Development of Colorectal Tumors in ... - Gastroenterology

GASTROENTEROLOGY 2007;133:1093–1098

JUKKA–PEKKA MECKLIN,* MARKKU AARNIO,* ESA LÄÄRÄ,‡ MATTI V. KAIRALUOMA,* KIRSI PYLVÄNÄINEN,* PÄIVI PELTOMÄKI,§ LAURI A. AALTONEN,§ and HEIKKI J JÄRVINEN㛳 *Department of Surgery, Jyväskylä Central Hospital, Jyväskylä, Finland; ‡Department of Mathematical Sciences, University of Oulu, Oulu, Finland; §Department of Medical Genetics, Biomedicum Helsinki, Helsinki, Finland; and the 㛳Department of Surgery, Helsinki University Central Hospital, Helsinki, Finland

See Durno C et al on page 1190 in the October 2007 issue of CGH.

Background & Aims: Mutation carriers in Lynch syndrome families have a high risk for developing colorectal cancer during their lifetime. This study was designed to assess the cumulative risk for the development of colorectal adenoma or carcinoma in prospective colonoscopic surveillance. Methods: Data from the Finnish Hereditary Colorectal Cancer Registry electronic database on 420 Lynch syndrome mutation carriers without previous colorectal tumors were reviewed. Between March 1982 and May 2005 the mutation carriers underwent a total of 1252 colonoscopies. The total follow-up time was 3150 years (mean, 6.7 y/patient). Results: The cumulative risk of adenoma by age 60 was estimated as 68% (95% confidence interval [CI], 50%– 80%) in men and 48% (95% CI, 29%– 62%) in women. The estimated cumulative risk up to age 60 years for the development of cancer found as a result of surveillance at an interval of 2–3 years was 35% (95% CI, 16%– 49%) in men and 22% (95% CI, 7%–34%) in women. Half of the adenomas were located proximal to the splenic flexure. Extracolonic cancer was diagnosed in 73 patients (18%). Conclusions: Adenoma would appear to be the most important lesion preceding cancer formation in Lynch syndrome and removal of adenomas decreases the risk for colorectal cancer (CRC). The Finnish surveillance protocol of colonoscopies at 2- to 3-year intervals facilitates patient adherence but includes an essential risk for CRC up to 60 years of age, but without CRC-related mortality when the surveillance instructions are followed.

L

ynch syndrome, previously known as hereditary nonpolyposis colorectal cancer, is an autosomal-dominant disorder caused by germline mutations in DNA mismatch repair (MMR) genes. The syndrome is characterized by high penetrance and early onset of colorectal tumors and has an association with a broad spectrum of extracolonic cancers,

including gynecologic, urothelial, small-bowel, and brain tumors.1 Mutation carriers have an up to 80% risk for the development of colorectal cancer in Amsterdam-positive families, and therefore endoscopic screening of colorectal tumors has been recommended. The screening should be started at the age of 20 –25 years and continued at 2-year intervals.2 It has been advocated that colorectal carcinomas in Lynch syndrome develop via the adenoma– carcinoma sequence. Regular colonoscopic screening has proved to diminish the incidence of colorectal carcinomas (CRCs) and this was thought to be a sequel to the removal of adenomas.3 However, even with regular colonoscopic surveillance, mutation carriers still have a substantial risk for the development of colorectal cancer. In a recently published Dutch study this risk was estimated to be 10%–15% at 10 years.4 The explanations offered for interval cancers are an accelerated adenoma– carcinoma sequence and/or the de novo development of a carcinoma without a preceding adenoma. Only a few reports have been published on colonoscopic surveillance in Lynch syndrome, and therefore the occurrence and significance of adenomas as a precursor of CRC has remained unclear.5 We previously published results on surveillance over 10 and 15 years in a limited number of families.3,6 These 2 studies showed the efficacy of surveillance in decreasing mortality and the rate of CRC. The present study assessed the incidence of colorectal tumors during long-standing surveillance of 117 Lynch syndrome families by analyzing the index colorectal manifestations observed during regular endoscopic screening in asymptomatic mutation carriers with a previously clean bowel. Special emphasis was placed on the cumulative age-related incidence of adenomas and adenocarcinomas in an attempt to clarify the significance of adenoma as a precursor of colorectal carcinoma.

