Background
Over the past two decades, the global incidence of pediatric inflammatory bowel disease (IBD) has steadily increased, making it an increasingly important public health concern [1]. In China, the incidence of Crohn’s disease (CD) in children is increasing annually. Compared to that of ulcerative colitis, pediatric CD more frequently involves extensive gastrointestinal tract involvement and tends to exhibit a more severe and rapidly progressive course [2].
Perianal disease (PD) is a distinct and aggressive phenotype associated with considerable morbidity and disability, with manifestations such as perianal skin tags, anal fissures, perianal abscesses, and fistulas. The reported prevalence of PD among adult patients with CD ranges from 8% to 62% [3], whereas that among pediatric patients with CD ranges from 13.6% to 62%, with Asian populations potentially exhibiting even higher rates [4, 5]. Very early-onset IBD (VEO-IBD), typically defined as a disease diagnosed before 6 years of age, often exhibits unique clinical features and is more likely to be attributable to monogenic etiologies [6], with notable differences from CD in older children in terms of disease phenotype, treatment response, and prognosis.
To date, data regarding the epidemiology, clinical phenotypic characteristics, and prognosis of pediatric CD complicated by PD in China remain sparse, and existing findings are inconsistent, particularly regarding systematic comparisons of perianal manifestations between older children and those with VEO-IBD. Therefore, the present study aimed to retrospectively analyze clinical data from pediatric inpatients diagnosed with CD at our center over the past decade, with the goal of systematically characterizing the clinical phenotypes, treatment responses, and outcomes of patients with concurrent PD, thereby providing evidence-based insights to facilitate early identification and inform precision management in clinical practice.
Methods
Study population
Data from pediatric patients diagnosed with CD who were hospitalized at the Capital Center for Children’s Health, Capital Medical University (Beijing, China) between January 2015 and December 2024 were retrospectively collected and analyzed. The inclusion criteria were as follows: age at diagnosis between 0 and 18 years, with a confirmed diagnosis of CD based on the 2014 European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (i.e., “ESPGHAN”) diagnostic guidelines, and the 2019 Chinese Expert Consensus on the Diagnosis and Treatment of Pediatric Inflammatory Bowel Disease; and a follow-up duration ≥ 12 months. The exclusion criteria were as follows: revised diagnosis during follow-up, incomplete clinical data, coexisting severe infections, tuberculosis, lymphoma, autoimmune diseases, or other significant comorbidities during the disease course.
PD was classified according to the 2003 American Gastroenterological Association (AGA) criteria as follows [7]: perianal skin lesions (skin tags and hemorrhoids), anal canal lesions (anal fissures, anal ulcers, and anorectal strictures), perianal abscesses and fistulas, rectovaginal fistulas, and malignancy. The patients were divided into PD and non-PD groups based on the presence of perianal involvement. To further explore the distribution of perianal manifestations across different age-at-onset subgroups, all patients were stratified according to age at diagnosis into two groups: very-early onset (VEO [diagnosed at < 6 years of age]) and late-onset (LO [diagnosed at 6–18 years of age]).
The clinical characteristics were compared between patients with and without PD within each subgroup.
At our center, the treatment algorithm for pediatric CD follows the ESPGHAN/European Crohn’s and Colitis Organization (ECCO) guidelines. For patients with mild to moderate disease, exclusive enteral nutrition (EEN) and/or corticosteroids are used as induction therapy, followed by immunomodulators (azathioprine or methotrexate). Anti-TNF therapy (infliximab) is recommended as the first-line biological treatment for patients with moderate-to-severe disease or complex perianal disease.
The standard induction regimen for infliximab was 5–10 mg/kg at weeks 0, 2, and 6, followed by maintenance every 8 weeks. Therapeutic drug monitoring (TDM) was performed in patients with a loss of response or suspected adverse reactions, although routine proactive TDM was not uniformly implemented during the study period because of its limited availability in China in earlier years. This limitation has been acknowledged in the discussion section.
All patients underwent a complete colonoscopy. Esophagogastroduodenoscopy (EGD) was performed in all LO patients and in 34 of 44 VEO-IBD patients (77.3%); the remaining VEO-IBD patients did not undergo EGD due to neonatal-onset disease or a very young age at diagnosis, which made general anesthesia or endoscopic procedures technically difficult or clinically inadvisable.
This study was approved by the Ethics Committee of the Capital Center for Children’s Health, Capital Medical University (approval number: SHERLL2025097).
Data collection and clinical parameters
Data collection
Clinical data were collected for each patient, including age, sex, diagnostic delay, laboratory parameters, Paris classification, Pediatric Crohn’s Activity Index (PCDAI), Simplified Endoscopic Activity Score for CD (SES-CD), perianal magnetic resonance imaging (MRI) findings, treatment regimens, and follow-up outcomes. Anthropometric measurements at diagnosis, including height, weight, and body mass index, were converted to Z-scores using the World Health Organization child growth standards adjusted for age and sex.
The following data were extracted from the electronic medical records during follow-up: disease phenotype, genetic findings, pharmacological treatments, surgical interventions, and clinical outcomes. Genetic diagnosis of the pathogenic variants was confirmed using targeted genome sequencing or whole-exome sequencing, and variant interpretation was performed according to the American College of Medical Genetics and Genomics/Association for Molecular Pathology guidelines. Data were collected from the first hospitalization to December 2024 (the data cutoff date); for patients lost to follow-up or deceased before this date, the date of last contact or death was used as the censoring point.
Clinical phenotype classification: Clinical phenotypes were characterized according to the Paris classification for pediatric CD, which encompasses four domains: age at diagnosis, disease location, disease behavior, and growth status. Specifically, age at diagnosis was classified as A1a (0 to < 10 years) or A1b (10 to < 17 years); disease location as L1 (distal one-third of ileum ± limited cecal involvement), L2 (colonic), L3 (ileocolonic), or L4 (upper gastrointestinal tract, further subdivided into L4a as proximal to the ligament of Treitz [esophagus, stomach, duodenum] and L4b [distal to the ligament of Treitz, limited to the proximal jejunum]); disease behavior as B1 (non-stricturing, non-penetrating), B2 (stricturing), B3 (penetrating), or B2 + B3 (stricturing and penetrating); growth status as G0 (no growth delay) or G1 (growth delay); and perianal fistula involvement as P0 (no perianal fistula) or P1 (perianal fistula).
