screening and preventive behaviors one year after predictive genetic testing for hereditary...

Screening and Preventive Behaviors One Year afterPredictive Genetic Testing for Hereditary NonpolyposisColorectal Carcinoma

Veronica Collins, Ph.D.1

Bettina Meiser, Ph.D.2,3

Clara Gaff, Ph.D.4,5

D. James B. St. John, M.D.5,6

Jane Halliday, Ph.D.1

1 Public Health Genetics, Murdoch Childrens Re-search Institute, Parkville, Victoria, Australia.

2 Department of Medical Oncology, Prince of WalesHospital, Randwick, New South Wales, Australia.

3 School of Psychiatry, University of New SouthWales, New South Wales, Australia.

4 Genetic Health Services Victoria, Royal Children’sHospital, Parkville, Victoria, Australia.

5 Familial Cancer Centre, The Royal MelbourneHospital, Victoria, Australia.

6 National Cancer Control Initiative, Carlton, Victo-ria, Australia.

Supported by a three-year Project Grant from TheCancer Council Victoria. Dr. Bettina Meiser is sup-ported by Public Health Australia Fellowship007079 from the National Health and MedicalResearch Council of Australia (NHMRC). Dr. JaneHalliday is supported by an NHMRC Career Devel-opment Award (ID 216741).

The authors thank the following individuals fortheir assistance with patient recruitment, data col-lection, and the ethics application process: Ms.Rebecca Doherty, Ms. Elly Lynch, and Ms. MashaSlattery, from the Familial Cancer Centre, TheRoyal Melbourne Hospital, Victoria, Australia; Mr.Ivan Macciocca from Genetic Health Services Vic-toria, Victoria, Australia; Ms. Mary-Anne Young,Ms. Linda Warwick, Ms. Lisette Curnow, and Ms.Anna Henry from the Familial Cancer Centre, PeterMacCallum Cancer Institute, Melbourne, Australia;Dr. Jack Goldblatt, Dr. Ian Walpole, Ms. JillianKennedy, and Ms. Karen Harrop from GeneticsServices of Western Australia; Dr. Michael Gattas,Dr. Jenny Kromberg, Ms. Julie White, and Ms. KayPodger from Queensland Clinical Genetics Service,Queensland, Australia; Dr. Allan Spigelman, Ms.Bronwyn Burgess, and Ms. Claire Groombridgefrom Hunter Family Cancer Service, New South

Wales, Australia; Dr. Kathy Tucker, Ms. MonicaTucker, and Ms. Margaret Gleeson from the He-reditary Cancer Clinic, Prince of Wales Hospital,New South Wales, Australia; and Ms. RosemaryWarren from Public Health Genetics, Murdoch Chil-drens Research Institute, Victoria, Australia. Theyare also most grateful for the valuable contributionof all the individuals who participated in the cur-rent study.

Address for reprints: Veronica Collins, Ph.D., PublicHealth Genetics, Murdoch Childrens Research In-stitute, 10th Floor, Royal Children’s Hospital, Flem-ington Road, Parkville, Victoria 3052, Australia;Fax: (011) 61-3-8341-6212; E-mail: [email protected]

Received September 14, 2004; revision receivedJanuary 27, 2005; accepted March 22, 2005.

BACKGROUND. Prevention benefits from predictive genetic testing for cancer will

only be fully realized if appropriate screening is adopted after testing. The current

study assessed screening and preventive behaviors during 12 months after predic-

tive genetic testing for hereditary nonpolyposis colorectal carcinoma (HNPCC) in

an Australian clinical cohort.

METHODS. Participants received predictive genetic testing for HNPCC at one of five

Australian familial cancer clinics. Data on self-reported screening behaviors

(colonoscopy, and endometrial sampling and transvaginal ultrasound for women)

and prophylactic surgery (colectomy, and hysterectomy and bilateral oophorec-

tomy for women) were collected using postal questionnaires before (baseline) and

12 months after receipt of genetic test results. Age, gender, perceived risk of cancer,

and cancer-specific distress were assessed as predictors of colonoscopic screening.

RESULTS. In the current study, 114 participants returned baseline questionnaires

(32 carriers and 82 noncarriers of an HNPCC mutation). Ninety-eight participants

also returned a 12-month follow-up questionnaire. Of those � 25 years, 73%

reported having had a colonoscopy before genetic testing. At follow-up, 71% (15 of

25) of carriers and 12% (8 of 65) of noncarriers reported having a colonoscopy in

the 12 months after receipt of test results. The reduction in colonoscopy among

noncarriers was statistically significant (P � 0.001). High perceived risk was asso-

ciated with colonoscopy at baseline. At follow-up, mutation status was the only

variable significantly associated with colonoscopy. Among female mutation carri-

ers, 47% reported having transvaginal ultrasonography and 53% endometrial sam-

pling during follow-up. There was low uptake of prophylactic surgery for colorec-

tal, endometrial, or ovarian carcinomas.

CONCLUSIONS. The majority of individuals reported appropriate screening behav-

iors after predictive genetic testing for HNPCC. The small group of noncarriers who

had screening after genetic testing might benefit from additional counseling.

Cancer 2005;104:273– 8. © 2005 American Cancer Society.

KEYWORDS: hereditary nonpolyposis colorectal carcinoma, screening, genetic test-ing, prospective study, colonoscopy, endometrial carcinoma.

273

© 2005 American Cancer SocietyDOI 10.1002/cncr.21183Published online 9 June 2005 in Wiley InterScience (www.interscience.wiley.com).

