Mesenteric granulomas independently predict long‐term risk of surgical recurrence in Crohn's disease

Abstract Aim The risk factors that predict surgical recurrence in Crohn's disease (CD) remain controversial. Postoperative anti‐tumour necrosis factor (anti‐TNF) therapy might lower recurrence rates whilst the presence of mesenteric granulomas has been postulated to increase the risk. We hypothesized that mesenteric granulomas indicate disease severity and might predict the risk of surgical recurrence, irrespective of immunosuppressive therapy. Method We performed a retrospective review of all consecutive patients undergoing operations for CD between January 2000 and December 2014 at a single tertiary referral centre and assessed the perioperative factors and histological findings at the time of surgery. Surgical recurrence rates and the immunosuppressive regimen were assessed through retrospective chart review and telephone interviews. Results A total of 274 patients were eligible for analysis. Median follow‐up was 8.54 (5.48–14.42) years. A total of 63 patients (23.0%) underwent surgery for recurrent CD after a median of 4.75 (2.10–7.96) years. In final histology, 35 (12.8%) patients had mesenteric granulomas. TNF inhibitors were administered postoperatively in 104 (38.0%) and thiopurines in 137 (50.0%) patients. In univariate analysis, only the presence of mesenteric granulomas [hazard ratio (HR) 1.95; 95% CI 1.05–3.62; P = 0.035] significantly increased the risk for recurrent surgery while postoperative anti‐TNF (HR 0.85; 95% CI 0.49–1.50; P = 0.581) or thiopurine therapy (HR 1.03; 95% CI 0.61–1.73; P = 0.916) did not. In multivariate analysis, only the presence of mesenteric granulomas significantly influenced the risk of surgical recurrence (HR 1.94, 95% CI 1.04–3.60; P = 0.037). Conclusion Intestinal and mesenteric granulomas should be differentiated in pathology reports, because mesenteric, but not intestinal, granulomas may be associated with an increased risk of surgical recurrence.


Introduction
Medical treatment is the main strategy for achieving disease control in Crohn's disease (CD), but the lifetime risk of CD-related complications that require surgical intervention is still high [1]. Owing to the use of thiopurines and biologicals as well as interventional endoscopy, the 5-year risk of undergoing surgery after diagnosis has significantly declined and is currently around 30% [2]. This is partly due to the fact that the gap between diagnosis and initiation of anti-tumour necrosis factor (anti-TNF) therapy has decreased [3]. Recently, several studies have suggested a favourable outcome with regard to endoscopic and surgical recurrence in patients who receive postoperative anti-TNF therapy [4][5][6][7]. However, these studies were conducted mainly after index ileocolic resection and comprise relatively small sample sizes, warranting further proof-of-concept work. Additionally, immunosuppression bears the risk of infection and other adverse events. Therefore, an improved risk assessment for surgical recurrence would potentially allow a more patient-tailored selection of postoperative immunosuppression.
Several established risk factors such as age, smoking and prior resection contribute to the risk of endoscopic disease recurrence and can only be partially influenced by clinicians [8,9]. Granulomas have been postulated as objective histological indicators of disease severity and thus might indicate an increased risk of postoperative endoscopic and surgical recurrence. However, results remain inconclusive in the available literature [10][11][12]. These discrepant findings might be due to a lack of differentiation between granuloma site (e.g. intestinal or mesenteric). In a recent work by Li et al. [11], postoperative recurrence in 194 patients who underwent index ileocaecal resection for CD-related complications was analysed according to granuloma site. The authors identified mesenteric lymph node granulomas as an independent prognostic factor for endoscopic as well as surgical recurrence, while intestinal granulomas did not prove valuable as a prognostic factor. These findings are of great interest, as lymphatic vessels are altered in CD and might play a key role in pathogenesis [13]. Moreover, mesenteric resection has recently been suggested to potentially reduce surgical recurrence rates [14]. Unfortunately, the generalizability of results from index ileocaecal resection is limited. In addition, the utilization of postoperative immunosuppression was relatively low in the published literature. Thus, more data on the prognostic value of objective histological hallmarks in a more general, well-characterized population are warranted.
Given the available data, we hypothesized that the presence of mesenteric granulomas might serve as an objective marker in a general retrospective cohort of patients and independently predict the risk of surgical recurrence.

