Mutations in the NOD2/CARD15 gene in Crohn's disease are associated with ileocecal resection and are a risk factor for reoperation


Dr C. Büning, Department of Gastroenterology, Hepatology & Endocrinology, Charité, Campus Mitte, Schumannstr. 20/21, 10117 Berlin, Germany.


Background : Mutations within the NOD2/CARD15 gene have recently been shown to be associated with Crohn's disease.

Aims : To investigate the clinical impact of the three common NOD2/CARD15 mutations in patients with Crohn's disease.

Methods : We investigated the prevalence of the three common NOD2/CARD15 mutations (Arg702Trp, Gly908Arg, 3020insC) in 180 patients with Crohn's disease, 70 patients with ulcerative colitis and 97 controls. In patients with Crohn's disease, prevalence of NOD2/CARD15 mutations were correlated to clinical and demographical parameters.

Results : In Crohn's disease patients, 35.6% carried at least one mutant allele of NOD2/CARD15 mutations compared with 14.3% of patients with ulcerative colitis (P = 0.006) and to 15.5% of controls (P = 0.0001). Genotype phenotype analyses revealed that NOD2/CARD15 mutations determined younger age at disease diagnosis (P = 0.03), ileal disease location (P = 0.01) and ileocecal resections (P = 0.0002). Interestingly, reoperation with resection of the anastomosis was significantly more frequent in patients with NOD2/CARD15 mutations (P = 0.01).

Conclusions : Our investigations support the current hypothesis that NOD2/CARD15 mutations are associated with a phenotype of Crohn's disease with younger age at diagnosis, ileal involvement, ileocecal resections and a high risk of postoperative relapse and reoperation. NOD2/CARD15 mutations might therefore be used to identify high risk patients for relapse prevention strategies.


The current working hypothesis for the pathogenesis of Crohn's disease (CD) suggests that it is caused by a dysregulated immune response towards antigens of faecal bacteria in a genetically susceptible host. Although several loci within the human genome have been linked to CD, the identification of mutations in the NOD2/CARD15 gene situated at chromosome 16q12 within the inflammatory bowel disease (IBD)1 region and their association to CD has been a major step forward in understanding the disease pathophysiology.1–3

NOD2/CARD15 is known to act as an intracellular receptor in monocytes for bacterial components triggering activation of NFkappa-B and thus leading to subsequent activation of the inflammatory response. Within the NOD2/CARD15 gene, there are three independent mutations recently found to be associated with CD: two missence mutations (Arg702Trp and Gly908Arg) and one frameshift mutation (3020insC). The 3020insC mutation results in a truncated protein leading to an altered stimulation of NF-kappaB after bacterial activation.1

Although these mutations fit very well into the present pathogenetic hypothesis, they are only found in approximately 30% of CD patients. Therefore genotype phenotype correlations were conducted to find out if NOD2/CARD15 mutations are associated with a distinct clinical subtype of CD. Mutations within the NOD2/CARD15 gene and their correlation to clinical data have been performed by others with differing results. Associations were found to ileal involvement,4–7 fibrostenotic or fistulizing behaviour8 and younger age at diagnosis,9 whereas other investigators failed to notice a relationship to the subgroups of the Vienna classification.10 However, up to now, no clinically relevant role for the NOD2/CARD15 mutations has been described.

Thus, an accurate genotype/phenotype analysis is required to elucidate the role of mutations in the NOD2/CARD15 gene and to characterize in more detail their clinical contribution to the course of CD. We therefore investigated the three common NOD2/CARD15 mutations (Arg702Trp, Gly908Arg, 3020insC) in 180 patients with CD and related the results to the demographic and disease phenotype.



The study was approved by the ethics committe of the Charité in Berlin and informed consent was obtained from each participant.

