• 6-mercaptopurine;
  • azathioprine;
  • pediatric;
  • dose escalation;
  • immunomodulator


  1. Top of page
  2. Abstract

Background: 6-Mercaptopurine (6-MP) and its prodrug azathioprine (AZA) are effective for the induction and maintenance of remission and reduction of corticosteroid exposure for pediatric inflammatory bowel disease (IBD). The standard dose of 6-MP is 1.0–1.5 mg/kg/day and for AZA is 2.0–2.5 mg/kg/day. The aim of this study was to determine whether IBD patients 6 years of age and younger require higher than standard doses of 6-MP/AZA to achieve clinical remission.

Methods: Clinical data was collected retrospectively for all IBD patients 6 years of age or younger treated with 6-MP/AZA at The Children's Hospital of Philadelphia.

Results: Thirty patients met the inclusion criteria. IBD was diagnosed at a median age of 3.3 years (25–75th %ile 2.3–4.6 years) and 6-MP/AZA was initiated at a median age of 3.9 years (range 0.8–6.8 years). After dose escalation, the median AZA-equivalent dose was 3.1 mg/kg/day (25–75th %ile 2.5–3.5, max. dose 5.1 mg/kg/day). At the final recorded dose, 8/13 (62%) patients receiving AZA >3.0 mg/kg/day achieved clinical remission, compared to 2/12 (17%) receiving 2–3 mg/kg/day (P = 0.02). The risk of having active disease was on average 85% lower if the AZA-equivalent dose was >3.0 mg/kg/day (95% confidence interval [CI] 72%–93%). Adverse events were experienced by 4/30 patients (hepatitis, n = 2; leukopenia, n = 2). No patients had to discontinue 6-MP/AZA, and all laboratory abnormalities improved spontaneously or with dose reduction.

Conclusions: The standard dose of 6-MP/AZA may not be adequate for IBD patients 6 years of age and younger. Closely monitored dose escalation beyond the standard dosing range is effective and well-tolerated.

(Inflamm Bowel Dis 2008)

6-Mercaptopurine (6-MP) and its prodrug azathioprine (AZA) are potent immunomodulators that have become a standard component of therapy for inflammatory bowel disease (IBD). They are effective for the induction and maintenance of remission and have been shown to contribute to a reduction in corticosteroid exposure for both Crohn's disease (CD) and ulcerative colitis (UC).1–4 6-MP and AZA are generally well tolerated, although dose-related adverse reactions such as hepatitis and myelosuppression are recognized to occur in up to 2%–8% of patients, while dose-independent hypersensitivity reactions such as fever and pancreatitis each occur in approximately 4% of patients.5, 6

6-MP is a thiopurine that is converted through a series of reactions to its active metabolites, 6-thioguanine nucleotides (6-TGNs).7 6-TGNs interfere with both DNA and RNA synthesis, exerting an antiproliferative effect on active lymphocytes.8 Additionally, 6-TGNs are thought to inhibit cytotoxic T-cell and natural killer (NK) cell function9 and induce apoptosis of T cells.10 Dubinsky et al11 concluded that erythrocyte 6-TGN levels are significantly associated with therapeutic response in IBD, with 6-MP efficacy optimized at 6-TGN levels greater than 235 pmol/8 × 108 red blood cells (RBC). Although subsequent studies have demonstrated conflicting results regarding the correlation between 6-TGN levels and clinical response to 6-MP/AZA,12–14 a recent meta-analysis supported the finding that higher 6-TGN levels are associated with clinical remission.15

The established target daily doses for 6-MP and AZA in inflammatory bowel disease are 1.0–1.5 mg/kg and 2.0–2.5 mg/kg, respectively.16 It was clinically observed that, when treated with standard doses of 6-MP/AZA, our youngest IBD patients experience poor clinical response and do not achieve therapeutic 6-TGN levels. The primary objective of this study was to determine whether IBD patients who are 6 years of age and younger require higher than standard doses of 6-MP/AZA to achieve clinical improvement and remission. The secondary objective was to determine if these patients are more likely to achieve 6-TGN levels >235 pmol/8 × 108 RBC when they receive higher than standard doses of 6-MP/AZA.


