Mercaptopurine metabolite results in clinical gastroenterology practice


Correspondence to: Dr R. S. Bloomfeld, Section of Gastroenterology, Wake Forest University Baptist Medical Center, Medical Center Boulevard, Winston-Salem, NC 27157, USA. E-mail:


Background : Azathioprine (AZA) and its active metabolite mercaptopurine (MP) are frequently used in the management of inflammatory bowel disease. Measurement of the AZA/MP metabolites, thioguanine (TG) and methylmercaptopurine (MMP), has been suggested as a means to optimize therapy with AZA/MP in inflammatory bowel disease.

Aim : To evaluate the results of initial AZA/MP metabolite panels sent by gastroenterologists during the first year of its widespread availability.

Methods : Initial AZA/MP metabolite panels sent by gastroenterologists to a single laboratory were reviewed and the metabolite panels were interpreted.

Results : Initial metabolite levels were reviewed for 9187 patients. Noncompliance was detected in 263 patients (3%) and under-dosing in 4260 patients (46%). 534 patients (6%) had levels that were consistent with preferential metabolism via the TPMT pathway. The therapeutic goal was achieved in 2444 patients (27%) and an additional 552 patients (6%) had appropriate TG levels but potential hepatotoxicity. 936 patients (10%) had potential TPMT deficiency, and 58 patients (1%) had potential TPMT absence and were at risk for leukopenia. 140 patients (2%) had too high a dose.

Conclusions : Measurement of AZA/MP metabolites can be used by practising gastroenterologists to identify potential reasons for nonresponse to AZA or MP, and to identify patients at risk for certain drug-related toxicities.


Azathioprine (AZA) and its active metabolite mercaptopurine (MP) are frequently used in the management of inflammatory bowel disease. These drugs are highly effective in steroid refractory and steroid dependent ulcerative colitis and Crohn's disease. Azathioprine is able to induce remission in most ulcerative colitis patients who are refractory to steroids.1, 2 In ulcerative colitis patients in remission on steroids, azathioprine and mercaptopurine have a steroid sparing effect.2, 3 Withdrawal of azathioprine leads to an increased relapse rate for up to 2 years.4 Similarly, in Crohn's disease, azathioprine and mercaptopurine are effective in steroid dependent and steroid refractory disease.5 In steroid refractory Crohn's disease, remission is achieved in 60% to 70% of patients.5 These drugs also maintain remission in Crohn's disease, as was shown by an increased relapse rate following drug withdrawal after 4 years.6 Azathioprine has also been shown to increase quality-adjusted life expectancy in patients with steroid induced remission of Crohn's disease.7

Despite proof of efficacy, however, these drugs are rarely chosen as first line therapy. Possible reasons for this include delay before efficacy, the potential for adverse events such as hepatotoxicity and leukopenia, and a previously reported 20–40% nonresponse rate. The reasons that azathioprine/mercaptopurine may be ineffective include a lack of compliance, inadequate dosing, pharmacogenetically altered drug metabolism, and azathioprine/mercaptopurine refractory disease.

Recently, a better understanding of the metabolism of these drugs and their mechanism of action has developed. Oral azathioprine is rapidly converted to mercaptopurine. Mercaptopurine may be catabolized into inactive metabolites such as thiouric acid (by xanthine oxidase) and methylmercaptopurine (MMP) (by thiopurinemethyltransferase (TPMT)), or it may be anabolized to the active thioguanine nucleotides.8 Clinical response is highly correlated with thioguanine level, but not with any other variable including MMP level or drug dose.9 A recent study in children showed that the frequency of therapeutic response increased with a thioguanine level > 235 pmol/8 × 108 erythrocytes.9 Another study in adults found that the optimal cut-off between responders and nonresponders was at a thioguanine level > 260.10 However, thioguanine levels > 450 may be associated with leukopenia, and MMP levels > 5700 pmol/8 × 108 erythrocytes may be associated with hepatotoxicity.9

Measurement of thioguanine and MMP levels in inflammatory bowel disease patients on azathioprine/mercaptopurine may allow the identification of patients at risk for toxicity and provide an explanation for the ineffectiveness of the drug in certain patients.11 Patients with no detectable metabolites are likely to be noncompliant, while patients with low levels of metabolites may be suboptimally compliant or under-dosed. Some patients may have high levels of MMP but low levels of thioguanine. It has been postulated that these patients have a preferential metabolism of drug via the TPMT pathway.11 Additionally, some patients may have high levels of thioguanine and low or absent levels of MMP. Such patients may have heterozygous or homozygous mutations of the TPMT gene resulting in low or absent TPMT activity and a high risk of bone marrow toxicity.11

Knowledge of metabolite levels may thus significantly alter the clinical care of patients with inflammatory bowel disease. Those at risk for toxicity may be subject to closer monitoring or dose reduction. Patients who are nonresponders with low thioguanine levels may benefit from improved compliance or dose-escalation.12 Dubinsky et al. suggested that an alternative strategy in patients with a preferential metabolism via the TPMT pathway may include the therapeutic administration of thioguanine.9, 13

Measurement of mercaptopurine metabolites became available to practising gastroenterologists in 1999. The aim of this study was to examine the results of initial mercaptopurine metabolite tests in patients during the first year of widespread test availability.


