The cause of ulcerative colitis (UC) and Crohn s disease

GASTROENTEROLOGY 2003;124:1767 1773 Inflammatory Bowel Disease in a Swedish Twin Cohort: A Long- Term Follow-up of Concordance and Clinical Characteristics JONAS HALFVARSON,* LENNART BODIN, CURT TYSK,* EVA LINDBERG, and GUNNAR JÄRNEROT* *Division of Gastroenterology, Department of Medicine, Clinical Research Center, and Department of Pediatrics, Örebro University Hospital, Örebro, Sweden Background & Aims: In 1988, we reported the first twin study in inflammatory bowel disease. The aim of the current study was to follow up these twins regarding new cases of inflammatory bowel disease and Crohn s disease characteristics using the Vienna classification. Methods: The official Swedish population register and the cause of death register were used to search for the twins. All living patients were interviewed. Results: Three monozygotic twins earlier classified as healthy had been diagnosed with inflammatory bowel disease (ulcerative colitis, n 2; Crohn s disease, n 1). Retrospectively, all 3 were symptomatic at the original survey. This changed the pair in monozygotic twins from 6.3% to 18.8% in ulcerative colitis and from 44.4% to 50.0% in Crohn s disease. A high degree of regarding age at, disease location at and during the course, and disease behavior was found in concordant monozygotic twin pairs with Crohn s disease. Seven of 9 pairs were identical in 3 or more of these disease characteristics compared with an expected number of 1.5 (P 0.000076). Conclusions: This study confirms that the genetic influence is stronger in Crohn s disease than in ulcerative colitis. A remarkable phenotype similarity within concordant pairs with Crohn s disease was found using the Vienna classification. The cause of ulcerative colitis (UC) and Crohn s disease (CD) is still unknown. Genetic factors are of importance in both diseases, although more in CD than UC. In 1988, we published a study of unselected monozygotic or dizygotic twins with UC or CD that showed a significantly different heritability in the 2 diseases with much stronger genetic influence in CD than UC. 1 This finding has been confirmed in another unbiased twin study from Denmark 2 and supported by a British twin study. 3,4 The aim of the present study was to follow up our previous twin material to investigate whether further twins had been affected by inflammatory bowel disease (IBD), which could affect the reported rate. Furthermore, we wanted to compare the clinical phenotype according to the new Vienna classification within concordant monozygotic twin pairs with CD. 5 Materials and Methods The twins in the study were born between 1886 and 1958 and found by running the Swedish twin registry against the Swedish hospital discharge register; only twins of the same sex within each pair were included. The twin material has previously been presented in detail. 1 In short, we found 80 twin pairs with IBD for which at least one twin in each pair was affected. In 16 monozygotic twin pairs with UC, only one pair was concordant for the disease compared with 8 of 18 monozygotic pairs with CD. Among 17 dizygotic pairs and 3 pairs with unknown zygosity with UC, no for the disease occurred. In 26 dizygotic pairs with CD, only one pair was concordant. Using the unique 10-digit registration number for each Swedish inhabitant, the official Swedish population register was used to search for the twins during 1999 and 2000; the register also contains each inhabitant s present address. Twins who no longer could be found in this register had either died or emigrated and were searched for in the cause of death register. If not found there, the last known hospital clinic was approached for further information. All living twins were sent a questionnaire regarding a possible of IBD and more general gastrointestinal symptomatology. If a twin did not answer the questionnaire after a reminder letter, attempts were made to contact the twin for a telephone interview. When a twin who had been unaffected at the earlier survey was suspected to have acquired IBD, the medical notes from the local hospital were requested to verify or refute the. Also, the medical notes from all twins with earlier known IBD were scrutinized up to the time of the present survey. This was done to confirm that the earlier of IBD had not been changed. In twins with CD, the aim was also to state the age at, location of disease at onset and during the course of illness, and CD behavior 2003 by the American Gastroenterological Association 0016-5085/03/$30.00 doi:10.1016/s0016-5085(03)00385-8

