Introduction

The rate of surgery for Crohn’s disease (CD) is high, with reported cumulative surgery rates of 16.3% at 1 year, 33.3% at 5 years, and 46.3% at 10 years following diagnosis, respectively [1]. Furthermore, many cases require repeat surgery, and the cumulative repeat surgery rates are as high as 24.2 and 35.0% at 5 and 10 years following their first surgery, respectively [2]. We have also reported that 48.7% of patients require first surgery at 5 years following diagnosis, while 36.0% require repeat surgery at 5 years after their first surgery [3]. In order to optimize postoperative maintenance therapy depending on the activity of each case, endoscopic recurrence preceding clinical recurrence has been recognized to be important [4, 5]. Endoscopic recurrence rates in patients without symptoms have been reported to be as high as 70–90% and 85–100% at 1 and 3 years following surgery, respectively [6]. Recently, in order to utilize endoscopic recurrence for improving postoperative treatments, De Cruz et al. demonstrated in a multicenter randomized trial that it is useful to intensify treatments according to the endoscopic findings at 6 months following surgery [7]. The prospective detection of endoscopic recurrence may enable us to reconsider treatment regimens prior to clinical recurrence, thereby improving the long-term prognosis [4, 5].

However, we have some concern about the aforementioned study [7]. First, only cases following ileocecal resection were included and only colonoscopy was performed. The rate of small bowel lesions not found on ileocolonoscopy was quite high [8, 9]. Second, it is difficult to distinguish endoscopic recurrence from residual lesions. We often find residual lesions (mostly minor lesions) even after surgery; however, we have little knowledge about how high the rate of residual lesions is and whether these lesions have an influence on postoperative course or not. Therefore, prior to the evaluation of postoperative endoscopic recurrence, it is necessary to perform an accurate assessment of residual lesions in small bowel immediately after surgery.

In recent years, the usefulness of capsule endoscopy (CE) has been reported in the diagnosis of small bowel diseases. Since CE is less invasive and less painful than small bowel follow-through and balloon-assisted small bowel enteroscopy, and CE has a higher detection rate of lesions compared to CT enterography (CTE) and MR enterography (MRE) [10,11,12,13], it is considered to be an important modality in the diagnosis of CD. However, there have so far been few studies regarding the role of CE for CD following surgery. Thus, in order to assess any residual small bowel lesions accurately, we performed capsule endoscopy (CE) immediately after surgery in patients with CD. We also investigated the relationship between such residual lesions and postoperative clinical recurrence prospectively.

Materials and Methods

Subjects

We included patients with CD who had undergone surgery for remission induction at Tohoku University Hospital between December 2013 and February 2016. CE was performed immediately following surgery (< 3 months after surgery), and the patients’ clinical courses were followed prospectively. CD was diagnosed based on the diagnostic criteria from the Research Group of Intractable Inflammatory Bowel Disease organized by the Ministry of Health, Labour, and Welfare of Japan [14]. Regarding surgical procedures, only patients who had undergone intestinal resections were included, excluding those who had only undergone surgery for anal lesions. Patients who had undergone colostomy or ileostomy were also included. As the disease location at diagnosis may change over time, various types of disease location were included.

The present study was approved by the ethics committee of our hospital on September 30, 2013. Written informed consent was obtained in all cases. All CEs in the present study were performed on an outpatient basis, and the examinations were covered by the patients’ health insurance.

CE Procedure

In the present study, CE was performed using either PillCam® SB2 or SB3 (Given Imaging, Covidien Ltd., Yokneam, Israel). Prior to CE procedure, patency was assessed with a patency capsule (PC) being excreted in its original shape within 33 h after swallowing. In patients who were judged to have no patency, other gastrointestinal examinations were performed.

Regarding a bowel preparation for CE, 50 mg of magnesium citrate and 150 mg of sodium picosulfate hydrate were orally administered at 14 and 13 h before CE, respectively. A high concentration of polyethylene glycol was also administered 1 h before CE. Two and 4 h after swallowing CE, the patients were allowed to drink water and to eat, respectively. In cases in which the capsule reached colon within 8 h after swallowing, the examination was ended. If not, the examination was extended to 10 h.