The discovery of genes predisposing to hereditarycancer syndromes allows the option of predictive

genetic testing in families where a germline pathologicmutation has been identified. Several mismatch repairgenes are known to predispose to hereditary nonpol-yposis colorectal carcinoma (HNPCC), a condition es-timated to account for � 5% of all colorectal carcino-mas.1,2 The lifetime risk of colorectal carcinoma formen carrying 1 of these mutations is believed to beapproximately 80%.3 For women, the risk of colorectalcarcinoma is lower, but their lifetime risk for endome-trial carcinoma is approximately 45% and ovarian car-cinoma approximately 10%.3,4 Predictive genetic test-ing for cancer may provide benefits. Those notcarrying a mutation can be relieved from intensivescreening, as can their descendants, with associatedsavings to health services.5,6 Individuals carrying amutation can be monitored closely so that preventionand early treatment strategies are implemented. How-ever, there may be psychological harm resulting fromcurrently healthy people knowing that they carry apotentially pathogenic mutation. Moreover, preven-tion benefits will only be realized if appropriate pre-ventive behaviors are adopted or continued after ge-netic testing.

Data on the psychological impact of predictivetesting in hereditary breast and ovarian carcino-mas7–10 and HNPCC11,12 have been accumulating.However, few studies have documented preventivebehaviors after genetic testing and a genetic test resultin itself does not necessarily lead to appropriate risk-reducing behavior.13 Several studies describe screen-ing for breast and/or gynecologic carcinomas aftertesting for BRCA1/BRCA2 mutations,14 –17 despite littleevidence for benefit from screening in carriers of thesemutations. In HNPCC, the efficacy of colonoscopy inreducing incidence and mortality from colorectal car-cinoma has been demonstrated,18 but only two stud-ies report screening practices after predictive genetictesting for HNPCC, neither of which assessed psycho-logical variables.19,20 Moreover, there are no reports ofgynecologic screening behaviors after genetic testingfor HNPCC.

To help individuals adopt appropriate preventivebehaviors, it is important to understand predictors ofcompliance. In the general population, factors relatedto compliance with colorectal carcinoma screeninginclude physician recommendations, perceived bene-fits from screening, age, having relatives with colorec-tal carcinoma, perceived susceptibility to colorectalcarcinoma, perceptions of self-efficacy, and salienceand coherence of screening.21–23 Studies of preventivebehaviors in high-risk populations, after predictivetesting for BRCA1/BRCA214 –17,24 or HNPCC,19 suggest

that the main predictors of uptake of screening andprophylactic surgery are genetic test result and rec-ommendations from health professionals.

Theoretic models of compliance with health be-haviors can be useful to consider in understandinguptake of preventive behaviors. Self-regulation theory,in this context, would suggest that individuals who aremotivated to be reassured will engage in screening toreduce emotional distress.25 However, others maycope with distress by avoiding screening if they areafraid of what might be revealed. Guided by self-reg-ulation theory and using validated psychological mea-sures, we have followed individuals from a clinicalcohort undergoing predictive genetic testing forHNPCC, for a period of 12 months. Our findings re-lated to the psychological impact of predictive genetictesting have been published.12 In the current reportwe describe screening and preventive behaviors be-fore and after genetic testing.

MATERIALS AND METHODSParticipants and RecruitmentThe sample included individuals undergoing predic-tive genetic testing at 1 of 5 familial cancer clinics inAustralia, between June 1998 and August 2002. Thestudy was approved by five institutional ethics com-mittees and signed informed consent was obtainedfrom all participants. Only individuals who had neverhad colorectal carcinoma or any of the cancers asso-ciated with HNPCC were included in the study. Indi-viduals were ineligible if they were unable to giveinformed consent or had limited literacy in English,because data were collected using self-report ques-tionnaires. Staff of the familial cancer clinics invitedindividuals to participate in the study during the pre-clinic telephone call. All subsequent correspondencewas via the coordinating research center. The baselinequestionnaires and consent forms were sent with areply-paid envelope as soon as the research centerreceived recruitment details from the familial cancercenters. Participants were asked to return the com-pleted questionnaire and a consent form before at-tending their appointment. Follow-up questionnaireswere mailed 2 weeks, 4 months, and 12 months afterparticipants received their genetic test results.

Genetic Counseling and TestingPredictive genetic testing is only offered to familymembers once a pathogenic HNPCC mutation hasbeen identified in an individual affected by cancer. InAustralia, most of those wanting predictive testingattend a familial cancer service for at least one sessionof genetic counseling before collection of blood spec-imens.26 The predictive test result is then given to the

274 CANCER July 15, 2005 / Volume 104 / Number 2

individual within a genetic counseling protocol, usu-ally by the person who has conducted the initial coun-seling. Screening recommendations, based on the testresult and any other specific medical details, are givenby the relevant medical specialist. Decisions regardingthe timing and arrangements for screening are madeat the clinic in consultation with the person receivingthe test result. Letters outlining the decisions and rec-ommended actions are then sent to the individualsconcerned, as well as to other practitioners involved intheir care.

