Fistulizing Anorectal Crohn Disease in a Child With Severe HLA-B27-associated Enthesitis-related Arthritis

Juvenile idiopathic arthritis (JIA) represents a family of childhood arthropathies characterized by chronic synovial inflammation. Enthesitis-related arthritis (ERA), a subgroup of JIA and formerly known as juvenile ankylosing spondylitis, is frequently associated with inflammatory bowel disease (IBD). ERA represents the pediatric form of spondyloarthropathy (SA) in adults. The typical presentations of ERA are seronegative oligoarthritis associated with enthesitis affecting the joints of the lower extremities. Abnormalities of the axial skeleton, including sacroiliitis, are absent in the early stages of the disease ⁽¹⁾. The histocompatibility antigen HLA-B27 is strongly linked with adult ankylosing spondylitis (AS) (90%) and with the risk for the development of AS in IBD ⁽²⁾. This antigen is associated with 25% of cases of IBD and is present in 70% of patients with IBD having inflammation of the axial skeleton ⁽³⁾. The prevalence of the HLA-B27 antigen in adult patients with Crohn disease (CD) has been found to be 12%, although it is 78% in patients with CD and AS ⁽⁴⁾.

The prevalence of IBD is increased in adults with other autoimmune diseases, particularly AS, and in those with reactive and psoriatic arthritis ⁽⁵⁾. Indeed, in 1 study, ileocolonoscopies confirmed gut inflammation in 57% of adults with SA ⁽⁶⁾. Another study also reported the eventual development of CD in patients with SA ⁽⁷⁾. Conversely, arthritic manifestations are common extraintestinal symptoms of CD in children ⁽⁸⁾. In another study, 35% of patients with CD fulfilled criteria consistent with the European Spondyloarthropathy Study Group for SA ⁽¹⁾. To our best knowledge, this is the first report to describe the development and successful treatment of fistulizing CD in a child with long-standing ERA.

CASE REPORT

In February 2006, a 15-year-old boy with an established diagnosis of ERA presented with 1 episode of acute-onset nonbilious and nonbloody vomiting, epigastric pain, and liquid bloody stools of 8 days' duration. He was afebrile and also described worsening joint pains and a weight loss of 5 kg during the same period.

The patient's medical history was significant for established HLA-B27-positive JIA, which had been diagnosed 4 years previously. With each successive joint involvement his diagnosis ascended from oligoarthritis through to ERA. The results of repeated full metabolic and rheumatology panels, blood count with differential, liver function tests, inflammatory markers, urinalysis, and microbiology analyses were normal, as were abdominal ultrasound and ophthalmological investigations. However, in May 2004 and again in January 2006, he presented with right-sided and then left-sided iridocyclitis. His maintenance medication included nonsteroidal anti-inflammatory drugs (NSAIDs), sulfasalazine, methotrexate, and corticosteroids. There was no family history of rheumatic diseases, allergies, psoriasis, or IBD. His growth and development during the 4 years were normal.

On physical examination, his abdomen was nondistended and mildly tender to deep palpation. No organomegaly or masses were found. Joint pain and stiffness were present in the toes and ankles on both sides, the right hip, both sacroiliac joints, and the right elbow.

Laboratory findings included the following: erythrocyte sedimentation rate 52 mm/hour; C-reactive protein 195 mg/L, and leukocytes 13 × 10⁹/L with 31% neutrophil bands. Infectious causes were excluded. Abdominal ultrasound showed no abnormalities, and esophagogastroduodenoscopy revealed a normal macroscopic appearance. Although duodenal and esophageal biopsy specimens seemed were normal, gastric biopsies revealed a mild Helicobacter pylori-negative gastritis.

Although the gastrointestinal symptoms responded to histamine-2 receptor blocker (ranitidine), the patient was readmitted in May 2006 with a perianal abscess, which had to be repeatedly drained. Five months later, ileocolonoscopy and subsequent fistulography revealed inflammatory changes in the ileum and a complete external anorectal fistula (Fig. 1). Abdominal ultrasound and magnetic resonance imaging enterography confirmed enlargement and inflammation of the terminal ileum and also sacroiliitis (Figs. 2 and 3).

