Prognostic Value of the Modified Rutgeerts Score for Long-Term Outcomes After Primary Ileocecal Resection in Crohn's Disease

INTRODUCTION: The prognostic value of the modified Rutgeerts score (mRS) in patients with Crohn's disease (CD) needs to be further elucidated. This study assessed the prognostic value of the mRS for long-term outcomes after primary ileocecal resection in patients with CD. METHODS: Patients with CD after primary ileocecal resection with an available mRS at first postoperative ileocolonoscopy (index mRS) were retrospectively included. The primary outcome was surgical recurrence. Secondary outcomes were clinical recurrence and progression to severe endoscopic recurrence (≥i3). Cox proportional hazard models were used to assess the association between index mRS and outcomes. RESULTS: Six hundred fifty-two patients were included (mean follow-up: 6.4 years, SD: 4.6). Surgical recurrence rates were 7.7%, 5.3%, 12.9%, 19.1%, 28.8%, 47.8% for index mRS i0, i1, i2a, i2b, i3, and i4, respectively. Clinical recurrence occurred in 42.2% (i0), 53.7% (i1), 58.5% (i2a), 80.2% (i2b), 79.4% (i3), and 95.3% (i4) of patients. Progression to severe endoscopic recurrence occurred in 21.1% (i0), 33.9% (i1), 26.8% (i2a), and 33.3% (i2b) of patients. An index mRS of i2b (adjusted hazard ratio [aHR] 3.0; 1.5–5.6), i3 (aHR 4.0; 2.0–7.9) and i4 (aHR 8.0; 4.0–16.0) were associated with surgical recurrence. An index mRS of i1 (aHR 1.7; 1.2–2.4), i2a (aHR 1.7; 1.2–2.4), i2b (aHR 4.4; 3.2–6.0), i3 (aHR 3.6; 2.5–5.2), and i4 (aHR 7.3; 4.8–10.9) were associated with clinical recurrence. An index mRS of i1 (aHR 2.0; 1.1–3.7) or i2b (aHR 2.5; 1.4–4.6) was associated with progression to severe endoscopic recurrence. DISCUSSION: The increasing mRS corresponds closely with the risk of surgical and clinical recurrence. An index mRS ≥ i2b is associated with surgical recurrence, an index mRS ≥ i1 is associated with clinical recurrence, and i1 or i2b with progression to severe endoscopic recurrence. These results support tight monitoring of disease activity and treatment optimization in patients with ileal lesions and a more conservative management in patients with anastomotic lesions.


INTRODUCTION
Patients with Crohn's disease (CD) are still at considerable risk of an intestinal resection although the risk has declined over the past decades (1).An intestinal resection is an important treatment modality, which is performed in approximately 25% of patients within 10 years after CD diagnosis (2).An ileocecal resection (ICR) is the most common surgical procedure in CD (3).Despite an intestinal resection may induce disease remission and provide relief of CD symptoms, surgery is not curative and recurrence at the ileocolic anastomosis and/or in the neoterminal ileum is common (4)(5)(6).
Ileocolonoscopy is considered the golden standard for the diagnosis of postoperative recurrence in patients with CD (7).The Rutgeerts score (RS) was developed as an endoscopic scoring system to assess the severity of recurrence of inflammation at the ileocolic anastomosis and in the neoterminal ileum.The original RS stratifies the endoscopic severity into 5 groups (i0-i4) (8).High indices of the RS ($i2) are associated with a higher risk of clinical recurrence and a re-resection when compared with a lower RS (i0-i1) (9).However, the prognostic value per index score of the RS is unknown.
The modified Rutgeerts score (mRS) was proposed to differentiate i2 into lesions confined to the anastomosis (i2a) versus lesions in the neoterminal ileum (i2b) and is currently used to assess the severity of postoperative endoscopic recurrence (10).The nature of anastomotic lesions (i2a) is unknown and may be related to a postischemic surgical phenomenon or related to staples, instead of CD recurrence (11).Several studies have reported conflicting clinical outcomes of anastomotic lesions on several measures of postoperative recurrence (clinical recurrence, surgical recurrence, and/or progression to [severe] endoscopic recurrence) (12)(13)(14)(15)(16)(17)(18).In a recently published individual participant data meta-analysis, no difference was observed between i2a and i2b lesions for clinical recurrence and/or a surgical reintervention (19).However, no adjustment for known risk factors was conducted for the latter outcome.In addition, progression to severe endoscopic recurrence was not assessed.Therefore, the initiation or optimization of medication after an endoscopic diagnosis of ulcerations at the ileocolic anastomosis remains a matter of debate.
In this cohort study, we assessed the prognostic value of the mRS (per index score), after correction for known clinical risk factors, to predict the risk of surgical and clinical recurrence, and progression to severe endoscopic recurrence after primary ICR in patients with CD.