Definitions of key outcome measures
All outcome measures were assessed at the last follow-up visit (December, 2024). For the Kaplan–Meier analysis, the time cutoff was set at 52 weeks (1 year) after diagnosis to permit a standardized comparison of event occurrence between the two groups.
Endoscopic remission: Defined as SES-CD score of 0–2. Scores of 3–6, 7–15, and ≥ 16 were classified as mild, moderate, and severe endoscopic activity, respectively. Disease activity was classified based on PCDAI scores, which were interpreted as follows: < 10, remission; 10–30, mild activity; 30–40, moderate activity; and 40–100, severe activity.
Clinical remission of PD
Defined as the absence of related clinical symptoms, including fistulas, abscesses, or discharge following treatment. Clinical non-remission was defined as persistent active perianal CD after treatment, as evidenced by the presence of visible fistulas, discharges, abscesses, ulcers, or setons [8].
Diagnostic delay: assessed at two time intervals: time from symptom onset to first medical visit, defined as the interval between symptom onset and the patient’s first consultation at a healthcare facility (outpatient or emergency department) for presenting symptoms, and time from first medical visit to diagnosis, defined as the interval between the first healthcare consultation and the confirmed diagnosis.
Biologic switching
Defined as the transition from one biological agent to another owing to a primary nonresponse, secondary loss of response, or adverse drug reactions.
Statistical analysis
All statistical analyses were performed using SPSS version 30.0 (IBM Corp., Armonk, NY, USA). Continuous variables were tested for normality. Data exhibiting a normal distribution were expressed as mean ± standard deviation (SD) and were compared between groups using an independent-samples t-test.
Non-normally distributed data are expressed as median (interquartile range [IQR], i.e., P25–P75) and were compared using the Mann–Whitney U test for between-group comparisons and the Wilcoxon signed-rank test for pre- versus post-treatment comparisons. Categorical variables were expressed as frequencies (percentages) and compared using the chi-square test or Fisher’s exact test, as appropriate. Survival analysis was performed using the Kaplan–Meier method to generate survival curves, and between-group comparisons were assessed using the log-rank test.
All statistical tests were two-sided, and differences with P < 0.05 were considered to be significant.
Results
Baseline characteristics
Fig. 1Annual number of children diagnosed with CD at our center from 2015 to 2024, stratified by PD status. Dark gray bars, PD; light gray bars, non-PD.
A total of 161 children were diagnosed during the study period; after excluding 17 patients, 144 were included in the final analysis
During the study period, 161 children were diagnosed with CD (Fig. 1). After excluding 17 patients with irregular follow-up or incomplete data, data from 144 patients (89/144 [61.81%] male and 55/144 [38.19%] female; male-to-female ratio, 1.62:1) were included in the final analysis.
The median age at diagnosis was 9 years (IQR, 3.85–12.00 years), and the median follow-up was 16 months (IQR, 12–36) months (Table 1).
Table 1Baseline characteristics of CD patients at diagnosisTotal(N = 144)Sex, n (%)Male89(61.81)Female55(38.19)Age at diagnosis[Median (P25, P75), years]9.00 (3.85,12.00)Nutritional parametersBAZ (x̄ ± s)-0.07 ± 1.44HAZ [Median (P25, P75)]0.20 (-0.50,1.10)WAZ [Median (P25, P75)]-0.10 (-0.80,0.80)Age at diagnosis n (%)VEO-IBD (< 6 years)44 (30.56)A1a (0–10 years)79 (54.86)A1b (≥ 10 to < 17 years)65 (45.14)Medications, n (%)Biologics87(60.42)— Infliximab84(58.33)— Adalimumab3(2.08)Steroids15(10.42)Thalidomide31(21.53)ASA19(13.19)EEN4(2.78)Follow-up time[Median (P25, P75), months]16.00(12.00,36.00)
Fig. 2Types of perianal lesions according to the AGA classification
The overall prevalence of PD was 57/144 (39.58%), with 50/57 (87.72%) perianal abscesses and 38/57 (66.67%) perianal fistulas. The distribution of these perianal lesion types according to the AGA classification is shown in Fig. 2.
Other perianal manifestations included rectoperineal fistulas (5/57 [8.77%]), skin tags (3/57 [5.26%]), and anal fissures (2/57 [3.51%]). Among the 57 patients with PD, 35 (61.40%) presented with perianal manifestations as the first symptom of CD, and the remaining 22 (38.60%) had PD detected at the time of CD diagnosis or developed PD during follow-up.
The distribution of disease location, disease behavior, disease activity, and intestinal surgery is summarized in Table 2, and diagnostic delay and clinical outcomes are presented in Table 3. In brief, colonic involvement (L2), severe PCDAI, intestinal surgery, and a longer time from symptom onset to first medical visit were more common in the PD group, whereas the clinical remission rate was lower in the PD group (all P < 0.05).