Materials and Methods The Finnish Hereditary Colorectal Cancer Registry was established in 1982 in an attempt to organize nationwide screening and management for Lynch synAbbreviations used in this paper: CI, conﬁdence interval; CRC, colorectal cancer; MMR, mismatch repair. © 2007 by the AGA Institute 0016-5085/07/$32.00 doi:10.1053/j.gastro.2007.08.019

CLINICAL– ALIMENTARY TRACT

Development of Colorectal Tumors in Colonoscopic Surveillance in Lynch Syndrome

1094

MECKLIN ET AL

GASTROENTEROLOGY Vol. 133, No. 4


drome families in a country with a population of 5.5 million.7 The registry contains complete endoscopic and hospital follow-up data on family members. After discovery of the MMR genes in 1993 and 1994, molecular genetic diagnosis was organized rapidly in Finland because the 2 founding mutations together accounted for more than 60% of kindreds meeting the international diagnostic criteria.8,9 Since then, genetic counseling and predictive testing has been provided for at-risk family members. Regular colonoscopic screening has been recommended for mutation carriers. The implementation of the follow-up procedures was controlled and endoscopy and pathologic records were obtained immediately by the Registry to ensure the continuation of adequate surveillance. Thus, the study material has been collected prospectively. The screening interval was 5 years in the 1980s and 3 years between 1989 and 1994. As soon as the predictive testing enabled mutation-negative family members to be excluded and surveillance to be focused on true mutation carriers the interval then was changed to 2 years for individuals reaching the age of 35 years. By May 2005, the Finnish Hereditary Colorectal Cancer Registry had data on 117 families, in whom a mutated human MutL homolog 1 gene has been detected in 97 families, human MutS homolog 2 in 13 families, and human MutS homolog 6 in 7 families. Altogether 740 DNA-tested mutation carriers were registered in the database. The inclusion criteria for the present study was no diagnosis or removal of a colorectal tumor before regular colonoscopic screening organized by the Registry. Of the 740 mutation carriers, 320 were excluded from the study: 211 patients had been diagnosed with colorectal cancer before Lynch syndrome was ascertained in the family and systematic screening was organized, 83 had not started surveillance by May 2005, 11 patients died before surveillance, and 15 patients declined screening (7 lived permanently abroad). Thus, 420 mutation carriers (197 men and 223 women) were included in the analysis (median age, 36.0 y; range, 20 –74 years at the beginning of surveillance). Between March 1982 and May 2005 a total of 1252 colonoscopies (mean, 2.1 per person) were performed. The median follow-up time was 6.7 years (range, 1 mo to 23.5 y). The number of colonoscopies during the screening is presented in Table 1.

Table 1. Numbers of Colonoscopies During the Screening Period Number of colonoscopies

Number of patients (%) aTwenty-six

1

2

3

4

5

6a

421 (100)

315 (84)

222 (53)

128 (30)

70 (17)

42 (10)

individuals had more than 6 colonoscopies (range, 7–12).

Figure 1. Lexis-diagrams showing the individual prevalent (found at the first colonoscopy, marked by bullets) and incident cases (found during the follow-up period, marked by squares) of adenoma (left panel) and carcinoma (right panel) as well as the individual follow-up times (gray lines) from the first colonoscopy until an incident case is found, by calendar time and age.