1768 HALFVARSON ET AL. GASTROENTEROLOGY Vol. 124, No. 7 Table 1. Clinical Characteristics of the Reference Population of Twins with CD According to the Vienna Classification No. of twins Location a Terminal ileum 15 Colon 8 Ileocolon 11 Upper GI 2 Behavior Nonstricturing Nonpenetrating 18 Stricturing 10 Penetrating 8 Progression b Yes 9 No 27 Age at Younger than 20 2 20 24 16 25 29 6 30 34 5 35 39 1 40 44 4 45 49 1 50 54 0 55 59 1 Older than 60 0 NOTE. n 36. a Location of inflammation at. b Progression in extent of disease. according to the Vienna classification. The time of was used instead of onset of symptoms, because the latter is difficult to exactly establish retrospectively on the basis of medical notes. Extraintestinal manifestations were also recorded. Data on smoking were known from the earlier survey, 1 and the accuracy of these data could be evaluated by the present questionnaire. In twins with UC, age at, disease location, surgery, and extraintestinal manifestations were recorded in a similar manner. and proband, respectively, were calculated according to a previously reported formula. 1 Clinical in concordant monozygotic pairs with CD was evaluated regarding 4 characteristics: (1) age at ( 5 years difference between twins was considered as from a statistical point of view), (2) location of disease at, (3) progression in extent of disease, and (4) disease behavior. We tested the hypothesis that the outcome in these 4 characteristics is independent within the individuals in each twin pair. This hypothesis was analyzed in 2 ways. In the first, each characteristic was evaluated individually (a univariate perspective); in the second, all 4 characteristics were considered simultaneously (a multivariate perspective). The twins with CD in the discordant monozygotic and the dizygotic pairs constituted the reference population (Table 1). The 4 characteristics and their distributions in the reference population are shown in Table 1. Based on these figures, the estimated probability that 2 individuals in a twin pair are concordant can be calculated, given that the distribution of these clinical characteristics is a random assortment (i.e., according to the hypothesis of independence). For instance, the probability for in location is 15/36 15/36 8/36 8/36 11/36 11/36 2/36 2/36 or 0.319. The estimated probabilities for all 4 characteristics were calculated in a similar manner. As a second step, using the binomial distribution, the probability for the observed degree of clinical among the monozygotic twin pairs for which both twins have CD was calculated for each characteristic separately based on the estimated probabilities. This is in fact equivalent to using a one-sided sign test. The estimated probabilities were also tested in a multivariate perspective. A twin pair can be concordant in 0 up to 4 of the 4 clinical characteristics. The estimated probability for each degree of was calculated using the estimated probability for each clinical characteristic. Based on these estimated probabilities for each degree of, the probability for the observed overall in the monozygotic twin pairs for which both twins have IBD was calculated in a third step using the multinomial distribution. In a fourth step, the multivariate approach based on the multinomial distribution was somewhat simplified using the binomial distribution. Using this, the estimated probability for 4 or 3 clinical s within pairs was compared with the estimated probability for 3 clinical s. From this, the P value for the observed number (or more) of twin pairs with in at least 3 of the 4 clinical characteristics among the total number of monozygotic twin pairs for which both twins have CD was calculated. The Örebro County Ethical Committee approved the study. Results At follow-up, 136 twins were still alive and 24 had died. Causes of death and data regarding whether or not they had acquired IBD were obtained in all 24 dead twins. Of the 136 living twins, 125 responded to the questionnaire either by mail (n 100) or telephone (n 25). Three twins were lost to follow-up, and 8 twins declined participation. However, for one of these twins, his twin brother informed us about his IBD status (not having the disease). Thus, the follow-up was successful in 126 of 136 living twins (92.6%), in 100% of the dead twins, and in 93.8% of the whole cohort. Follow-up data could not be recorded for 10 living twins. Five of the twins were previously known to have IBD, and the remaining 5 twins (who did not have IBD at the first survey) were searched for in the Swedish hospital discharge register at the National Board of Health and Welfare. None of them were found there and therefore considered as still free from IBD. The of IBD had not been changed in any of the twins.