Table 2Clinical characteristics of pediatric CD patients at diagnosisPD(N = 57)NPD(N = 87)χ2/t/ZPSex, n (%)χ2 = 0.9440.331Male38(66.67)51(58.62)Female19(33.33)36(41.38)Age at diagnosis[Median (P25, P75) years]9.00(5.50,12.00)9.00(1.60,12.50)Z=-0.1510.880Nutritional parametersBAZ (x̄ ± s)-0.14 ± 1.54-0.01 ± 1.38t = 0.5290.598HAZ [Median (P25, P75)]0.20(-0.67,1.30)0.20(-0.50,0.90)Z=-0.0610.951WAZ [Median (P25, P75)]-0.20(-1.02,0.80)-0.10(-0.50,0.80)Z=-1.1000.272Age at diagnosis n (%)χ2 = 0.6050.437A1a (0–10 years)29(50.88)50(57.47)A1b (≥ 10 to < 17 years)28(49.12)37(42.53)Location n (%)L1 ileal0(0.00)5(5.75)χ2 = 1.8960.169L2 colonic25(43.86)17(19.54)χ2 = 9.8590.002L3 ileocolonic30(52.63)58(66.67)χ2 = 2.8540.091L4 (L4a and/or L4b)29(50.88)57(65.52)χ2 = 3.0680.080Behavior n (%)χ2 = 2.9390.401B1 inflammatory51(89.47)72(82.76)χ2 = 1.2470.264B2 stricturing5(8.77)9(10.34)χ2 = 0.0970.755B3 penetrating1(1.75)2(2.30)χ2 = 0.0001.000B2B30(0.00)4(4.60)χ2 = 1.2620.261PCDAI n (%)χ2 = 8.6500.013Mildly active17(29.82)45(51.72)χ2 = 6.7360.009Moderately active15(26.32)22(25.29)χ2 = 0.0190.890Severely active25(43.86)20(22.99)χ2 = 6.9820.008Intestinal surgery n (%)9(15.79)5(5.75)χ2 = 3.9570.047
Table 3Clinical outcomes and diagnostic delay of PD versus non-PD in the full cohortVariablePD(N = 57)NPD(N = 87)χ2/t/ZPTime from symptom onset to presentation [Median (P25, P75), months]0.75(0.40,1.55)0.50(0.20,1.00)Z=-2.1630.031Time from presentation to diagnosis[Median (P25, P75) months]2.50(1.50,8.35)2.00(1.15,6.30)Z=-1.2140.225Time from symptom onset to diagnosis[Median (P25, P75), months]4.00(2.30,10.10)3.30(1.60,10.40)Z=-1.6810.093Follow-up time[Median (P25, P75), months]13.00(12.00,32.75)18.00(12.00,38.00)Z=-1.9400.052Prognosis n (%)Clinical remission41(71.93)83(95.40)χ2 = 15.8640.000Active disease13(22.81)4(4.60)χ2 = 10.9670.001
Table 4Clinical features and outcomes of PD versus NPD in the VEO groupSex, n (%)PD(N = 22)NPD(N = 22)χ2/t/ZPχ2 = 0.0910.763Male12(54.55)11(50.00)Female10(45.45)11(50.00)Age at diagnosis[Median (P25, P75), years]0.78(0.25,2.23)1.70(1.00,3.85)Z=-1.9980.046Nutritional parametersBAZ (x̄ ± s)-0.53 ± 0.910.27 ± 0.82t = 3.0540.004HAZ [M(P25,P75)]-0.34(-1.20,0.38)-0.30(-0.83,0.15)Z=-0.3290.742WAZ(x̄ ± s)-0.66 ± 1.120.00 ± 0.57t = 2.4680.018Location, n (%)L1 ileal0(0.00)0(0.00)--L2 colonic21(95.45)13(59.09)χ2 = 8.2820.004L3 ileocolonic1(4.55)9(40.91)χ2 = 8.2820.004L4 (L4a and/or L4b)2(9.09)2(9.09)χ2 = 0.0001.000Behavior, n (%)χ2 = 1.4450.607B1 inflammatory19(86.36)21(95.45)χ2 = 0.2750.600B2 stricturing2(9.09)1(4.55)χ2 = 0.0001.000B3 penetrating1(4.55)0(0.00)/1.000Growth delay, n (%)χ2 = 5.9390.015G013(59.09)20(90.91)G19(40.91)2(9.09)PCDAI, n (%)χ2 = 9.7790.005Mildly active1(4.55)3(13.64)χ2 = 0.2750.600Moderately active2(9.09)10(45.45)χ2 = 7.3330.007Severely active19(86.36)9(40.91)χ2 = 9.8210.002Intestinal surgery, n (%)7(31.82)2(9.09)χ2 = 3.4920.062— Bowel resection + stoma5(71.43)*2(100)*— Stoma alone2(28.57)*0(0.00)Medications, n (%)Steroids1(4.55)11(50.00)χ2 = 11.4580.001Thalidomide20(90.91)10(45.45)χ2 = 10.4760.001Adalimumab1(4.55)0(0.00)/1.000Infliximab1(4.55)4(18.18)χ2 = 0.9030.342ASA0(0.00)8(36.36)χ2 = 7.4860.006Outcome, n (%)χ2 = 9.8270.006Disease activity10(45.45)3(13.64)χ2 = 5.3500.021Clinical remission9(40.91)19(86.36)χ2 = 9.8210.002Death3(13.64)0(0.00)χ2 = 1.4310.232Monogenic, n (%)17(77.27)7(31.82)χ2 = 9.1670.002* Percentages of surgical types were calculated based on the number of patients who underwent intestinal surgery in each group (PD, n = 7; non-PD, n = 2)
Table 5Medication use, surgery, and outcomes in monogenic versus non-monogenic VEO-IBD patientsMedications, n (%)Monogenic(n = 24)Non-monogenic(n = 20)χ2PSteroids5(20.83)7(35.00)χ2 = 1.1040.293Thalidomide16(66.67)14(70.00)χ2 = 0.0590.808Adalimumab0(0.00)1(5.00)/0.455Infliximab1(4.17)4(20.00)χ2 = 1.3710.242ASA0(0.00)8(40.00)χ2 = 9.1980.002Intestinal surgery, n (%)8(33.33)1(5.00)χ2 = 3.7820.052Outcome, n (%)Disease activity7(29.17)6(30.00)χ2 = 0.0040.952Clinical remission14(58.33)14(70.00)χ2 = 0.6420.423Death3(12.50)0(0.00)χ2 = 1.0760.300
Table 6Laboratory findings and diagnostic delay of PD versus NPD in the VEO cohortHB(g/L)༈x̄ ± s༉PD(N = 22)NPD(N = 22)χ2/t/ZP98.45 ± 16.91114.18 ± 16.70t = 3.1030.003CRP(mg/L) [Median (P25, P75)]20.75(12.38,44.63)2.65(0.87,8.00)Z=-3.6270.000ESR(mm/h) [Median (P25, P75)]23.50(15.00,33.25)10.00(3.00,24.00)Z=-2.270.023ALB(g/L) [Median (P25, P75)]34.00(32.03,38.68)42.00(37.60,43.33)Z=-2.8410.004Time from symptom onset to presentation[Median (P25, P75), months]0.50(0.28,0.80)0.50(0.30,0.85)Z=-0.2730.785Time from presentation to diagnosis[Median (P25, P75), months]2.70(1.50,13.70)2.15(1.43,5.60)Z=-1.0820.279Time from symptom onset to diagnosis[Median (P25, P75), months]3.80(2.23,14.05)2.65(1.85,6.35)Z=-1.3980.162