The first neoplastic tumor (adenoma or colorectal carcinoma) observed in the colonoscopic examination after the first colonoscopy was considered as the index finding. If both an adenoma and a carcinoma were diagnosed metachronously during the surveillance period, the index lesion was the one diagnosed earlier. If several synchronous lesions were observed during the same screening procedure, the most advanced one (carcinoma) was taken as the index lesion. Although in some persons an adenoma was diagnosed as the index tumor before CRC, in calculating the risk for CRC with the surveillance protocol, all CRC cases observed during the follow-up period were included. The follow-up by calendar time and age is illustrated in Figure 1 for both adenomas and carcinomas, and includes both prevalent and incident cases.

Statistical Analysis The cumulative risk for (1) incident adenoma, (2) incident carcinoma, and (3) the combined risk for both tumors by age for both sexes was estimated using the Kaplan–Meier method. For all 3 outcomes the follow-up evaluation was started at the individual date of the first colonoscopy for each subject, at which point he/she be-

COLORECTAL ADENOMAS IN LYNCH SYNDROME

came a member of the relevant risk group, and remained there until the outcome diagnosis or death, or until April 30, 2005. Persons in whom an adenoma, carcinoma, or any tumor was found at the first colonoscopy were excluded from the analysis of incident adenomas, carcinomas, or both tumors. Approximate 95% confidence intervals (CIs) for the age-specific cumulative risks were computed according to the standard error of the estimated cumulative hazards. The relative hazards between men and women were estimated by the proportional hazards (or Cox) model. The computations were performed using tools of the survival in the R environment of statistical computing.10

Pathology The histologic classification of adenomas and staging of dysplasia were based on the World Health Organization classification of tumors.11

Results During the surveillance a total of 116 initial adenomas were removed in 100 (28%) subjects (56 men, 44 women; mean age, 48 y). In 16 cases more than one adenoma was encountered. In these cases it was impossible to decide which of the synchronously occurring adenomas was the most advanced. Therefore, the locations of all these adenomas were taken into account in the study (Table 2). In 39 subjects the first adenoma was found in the first colonoscopy. After excluding these cases, the risk for incident adenoma increased steadily over the age range starting from age 20 (Figure 2), the estimated risk by age 60 was 68.5% (95% CI, 50.5%– 80.0%; based on 32 cases) in men and 48.3% (95% CI, 29.1%– 62.2%; based on 29 cases) in women. The relative hazard for incident adenoma between men and women was 1.40 (95% CI, 0.84 –2.32). Tubular adenomas were the most common histologic type (82%), and the majority of adenomas (89%) showed only low-grade dysplasia (Table 2). The mean adenoma size was 7.4 mm (range, 1– 40 mm). Although located throughout the colon and rectum,

1095

nearly 50% of the adenomas were situated proximal to the splenic flexure (Table 2). During the surveillance period a total of 41 CRCs were observed in 40 patients (21 men, 19 women; mean age, 46 y). Fifteen CRCs (in 14 patients) were diagnosed at the first screening colonoscopy (mean age, 46.4 y; range, 27– 69 y) and the remaining 26 carcinomas were diagnosed during the follow-up period (mean age, 44.0 y; range, 29 –75 y). The estimated cumulative risk for developing an incident CRC from 20 years onward (Figure 2) until age 60 was 34.6% (95% CI, 16.0%– 49.0%; 14 cases) in men and 22.1% (95% CI, 7.2%–34.5%; 12 cases) in women. The relative hazard of incident CRC between men and women was 1.46 (95% CI, 0.67–3.19). Carcinomas were situated mainly in the proximal colon (Table 2) and 80% of them were only locally advanced Dukes A or B tumors. In 3 subjects, the patient-related delay between colonoscopies was considerably long, ranging from 5 to 9 years. Insufficient cooperation (eg, owing to alcoholism) was the main reason for the delay. In these 3 patients an advanced carcinoma was diagnosed, resulting in premature death. Four of the carcinomas diagnosed within the regular surveillance protocol were Dukes C carcinomas. The combined risk of an incident adenoma or carcinoma by 60 years of age (Figure 2) was estimated to be 81.8% (95% CI, 67.3%– 89.9%; 44 cases) in men and 62.9% (95% CI, 45.2%–74.9%; 41 cases) in women. The hazard ratio between men and women was 1.40 (95% CI, 0.91– 2.17).

