Characteristics and long-term outcomes of children with perianal Crohn’s disease

1. Ethical Statements

This study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of the National Center for Child Health and Development (approval number: 2023-077). The requirement for written informed consent was waived because of the retrospective design of the study.

2. Study Population

Until March 2023, we retrospectively reviewed the medical records of a cohort of patients younger than 18 years old who were diagnosed with CD based on the revised Porto criteria [12] between January 2010 and December 2020 at our institution (a tertiary children’s hospital in Japan). We included patients with PL, which were defined according to the Paris classification [2] and confirmed either at the time of diagnosis or later during the disease course. Patients initially thought to have CD but were diagnosed to have monogenic IBD were excluded from the study.

3. Study Design and Data Collection

The following parameters were extracted from the medical records: age at CD diagnosis, age at PL diagnosis, Paris classification (age at diagnosis, disease location, and disease behavior), pharmacological treatments for CD and PL (antibiotics, corticosteroids, immunomodulators, biologics, and small-molecule Janus kinase inhibitors [JAKi]), surgical treatments for PL (incision and drainage, seton placement, diversion surgery with ostomy placement, and proctocolectomy), and clinical course (PL recurrence and continuation of biologics). JAKi was counted as a kind of biologic because it was used in only 1 patient in this study.

Patients were classified according to age and year of diagnosis, respectively. We defined the very early-onset (VEO) group as those diagnosed with CD when they were < 6 years of age and the older age group as those diagnosed at ≥ 6 years of age. Those diagnosed in 2010–2015 were defined as the early diagnosis group, whereas those diagnosed in 2016–2020 were defined as the late diagnosis group. Classification by the year of diagnosis was established because ustekinumab, vedolizumab, risankizumab, and upadacitinib were approved for CD after 2017, and the refractory PL management strategy may have changed after that.

Descriptive statistics were expressed as mean with standard deviation, median with interquartile range, and proportions as appropriate. Categorical variables between the groups were compared using the chi-square test. The Fisher exact test was used to evaluate the variables in these groups while comparing the ratio of surgical interventions, as no patient diagnosed after 6 years of age underwent ostomy placement.

1. Patients

Between January 2010 and December 2020, 112 patients were newly diagnosed with CD at our institution. Of the 112 patients, 36 (32.1%) had PL during the disease course. All patients with PL were of Japanese ethnicity. Disease characteristics at the time of CD diagnosis are summarized in Table 1. Among the 36 patients, the median age at diagnosis and the follow-up period were 131 ± 28 and 70 ± 28.3 months, respectively. Seven patients (19.4%) were in the VEO group, and 29 (80.6%) were in the older age group. Our study included more males (63.9%) than females (36.1%). PL were confirmed in 30 patients at the time of CD diagnosis (83.3%) and in 6 patients during follow-up (16.7%). Regarding the disease location, 26 patients (72.2%) had L3, and 27 had L4b (75.0%). For small bowel involvement, represented by L1, L3, or L4b, 4 of the 7 patients in the VEO group had L2, indicating colonic disease without small bowel involvement. The remaining 3 had L3, with 2 also exhibiting L4b involvement. In the older age group, only 2 of the 29 patients lacked small bowel involvement. Three patients in this group had L1 with L4b involvement, while 23 patients had L3, of whom 22 also had L4b. One patient had an isolated L4b lesion in the small intestine. Two of the 36 patients in the VEO group had stenotic lesions (B2) at the time of CD diagnosis; however, none had penetrating bowel lesions (B3). Six of the 7 patients in the VEO group and 10 of the 29 patients in the older age group underwent genetic testing, with 8 patients undergoing whole exome sequencings and 8 undergoing panel sequencing (covering 17 known monogenic IBD genes). However, none of the patients was diagnosed with a known monogenic form of IBD.

Table 1.

Demographics of Children with CD Complicated with PL

2. Medical Treatments

Almost all of the patients with PL received biologics during the course of their disease (34/36, 94.4%). More than half had received more than 2 biologics in the VEO-CD group (57.1%) and the older age group (55.2%) owing to intolerance or primary or secondary failure. Two patients with refractory disease received 4 or 5 different biologics (Figs. 1 and 2).

Fig. 1.

Biologics used by different age groups. The entire very early onset (VEO) group and 93.1% of the older age group have received biologics, with more than half of them had received more than 2 biologics.

Fig. 2.

The number of biologics used in each patient (n=36). Single and 2 biologics were used in more than 70% of the patients. One patient had received 5 biologics.

Infliximab and adalimumab were given to 23 (67.7%) and 20 (58.8%) patients, respectively, covering all patients who had ever been treated with biologics. Ustekinumab was given to 15 patients (44.1%) for secondary failure or intolerance to anti-TNF-α drugs. Risankizumab and vedolizumab were administered to 4 (11.8%) and 2 patients (5.9%), respectively. An oral JAKi drug, upadacitinib, was administered to 1 patient.