Table 7Clinical features and outcomes of PD versus NPD in the LO groupSex, n (%)PD(n = 35)NPD(n = 65)χ2/t/ZPχ2 = 1.6470.199Male26(74.29)40(61.54)Female9(25.71)25(38.46)Age at diagnosis12.20(10.11,13.20)10.30(8.11,12.90)Z=-2.3960.017[Median (P25, P75), years]Nutritional parametersBAZ (x̄ ± s)-0.2(-1.8,1)-1.16(-1.88,0.06)Z=-1.9870.047HAZ (x̄ ± s)0.71 ± 1.710.36 ± 1.29t = 1.0740.288WAZ [Median (P25, P75)]0.21(-1.3,1.61)-0.64(-1.25,0.45)Z=-1.6430.100Age at diagnosis5.2410.0226 to < 10 years11(31.43)36(55.38)10 to < 17 years24(68.57)29(44.62)Location, n (%)L1 ileal0(0)5(7.69)1.4460.229L2 colonic4(11.43)4(6.15)0.2930.589L3 ileocolonic29(82.86)49(75.38)0.7400.390L4a6(17.14)29(44.62)7.5470.006L4b27(77.14)45(69.23)0.7060.401Behavior, n (%)3.0950.372B1 inflammatory32(91.43)51(78.46)2.7110.100B2 stricturing3(8.57)8(12.31)0.0550.815B3 penetrating0(0)2(3.08)/0.540B2B30(0)4(6.15)0.9270.336Growth delay, n (%)0.3510.554G029(82.86)58(89.23)G16(17.14)7(10.77)PCDAI, n (%)χ2 = 4.5780.101Mildly active16(45.71)42(64.62)χ2 = 3.3360.068Moderately active13(37.14)12(18.46)χ2 = 4.2340.040Severely active6(17.14)11(16.92)χ2 = 0.0010.978Medications (n,%)Steroids3(8.57)0(0.00)χ2 = 3.1760.075Thalidomide0(0.00)1(1.54)/1.000Adalimumab1(2.86)1(1.54)/1.000Infliximab31(88.57)48(73.85)χ2 = 2.9730.085ASA0(0.00)11(16.92)χ2 = 5.0390.025EEN0(0.00)4(6.15)χ2 = 0.9270.336[L4a] upper disease proximal to ligament of Treitz༻L4b༽upper disease distal to ligament of Treitz and proximal to distal 1/3 ileum
Table 8Laboratory findings and diagnostic delay of PD versus NPD in the LO groupHB(g/L)(x̄ ± s)PD(n = 35)NPD(n = 65)χ2/t/ZP112.74 ± 16.79111.74 ± 17.16t = 0.2810.779CRP(mg/L)[M(P25, P75)]33.2(11.1,71)37.5(11.35,66.8)Z=-0.0360.971ESR(mm/h)[M(P25, P75)]38(16,58)31(15,47.5)Z=-1.6740.094ALB(g/L)༈x̄ ± s༉38.77 ± 5.1438.24 ± 5.75t = 0.4520.652Time from symptom onset to presentation[Median (P25, P75), months]1.10(0.50,2.00)0.50(0.20,1.20)2.7550.006Time from presentation to diagnosis[Median (P25, P75), months]2.50(1.20,6.50)2.00(1.10,7.05)-0.5030.615Time from symptom onset to diagnosis[Median (P25, P75), months]4.00(2.30,9.70)3.90(1.55,11.53)-1.0260.305
Table 9Comparison of PCDAI scores before and after treatment and perianal remission in the LO groupVariablePD (n = 35)NPD(n = 65)PPre-treatment PCDAI, [Median (P25, P75)]32.5 (20, 37.5)27.5 (17.5, 38.13)—Post-treatment PCDAI, [Median (P25, P75)]0 (0, 6.25)0 (0, 0.63)—Z (pre- vs. post-treatment)-4.710-5.5840.000Clinical remission of perianal lesions, n (%)30/35 (85.71%)——
Discussion
Pediatric CD is a chronic, progressive inflammatory bowel disease that can affect any segment of the gastrointestinal tract, from the oral cavity to the anus [1]. Perianal disease (PD) is one of the most challenging pediatric CD phenotypes. In our cohort, the overall prevalence of PD was 57/144 (39.58%), with particularly high prevalence in patients with VEO-IBD (22/44, 50.00%).
PD is consistently associated with extensive colonic involvement and greater disease severity, particularly in the VEO subgroup. The critical question remains: Does PD signify a more severe form of CD, or does it represent a distinct disease subtype? Through age-stratified analysis, our study revealed that in VEO-IBD, PD is closely associated with monogenic defects, whereas in older children, PD is characterized by prominent colonic involvement but a paradoxically milder upper systemic inflammatory burden, suggesting that PD may represent a more localized disease subtype.
In our cohort of 144 pediatric patients with CD, the overall prevalence of PD was 57/144 (39.58%), with 50/57 (87.72%) perianal abscesses and 38/57 (66.67%) perianal fistulas. This prevalence is consistent with those reported in previous studies conducted in East Asia [9]. A meta-analysis of 525,425 adult and pediatric patients with IBD noted that Asian patients with CD are more likely to have concurrent PD9.
Previous studies have reported PD rates of 34.1% and 33.3% in Japanese and Korean pediatric CD populations, respectively, whereas the prevalence in Western pediatric CD populations ranges from 8% to 20% [10]. These findings further suggest that Asian pediatric patients with CD may have a higher susceptibility to PD, a difference potentially attributable to a combination of genetic predispositions, environmental factors, and gut microbial composition.
In our full-cohort analysis, the proportion of patients with colonic involvement (L2) was significantly higher in the PD group than in the non-PD group (25/57 [43.86%] vs. 17/87 [19.54%], respectively; P < 0.05). Patients with VEO-IBD accounted for 44/144 (30.56%) of the cohort, of whom 22/44 (50.00%) had concurrent PD, a substantially higher rate than that observed in older children.
Among the 22 patients with VEO-IBD and PD, 17/22 (77.27%) harbored monogenic variants, with IL-10RA mutations being the most common (13/17 [76.47%]). Consistent with our findings, Chinese investigators reported that up to 87.7% of patients with VEO-IBD and PD had colonic involvement, with a monogenic mutation detection rate of 94.9% (74/78),among which IL-10RA mutations were predominated [11]. These observations further support a close association between PD and distal colonic involvement.
VEO-IBD resulting from IL-10RA mutations frequently presents with severe perianal manifestations, a phenomenon that may be mechanistically linked to defects in the IL-10 signaling pathway, leading to localized intestinal immune dysregulation and impaired barrier function [12]. Previous studies have reported that VEO-IBD exhibits a more aggressive disease course and significantly lower response rates to conventional therapies than LO [13],[14] In our study, the thalidomide utilization rate in patients with VEO-IBD and PD was as high as 20/22 (90.91%), which significantly exceeded that in the non-PD group. Bramuzzo et al. similarly confirmed that thalidomide is both efficacious and well-tolerated in children with VEO-IBD, with a lower incidence of adverse events than that observed in older IBD populations.
