It is not NOD2 — genetic and clinical risk factors for postoperative complications following ileocolic resection in Crohn’s disease

Crohn’s disease (CD) is characterized by a chronic relapsing disease course based on a chronic inflammation of the intestine, frequently resulting in a stricturing or penetrating disease phenotype [1]. Despite encouraging improvements in medical treatment of CD by immunomodulators and biologicals [2], surgical therapy is still required in the majority of patients [2‐4]. About one-third of all CD patients eventually undergo surgery for CD [5]. The most common procedures represent surgery for perianal fistulas and bowel resections [6, 7]. Compared to other benign disease, in CD, the rate of postoperative morbidity remains high: a recent prospective study reports a postoperative morbidity rate of 36% [8]. Furthermore, not only in short-term but also in long-term outcome patients experiencing postoperative complications are at risk for undergoing repeated abdominal surgery in the future [5].

Thus, it is important to identify high-risk patients in order to reduce morbidity. So far, several clinical, serological, and genetic factors have been investigated regarding prognosis in CD. Specific polymorphisms of the NOD2 gene have been associated with younger CD onset, ileal disease and ileocolectomies, and some with increased recurrence after surgery [9, 10]. Among these, three variants of the NOD2 gene, p.Arg702Trp, p.Gly908Arg, and p.Leu1007fsX1008, have been reported to be associated with a complicated disease course particularly with a stricturing and penetrating disease behavior [1, 11]. In this context, the frameshift mutation p.Leu1007fsX1008 in exon 11 of the NOD2 gene has been successfully used as a marker for therapeutic decisions [1]. In the gastrointestinal mucosa, NOD2 is widely expressed on various cell types [12]. It is part of a family of intracellular pathogen recognition receptors (PRRs) and its signaling influences microbial activity and prevents pathogenic invasion [12].

Whether it might be a useful predictor for postoperative morbidity has not been conclusively elucidated yet. NOD2-deficient mice exhibit impaired healing of the ileocolonic anastomosis with an intestine-specific effect [13]. Therefore, the aim of this study was to investigate the role of the high risk frameshift mutation of NOD2 p.Leu1007fsX1008 in the context of perioperative morbidity in CD patients undergoing ileocolic resections.

To our knowledge, this is the first study to investigate the impact of high-risk SNP p.Leu1007fsX1008 (rs2066847) status on postoperative surgical outcome in a CD cohort undergoing ileocecal resection. In a recent study, Schnitzler et al. showed that homozygosity of this variant together with active smoking is associated with a 100% risk for developing ileal stenosis requiring surgery [18]. In our analysis, previous immunosuppressive therapy, the homozygous NOD2 variant p.Leu1007fsX1008 and positive microscopic inflammation margins were not associated with increased risk for severe complications. This study identified male gender, the duration of CD, previous abdominal surgery, and the presence of fistulas as independent risk factors for the incidence of major postoperative complications following ileocolic resection in CD patients. One study by Giudici et al. also investigating NOD2 polymorphisms rs2066844, rs2066845, and rs2066847 in CD patients undergoing abdominal surgery in general could not identify homozygosity as risk factor for postoperative complications but a longer time period for disease onset in the p.Leu1007fsX1008 (rs2066847) cohort [19], whereas a study by Geramain et al. found the NOD2 SNP rs5743289 to be an independent risk factor for postoperative abdominal septic complications [20]. A third study by Kline et al. assessed the impact of NOD2 SNPs rs2076756, Arg702Trp (rs2066844), and p.Gly908Arg (rs2066845) [21]. In our cohort, rs2076756 was not assessed and the others were quite rare with only one homozygote Arg702Trp (rs2066844) and one p.Gly908Arg (rs2066845), respectively, but Kline et al. also did not find one of the NOD2 SNPs to be associated with postoperative complications after ileocecal resection [21]. In our follow-up data, we could not detect significant differences in the need for further surgical therapy or increased need for immunosuppressive or immune-modulatory therapy over time regarding the risk factors identified by others [18‐20]: homozygous NOD2 variant, smoking, or fistulating disease.

There is an ongoing debate about the impact of immunosuppressive reagents on postoperative morbidity and the effect of preoperative steroids in particular. In this regard, the data remain rather inconsistent: in some studies, the use of steroids has been observed to increase the risk of postoperative septic complications [8, 12‐14], whereas in other studies, including ours, this was not the case [6, 15‐19]. This inconsistency may be explained by differences in the indications, duration, and dose of steroid treatment [14].

In addition to steroids, other immunomodulators have become standard therapy for CD [22]. In the present study, the influence of azathioprine and anti-TNF antibodies (infliximab, adalimumab) was investigated, and no increased risk was found. A recent systematic review by Law et al. [23] analyzed 63 studies regarding the effects of immunomodulators on postoperative complications after surgery for inflammatory bowel disease and found an increased risk for abdominal sepsis for anti-TNF medication and corticosteroids, whereas the risk was not increased for other immune-modulators and anti-integrins [23]. On the contrary, early initiation of immunomodulatory therapy postoperatively is recommended to reduce the risk of disease recurrence after surgery [24]. The initiation of anti-TNF therapy even in a period less than 2 weeks after surgery did not show an increased complication rate in a big retrospective cohort [25]. This seems to be important as a recent study has demonstrated an association between low anti-TNF drug levels and unfavorable postoperative outcomes in CD patients [26].

