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

Purpose To evaluate the role of the nucleotide oligomerization domain 2 (NOD2) mutation status and other risk factors for the incidence of postoperative complications after ileocolic resection for Crohn’s disease (CD). Methods Data of 138 patients consecutively undergoing ileocolic resection for CD at a tertiary academic referral center were retrospectively analyzed including single nucleotide polymorphism (SNP) data of the NOD2 gene. Uni- and multivariate regression analysis was performed to identify factors associated with increased risk of severe postoperative complications. Results From 114 patients (83%), the NOD2 mutation status was available. Of these, 60 (53%) had a NOD2 wildtype, whereas eleven (10%) were homozygous for the high risk p.Leu1007fsX1008 (rs2066847) variant. Major postoperative complications occurred in 28 patients (20%). Twenty-seven of these (96%) were intraabdominal septic complications such as anastomotic leakage or abscess. Male gender (P = 0.029; OR 3.052, the duration of CD (time [months] from initial diagnosis of CD to surgery; P = 0.001; OR 1.009), previous abdominal surgery for CD (P = 0.017; OR 3.49), and the presence of enteric fistulas (P = 0.023; OR 3.21) were identified as independent risk factors for major postoperative complications. Homozygosity for the NOD2 high-risk variant p.Leu1007fsX1008 did not show increased postoperative morbidity in the short and long-term outcome. Conclusions We could detect independent risk factors for major postoperative complications after ileocolic resection for Crohn’s disease. However, patients with the high-risk variant p.Leu1007fsX1008 of the NOD2 gene did not show increased postoperative morbidity.


Introduction
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][3][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].

3
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. NOD2deficient 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.

Design and study population
Demographic and perioperative data of a previously welldescribed cohort [1] of patients undergoing ileocolic resection for histologically proven CD between 2001 and 2010 at the Department of Surgery, Ludwig-Maximilians-University of Munich, Germany, were retrospectively analyzed. A follow-up of these patients was performed until 2021. Patients undergoing repeat (neo-)ileocolic resection were excluded from the study. This retrospective study was approved by the Ethics committee of the Ludwig-Maximilians-University of Munich, Germany (723-16). This retrospective cohort study was conducted according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [14].

Data collection
Preoperative assessment included patient demographics (age, gender) and relevant clinical characteristics such as immunemodulatory or immunosuppressive medication and smoking history. Furthermore, for genotype-phenotype analysis, mutations of the NOD2 gene (the three common CD-associated NOD2 mutations (p.Arg702Trp (rs2066844), p.Gly908Arg (rs2066845), and p.Leu1007fsX1008 (rs2066847)) were derived from the local CD database as previously described [1]. Patient's disease phenotype was classified according to Montreal classification [15]. Among perioperative parameters, the technique of surgery (open vs. laparoscopic assisted) and the estimated intraoperative blood loss (EBL) were assessed. The standard reconstruction technique at our institution is a handsewn double-layer sideto-side anastomosis with a running suture in open and laparoscopic assisted ileocolic resection. From the standardized pathology reports, the resection margins were classified as inflammation-free or not. Postoperative complications were assessed according to the validated Clavien-Dindo classification [16]. Major postoperative complications were defined as those greater than Clavien-Dindo II, and this was defined as main outcome measure as at least some kind of invasive re-intervention was required in these patients. Septic intraabdominal complications were classified as described: anastomotic leakage [17], intraabdominal abscess, and intestinal fistula [6]. For follow-up analysis, a number of different immunosuppressant and immunomodulatory therapies as well as further abdominal surgeries were assessed.

Statistical analysis
Results were expressed as numbers and percentages or median and range (minimum and maximum). In univariate analysis, dependent on the variable's character, chi-squaretest, Fisher's exact test (low frequency), or Mann-Withney test for non-normally distributed values were applied. Significant variables were entered into the multivariate analysis; for this purpose, stepwise logistic regression analysis was performed. Odds ratios (OR) with their 95% confidenceintervals (95% CI) were calculated. Concerning significance levels, P < 0.05 was regarded as statistically significant. For statistical analysis, SPSS software (version 28.0, IBM, Chicago, USA) was used.