  1. Top of page
  2. Abstract

A retrospective chart review of all patients 6 years of age and younger who were treated with 6-MP/AZA for IBD at The Children's Hospital of Philadelphia from January 1, 2001 through November 30, 2006 was performed. Patients were identified by a search of our clinical database for the corresponding diagnosis codes for CD, UC, and indeterminate colitis (IC) as well as by personal inquiry of all practitioners. The inclusion criteria were an endoscopically confirmed diagnosis of CD, UC, or IC and age 6 years old or younger at the time of 6-MP/AZA initiation. The exclusion criterion was inadequate records. During the study period, there was no standardized protocol for 6-MP/AZA administration or monitoring, so medication dosing and 6-TGN measurement were performed at the discretion of each clinician.

All outpatient and inpatient charts were reviewed by a single abstractor using a standardized form. Data were abstracted from visits that corresponded to the initiation of 6-MP/AZA therapy, at 3, 6, and 12 months after start of therapy, and at any additional visits when a 6-TGN level was measured, assuming the patient was still 6 years of age or younger on the visit date. Relevant clinical data were recorded, including patient age, diagnosis, age at diagnosis, height and weight, thiopurine methyltransferase (TPMT) phenotype/genotype (if available), 6-MP/AZA dose and formulation, 6-TGN and 6-MMP levels (if available), adverse effects, clinical status of disease, laboratory results, and concomitant IBD medications. The AZA dose by weight (mg/kg) was calculated; for patients receiving 6-MP, an “AZA equivalent dose” was determined by multiplying the 6-MP dose by a conversion factor of 2.08.17 Clinical disease activity was assessed by physician global assessment (PGA) and an abbreviated Pediatric Crohn's Disease Activity Index (aPCDAI), which includes measurements of abdominal pain, stool characterization, general well-being, weight trend, abdominal examination, perianal disease, and extraintestinal manifestations of IBD.18, 19 A “final AZA-equivalent dose” was defined as the patient's dose at the time of the final recorded visit.

Patient demographic characteristics are presented with descriptive statistics. The effect of dose escalation was determined by dividing patients into standard dosing (2–3 mg/kg/day) and higher dosing (>3.0 mg/kg/day) groups based on the final AZA-equivalent dose at their last visit. Group comparisons were made by Fisher's exact and Mann–Whitney U-tests. Patients were excluded from this analysis if adverse events prevented dose escalation or if they never received the minimum recommended dose for patients with normal TPMT activity of AZA ≥2.0 mg/kg/day. The influence of 6-MP/AZA dose, dose escalation, and 6-TGN levels on disease activity over time was determined by generalized estimating equation (GEE) models. Patients were excluded from the GEE models if adverse events prevented dose escalation. For all statistical analyses remission was defined as an aPCDAI ≤10 (for CD and IC patients only), physician global assessment of inactive disease, steroid dosing <0.25 mg/kg/day, and no concurrent infliximab therapy. Adverse events associated with 6-MP/AZA are documented qualitatively. All statistical analyses were performed with Stata 9.0 (College Station, TX).

This study was approved by the Institutional Review Board of The Children's Hospital of Philadelphia.


  1. Top of page
  2. Abstract

A total of 105 children diagnosed with endoscopically confirmed CD, UC, or IC at age 6 years or younger were evaluated at our center during the study period. Forty-three of these patients began 6-MP/AZA therapy when 6 years of age or younger. Thirty patients met the inclusion criteria, while 13 were excluded from this study because of inadequate records (10 were evaluated only once for a second opinion, 1 discontinued 6-MP prior to transfer of care to our center, and 2 were lost to follow-up.) The 30 patients included in this study were treated by 10 clinicians.

The median age at time of IBD diagnosis was 3.3 years (25th–75th %ile 2.3–4.6 years). Other baseline demographic data are presented in Table 1. TPMT phenotypic or genotypic results were available for 25/30 (83%) patients; 24/25 (96%) had normal enzymatic activity or genotype while 1/25 (4%) demonstrated intermediate enzymatic activity. The median age at initiation of 6-MP/AZA was 3.9 years (25th–75th %ile 3.0–5.9 years, range 0.8–6.8 years). The median time from IBD diagnosis until initiation of 6-MP/AZA was 6 months (25th–75th %ile 3–11 months). Seventeen patients (57%) were treated with 6-MP and 13 (43%) with AZA. Six patients (20%) received pills and 9 (30%) liquid formulation, while the method of administration was unclear for 15 (50%). The median initial AZA-equivalent dose was 2.2 mg/kg/dose (25th–75th %ile 1.6–2.6 mg/kg/dose). The median follow-up period from initiation of 6-MP/AZA was 12.6 months (range 4–59 months).

