Crohn's disease (CD) is a chronic, relapsing inflammatory disorder potentially involving the entire gastrointestinal tract. The chronic inflammation in genetically susceptible individuals is probably related to a disturbed interaction between the intestinal microbiome and the immune system.1, 2 Its natural history is not well defined and generally difficult to predict, with some patients remaining in prolonged remission, while others have recurrent episodes of active disease.3 The chronic inflammatory state and recurrent relapses can lead to major bowel damage, such as strictures and fistulas, often requiring surgical interventions and severely impairing patients' quality of life.
The major objective of medical therapy in the care of CD is to induce and to maintain mucosal healing leading to long-term remission with normal life quality and optimal growth in children. The recently updated Cochrane Reviews4, 5 show that thiopurines (azathioprine [AZA] and its metabolite 6-mercaptopurine [6-MP]) are effective in inducing and maintaining remission in adult CD patients, supporting its wide use in the treatment of inflammatory bowel diseases (IBDs).6 Individual randomized clinical trials in adult cohorts indicate that thiopurines are efficacious in maintaining remission in patients with CD7–10 with remission rates of 40%–70%, at 1 year of follow-up and a number needed to treat of four to six patients for preventing one relapse. Few data exist in pediatric populations11–13: the only pediatric randomized multicenter placebo-controlled trial12 confirmed the efficacy of thiopurines to spare steroids and to prevent disease relapse in children with CD. The maintenance of remission was reported to be 90% at 18 months, which is markedly higher compared to studies in adult populations. This might reflect a particular sensitivity of pediatric CD patients to AZA/6-MP or simply the clinical setting of this particular study. There are additional data indicating that AZA is more efficient in preventing disease relapse when administered early after disease onset.11, 13
Several side effects of thiopurine therapy were reported in patients with IBD, such as nausea, allergic reactions, flu-like illness, malaise, fever, rash, abdominal pain, pancreatitis, hepatotoxicity, myelosuppression, and an increased risk of lymphoma, mainly Epstein-Barr virus (EBV)-related.14–17 The occurrences of leukopenia and hepatotoxicity are probably related to inherited variations in the patient's ability to metabolize thiopurines, suggesting that thiopurine methyl transferase (TPMT) activity dosage is useful for predicting and preventing some of these complications.18, 19
Given the lack of data in the pediatric literature and the marked difference between the results of the study of Markowitz et al12 and trials in adult populations suggesting higher efficacy of AZA/6-MP in pediatric CD patients, we intended to analyze tolerance and efficacy of AZA to maintain remission in pediatric CD in a homogenous, single-center pediatric IBD cohort. This retrospective approach allowed to analyze the impact of early to late AZA introduction, the type of induction therapy (surgical, steroids, enteral nutrition), or different disease presentations on the efficacy of AZA maintenance therapy.
MATERIALS AND METHODS
- Top of page
- MATERIALS AND METHODS
The clinical charts of 105 CD patients on AZA maintenance therapy were reviewed in a retrospective manner. All children were followed in the pediatric IBD clinics at Necker–Enfants Malades Hospital, Paris, between January 1998 and January 2008. The diagnosis of CD was based on the Porto criteria20 with established clinical, endoscopic, radiological, and histological items. The patients' clinical data evaluating sex, age at diagnosis, disease location, active perianal disease (fissures, fistulas, abscess) were recorded and patients were classified according the Montreal criteria for location and disease behavior.23 The following parameters were monitored at treatment onset and all follow-up visits at months 6, 12, 18, and 24: pediatric Crohn's disease activity index (PCDAI), blood count (leukocytes, hemoglobin, hematocrit, and platelets), albumin, C-reactive protein, and erythrocyte sedimentation rate (ESR). We also recorded the type of induction therapy (enteral nutrition, steroids, parenteral nutrition, salicylates [Salazopyrine or 5-aminosalicylate [5-ASA] derivates), antibiotic use, resection-surgery, early or late introduction of AZA (before or later than 6 months from the diagnosis), and other concomitant treatments.
Screening criteria for eligibility to be included in this study were age: 6–17 years, AZA as maintenance treatment for CD with a follow-up of at least 2 years, AZA introduction at diagnosis or first relapse in patients with a PCDAI >30 (moderate-to-severe active disease21, 22) severe and extensive mucosal lesions, or in patients who were dependent on steroids or enteral nutrition. An additional indication was immunomodulator treatment after resection surgery. Upon screening for eligibility (n = 105) only CD patients with a complete chart and who responded to induction therapy with remission (PCDAI <10) were included in this study (n = 93). Induction therapy had to be limited to a maximum of 4 months, with reintroduction of food in patients on exclusive enteral nutrition after 4 months. Similarly, steroid medication had to be ≤0.3 mg/kg at 4 months. In the following, to evaluate efficacy of AZA remission was defined as a PCDAI ≤10 without any steroid medication and all patients were evaluated at 6, 12, 18, and 24 months. A relapse was defined as PCDAI >10 and/or when patients required anti-TNF treatment, other immunomodulators, reintroduction of exclusive enteral nutrition, or steroids. Also, patients who needed surgery were considered to experience a relapse if AZA medication was maintained at stable and optimal doses for at least 4 months.24 Only salycilate or antibiotic use was allowed. Since 2003, TPMT activity was tested in all patients prior to starting AZA. AZA was not introduced in the case of low TPMT activity (<5 nmol/h/mL erythrocytes confirmed by a subsequent genetic analysis revealing a mutation in one of the exons 5, 7, 10).