We included 180 patients in the study with a diagnosed CD including clinical, endoscopic, radiological and histological findings according to standardized criteria.11 All patients were recruited from the Charité Berlin (Campus Mitte and Campus Virchow), a tertiary referral centre. All patients from the disease and control groups were Caucasians. Data were obtained through retrospective collection from the patients clinical charts.

The following data of patients with CD were obtained: age, age at diagnosis, gender, familial or spontaneous disease (familial disease was considered if one first- or second-degree relative had IBD), smoking habits (current smoking/history of smoking/never smoked), disease localization, disease behaviour, extraintestinal manifestations [type 1 peripheral arthralgia, affections of eyes or skin, primary sclerosing cholangitis (PSC)], type and date of surgery (e.g. ileocecal resection, small or large bowel resection, reoperation).

The following definitions were taken into account: disease localization was defined as the maximum extent of digestive involvement at the latest follow-up. Information was obtained through endoscopic (including upper endoscopy, capsule endoscopy or colonoscopy), radiological [small bowel X-ray or computerized tomography (CT) enteroclysma] or histological examination. Stenotic CD was considered if persistent intestinal obstruction was found either in small bowel X-ray, CT enteroclysma, ultrasonography or colonoscopy. Perforating disease was recorded if patients had enterocutaneous, enteroenteric, enterovesical or enterovaginal fistula, intraabdominal abscess or small bowel perforation. Perianal disease was diagnosed if perianal fistula, ulcers or abscesses were present. Extraintestinal manifestations were defined as follows: Typ 1 peripheral arthralgia (defined in Orchard et al.12), presence of PSC diagnosed through endoscopic cholangiography, affections of skin (e.g. presence of erythema nodosum or pyoderma gangrenosum) or eye (e.g. presence of episcleritis or anterior uveitis).

Genotyping for the NOD2/CARD15 mutations (Arg702Trp, Gly908Arg, 3020insC) mutations as also performed in 70 patients with ulcerative colitis (UC) (m:f 25:45; age 20–73) that were also recruited from the Charité Berlin (Campus Mitte). In total, 97 healthy and unrelated spouses of patients served as controls (Co) (m:f 40:57; age 19–68).

Polymerase chain reaction and sequencing

Genomic DNA from index patients and family members’ blood were prepared using commercially available extraction columns (QIAmp Blood Kit; Qiagen, Hilden, Germany).

Exon 11 of the NOD2 gene was amplified using the following primers: forward primer 5′-ggg aca ggt ggg ctt cag ta-3′, reverse primer 5′-cca ttc ctc tct ccc gtc ac-3′; the annealing temperature was 62 °C. DNA sequencing of the amplified exon was performed by cycle sequencing with fluorescent dye terminators. Analysis was performed using an ABI 310 automatic sequencer (Applied Biosystems, Weiterstadt, Germany). The analysis was confirmed by sequencing in both directions.

Single nucleotide polymorphism genotyping

Genotyping of the three most common NOD2/CARD15 single nucleotide polymorphisms (Arg702Trp, Gly908Arg, 3020insC) were performed using the fluorogenic 5′-nuclease assay with the primers and probes listed in Table 1.

Table 1.  Primers and probes used in genotyping of NOD2/CARD15 genetic variants
Single nucleotide polymorphismPrimersProbesAnnealing

Each assay were performed using 25 μL IQ Supermix (BioRad, München, Germany), 0.6 μL of each primer (100 μm), 0.1 μL of each probe (100 μm) (TibMolBiol, Berlin, Germany), 3 μL genomic DNA, and 20.6 μL H2O. Fluorescence were measured over 40 cycles, and allelic discrimination were performed with the ICycler IQ Real-Time PCR Detection System and the appropriate software (BioRad).

Statistical analysis

Comparison of the frequency of the NOD2/CARD15 mutations between patients and Co and the association to the phenotype was performed by Chi-Square test or Fisher's exact test where appropriate. For analysis of age, age at diagnosis and age of ileocecal resection Wilcoxon-U-Mann–Whitney test was applied. P-values <0.05 were considered to be significant. The data were analysed using SPSS/PC+V10.01 software (SPSS, Chicago, IL, USA).