Participants and study design
Consecutive patients who underwent a primary ICR for the indication of CD between 2000 and 2019 were identified from a multicenter, retrospective database from 6 academic and 4 teaching hospitals in the Netherlands.All patients with CD (i) aged 16 years or older, (ii) who underwent ICR with restoration of the intestinal continuity, and (iii) who had $ 1 postoperative ileocolonoscopy assessed with the use of the mRS were included.Exclusion criteria were a permanent stoma, a re-resection before the first postoperative endoscopic assessment, prior intestinal resections, other indications for ICR (e.g., gastrointestinal malignancy), and/or absence of follow-up data.
Copyright © 2023 The Author(s).Published by Wolters Kluwer Health, Inc. on behalf of The American College of Gastroenterology The American Journal of GASTROENTEROLOGY

Outcomes
The primary outcome of this study was surgical recurrence (i.e., re-resection of the small bowel and/or colon) for CD recurrence during follow-up.Surgical recurrence within 3 months from primary ICR was considered as a re-resection due to postoperative complications and not considered as surgical recurrence.The secondary outcomes were (i) clinical recurrence defined as CD-related complaints with subsequent endoscopic recurrence (mRS $ i2b), surgical recurrence, radiologic recurrence (assessed by a local radiologist on ultrasonography, computed tomography, or magnetic resonance imaging), and/or therapeutic optimization (i.e., initiation of corticosteroids, immunomodulators, or biologicals for symptomatic disease) and (ii) progression to severe endoscopic recurrence (mRS $ i3) in patients with an index mRS i0-i2b.

Data collection
Baseline and clinical data were retrieved from individual medical charts including demographics, surgical and disease characteristics, and prior medical treatment.The date of index ileocolonoscopy (i.e., first operative ileocolonoscopy) was set as start of the follow-up and time at risk of this study.The mRS at the first postoperative ileocolonoscopy (i.e., index mRS) was used to assess the outcomes.The mRS was graded separately by 4 trained physicians (S.B., J.A., E.B., and J.S.) based on available photographs and/or the endoscopy report for all patients.Follow-up time was defined as the interval between the index ileocolonoscopy (t0) and time to event.Patients were censored in case of the event was not observed (i.e., end of follow-up or lost to follow-up).

Statistical analyses
Descriptive statistical analysis (frequency, percentage, mean, SD, median, and interquartile range [IQR]) was used to describe the research sample.Categorical variables were quoted as the number and percentage.Continuous variables were tested for normality using the Shapiro-Wilk test.Normal distributed variables were presented as mean and SD, while non-normal distributed variables were presented as median and IQR.Kaplan-Meier curves, with log-rank test for significance, were used to describe and compare survival probabilities between individual mRS.
Associations between index mRS and known clinical risk factors (according to the current guidelines) and the 3 time-to-event outcomes (surgical and clinical recurrence, and progression to severe endoscopic recurrence) were investigated using Cox proportional hazard models (7,20,21).The following variables were included for multivariable analysis: age at diagnosis, penetrating disease behavior at surgery (according to the Montreal classification), maintenance therapy during follow-up (i.e., continuation of postoperative prophylactic medication or start of medication within 6 weeks after index ileocolonoscopy with an antitumor necrosis factor agent [anti-TNF] and/or an immunomodulator), and time to index ileocolonoscopy (7,20,21).The models included a random effect for the study center to take potential correlation into account between patients treated in the same hospital.
Because severe endoscopic recurrence is not observed directly and only known to lie within the interval between the first ileocolonoscopy at which it was not yet present and the last ileocolonoscopy at which it was diagnosed, sensitivity analysis with interval censoring for severe endoscopic recurrence was performed.Analyses were performed in R version 4.1.3(R Core Team 2022) with the help of the packages icenReg (version 2.0.15) and survival (22).

Ethics
This study was performed in accordance with the Declaration of Helsinki and approved by the Medical Ethical Research Committee of the Erasmus University Medical Centre Rotterdam (MEC-2017-1151).

DISCUSSION
In this study, the increasing mRS corresponds closely with the risk of surgical and clinical recurrence in patients with CD after a primary ICR, but not with the risk of progression to severe endoscopic recurrence.In multivariable analysis, anastomotic lesions (i2a) were not associated with a re-resection, in contrast to an index mRS $ i2b.Similarly, anastomotic lesions were not associated with severe endoscopic recurrence, in contrast to mild lesions in the neoterminal ileum (index mRS of i1 or i2b).An index mRS $ i1 is associated with clinical recurrence.Tight monitoring to timely optimize medication seems indicated in patients with inflammation in the ileum (index mRS of i1 and $ i2b) to prevent progression to severe endoscopic recurrence and/ or surgical recurrence.In patients with inflammation confined to the anastomosis, a more conservative approach seems appropriate.

