agencies(7 cases), as well as other professionals(table 6). It can be seen that the majority of the39 cases came from the Cardiff series; it ispossible that comparable referrals to the othercentres may have been filtered out at an earlystage.

Example 1 (Leuven)A 45 year old mother whose husband hadcommitted suicide requesting testing for her14 year old daughter. At the time her husbandcommitted suicide, they were divorced. Themother was only recently informed about thenature of the diagnosis of HD in her father inlaw (she had had no contact with her hus-band's parents for the last eight years).

Thereafter, she started to have doubtswhether her husband might have committedsuicide in association with the onset of HD.Her daughter had no contact at all with heraffected grandfather. According to the mother,the girl was not aware of the presence of anyserious disease in her grandfather or of thesuicide of her father.This woman was furious and indignant

about our refusal to test her 14 year olddaughter. She was convinced that it was herright to know her daughter's status. However,she was only considering the possibility of afavourable result and did not want to thinkabout the consequences of an unfavourableresult in her daughter. It was very difficult forthe mother to accept that a predictive testshould only be performed if the at risk personcan make an informed decision and that testingis never performed below 18 years. Follow upcounselling was offered.

Example 2 (Cardiff)A social worker from an adoption agency askedfor a baby to be tested. The child's mother,aged 19 years, was at 50% risk and there wassome concern that her erratic life style andpromiscuous behaviour might herald the onsetof HD. The child's father was unknown. Thebaby was said to be the result of a briefencounter and the mother had no interest inthe child. The adoption agency felt that theprospective adoptive parents had the right toknow whether the baby was likely to be a genecarrier. Counselling was accepted by the socialworker, the adoptive parents, and the mother,who was totally opposed to being tested her-self. However, she was prepared to considervoluntary surveillance so that, if she were to

Table 6 Referral of minors.

Referral source Cardiff Leiden Leuven Rome Total

Parents 16 0 4 3 23Professionals* 10 0 1 0 11Self 3 2 0 0 5Total 29 2 5 3 39

*Social worker 7, psychiatrist/paediatrician 4.

develop symptoms, the adoptive parents couldbe told.

In fact, owing to the poor family structure,the chance of obtaining an informative testresult was low, but the situation could havecreated great difficulties had a specific mutatio-nal test been available.

REFERRAL UNDER THIRD PARTY PRESSUREThis indicated that testing had been sought,not primarily by the person to be tested, but byanother person, though with permission. Sub-jects involved were family members (mostcommonly spouse or partner) in 16 cases,professionals in eight cases.

Example 1 (Leuven)A young woman was living with her partnerwho is at risk for Huntington's disease.Initially, she completely denied the fact thather partner had a 50% risk of becoming affec-ted. When her parents became aware of thisrisk, they had very strong objections to therelationship.As a consequence of her own mixed feelings,

in combination with the parental objections,she could not cope any longer with the at riskstatus of her friend and became depressed.Under the influence of her parents, she con-sidered that predictive testing could offer away out of this situation. During the initialsession in the genetic centre, it was clear thatthe young man who was at risk was not at allmotivated to have a predictive test. Because ofthe relationship problems, associated partiallywith the risk for Huntington's disease, andbecause of the depression of the partner, thecouple was immediately referred to the psy-chiatrist of the team for a number of thera-peutic sessions.

Example 2 (Rome)A 31 year old woman at 50% risk came to ourcentre in 1987 for genetic counselling. At thattime, presymptomatic testing was not yetavailable in Italy and DNA banking was pro-posed of samples from her affected father anduncle, which was accepted. In 1991, she askedfor presymptomatic testing and came to thefirst session accompanied by her fianc& Afterthe introductory part, the psychologist askedto speak individually to the applicant. Whenasked why she decided to be tested, she saidthat she could not see how to escape the test,since she had been engaged to her fiance for 10years. Both were living in a small village whereeveryone knew everything about everybodyand the disease in her relatives was well knownto the fianc&'s family; in addition, his sister wasa medical doctor and had a professional know-ledge of the disease. After initial opposition tothe engagement, owing mainly to the lowersocial class of her family, the fianc&'s relativesaccepted her and she had a good relationshipwith them. However, when they learned that apresymptomatic HD test was available, they