Distribution of NOD2/CARD15 mutations

First, we investigated the frequencies of the three NOD2/CARD15 mutations (Arg702Trp, Gly908Arg, 3020insC) in 250 patients with inflammatory bowel disease and compared them with 97 Co. Genotypes and allele frequencies are shown in Tables 2 and 3. In total, 35.6% of CD patients carried at least one mutant allele within NOD2/CARD15 compared with 14.3% of patients with UC (P = 0.006) and to 15.5% of Co (P = 0.0001, Table 3). Differences between patients with UC and Co were not significantly different. 5.6% of CD patients whereas none of the patients with UC or Co were homozygous for any of the three NOD2/CARD15 mutations (Table 3). Although the two missense mutations (Arg702Trp and Gly908Arg) were found in higher allele frequencies in patients with CD, statistical analysis did not reach significance (Table 2).

Table 2.  Genotypes and allele frequencies for the three NOD2/CARD15 mutations
PatientsGenotypes (%)Allele frequency (%)P-value*
  1. * Fisher's exact test was performed comparing percentages of patients positive or negative for the corresponding NOD2/CARD15 mutation between CD and controls1, CD and UC2, UC and controls3.

  2. Crohn's disease (CD, n = 180), ulcerative colitis (UC, n = 70) and controls (n = 97); genotypes: −/− homozygous wild-type, −/+ heterozygous mutant, +/+ homozygous mutant; absolute numbers (percentages) of individuals negative (−/−), heterozygous (−/+), homozygous (+/+) for the different NOD2/CARD15 mutations (Arg702Trp, Gly908Arg and 1007finsC) are shown.

  3. n.s., Not significant.

 CD156 (86.7)22 (12.2)2 (1.1)26 (7.2)0.161
 UC67 (95.7)3 (4.3)03 (2.1)0.042
 Controls90 (92.8)7 (7.2)07 (3.6)n.s.3
 CD165 (91.7)15 (8.3)015 (4.2)0.221
 UC67 (95.7)3 (4.3)03 (2.1)0.412
 Controls93 (95.9)4 (4.1)04 (2.1)n.s.3
 CD144 (80.0)28 (15.6)8 (4.4)44 (12.2)0.00021
 UC64 (91.4)6 (8.6)06 (4.3)0.032
 Controls93 (95.9)4 (4.1)04 (2.1)n.s.3
Table 3.  NOD2/CARD15 genotypes in inflammatory bowel disease patients and controls
 TotalSimple heterozygote (%)Compound heterozygote (%)Homozygote (%)Any variant (%)No variant (%)P-value*
  1. * Fisher's exact test was performed comparing numbers of individuals positive or negative for any of the three NOD2/CARD15 variants: Crohn's disease and controls1, Crohn's disease and ulcerative colitis2, ulcerative colitis and controls3.

  2. Numbers (percentages) of NOD2/CARD15 genotypes in patients with Crohn's disease (CD), ulcerative colitis (UC) and controls (Co).

  3. n.s., Not significant.

CD18044 (24.4)10 (5.6)10 (5.6)64 (35.6)116 (64.4)0.00011
UC7010 (14.3)0010 (14.3)60 (85.7)0.0062
Co9714 (14.4)1 (1.0)015 (15.5)82 (84.5)n.s.3

Genotype phenotype analysis

In the next step a detailed genotype/phenotype analysis was performed in the 180 patients with CD. The percentages of CD patients carrying at least one NOD2/CARD15 variant were correlated to demographic data, disease localization and behaviour (Table 4). NOD2/CARD15 mutations showed a significant association with younger age at disease diagnosis (P = 0.03) and ileal involvement (P = 0.01; Table 4). With respect to all other clinical and demographic data, including extraintestinal manifestations and fistulizing disease, no significant differences were found in respect to the NOD2/CARD15 genotype (Table 4). Interestingly, no difference was observed when smoking behaviour in both groups was analysed. Moreover the frequency of stenotic disease was not different between the two groups (P = 0.15).