INFLAMMATORY BOWEL DISEASE
Bak et al (14,17,18).The more indolent disease course in patients with anastomotic lesions when compared with ileal inflammation regarding progression to severe endoscopic lesions has also been shown in 2 retrospective multicenter studies (14,17).In addition, Hammoudi et al (18) reported a shorter clinical recurrence-free survival in patients with ileal lesions at index ileocolonoscopy when compared with patients with lesions confined to the anastomosis.These findings are in line with our results showing that an index mRS of i1 is associated with both clinical recurrence and progression to severe endoscopic recurrence, whereas an index mRS of i2a is merely associated with clinical recurrence.These outcomes may be explained by a distinct pathological mechanism of anastomotic lesions when compared with ileal lesions, in which the role of ischemia is debated (18,23).A recent published metaanalysis with individual patient data reported no difference was observed between i2a and i2b lesions on the outcomes of clinical recurrence and/or a surgical reintervention (19).However, the analyses for a surgical reintervention were not corrected for known risk factors associated with recurrence.In this study, after adjusting for known clinical risk factors, an index mRS $ i2b was found to be independently associated with surgical recurrence and progression to severe endoscopic recurrence, which supports the recommendation to consider therapy optimization in patients with an index mRS of $ i2b after primary ICR.Despite the lack of a statistically significant association between anastomotic lesions and surgical recurrence and progression to severe endoscopic recurrence, the risks for both outcomes were still as high as 12.7% and 26.8% during follow-up.Further research to identify risk factors and/or biomarkers for postoperative recurrence is warranted to appropriately manage patients with anastomotic lesions.The need for more accurate biomarkers seems underscored by the lack of association between clinical risk factors, except for active smoking and maintenance therapy with an immunomodulator with clinical recurrence and long-term outcomes in multivariable analysis in this study.
Recently, a new endoscopic scoring system has been proposed in which endoscopic scoring should be adapted to the anastomotic technique (24).The (m)RS has been developed for the assessment of an end-to-end anastomosis.In the modern era, a wide lumen stapled side-to-end or side-to-side anastomosis has been preferred over the end-to-end anastomosis to prevent anastomotic leakage, fecal stasis, and stenosis of the anastomosis.When the (m)RS is applied to endoscopically assess these anastomotic techniques, anatomic locations such as the ileal blind loop and ileal body are disregarded (24,25).Prospective analysis of inflammation at these locations and subsequent refinement of the endoscopic score is awaited.
Our study was the first to assess the predictive value of the mRS on long-term outcomes in postoperative patients with CD.Despite the consideration of objective outcome measures in a large population (from both academic and nonacademic hospitals) of patients who underwent a primary ICR with long-term follow-up, limitations of this study need to be taken into consideration.First, because the indication of subsequent ileocolonoscopies could not be assessed, due to the retrospective design, confounding by indication may be present.Second, because our study concerns a wide period, several changes of postoperative management may have influenced the outcomes including improved access to endoscopy, development of strict and noninvasive monitoring, and medication strategies.This study design did not allow to correct for all these potential confounding factors.Regarding the changes in the postoperative endoscopic strategy, a substantial number of patients (40%) did not undergo an index ileocolonoscopy within 1 year postoperatively, which is recommended by the current guidelines (7,20,21).To adjust for potential confounding, we have included time to index ileocolonoscopy in the multivariable analysis.Finally, perianal fistulizing disease, plexitis, and/or granulomas in the resection specimen are considered risk factors of postoperative recurrence in current guidelines (7,20,21).Due to the restriction of number of variables that could be included in multivariable analysis, the findings are not corrected for the presence of perianal fistulas.In addition, standardized data on the presence of plexitis and/or granulomas in the resection specimen were unavailable in the pathology reports.
In conclusion, the increasing mRS at index ileocolonoscopy corresponds closely with the risk of surgical and clinical recurrence after primary ICR.Anastomotic lesions (i2a) are not associated with surgical recurrence and progression to severe endoscopic recurrence, in contrast to lesions in the neoterminal ileum ($i2b).An index mRS $ i1 is associated with clinical recurrence.In addition, i1 lesions are associated with progression to severe endoscopic recurrence.These results support conservative management and no need for escalation of therapy in patients with anastomotic lesions and tight monitoring of disease activity and treatment optimization in patients with ileal lesions.

Table 1 .
Baseline characteristics of the study cohort (n 5 652)

Table 2 .
Multivariable Cox proportional hazard models