The erythrocyte-free bases thioguanine and MMP can be measured using reverse phase high performance liquid chromatography.14 Erythrocyte thioguanine and MMP levels have been shown to correlate with leukocyte DNA levels.15 Mercaptopurine metabolite panels reporting the levels of thioguanine and MMP (pmol/8 × 108 erythrocytes) became commercially available through Prometheus Laboratories (San Diego, CA) in February 1999 and achieved widespread use in the autumn of 1999.

In this observational study, the results of all initial mercaptopurine metabolite panels sent to Prometheus Laboratories by gastroenterologists across the USA through September 2000 were reviewed without patient identifiers. The data were provided by Prometheus Laboratories for this observational analysis. All values from patients who were reported to be receiving thioguanine therapy were excluded. The metabolite panels were interpreted as shown in Table 1 for the purpose of this analysis. The clinical care of the patients remained with the primary physicians.

Table 1.  Interpretation of mercaptopurine metabolite panels
Thioguanine (pmol/8 × 108 erythrocytes)MMP (pmol/8 × 108 erythrocytes)Interpretation
< 230< 5700Under-dosed
< 230> 5700Preferential metabolism  via TPMT pathway
230–450< 5700Therapeutic goal
230–450> 5700Potential hepatotoxicity
> 450< 5700Potential TPMT deficiency  (potential bone marrow  toxicity)
> 1000UndetectablePotential TPMT absence  (potential bone marrow  toxicity)
> 450> 5700Over-dosed


Initial mercaptopurine metabolite panels were reviewed for 9187 patients. In a sample of 733 panels sent from the state of North Carolina, 34% came from academic medical centres and 66% were from private practice gastroenterologists. Repeat metabolite panels were sent on 28% of the patients during the time frame of the study, but only the initial metabolite panels were considered in the analysis.

As shown in Table 2, noncompliance was detected in 263 patients (3%), and under-dosing in 4260 patients (46%). 534 patients (6%) had levels which were consistent with preferential metabolism via the TPMT pathway. The therapeutic goal was achieved in 2444 patients (27%) and an additional 552 patients (6%) had appropriate thioguanine levels but potential hepatotoxicity. Similar to previously reported rates,16 936 patients (10%) had a potential TPMT deficiency and 58 patients (1%) potential TPMT absence and were at risk for leukopenia. 140 patients (2%) had metabolite profiles which were consistent with a dose that was too high.

Table 2.  Results of initial mercaptopurine metabolite panels
InterpretationPatients n (%)
Non-compliance263 (3%)
Under-dosed4260 (46%)
Preferential metabolism via TPMT pathway534 (6%)
Therapeutic goal2444 (27%)
Potential hepatotoxicity552 (6%)
Potential TPMT deficiency  (potential bone marrow toxicity)936 (10%)
Potential TPMT absence  (potential bone marrow toxicity)58 (1%)
Over-dosed140 (2%)


Assays for the mercaptopurine metabolites thioguanine and MMP became commercially available in 1999. This is the first report of the results of mercaptopurine metabolite testing obtained during routine clinical practice.

To date, studies demonstrating the effectiveness of azathioprine and mercaptopurine in the management of inflammatory bowel disease have relied upon weight-based dosing. Recent evidence, however, suggests that the clinical response is not correlated with dose but is correlated with levels of the active metabolite thioguanine.9 Dubinsky et al. showed that 78% of mercaptopurine treated inflammatory bowel disease patients in remission had thioguanine levels > 225 pmol/8 × 108 erythrocytes, compared with only 36% of patients who were nonresponders.17 Additionally, levels of MMP > 5700 were associated with hepatotoxicity. Checking metabolite levels could therefore be used to assist in optimizing the dosing of azathioprine or mercaptopurine.

A survey of paediatric gastroenterologists found that the principle reasons for physicians to measure mercaptopurine metabolites were to assess the cause for an inadequate response and to monitor for drug toxicity.18 We propose that most of the metabolite tests sent by gastroenterologists in the present study were ordered for the same reasons. Several possible explanations for drug failure or inadequate response can be identified based on the metabolite profile results.