At the end of the follow-up period, the PL were well controlled with continued biologic therapies in most cases. Ten patients continued with ustekinumab, 9 with infliximab, 9 with adalimumab, 3 with risankizumab, 1 with vedolizumab, and 1 with upadacitinib. Infliximab was discontinued after 3 doses in 1 patient because of continuous remission.

Thiopurines and 5-aminosalicylic acid were given to 27 (75.0%) and 35 (97.2%) patients, respectively. Corticosteroids were administered to 13 patients, representing 36.1% of our study population. All patients, except for 1 patient who refused to take it, received an elemental formula during their disease course.

Corticosteroids were given to more patients in the early diagnosis group (56.3% vs. 20.0%, P< 0.05). More patients in the late diagnosis group were given biologics within 1 year of diagnosis (89.5% vs. 40.0%, P< 0.05). Ustekinumab was given to more patients in the early diagnosis group (50.0% vs. 10.0%, P< 0.05) (Table 2).

Table 2.

Shift in Treatment by the Year of Diagnosis

Six of the 7 patients in the VEO group (85.7%) and 24 of the 29 patients in the older age group (82.8%) had PL at the time of CD diagnosis. There were no significant differences between the VEO and older age groups regarding corticosteroid or biologic utilization. The ratio of patients introduced to biologics within 1 year of diagnosis was higher in the older age group; however, no statistical significance was observed (42.9% vs. 69.0%, P= 0.23) (Table 3). Regarding the continued use of biologics, there was no significant difference between the VEO and older age groups.

Table 3.

Treatment Outcome of Perianal Disease by Age Group

3. Surgical Intervention

Five patients (13.9%) did not undergo examination under anesthesia, while 21 of the 31 patients who underwent examination under anesthesia required either incision with drainage or seton placement. Over the course of the disease, 28 patients (77.8%) underwent surgical intervention for PL. The procedures ranged from incision with drainage and seton placement to more aggressive interventions such as diversion with ostomy placement or proctocolectomy. Among these 28 patients, 2 of the 4 patients in the VEO group and 18 of the 24 patients in the older age group underwent surgical intervention before the use of biologics. Additionally, 7 of the 12 patients diagnosed before 2016 and 13 of the 16 patients diagnosed in 2016 or after had surgical intervention before the introduction of biologics.

Seton placement was performed in 24 patients (66.7%). Seventeen patients (70.8%) had the seton removed or detached by the end of the follow-up period. Repeated seton placement was required by 2 patients because of recurrent PL. Three of the 7 patients diagnosed with VEO-CD underwent diversion surgery with ostomy placement (sigmoid colon, transverse colon, and terminal ileum). One patient eventually required proctocolectomy and a permanent stoma at the sigmoid colon level. None of the 29 patients in the older age group required bowel surgery.

Twelve patients in the early diagnosis group (75.0%) and 16 in the late diagnosis group (80.0%) had PL at the time of CD diagnosis. Ten patients in the early diagnosis group (62.5%) and 14 in the late diagnosis group (70%) underwent seton placement. The rate of total surgical intervention did not differ significantly between both groups (75.0% vs. 80.0%, P= 0.72) (Table 2). Five of the 7 patients with VEO-CD (71.4%) were in the early diagnosis group, as well as 3 patients who underwent diversion surgery with ostomy placement with/without proctocolectomy.

With the further comparison of the VEO and older age groups, more patients in the older age group required surgical intervention, mainly seton placement (57.1% vs. 82.7%, P=0.14). Upon closer examination of the surgical interventions, we observed that more patients in the older age group underwent less invasive procedures, such as seton placement or incision with drainage (14.3% vs. 82.8%, P< 0.05), whereas patients in the VEO group who required surgical intervention eventually underwent more invasive procedures, such as diversion surgery with ostomy placement with/without proctocolectomy (42.9% vs. 0.0%, P< 0.05).

1. Biologics Play a Key Role in Pediatric CD Patients with PL

Recent medical management guidelines for CD recommend the early introduction of anti-TNF-α drugs for patients with poor outcome predictors, which include perianal disease, stricturing or penetrating disease, extensive disease, deep colonic ulcers, or growth delay [13,15].

Biologics have become a standard treatment for CD since their introduction more than 20 years ago. In the management of pediatric CD patients with PL, recent studies have suggested a clear trend towards the early introduction and increased utilization of biologic therapies, which has shown promising results [6,15]. Various studies have highlighted positive treatment outcomes associated with biologic therapies for PL in pediatric patients with CD [16,17], including higher rates of fistula closure, reduced need for surgical interventions, and enhanced quality of life [17,18]. In our study, 94.4% of patients were given biologics, and in the late diagnosis group (those diagnosed in 2016–2020), up to 90% of patients with PL initiated biologics within 1 year. Biologics have been prescribed for children with VEO-CD who are more vulnerable to the adverse effects of corticosteroids and tend to have more aggressive disease courses that require immediate disease control [19,20].