Table 1. Baseline Characteristics (n=30)
  • a

    N=27, 3 UC patients not included.

  • b

    N=27, missing baseline data for 3 patients.

Age at Diagnosis (Years, 25th−75th %ile)3.3 (2.3–4.6)
Disease classification (n, %) 
 Crohn's disease16 (53%)
 Indeterminate colitis11 (37%)
 Ulcerative colitis3 (10%)
Crohn's disease location (n, %) 
 Upper GI involvement10 (63%)
 Ileocolonic9 (56%)
 Colonic7 (44%)
Gender (n, %) 
 Male19 (63%)
 Female11 (37%)
Other medical treatment (n, %) 
 Corticosteroids18 (60%)
 5-ASA21 (70%)
 Antibiotics4 (13%)
 Infliximab3 (10%)
Disease activity 
 aPCDAI (median, 25th−75th %ile)a7.5 (1.3–15)
 Active disease by PGA (n, %)b19 (70%)
Labs (median, 25th−75th %ile) 
 WBC [5.5–15.5 103/μL]9.3 (8.3–13.4)
 Hemoglobin [11.5–15.5 g/dL]11.0 (10.0–12.7)
 ESR [0–20mm/hr]12.5 (4.3–33)
 CRP [0–1 mg/dL]0.8 (0.5–3.1)
 Albumin [3.5–5.2 g/dL]4.1 (3.6–4.4)

Effect of Dose Escalation

Three patients were excluded from the group analysis because adverse events prevented further dose escalation, and 2 were excluded because they were subtherapeutically dosed based on study criteria (AZA <2.0 mg/kg/day). For the 25 remaining patients the final AZA-equivalent dose was 3.1 mg/kg/day (25th–75th %ile 2.5–3.5, max. 5.1 mg/kg/day). The mean dose increase from 6-MP/AZA initiation until the final visit was 0.9 mg/kg, with 13/25 (52%) patients requiring higher than standard doses (Table 2). 6-TGN and 6-MMP levels were higher for those patients who received >3.0 mg/kg/day final AZA-equivalent dose as compared to those who received <3.0 mg/kg/day. Five patients (20%) achieved 6-TGN levels >235 pmol / 8 × 108 RBC at their final visit, all from the >3.0 mg/kg/day dosing group.

Table 2. Baseline and Final AZA Dose Group Comparisons
 AZA 2–3 mg/kg/d N=12AZA >3 mg/kg N=13
  • a

    P=0.02 (difference between baseline and final visits).

  • b

    P=0.002 (difference between AZA 2-3mg/kg/day and AZA > 3 mg/kg/day).

  • c

    P=0.02 (difference between AZA 2-3mg/kg/day and AZA > 3 mg/kg/day).

Follow-up from 6-MP/AZA start (months) (median, 25–75th %ile)11.4 (7.4, 17.0)13.2 (6.8, 19.4)
Final AZA-equivalent dose (mg/kg/d) (median, 25–75th %ile)2.46 (2.33, 2.73)3.51 (3.25, 4.08)
aPCDAI (Median, 25–75th %ile)  
 Baseline10 (5, 15)15 (5, 20)
 Final visit10 (5, 15)0 (0, 5)a
 Median difference0 (−10, 10)−5 (−20, 0)
Inactive disease by PGA (n)  
 Final visit312b
Remission at last F/U (n)  
 Final visit28c
(Median, 25–75th %ile)154 (103, 186)219 (180, 375)
(Median, 25–75th %ile)1007 (521, 2485)2585 (1334, 5094)
 Baseline (number)78
  Dose mg/kg (median, 25–75th %ile)0.14 (0, 0.69)0.93 (0, 1.3)
 Final visit (number)43
  Dose range mg/kg0.38–1.00.38–1.25
 Final visit43
 Baseline (median, 25–75th %ile)10.7 (7.6, 13.8)9.4 (8.7, 12.2)
 Final visit (median, 25–75th %ile)6.9 (6.5, 13.5)7.3 (5.7, 9.6)
 Baseline (median, 25–75th %ile)10.8 (9.8, 12.3)10.9 (10,0, 13.2)
 Final visit (median, 25–75th %ile)11.6 (11, 12.4)12.4 (10.9, 12.9)
 Baseline (median, 25–75th %ile)14 (3.5, 33.5)14.5 (3.8, 48.3)
 Final visit (median, 25–75th %ile)20 (9.5, 32.8)12.5 (6.3, 49.5)
 Baseline (median, 25–75th %ile)0.7 (0.3, 0.7)3.1 (1.1, 6.3)
 Final visit (median, 25–75th %ile)1.1 (0.7, 3)0.6 (0.3, 3.3)
 Baseline (median, 25–75th %ile)4.2 (3.7, 4.5)3.9 (3.1, 4.2)
 Final visit (median, 25–75th %ile)4.2 (4.0, 4.3)4.4 (4, 4.7)

