Analysis of the risk of future gastrointestinal surgery in Crohn’s disease with stricture

Abstract

Background:

Stricture in patients with Crohn’s disease (CD) carries a high risk of CD-related surgery in the course of the disease. The aim of this study was to assess the rate of occurrence of CD-related surgery and to determine baseline risk factors predicting subsequent surgery in this patient group.

Methods:

Patients registered with stricturing CD were included. All baseline and follow-up data were collected retrospectively. Patients attended the clinic for follow-up at week 14 to assess their response to infliximab (IFX). CD-related surgery was the observational endpoint. Univariate and multivariate Cox regression analyses were used.

Results:

A total of 123 patients with stricturing CD were included in this study. The cumulative risk of CD-related surgery for years 1–5 after diagnosis was 18.0%, 26.7%, 32.6%, 40.7%, and 46.4%, respectively. Prior gastrointestinal (GI) surgery, low body mass index (BMI), and high platelet count might be risk factors for future CD-related surgery. With 97 participants treated by IFX, prior GI surgery and primary non-response (PNR) to IFX correlated with future CD-related surgery.

Conclusion:

Prior GI surgery, BMI, and platelet counts were related to future CD-related surgery. Patients who were PNR to IFX had a higher risk of CD-related surgery in the future.

INTRODUCTION

Crohn’s disease (CD) is a subtype of inflammatory bowel disease (IBD), which also includes ulcerative colitis (UC). CD is a transmural nonspecific inflammatory intestinal disorder with an unknown etiology.[1] It has different phenotypes or disease behaviors: non-stricturing and non-penetrating, stricturing, and penetrating.[2] Intestinal resection is one of the most common complications of CD. Contemporary epidemiological data show that the overall risk of surgery at 1, 5, and 10 years was 18.7%, 28.0%, and 39.5%, respectively.[3] Some patients with CD may need repeated gastrointestinal (GI) surgery.[3] Approximately one-third of patients undergo surgery because of stricturing, which is the most common indication for GI surgery.[4]

Treatment with biologics might reduce the occurrence of CD-related surgery in CD patients with stricture.[5] Approximately half of the patients with symptomatic small bowel stricture were free of surgery 4 years after anti-tumor necrosis factor (anti-TNF) treatment, including infliximab (IFX) and adalimumab, initiation.[6] Nevertheless, several patients still required CD-related intestinal resection after anti-TNF therapy.[7] Even combination therapy with two biologic medications had treatment failures, with a few patients needing surgery.[8] The loss of response to biologics is one of the tougher challenges in CD.[9] The strategies for decreasing the need for CD-related surgery are an ongoing area of research in the CD field.

The objective of this study was to assess the rate of occurrence of CD-related surgery and determine baseline risk factors predicting subsequent surgery for overall patients with stricturing CD: the higher hazard in intestinal resection.

PATIENTS AND METHODS

Patients

In this retrospective study, patients with CD[10] who registered as either an outpatient or inpatient, at the Department of Gastroenterology and the Department of General Surgery, Central Hospital of Shaoyang, between January 2014 and June 2022 were enrolled.

Inclusion and exclusion criteria

The inclusion criteria were as follows: 1) aged between 18 and 60 years; 2) initially diagnosed with stricturing disease behavior (B2 on the Montreal classification)[11]; and 3) longer than 3 months of follow-up. The exclusion criteria were as follows: 1) combined with penetrating CD; 2) accompanied by other autoimmune diseases; 3) accompanied by malignant tumors or CD-related colorectal cancers; and 4) females who were pregnant or breastfeeding.

Definitions

CD-related surgery: According to clinical practice guidelines and consensus of surgical managements in CD,[12,13] the operative indications were as follows: Medically refractory disease, stricture, penetrating disease, hemorrhage, and colorectal dysplasia and cancer.[13] There are no requirements for the anastomosis of the intestine and colon. Both laparoscopic surgery and open surgery patients could be included, while exploratory laparotomy and stoma closure without intestinal resection were excluded. Moreover, endoscopic and perianal surgery were excluded.

Primary non-response (PNR): PNR is defined as a lack of improvement in clinical symptoms after induction therapy.[9] However, there is no standardized definition.[9] We referred to a previous study,[14] in which PNR to IFX was defined as the Crohn’s Disease Activity Index (CDAI) dropping less than 70 points at week 14 after administration; having CD-related surgery before week 14 was also defined as PNR.