Study population and perioperative characteristics
Overall, 138 patients were included in the present study (75 women and 63 men). The clinical baseline and perioperative characteristics are shown in Table 1. The median age at the time of surgery was 35 years (15-69), and the median time span from initial diagnosis of CD to surgery was 50 (0-368) months. Disease localization was limited to ileal (L1) and ileocolic (L3) phenotype in 78 (57%) and 60 (44%) patients, respectively. Overall, 86 patients (62%) had immunosuppressive therapy preoperatively (Table 1). Thirty-one 1 3 patients (22%) had previous abdominal surgery for CD other than ileocolic resection, and 45 patients (33%) had fistulating disease. Among these, entero-enteric and entero-vesical fistulas were the most common (N = 33; 73% of all fistulas).
Ten patients (7%) were observed to have a preoperative hemoglobin level below 10 g/dL, and two patients (1%) below 8 g/dL. No patient exhibited hypoalbuminemia. In eight patients (6%), ileocolic resection was performed as emergency surgery, and all of these procedures were carried out as open surgery. No patient received a primary ostomy. In ; P = 0.034). The overall median blood loss was 200 mL (20-1200 mL). In 30 patients (22%), the resection margin of the bowel showed signs of CD-associated inflammation, but this was not significantly correlated to ileocolic disease localization (P = 1.0) ( Table 1).

Postoperative morbidity
There was no postoperative mortality in our study. Overall, 55 patients (40%) experienced postoperative complications, 28 of these (20% overall) major postoperative morbidity (Clavien Dindo ≥ 3a) ( Table 3). Besides one patient with postoperative hemorrhage and minor septic complication, all of these patients had developed severe intrabdominal septic complications. In 22 patients, re-operation was necessary due to these complications mostly due to anastomotic leak, abscess, and fistula formation. Consecutively, two of them suffered of multi-organ failure and had to be referred to an intensive care unit (Table 3). In five patients, surgical re-intervention led to the creation of an ileostomy. In six patients, CT-guided intervention was performed due to intraabdominal abscess (Table 3).

Univariate and multivariate risk factor analysis
The univariate risk factor analysis for severe postoperative complications is presented in Table 1. Duration of CD, male gender, history of previous abdominal surgery, ileocolic localization (L3), preoperative Hb concentrations, and fistulizing disease were factors significantly associated with major morbidity (Mann-Whitney test, Fisher's exact/χ 2 all P < 0.05), while there was a trend towards an increased risk in patients who underwent emergency surgery (P = 0.053). Immunosuppressive therapy or immunomodulators (steroids, azathioprine, infliximab, adalimumab), smoking, having the homozygous high-risk NOD2 variant p.Leu1007fsX1008 (rs2066847), and resection margins with inflammation did not predict major postoperative morbidity.
In multivariate analysis, only male gender, duration of CD (time span from initial diagnosis of CD to surgery), previous abdominal surgery for CD other than ileocolic resection, and the presence of fistulas were independent risk factors for the incidence of severe postoperative complications (Table 4). This risk was about threefold higher in men (OR 3.1), patients undergone previous surgery (OR 3.49), or with fistulas (OR 3.2). Every year of CD duration increased the odds of morbidity by 11% (0.9% per month).

Discussion
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][19][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][13][14], whereas in other studies, including ours, this was not the case [6,[15][16][17][18][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 immunemodulators 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 highrisk 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, redosurgery 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][36][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][40][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.

Conclusion
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. Department of General, Visceral and Transplant Surgery of LMU clinic. No further funding was received for the present study.

Data availability statement
The data underlying this article cannot be published for ethical reasons to protect the privacy of the subjects included in the study. The data will be released in irreversibly anonymized form to the corresponding author upon justified request.

Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.