June 2003 INFLAMMATORY BOWEL DISEASE IN TWINS 1769 Table 2. and Concordance Rates in the Swedish Cohort of Twins in 1985 and 2000, the Danish Cohort of Twins, and the British Cohort of Twins With IBD UC CD Cohort Monozygotic twins Dizygotic twins Monozygotic twins Dizygotic twins Swedish, 1985 (n 80 twin pairs) 6.3 6.3 0 0 44.4 58.3 3.8 3.8 Swedish, 2000 18.8 18.8 0 0 50.0 62.5 3.8 3.8 Danish (n 103 twin pairs) 14.3 18.2 4.5 4.5 50.0 58.3 0 0 British (n 249 twin pairs) 13.4 Not published 4.4 Not published 33.3 Not published 9.9 Not published Concordance UC. Two new twins with UC were identified within the monozygotic twin cohort. In one, UC was suspected at the first survey; however, the evidence for a definite was considered insufficient at that time. Since then, UC complicated by primary sclerosing cholangitis has been established. The first suspicion of UC arose at 36 years of age, whereas his monozygotic twin brother developed symptoms at 33 years of age and UC was diagnosed the same year. Among the dead twins, one earlier unknown case of UC was found when reading the medical notes to establish the cause of death. A monozygotic woman had had definite UC not found in the first survey, because at that time the Swedish hospital discharge register did not cover the area of Sweden in which she lived. Furthermore, she denied any gastrointestinal disease when directly questioned at the first survey. UC was diagnosed at the age of 68 years and in her twin sister at the age of 66 years. In dizygotic twins, no concordant pair with UC was found in the original survey; at follow-up, this lack of remained unchanged. In the 3 pairs with unknown zygosity, no new cases of IBD had occurred. These findings changed the rates in the monozygotic but not in the dizygotic twins or in the twins with unknown zygosity compared with the first survey. Both the pair and proband in the monozygotic twins with UC increased from 1 of 16 (6.3%) to 3 of 16 (18.8%). The rates in the dizygotic twins and in those with unknown zygosity remained nil (Table 2). The mean observation time in the healthy twin in the group of discordant monozygotic twins with UC was 32.2 years and the median was 34 years (range, 7 53 years) after the of UC in the twin sibling. In the total group (monozygotic dizygotic unknown zygosity) of discordant twins with UC, the mean observation time in the healthy twin was 32.3 years and the median was 30 years (range, 7 53 years) after the of UC in the twin sibling. CD. At follow-up, only one new case of CD was found. An earlier healthy monozygotic twin brother to a twin with established CD was also found to have the disease. He was suspected of having CD at the first survey, but the evidence for the was considered insufficient. The suspicion of CD arose when he was 27 years of age but did not become definite until 52 years of age. His twin brother was diagnosed at 26 years of age. No new cases of CD were found in dizygotic twins. This changed the pair in monozygotic twins with CD from 8 of 18 (44.4%) to 9 of 18 (50.0%) and the proband from 58.3% to 62.5%. In dizygotic twins, both the pair (3.8%) and the proband (3.8%) remained unchanged (Table 2). The mean observation time in the healthy twin in the group of discordant monozygotic twins with CD was 27.4 years and the median was 27 years (range, 23 33 years) after the of CD in the twin sibling. In the total group (monozygotic dizygotic) of discordant twins with CD, the mean observation time in the healthy twin was 27.8 years and the median was 27 years (range, 3 42 years) after the of CD in the twin sibling. Clinical Comparison in Concordant Monozygotic Twins Three monozygotic twin pairs were concordant for UC (Table 3). The location of the disease at was similar in 2 pairs but differed slightly in the third. In 2 of the 3 pairs, for age at was observed. One twin had had a proctocolectomy and an ileostomy, but the remaining 5 twins had not needed an