Assessment of Residual Lesions

The endoscopic activity immediately after surgery was assessed by two endoscopists using the Lewis score (LS) which was reported by Gralnek et al. [15]. The small intestine is divided into three parts from the oral side: first tertile, second tertile, and third tertile. Then, the tertile score is calculated as the sum of the inflammatory parameter scores (villous appearance and ulcers) in each tertile. On the other hand, the stenosis score is calculated in whole small intestine (not in each tertile). The LS is the sum of the highest tertile score and the stenosis score. The LS classifies the endoscopic severity into three grades: normal or physiological inflammation (< 135), mild inflammation (≥ 135 and < 790), and moderate-to-severe inflammation (≥ 790).

Primary and Secondary Outcomes

The primary outcome was the endoscopic activity immediately after surgery (i.e., residual lesions) using the LS. The secondary outcomes included the relationship between clinical characteristics and residual lesions, and that between residual lesions and postoperative clinical recurrence. Postoperative clinical recurrence was defined as repeat surgery, hospitalization associated with CD or changes in the medical treatments. In order to include patients with ileostomy or colostomy, we did not use the Crohn’s Disease Activity Index.

Statistical Analysis

Quantitative data are shown as the mean value and standard deviation. For comparisons between groups, t test, Mann–Whitney U test, or correlation analysis was used as appropriate. The cumulative clinical recurrence rate was calculated using the Kaplan–Meier method, and comparisons between groups were performed using the log-rank test. These analyses were all performed using the JMP Ver. 11 software program (SAS Institute Inc., Cary, NC, USA). A P value of less than 0.05 was considered to be statistically significant.

Results

Clinical Characteristics and Medical Treatments

There were 46 patients with CD who underwent surgery during the study period. CE was attempted in 27 of these patients, excluding any patients who changed hospitals following surgery, patients who did not provide their consent to undergo CE, and patients in whom more than 3 or more months had passed following surgery.

The clinical characteristics of 27 patients are shown in Table 1. These patients included 24 males (88.9%) and 3 females (11.1%), with a mean age at diagnosis of 24.7 (SD, 8.0) years, age at surgery of 35.6 (8.5) years, and disease duration at surgery of 12.0 (8.5) years. The number of previous surgeries was 2.1 (1-7), and 14 patients (51.9%) had no history of previous surgery for CD. The disease type included 9 ileitis type (33.3%), 17 ileocolitis type (63.0%), and 1 colitis type (3.7%), while the disease behavior included 2 penetrating type (7.4%) and 25 non-penetrating type (92.6%). There were 17 patients (63.0%) with anal lesions. Six patients (22.2%) had a smoking history, while 21 (77.8%) did not. The surgical procedures included 10 partial resections of the small intestine (37.0%), 8 ileocecal resections (29.6%), 6 anastomotic resections (22.2%), 3 colostomies (11.1%), 3 strictureplasties (11.1%), 2 partial colectomies (7.4%), and 2 subtotal colectomies (7.4%) with some overlapping.

Table 1 Clinical characteristics
Full size table

Following surgery, same medical treatments were continued in all cases except for two patients in whom anti-tumor necrosis factor agent or immunomodulator was newly added. There were no significant differences in clinical characteristics or medical treatments between the 27 patients who underwent CE and other 19 patients who were excluded.

CE Procedure

Twenty-five of 27 patients (92.6%) were diagnosed to have patency using a PC, and CE was performed at 1.8 (1.1) months following surgery. Of 25 patients in whom CE was performed, we could observe the entire small intestine in 20 patients (80.0%), and the average observation time of the small intestine was 212 (103) minutes. No adverse events such as aspiration or abdominal pain were observed. In 24 patients, the capsule endoscope was excreted within 2 weeks. Although the excretion of the capsule endoscope was not confirmed within 2 weeks in one case, an X-ray examination at 2 weeks showed that it had already been excreted.