1031

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

European Community, Huntintgton's Disease Collaborative Study, Group

became violently opposed to their marriageunless a presymptomatic test was done.The applicant thought that she had no

choice, although knowledge of her genetic sta-tus was not relevant for her. Her religiousbeliefs led her to accept everything that canhappen in life, and the disease was not worsethan other possible events. However, havingbeen engaged for such a long time, she couldnot break the engagement: "the gossip wouldbe awful and my mother would die of it". Inaddition, she felt very deceived by her fianc&'sinability to oppose his family situation.During the lengthy session with the psycho-

logist, the applicant realised that she did notwant to submit to the test, not because of fearof a 'bad' result, but because she could notaccept the test as a condition for marriage aftersuch a long engagement. Contact with thepsychologist was kept by the applicant, duringwhich her level of self-esteem increased. Somemonths later, she announced that she wasgoing to marry her fiance, not knowing, how-ever, what made his family withdraw theiropposition. She intended to undergo testingafter marriage without telling the husband.When the psychologist made her think aboutthe difficulty of hiding the emotional impact ofa result, either positive or negative, she askedfor more time to decide.

This case exemplifies: (1) how availability ofa test can induce social pressure on at risksubjects to undertake it; (2) the necessity oflong and patient psychological work to unraveldeep attitudes towards the test and towards theenvironment that stimulated its request, inorder to protect a subject's right to an autono-mous decision; (3) how the decision processcan be long and time consuming both for thesubject and the genetic counselling team; (4)how the whole procedure, for the above rea-sons, is not easily transferable from researchcentres to a routine National Health Servicesetting.Although most examples of third party pres-

sure have related to family members, we havealso encountered pressure from professionals;in some cases, the distinction from the pre-viously discussed situation of 'referral withoutconsent' is a fine one.

Problems relating to other familymembersOne of the principal limitations of geneticprediction based on closely linked geneticmarkers is that a result depends on the analysisof the genetic pattern in a family, not on thegenotype of the applicant alone. This can re-sult in a variety of difficulties, including thoseoutlined below.

THREAT TO PRIVACY OF INFORMATIONTo undergo presymptomatic testing for a per-son, it is essential to have accurate diagnosticinformation on those who are affected, and tobe certain whether a subject whose sample isbeing used is affected or clinically normal. In anumber of instances, relatives were reported as

affected by the family, but had never beenformally diagnosed; in some cases, they wereunaware of their condition and had avoidedmedical contact. Others were noted as havingshown equivocal features. In such instances,the use of information would involve not onlyproviding a sample, but the establishing of adiagnostic situation which had not been soughtby the subject, and which would result ininvasion of privacy. We avoided the use ofsamples in such situations.

Example (Cardiff>A 35 year old woman was distressed to receivewhat she described as a 'peremptory' requestfrom her sister, living in another part of theUK, for a blood sample. The woman hadapplied for testing and had, apparently, beenasked by her local genetic centre to obtain thissample. Her sister was worried and very de-pressed about marital problems and vaguenon-specific symptoms which, she felt, couldherald the onset of HD. She had firmly de-cided not to be tested as she felt that an adverseresult could cause her to become suicidal. Shewas reassured that she did not need to providea blood sample and was offered support, ifneeded, in any further communications withher sister.

REFUSAL TO PROVIDE A SAMPLEThe trauma and disruption often produced bythe occurrence of HD in a family may bereflected in strained family relationshipswhich, in turn, may result in a family member,necessary to provide a risk estimate for theapplicant, refusing to cooperate. While sensi-tive discussion and explanation will resolvemost such cases, a substantial number ofremaining problems were encountered in thisstudy.

Example 1 (Cardiff)A not at risk parent of an applicant refused togive a blood sample to enable three of her adultchildren to be tested. She felt that her childrenhad neglected her for many years and that shewas under no obligation to them. She haddivorced their father many years ago, before hehad developed HD, and subsequently had hadno contact with him. The family, therefore,had divided loyalties. Eventually, after familycounselling sessions, relationships wereresumed and the mother gave a blood sample.

Example 2 (Leuven)A mother with early symptoms of Hunt-ington's disease could not cope with the psy-chological problems associated with the onsetof the disease. In this context, she was not yetprepared to have the diagnosis confirmed by aneurologist and she refused to supply a bloodsample to allow testing of her two adult chil-dren who entered the test programme. Themain reason given was that she could not cope

1032

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

Ethical and social issues in presymptomatic testing for Huntington's disease

with the implications of the test request by herchildren in this particularly difficult period.