We know from several studies that poor nutritional status, such as low serum albumin or protein levels [27], low hemoglobin [28], or low BMI [29, 30], may put CD patients at risk for postoperative complications after abdominal surgery. As our cohort did not show clinical and laboratory signs of malnutrition, this can rather be excluded as a possible bias.

In our study, 22% of patients had a microscopically inflamed resection margin, although this was not a risk factor for complications in this analysis. Again, the data are rather heterogeneous: a recent observational study showed that CD patients with anastomotic leakage after elective ileocecal resection were more likely to have a positive resection margin [31], but our and other data [32] suggest that it is not necessary to achieve completely inflammation-free margins.

Homozygosity for the NOD2 frame shift mutation p.Leu1007fsX1008 has been described as a strong marker for a severe clinical phenotype of CD associated with more severe and penetrant disease behavior [1]. In children with CD, NOD2 mutation carrier status was associated with the need for surgery until 17 years of age [33]. Of the 114 patients with available NOD2 SNP status in our current study, 11 (9.6%) had the homozygous variant p.Leu1007fsX1008. This is twice as many as in a CD cohort from our colleagues in the Division of Gastroenterology [1], underscoring that surgery is more likely in this high-risk population of CD patients with this particular NOD2 frameshift mutation. However, this variant itself was not associated with higher morbidity in our study, indicating that the variant does not have an adverse effect on short-term postoperative outcome in these patients.

This study is subject to some limitations. Due to the retrospective design, the generalizability is limited as our data represent the experience from a single-center population and thereby reflect a small patient cohort. However, our study reports data from a genetically well-characterized CD population undergoing ileocolic resection with a long-term follow-up of median 123 month. Due to the data collecting period from 2001 to 2010, the patients did not receive further antibody treatment than anti-TNF. Therefore, this does not necessarily reflect todays situation, but this does not seem to be a strong bias since the overall complication rate of 40% is similar to those reported by others [8]. Nevertheless, redo-surgery was more frequent in our cohort with about 20% of cases. This can be explained by the high rate of fistulating disease in one-third of the cohort, which was significantly associated with the occurrence of postoperative complications. It is known that CD patients operated on in the acute stage are more likely to have an unfavorable postoperative course with about 40% increase of postoperative complications [8, 34]. The presence of preoperative fistulas as another surrogate parameter for severe and advanced disease progression has been identified by us and others [35‐37] as a risk factor, as well as the duration of disease. It had been shown that patients undergoing early surgery had favorable disease course [38]. However, in the era of effective treatment options and interdisciplinary approaches, further studies will be required to figure out which patients will benefit from early surgical intervention and which will manage for a long time with immune-modulatory or immunosuppressive treatment. Therefore, the problem remains that surgeons often see patients too late when they are in a situation where only surgery can be recommended. In this regard, Iesalnieks et al. [6] and Alves et al. [39] reported that “recurrent clinical episodes” and “duration of symptoms leading to surgery” were associated with postoperative morbidity. Similarly, in our study, the time from initial diagnosis of CD to surgery was identified as an independent risk factor, with each year of CD duration increasing the risk by 11% per year. This confirms the importance of a multidisciplinary approach in which surgeons see patients early and repeatedly throughout the course of their disease.

In this study, fistulating disease was an important factor for performing open rather than laparoscopic resection. In our population, 59% of the open group and only 9% of the laparoscopic group had fistulating disease. Laparoscopic resection of the ileocolic is a commonly performed operation for CD that provides short-term benefits [39‐41]. The laparoscopic approach for ileocecal CD has been increasingly used in recent years, especially in patients without fistulas or prior abdominal surgery. In our population, 65 patients (47%) underwent laparoscopic surgery, and 7 patients required conversion to open surgery. Overall, probably because of patient selection for the laparoscopic procedure (e.g., intestinal fistulas, previous abdominal surgery, age, and emergencies), patients with open surgery had a significantly higher morbidity rate (27% versus 12%). Of the 7 patients who required conversion, all had extensive abdominal adhesions that technically interfered with laparoscopy, and three had fistulas that had not been apparent preoperatively. No major complications occurred in this group. This is consistent with other publications in which no increased complication rate was observed in converted patients [6, 39]. Since adverse short- and long-term sequelae (especially recurrence) have been observed in patients with postoperative complications after abdominal surgery in CD, the identification of risk factors is of great importance, especially in the era of personalized medicine.

Male gender, duration of CD, previous abdominal surgery, and fistulating disease but not NOD2 polymorphisms were identified as independent risk factors for the development of postoperative complications after ileocolic resection for CD. All of these risk factors should be considered in risk stratification to choose the best surgical approach (e.g., laparoscopic vs. open, stoma) for the individual patient.