Mortality During the surveillance 26 (6.2%) individuals died, 19 from cancer and 7 from other causes. Five persons died from CRC. In 4 cases this was owing to patientrelated delayed diagnosis. One patient postponed the first screening visit for several years because of childbirth and care of small twins and 3 persons did not follow the surveillance instructions and postponed colonoscopy until symptoms forced them to contact

Table 2. Characteristics and Localization of Adenomas and Carcinomas in 420 Mutation Carriers

Adenomas Histology Tubular Tubulovillous Dysplasia Low grade High grade Carcinomas Dukes A Dukes B Dukes C Dukes D

Cecum and ascending colon, n (%)

Transverse colon, n (%)

Descending and sigmoid colon, n (%)

Rectum, n (%)

Total, n (%)

43 (37)

15 (13)

31 (27)

27 (23)

116 (100)

36 7

12 3

25 6

22 5

95 (82) 21 (18)

39 4 21 (51) 16 1 3 1

12 3 10 (24) 7 1 0 2

25 2 2 (5) 2 0 0 0

103 (89) 13 (11) 41 (100) 29 (71) 4 (10) 5 (12) 3 (7)

27 4 8 (20) 4 2 2 0


October 2007

1096

MECKLIN ET AL



Figure 2. Kaplan–Meier curves showing the estimated cumulative risks of developing adenoma (A, B), carcinoma (C, D), and either tumor in male (E) and female (F) Lynch syndrome mutation carriers by age. The solid lines describe the point estimates, and the dashed lines show the 95% confidence limits of the risks at each age.

us. The fifth individual was a 73-year-old woman who had an early carcinoma in a polyp 9 mm in diameter. This was treated by endoscopic polypectomy. She had annual follow-up colonoscopies until age 81, at which point she reported abdominal symptoms and multiple liver metastases were diagnosed. The origin of the metastases remained unclear because no surgery or autopsy was performed. Extracolonic cancer was observed in 73 (18%) patients: 34 endometrial cancers (1 patient had concomitant ovarian cancer); 8 ovarian cancers, 5 urothelial cancers, 5 skin tumors, 3 gastric cancers, 3 brain tumors, 1 prostate cancer, 1 lung cancer, 1 small intestine cancer, 1 breast cancer, and 1 lymphoma were diagnosed on the basis of

symptoms. Ten of these patients with extracolonic cancer also had CRC. Fourteen persons died from extracolonic cancer (mean age, 62.2 y): pancreas or biliary tract, 5; stomach, 2; lung, 2; urinary tract, 1; prostate, 1; uterus, 1; brain, 1; and lymphoma, 1.

Discussion The results of surveillance in a prospective casecontrolled study of 22 hereditary nonpolyposis colorectal cancer families after follow-up evaluations of 10 and 15 years have been published previously. These studies indicated that CRC-related mortality can be avoided and a 60% decrease in the risk for developing CRC can be