Screening RecommendationsThe Australian guidelines for the management of colo-rectal carcinoma form the basis of screening recom-mendations for those at risk of HNPCC.27 Annualcolonoscopy from the age of 25 years or 5 years earlierthan the youngest affected family member is recom-mended for those carrying a mutation. Those not car-rying the mutation identified in an affected familymember are advised to follow population screeningguidelines, i.e., annual (or biennial) fecal occult bloodtesting possibly with 5-yearly sigmoidoscopy, from theage of 50 years. Due to the lack of proven benefit, theguidelines for endometrial or ovarian carcinomascreening in women carrying a mutation for HNPCCare not as prescriptive as for colorectal carcinomascreening. Endometrial screening can be consideredfrom the age of 30 –35 years,27 or even 25 years.28 Thefamilial cancer clinics generally recommend annualtransvaginal ultrasound and endometrial sampling inpremenopausal women, and annual transvaginal ul-trasound and CA 125 testing for postmenopausalwomen. Preventive colectomy is not recommendedroutinely for those carrying a mutation but may bediscussed as an option, depending on individual cir-cumstances. For women carrying a mutation, prophy-lactic hysterectomy and oophorectomy might also beconsidered from the age of 30 –35 years, or whenchildbearing is complete.27

MeasuresDemographic characteristicsAt baseline, age, gender, educational level, maritalstatus, and number of children were recorded.

Mutation carrier riskTo provide an estimate of risk before genetic test-

ing, individuals were given a 50%, 25%, 12.5%, or6.25% risk of carrying a mutation, based on their re-lationship to the person in whom the mutation wasidentified. Predictive genetic testing is generally onlyoffered to those at 50% risk of carrying a mutation(first-degree relative). However, others at 25%, 12.5%,

or 6.25% mutation carrier risk might be offered pre-dictive testing if the intervening person/s were de-ceased or if there was some other reason why theperson at 50% risk had not been tested.

Impact of event scaleThis is a 15-item validated scale that measures intru-sive and avoidant thoughts about a specific stressfulevent29 and has been validated for use with womenwith a family history of breast carcinoma.30 In thecurrent study, the particular stressor was concernabout being at risk of developing colorectal carci-noma. Total scores (range, 0 –75) were used in theanalysis. The Impact of Event Scale was included inthe questionnaires at each time point.

Risk perceptionTwo questions assessed subjective perception of riskfor colorectal carcinoma at each time point. One ques-tion asked participants to indicate on a scale from 0(no chance) to 100 (definitely) what they believe istheir chance of developing bowel carcinoma. Theother question asked whether individuals perceivedtheir risk as much lower, lower, the same, higher, ormuch higher than the general population. Womenwere asked similar questions to assess perceived riskof endometrial carcinoma.

Beliefs about screeningThe baseline questionnaire included questions to as-sess beliefs about bowel carcinoma and effectivenessof screening. These were based on constructs postu-lated by the health belief model and self-motivationtheory, previously used in a model to describe adher-ence to colorectal carcinoma screening.21 Similarquestions were included in the 12-month follow-upquestionnaire to assess motivations for undergoingscreening, or not, in the previous year. In addition,there were questions on beliefs about genetic testresults (see Table 4). Women were asked the samequestions in relation to endometrial carcinoma.

Outcome MeasuresScreening behaviorsInformation on colonoscopy was collected by self-report on the baseline (pretest result) questionnaireand on the questionnaire that was sent 12 monthsafter receiving the predictive genetic test result.Women were asked similar questions regarding trans-vaginal ultrasound or endometrial sampling for endo-metrial or ovarian carcinoma screening.

Preventive Behaviors in HNPCC/Collins et al. 275

Prophylactic surgeryAt baseline, a question assessed interest in prophylac-tic colectomy should a mutation be identified. In the12-month follow-up questionnaire, participants wereasked whether a prophylactic colectomy had beenperformed or if there was an intention to have surgery.Women were asked similar questions regarding pro-phylactic hysterectomy and bilateral oophorectomy.

Statistical AnalysisData were analysed using SPSS 11.5 (Statistical Pro-gram for the Social Sciences, Chicago, IL). Carriers andnoncarriers were compared with respect to sociode-mographic variables at baseline, using chi-square testsfor categorical variables and independent samples ttests for continuous data. Analyses of use of colonos-copy and variables associated with colonoscopy in-cluded only participants � 25 years at baseline, therecommended age for beginning colonoscopic screen-ing in most families with HNPCC. The proportionreporting screening at follow-up was compared be-tween carriers and noncarriers with a chi-square test.When an expected cell frequency was less than five,Fisher exact tests were used to compare proportions.

Bivariate associations between screening behav-iors and possible predictors (see Tables 2 and 6 forvariables included) were examined. Significance ofdifferences between groups was assessed using t testsor Mann–Whitney U tests, as appropriate, for contin-uous variables, and chi-square tests for categoricalvariables. These bivariate analyses were followed bylogistic regressions using a backward elimination pro-cess. All variables with a bivariate association with P� 0.2 were entered into the regression model andprogressively eliminated until the only remaining vari-ables were those with P � 0.05. Separate analyses weredone to assess factors independently associated withcolonoscopy reported at baseline and at 12 months offollow-up. The proportion of participants reportingscreening behaviors at baseline was compared withthe proportion at 12 months of follow-up using theMcNemar test for related samples, for carriers andnoncarriers separately.

RESULTSThe SampleOne hundred thirty-four individuals were invited toparticipate and 114 (85.1%) returned the baselinequestionnaire before attending the familial cancerclinic. Of these 114 participants, 32 were identified ascarriers and 82 as noncarriers of an HNPCC mutation.Twelve months later, 98 (86.0%) participants returnedthe follow-up questionnaire, including 84.4% of carri-

ers and 86.6% of noncarriers. A comparison of the 98responders to the follow-up questionnaire to thoselost to follow-up (n � 16) showed no statistically sig-nificant differences in carrier status, demographic orpsychological variables, or the proportion that hadever had a colonoscopy (reported at baseline). Partic-ipants received test results 1 week to 19 months aftercompletion of the baseline questionnaire (a median of6 weeks). Table 1 shows that noncarriers were older,more likely to be married or in a de facto relationship,and were more likely to have children. Eighty-onepercent of carriers and 93% of noncarriers were � 25years at baseline (n � 102).

