Mucosal healing is not associated with better outcome during seven years of follow-up in pediatric patients with Crohn's disease

Mucosal healing (MH) has become a perspective treatment target in patients with Crohn’s disease (CD). Data about the impact of MH on long-term outcome in pediatric patients are still scarce. Methods 76 pediatric patients with CD were evaluated retrospectively (2000–2015) in a tertiary care center. Based on MH achievement, they were divided into two groups (MH, n = 17 and No MH, n = 59). The primary endpoint was to assess the association of MH and the need for CD-related hospitalizations or surgery in pediatric patients with CD.


Results
The number of hospitalized patients was 24% in the MH group and 42% in the No MH group, P = 0.26.
The total number of CD-related hospitalizations was not signi cant between the MH group and the No MH group (5 vs Background It seems to be promising to accomplish mucosal healing (MH) in patients with Crohn's disease (CD).
However, the absence of a unique MH de nition can make the comparing of data from clinical trials more challenging. It is known that treatment based only on the disappearance of the symptoms is inadequate for the proper monitoring of the disease activity (1). It has been shown in adult patients with CD that MH is associated with a lower risk of surgical resection, lower hospitalization rate, decreased risk of relapse, and prolonged steroid-free remission (2)(3)(4)(5). However, some recent pediatric data indicate no bene t of MH on CD-related hospitalizations and surgery (6). More extensive pediatric data about MH and longterm outcome are still sparse. MH achieved with exclusive enteral nutrition (EEN) in children showed potential bene t for the relapse rate in comparison with corticosteroid (CS) treatment during a one-year follow-up (7). Complete endoscopic healing in children correlated with a decreased clinical relapse rate and lower intestinal resection during three years of follow-up (8). Our study aimed to assess the long-term outcome in pediatric patients with CD according to MH.

Design
The primary goal was to assess the rate and time interval of CD-related hospitalizations and operations after the follow-up (2nd ) colonoscopy. The rst colonoscopy was performed during the initial diagnostic workup at the time of diagnosis. Furthermore, the Pediatric CD Activity Index (PCDAI) was recorded. The second colonoscopy was indicated after the introduction of therapy within week 16 up to week 38 of new treatment according to clinical symptomatology. At the time of the 2nd colonoscopy, MH was assessed (with histology), and laboratory parameters of in ammation (C-reactive protein and erythrocyte sedimentation rate (ESR)) were recorded. According to MH achievement, the patients were divided into two groups: "MH" and "No MH." Patients Pediatric patients with CD from a single tertiary care center of the University Hospital in Hradec Králové were evaluated retrospectively (2000-2015) and followed till January 2019. CD diagnosis was made according to the Porto criteria (9). The phenotype of the CD was estimated by the pediatric modi cation of the Montreal classi cation (10). The upper gastrointestinal tract was examined by esophagogastroduodenoscopy, while magnetic resonance enterography was used to examine the small bowel. In ammatory bowel disease in a rst-degree relative was marked (Table 1). Perianal disease included anal ssure. PCDAI was used for the evaluation of the clinical and laboratory activity of the disease (11). The study follows the principles outlined in the Declaration of Helsinki. Informed consent was not obtained from the patients or parents because of the retrospective character of the study.

MH de nition
We de ned MH as a combination of total endoscopic and histological healing. MH was determined as complete endoscopic disappearance of all mucosal ulcerations (including aphthous ulcers) and the absence of any signs of mucosal in ammation (erythema, edema) in the terminal ileum and large bowel. A complete colonoscopy with terminal ileum intubation was performed on all MH patients. In 6 of the No MH patients, the terminal ileum was not intubated, and in 11 patients, a histology sample was not obtained. In all of these patients, endoscopic ndings showed apparent signs of in ammation. In most patients (and all MH patients), at least one histologic sample from the terminal ileum and one from the colon were taken. In the normal-appearing bowel segments, the biopsies were taken randomly. Biopsy specimens were stained with hematoxylin and eosin and read by experienced pathologists. Histologically, we distinguished low, medium, and high in ammatory activity based on the description by the pathologists. In the case of a normal endoscopic assessment but with histological signs of high in ammatory activity, the patient was classi ed as "No MH." Operations and hospitalizations CD-related surgical interventions were represented mainly by major abdominal surgery, then anal stula surgery, and anal divulsion. In some patients, a combined surgical procedure was performed. CD-related hospitalizations included only admissions for deterioration in disease activity. Hospitalizations for operation were assessed separately. Repetitive surgical procedures and hospitalizations were recorded during the whole follow-up period.

Therapy
Any therapy between the 1st and 2nd colonoscopy and immediately after the 2nd colonoscopy was registered. EEN given for at least four weeks was recorded. Treatment with 5-aminosalicylic acid (

Baseline characteristics
From 146 pediatric patients with CD, we excluded patients with missing data (n = 42), patients who had undergone an operation for CD before the 2nd colonoscopy (n = 24), patients who appeared to be noncompliant (n = 4), and those with con rmed infectious gastroenteritis at the time of colonoscopy (n = 1

Operations
Four patients (24%) from the MH group underwent CD-related surgery, which was not signi cantly different from the No MH group (23 patients, 39%) P = 0.39 (Table 2), as shown in the survival curve ( Fig. 2). In total, there were six operations (in 4 patients) in the MH group and 33 operations in 23 patients from the other group, P = 0.37 (Table 2). In some of the patients, the procedures were combined (Additional le 3: Table S3). All operated patients from the No MH group underwent a major operation. One patient from the MH group had only a perianal operation. The time to the rst operation did not achieve statistical signi cance between the groups (MH 43 months  and No MH 19 months [1.0-45], P = 0.13) ( Table 2).