Collectively, these findings suggest that thalidomide may be a feasible option in selected patients with VEO-IBD and concurrent PD, particularly those with monogenic defects such as IL-10RA deficiency [15]. However, this finding cannot be generalized to all patients with VEO-IBD. The high utilization of thalidomide in our cohort was largely driven by the predominance of IL-10-related mutations in our monogenic-PD VEO-IBD subgroup (14/17, 82.35%), together with limited biologic availability and financial constraints in China during the study period, as discussed in detail below.
To exclude the confounding effects of genetic heterogeneity inherent to VEO-IBD, we performed a subgroup analysis of 100 patients diagnosed at ≥ 6 years of age. The results revealed that in the LO group, patients with PD predominantly presented with ileocolonic involvement at diagnosis, with a significantly lower proportion of upper gastrointestinal involvement (L4a) and lower rates of gastric and deep colonic ulcerations than those in non-PD patients. However, no significant differences were observed in any of the laboratory inflammatory markers between the two groups.
Similarly, a nationwide Korean pediatric CD cohort study found that patients with perianal fistulizing disease exhibited lower CRP and albumin levels, a higher proportion of B1-type behaviors, and better nutritional status [16]. Furthermore, a large-scale multicenter study including 6679 pediatric patients with CD reported that Asian children had a significantly higher prevalence of PD than non-Asian children, with Asian patients aged > 10 years at a particularly high risk for early onset perianal manifestations [17]. Collectively, these findings led us to hypothesize that PD may represent a relatively localized disease subtype and that its occurrence is not significantly correlated with the overall degree of gastrointestinal inflammation.
In older children presenting with perianal manifestations as the initial symptoms, a high index of suspicion for CD should be maintained, whereas monogenic etiologies should be prioritized in children aged < 6 years.
Previous studies have described pediatric CD with PD as an aggressive phenotype with high rates of surgical intervention. However, our LO-PD cohort appeared to be relatively milder, with lower rates of upper gastrointestinal involvement (L4a: 17.14% vs. 44.62%, P = 0.006) and no significant elevation in systemic inflammatory markers compared to non-PD patients.
This discrepancy could be attributed to several factors. First, the early adoption of biological therapy in our patients with LO-PD (infliximab: 31/35, 88.57%) may have modified the natural history of the disease and reduced progression to complicated phenotypes. Second, the relatively short follow-up period (median, 16 months) may not have captured long-term complications, including intestinal resection, which typically occur later in the disease course.
Third, differences in study populations, disease definitions, and classification systems across studies may have contributed to phenotypic heterogeneity. Notably, the milder phenotype observed in our LO-PD cohort supports the hypothesis that PD in older children may represent a more anatomically confined disease subtype, with the occurrence of PD not necessarily correlating with the overall degree of gastrointestinal inflammation. However, it must be acknowledged that the median follow-up in our cohort was only 16 months, whereas Kori et al. [18] demonstrated that progression from inflammatory (B1) to complicated behavior (B2/B3) required a mean follow-up of > 10 years (12.10 years in progressors vs.
9.64 years in non-progressors, P = 0.018). Therefore, we can only conclude that LO-PD appears relatively mild during the short-term follow-up, and its long-term prognosis requires further validation with extended observations.
In the LO group, the patients in the PD group were significantly older at diagnosis than those in the non-PD group, and the median time from symptom onset to the first medical visit was significantly longer in the PD group. However, no significant difference was observed between the two groups in terms of time from the first medical visit to diagnosis. This finding suggests that the older age at diagnosis in patients with PD reflects, at least in part, the cumulative effect of diagnostic delay.
When PD is the presenting symptom and the gastrointestinal manifestations are subtle, patients may initially present to the surgical or dermatology department, thereby missing timely gastroenterological evaluations. A multicenter study by Roskam et al. further revealed age-related disparities in diagnostic delay. The median time from perianal abscess/fistula to CD diagnosis was 9.0 months in the 6–12 years’ age group, which was significantly longer than the 2.0 months observed in the 12–18-year age group, highlighting that diagnostic delay is particularly pronounced in younger children [19].
Therefore, early endoscopic evaluation should be performed in children with unexplained perianal lesions irrespective of the presence or absence of gastrointestinal symptoms.
Both domestic and international expert consensus guidelines recommend early biological therapy for pediatric patients with CD and complex perianal fistulas [20, 21]. In accordance with these recommendations, we observed a significantly higher utilization rate of infliximab in the LO group than in the VEO group. In the present study, the Kaplan–Meier analysis revealed that, at 52 weeks’ follow-up, the median time to biological initiation was significantly shorter in LO patients with PD than in those without PD, underscoring the urgent need for intensified therapy in patients with PD.
Following treatment, PCDAI scores decreased significantly from baseline in both groups. In the LO group, the clinical remission rate for PD was 30/35 (85.71%), and the endoscopic remission rate was higher in the PD group than in the non-PD group (23/30 [76.67%] vs. 24/42 [57.14%], respectively).
Although this difference was not statistically significant, this trend may be attributable to earlier biological interventions. A notable finding in our study was the substantially lower infliximab utilization rate in the VEO group (5/44 [11.36%]) compared with the LO group (79/100 [79.00%]), despite VEO-IBD patients having more severe disease phenotypes and a higher burden of monogenic mutations. This apparent paradox warrants further study.
First, thalidomide was the predominant therapy in patients with VEO-IBD (30/44 [68.18%]), which is supported by our center’s previous studies demonstrating its efficacy in IL-10RA-deficient patients [22, 23] as well as by independent studies confirming its immunomodulatory role in VEO-IBD [24]. Second, patients with VEO-IBD are known to have poorer responses to anti-TNF agents than older children, and a recent multicenter real-world study reported that infliximab as a first-line therapy achieved only 19% steroid-free clinical remission and 36% persistence at one year [25]. Third, during the study period, infliximab faced insurance coverage limitations in China for children under 6 years of age, whereas thalidomide was more accessible.
Fourth, 4/17 (23.5%) patients with monogenic VEO underwent allogeneic hematopoietic stem cell transplantation (HSCT) as a curative treatment, which has been shown to achieve favorable survival outcomes in IL-10RA-deficient VEO-IBD [26], further reducing biological exposure in this subgroup.