In this study, 3% of patients had undetectable metabolites. This strongly suggests noncompliance or conditions that might impair absorption of the drug in the small intestine. It has been proposed that metabolite measurement be used to monitor patient compliance.15

Another potential explanation for drug failure is under-dosing. This can result from a prescription regimen that does not yield adequate levels of the active metabolites, inadequate compliance by the patient, a problem with drug absorption, or some combination of the above. Patient history and pharmacy records can be used to help identify the contribution of noncompliance to under-dosing. Under-dosing was the most common metabolite profile identified in our study, comprising 46% of the profiles. Under-dosing in the setting of remission does not require dose adjustment. However, if metabolite levels are low in a patient with an inadequate response to azathioprine/mercaptopurine in the setting of good compliance, the physician should consider increasing the dose. Cuffari et al. reported preliminary data showing that dose-escalation could lead to higher thioguanine levels and an improved clinical response in children.12 More recently, Cuffari et al. reported on adult inflammatory bowel disease patients with an inadequate response to azathioprine and thioguanine levels < 250; dose-escalation led to a remission of disease in 18 of 22 patients.19

A subset of patients not responding to azathioprine/mercaptopurine may have high TPMT activity, resulting in preferential shunting to MMP and thus high MMP but low thioguanine levels. This group composed 6% of all profiles analysed in our study. When Dubinsky et al. performed a dose-escalation study in inflammatory bowel disease patients who did not show an initial response to mercaptopurine, 86% remained nonresponders and MMP levels increased to a median of 11 990. The median thioguanine level increased to only 171, and 72% of the thioguanine levels remained below 230.20 This may be an important group of nonresponders to identify, as dose-escalation may lead to drug toxicity. Additionally, these patients may be candidates for clinical trials of thioguanine.13

In our analysis, 46% of all metabolite panels had thioguanine levels > 230, although the proportion of nonresponders is not known. Recent data suggest that a higher threshold of 260 may be a better predictor of response.10 While some nonresponders are likely to have inflammation refractory to azathioprine/mercaptopurine, others may have stricturing disease or coexistent conditions such as irritable bowel syndrome that would not be expected to respond to immunomodulator therapy. When active inflammation is documented and thioguanine levels are > 230 in a patient who has been on azathioprine/mercaptopurine for an adequate duration of time, dose escalation (only if thioguanine levels < 450 and there is no leukopenia) or alternative therapies should be considered.

Metabolite profiles may also be obtained to identify those at risk for certain drug-related toxicities. While some toxicities such as pancreatitis appear to be idiosyncratic, others such as hepatotoxicity and leukopenia are associated with certain metabolite profiles. In our study, 6% of patients had thioguanine levels within the target range but potential hepatotoxicity based on MMP levels > 5700. An additional 2% had elevated levels of both metabolites. The proportion of these patients experiencing drug-related toxicity in this series is not known, but it would seem prudent to monitor these patients closely.

As noted previously, genetic mutations of TPMT, the enzyme responsible for the metabolism of azathioprine/mercaptopurine to MMP, result in considerable variation in enzyme activity. Weinshilbourn & Sladek found that 0.3% of individuals tested had low-to-absent enzyme activity, 11% had intermediate activity, and 89% had normal-to-high activity.16 A diminished TPMT activity results in high levels of thioguanine and low levels of MMP. While patients with reduced TPMT activity can be successfully treated with azathioprine/mercaptopurine, they should be started on low doses and monitored closely for bone marrow toxicity.11 Some investigators have proposed measuring TMPT activity prior to the initiation of immunomodulator therapy in all patients with inflammatory bowel disease.21 Although TPMT activity was not measured in this study, an analysis of the profiles suggested that 1% had a potential absence of TPMT activity (homozygous for mutant TPMT) and 10% had a possible decrease in TPMT activity (heterozygous for mutant TPMT), results which are similar to those reported previously.16 In short, 11% of patients in this series were at risk for leukopenia with a standard dose of azathioprine or mercaptopurine, a finding that justifies further investigation into the use of TPMT testing prior to initiation of immunomodulator therapy.

While the measurement of metabolites may identify some patients at risk for hepatotoxicity and leukopenia, it does not obviate the need to monitor all patients on azathioprine/mercaptopurine therapy. Colombel et al. recently found that 73% of patients with Crohn's disease and bone marrow toxicity (leukopenia or thrombocytopenia) associated with azathioprine had no identifiable TMPT mutation.22 This suggests that metabolite testing should not replace laboratory monitoring for toxicity.

The results of this study are limited by the absence of associated clinical information. Nevertheless, the fact that 73% of samples analysed yielded results outside what is considered the ideal therapeutic range suggests that this information could have been useful in clinical care. Another potential drawback is selection bias. Clinicians may be more likely to order metabolite testing in the subset of patients who are not responding to therapy with azathioprine/mercaptopurine. Thus, these results may include a higher proportion of nonresponders than the general inflammatory bowel disease population. Nevertheless, if the tests were obtained to investigate nonresponse, several possible explanations were identified, including noncompliance, under-dosing, and pharmacogenetic resistance to azathioprine/mercaptopurine. If the tests were obtained to find those at risk for drug toxicity, a sizeable group of patients was identified.

Future studies are needed to determine the value of checking TPMT activity prior to initiating immunomodulator therapy, to define the optimal metabolite levels, to assess whether dose-escalation and repeat metabolite testing improves patient outcomes, and to study the natural history of patients with high metabolite levels but no clinical evidence of drug-related toxicity.


We thank Amy Landon for assistance with the data management. This work was presented in part at Digestive Diseases Week 2001, Atlanta, Georgia.