In two patients who were diagnosed to have no patency, narrowing or stenosis of the intestinal anastomosis was found by other examinations. Since both patients had no symptoms, they preferred to continue monitoring without any intervention. There has been no change in their symptoms or blood tests at 23 and 13 months following CE, respectively.

Primary Outcome

Regarding villous edema of which the tertile score was composed, the scores were 5.8 (21.9), 9.6 (30.3), and 6.7 (21.8) in the first, second, and third tertile, respectively. Regarding ulcers of which the tertile score was composed, the scores were 176.4 (356.6), 160.2 (306.6), and 177.6 (288.9) in the first, second, and third tertile, respectively. As the sum of the parameter scores (villous edema and ulcers), the tertile scores in the first, second, and third tertile were 182.2 (371.6), 169.8 (313.1), and 184.3 (290.1), respectively; no significant difference was found among the three groups. The stenosis score was 182.2 (371.6). Using the highest tertile score and the stenosis score, the LS was 751.3 (984.0). According to the LS, 16.0, 56.0, and 28.0% were classified as normal, mild, and moderate-to-severe inflammation, respectively. The details of the scores are shown in Table 2.

Table 2 Lewis scores of all cases
Full size table

Prior to surgery, we performed CT in 25 patients (100%), small bowel follow-through in 22 patients (88%), colonoscopy in 12 patients (48%), and colonoscopy followed by retrograde gastrointestinal series in 11 patients (44%) with some overlapping. However, the recognition rate of lesions other than those resected by surgery was 0%. During surgery, residual lesions were reviewed based on the findings of a serosal side view; however, the recognition rate was only 16.0% (4/25).

Secondary Outcomes

Relationship Between Clinical Characteristics and Residual Lesions

Among the clinical characteristics, patients with anal lesions had significantly higher LS (P = 0.024) (Fig. 1). There were no significant differences for the LS concerning age at diagnosis, age at surgery, disease duration at surgery, number of previous surgeries, disease type, disease behavior, extraintestinal manifestations, or smoking history. Regarding medical treatments, no significant difference for the LS was observed.

Fig. 1
figure 1

Lewis scores of patients with and without anal lesions. Patients with anal lesions had significantly higher Lewis score than those without anal lesions (P = 0.024)

Full size image

Relationship Between Residual Lesions and Postoperative Clinical Recurrence

Postoperative clinical recurrence was observed in 5 of 25 patients (Nos. 2, 9, 13, 15, and 16 in Table 2), and the time from surgery to clinical recurrence was 9.8 (4.5) months. Regarding the cumulative clinical recurrence rate, no significant differences were observed in clinical characteristics and medical treatments. When the cumulative clinical recurrence rate was compared among the endoscopic severities classified by the LS (normal, mild, and moderate-to-severe), again no significant difference was found. However, when comparing groups divided according to the highest tertile score, the cumulative clinical recurrence rate was significantly higher in the group with the highest third tertile score (P = 0.046) (Fig. 2). Furthermore, patients with ulcers in the third tertile had a significantly higher clinical recurrence rate in comparison with the others (P = 0.045) (Fig. 3).

Fig. 2
figure 2

Cumulative recurrence rates comparing groups divided according to the highest tertile score. When comparing groups divided according to the highest tertile score, the cumulative recurrence rate was significantly higher in the patients with the highest third tertile score (P = 0.046)

Full size image
Fig. 3
figure 3

Cumulative recurrence rates comparing groups divided according to the presence of ulcers in the third tertile. Patients with ulcers in the third tertile had a significantly higher recurrence rate in comparison with the others (P = 0.045)

Full size image

Discussion

Since CD often requires repeat surgery, the prevention of postoperative recurrence is an important issue. In order to prevent postoperative clinical recurrence, it is useful to detect endoscopic recurrence preceding clinical symptoms. However, even if active lesions are detected endoscopically in postoperative follow-up study, it cannot be determined with certainty that those lesions actually represent endoscopic recurrence. This is because a certain number of lesions may remain even after surgery. Therefore, the present pilot study was conducted in order to accurately assess the presence of any residual lesions immediately after surgery and to elucidate the relationship between these residual lesions and postoperative clinical recurrence prospectively.