Importantly, the treatment pattern in our VEO cohort should be interpreted with caution. Although our center generally follows ESPGHAN/ECCO guidelines, thalidomide was used in 30/44 (68.2%) VEO-IBD patients overall, including 16/24 (66.7%) patients with monogenic disease and 14/20 (70.0%) without identified monogenic variants, whereas infliximab was used in only 5/44 (11.4%) patients overall (1/24 [4.2%] with monogenic disease and 4/20 [20.0%] without identified monogenic variants) (Table 5). This treatment pattern indicates that thalidomide was frequently used at an early stage of therapy in our cohort, which differs from the 2014 ECCO-ESPGHAN and subsequent guideline recommendations that thalidomide generally be reserved for selected patients who have failed or cannot tolerate anti-TNF therapy.
Given its potential adverse effects, including peripheral neuropathy, sedation, and teratogenicity, thalidomide should be used cautiously with appropriate clinical monitoring. The relatively frequent use of thalidomide in our cohort may reflect several real-world constraints in China during the study period, including limited insurance coverage of infliximab for children younger than 6 years, limited access to proactive therapeutic drug monitoring (TDM), restricted availability of HSCT, and financial barriers to biologic therapy. In addition, the high proportion of IL-10-related variants (14/17, 82.35%) among patients with monogenic VEO-IBD may have further contributed to the empirical use of thalidomide in clinical practice.
Similar treatment patterns have been reported in other Chinese pediatric cohorts. Importantly, these findings should not be interpreted as evidence that thalidomide is superior to or can replace anti-TNF therapy in VEO-IBD; rather, they reflect treatment practices shaped by the clinical and healthcare context during the study period and highlight the potential role of thalidomide in selected patients in real-world practice.
Importantly, given the low infliximab utilization rate in our VEO cohort (5/44, 11.36%), we were unable to make recommendations for early biological therapy in this subgroup based on the present data. Therefore, based on our observational data, no conclusion can be drawn regarding the comparative effectiveness of thalidomide versus anti-TNF therapy in VEO-IBD. In selected monogenic cases, HSCT may be considered according to genetic diagnosis and clinical eligibility.
In the present study, perianal surgical intervention was common in patients with LO and PD, whereas intestinal surgery was uncommon and did not differ significantly between the two groups.However, Kaplan–Meier analysis revealed no significant differences between the two groups in the probability of surgical intervention or biologic switching, These findings may partly reflect the relatively early use of biologic therapy in the LO-PD group.Notably, 12 patients with PD in the LO group underwent perianal surgery (primarily for drainage and infection control) before CD diagnosis, further corroborating the clinical reality that perianal manifestations can precede and serve as presenting features of CD. In terms of imaging assessment, 11 patients in the LO group with PD underwent both perianal MRI and perianal ultrasound. Among those who underwent both modalities, perianal abscesses were detected by both MRI and ultrasound in six (54.5%) cases, and complex fistulas were identified in four (36.4%), suggesting that MRI confers a diagnostic advantage in identifying complex perianal fistulas.
Similarly, Akkelle et al. found that more than half of the patients without visible external openings were diagnosed with perianal CD by MRI, of whom 77% had intersphincteric fistulas, confirming the critical value of MRI in delineating anatomical structures and classifying fistula types27.
This study has several limitations, the first of which is its single-center, retrospective design, relatively small sample size (n = 144), and median follow-up of only 16 months. Therefore, the findings may not be sufficient to rigorously evaluate long-term outcomes, and the generalizability of our conclusions is limited. Second, incomplete endoscopic follow-up data (available for only 30 patients in the PD group and 42 in the non-PD group) may have reduced the statistical power and introduced a potential bias.
Third, genetic testing was systematically performed only in patients with VEO-IBD, and older children did not routinely undergo next-generation sequencing, which may have resulted in missed monogenic etiologies. Fourth, routine proactive therapeutic drug monitoring (TDM) was not uniformly performed during the study period; it was only conducted in patients with a loss of response or suspected adverse reactions. This may have affected the optimization of biologic therapy, although this practice reflects the limited availability of TDM in China during the earlier years of the study period.
Fifth, EGD was not performed in all VEO-IBD patients (77.3% completion rate). Sixth, treatment selection in the VEO cohort was influenced by real-world resource constraints and was not fully guideline-concordant, as detailed above. Treatment responses may therefore have been biased by biologic availability, clinician discretion, and other unmeasured factors, limiting the comparability of our treatment outcomes with those of guideline-concordant cohorts.
In addition, limited availability and tolerability concerns regarding conventional immunomodulators (e.g., azathioprine, methotrexate) in very young children may also have contributed to the frequent empirical use of thalidomide in this cohort. Finally, survival analysis was not performed for the VEO-IBD subgroup because of the limited sample size (n = 44) and incomplete follow-up data for some patients who were followed up in outpatient clinics.
Conclusions
In this single-center retrospective study of 144 pediatric patients with CD, PD was present in 57/144 (39.58%) of cases and was associated with colonic involvement and greater disease severity. Age-stratified analysis revealed distinct characteristics: VEO-PD was strongly associated with monogenic variants, predominantly involving the IL-10 signaling pathway, particularly IL-10RA, and was characterized by poor remission rates (9/22, 40.91%), whereas LO-PD responded favorably to early biologic therapy, with a clinical remission rate of 30/35 (85.71%).In selected patients with monogenic VEO-IBD, thalidomide and, in eligible cases, HSCT may represent feasible individualized therapeutic options; however, thalidomide is not appropriate for all causes of VEO-IBD and should not be generalized as a first-line therapy for the entire VEO-IBD population.Early recognition of perianal manifestations, timely biologic intervention in older children, and genetic testing in VEO-IBD patients are key to improving clinical outcomes. Longer-term prospective studies with extended follow-up are needed to optimize personalized management strategies for pediatric CD patients with PD.
Comparison of clinical characteristics between patients with VEO-IBD, with and without PD
Of the 44 patients in the VEO group, 22 each were classified into the PD group and 22 into the non-PD group. There was no significant difference in the sex distribution between the two groups (Table 4). The median age at diagnosis was significantly lower in the PD group (0.78 years) than in the non-PD group (1.70 years; P < 0.05).