Baseline Colonoscopic ScreeningAt baseline, 3 of 12 (25%) of those � 25 years and 74 of101 (73%) of those � 25 years reported ever having acolonoscopy. (Note that 1 person � 25 years did notreport a baseline colonoscopy history.) For those � 25years, the only factor significantly associated with everhaving a colonoscopy before receiving a genetic testresult was perceived risk of developing bowel carci-noma (0 –100 scale) (Table 2). Using logistic regres-sion, perceived risk of developing bowel carcinomawas still the only variable significantly associated withever having a colonoscopy (odds ratio [OR] � 1.03;95% confidence interval [CI], 1.00 –1.05; P � 0.03).

TABLE 1Sociodemographic Characteristics, Mutation Carrier Risk,and Family History of Carriers and Non-carriers

VariablesCarriers(n � 32)

Non-carriers(n � 82)

Mean age (yrs) 36.8 43.1a

No. (%) No. (%)

Gender female 21 (65.6) 48 (58.5)Age � 25 yrs at baseline 26 (81.2) 76 (92.7)Have children 21 (65.6) 68 (82.9)a

Post-school education 20 (71.4) 51 (64.6)Married-/-de facto 18 (56.3) 68 (82.9)a

Pre-test mutation carrier risk50% 29 (90.6) 60 (73.2)a

25% 3 (9.4) 19 (23.2)12.5% 0 (0.0) 1 (1.2)6.25% 0 (0.0) 2 (2.4)

Family history of cancerColorectal carcinoma only 12 (37.5) 39 (46.4)Endometrial carcinoma included 16 (50.0) 26 (31.0)Ovarian carcinoma included 3 (9.4) 9 (10.7)Endometrial and ovarian carcinoma included 1 (3.1) 10 (11.9)

yrs: years.a P � 0.05 in a comparison of carriers and non-carriers.


Colonoscopy at 12-Month Follow-UpEighty-eight individuals completed the follow-upquestionnaire, and all were � 25 years. Table 3 showsthat 71% of carriers reported having had a colonos-copy examination in the 12 months since receivingtheir genetic test result. Among noncarriers, 8 (12%)reported a colonoscopy in the same period, a signifi-cantly lower proportion compared with the carriergroup (�2 � 28, P � 0.001).

When asked to endorse statements describing var-ious reasons that may have influenced the decision tohave a colonoscopy, there was little difference in theresponses of carriers compared with noncarriers (Ta-

ble 4). The reasons chosen as important included be-lieving the risk for bowel carcinoma was high, wantingreassurance, and believing colonoscopy was effectivein finding bowel carcinoma early. Additional reasonsoffered by noncarriers included having a high-riskfamily history, having had bowel problems in the past,and helping to decide whether frequency of colonos-copy should be 3 or 5 years apart.

The predominant reason for noncarriers not hav-ing a colonoscopy was the belief that their risk was low(Table 5). This was not the case for 6 carriers who didnot have a colonoscopy in the 12 months after receiv-ing their test result, none of whom endorsed the rea-son that their risk was low. Four of the six carriers whodid not have colonoscopy offered their own reasons:two were pregnant, one was undergoing biennialcolonoscopy, and the other was waiting for notifica-tion from the physician.

Table 6 shows bivariate associations with poten-tial predictors of colonoscopy in the 12 months afterreceipt of the test result. Significantly higher propor-tions of those who reported having colonoscopy weremutation carriers or women. This group also hadhigher levels of perceived risk (using either risk mea-sure) and higher scores on the Impact of Event Scale,compared with those who didn’t have a colonoscopy.However, logistic regression analysis of predictors ofcolonoscopy reported at follow-up showed carrier sta-tus to be the only significant predictor (OR � 20; 95%CI, 5.8 – 68; P � 0.001).

TABLE 2Bivariate Associations between Colonoscopy Reported at Baseline andPossible Predictors Measured at Baseline (Age > 25 Yr)

Characteristics

Ever had a colonoscopy beforebaseline questionnaire

Yes(n � 74)

No(n � 27) P value

Females (%) 59.5 51.9 0.49a

Percentage at 50% risk 79.7 63.0 0.08a

Percentage with perceived risk “muchhigher” than general population 31.5 29.6 0.86a

Mean age (yrs) 44.2 41.6 0.33b

Mean perceived lifetime risk of bowelcancer (0–100) 49.2 39.8 0.04b

Median Impact of Event Scale totalscore 7.0 1.0 0.09c

yrs: years.a Chi-square test.b t test.c Mann–Whitney U-test.

TABLE 3Colonoscopy in the Previous Year, Reported on the 12-MonthPost-test Result Questionnaire, According to Colonoscopyat Baseline (Age > 25 Yrs at Baseline)

Ever had a colonoscopybefore baseline

Colonoscopy in past yr(recorded at follow-up)

Carriers No. (%)Yes (n � 16) 11 (68.8)No (n � 5) 4 (80.0)All carriers (n � 21) 15 (71.4)

NoncarriersYes (n � 48) 8 (16.7)No (n � 17) 0 (0)All non-carriers (n � 65) 8 (12.3)

yrs: years.a Note that only 86 of 88 eligible participants are included. One person did not report their colonoscopy

history at baseline and one reported “don’t know” to the colonoscopy question on the follow-up

questionnaire.