obtained with colonoscopic screening at 3-year intervals.3,6 These studies remain the only controlled cohort studies that provide statistical evidence that surveillance colonoscopies reduce the risk for developing CRC and the risk for death.5 The Finnish surveillance protocol differs from the general recommendations. The 3-year interval is longer than the 2-year interval recommended by the International Collaborative Group on hereditary nonpolyposis colorectal cancer,2 and even more frequent follow-up evaluations have been proposed.1,12 No randomized controlled trials have been published regarding the optimal frequency of surveillance.5 The present study sought to elucidate the cumulative risk of colorectal adenomas and adenocarcinomas during regular follow-up evaluation so as to provide evidence of use in improving colonoscopic surveillance. There is a limited amount of information on the agerelated cumulative risk for adenomas in Lynch syndrome. In previous studies, the prevalence of adenomas has ranged from 17% to 49%. However, these studies were published mainly in the 1980s and early 1990s before the MMR genes were characterized and predictive testing became possible.13–18 These previously published prevalence rates therefore do not reflect the real cumulative risk of mutation carriers. In a more recent study by Lindgren et al19 the relative risk for Lynch syndrome mutation carriers to develop an adenoma was 4.5 times greater than that of the general population. However, it was not possible to estimate the cumulative age-related risk in this study. de Jong et al20 compared the occurrence of adenomas between mutation carriers and noncarriers in 86 Lynch syndrome families. The number of follow-up colonoscopies performed in the carrier group was 540 and a total of 83 adenomas were detected during the surveillance period. On the basis of their analysis, 29.7% of carriers were free of an adenoma at age 60, but no gender differences in the risk for developing adenoma were found. In our study, the cumulative risk of incident adenoma increased steadily over the relevant age range, reaching 68% by age 60 in men and 48% in women. CRC was observed in 40 of 420 persons in the present study. When the carcinomas observed during the first screening visit were excluded, the estimated cumulative risk presenting with CRC during surveillance increased gradually to 35% in men and 22% in women by age 60. The present results show the cumulative risk for the development of a colorectal adenoma and show a surveillance-related reduction in the occurrence of CRC. The combined age-related risk for adenoma and carcinoma (81.8% for men and 62.9% for women at age 60) is in accordance with earlier estimates based on Amsterdampositive families.1,21 In 3 recent studies the cumulative risk of CRC has been estimated using genotype restricted likelihood to avoid ascertainment bias caused by the selection of families

COLORECTAL ADENOMAS IN LYNCH SYNDROME

1097

according to the Amsterdam criteria and the obvious high penetrance of the genotype.22–24 This kind of calculation gives markedly lower estimates but with wide error margins. It can be assumed that the genotype restricted likelihood method would give more reliable and unbiased estimates in unselected and population-based series, especially if, as has been suggested, low penetrance of CRC truly occurs in mutations of some MMR genes, such as human MutS homolog 6 or PMS2. When families with a much lower risk of CRC are in follow-up programs, the instructions for surveillance should be reconsidered accordingly and adapted to the estimated low risk of cancer. In the study by Quehenberger et al22 the lifelong risk for CRC was between 12.6% and 51.0% for men and 10.6% and 43.8% for women, but the investigators did not recommend changes in screening practices. Our opinion is that these calculations based on the genotype restricted likelihood method underestimate the risk of adenoma and CRC in families meeting the Amsterdam criteria. The question of whether adenomas are the only type of preceding lesion required for cancer formation is of interest. The fact that the observed cumulative risk for the development of an adenoma in our study reaches the same high level as has been earlier estimated for cancer, together with the predominantly proximal location of adenomas and carcinomas, indicates that adenomas are an important precedent lesion for CRC. A similar conclusion was presented in the analysis by de Jong et al,20 who showed an absence of MMR proteins in most adenomas obtained for immunohistochemical staining of human MutL homolog 1, human MutS homolog 2, and PMS2 proteins, implying an association with MMR protein deficiency. To our knowledge, there is no ongoing randomized trial which could, within the next few years, provide an answer to the question of the optimal follow-up interval in Lynch syndrome. For the time being, follow-up evaluation should be based on a balance between the acceptable risk for cancer and patient compliance. The low risk of adenomas and carcinomas before age 30 may justify a longer interval in the early stages of lifelong surveillance. Also, complications related to the procedure play a role in the dilemma.5 In a recent trial, the number of patients lost to follow-up evaluation after endoscopic polypectomy was higher among those who were randomized for annual screening compared with those who were screened every 2 or 5 years.25 The risk of premature death in Lynch syndrome may not be so highly dependent on occurrence of CRCs found in regular surveillance (mostly of an early stage) but rather on the risk person’s adherence to the life-long follow-up. In many centers, adherence has been alarmingly low, even decreasing to less than 50%.26 –28 In Finland, the surveillance interval mainly has been 3 years. With the availability of predictive DNA testing, the interval has been reduced gradually to 2 years for individuals reaching the age of 35 years.