Regarding nutritional indices, the PD group had significantly lower WAZ and BAZ scores than the non-PD group (P < 0.05), whereas no significant difference was found in HAZ scores between the two groups (Table 4). In terms of the disease location, neither group exhibited L1 involvement. The proportion of L2 involvement was significantly higher in the PD group than in the non-PD group (21/22 [95.45%] vs.
13/22 [59.09%], respectively, P < 0.05), whereas the proportion of L3 involvement was significantly lower in the PD group (1/22 [4.55%] vs. 9/22 [40.91%], respectively, P < 0.05). Both groups had an L4 involvement rate of 2/22 (9.09%), with no significant difference between the two groups (Table 4).
Regarding disease behavior, type B1 was the predominant type in both groups, with no significant differences in the distribution of other behavior types (Table 4). The proportion of patients with growth delay was significantly higher in the PD group than in the non-PD group (9/22 [40.91%] vs. 2/22 [9.09%], respectively; P = 0.015).
The PD group had a significantly higher proportion of patients with severe disease activity (19/22 [86.36%] vs. 9/22 [40.91%]; P = 0.002) and a significantly lower proportion of patients with moderate activity (2/22 [9.09%] vs. 10/22 [45.45%]; P = 0.007) than the non-PD group, with no significant difference in the proportion of patients with mild activity between the two groups (Table 4).
The utilization rate of thalidomide was significantly higher in the PD group than in the non-PD group (20/22 [90.91%] vs. 10/22 [45.45%]; P < 0.001). However, thalidomide was frequently used as a first-line or early-line therapy in this VEO cohort, including in non-monogenic patients, which is not fully consistent with the 2014 ECCO-ESPGHAN and subsequent guideline recommendations that thalidomide be reserved for selected patients who have failed or cannot tolerate anti-TNF agents.
This practice reflects real-world constraints in China, including limited insurance coverage of infliximab for children younger than 6 years and limited access to proactive TDM, rather than a deliberate deviation from guideline recommendations (see Discussion and Limitations). The utilization rates of corticosteroids (1/22 [4.55%] vs. 11/22 [50.00%]; P = 0.001) and 5-aminosalicylic acid (ASA; 0/22 [0.00%] vs.
8/22 [36.36%]; P = 0.006) were significantly lower in the PD group.There was no significant difference in the utilization rate of infliximab between the two groups (1/22 [4.55%] vs. 4/22 [18.18%]; Table 4). Endoscopic follow-up data were available for 12 patients in the PD group and 15 patients in the non-PD group.
Regarding prognosis, the PD group had a significantly higher rate of active disease (10/22 [45.45%] vs. 3/22 [13.64%]; P = 0.021) and a significantly lower clinical remission rate (9/22 [40.91%] vs. 19/22 [86.36%]; P = 0.002) than the non-PD group.
No significant difference in mortality was found between the two groups (3/22 [13.64%] vs. 0/22 [0.00%]; Table 4). Among the 22 VEO-IBD patients with PD, seven (31.82%) underwent intestinal surgery, all of whom received thalidomide in combination with surgical intervention.
In the VEO-PD subgroup, thalidomide monotherapy (without surgery) was administered to 13 patients (13/22, 59.09%).
One patient received infliximab alone without surgery (1/22, 4.55%) and one patient received adalimumab alone without surgery (1/22, 4.55%). No patient in the VEO-PD group received anti-TNF therapy combined with surgery.
In terms of genetic testing, all 44 patients in the VEO group underwent genetic testing, of whom 24/44 (54.55%) underwent whole-exome sequencing and 20/44 (45.45%) underwent targeted genome sequencing. Monogenic mutations were identified in 24/44 (54.55%). The detection rate of monogenic mutations in the PD group was 17/22 (77.27%), which was significantly higher than that in the non-PD group (7/22 [31.82%]; χ² = 9.167; P = 0.002).
Among the 17 positive cases in the PD group, interleukin-10 receptor alpha (IL-10RA) mutations were most common (13/17 [76.47%]), followed by interleukin-10 receptor beta (IL-10RB) mutation (1/17 [5.88%]), XIAP mutations (1/17 [5.88%]), and CYBB mutations (2/17 [11.76%]) (Table 4). We further compared the medication use between monogenic (n = 24) and non-monogenic (n = 20) patients in the VEO-IBD group (Table 5). Thalidomide use was comparable between the two groups (16/24 [66.67%] vs.
14/20 [70.00%]; P = 0.808). In contrast, ASA was more commonly used in the non-monogenic group (8/20 [40.00%] vs. 0/24 [0.00%]; P = 0.002).
The use of steroids (5/24 [20.83%] vs. 7/20 [35.00%]; P = 0.293) and infliximab (1/24 [4.17%] vs. 4/20 [20.00%]; P = 0.242) was comparable between the two groups, with no statistically significant differences.
We also compared intestinal surgery and clinical outcomes between the two subgroups (Table 5). All intestinal surgeries were bowel resection with stoma. Intestinal surgery was performed in 8/24 (33.33%) monogenic versus 1/20 (5.00%) non-monogenic patients (P = 0.052).
Clinical remission was achieved in 14/24 (58.33%) versus 14/20 (70.00%) patients (P = 0.423), and death occurred in 3/24 (12.50%) versus 0/20 (0.00%) patients (P = 0.300). Although these differences were not statistically significant, monogenic patients showed a numerically higher surgery rate and a lower remission rate, which should be interpreted with caution given the small sample size.These findings suggest that the genetic background may influence therapeutic decisions in patients with VEO-IBD, particularly with regard to ASA use. Among the 17 patients with monogenic mutations in the VEO-PD group, 4/17 (23.5%) with interleukin-10 receptor alpha (IL-10RA) mutations underwent hematopoietic stem cell transplantation (HSCT) at the Department of Hematology at our center during follow-up.
By the end of the follow-up period, three of the four patients achieved both clinical remission and mucosal healing. The remaining 13 patients did not undergo HSCT because of limited availability during the earlier years, family refusal, or clinical improvement with thalidomide. All 44 patients (100%) in the VEO group underwent complete colonoscopy, and 34 of the 44 (77.3%) underwent esophagogastroduodenoscopy.
All patients (100%) underwent intestinal ultrasonography. On gastroscopy, the prevalence of erosions was significantly higher in the PD group than in the non-PD group (66.7% vs. 25.0%, P = 0.015), whereas no other endoscopic findings differed significantly between the two groups.