TABLE 4Factors Influencing Decisions to Have Colonoscopyin the Past 12 Months According to Carrier Status,Recorded at 12 Months of Follow-Up

Importance of factor in influencingdecision to have colonoscopy

Percentage choosing somewhator definitelya

Carriers(n � 15)


Believe risk for bowel carcinoma is high 100 100Believe gene tests for HNPCC are

inaccurate 20.0 37.5Wanted to be reassured did not have

bowel carcinoma 100 100Believe colonoscopy is effective in

finding bowel carcinoma early 100 100Colonoscopy is an easy thing for me to

do 78.6 57.2

HNPCC: hereditary nonpolyposis colorectal carcinoma.a Response options were: 1, definitely not a factor; 2, somewhat influenced my decision; 3, yes,

definitely an important factor.


Comparison of Baseline with Follow-UpIn Table 3, the proportion of carriers who reportedhaving colonoscopy at baseline was not significantlydifferent to the proportion having a colonoscopy bythe 12-month follow-up (76% [16 of 21] at baseline vs.71% [15 of 21] at the 12-month follow-up, P � 0.75).For noncarriers, there was a significant reduction in

colonoscopic screening between baseline (48 of 65[74%]) and follow-up (8 of 65 [12%]; P � 0.001).

Gynecologic ScreeningAt baseline, 19 of 63 women (30%) reported ever hav-ing a transvaginal ultrasound, and 4 of 58 women (7%)reported ever having endometrial sampling (somewomen were not eligible for endometrial samplingbecause they had received a hysterectomy). Beliefsabout the severity or curability of endometrial carci-noma, worry about the possibility of finding a cancerwith screening, or the effectiveness or ease of havingscreening were not associated with reported gyneco-logic screening at baseline. At 12 months of follow-up,47% (8 of 17) of carriers reported having had transvag-inal ultrasound and 53% (9 of 17) reported having hadendometrial sampling since receiving their genetictest result. The proportions of the 39 noncarriers re-porting either transvaginal ultrasound (10%) or endo-metrial sampling (5%) were significantly lower thanfor carriers (P � 0.004 and P � 0.001 for each proce-dure, respectively). All carriers who had endometrialsampling believed that their risk of endometrial carci-noma was high and wanted reassurance that they didnot have endometrial carcinoma. Eight of the ninecarriers also believed that endometrial sampling waseffective at detecting endometrial carcinoma early andthat it was an easy thing for them to do. Neither of thetwo noncarriers who reported endometrial samplingbelieved that their risk for endometrial carcinoma washigh, but they indicated that the other factors given asoptions (reassurance, effectiveness of endometrialsampling, and endometrial sampling was easy to do)were important in the decision.

Eight female carriers reported not having endo-metrial sampling by the 12-month follow-up. Threewere in their early 20s and had not yet been advised tohave gynecologic screening. Another two carrierwomen said that gynecologic tests had not been rec-ommended by their physicians, whereas another be-lieved that endometrial sampling would be difficult todo and was worried that it might show cancer. Of thenoncarriers, the main reasons given for not havingendometrial sampling was that their risk for endome-trial carcinoma was low, or that it had not been rec-ommended by their physician.

Prophylactic SurgeryAt baseline, no one reported having a prophylacticcolectomy, but 6 (5%) individuals said they wouldconsider a colectomy if found to be carriers of a mu-tation. By 12 months after receiving genetic test re-sults, none had undergone a prophylactic colectomyalthough 1 person indicated an intention to do so. At

TABLE 6Bivariate Associations between Colonoscopy Reported at 12 Monthsof Follow-Up and Possible Predictors (> 25 Yrs at Baseline)

Characteristics

Had a colonoscopy inpast yr

P valueYes(n � 23)

No(n � 64)

Carriers (%) 65.2 9.4 � 0.001a

Females (%) 78.3 53.1 0.035a

Screened at baseline (%) 82.6 71.4 0.293a

Perceived risk “much higher” thanpopulation (at 12 mos) (%) 47.8 9.5 � 0.001a

Mean age at baseline (yrs) 40.1 44.6 0.095b

Mean perceived lifetime risk (0–100) (at 12 mos) 64.8 37.5 � 0.001b

Median Impact of Event Scale totalscore (at 12 mos) 7.0 0.0 0.001c

yrs: years; mos: months.a Chi-square test.b t test.c Mann–Whitney U-test.

TABLE 5Factors Influencing Decisions to not Have Colonoscopyin the Past 12 Months According to Carrier Status, Recordedat 12 Months of Follow-Up

Importance of factor in influencingdecision to not have colonoscopy

Percentage choosing somewhator definitelya

Carriers(n � 6)


Believe risk for bowel carcinoma is low 0 69.1b

Believe gene tests for HNPCC areinaccurate 0 9.4

Worried colonoscopy will show bowelcarcinoma 0 1.9

Believe colonoscopy ineffective infinding bowel carcinoma early 16.7 3.8

Colonoscopy is a difficult thing for meto do 16.7 13.4

HNPCC: hereditary nonpolyposis colorectal carcinoma.a Response options were: 1, definitely not a factor; 2, somewhat influenced my decision; 3, yes,

definitely an important factor.b P � 0.001, Fisher exact test comparing carriers with noncarriers.


baseline, two women reported having had an oopho-rectomy and five reported having had a hysterectomy,but it is likely that these were done for reasons otherthan cancer prevention. Three of 65 individuals indi-cated at baseline that they would consider a prophy-lactic hysterectomy and a prophylactic bilateral oo-phorectomy if found to be carriers. In the 12 monthsafter receiving the test result, none had gone aheadwith a prophylactic hysterectomy but 2 women whoreported having had a hysterectomy on the baselinequestionnaire, reported having bilateral oophorec-tomy since receiving genetic test results.