October 2007

1098

MECKLIN ET AL


With this protocol the adherence has been high. In our recent study of 664 mutation carriers who started the endoscopic surveillance program, the compliance rate was more than 98%, with only 8 (1.2%) persons interrupting the follow-up evaluation.29 In addition, all at-risk family members both in Finland and abroad have been actively contacted and the proportion of nonresponders has remained low.30 On the basis of our results, it can be concluded that adenoma is an important lesion preceding CRC in most cases of Lynch syndrome, and that the removal of adenomas decreases the risk for CRC. In patients younger than 30 years of age, the risk for the development of colorectal tumors is low and this may justify a longer screening interval. The Finnish surveillance protocol of colonoscopies at 2to 3-year intervals facilitates patient adherence but includes essential risk for CRC up to 60 years of age, although fortunately without CRC-related mortality when the instructions are followed. In the future we intend to opt for a 2-year interval after 30 years of age with frequent reminders to those failing to comply. References 1. Lynch HT, Riley BD, Weismann S, et al. Hereditary non-polyposis colorectal carcinoma (HNPCC) and HNPCC-like families; problems in diagnosis, surveillance and management. Cancer 2004;100: 53– 64. 2. Vasen HF, Mecklin JP, Khan PM, et al. The International Collaborative Group on Hereditary Nonpolyposis Colorectal Cancer (ICGHNPCC). Dis Colon Rectum 1991;34:424 – 425. 3. Järvinen HJ, Aarnio M, Mustonen H, et al. Controlled 15 year trial on screening for colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 2000;118:828 – 834. 4. Vasen HF, Nagengast FM, Khan PM. Interval cancer in hereditary non-polyposis colorectal cancer (Lynch syndrome). Lancet 1995; 345:1183–1184. 5. Johnson PM, Gallinger S, McLeod RS. Surveillance colonoscopy in individuals at risk for hereditary nonpolyposis colorectal cancer: an evidence-based review. Dis Colon Rectum 2005;49:80 –95. 6. Järvinen HJ, Mecklin J, Sistonen P. Screening reduces colorectal cancer rate in families with hereditary non-polyposis colorectal cancer. Gastroenterology 1995;108:1405–1411. 7. Mecklin J-P, Järvinen HJ, Aukee S, et al. Screening for colorectal carcinoma in cancer family syndrome kindreds. Scand J Gastroenterol 1987;22:449 – 453. 8. Nyström-Lahti M, Sistonen P, Mecklin JP, et al. Close linkage to chromosome 3p and conservation of ancestral founding haplotype in hereditary nonpolyposis colorectal cancer families. Proc Natl Acad Sci U S A 1994;21:6054 – 6058. 9. Nyström-Lahti M, Kristo P, Nicolaides NC, et al. Founding mutations and Alu-mediated recombination in hereditary colon cancer. Nat Med 1995;11:1203–1206. 10. R Development Core Team. R: A language and environment for statistical computing. R Foundation for statistical computing, Vienna, Austria. Available at: http://www.r-project.org/. Accessed August 23, 2007. 11. Hamilton SR, Aaltonen LA. Word Health Organization classification of tumours. Pathology and genetics of tumours of the digestive system. Lyon: IARC Press, 2000. 12. Möslein G, Pistorius S, Saeger H-D, et al. Preventive surgery for colon cancer in familial adenomatous polyposis and hereditary nonpolyposis colorectal cancer syndrome. Langenbecks Arch Surg 2003;388:9 –16.