1770 HALFVARSON ET AL. GASTROENTEROLOGY Vol. 124, No. 7 Table 3. Clinical Comparison in Concordant Monozygotic Twins With UC no. birth Onset of symptoms Location of inflammation a Surgery Smoking at time of Associated diseases 1 1911 1977 1977 Left colon 3 Ileostomy Never smoked Deep venous thrombosis? 1979 Left colon None Never smoked None 2 1937 1973 1973 Left colon 3 None Never smoked Primary sclerosing cholangitis, arthritis, osteoporosis 1970 1970 Extensive None Never smoked Abnormal liver function tests, arthritis 3 1938 1965 1965 Left colon 3 None Never smoked None 1973 1974 Left colon 3 None Former smoker None a Location of inflammation at onset 3 maximum location of inflammation during follow-up, if increased. operation. No definite conclusions can be drawn regarding associated diseases. Nine pairs of monozygotic twins were concordant for CD. Details are given in Table 4. In 6 pairs, the was made 2 years or less apart in each twin, but in 3 pairs the time interval was considerably longer (up to 26 years). The onset of symptoms showed similar time aspects. The location of disease at according to Table 4. Clinical Comparison in Concordant Monozygotic Twins With CD no. birth Start of symptoms Location a (Vienna) 1 1920 1970 1977 Terminal ileum 3 1969 1977 Terminal ileum 3 Progression of disease extent Behavior (Vienna) Smoking at time of Associated diseases Yes Penetrating Ankylosing spondylitis Yes Penetrating Former Ankylosing spondylitis, primary sclerosing cholangitis, iritis, arthritis Yes Stricturing Kidney stones, coagulopathy 2 1929 1963 1964 Terminal ileum 3 ileocolon 1974 1974 Terminal ileum 3 Yes Stricturing Perianal fistulae ileocolon 3 1940 1960 1966 Terminal ileum No Stricturing Arthritis, abnormal liver function tests end of 1992 Ileocolon No Penetrating Arthritis, abnormal liver 1960s function tests 4 1946 1964 1967 Terminal ileum 3 Yes Penetrating None 1968 1969 Terminal ileum 3 Yes Stricturing Arthritis, Hodgkin s lymphoma 5 1947 1965 1973 Terminal ileum No Stricturing None 1974 1974 Terminal ileum No Stricturing None 6 1948 1973 1974 Ileocolon No Nonstricturing, Perianal fistulae, kidney stones 1974 1975 Colon No Nonstricturing, Arthritis, eczema, lichen ruber 7 1949 1966 1966 Ileocolon No Penetrating Perianal fistulae 1966 1966 Ileocolon 3 Yes Penetrating Perianal fistulae, arthritis 8 1953 1971 1976 Terminal ileum No Nonstricturing, None 1972 1976 Terminal ileum No Nonstricturing, None 9 1954 1965 1968 Ileocolon No Nonstricturing, Erythema nodosum 1975 1982 Ileocolon No Penetrating Former None, smoker;, never smoked, former, former smoker. a Location of inflammation at onset 3 maximum location of inflammation during follow-up, if increased.

June 2003 INFLAMMATORY BOWEL DISEASE IN TWINS 1771 Table 5. Estimated Probability That 2 Individuals in a Twin Are Concordant/Discordant Age at Location Behavior Progression Estimated probability for 0.262 0.319 0.377 0.625 Estimated probability for discordance 0.738 0.681 0.623 0.375 the Vienna classification was identical in 7 of the 9 pairs but differed slightly in 2. During the observation period, the extent of the disease in each twin could change or remain within the same part of the gastrointestinal tract. In 8 pairs, the twins showed the same pattern in this aspect. The twins whose extent of disease increased also showed a remarkable similarity in disease progress. The Vienna classification divides disease behavior into stricturing, penetrating, or nonstricturing. Six of the 9 pairs were concordant