DISCUSSIONThis is only the second report documenting the effectof predictive genetic test results on screening behav-iors in both unaffected carriers and noncarriers ofHNPCC mutations, and the first in a clinic-based sam-ple that also reports on gynecologic screening.

The baseline participation rate in our study washigh (85%), but some individuals did not complete the12-month follow-up questionnaire. However, thisgroup was small (n � 16) and their baseline charac-teristics were similar to those of the participants, sug-gesting that this should not have resulted in biasedfindings. Data on screening were based on self-report,which has been shown to be fairly reliable for colonos-copy,31,32 but there are no data published on the reli-ability of self-reported endometrial sampling or trans-vaginal ultrasound.

A notable feature of the study sample was thesmaller number of carriers than was expected basedon pretest mutation risks. This was because carrierswho had already developed cancers associated withHNPCC were not eligible to participate and this isreflected in the younger mean age of carriers. The agedifferences probably explain the lower proportion ofcarriers who were married or in a de facto relation-ship, or had children.

Our study showed that the majority of individualsadopted appropriate colorectal carcinoma screeningin the year after receipt of HNPCC mutation test re-sults. It has been suggested that individuals found tobe noncarriers might be reluctant to stop screeningbecause of residual anxieties about their risk for can-cer, or failure to fully understand the significance ofthe genetic test result.6,33–35 This was not strongly sup-ported by our data, as only a small proportion (12%) ofnoncarriers reported having a colonoscopy in the yearafter receipt of their test result.

The findings in two previous studies on screeningpractices after predictive genetic testing for HNPCCdiffered from our own.19,20 One included researchstudy participants unaffected by cancer. Adherence to

screening guidelines was slightly better in mutationnoncarriers than carriers, although nonadherence in-cluded overscreening as well as underscreeningamong carriers.19 The primary aim of the second studywas to assess uptake of genetic testing, although ahigh uptake of colonoscopy in mutation carriers wasreported as a secondary finding.20 Screening in non-carriers was not fully described in that article andneither study assessed psychological variables.

Noncarriers in our study who reported having acolonoscopy indicated that a belief that their risk ofdeveloping bowel carcinoma was high, and wantingreassurance that they did not have bowel carcinoma,influenced their decision to have screening, lendingsupport to self-regulation theory.25 It is noteworthythat these noncarriers appeared to be aware that theirrisk was that of the general population (ascertainedfrom another question in the questionnaire), but theynonetheless opted for a colonoscopy. These findingsindicate that a few noncarriers who undergo colonos-copy to reduce distress about their risk may benefitfrom interventions aimed at decreasing anxiety, suchas supportive counseling or brief problem-solving in-terventions,36 or communication tools designed tocorrect inflated risk perceptions.37

The main change in screening practices after re-ceipt of test results was a significant decrease in theproportion of noncarriers reporting a colonoscopyfrom baseline to 12 months, whereas for carriers, theproportion reporting colonoscopy did not change sig-nificantly. Similar findings have been reported in stud-ies of HNPCC19 and BRCA1/BRCA2 testing.14 Beforegenetic testing, most of the participants had a 50% riskof carrying a mutation with similar screening recom-mendations to those for mutation carriers. However,the recommendation for those found to be noncarri-ers is to reduce screening considerably—in the ab-sence of other indications—from what was recom-mended before testing. For carriers, screening is ofcourse recommended and the finding that 4 of the 5carriers who had not had a colonoscopy before base-line reported having colonoscopy during the 12-month period after testing suggests that the genetictest result had prompted this screening procedure.

At baseline, the perceived risk of developing colo-rectal carcinoma was the only variable found to besignificantly associated with ever having screening,confirming previous findings of the association be-tween perceived risk and colorectal carcinoma screen-ing in the general population21–23 and studies ofscreening behavior after predictive testing for BRCA1/BRCA2.17 By the 12-month follow-up, carrier statuswas the major factor associated with screening, whichcorresponds with findings from studies on predictors


of screening after genetic testing for HNPCC19 andBRCA1/BRCA2 mutations.14 Once this variable wasentered into the logistic regression, the perceived riskand Impact of Event Scale score were not significant,probably due to the high correlation between thesevariables and mutation carrier status.

The few individuals who were either carrying amutation and did not have a colonoscopy during fol-low-up or were not carrying a mutation and did havea colonoscopy between baseline and follow-up (de-fined loosely as nonadherence) precluded identifyingpredictors of adherence. That there were few nonad-herers indicates that the mutation test result was themost important predictor. It is difficult to separate theeffect of the test result itself from the accompanyinggenetic counseling but the combination appeared tobe effective in motivating appropriate screening be-haviors.