Method Setting and participants
We performed a retrospective review of all consecutive patients operated on for CD between 1 January 2000 and 31 December 2014 at the Division of General Surgery, Department of Surgery, Medical University of Vienna and assessed perioperative factors and histological findings at the time of surgery.
After obtaining approval from the Medical University of Vienna's institutional review board, patients were retrospectively identified, and all patients were contacted by post. All eligible subjects were asked to supply written informed consent to participate in the study, as required by the local ethics committee. Telephone interviews and/or questionnaires assessing long-term surgical recurrence rates and immediate postoperative immunosuppressive regimens were conducted between May 2017 and August 2018.
Of the 445 patients eligible for inclusion, we had to exclude a total of 171, mainly due to missing informed consent. Detailed information is available in Fig. 1.

Variables
In order to investigate the influence of several (established) risk factors on long-term surgical recurrence rates, we assessed sex, age, body mass index (BMI), American Society of Anesthesiologists (ASA) grade, family history of inflammatory bowel disease, number of Crohn's-related operations and smoking status at the time of surgery between January 2000 and December 2014.
Operative factors such as surgical approach (laparoscopic versus open), urgency of surgery (emergency versus elective), intra-operative findings (perforation, abscess, stricture) and type of resection were extracted from surgery reports. For analysis of surgical complexity, we defined 'simple resection' as only one resection without any additional resection and/or strictureplasty. The presence or absence of mesenteric and/or intestinal granulomas was extracted in a dichotomous fashion (present/absent) from full-length pathology reports.

Perioperative outcome and postoperative immunosuppression
Postoperative prophylactic immunosuppressive therapy (thiopurines, steroids, biologicals), postoperative complications (according to the Clavien-Dindo classification [15]) and length of hospital stay were extracted from discharge reports and/or outpatient records and data was verified by telephone interviews/questionnaires.
For uni-and multivariate analysis, minor complications were defined as Clavien-Dindo grades I and II and major complications were considered as Clavien-Dindo grades III, IV and V.

Definition of surgical recurrence
Surgical recurrence was defined as the need for any reoperation under general anaesthesia due to recurrence of complicated CD. Balloon dilatations were not considered as surgical recurrence. The date of surgery was assessed by individual chart review and verified via telephone interview/questionnaire to assess surgical recurrence in patients who were followed in other centres.
All patients who were included in the study had been evaluated and discussed in a multidisciplinary team meeting for inflammatory bowel disease prior to surgery.

Ethical considerations and informed consent
The retrospective study with written informed consent and data verification was conducted in accordance with the Declaration of Helsinki and approved by the local review board (EK numbers 1700/2016 and 1881/2017; access https://ekmeduniwien.at/core/catalog/2016/).
The STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) checklist was used to assure appropriate quality reporting [16].