for disease behavior. Four twin pairs were concordant for either penetrating or stricturing disease, and 2 pairs were concordant for nonstricturing disease. In 2 other pairs, one twin had stricturing and the other had penetrating disease. It may be questioned whether penetrating disease occurs without a certain degree of stricturing influence. There seemed to be a definite discordance in disease behavior in only one pair (no. 9). In 7 pairs, the smoking pattern at the time of was identical within each pair. The occurrence of or lack of associated diseases was similar in 5 of the 9 pairs (Table 4). With this background, the clinical similarities within concordant monozygotic twins with CD were evaluated using the statistical methods described. In the first step, the expected probabilities for in each clinical characteristic were estimated (Table 5). As a second step, the probability for the observed degree of in Table 4 was calculated for each characteristic separately based on the probabilities in Table 5. Six concordant pairs were observed for age (P 0.013), 7 concordant pairs for location (P 0.00643), 8 concordant pairs for progression (P 0.093), and 6 concordant pairs for behavior (P 0.0764). The estimated probabilities in Table 5 were also tested in a multivariate perspective. A twin pair can be concordant in 0 up to 4 of the 4 clinical characteristics. The estimated probability for in all 4 characteristics is 0.377 0.319 0.262 0.625 0.0196931. Concordance in 3 of 4 characteristics can occur according to 4 different patterns (CCCD, CCDC, CDCC, and DCCC). The estimated probability for each alternative can be calculated in the same way. The sum of the estimated probability for the 4 different patterns gives the estimated probability for in 3 of the 4 characteristics (in this example, 0.1418710). The same procedure can be applied for 2, one, and no similarities. The probabilities for, the expected and observed number of twin pairs with in the different number of characteristics, are shown in Table 6. In the third step, the probability for the observed overall in the 9 monozygotic twin pairs for which both twins had IBD was calculated based on the results in Table 6. We found 3 pairs with in all 4 characteristics, 4 pairs with in 3 characteristics, one pair with in 2 characteristics, one pair with in only one characteristic, and no pairs with no at all. The probability was calculated as follows: 9!/(3! 4! 1! 1! 0!) 0.0196931 3 0.1418710 4 0.3575936 1 0.3634276 1 0.1174148 0, which gives a very low probability of 1.0 10 6. This gives a basis for rejecting the hypothesis of independence for the outcome in the 4 evaluated characteristics in the concordant twin pairs. In the fourth step, the estimated probability for 4 or 3 s within pairs was compared with the probability for 3 s. This probability is 0.0196931 0.1418710 0.16156409 (Table 6) and for 3 s is 1 0.16156409 0.83843591. From this, the P value for observing 7 pairs (or more) with in at least 3 of the 4 clinical characteristics among the total number of 9 monozygotic twin pairs for which both twins had CD was calculated. This was found to be 7.6 10 5 (P 0.000076), and the hypothesis of independence was thus rejected. Table 6. Probabilities for Concordance and Expected and Observed Number of Twin s With Concordance in Different Number(s) of Clinical Characteristics Concordance in clinical characteristics Probability Expected no. a Observed no. b 0 0.1174148 1.1 0 1 0.3634276 3.3 1 2 0.3575936 3.2 1 3 0.1418710 1.3 4 4 0.0196931 0.2 3 a Expected number of twin pairs among the 9 monozygotic twin pairs in which both twins have CD. b Observed number of twin pairs among the 9 monozygotic twin pairs in which both twins have CD.