USE OF RESEARCH SAMPLESMany centres involved in HD presymptomatictesting, including the four in this study, havebeen involved in previous molecular researchand have maintained stores of DNA on HDkindreds, whose members may later requesttesting as a service. The use of such samples,especially from subjects at risk, may be con-tentious unless it is completely clear whatthose sampled have consented to.While the use of samples from dead subjects

may be vital in allowing prediction, we havegenerally avoided the use of research sampleswhenever possible. In Leuven, researchsamples were only used when dead patientswere vital in allowing prediction and in agree-ment with the family. Moreover, this decisionwas made at the time of DNA storage.

Problems of disclosure andconfidentialityDISCLOSURE TO PHYSICIANInformation on a person's risk status regardingHD is a matter of the greatest sensitivity andpersonal importance, with implications forinsurance, employment, and marriage amongothers. All centres undertaking presymptoma-tic testing have found that an assurance of totalconfidentiality is given the highest priority byapplicants, and the framework of testing isconstructed accordingly in terms of both theactual result and the associated counselling.The practice in the four centres differed

considerably on this point, one of the few areaswhere major differences existed. In Leuvenand Rome, the majority of applications camedirectly from the subject at risk, as notedearlier, not involving any medical professional.In Leuven, results were given face to face withno written result that could be sent either tothe applicant or to anyone else. Upon requestof the applicant, a telephone call with theapplicant's physician was made in a number ofcases. In Rome, written results were given tothe applicant in the same session as they werecommunicated verbally. In the Cardiff series,referrals were channelled through a profes-sional, usually a clinical geneticist, while theresult, initially given face to face, was con-firmed to the applicant in writing, and also tothe applicant's family doctor, and referringclinician, if permission for this was granted. InLeiden also, the family doctor was usuallyinvolved at the referral stage and afterwardswith the permission of the applicant; sevenapplicants from Cardiff and Leiden wereunwilling for disclosure to occur.

REFUSAL OF DISCLOSURE BY THE APPLICANT TOFAMILY MEMBERSThis was encountered in 14 instances, princi-pally involving spouses and children; suchsharing of risk information was recommendedin the counselling protocols of all the centres,

but if a subject was insistent in not sharing theresult or the fact of being tested, this was notregarded as a reason against testing.

It is interesting to note that no direct appli-cation from a relative to have access to a testresult was received, even though this was anitem specifically enquired for on the question-naire. The consequences of non-disclosure ormisleading information on family memberscan be seen from the following examples.

Example 1 (Leiden)A 48 year old man with three children enteredthe presymptomatic testing programme, butdid not respond to the invitation to receive thetest result. Although he has still not been giventhe results, he has informed his children andother relatives that his test was normal.

Example 2 (Leiden)A 25 year old woman, at 25% prior risk,applied for testing and found that her mother(at 50% prior risk) was willing to cooperateand wished to learn her own result. Bothwomen had a high risk result and wereinformed of this separately. Subsequently, thedaughter told her mother and sibs that she hadreceived her result in a sealed envelope andpretended that she did not wish to know theresult until a later, more appropriate time(when considering having a family).

Non-compliance with protocolThe final items listed, non-compliance withthe protocol and 'other problems', contained(not surprisingly) a variety of problems relat-ing to attendance at counselling and follow upinterviews, participation in neurological andpsychological evaluations, and other aspects.Considering the stressful nature of HD pre-symptomatic testing and the expense and in-convenience of travel often involved, it is per-haps surprising that the number is not greater.

Testing subjects at 25% prior riskThe great majority of subjects in this studyhad a parent affected with HD, only 11% ofapplicants and 7% of those tested being seconddegree relatives, with a prior risk of 25%.Although, in some cases, the intervening par-ent had died of an unrelated cause, the situa-tion where a living parent exists, whose childrequests testing, creates a particularly difficultsituation, likely to be encountered more fre-quently now that testing for a specific mutationis possible.Two principal types of situation were

encountered in the study. (1) A parent at riskdoes not wish to know their genotype, whiletheir child requests a presymptomatic test.The child's status, if affected, will reveal thatof the at risk parent. (2) A more distant relativehas agreed to give a blood sample to help arelative and, as a result of family genotyping,their own risk may be increased. In situation 2,it is often possible to do a presymptomatic test

1033

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

European Community Huntington's Disease Collaborative Study Group

for the applicant, while avoiding sampling therelative; in situation 1, only exclusion testingwill ensure the parent's right 'not to know'.