In spite of having undergone surgery, 84% with CD had endoscopic activity immediately after surgery, and the average endoscopic activity was almost moderate. Such a high prevalence of residual small bowel lesions using perioperative endoscopy has been reported in some previous studies [16, 17]. Kono et al. also reported that 78.0% of the subjects had endoscopic activity on very early postoperative endoscopy [18]. These residual lesions probably have some influence on the findings of follow-up endoscopy after surgery. On the other hand, most of these residual lesions could not be detected by preoperative examinations. Preoperative gastrointestinal examinations are limited due to the presence of stenosis or fistula, and other minor lesions are thus often missed. For example, the detection rate of lesions on small bowel follow-through, which is often performed in cases with stenosis or fistula, is very low compared to CE [19]. The detection rate of minor lesions on CTE and MRE is also low. It is difficult to perform CE preceding surgery due to the risk of capsule retention, while residual lesions could be non-invasively assessed on CE immediately after surgery without any stenosis or fistulas.

Among the clinical characteristics, only anal lesions were significantly associated with higher endoscopic activity immediately after surgery. Anal lesions have been reported to be a risk factor for recurrence following surgery [4, 5]. Patients with some risk factors are likely to already have a high endoscopic activity (i.e., many residual lesions) immediately after surgery. These patients are considered to require careful observation following surgery.

A high rate of residual lesions may lead to overestimation of the risk of postoperative clinical recurrence [17]. However, there are so far no available data regarding the course of residual lesions. Thus, we also followed clinical courses prospectively. As a result, postoperative clinical recurrence was observed in 20% at about 10 months following surgery. Clinical characteristics and postoperative medical treatments did not affect clinical recurrence. There was also no significant association between the endoscopic severity as classified by the LS and postoperative clinical recurrence. However, when grouped according to the highest tertile score, the cumulative clinical recurrence rate was significantly higher in the group with the highest third tertile score. In addition, when grouped according to the presence of the parameters (villous edema and ulcers) which compose the tertile score, the cumulative recurrence rate was significantly higher in the group with ulcers in the third tertile. Residual lesions, especially in the distal small intestine, were associated with clinical recurrence. The location as well as the presence of residual lesions immediately after surgery was associated with postoperative clinical recurrence. Of course, CE is accompanied by some risk of retention and expensive cost in follow-up study. Thus, ileocolonoscopy or close monitoring of biomarkers (treat to target approach) may be a better alternative [20].

We acknowledge that there are several limitations in the present study. First, the small number of patients and the short observation period might have influenced the aforementioned outcomes. Second, we did not show any data about repeated CEs. Up to now, there are no available data regarding the optimal timing to detect endoscopic recurrence. Therefore, we conducted the present study as the basic data to detect recurrent lesions. With a larger number of patients and a longer observation period, further studies performing CE repeatedly are required. Third, we could not show what the inflammation in the first or second tertile is associated with. Some factors such as non-steroidal anti-inflammatory drugs (NSAIDs) or proton pump inhibitors (PPIs) may have been related to the inflammation in these sites. NSAIDs are known to have a strong potential to induce inflammation in the upper small intestine, especially when used with PPIs [21, 22]. However, there was no significant association between the endoscopic activity in the upper small intestine and concurrent use of NSAIDs and/or PPIs in the present study (data not shown).

In conclusion, many cases with CD have endoscopic activity despite having undergone surgery. These residual lesions, especially in the distal small intestine, were associated with postoperative clinical recurrence. Optimizing medical treatments based on the assessment of both residual lesions and recurrent lesions may improve the long-term prognosis. With a larger number of patients and a longer observation period, further studies are required.