On colonoscopy, deep ulcerations were more common in the PD group than in the non-PD group (45.5% vs. 18.2%); however, this difference was not statistically significant (P = 0.052), and no other endoscopic features showed significant differences between the two groups. Intestinal ultrasonography revealed no significant differences in bowel wall thickening, luminal strictures, or bowel stiffness between the two groups (all P > 0.05).
Regarding laboratory parameters, the PD group exhibited significantly lower median hemoglobin and albumin levels than the non-PD group (P < 0.05), whereas the median C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) levels were significantly higher in the PD group (P < 0.05). No significant differences were observed in any of the diagnostic delay intervals between the two groups (Table 6).
Comparison of clinical characteristics between patients with LO disease, with and without PD
To exclude the potential confounding effects of genetic heterogeneity inherent to VEO-IBD, a subgroup analysis restricted to patients ≥ 6 years of age was performed. Data from 100 patients were included in this subgroup, with 35/100 (35.00%) patients in the PD group and 65/100 (65.00%) patients in the non-PD group.
The age at onset in the PD group was significantly higher than in the non-PD group, with the majority of patients aged 10–17 years. The proportion of patients with upper gastrointestinal tract involvement was significantly lower in the PD group than in the non-PD group (6/35 [17.14%] vs. 29/65 [44.62%], respectively; P < 0.05).
In both the groups, L3 was the predominant disease location. The disease behavior classification, growth delay, and laboratory parameters were comparable between the two groups (Table 7).
Laboratory parameters, including hemoglobin, CRP, ESR, and albumin levels, were comparable between the two groups (Table 8). Regarding diagnostic delay, the PD group had a significantly longer median time from symptom onset to first medical visit compared with the non-PD group (1.10 months [IQR, 0.50–2.00 months] vs. 0.50 months [IQR, 0.20–1.20 months]; P < 0.05).
However, there was no significant difference between the two groups in terms of the time from the first medical visit to diagnosis (P > 0.05). Similarly, the total time from symptom onset to diagnosis did not differ significantly between the two groups (4.00 months [IQR, 2.30–9.70 months] vs. 3.90 months [IQR, 1.55–11.53 months]; P > 0.05) (Table 8).
All 100 patients underwent a complete colonoscopy. The incidence of deep ulcerations was significantly higher in the non-PD group than in the PD group (35/65 [53.85%] vs. 11/35 [31.43%]; P < 0.05), whereas no other endoscopic findings differed significantly between the two groups.
All patients underwent esophagogastroduodenoscopy, and the incidence of gastric ulcers was significantly higher in the non-PD group than in the PD group (26/65 [40.00%] vs. 7/35 [20.00%], P < 0.05). Capsule endoscopy was performed in 71 patients, with no significant differences between the two groups in the findings assessed (P > 0.05).
Detailed comparisons of the baseline PCDAI and medication use between PD and non-PD patients in the LO group are presented in Table 7.
Perianal magnetic resonance imaging and ultrasound findings in LO patients with PD
Among the 35 patients ≥ 6 years of age with PD, 33/35 (94.29%) underwent perianal ultrasonography at the time of initial diagnosis, of whom 16/33 (48.48%) were diagnosed with perianal fistulas. Eleven patients underwent perianal magnetic resonance imaging (MRI) and ultrasonography simultaneously. Among those who underwent both imaging modalities, perianal abscesses were detected by both MRI and ultrasound in 6/11 (54.55%) cases, and complex perianal fistulas were identified in 4/11 (36.36%) cases.
Representative imaging findings from patients with perianal disease are presented in Fig. 3.
Fig. 3Representative imaging findings in PD. (B) and, (C) (A) CT enterography (CTE) showing small bowel involvement; (B) pelvic MRI; (C) pelvic MRI
Treatment and surgical interventions in patients with LO disease, with and without PD
Kaplan–Meier survival curves were constructed, and the log-rank test was used to compare the cumulative probabilities of patients remaining free from the three outcome events (i.e., biologic-naïve status, surgical intervention, and biologic switching) between the PD and non-PD groups. By week 52 of follow-up, the cumulative probability of remaining biologic-naïve was significantly lower in the PD group than in the non-PD group (P < 0.05). No significant differences were observed between the two groups in the cumulative probability of undergoing surgical intervention or biological switching (P > 0.05) (Figs.
4, 5 and 6).
In the PD group, seven patients underwent surgical intervention at or after diagnosis. The predominant surgical procedures were non-cutting seton drainage for perianal fistulas (4/7 [57.14%]) and incision and drainage for perianal abscesses (2/7 [28.57%]), with only 1/7 (14.29%) patients undergoing intestinal surgery. In the non-PD group, five patients underwent surgical intervention, all of whom (5/5 [100.00%]) underwent intestinal surgery.
Fig. 4Kaplan-Meier curves comparing biologic-free survival between PD and NPD groups in the LO group Note: The y-axis shows the cumulative probability of remaining biologic-naïve and the x-axis shows the follow-up time in weeks. The event of interest was the initiation of the biological therapy.
The log-rank test demonstrated a significantly lower cumulative probability of remaining biologic-naïve in the PD group than in the non-PD group (P < 0.05)
Fig. 5Kaplan-Meier curves comparing surgery-free survival between PD and NPD groups in the LO group Note: The y-axis shows the cumulative probability of remaining surgery-free and the x-axis shows the follow-up time in weeks. The event of interest was the first CD-related surgery (including intestinal and perianal surgeries).
The log-rank test demonstrated no significant difference in surgery-free survival between the two groups
Fig. 6Kaplan-Meier curves comparing biologic-switch-free survival between PD and NPD groups in the LO group Note: The y-axis shows the cumulative probability of remaining biologic-switch-free and the x-axis shows the follow-up time in weeks. Observations began from the date of the first biological initiation.
The event of interest was the first biological switch (due to a primary or secondary loss of response). The log-rank test demonstrated no statistically significant difference in biologic switch-free survival between the two groups
Clinical remission in patients with LO disease, with and without PD
The PCDAI scores decreased significantly from baseline after treatment in both groups (P < 0.05). Owing to incomplete endoscopic follow-up in some patients, endoscopic remission was assessed in 30 patients in the PD group and 42 in the non-PD groups, respectively. The endoscopic remission rate was higher in the PD group (23/30 [76.67%]) than in the non-PD group (24/42 [57.14%]), although the difference was not statistically significant (Table 9).
The clinical remission rate of perianal lesions in the PD group was 30/35 (85.71%) (Table 9).
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1