Table 4.  Clinical characteristics of Crohn's disease patients
  1. * Younger age at diagnosis was significantly associated with NOD2/CARD15 mutations (P = 0.03, Wilcoxon-U-Mann–Whitney test).

  2. † Furthermore, ileal involvement was significantly associated with NOD2/CARD15 mutations (P = 0.01, cross table, Fisher's exact test). No other significant differences were found.

  3. Correlation of the demographic and clinical pattern of 180 patients with Crohn's disease positive or negative for NOD2/CARD15 mutations. Absolute numbers (percentages) are illustrated.

Total number64116
Age at diagnosis in years
 Mean26 ± 9.9*30 ± 11.7
Disease duration in years 7 ± 8.7 6 ± 7.0
Familial Crohn's disease (%)4 (6.3)11 (9.5)
Sporadic Crohn's disease (%)60 (93.7)105 (90.5)
Smoking habits (%)
 Never44 (68.8)83 (71.6)
 Current17 (26.6)31 (26.7)
 Ex-smoker3 (4.7)2 (1.7)
Localization (%)
 Upper GI15 (23.4)20 (17.2)
 Ileum60 (93.8)†93 (80.2)
 Right colon48 (75.0)84 (72.4)
 Left colon37 (57.8)88 (75.9)
 Colon only4 (6.3)11 (9.5)
 Perianal disease23 (35.9)43 (37.1)
Behaviour (%)
 Non-stricturing non-penetrating14 (21.9)36 (31.0)
 Stricturing44 (68.8)67 (57.8)
 Penetrating29 (45.3)48 (41.4)
Extraintestinal manifestation (%)
 Typ 1 Arthralgia (peripheral)25 (39.1)38 (32.8)
 Primary sclerosing cholangitis1 (1.6)2 (1.7)
 Skin6 (9.4)5 (4.3)
 Eye3 (2.6)6 (5.2)

We furthermore analysed frequency and location of surgical resections in patients with CD (Table 5). In total, 45.3% of the NOD2/CARD15 positive patients who underwent an ileocecal resection, while only 19.0% of patients without mutations were operated (P = 0.0002; Table 5). This difference could not be explained by different observation times or different age at disease diagnosis (Table 4). The difference was still significant if corrected for the higher frequency of ileal involvement in patients with the NOD2/CARD15 mutations by logistic regression analysis. Time between age at diagnosis and first ileocecal resection did not also differ between the two groups (NOD2/CARD15+: 2.5 ± 8.2 years, NOD2/CARD15-: 3.5 ± 4.5 years, P = 0.76, Wilcoxon-U-Mann–Whitney). Within the patients with stricturing disease behaviour (n = 111) in our cohort, ileocecal resections were performed in 56.8% of patients positive for NOD2/CARD15 mutations compared with 29.9% of patients negative for NOD2/CARD15 mutations (P = 0.006).

Table 5.  Surgery in Crohn's disease patients correlated to NOD2/CARD15 mutations
  1. * In this analysis, only patients with a first ileocecal resection were included (see line ileocecal resection). Fisher's exact test was performed.

  2. Absolute numbers of Crohn's disease patients positive or negative for NOD2/CARD15 mutations and the numbers of resections depending on the localization.

Total number64116 
Surgery (%)
 Small bowel resection4 (6.3)4 (3.4)0.45
 Ileocecal resection29 (45.3)22 (19.0)0.0002
 Colonic resection12 (18.6)13 (11.2)0.18
 Reoparation neoterminal*12 (41.4)2 (9.1)0.01

Interestingly, in 12 of 29 patients (41.4%) with NOD2/CARD15 mutations and only in two of 12 (16.7%) patients without NOD2/CARD15 mutations a second resection of the neoterminal ileum was performed(P = 0.01, Table 5).