Statistical analysis
Continuous variables are reported as median (Q1-Q3) and categorical variables are reported as number (n) of patients with the certain characteristic (proportion of patients with the certain characteristic in %). The Mann-Whitney U-test was used as numerical variables were not normally distributed.
Pearson's chi-square test or Fisher's exact test were used for group comparisons. Characteristics that might affect surgical recurrence-free survival were analysed using semiparametric Cox proportional hazard models. Univariate Cox regression was carried out to determine factors showing a trend towards improved/impaired surgical recurrence-free survival. Then all univariate factors with a P-value < 0.1 were included in a separate multivariate model. Patients entered the model on the day of the respective first surgery within the study period and were followed until either the time of CD-related reoperation (i.e. surgical recurrence; defined as event) or last patient contact (defined as date last seen within the hospital or date of questionnaire return).
The log-rank test was used to identify differences in surgical recurrence-free survival times between patients with or without mesenteric granulomas ( Fig. 2) and patients on different therapies in the respective subgroups ( Fig. S1 in the online Supporting Information).
Two-sided P-values < 0.05 were considered statistically significant. In addition, the majority of surgical resections were considered simple (195/274, 71.2%), defined as only one resection without any additional resection and/or strictureplasty. In histological work-up, 50% of the patients (137/274) had intestinal granulomas in the specimen, while only 12.8% (35/274) showed mesenteric granulomas.
A total of 63 patients (23.0%) had to undergo recurrent surgery during follow-up for CD-related complications after a median of 4.75 (2.10-7.96) years. Of the 63 patients who were considered as undergoing 'recurrent surgery during follow-up', 3 required subtotal colectomy, 15 underwent subsequent ileocaecal resection, 8 had to undergo anastomotic resection, 18 underwent small bowel resection with or without additional resection (1 had an additional colonic resection due to a stricture and 3 required stoma formation due to perianal disease but without abdominoperineal excision), 13 underwent colonic segmental resection, 3 underwent abdominoperineal excision due to perianal disease and 3 had multiple strictureplasties without resection. The patient characteristics of the respective subgroups are presented in Table S1.
Mesenteric granulomas as the only independent risk factor for recurrent surgery To assess risk factors for surgical recurrence, univariate and multivariate Cox regression analyses were conducted (Table 2). Published factors that influence postoperative recurrence risk were included in univariate analysis and all variables that showed a trend to influence postoperative surgical recurrence (P < 0.1) were included in the multivariate analysis.  In univariate analysis, only mesenteric granulomas in final histology [hazard ratio (HR) 1.95; 95% CI 1.05-3.62; P = 0.035] and smoking at the time of surgery (HR 1.58; 95% CI 0.92-2.70; P = 0.097) showed a trend to affect risk of postoperative recurrence. Postoperative immunosuppressive regimens (anti-TNF therapy, azathioprine/6-mercaptopurine or steroids) did not affect recurrence risk, neither did prior operations due to CD, surgical approach or postoperative complications. Therefore, only mesenteric granulomas and smoking at the time of surgery were included in the multivariate analysis. Finally, only the presence of mesenteric granulomas was independently associated with risk of surgical recurrence (HR 1.94; 95% CI 1.04-3.60; P = 0.037).

Differences in patients with or without mesenteric granulomas
To assess differences in surgical recurrence-free survival in patients with and without mesenteric granulomas, a Kaplan-Meier survival curve was plotted to compare groups (Fig. 2). Recurrence-free survival after 1, 5 and 10 years was 98.7%, 90.3% and 80.0%, in patients without granulomas and 100%, 82.8% and 68.0%, in patients with mesenteric granulomas in histological work-up (log-rank P = 0.038). To further identify differences between groups, we outline the baseline characteristics in Table 3. Patients with mesenteric granulomas had a tendency towards younger age compared to patients without granulomas [31.32 (24.10-38.35) years vs 36.1 (26.93-46.73) years, respectively; P = 0.052] and significantly higher rates of concomitant intestinal granulomas (71.4% vs 46.9% in patients with granulomas versus those without granulomas; P = 0.010). Moreover, we found a tendency towards more intense immunosuppression postoperatively, reflected by more common use of anti-TNF therapy (52.9% in patients with mesenteric granulomas versus 36.4% in those without mesenteric granulomas; P = 0.089). There were no differences in utilization of azathioprine or steroids. No differences could be found when subgroup analyses of patients on postoperative monotherapy with and without mesenteric granulomas were compared via a logrank test (Fig. S1).