Follow-up

All participants were followed up until CD-related surgery occurred. Those who did not undergo surgery would be followed up until the end of the observation period. The last date of follow-up was in January 2023.

Data collection

Participants’ baseline and follow-up data were retrieved from the electronic medical record system. For patients who had ever been diagnosed and treated in another hospital, study physicians would contact the patients and acquire all clinical data by e-mail or telephone.

Observational endpoint

In this retrospective study, the need for surgery was considered an adverse event and defined as the observational endpoint in this study.

Statistical analysis

Continuous and categorical variables are described as medians with interquartile ranges and frequency with percentages, respectively. To identify risk factors, univariate and multivariate Cox regression analyses were used to assess the relationship between parameters and surgery-free survival. A P value of less than 0.05 was considered statistically significant. Statistical analysis was conducted using the Statistical Package for the Social Sciences (SPSS) version 26.0 [IBM Corp, Armonk, NY, USA], and GraphPad Prism 9.0.0 [GraphPad Software, San Diego, CA, USA] was used for statistical analyses and graphics.

Ethical considerations

The study was approved by the ethical committee of Central Hospital of Shaoyang (No. 20220718), dated July 18, 2022.

RESULTS

Characteristics of subjects

A total of 308 patients with CD were initially identified and reviewed for inclusion. A total of 139 patients were classified as non-stricturing and non-penetrating, and 36 patients were classified as penetrating or combined with penetrating. One patient with psoriasis, one with ankylosing spondylitis, four patients diagnosed with CD-related colorectal cancer, and four pregnant or breastfeeding women were excluded. Finally, 123 patients were eligible for analysis in our study [Figure 1]. Their median age was 25. The percentage of males was 75.6%. Ninety-seven (78.9%) of them were treated with IFX. Other basic clinical characteristics of the study participants are summarized in Table 1.

Figure 1.

Patient selection flowchart of the study. CD: Crohn’s disease

Table 1.

Baseline characteristics of 123 subjects

Variables	Median (range) or frequency (percentage)
Age (years)	25.00 (21.00–32.00)
Male (%)	93 (75.6%)
BMI	17.91 (15.74–19.53)
Smoking history (%)*	27 (22.0%)
Prior GI surgery (%)	33 (26.8%)
Treated with IFX (%)	97 (78.9%)
Locations (L)†
L1	1 (0.8%)
L2	7 (5.7%)
L3	36 (29.3%)
L4	37 (30.1%)
L1+L4	14 (11.4%)
L2+L4	9 (7.3%)
L3+L4	19 (15.4%)
Perianal disease (%)	54 (43.9%)
CDAI >220 (%)	57 (46.3%)
Laboratory examination
White blood cells (109/L)	6.60 (5.00–8.30)
Hemoglobin (1012/L)	118.00 (102.00–132.00)
Platelet counts (109/L)	339.00 (255.00–401.00)
Albumin (g/L)	33.50 (28.60–39.00)
Globulin (g/L)	28.40 (25.20–33.20)
C-reactive protein (mg/L)	17.05 (5.70–46.72)

*: smoking history included former and current. ^†: classifications of locations in CD were based on the Montreal classification. L1: terminal ileum; L2: colon; L3: ileocolonic; L4: upper gastrointestinal involvement. L4 can be added to L1, L2, and L3. BMI: body mass index; GI: gastrointestinal; IFX: infliximab; CDAI: Crohn’s Disease Activity Index

The rate of surgery in CD with stricture added up yearly

We followed up the participants and recorded the relevant clinical data and adverse events. Three subjects were excluded because of shorter than 3-month follow-up. There were three, 12, 16, 13, and 20 patients who were lost to follow-up at 12, 24, 36, 48, and 60 months, respectively. Perianal disease-related surgery, such as perianal fistula and perianal abscess, was not an observational endpoint in this study. The cumulative risk of CD-related surgery in 1, 2, 3, 4, and 5 years after diagnosis was 18.0%, 26.7%, 32.6%, 40.7%, and 46.4%, respectively [Figure 2].

Figure 2.