Example 1 (Rome)A 24 year old man with an affected paternalgrandfather and an at risk father asked for apresymptomatic test. The team asked to speakto the father in order to clarify the fact that araised risk for the son would increase his ownrisk and ask him whether he would wish toknow his own status. The father was notwilling to know, although he understood hisson's motives in asking for the test and waswilling to cooperate with a blood sample. Theteam decided that, in order to protect thefather's right not to know, only exclusion test-ing could be offered to the son.

Example 2 (Cardiff)A man at 25% risk of HD asked for exclusiontesting as his father, at 50% prior risk, did notwish in the first instance to be tested himself.They came from a large pedigree and it wouldhave been easy to alter the father's risk inad-vertently by full genotyping of other familymembers. This was avoided by deliberatelynot typing an affected sib of the father so thatphase remained unknown, thus leaving thefather's risk unaltered, while giving the possi-bility of either a very low risk for the son or ofraising his risk to that of the father's presentlevel. Careful forethought and discussion withthe family and with the laboratory was neededfor this approach to be achieved successfully.

DiscussionOur study, along with the individual pro-grammes on which it is based, has shown thatpresymptomatic testing for HD has been car-ried out effectively in four different Europeancountries, using broadly similar protocols forthe preparation, counselling, and support ofthose undergoing testing. However, thenumber and variety of significant problemsthat have been encountered, despite the ab-sence of major adverse effects, raises majordoubts as to whether testing without an estab-lished protocol and the availability of exper-ienced staff is advisable. This will become animportant issue now that simpler and probablycheaper mutational testing, without the needfor sampling other family members, is pos-sible.The specific problems here are not unique to

HD, but represent items that could arise in thecontext of testing for many late onset, progres-sive genetic disorders. Those relating to refer-ral and consent are of special relevance andillustrate the importance of the testing processbeing a joint clinical and laboratory activity.An isolated laboratory that analysed withoutquestion all samples referred to it could havecreated serious ethical dilemmas by the testingof minors or of subjects who had not givenconsent.The group of problems involving the sam-

pling of family members is also highly relevantto testing in other disorders based on geneticlinkage. Not only does the linkage approachreduce the accuracy of prediction owing torecombination between marker and disease,but many applicants have a pedigree structurethat makes linkage based testing impossible,with key family members dead or unavailable.Most of such subjects would not have enteredthe current series. The problems discussed inthis group raise difficult ethical issues, whoseanswers may not be clear cut; thus whether ornot to use a stored research sample, or toapproach a family member suspected of beingaffected, will require careful individual de-cisions. However, unless the general principlesand the possibility of problems arising areappreciated, serious difficulties could becreated for both the testing centre and thesubjects concerned.The third group of problems, concerned

with confidentiality and disclosure, illustratethe importance not only of scrupulous atten-tion to confidentiality of laboratory and clinicalrecords, but also of a clear policy as to how testresults will be handled. All four centres had apolicy of giving results face to face, never byletter or telephone; the policy of one centre(Leuven) not to generate a written resultavoids the possibility of accidental disclosure.The theoretical risk in this centre of latermisinterpretation or forgetfulness by thosegiven the result verbally was considered to beminimal for two reasons. First the applicantwas always accompanied during the communi-cation of the test result. Secondly, the testresult was discussed again during at least twopost-test counselling sessions. The question ofthird party access to results by bodies, such asinsurance agencies, is an area of considerableconcern. There is no doubt that, in the UK atleast, insurance companies expect to have ac-cess to such results and consider genetic testinformation as no different from other medicaltest results.20The rapid pace of advances in mapping and

isolating disease genes means that already thepossibility of prediction now exists for numer-ous other important disorders of late onset;inherited prion dementias, adult polycystickidney disease, and the familial cancer syn-dromes are examples where this is alreadyavailable.For some of these conditions, specific muta-

tion testing is a possibility, while for otherslinkage based analysis is required, but for all ofthem, whether involving the nervous system ornot, the problems outlined in this paper are ofrelevance.Now that a specific mutation has been iden-

tified for HD,2 with the great majority of casesstudied so far showing a clear expansion of theunstable CTG repeat in the HD gene, it islikely that this will rapidly supersede the use oflinked genetic markers. Some of the ethicaland practical problems encountered in thisstudy will be reduced by the use of mutationanalysis; thus privacy should be enhanced bysamples from multiple family members nolonger being needed. Other problems may be

1034

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from

Ethical and social issues in presymptomatic testing for Huntington's disease

increased; in particular, the specificity andrelative ease of mutation testing may result init being requested by clinicians without proper

information and counselling being given to thesubject, or even without consent. Requestsfrom those at 25% risk, rare in the presentseries, are likely to become more frequent; byno longer requiring the direct involvement ofthe intervening parent, serious conflictsbetween right to privacy and right to know ofthe subjects concerned could arise.We consider it essential that the ethical and

social aspects of HD presymptomatic testingcontinue to be monitored closely, and forproblems as well as benefits arising from use ofmutation analysis to be compared with pre-

vious forms of testing. We hope that the infor-mation from this study, summarising the ex-

perience of four European centres with majorinvolvement of presymptomatic testing, willbe of value in ensuring that this is carried outto the highest possible standard, not only forthose at risk for HD, but for other geneticdisorders of late onset.