2. Types of Biologics

Among the biologics used in treating pediatric CD patients with PL, infliximab and adalimumab have demonstrated good efficacy [17,18,21]. Ruemmele et al. [17] highlighted the significance of adalimumab in achieving fistula closure and improving disease outcomes in the pediatric population. The efficacy of infliximab in inducing and maintaining remission in pediatric CD patients with PL has been evaluated [13,22,23].

The role of vedolizumab in managing IBD, including in adult CD patients with PL, has been investigated [7]. Atia et al. [24] presented data on its effectiveness and dosing considerations for pediatric patients with CD but with little evidence on PL. Two of our patients received vedolizumab, and 1 had it discontinued due to worsening scalp folliculitis.

Ustekinumab has been used as a superior therapeutic option in adults with CD, with or without PL [7,25,26]. It has also been introduced to pediatric patients [27].

Risankizumab effectively induced remission in moderate-to-severe PL in CD [28]; however, more research is needed to establish its specific role in pediatric patients with PL. Four of our 36 CD patients with PL received risankizumab to control the CD, and not primarily for the PL; nevertheless, 3 of them achieved better control of the CD without worsening the PL.

Upadacitinib, used for atopic dermatitis in children ≥ 12 years old, was approved for adult ulcerative colitis and CD. It is occasionally introduced to pediatric patients with refractory disease; however, it still requires further investigation [29,30]. Symptomatic control of CD with upadacitinib was achieved in one of our patients, who experienced intolerance to most previously administered biologics, and PL was controlled. There are limited published data on the effects of JAKi on PL [31]. Overall, the choice of biologics for pediatric CD patients with PL should be based on individual patient factors, disease severity, and response to previous treatments, while also considering the available evidence regarding the efficacy and safety profile of each biologic agent. The availability of newer biologics with distinct mechanisms of action offers additional therapeutic options and the potential for treatment approaches tailored to individual patient characteristics and preferences.

Furthermore, darvadstrocel, a suspension of allogenic adipose-derived mesenchymal stem cells developed as a one-time local injection for the treatment of complex perianal fistulas in adults with CD, has been reported to be safe and effective [32-34]. However, there are few reports available to date on its use in pediatric patients. Close monitoring and multidisciplinary management are essential to optimize treatment outcomes in this challenging patient population.

3. VEO-CD

Patients with VEO-CD exhibit distinct characteristics compared to those with older-onset pediatric CD [35,36]. The disease behavior in patients with VEO-CD is usually more inflammatory and non-stricturing than that in the older-onset group and often presents with a more isolated colonic disease [37]. In contrast, the disease course in VEO-CD could be more aggressive with greater perianal involvement, eventually causing more severe growth retardation [37]. Patients with VEO-CD may require aggressive medical therapy, including early initiation of biologics such as anti-TNF-α or IL-12/23 [35]. Despite aggressive medical management, patients with VEO-CD may require invasive surgical intervention, usually at a higher rate than patients with older-onset pediatric CD [35,38]. Due to refractory responses to medical therapy or disease complications, studies have reported a higher likelihood of patients with VEO-CD requiring early surgical intervention, including bowel resection, stoma formation, or perianal surgeries [38].

As the majority of the patients with VEO-CD received biologics, with over half being treated with corticosteroids, the requirement for surgical intervention may be invasive, such as ostomy or colectomy. The fact that our patients with VEO-CD requiring ostomy placement with/without proctocolectomy came from the early diagnosis group implies that earlier initiation of biologics with additional choices if it fails might lessen the need for invasive surgical management, such as ostomy placement.

Bor et al. [39] demonstrated the efficacy of combined anti-TNF-α and surgical therapy in adults, such as seton placement, in treating perianal and enterocutaneous fistulizing CD. Half of the patients achieved complete fistula closure within 1 year, and its success has been recognized in the pediatric population [22,40]. Our results are consistent with these findings. Ashton et al. [23] noted that an increased utilization of anti-TNF therapy was associated with a decline in surgical resection in pediatric patients with IBD. This trend suggests the potential impact of medical therapy in reducing the need for surgical intervention.

4. Limitation and Conclusion

The limitations of our study include its retrospective nature, limited patient enrollment from a single center, the lack of therapeutic drug monitoring, and changes in evolving treatment strategies, especially with the introduction of newly approved biologics. However, these management changes have improved the outcomes of pediatric CD patients with PL. Although we defined PL according to the Paris classification, further details of PL were not described because of the study’s retrospective nature and the non-standardized description of PL. Additionally, this study did not compare pediatric CD patients with PL to that without PL, as it was beyond the scope of the study. Further research involving data from a larger cohort of pediatric-onset CD with and without PL would help clarify the factors that influence the outcome of PL.

The early use of biologics and appropriate surgical intervention could improve the outcome of children with CD and PL. The use of newly approved biologics offers the option of controlling PL in cases refractory to anti-TNF-α drugs. VEO-CD complicated by PL should be considered a high-risk group for invasive surgical procedures.