Therapy
Between the 1st and 2nd colonoscopy, 9 (53%) patients in the MH group were treated with IFX in comparison with 8 (14%) in the No MH group, P = 0.002 (Additional le 1: Table S1). There was no statistical difference between the groups for any of the other therapies. In 10 (59%) patients, therapy with AZA or IFX was interrupted immediately after achievement of MH assessed by the 2nd colonoscopy. However, in most patients N = 4 (80%) initially treated with both IFX and AZA, either of the drugs was continued after the 2nd colonoscopy. IFX or adalimumab was introduced to 12 (20%) patients in the No MH group (Additional le 2: Table S2). Unhealed patients who were introduced to anti-TNF therapy did not require less CD-related hospitalization (P = 0.21) or operation (P = 0.33) than patients with no anti-TNF therapy (Table 3).

Discussion
In our study of pediatric patients with CD, we proved that MH at the time of the rst follow-up endoscopy is not associated with a more favorable outcome in the number of CD-related hospitalizations or operations during the next seven years of observation. In adult patients, a lower hospitalization rate was observed during eight years (96 months) of follow-up (3). An endoscopic substudy of a Step-up/Topdown trial on 49 pediatric patients did not nd a difference in the time to the rst CD-related hospitalization (P = 0.67) (6), which is in line with our results. It has been shown in adult patients with CD that MH is linked to a lower risk of surgical resection (3). We observed that 24% patients from the MH group and 39% from No MH group required a CD-related surgery during 7.5 resp. 6.6 years of observation, which was not signi cantly different. The time to the operation was also not statistically signi cantly different between the groups. This nding is consistent with data from the previously mentioned pediatric study from Hoekman et al. (6) (P = 0.5). We propose two possible explanations for the similarity of the results in the number of operated as well as hospitalized patients in our cohort. First, 59% of the patients in the MH group were weaned from AZA or IFX immediately after the 2nd colonoscopy after MH attainment, whereas in 80% of patients initially treated with both IFX and AZA, one of the drugs was continued after MH achievement. Furthermore, the effect of withdrawal of either AZA or IFX in patients with quiescent CD is still uncertain (12). This decision could also be justi able, especially in children, due to the possible adverse effects of combination therapy (13)(14)(15). Importantly, in 20% of patients from the No MH group, IFX or adalimumab was introduced after endoscopic proof of activity. In total, in 47% of patients, some treatment was launched (5-ASA, ADA, AZA, CS, EEN, IFX, mycophenolate mofetil) which could have had a positive in uence on the disease course and consequent prognosis. However, we did not prove that the launch of anti-TNF alone would have had a positive effect on hospitalizations or operations in unhealed patients. Paradoxically, we observed initially higher CD activity in patients who later attained MH (Table 1). Maybe the more aggressive therapy at the beginning can add to the higher rate of MH, as has been demonstrated in adults (16). There was greater use of IFX between the 1st and 2nd colonoscopy in the group that later reached MH. The premise that IFX is a potent drug in the induction of MH in patients with CD has been proven in children as well as in adults (17,18). Still debatable is a widely-adopted de nition of MH. Because of the retrospective character of the study, we did not use either of the indexes for the assessment of MH (SES -CD, Simple Endoscopic Score for Crohn's Disease, or CDEIS, Crohn's disease endoscopic index of severity). Nevertheless, MH, as de ned by us, corresponded with SES-CD = 0. This de nition is used frequently (5,19). We are aware of other limitations of our research. The 2nd colonoscopy was not scheduled, and so the time interval between colonoscopies is vast. The indication for the 2nd colonoscopy was not always traceable, and hence it was not possible to subcategorize the patients according to this. Some patients required colonoscopy because of clinical symptoms. In some patients drug reduction was intended in case of the favorable nding. In others, it was part of a scheduled checkup. Furthermore, the number of patients allocated to the MH group was quite low. At the time of MH assessment, we did not search for MH in the whole GIT tract but only in the colon and terminal ileum. We also would have appreciated having the values from the fecal calprotectin test. Finally, the approach to therapy has changed during the 16 years (2000-2015).
In conclusion, the use of IFX was associated with an increased probability of MH achievement in our group. We showed that patients with MH do not undergo fewer CD-related hospitalizations or operations, and there are indications that even later optimization of the therapy can still lead to a good clinical outcome. On the other hand, these results can also indicate the necessity for prolonged treatment, despite MH having been achieved. Ethics approval and consent to participate: The study follows the principles outlined in the Declaration of Helsinki. Informed consent was not obtained from the patients and parents because of the retrospective character of the study.

Consent for publication:
Not applicable.
Availability of data and materials: Survival curve. CD-related operations between MH and No MH groups