Fig. 1

Fig. 2

Fig. 3

Laboratory values were as follows: erythrocyte sedimentation rate 15 mm/hour, leukocytes 12.5 × 10⁹/L with 7% of band neutrophils, C-reactive protein 24 mg/L. Fecal calprotectin was elevated at >150 μg/g. Pancreatic organ-specific and anti-Saccharomyces cerevisiae (ASCA) antibodies were positive, whereas anti-goblet cells antibodies and celiac antibodies were negative. The HLA serotyping and genotyping of the patient were as follows: HLA class I-A2, 32, B27, 51 Cw1, 2; HLA class II alleles HLA-DRB1*11, 01; HLA-DQB1*03, 05. Tumor necrosis factor (TNF)-α 308 G/G genotype and heterozygous 4G/5G genotype for the type 1 plasminogen activator inhibitor were observed. The Pediatric Crohn Disease Activity Index score was 32.5.

The overall treatment strategy included mesalamine (5-ASA), metronidazole, methotrexate, and infusions of infliximab administered at weeks 0, 2, and 6 at a dosage of 5 mg/kg of body weight and again every 8 weeks until week 54. To prevent the formation of human anti-chimeric antibodies, hydrocortisone was given immediately before each dose of infliximab. Complete and permanent fistula closure was achieved in 6 weeks without surgical intervention. The patient's joint manifestations also completely resolved.

DISCUSSION

This study has provided new evidence regarding pediatric-onset anorectal fistulizing CD and long-standing JIA. Treatment with infliximab, a chimeric anti-TNF-α monoclonal antibody, was effective in controlling the activity of both inflammatory diseases.

To date, the coexistence of ERA and CD has not been reported in children as far as we are aware. Clinical and epidemiological evidence supports the hypothesis that common genetic pathways may underlie both inflammatory diseases ^(9,10). T cells play a critical role in the early stages of JIA. T cell activation triggers immunopathological mechanisms, including B cell activation, complement consumption, interleukin release, and TNF-α, leading to damage of joints and connective tissue ⁽¹¹⁾. Polyarticular JIA is associated with HLA alleles at DR4, and oligoarticular JIA has been associated with HLA alleles at DR8 and DR5 ⁽¹²⁾. ERA as a subtype of JIA is considered to be strongly associated with HLA-B27 ⁽²⁾. CD is influenced by a complex interaction of genetic, immunologic, and environmental triggers and is thought to result from a sustained and inappropriate activation of the mucosal immune response. TNF-α is centrally involved in a cytokine cascade, resulting in irreversible damage to affected tissue ⁽¹³⁾. CD has a highly positive association with the HLA-DRB3*0301 ⁽¹⁴⁾. Patients with phenotype B27, B44 are at high risk for development of the common manifestations of CD and AS ⁽¹⁵⁾. Our patient also demonstrated HLA-B27 positivity.

Of children with IBD, 20% may present with peripheral arthritis, spondylitis, and sacroiliitis, although only the latter 2 conditions are associated with HLA-B27 and IBD ^(2,8). Although data regarding the prevalence of ERA in children with CD are not available, AS and IBD are apparently linked in adulthood, when the total prevalence of AS in patients with IBD was found to be 3.7%, and in patients with CD 6.0% ⁽¹⁶⁾. The onset of AS precedes the development of clinical findings of IBD in all documented cases in adults, although no correlation between the distribution of intestinal inflammation and the occurrence of AS has been reported ⁽¹⁶⁾.

Genetic susceptibility has been associated with pediatric-onset rather than adult-onset CD ⁽¹⁷⁾. TNF-α promoter polymorphisms are associated with inducible levels of TNF-α in vitro. Few reports are available regarding TNF-α polymorphism in pediatric-onset CD. Recently, our group demonstrated the association of the 308 A polymorphism with stricturing/penetrating behavior in children with CD and with increased disease activity in patients with CD and UC ⁽¹⁸⁾. The TNF-α-308A polymorphism promoter is significantly involved in the predisposition to both CD and UC ⁽¹⁹⁾. It is also suggested that the CARD15 and TNF-α 308 polymorphisms are associated with increased risk for CD, displaying distinct clinicopathological profiles ⁽²⁰⁾. The association of TNF polymorphisms with JIA is still controversial ^(21-23).