Among the patients homozygous for NOD2/CARD15 mutations, an initial ileocecal resection was performed in six of the 10 patients (60.0%) compared with 18 of 44 (40.9%) patients simple heterozygous for NOD2/CARD15 mutations. A neoterminal reoperation was performed in two of six (33.3%) homozygous vs. eight of 18 (44.4%) simple heterozygous patients.

In our cohort, there were four patients with a third resection of the neoterminal ileum and all these patients were positive for NOD2/CARD15 mutations. The frequency of resection of other parts of the intestine than terminal ileum did not differ between the two groups (Table 5). Only few patients with resections in the upper small bowel or colon had reoperations with no difference in frequency between the two groups (data not shown).


Clinical decisions in CD patients are still based on phenotypic disease characterization, which does in most cases neither allow prediction of the course of disease, nor response to a certain therapy. In our study for the first time, we could show that NOD2/CARD15 mutations are not only associated with ileocecal resections in CD but also with a significantly higher risk for reoperation. NOD2/CARD15 mutations might therefore serve as a parameter to select patients for post-operative maintenance therapy.

Studies investigating the frequency of reoperations in patients with mutations in the NOD2/CARD15 gene are very limited. Recently, Helio et al. described a patient homozygous for the 3020insC mutation who had three bowel resection in agreement with our observation.13 Our findings appear to be in contrast to the results reported by Ahmad et al.,5 who could not identify NOD2/CARD15 mutations to be associated with a higher frequency of surgical reoperations. However, they analysed ileal stenotic reoperations without subdividing into reoperations at the neoterminal site and other reoperations within the small bowel. In our analysis, we detected the higher frequency of reoperations associated with NOD2/CARD15 mutations specific for the neoterminal site but not for reoperations at other locations within the small bowel.

In agreement with previous findings, we identified a significantly higher proportion of NOD2/CARD15 mutations (35.6%) in patients with CD compared with patients with UC (14.3%) and Co (15.5%). The prevalence of NOD2/CARD15 mutations described by other investigators showed regional differences among Caucasian populations but were found in similar ranges compared with our study.1–3 In contrast, in a large Japanese cohort of CD patients, none of the three NOD2/CARD15 mutations were found.14 This strengthens the importance of regional differences regarding the contribution of NOD2/CARD15 mutations to CD.

We found an association between NOD2/CARD15 mutations and younger age at disease diagnosis. Similar findings were reported from two other large studies,5, 9 whereas others did not detect this association.4, 7 We also observed a significant association between the NOD2/CARD15 mutations and ileal involvement. This ileal involvement seems to be the most consistent phenotype associated with NOD2/CARD15 mutations.15 The association of ileocecal resection and mutations in the NOD2/CRAD15 gene has been reported by two other studies.5, 8 In the study by Ahmad et al., this association was shown to be a consequence of the higher frequency of ileal involvement.5 Performing logistic regression in our analysis, association with ileocecal resection and NOD2/CARD15 mutations could not be explained by the higher frequency of ileal involvement.

We propose that genotyping for NOD2/CARD15 mutations could have clinical applications. NOD2/CARD15 mutations might help to identify patients with a higher risk for reoperation. We therefore suggest that in further clinical studies investigating post-operative treatment strategies stratifying for NOD2/CARD15 mutations should be undertaken to specifically analyse this subgroup of patients.

In summary, our data do not only confirm other observations of younger age at diagnosis and an ileal involvement of patients with NOD2/CARD15 mutations but also show that these mutations lead to a severe involvement of the ileal region requiring surgery of the terminal ileum in almost two-thirds of the patients in our cohort. In addition, the genotyping for NOD2/CARD15 mutations might serve as potential prognostic in stratifying CD patients with high risk of reoperations.


We thank all patients who have been involved in this evaluation.