Discussion and conclusion
In this retrospective single-centre cohort study we have investigated several established as well as recently suggested risk factors for surgical recurrence of CD. In the general cohort of patients investigated, only granulomas found in the mesentery of resected specimens predicted the risk of surgical recurrence while other patient-specific and treatment-specific factors did not. As utilization of this recently suggested risk factor could help to identify high-risk patients, these findings must be put into the context of the available literature.
Due to the specialized care required for CD patients, surgery is commonly deferred. Abdominal operations remain the 'last resort' for more complex disease courses only after several conservative approaches have failed [17]. Despite common utilization of minimally invasive approaches [18], surgery-related morbidity remains. In addition, while many studies have investigated immunosuppressants after index ileocaecal resection, limited evidence is available on the optimal immunosuppressive regimen in CD patients with refractory disease who require recurrent surgery. Moreover, trials that evaluate anti-TNF therapy tend to be focused on endoscopic rather than surgical recurrence [19].
This gap in available data highlights the need for evidence-based treatment approaches since (cumulative) digestive damage might be underestimated by established scores such as the CD activity index [20]. Thus, due to the variance in disease course, objective histological hallmarks for grading disease severity would be favourable for decision-making [8]. A recent study conducted by Randolph et al. [13] has revived interest in the alteration of lymphatic vessels in CD and shown that B-cell-rich aggregates resembling tertiary lymphoid organs impinge on lymphatic collecting vessels adjacent to mesenteric lymph nodes. These alterations might be causative or a result of secondary changes due to chronic inflammation. With respect to this, more extensive resection of the mesentery has been suggested to decrease rates of surgical recurrence after ileocolic resection, although a prospective evaluation of these results is lacking to date [14]. Taken together, granulomas in the mesentery might be caused by excessive inflammation and their presence or absence could allow 'grading' of disease severity. In contrast, granulomas in the bowel wall seem to be less specific for a complicated disease course/disease duration. This is in line with a meta-analysis that found higher reoperation rates and faster recurrence in granulomatous CD [12]. Using a more differentiated approach, Li et al. [11] recently showed that mesenteric granulomas had a prognostic impact on endoscopic and surgical recurrence after index ileocaecal resection while intestinal granulomas did not. However, the postoperative utilization of immunosuppression was relatively low in their study and the impact on recurrence rate is not stated. In our study we have independently validated the findings of Li et al. in a larger and more generalized cohort of CD patients. However, the higher utilization rate of anti-TNF therapy in patients with mesenteric granulomas did not reduce the risk of recurrence in our retrospective cohort. These findings are in line with a nationwide cohort study from Denmark that evaluated a total of 48 967 individuals and did not find a 'convincing surgery-sparing effect' of higher rates of use of immunosuppression [21]. In contrast, several prospective studies in small cohorts of patients undergoing index ileocaecal resection have shown benefits for early initiation of anti-TNF therapy with regard to endoscopic and surgical recurrence rates [5][6][7]22]. These discrepant findings in the published literature can probably be attributed to the different inclusion criteria, and collectively indicate that future studies should focus on improving reoperation rates in high-risk patients. In addition, these discrepancies between 'real-world' studies and controlled trials emphasize the need for studies in patients suffering from more complicated diseases (e.g. a second or third surgical intervention). Of note, our study has some limitations that need to be addressed. Firstly, we had to exclude a significant number of patients from the analysis (171/445, 38.4%) due to missing informed consent to participate in the study, which was required by the institutional review board. Therefore, we cannot rule out a selection bias. Secondly, despite prospective telephone interviews and confirmation of our compiled data, the study remains a retrospective one with all the associated limitations. In particular, switching of immunosuppressants within the follow-up period and patient noncompliance cannot completely be ruled out. Due to the impossibility of reliably assessing endoscopic recurrence/interventions in this setting, only surgery under general anaesthesia was considered as 'surgical recurrence' while balloon dilatation, for example, was not. Thirdly, although CD histology is evaluated by specialized colorectal pathologists in our centre and granulomas are usually reported in the final reports, the mesentery is not extensively resected in our centre and some granulomas that are located more distant to the bowel wall could potentially be missed. Finally, our tertiary referral centre has a high rate of complex disease courses and patients with medically refractory disease, potentially limiting the applicability to other centres.
In conclusion, mesenteric but not intestinal granulomas reflect disease severity and should trigger closer monitoring of affected patients. While immunosuppression did not alter the risk of surgical recurrence in the presented population, future studies must focus on patient-tailored approaches to further reduce recurrent surgery rates.