Introduction

Thiopurine drugs, 6-Mercaptopurine (6-MP) and its prodrug azathioprine (AZA), are effective for the induction and maintenance of remission in patients with inflammatory bowel disease (IBD).1, 2 There has been tremendous interest with regard to thiopurine metabolism as a means of identifying ways in which individual therapy might maximize clinical response and minimize adverse effects. Genetic polymorphisms in the thiopurine S-methyl transferase (TPMT) gene have been associated with decreased TPMT activity and the development of myelotoxicity due to high thioguanine metabolite concentrations.3, 4 Moreover, inosine triphosphate pyrophosphatase (ITPA) deficiency has recently been associated with AZA toxicity because of the accumulation of 6-thioinosine-triphosphate.5, 6 Theoretically, low TPMT activity might also augment the therapeutic efficacy of thiopurines and lead to better long-term clinical outcomes. To confirm the usefulness of the TPMT/ITPA genotype or TPMT activity in clinical prediction, research regarding the relevance of measurement for prevention of both adverse effects and clinical outcomes is necessary. However, the impact of the TPMT/ITPA genotype on long-term clinical response of IBD patients treated with AZA/6-MP has not yet been examined. In this study, we investigated whether TPMT/ ITPA genotypes could affect long-term clinical outcomes as measured by disease relapse in Korean patients with IBD undergoing AZA or 6-MP treatment.

Materials and methods

A total of 204 patients with IBD (105 patients with Crohn's disease, 59 with ulcerative colitis and 40 with intestinal Behcet's disease) were initiated on AZA treatment during the study period from June 1999 to November 2008 at Severance Hospital (Seoul, Korea). Patients were initially administered AZA at a fixed dose of 50 mg daily. The dose of AZA might be further increased to 2–3.5 mg kg−1 every 2 weeks if no obvious therapeutic response was observed, provided the patient continued to tolerate therapy and white blood cell counts remained >3 × 103 per l and serum alanine transaminase levels were less than twice the upper limit of normal (50 IU l−1). If the patient could not tolerate AZA because of adverse effects such as severe gastrointestinal symptoms, 6-MP was substituted for AZA. 6-MP was initially started at 25 mg daily. Similarly, the dose was increased to 1–1.5 mg kg−1 every 2 weeks if no adverse effects such as neutropenia and hepatotoxicity were observed.

Pretreatment and posttreatment clinical data were recorded. The investigators evaluating patients' clinical data were blinded to the results of TPMT and ITPA genotypes. Disease activity was evaluated using the Crohn's disease activity index for Crohn's disease and the partial Mayo score for ulcerative colitis.7, 8 Patients with intestinal Behcet's disease were considered to have achieved remission or to have relapsed as described previously.9

Blood samples for TPMT/ITPA genotyping were collected from patients at the initial clinic visit. Genomic DNA was extracted from blood after informed consent was obtained. Single-nucleotide polymorphisms in the TPMT (exon 5, 7, 10) and ITPA (exon 2) genes flanking their exon–intron junctions were identified using denaturing high-performance liquid chromatography and subsequent direct sequencing as previously described with some modifications.10, 11 At the beginning of this study, we conducted a preliminary study by sequencing TPMT/ITPA genes with pilot study samples to discover other polymorphisms, but found that minor allele frequencies were <5%, which was considered clinically insignificant. Relapse-free survival (time to first relapse) according to TPMT/ITPA genotypes was compared by Kaplan–Meier curves using log-rank analysis. Ethical approval for this study was obtained from the Human Research Review Board of Severance Hospital.