We thank our laboratory colleagues in the fourcentres who were responsible for undertakingthe molecular analyses and without whoseinterest and close collaboration this studywould not have been possible. We are gratefulalso to all our colleagues involved in the pre-

and post-test counselling. The study was

funded by an award from the European Com-munity programme on the ethical and socialaspects of human genome research.

1 Harper PS. Huntington's disease. London: Saunders, 1991.2 The Huntington's Disease Collaborative Research Group.

A novel gene containing a trinucleotide repeat that is

expanded and unstable on Huntington's disease chromo-somes. Cell 1993;72:971-83.

3 Duyao M, Ambrose C, Myers R, et al. Trinucleotide repeatlength instability and age of onset in Huntington's dis-ease. Nature Genet 1993;4:387-92.

4 Snell RG, MacMillan JC, Cheadle JP, et al. Relationshipbetween trinucleotide repeat expansion and phenotypicvariation in Huntington's disease. Nature Genet1993;4:393-7.

5 Andrew SE, Goldberg YP, Kremer B, et al. The relation-ship between trinucleotide (CAG) repeat length and clini-cal features of Huntington's disease. Nature Genet1 993;4:398-403.

6 Harding AE. The gene for Huntington's disease. BMJ1993;307:396-7.

7 Harper PS. Clinical consequences of isolating the gene forHuntington's disease. BMJ 1993;307:397-8.

8 Meissen GJ, Myers RH, Mastromauro CA, et al. Predictivetesting for Huntington's disease with use of a linked DNAmarker. N Engl Med 1988;318:535-42.

9 Hayden MR, Robbins C, Allard D, et al. Improved predic-tive testing for Huntington disease by using three linkedDNA markers. Am J Hum Genet 1988;43:689-94.

10 Brandt J, Quaid K, Folstein SE, et al. Presymptomaticdiagnosis of delayed-onset disease with linked DNAmarkers: the experience in Huntington's disease. JAMA1989;261:3108-14.

11 Craufurd D, Dodge A, Kerzin-Storrar L, et al. Uptake ofpresymptomatic predictive testing for Huntington's dis-ease. Lancet 1989;ii:603-5.

12 Brock DH, Mennie M, Curtis A, et al. Predictive testing forHuntington's disease with linked DNA markers. Lancet1989;ii:463-6.

13 Tyler A, Morris M, Lazarou L, Meredith L, Myring J,Harper P. Presymptomatic testing for Huntington's dis-ease in Wales 1987-90. Br

_Psychiatry 1992;161:481-8.

14 Tibben A, Vegter-vd Vlis M, Skraastad MI, et al. Presymp-tomatic DNA testing for Huntington disease in the Neth-erlands. Birth Defects 1992;28(1):127-31.

15 Morris M, Tyler A, Lazarou L, Meredith L, Harper PS.Problems in genetic prediction for Huntington's disease.Lancet 1989;ii:601-3.

16 World Federation of Neurology Research Group on Hunt-ington's Disease. Ethical issues policy statement on Hunt-ington's disease molecular genetics predictive test. J MedGenet 1990;27:34-8.

17 Bloch M, Hayden MR. Predictive testing for Huntington'sdisease in childhood: challenges and implications. Am

_

Hum Genet 1990;46:1-4.18 Demyttenaere K, Evers-Kiebooms G, Decruyenaere M.

Pitfalls in counseling for predictive testing in Huntingtondisease. Birth Defects 1992;28(1):105-1 1.

19 Jacopini GA, D'Amica R, Frontali M, Vivona G. Attitudesof persons at risk and their partners towards predictivetesting. Birth Defects 1992;28(1):113-17.

20 Harper PS. Insurance and genetic testing. Lancet1993;341 :224-7.

1035

on Novem

ber 12, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.30.12.1028 on 1 D

ecember 1993. D

ownloaded from