At the final visit 12/13 (92%) from the AZA >3.0 mg/kg/day group had a physician global assessment of inactive disease, compared to only 3/12 (25%) from the AZA 2–3 mg/kg/day group (P = 0.002, Fisher's exact test). Additionally, 8/13 (62%) from the >3.0 mg/kg/day dosing group met the criteria for clinical remission, compared to 2/12 (17%) from the AZA 2–3 mg/kg/day group (P = 0.02, Fisher's exact test). There was a trend toward a greater decrease in aPCDAI score from baseline to final visit in the AZA >3.0 mg/kg/day group compared to the 2–3 mg/kg/day group, but this did not reach statistical significance (P = 0.13, Mann–Whitney U-test). There were no significant differences between groups for any measured laboratory parameters, the percentage of patients who were successfully weaned off of steroids, or the percentage of patients who required infliximab (Table 2).

Three patients were excluded from the GEE models because adverse events prevented dose escalation and 2 patients were excluded as they were significant outliers (only 1 recorded AZA-equivalent dose per patient of 0.42 mg/kg and 4 mg/kg, respectively). The GEE models demonstrated that the risk of having active disease was on average 85% lower if the AZA-equivalent dose is >3 mg/kg/day (95% confidence interval [CI] 72–93%). Additionally, the risk of active disease was on average 67% lower for each 1 mg/kg increase in the AZA-equivalent dose (95% CI 47–80%). For our cohort, the risk of active disease was on average 70% lower if the 6-TGN level was greater than 235, but this did not reach statistical significance (95% CI 0.10–1.09). Since few patients achieved a 6-TGN level >235, the sample size might have been too small for meaningful analysis of the effect of this variable.

Adverse Events

Adverse events from 6-MP/AZA were reported for 4/30 (7%) patients. Two patients experienced leukopenia and 2 patients had transaminase elevation (Table 3); no serious adverse events were reported. All 4 patients with adverse events had normal TPMT activity. Both cases of leukopenia and 1 case of hepatitis occurred after dose escalation; the other case of hepatitis occurred without deviation from the original dosing schedule. No patients had to discontinue 6-MP/AZA due to adverse events, although 3 did not tolerate dose escalation deemed necessary by their clinicians.

Table 3. Adverse Events: Leukopenia (2/31) and Hepatitis (2/31)
WBC (U/L)ALT/AST (103/μL)Duration of Therapy (mo)AZA-Equivalent Dose (mg/kg)6-TGN6-MMPTPMTOutcome
  • Enzyme activity not measured, but normal genotype.

2.1 7.33.45367570034.1Normalized with decreased dose
2.7 5.83.17180292628.7Normalized with deceased dose
 216/1515.52.55160579641.7Normalized; no dose escalation
 221/115174.521887485*1/*1Normalized with decreased dose


  1. Top of page
  2. Abstract

This is the first study to describe increased 6-MP/AZA dosing requirements for younger pediatric IBD patients. Our data suggest that the standard AZA dose of 2.0–3.0 mg/kg/day, and the equivalent 6-MP dose, is not adequate for IBD patients 6 years of age and younger. The patients from our study who were escalated to doses greater than AZA 3.0 mg/kg/day were significantly more likely to achieve clinical remission than those who remained on a standard dosing regimen. Additionally, these higher doses were generally well tolerated. Leukopenia and transaminase elevation were each experienced by 7% of our patients, all resolving either spontaneously or after 6-MP/AZA dose decrease. No serious adverse events were reported despite escalation of AZA-equivalent doses to as high as 5.1 mg/kg/day.