1772 HALFVARSON ET AL. GASTROENTEROLOGY Vol. 124, No. 7 Discussion The cause of UC and CD is unknown. These are multifactorial disorders caused by interplay of genetic and environmental factors. Complex segregation analyses have previously shown that the best model for inheritance of UC is a major dominant or additive gene with low penetrance; in CD, a major recessive gene with high penetrance is suggested. 6 9 Recently, 2 independent studies showed the importance of mutations in the CARD15/NOD2 gene on chromosome 16 for the development of CD. 10,11 However, these mutations only explain a minor part of the total CD population. The twin study reported by us in 1988 was the first unbiased twin report. 1 A second study came from Denmark in 2000. 2 Both studies showed a significantly stronger genetic influence in CD than in UC. A third study from Britain indicated the same finding. 3,4 However, the twins in the British study were found after inquiries in patient associations about twins with occurrence of IBD and in that sense was not strictly unbiased. After the present follow-up of our twin cohort, the rates in the Danish and Swedish twin groups are almost identical (Table 2). The new UC and CD cases found in the follow-up already existed at the first survey, although unidentified. Thus, no really new case of IBD has occurred during the 15 years between the 2 studies. Furthermore, genetics seem to influence the clinical characteristics of CD, at least in concordant identical twins. Despite the limited number of concordant monozygotic twins with CD in our study (n 9), we found that there was a phenotype similarity within twin pairs considering the age at (P 0.013), and location of the disease (P 0.00643). There was also a nonsignificant trend for similarity in disease behavior (P 0.0746) and progression in extent of disease (P 0.093). The probability for the observed overall in the 9 twin pairs was 1.0 10 6, which strongly speaks against an association by chance. The reference population used in our calculations consisted of only 36 twins with CD, but we do not know with certainty if twins with CD are different from patients with CD in the general population. Therefore, we performed sensitivity analyses with different assumptions concerning the estimated probabilities in Table 5. Because we do not know the true probabilities in the reference population, computer simulations were performed for which the probability for behavior varied between 0.33 and 0.60, for location between 0.28 and 0.55, for age between 0.23 and 0.50, and for progression between 0.53 and 0.80. In all of these analyses, we received probabilities for the overall that were small, generally of the same size as those previously listed. In the most extreme case, we assumed all estimated probabilities for clinical to be 50% higher than those given in Table 5 but still obtained a probability for overall 0.05. This supports that the results would not be radically different when other clinically reasonable assumptions on the probabilities in the reference population were introduced. The remarkable similarity is also supported by the fact that in at least 3 of the 4 evaluated characteristics could be expected in 1.5 of the 9 pairs, but the observed number was 7 (P 0.000076). The limited number of monozygotic concordant twins with UC makes phenotype comparison difficult. However, the extent of disease in the colon was more or less identical in 2 pairs and similar in the third. Both twins in one pair had arthritis and one of them had primary sclerosing cholangitis, whereas the other had abnormal liver function tests but was not investigated for primary sclerosing cholangitis. In the Danish twin study, 3 of 5 concordant monozygotic twin pairs with CD were diagnosed within 3 years of each other. The location at was identical in 2 of the pairs. However, location was not defined in exactly the same way as in the present study and behavior was not described, so definite comparisons cannot be made. 2 The British study gave no such clinical information, so comparison is impossible. 3,4 Similar data were found in family studies. Bayless et al. 12 found a greater than expected for location and clinical type in a large sample of families with CD. This group also found that early age at was associated with more complicated disease. 13 Peeters et al. 14 confirmed a significant for location but not for behavior. The findings by Colombel et al. 15 supported that the median age at onset was younger in familial CD than in sporadic cases and a fair similarity regarding disease location, which increased in families with more than 2 affected members. Satsangi et al. 16 noted good regarding location and extraintestinal manifestations in familial CD. On the other side, in a large study of families with 3 or more first-degree relatives, Lee and Lennard-Jones 17 could only support that onset of disease in familial CD occurred at a younger age in the second generation, whereas site, extent, and clinical type of CD did not differ from the expected. Genetics alone are obviously not sufficient to cause CD, because the pair rate in monozygotic pairs is only about 50%. Environmental factors are re-