Results

Table 1 shows the baseline characteristics of the 204 patients enrolled in the study. All patients were initially treated with AZA. However, six patients were unable to tolerate AZA because of adverse effects such as severe gastrointestinal trouble, and were eventually treated with 6-MP. TPMT and ITPA genotypes were assessed in all 204 patients (Table 2). Of the 204 patients studied, 13 were lost to follow-up or discontinued study medications because of side effects. Of the 191 remaining patients who had available medical records regarding clinical response, 164 achieved clinical remission. Of the 164 patients who achieved a remission, 162 received AZA or 6-MP for ⩾3 months and, were thus included in the final analysis of clinical outcomes in terms of relapse. Starting AZA dose, maximum AZA dose and duration of AZA use were similar between the ‘non-relapsers’ group and the ‘relapsers’ group (data not shown). To compare clinical outcomes between wild-type and mutant TPMT and ITPA, Kaplan–Meier survival curves using log-rank analysis were constructed on the basis of the time to first relapse in IBD patients who had received AZA or 6-MP for ⩾3 months (Figure 1). There were no significant differences in disease relapse-free survival between wild and mutant types of TPMT (P=0.903) or ITPA (P=0.392) according to the results of the log-rank analysis. Similarly, on subgroup analysis, there was no significant difference in disease relapse-free survival between wild and mutant types of TPMT or ITPA on each disease entity (P=0.787 or P=0.775 for Crohn's disease; P=0.883 or P=0.150 for ulcerative colitis; and P=0.526 or P=0.961 for intestinal Behcet's disease).

Table 1 Characteristics of patients enrolled in the study
Full size table
Table 2 Genotype frequencies of TPMT and ITPA by disease
Full size table
Figure 1
figure 1

Time to first relapse in IBD patients treated AZA or 6-MP according to the TPMT (a) or the ITPA (b) genotype. IBD, inflammatory bowel disease; AZA, azathioprine; 6-MP, 6-Mercaptopurine; TPMT, thiopurine methyltransferase; ITPA, inosine triphosphate pyrophosphatase.

Full size image

Discussion

To confirm the usefulness of the TPMT/ITPA genotype in clinical prediction, research regarding the relevance of measurement for prevention of both adverse effects and clinical outcomes is necessary. To our knowledge, this is the first study that is specifically focused on the relationship between TPMT/ITPA genotypes and clinical outcomes during thiopurine treatment.

We found that neither the TPMT nor the ITPA genotype influenced disease relapse in IBD patients treated with thiopurines. In Asian populations, including Koreans, the frequencies of TPMT mutation are lower than those reported for Western countries, but myelotoxicity develops more frequently.12, 13, 14 Therefore, determination of the TPMT genotype before the initiation of treatment does not seem to be of significant use for predicting adverse effects with AZA in Asian populations. On the contrary, several studies in Western samples have addressed the impact of TPMT activity and genotype on the clinical response of IBD patients treated with AZA/6-MP and suggested that measurement of TPMT activity might predict clinical response to AZA/6-MP.15, 16 At present, the reasons for discrepancies among these studies are unclear. However, one possible explanation for these discrepancies is that the Asian population has genetic backgrounds of TPMT that are different from those of the Caucasian population. On the basis of the international HapMap data, we compared constructed linkage disequilibrium structures of both TPMT and ITPA genes between Asians and Caucasians, which revealed that definite ethnic diversity was shown in the TPMT gene but not in the ITPA gene (Supplementary Figure 1).

With respect to ITPA, the allele frequency of the ITPA 94C>A mutation in Korean IBD patients was 0.140, higher than that found in Europeans (0.060–0.070) and similar to the frequencies found in Japanese (0.135) and Chinese (0.15) populations.17, 18 Nevertheless, we did not find any evidence of an association between the ITPA genotype and clinical outcomes. Further research is warranted to reveal the true impact of TPMT/ITPA on the clinical efficacy of thiopurines.

In conclusion, our study suggests that TPMT/ITPA genotypes do not predict relapse in IBD patients using thiopurine treatment. Further large-scale studies should be carried out to validate the usefulness of TPMT/ITPA genotypes for guiding clinicians formulating individualized treatments for patients requiring thiopurine therapy for the treatment of IBD.

Conflict of interest

The authors declare no conflict of interest.