It has been suggested that genotyping the 4G/5G polymorphism of type 1 plasminogen activator inhibitor, the major inhibitor of fibrinolysis, may be useful in identifying a subgroup of patients whose CD shows penetrating behavior ⁽²⁴⁾. The 4G/5G genotype increases the probability of the penetrating phenotype and was also found in our patient's case.

Clinical and histological similarities in gut inflammation exist in people with seronegative SA and CD ⁽⁶⁾. Asymptomatic ileitis in adults has been described in subclinical CD, NSAID use, and especially SA ^(6,25). Although as many as two-thirds of patients with undifferentiated SA may have macroscopic or microscopic subclinical gut inflammation, overt CD develops in only a fraction of them ^(5,7). Patients with seronegative SA and CD also share the HLA-BW62 genetic marker, thus implying that these individuals with asymptomatic SA may have an underlying subclinical form of CD in which the joint manifestations are the main clinical findings. Although little information is available in children, 1 study has revealed gut lesions in 9 of 12 children with late-onset juvenile chronic arthritis ⁽²⁶⁾. Of the 5 children with chronic arthritic forms of disease, 1 child also presented with CD. During the 4.5 years since the onset of joint inflammation, the patient did not experience any gut symptoms. Moreover, the patient did not demonstrate any indication of subclinical gut inflammation (anemia, sideropenia, or hypoalbuminemia). Thus, there was no indication for the performance of ileocolonoscopy.

Although their actual pathological role is unknown, ASCA are significantly raised in CD and SA, particularly in AS, compared with healthy control individuals ^(27,28). ASCA may link AS with IBD as a common immunological feature, although further studies in children are needed to explore this issue.

Since the first description of terminal ileitis it has become apparent that not all cases of ileitis represent CD ⁽²⁵⁾. NSAIDs are also known to cause adverse inflammatory effects in the gastrointestinal mucosa, although the prevalence and magnitude of inflammation are unrelated to the type and dose ⁽²⁹⁾. Our patient had been treated with NSAIDs for 4 years without any previous manifestations of gastrointestinal symptoms. Furthermore, all clinical and laboratory findings support our diagnosis of CD with terminal ileitis, rather than an NSAID-induced enteropathy.

Moreover, the positive ASCA and elevated C-reactive protein, erythrocyte sedimentation rate, and fecal calprotectin supported our diagnosis of fistulizing CD, although fecal calprotectin levels are not valuable for assessing NSAID-induced enteropathy ^(30,31). Our patient also demonstrated pancreatic organ-specific antibodies, which are significantly associated with CD in comparison with patients with UC and control individuals ⁽³²⁾. Thus, NSAIDs are unlikely to be the underlying cause of terminal ileitis in our patient. Bacterial gastrointestinal infections may also trigger mild or asymptomatic ileal inflammation ⁽²⁵⁾; however, none were identified in our patient.

Infliximab has proved effective for reducing inflammatory conditions in patients with rheumatic disorders and for healing perianal fistulas in CD ^(13,33,34). In combination with concomitant therapy, infliximab was effective for CD fistula closure and for bringing about a dramatic resolution of joint manifestations in our patient. This clearly reflects the central role of TNF-α and the etiopathogenetic link between ERA and CD.

In conclusion, we describe the first case of ERA complicated by CD of which we are aware, and we provide new insights into the pathophysiology of both disorders and assume that the prevalence of both concomitant diseases in childhood is rare. Underlying immune dysfunction in addition to genetic predisposition may play a role in the development of both inflammatory disorders. More work is required on this subject. Our clinical perspective suggests the diagnosis of new-onset CD in a patient with ERA rather than an exacerbation of already established CD with joint manifestations.