Statistics from Altmetric.com
We read with interest the recent article by Lowry et al (Gut 2001;49:665–70) on the role of thiopurine methyltransferase (TPMT) activity and thiopurine metabolites in patients with inflammatory bowel disease (IBD). Lowry et al concluded that 6-thioguanine nucleotide (6-TGN) concentrations do not correlate with disease activity or leucocyte counts. Although in general terms this is a well designed trial and includes a large number of patients, analysis of the results presents a number of problems.
Firstly, the results for metabolite concentrations and metabolic pathways deserve more attention. A number of competitive enzymes, including TPMT, are involved in the complex multistep thiopurine metabolism and result in the synthesis of varying concentrations of active or inactive end metabolites which have both toxic and non-toxic properties.1 Secondly, thiopurine users can be broadly divided into clinically overlapping responders, non-responders, and therapy intolerant groups which in turn correspond to concentrations of these end metabolites.2, 3 Exclusion of any of these end metabolite dependent patient subpopulations can significantly influence study outcome.
Failure to find a significant association between thiopurine related toxicity or disease activity and 6-TGN concentrations by Lowry et al could be explained by the median whole blood 6-TGN concentration (136 pmol/8×108 red blood cells (RBC)) in the study population. This is well below the pharmacologically suggested therapeutic concentration of 230 pmol/8×10 RBC.2, 3 In a separate study, Dubinsky and colleagues4 were able to show that significantly high isolated 6-TGN concentrations were not related to toxicity. They in fact demonstrated a clinical improvement with increasing 6-TGN concentrations while toxicity was directly proportional to increasing 6-methyl mercaptopurine ribonucleotide (6-MMPR) concentrations. Lowry et al ignored this important preferential metabolic pathway observed in certain patient subpopulations.
We feel that carefully designed future trials are needed for further elaboration of these interacting factors by inclusion of all patient subpopulations and consideration of important metabolites. Alternatively, 6-thioguanine, a thiopurine compound not subject to the 6-MMPR catabolic pathway, may offer a more realistic approach to study the correlation between 6-TGN concentrations and disease activity in IBD patients.4
There are numerous factual inaccuracies in Qasim et al's letter which merit response. Firstly, they state that “A number of competitive enzymes, including thiopurine methyltransferase (TPMT), are involved in the complex multistep thiopurine metabolism and result in the synthesis of varying concentrations of active or inactive end metabolites which have both toxic and non-toxic properties”. They reference a nine year old review article for this statement.1 While there is no argument that the metabolism of azathioprine and 6-mercaptopurine is complex, and that it involves multiple enzymes, there is very little valid information about which metabolites are toxic or non-toxic. The clear bias of Qasim et al is that various metabolites are either active or inactive and that some have toxic properties while others are non-toxic, but this has not been definitively proved in statistically valid clinical studies.
Secondly, Qasim et al state that “thiopurine users can be broadly divided into clinically overlapping responders, non-responders, and therapy intolerant groups, which in turn correspond to concentrations of these end metabolites. Exclusion of any of these end metabolite dependent patient subpopulations can significantly influence study outcome”. The two papers quoted have exactly the same limitations as the current study (exclusion of patients who received azathioprine or 6-mercaptopurine for less than four months, and exclusion of patients with leucopenia or increased serum hepatic or pancreatic enzymes before initiation of therapy).2–4 As nearly half of the cases of thiopurine related leucopenia, most cases of elevated liver transaminases, and virtually all cases of pancreatitis and allergic reactions occur within the first four months of therapy, the study design for both of these studies would have excluded many patients who were intolerant to azathioprine or 6-mercaptopurine. Thus it is unlikely that the exclusion criteria in our study, which were virtually the same as in the studies of Cuffari and colleagues2 and Dubinsky and colleagues,3 can account for the differences in the results and conclusions.
Thirdly, Qasim et al state that the lack of an association between thiopurine toxicity or disease activity and 6-thioguanine nucleotide (6-TGN) concentrations in our study could be explained by low whole blood 6-TGN concentrations. The authors should refer to paragraph 4 of the discussion on 699 of our paper that indicates that our group has performed a study that shows a strong direct correlation between the assay used by Cuffari and Dubinsky and the assay used in our study (unpublished data). A conversion factor of 1.6 is required to convert the results of our study to those of the Cuffari and Dubinsky studies. Once that conversion factor is applied, the results of our study would yield a mean red blood cell (RBC) 6-TGN concentration of 261 pmol/8×108 RBC, which is well above the pharmacologically suggested therapeutic concentration of 230 pmol/8×108 RBC referred to by Qasim et al. Thus, as we clearly stated in the discussion of our paper, differences in assay methodology are unlikely to account for the discrepant results.
Finally, Qasim et al state that “In a separate study, Dubinsky et al were able to show that significantly high isolated 6-TGN concentrations were not related to toxicity. They in fact demonstrated a clinical improvement with increasing 6-TGN concentrations while toxicity was directly proportional to increasing 6-methylmercaptopurine ribonucleotide concentrations. Lowry et al ignored this important preferential metabolic pathway observed in certain patients subpopulations.” We disagree strongly with this statement. The study by Dubinsky et al was an uncontrolled open label pilot study in 10 patients with Crohn's disease.5 It is unreasonable to compare the results of our study (a study involving 170 patients) with the preliminary results from an open label pilot study in 10 patients.
We agree with Qasim et al on one point—that is, that a prospective trial is needed in which patients who are candidates for thiopurine therapy are randomly assigned to full dose azathioprine (2–3 mg/kg/day) or 6-mercaptopurine (1.5 mg/kg/day) with downward adjustments in dose, as clinically indicated for toxicity, versus dose adjustment to a target 6-TGN concentration range. The issue of patient subpopulations based on metabolite profiles is very preliminary (one published study in 10 patients) and this hypothesis needs further characterisation before a controlled trial can even be designed. In the meantime, routine measurement of RBC 6-TGN in patients being treated with azathioprine or 6-mercaptopurine is not necessary.6
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.