Cumulative risk of CD-related surgery at 1, 2, 3, 4, and 5 years after diagnosis. CD: Crohn’s disease

Body mass index (BMI), GI surgery history, and platelet counts might be risk factors for future CD-related surgery

To determine the risks of future GI surgery related to baseline clinical characteristics, univariate and multivariate Cox regression analyses were used. In the univariate analysis, BMI (hazard ratio (HR): 0.819, 95% CI: 0.721–0.929, P = 0.002), prior GI surgery (HR: 3.657, 95% CI: 1.779–7.514, P < 0.001), and platelet counts (HR: 1.006, 95% CI: 1.002–1.010, P = 0.002) were significantly related to future surgery. However, several indexes that we focus on in clinical practice, such as anti-TNF therapy and C-reactive protein level, did not significantly correlate with our observational endpoint. The multivariate Cox regression analysis showed that BMI (HR: 0.848, 95% CI: 0.724–0.992, P = 0.039), prior GI surgery (HR: 4.461, 95% CI: 2.010–9.899, P < 0.001), and platelet counts (HR: 1.006, 95% CI: 1.001–1.010, P = 0.015) were associated with future CD-related surgery [Table 2]. The survival curve shows the relationship of “future CD-related surgery-free” with prior GI surgery in the cohort of patients with CD [Figure 3a].

Table 2.

Cox regression analysis of factors associated with future CD-related surgery

Variables	Univariable analysis		Multivariable analysis
	HR (95% CI)	P	HR (95% CI)	P
All subjects (n=123)
BMI	0.819 (0.721–0.929)	0.002	0.848 (0.724–0.992)	0.039
Treated with IFX	0.486 (0.170–1.391)	0.179	0.638 (0.210–1.941)	0.429
Smoking history*	1.197 (0.510–2.809)	0.680	1.463 (0.560–3.825)	0.438
Perianal disease	1.398 (0.690–2.834)	0.352	1.286 (0.608–2.723)	0.510
Prior GI surgery	3.657 (1.779–7.514)	< 0.001	4.461 (2.010–9.899)	< 0.001
Disease severity	0.707 (0.347–1.443)	0.341	1.793 (0.722–4.455)	0.209
Platelet counts	1.006 (1.002–1.010)	0.002	1.006 (1.001–1.010)	0.015
C-reactive protein	1.005 (0.998–1.013)	0.181	1.006 (0.996–1.017)	0.239
Treated with IFX (n=97)
BMI	0.859 (0.743–0.994)	0.041	0.929 (0.789–1.093)	0.374
Smoking history*	1.534 (0.639–3.684)	0.338	1.392 (0.526–3.683)	0.506
Perianal disease	0.939 (0.440–2.007)	0.871	1.053 (0.471–2.355)	0.900
Prior GI surgery	3.489 (1.610–7.558)	0.002	2.967 (1.246–7.064)	0.014
Disease severity	0.741 (0.346–1.588)	0.441	1.276 (0.447–3.636)	0.649
PNR to IFX	4.153 (1.862–9.260)	0.001	2.690 (1.136–6.372)	0.025
Platelet counts	1.005 (1.000–1.009)	0.028	1.004 (0.999–1.009)	0.103
C-reactive protein	1.003 (0.995–1.011)	0.477	1.003 (0.991–1.014)	0.658

*: smoking history included former and current. BMI: body mass index; GI: gastrointestinal; PNR: primary non-response; IFX: infliximab

Figure 3.

Survival curve shows the relationship of “future CD-related surgery-free” (a) with prior GI surgery in all participants; (b) with prior GI surgery in IFX-treating participants; and (c) with PNR to IFX. GI: gastrointestinal; IFX: infliximab; PNR: primary non-response

GI surgery history and PNR were correlated with future CD-related surgery in those treated by IFX

A total of 97 patients were included in the further analysis. Of the 97 patients with CD who received IFX, 36 (37.11%) were recognized as PNR. No side effects of IFX were recorded. In the univariate analysis, BMI (HR: 0.859, 95% CI: 0.743–0.994, P = 0.041), prior GI surgery (HR: 3.489, 95% CI: 1.610–7.558, P = 0.002), PNR to IFX (HR: 4.153, 95% CI: 1.862–9.260, P = 0.001), and platelet counts (HR: 1.005, 95% CI: 1.000–1.009, P = 0.028) were included as candidate variables in the multivariate model. We also included several clinically significant variables such as smoking history, perianal disease, and disease severity which did not show statistical significance in the univariate analysis. The subsequent multivariate Cox regression model demonstrated that prior GI surgery (HR: 2.967, 95% CI: 1.246–7.064, P = 0.014) and PNR to IFX (HR: 2.690, 95% CI: 1.136–6.372, P = 0.025) were significantly related to future CD-related surgery [Table 2]. The survival curve analysis revealed a relationship between the future CD-related surgery-free group and prior GI surgery and primary response to IFX in the cohort [Figure 3b and 3c].