June 2003 INFLAMMATORY BOWEL DISEASE IN TWINS 1773 quired, and we are presently studying those more thoroughly in the twin population. In summary, the present long-term follow-up study confirms previous twin data that the genetic influence is stronger in CD than in UC. Using the Vienna classification, a remarkable similarity was seen in concordant twin pairs with CD regarding age at, location at and during the course of disease, and disease behavior. References 1. Tysk C, Lindberg E, Järnerot G, Floderus-Myrhed B. Ulcerative colitis and Crohn s disease in an unselected population of monozygotic and dizygotic twins. A study of heritability and the influence of smoking. Gut 1988;29:990 996. 2. Orholm M, Binder V, Sorensen TI, Rasmussen LP, Kyvik KO. Concordance of inflammatory bowel disease among Danish twins. Results of a nationwide study. Scand J Gastroenterol 2000;35:1075 1081. 3. Subhani J, Montgomery SM, Pounder RE, Wakefield AJ. Concordance rates of twins and siblings in inflammatory bowel disease (IBD). Gut 1998;42(Suppl 1):A40. 4. Thompson NP, Driscoll R, Pounder RE, Wakefield AJ. Genetics versus environment in inflammatory bowel disease: results of a British twin study. BMJ 1996;312:95 96. 5. Gasche C, Scholmerich J, Brynskov J, D Haens G, Hanauer SB, Irvine EJ, Jewell DP, Rachmilewitz D, Sachar DB, Sandborn WJ, Sutherland LR. A simple classification of Crohn s disease: report of the Working Party for the World Congresses of Gastroenterology, Vienna 1998. Inflamm Bowel Dis 2000;6:8 15. 6. Kuster W, Pascoe L, Purrmann J, Funk S, Majewski F. The genetics of Crohn disease: complex segregation analysis of a family study with 265 patients with Crohn disease and 5,387 relatives. Am J Med Genet 1989;32:105 108. 7. Monsen U, Iselius L, Johansson C, Hellers G. Evidence for a major additive gene in ulcerative colitis. Clin Genet 1989;36: 411 414. 8. Monsen U. Inflammatory bowel disease. An epidemiological and genetic study. Acta Chir Scand Suppl 1990;(Suppl 559):1 42. 9. Orholm M, Iselius L, Sorensen TI, Munkholm P, Langholz E, Binder V. Investigation of inheritance of chronic inflammatory bowel diseases by complex segregation analysis. BMJ 1993; 306:20 24. 10. Hugot JP, Chamaillard M, Zouali H, Lesage S, Cezard JP, Belaiche J, Almer S, Tysk C, O Morain CA, Gassull M, Binder V, Finkel Y, Cortot A, Modigliani R, Laurent-Puig P, Gower-Rousseau C, Macry J, Colombel JF, Sahbatou M, Thomas G. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn s disease. Nature 2001;411:599 603. 11. Ogura Y, Bonen DK, Inohara N, Nicolae DL, Chen FF, Ramos R, Britton H, Moran T, Karaliuskas R, Duerr RH, Achkar JP, Brant SR, Bayless TM, Kirschner BS, Hanauer SB, Nunez G, Cho JH. A frameshift mutation in NOD2 associated with susceptibility to Crohn s disease. Nature 2001;411:603 606. 12. Bayless TM, Tokayer AZ, Polito JM II, Quaskey SA, Mellits ED, Harris ML. Crohn s disease: for site and clinical type in affected family members potential hereditary influences. Gastroenterology 1996;111:573 579. 13. Polito JM II, Childs B, Mellits ED, Tokayer AZ, Harris ML, Bayless TM. Crohn s disease: influence of age at on site and clinical type of disease. Gastroenterology 1996;111:580 586. 14. Peeters M, Nevens H, Baert F, Hiele M, de Meyer AM, Vlietinck R, Rutgeerts P. Familial aggregation in Crohn s disease: increased age-adjusted risk and in clinical characteristics. Gastroenterology 1996;111:597 603. 15. Colombel JF, Grandbastien B, Gower-Rousseau C, Plegat S, Evrard JP, Dupas JL, Gendre JP, Modigliani R, Belaiche J, Hostein J, Hugot JP, van Kruiningen H, Cortot A. Clinical characteristics of Crohn s disease in 72 families. Gastroenterology 1996;111: 604 607. 16. Satsangi J, Grootscholten C, Holt H, Jewell DP. Clinical patterns of familial inflammatory bowel disease. Gut 1996;38:738 741. 17. Lee JC, Lennard-Jones JE. Inflammatory bowel disease in 67 families each with three or more affected first-degree relatives. Gastroenterology 1996;111:587 596. Received January 10, 2002. Accepted March 6, 2003 Address requests for reprints to: Jonas Halfvarson, M.D., Division of Gastroenterology, Department of Medicine, Örebro University Hospital, S-70185 Örebro, Sweden. e-mail: jonas.halfvarson@orebroll.se; fax: (46) 19-6021774. Supported by grants from Bengt Ihre s Foundation, Örebro University Hospital Research Foundation, Örebro County Research Foundation, and the International Organization for Study of Inflammatory Bowel Diseases.