Despite sample size limitations, the reported fre-quency of endometrial sampling showed a significantincrease from baseline in those identified as mutationcarriers. Although some of the carriers who reportednot having endometrial sampling in the 12 monthsafter testing gave reasons such as being too young,others did not give a reason or said that the physicianhad not recommended it. As the screening guidelinesare not prescriptive and are evolving over time, it ispossible that different health care providers havegiven varying advice regarding gynecologic screening.Furthermore, in some areas outside of the major cit-ies, gynecologic screening services are not easily ac-cessible. The data did not provide much insight intoreasons why noncarriers received endometrial carci-noma screening but at least they did not indicate ahigh perceived risk of endometrial carcinoma. Gyne-cologic screening behaviors after genetic testing forHNPCC have not been reported previously, althoughone small study of people with a clinical diagnosis ora family history of HNPCC (without genetic testing)reported inadequate screening for both colorectal andgynecological carcinomas.38

There was no uptake of prophylactic colectomy inmutation carriers 12 months after receipt of test re-sults. Clinical guidelines recommend that prophylac-tic surgery should only be discussed as an optiondependent on individual circumstances.27 The currentstudy showed low uptake of these preventive mea-sures but whether this is due to the preference of theindividuals at risk, or whether these measures are notbeing presented as an option by health professionals,is not known. Longer follow-up may also show greateruptake because decisions concerning surgery can in-volve lengthy deliberation. Interest in bilateral pro-phylactic oophorectomy was limited, and there was

no apparent interest in prophylactic hysterectomy. Incontrast, uptake of bilateral prophylactic oophorec-tomy and mastectomy among women carrying aBRCA1/BRCA2 mutation is high,39,40 reflecting the lowefficacy of screening, the proven efficacy of prophy-lactic surgery,41,42 and the high lifetime risk of ovariancarcinoma among BRCA1 mutation carriers.

In summary, the current study has demonstrateda high uptake of colonoscopy (71%) in the 12 monthsafter receipt of a predictive genetic test result, in thoseidentified as carriers of an HNPCC mutation. Uptakeof endometrial carcinoma screening in carriers ofHNPCC mutations also was reasonably high withinthe first 12 months after genetic testing (approxi-mately 50%). There was a small group of noncarrierswho continued to receive screening despite beingfound not to have an HNPCC mutation. This groupmay benefit from additional counseling or educationalinterventions. Overall, the study has shown that pre-dictive genetic testing for HNPCC is leading to com-pliance with screening recommendations for the ma-jority of patients.

REFERENCES1. Lynch HT, de la Chapelle A. Genetic susceptibility to non-

polyposis colorectal cancer. J Med Genet. 1999;36:801– 818.2. Aaltonen LA, Salovaara R, Kristo P, et al. Incidence of he-

reditary non-polyposis colorectal cancer and feasibility ofmolecular screening for the disease. N Engl J Med. 1998;338:1481–1487.

3. Vasen HF, Wijnen JT, Menko FH, et al. Cancer risk in familieswith hereditary nonpolyposis colorectal cancer diagnosed bymutation analysis. Gastroenterology. 1996;110:1020–1027.

4. Dunlop MG, Farrington SM, Carothers AD, et al. Cancer riskassociated with germline DNA mismatch repair gene muta-tions. Hum Mol Genet. 1997;6:105–110.

5. Stanley AJ, Gaff CL, Aittomaki AK, Fabre L, Macrae FA, St.John DJB. Value of predictive genetic testing in managementof hereditary non-polyposis colorectal cancer (HNPCC).Med J Aust. 2000;172:313–316.

6. Watson P, Narod SA, Fodde R, et al. Carrier risk statuschanges resulting from mutation testing in hereditary non-polyposis colorectal cancer and hereditary breast-ovariancancer. J Med Genet. 2003;40:591–596.

7. Croyle RT, Smith KR, Botkin JR, Baty B, Nash J. Psycholog-ical responses to BRCA1 mutation testing: preliminary find-ings. Health Psychol. 1997;16:63–72.

8. Lerman C, Narod S, Schulman K, et al. BRCA1 testing infamilies with hereditary breast-ovarian cancer: a prospec-tive study of patient decision making and outcomes. JAMA.1996;275:1885–1892.

9. Lodder L, Frets PG, Trijsburg RW, et al. Psychological impactof receiving a BRCA1/BRCA2 test result. Am J Med Genet.2001;98:15–24.

10. Meiser B, Butow P, Friedlander M, et al. Psychological im-pact of genetic testing in women from high-risk breast can-cer families. Eur J Cancer. 2002;38:2025–2031.


11. Aktan-Collan K, Haukkala A, Mecklin J-P, Uutela A, Kaari-ainen H. Psychological consequences of predictive genetictesting for hereditary non-polyposis colorectal cancer(HNPCC): a prospective follow-up study. Int J Cancer. 2001;93:608 – 611.

12. Meiser B, Collins V, Warren R, et al. Psychological impact ofgenetic testing for hereditary nonpolyposis colorectal can-cer (HNPCC). Clin Genet. 2004;66:502–511.

13. Marteau T, Lerman C. Genetic risk and behavioural change.Br Med J. 2001;322:1056 –1059.

14. Lerman C, Hughes C, Croyle RT, et al. Prophylactic surgerydecisions and surveillance practices one year followingBRCA1/2 testing. Prev Med. 2000;31:75– 80.

15. Botkin JR, Smith KR, Croyle RT, et al. Genetic testing for aBRCA1 mutation: prophylactic surgery and screening be-haviour in women 2 years post testing. Am J Med Genet.2003;118A:201–209.

16. Scheuer L, Kauff N, Robson M, et al. Outcome of preventivesurgery and screening for breast and ovarian cancer inBRCA mutation carriers. J Clin Oncol. 2002;20:1260 –1268.

17. Tinley ST, Houfek J, Watson P, et al. Screening adherance inBRCA1/2 families is associated with primary physician’sbehavior. Am J Med Genet. 2004;125A:5–11.