DISCUSSION

Intestinal resection is one of the most important outcomes of CD for physicians and patients. Our retrospective study found several risk factors for future CD-related surgery in high-risk CD patients with the stricture disease phenotype. BMI, GI surgery history, platelet counts, and PNR to IFX were key clinical features physicians need to be aware of.

The occurrence rate of CD-related surgery in patients with stricturing CD in our study was higher than in a previous meta-analysis covering 22 studies.[3] The small sample size of this study may have led to overestimation or underestimation of some observational endpoints. Further, the discrepancy might be related to the high number of patients lost to follow-up in our retrospective cohort study.

Platelet counts and BMI were related to future CD-related surgery, even though this observation was not found in subjects treated with IFX in the subsequent subgroup analysis. Previously, it had been thought that an increased platelet count indicated more active inflammation.[15] More severe disease activity and endoscopic appearance were suggested as predictors of a poor prognosis.[16,17] In addition, BMI was an indicator of nutritional status, which was more impaired in CD patients with stricturing disease.[18] A study[19] showed that malnutrition was associated with a higher rate of surgery in patients with CD. Moreover, BMI was associated with moderate to severe disability in CD.[20] Thus, among patients with stricturing CD, both platelet counts and BMI should be focused on. Regarding the non-significant statistical relation between BMI and platelet counts and future CD-related surgeries, it might be related to anti-TNF being administered based on weight and super anti-inflammation,[21] which would diminish the impact of the variables.

Another retrospective cohort study[22] also found that prior GI surgery and initial disease behavior with strictures might increase the risk of having an operation after diagnosis. Nevertheless, no further discussion was presented in that paper. Prior GI surgery was different from early surgery after diagnosis. In our study, all patients who had a history of surgery were diagnosed after the operation by histology or that the CD-related diagnostic basis was insufficient. It is likely that these patients initially had CD-related symptoms, such as symptomatic stricture. In other words, they had a delayed diagnosis, and diagnostic delay in CD is associated with a significantly higher need for future surgery.[23] This subset of participants progressed rapidly from onset to surgery, which is also a risk factor for future enterectomy.[24] Strictures in the patient population consist of a mix of inflammation and fibrosis.[25] Patients who do not respond to anti-fibrotic treatment within 12 weeks may develop irreversible fibrosis.[25] It suggests that the patients with a history of GI surgery already had refractory intestinal fibrous stenosis, which needed a further resection procedure.

The rate of PNR in our study was 37.1%, higher than reported by Kennedy NA et al. (23.8%).[26] The mechanism of PNR remains unclear. However, it has been suggested that patients with severe disease and systemic inflammatory burden may require higher doses of anti-TNF to achieve clinical efficacy.[27] Evidence for this was reported in a study[28] that showed fibrotic stenosis strongly correlated with circulating fibrocytes in stricturing CD, with increased levels of cytokines. Fibrostenotic disease indicates a more severe disease condition. Therefore, a standardized dose of anti-TNF would not achieve therapeutic effectiveness for patients with stricturing CD, resulting in PNR. The reason we should pay attention to PNR has already been shown in IBD: The long-term outcomes are reported to be related to patients’ response to therapeutic biologics.[29,30] A retrospective, multi-center study[31] found that about half of ulcerative colitis (UC) patients with PNR to anti-TNF would undergo colectomy. The situation might be worse in CD according to online surveys on treatment satisfaction and health-related quality of life in the United States of America.[32] In our study, we showed the correlation between PNR and a higher risk of CD-related surgery. It could help us to recognize patients with a poor prognosis sooner.

Our study had several limitations. First, it was a single-center, retrospective study with a small sample size, and thus, the results need to be confirmed in larger, multicenter studies. Second, the institution conducting this study was not a regional medical center, and many patients were being managed by tertiary referral centers. This resulted in heterogeneity of the clinical data. Finally, these results would benefit with confirmation in animal models which would help us understand and validate the pathophysiological basis of our findings.

In conclusion, this retrospective study demonstrated the high risks of CD-related surgery in stricturing CD. Prior GI surgery, lower BMI, and higher platelet counts were related to future CD-related surgery. Prior GI surgery and PNR were also indicators in patients treated with IFX. Nevertheless, the mechanism underlying the association between GI surgeries, PNR, and blood parameters in stricturing CD should be further investigated.

Financial support and sponsorship

This work was supported by the grants from the Innovation Guidance Project of Shaoyang Science and Technology Bureau (No. 2022NS4144).

Conflicts of interest

There are no conflicts of interest.