Analyses were performed on an intention-to-treat basis for the impact on outcome of different induction therapies (enteral nutrition versus steroids, history of surgery), early or late introduction of AZA, disease localization, disease behavior, surgical intestinal resection before introduction of AZA, and presence of perianal involvement. In addition, a Kaplan–Meier model was used to avoid potential underestimation of relevant factors impacting the response to AZA based on an intention-to-treat model.
The statistical differences between groups were evaluated using Student's t-test or chi-square test (Excel for Windows) as well as by the use of a survival model with Mann–Whitney testing (GraphPad Prism 4, San Diego, CA). P-values <0.05 were considered statistically significant.
- Top of page
- MATERIALS AND METHODS
There is good clinical evidence indicating that thiopurines (AZA and 6-MP) are efficient at maintaining remission and preventing relapses in patients with CD, as demonstrated in several placebo-controlled randomized trials in adult patients.4–10 Since the pivotal trial of Markowitz et al12 in a pediatric cohort of CD patients, the use of AZA/6-MP markedly increased in children with CD. This 18-month multicenter clinical trial analyzed the efficacy of 6-MP in 55 children with newly diagnosed moderate-to-severe CD. Patients were randomized to receive prednisone with either 6-MP or placebo as initial therapy. No synergistic effect on the induction of remission was noted, as ≈90% of both groups achieved remission within 3 months. However, during and after prednisone weaning only 9% of children in the 6-MP group relapsed subsequently, compared to 47% in the placebo group. Remission at 18 months was documented in 89% of the 6-MP group, compared with 39% in the placebo group. This study, along with reports of uncontrolled pediatric clinical experience,13, 25, 26 has led most pediatric gastroenterologists to prescribe commonly thiopurines for children with CD. A survey ending in 2008 within the pediatric IBD Porto group of ESPGHAN indicated that up to 70% of pediatric CD patients receive thiopurines within the first year of diagnosis (pers. unpubl. data). They may be initiated either at diagnosis, postoperatively, or after a severe relapse in corticosteroid-dependent or -resistant patients.
It is interesting to note, that clinical trials in adult populations showed a markedly lower efficacy of AZA/6-MP to maintain remission compared to the study of Markowitz et al.12 This contrast raises many questions, such as a particular sensitivity of pediatric patients to thiopurines, or the role of introducing AZA/6-MP early in the course of CD. Thus, the results of our retrospective and observational study are close to the clinical experience reported in adult patients. In our cohort, 60% of patients remained in remission on AZA monotherapy at M6, and this number dropped to 40% at M12. Patients, who were in remission at M12 usually remained in remission, even after 2 years of follow-up. This indicates a subgroup of patients with an excellent response to thiopurine.
One major finding of the present study is that 12% of our patients did not tolerate AZA medication. Thiopurine intolerance occurred within the first 4 weeks of treatment. This strengthens the recommendation of a particular surveillance at introduction of AZA/6-MP therapy. After 4 weeks of AZA medication, no further withdrawal was necessary in this study. Generally, AZA was well tolerated and no drug-related mortality was noted in this study. The major adverse reactions were infections, which needed a specific therapy and a transient suspension of AZA medication. No case of lymphoma or malignancies was observed in this pediatric population during the 2-year follow-up; however, long-term follow-up data are still largely missing in the pediatric literature.
The efficacy of thiopurines (AZA or 6-MP) in preventing postoperative recurrence in CD remains debated but there are several studies supporting this indication for thiopurines27, 28: These studies show a clinical remission rate of 83% in the AZA group and of 49% in the 6-MP group versus 22.5% in the placebo group at 2 years of follow-up. In our population, patients who had a surgical intervention prior to AZA introduction were more likely to maintain remission compared to patients on AZA who did not require surgery, indicating the potential to prevent relapses in this subgroup characterized by a particularly severe disease evolution. However, the number of patients in the present study is too small to obtain significant results.
We were not able to confirm that early introduction of AZA after initial diagnosis is more efficacious compared to later introduction (using a 6-month cutoff). In our analysis, the number of relapsing patients was comparable in both groups, in contrast to the registry data reported by Punati et al.11 However, when analyzing the data with a survival model, we observed a trend towards a better response to early AZA introduction in the time interval of 6 to 12 months. To address this point of early AZA/6-MP as a more efficacious treatment option, well-designed prospective studies are needed.
Another important issue, which could not be addressed in the present study, is the adherence of patients to AZA treatment. As in all clinical trials, treatment compliance markedly impacts the outcome. In the present work, no monitoring of adherence was possible due to the retrospective character of the study.
In summary, AZA therapy is efficacious in maintaining remission in pediatric CD patients; however, probably to a markedly lesser degree than previously suggested. It seems to be more efficient in preventing relapses after surgery especially if associated with a postsurgical period of enteral nutrition. Patients in remission after 12 months of follow-up are very likely to maintain long-term remission. Overall tolerance of AZA in this cohort was good, as all adverse reactions probably related to AZA medication were reversible.