Previous adult studies have investigated the effect of increasing the 6-MP/AZA dose for nonresponders. Cuffari et al12 showed that 18/22 adult patients who were previously failing AZA therapy achieved remission after a dose increase of 25 mg, although the majority were initially dosed below the recommended range. In all, 67% of the responders had therapeutic 6-TGN levels after dose adjustment. Dubinsky et al20 retrospectively demonstrated that over 25% of the primarily adult patients who were failing 6-MP/AZA and had subtherapeutic 6-TGN levels achieved therapeutic efficacy with dose escalation. The majority of the responders were still receiving standard doses of 6-MP/AZA after dose escalation. Clinical improvement with dose escalation did correlate with an increase in 6-TGN levels, while the 6-TGN levels did not increase for most of the nonresponders to dose increase. No studies have described the effect of dose increase for exclusively pediatric IBD patients.

Most of the patients in our study, regardless of final dose, did not achieve final 6-TGN levels >235 pmol / 8 × 108 RBC. This is not dissimilar to results from a recent cross-sectional study of adult IBD patients treated with 6-MP, where only 34.7% of adult IBD patients treated with standard dose 6-MP and 47.1% who were treated with higher doses achieved 6-TGN levels >230 pmol / 8 × 108 RBC.21 In our study, however, all final 6-TGN levels in the therapeutic range occurred in the patients receiving AZA >3 mg/kg/day.

The etiology for the increased dosing requirement of 6-MP/AZA for IBD patients 6 years of age and younger is unclear. Variability in the bioavailability and degree of intestinal absorption of 6-MP/AZA are well documented; the oral bioavailability of 6-MP is less than 20% in pediatric oncology patients.22, 23 We hypothesize that decreased medication absorption in this age group is a likely explanation. Another possibility is that there are age-related differences in the metabolism of 6-MP/AZA with decreased production of 6-TGN, resulting in decreased therapeutic efficacy. For example, a proportion of primarily adult 6-MP/AZA nonresponders have been discovered to preferentially shunt 6-MP metabolism toward the production of 6-MMP, resulting in decreased 6-TGN levels that do not increase with dose escalation.20 Only 1 patient from our cohort had a minimal increase in 6-TGN level but rapid increase in 6-MMP levels with dose escalation. Most of our patients exhibited decreased levels of both 6-TGN and 6-MMP, suggesting that decreased absorption is a more likely explanation than differences in drug metabolism. Additionally, TPMT activity or genotype was available for 80% of our patients; all except 1 (4%) demonstrated normal enzymatic activity or genotype, which is similar to expectations for an adult sample. Whether the administration of 6-MP/AZA in liquid or pill form affects the therapeutic efficacy needs to be investigated. Unfortunately, we could not obtain information regarding the formulation of administration for half our patients, precluding our ability to analyze the effect of this variable.

A potential limitation of our study is the retrospective design, with all data collected as part of routine clinical practice. 6-MP/AZA administration, particularly the willingness to escalate to higher than standard doses, varied among practitioners, as did metabolite monitoring. The effect of other medications or interventions on achieving clinical remission was also difficult to assess due to the retrospective nature. However, we tried to control for this by requiring that our subjects could be on only very low-dose steroids (<0.25 mg/kg/day) and could not be on concurrent infliximab therapy to be considered in remission. Compliance is always difficult to assess, particularly with a retrospective study. However, we hypothesize that noncompliance is less prevalent in this age group, as the patients are dependent on their parents for medication administration. Another limitation is the small sample size, which was a consequence of the very specific patient population we described. Future prospective studies are needed to confirm these observations.

In summary, we observed that standard weight-based dosing of 6-MP/AZA is inadequate for many of our IBD patients 6 years of age and younger and that dose escalation results in improved clinical response. Dose escalation was also well tolerated. We suggest that if an IBD patient in this age group is a nonresponder to a standard dose of 6-MP/AZA, and 6-TGN and 6-MMP metabolites are low and compliance has been assessed, then carefully monitored dose increases beyond the standard range can be implemented prior to determining that a patient has failed this therapy. This might result in fewer younger IBD patients requiring initiation of infliximab, which is particularly desirable in lieu of recent concerns regarding its safety in pediatric patients.24


  1. Top of page
  2. Abstract