18. Jarvinen HJ, Aarnio M, Mustonen H, et al. Controlled 15-year trial on screening for colorectal cancer in families withhereditary nonpolyposis colorectal cancer. Gastroentero-ogyl. 2000;118:829 – 834.

19. Hadley DW, Jenkins JF, Dimond E, de Carvalho M, Kirsch I,Palmer CG. Colon cancer screening practices after geneticcounseling and testing for hereditary nonpolyposis colorec-tal cancer. J Clin Oncol. 2004;22:39 – 44.

20. Ponz de Leon M, Benatti P, Di Gregorio C, et al. Genetictesting among high-risk individuals in families with hered-itary nonpolyposis colorectal cancer. Br J Cancer. 2004;90:882– 887.

21. Myers RE, Ross E, Jepson C, et al. Modeling adherence tocolorectal cancer screening. Prev Med. 1994;23:142–151.

22. Vernon S. Participation in colorectal-cancer screening: areview. J Natl Cancer Inst. 1997;89:1406 –1422.

23. Levin B, Smith RA, Feldman GE, et al. Promoting earlydetection tests for colorectal carcinoma and adenomatouspolyps: a framework for action: the strategic plan of theNational Colorectal Cancer Roundtable. Cancer. 2002;95:1618 –1628.

24. Peshkin BN, Schwartz MD, Isaacs C, Hughes C, Main D,Lerman C. Utilization of breast cancer screening in a clini-cally based sample of women after BRCA1/2 testing. CancerEpidemiol Biomark Prev. 2002;11:1115–1118.

25. Leventhal H, Zimmerman R, Gutmann M. Compliance: aself-regulation perspective. In: Doyle Gentry W, editor.Handbook of behavioral medicine. New York: GuilfordPress, 1984:369 – 436.

26. Human Genetics Society of Australasia. Guidelines for DNApresymptomatic and predictive testing for genetic disorders.Human Genetics Society of Australasia, 2002. Available fromURL: http://www.hgsa.com.au.HCSAPolicies [accessed May2005].

27. National Health and Medical Research Council. Clinical

practice guidelines. The prevention, early detection andmanagement of colorectal cancer. Canberra: NationalHealth and Medical Research Council, 1999.

28. Brown GJE, St. John DJB, Macrae FA, Aittomaki K. Cancerrisk in young women at risk of hereditary nonpolyposiscolorectal cancer: implications for gynecologic surveillance.Gynecol Oncol. 2001;80:346 –349.

29. Horowitz M, Wilner N, Alvarez W. Impact of event scale: ameasure of subjective stress. Psychosom Med. 1979;41:209 –218.

30. Thewes B, Meiser B, Hickie IB. Psychometric properties ofthe Impact of Event Scale amongst women at increased riskfor hereditary breast cancer. Psycho-Oncology. 2001;10:459 –468.

31. Schoen RE, Weissfeld JL, Trauth JM, Ling BS, Hayran M. Apopulation-based, community estimate of total colon exam-ination: the impact on compliance with screening for colo-rectal cancer. Am J Gastroenterol. 2002;97:446 – 451.

32. Baier M, Calonge N, Cutter G, et al. Validity of self-reportedcolorectal cancer screening behavior. Cancer Epidemiol Bi-omark Prev. 2000;9:229 –232.

33. Michie S, Weinman J, Miller J, Collins V, Halliday J, MarteauTM. Predictive genetic testing: high risk expectations in theface of low risk information. J Behav Med. 2002;25:33–50.

34. Michie S, Smith JA, Senior V, Marteau TM. Understandingwhy negative genetic test results sometimes fail to reassure.Am J Med Genet. 2003;119A:340 –347.

35. Bleiker EMA, Menko FH, Taal BG, et al. Experience of dis-charge from colonoscopy of mutation negative HNPCC fam-ily members [electronic leter]. J Med Genet. 2003;40:e55.

36. Schwartz MD, Lerman C, Audrain JA. The impact of a briefproblem-solving training intervention for relatives of re-cently diagnosed breast cancer patients. Ann Behav Med.1998;20:7–12.

37. Lobb E, Butow P, Tucker K, et al. A communication aid tofacilitate risk communication with women from high-riskbreast cancer families [abstract]. Psycho-Oncology.2003;12(4 Suppl):S258.

38. Stoffel EM, Garber JE, Grover S, Russo L, Johnson J, Syngal S.Cancer surveillance is often inadequate in people at highrisk for colorectal cancer [electronic letter]. J Med Genet.2003;40:e54.

39. Tiller K, Meiser B, Butow P, et al. Psychological impact ofprophylactic oophorectomy in women at increased risk ofdeveloping ovarian cancer: a prospective study. Gynecol On-col. 2002;86:212–219.

40. Hatcher MB, Fallowfield L, A’Hern R. The psychosocial im-pact of bilateral prophylactic mastectomy: prospectivestudy using questionnaires and semistructured interviews.Br Med J. 2001;322:76 –79.

41. Meijers-Heijboer H, van Geel B, van Putten WLJ, et al. Breastcancer after prophylactic bilateral mastectomy in womenwith a BRCA1 or BRCA2 mutation. N Engl J Med. 2001;345:159 –164.

42. Rebbeck TR, Lynch HT, Neuhausen SL, et al. Prophylacticoophorectomy in carriers of BRCA1 or BRCA2 mutations.N Engl J Med. 2002;346:1616 –1622.


screening and preventive behaviors one year after predictive genetic testing for hereditary...

Documents