13. Love RR, Morrissey JF. Colonoscopy in asymptomatic individuals with a family history of colorectal cancer. Arch Intern Med 1984; 144:2209 –2211. 14. Houlston RS, Murday V, Harocopos C, et al. Screening and genetic counselling for relatives of patients with colorectal cancer in a family cancer clinic. BMJ 1990;301:366 –368. 15. Lanspa SJ, Lynch HT, Smyrk TC, et al. Colorectal adenomas in the Lynch syndromes. Results of a colonoscopy screening program. Gastroenterology 1990;98:1117–1122. 16. Hodgson SV, Harocopos C, Gaglia P. Screening results in a family cancer clinic: five years experience. Anticancer Res 1993;13: 2581–2585. 17. Jass JR, Stewart SM, Stewart J, et al. Hereditary non-polyposis colorectal cancer morphologies genes and mutations. Mutat Res 1994;310:125–133. 18. Vasen HF, Taal BG, Nagengast FM, et al. Hereditary nonpolyposis colorectal cancer: results of long-term surveillance in 50 families. Eur J Cancer 1995;31A:1145–1148. 19. Lindgren G, Liljegren A, Jaramillo E, et al. Adenoma prevalence and cancer risk in familial non-polyposis colorectal cancer. Gut 2002;50:228 –234. 20. de Jong AE, Morreau H, Nagengast FM, et al. Prevalence of adenomas among young individuals at average risk for colorectal cancer. Am J Gastroenterol 2005;100:139 –143. 21. Aarnio M, Sankila R, Pukkala E, et al. Cancer risk in mutation carriers of DNA-mismatch-repair genes. Int J Cancer 1999;81: 214 –218. 22. Quehenberger F, Vasen HFA, van Houwelingen HC. Risk of colorectal and endometrial cancer for carriers of mutations of the hMLH1 and hMSH2 gene: correction for ascertainment. J Med Genet 2005;42:491– 496. 23. Jenkins MA, Baglietto L, Dowty JG, et al. Cancer risks for mismatch repair gene mutation carriers: a population-based early onset case-family study. Clin Gastroenterol Hepatol 2006;4: 489 – 498. 24. Alarcon F, Lasset C, Carayol J, et al. Estimating cancer risk in HNPCC by the GRL method. Eur J Hum Genet 2007;8:1– 6. 25. Lund JN, Scholefield JH, Grainge MJ, et al. Risks, costs and compliance limit colorectal adenoma surveillance: lessons from a randomized trial. Gut 2001;49:91–96. 26. Lerman C, Hughes C, Trock B, et al. Genetic testing in families with hereditary nonpolyposis colorectal cancer. JAMA 1999;281: 1618 –1622. 27. Ponz de Leon M, Benatti P, Di Gregorio, et al. Genetic testing among high-risk individuals in families with hereditary non-polyposis colorectal cancer. Br J Cancer 2004;23:882– 887. 28. Church JM, Kay S, Shenae J. The benefits of counselling in families likely to have HNPCC and why families do not get it. Dis Colon Rectum 2005;48:616. 29. Pylvänäinen K, Kairaluoma M, Mecklin J-P. Compliance and satisfaction with long-term surveillance in Finnish HNPCC families. Familial Cancer 2006;5:175–178. 30. Aktan-Collan K, Mecklin J-P, Järvinen HJ, et al. Predictive genetic testing for hereditary non-polyposis colorectal cancer: uptake and long-term satisfaction. Int J Cancer 2000;89:44 –50.

Received August 8, 2006. Accepted July 12, 2007. Address requests for reprints to: Jukka-Pekka Mecklin, MD, Department of Surgery, Jyväskylä Central Hospital, Keskussairaalantie 19, FIN-40620 Jyväskylä, Finland. e-mail: jukka-pekka.mecklin@ksshp.ﬁ; fax: (358) 14-2693626. Supported by the Jyväskylä Central Hospital, the Sigrid Juselius Foundation, and the Finnish Cancer Foundation. The authors thank Tuula Lehtinen for the data collection and all the individuals and families who took part in this study.