ABOUT AZATHIOPRINE

Azathioprine's mode of action is not fully understood.4^,5 After the drug is taken orally, it is rapidly converted non-enzymatically to mercaptopurine by sulfhydryl compounds (e.g. cysteine, glutiathione).5 Mercaptopurine crosses cell membranes and is converted by the enzyme hypoxanthine phosphoribosyltransferase (HPRT), and subsequently by other enzymes, to 6-thioguanine nucleotides.5^,6 These nucleotides are thought to compete with endogenous nucleotides in many biochemical pathways (e.g. in the synthesis of DNA), so accounting for the immunosuppressive effects of azathioprine/mercaptopurine.5 Two other pathways compete with the activation pathway for azathioprine in cells: one is the conversion of mercaptopurine to methylmercaptopurine by TPMT; the other is conversion of mercaptopurine to inactive 6-thiouric acid by the enzyme xanthine oxidase5 (see figure 1).The relative contributions of these two alternative pathways determine how much azathioprine is inactivated and how much remains available for activation to 6-thioguanine nucleotides.5

• Download figure
• Open in new tab
• Download powerpoint

Unwanted effects of azathioprine

A minority of patients experience nausea when first given azathioprine, but this appears to be relieved by taking the tablets after meals.6 A very common unwanted effect (i.e. occurring in at least 1 in 10 people) is depression of bone marrow, most often manifesting as leucopenia; this is generally reversible.6 In a review of 66 studies including a total of 8,302 patients with inflammatory bowel disease treated with either azathioprine or mercaptopurine, the incidence rate (per person and year of treatment) of drug-induced myelotoxicity was 3% (95% CI 3% to 4%), with the risks being similar for both drugs. The risk of death among patients who developed myelotoxicity was 0.94% (95% CI 0.32 to 2.70).7 Bone marrow toxicity may develop at any time during therapy, but is most likely in the first months.8 Thrombocytopenia is common with azathioprine6 (occurring in between 1 in 10 and 1 in 100 people). Anaemia, pancreatitis, hypersensitivity reactions, cholestasis and degeneration of liver function tests are uncommon (occurring in between 1 in 100 and 1 in 1,000).

Dosing and other precautions

The dose of azathioprine for autoimmune diseases recommended in the drug's summary of product characteristics (SPC) is 1–3 mg/kg daily, adjusted according to clinical response and haematological tolerance.6 The SPC reccommends that, in patients with hepatic dysfunction, “regular” full blood counts and liver function tests should be undertaken, whereas in patients with renal deficiency, doses of azathioprine should be at the lower end of the normal range.6

The activity of the enzyme xanthine oxidase can be reduced by allopurinol, an interaction that results in reduced metabolism of mercaptopurine to inactive metabolites and resulting increased production of immunosuppressant thioguanine nucleotides. Consequently, the SPC for azathioprine advises that the dose of azathioprine or mercaptopurine needs to be reduced by three-quarters in patients also taking allopurinol.6

To detect signs of azathioprine-induced bone marrow suppression in patients on azathioprine, the British National Formulary (BNF) recommends weekly complete blood counts during the first 4 weeks, with more frequent monitoring for those with hepatic or renal impairment or on a higher dosage.8 By contrast, the SPC for azathioprine recommends weekly counts for the first 8 weeks of treatment, with more frequent monitoring for patients with severe renal or liver disorder or on a higher dosage,6 thereafter, it recommends checking blood counts at least 3-monthly.6 However, guidelines from the British Society of Gastroenterology state that there is no evidence this approach is effective, and that instead it may be sufficient to carry out less frequent monitoring (within 4 weeks of starting therapy and every 6–12 thereafter).9 The SPC recommends that the maintenance dosage should be decreased to the minimum required for a clinical response.6

ABOUT TPMT ACTIVITY

The level of TPMT activity varies between individuals, depending on whether both, one or neither of their two TPMT genes is a mutant allele associated with reduced or absent enzyme function. Several different mutant alleles (polymorphisms) have been identified, three of which appear to account for most cases of reduced TPMT activity.5 People who have inherited a pair of mutant alleles (i.e. up to 0.6% of the population)2^,3^,5 have TPMT deficiency (i.e. little or no activity); those with one mutant allele (heterozygous genotype; about 10% of people) have intermediate TPMT activity (sometimes known as ‘low’ activity); and those with a normal (‘wild type’) pair of alleles (about 90% of people) have normal (sometimes called ‘high’) TPMT activity.5 Confusingly, some laboratories and published papers reserve the term ‘high’ for activity above their quoted ‘normal’ range.

In general, people of African-Caribbean origin have lower TPMT activity (and are more likely to have low TPMT activity) compared with Caucasians and South Asians, and women have lower activity than men.3

Consequences of reduced tpmt activity

Bone marrow suppression

In people with lower than normal TPMT activity, more drug is available for conversion to immunosuppressant 6-thioguanine nucleotides. Therefore, patients with TPMT deficiency are at high risk of developing bone marrow toxicity. Patients with intermediate TPMT activity are also more likely than those with normal activity to develop such toxicity. For example, in an uncontrolled trial involving 394 patients with inflammatory bowel disease treated with azathioprine, those with intermediate TPMT activity (5–13.7 U/mL) were four times as likely as those with high TPMT activity (>13.8 U/mL) to develop bone marrow suppression (14.3% vs. 3.5% of patients).10 An uncontrolled study involving 33 patients with rheumatoid arthritis found that those with intermediate TPMT activity (8.3–18.0 pmol/106/erythrocytes/hour) were more likely than those with high TPMT activity (18.1–39.4 pmol/106/erythrocytes/hour) to develop severe side effects with azathioprine (relative risk 3.1, 95% CI 1.6 to 6.2).11

In patients deficient in TPMT, the onset of bone marrow suppression tends to occur earlier in azathioprine therapy. Evidence for this includes a study analysing TPMT genotype in 41 patients with Crohn's disease who developed leucopenia or thrombocytopenia while being treated with azathioprine or mercaptopurine.12 Bone marrow toxicity occurred in the first 1.5 months of treatment in patients with two mutant alleles (4 patients); at 1–18 months in those with 1 mutant allele (7 patients); and at 0.5–87 months in those where no known TPMT mutation was found (30 patients). However, reduced TPMT activity does not account for all bone marrow suppression associated with azathioprine: patients with normal or high TPMT activity may also develop this effect.13^,14

Risk of other unwanted effects

Patients with reduced TPMT activity may also be more likely to develop other unwanted effects as well as bone marrow toxicity. Evidence for this includes an uncontrolled trial, in which 207 patients with inflammatory bowel disease received azathioprine.15 A heterozygous TPMT genotype strongly predicted withdrawal of azathioprine due to unwanted effects (79% did not tolerate 6 months' treatment with azathioprine vs. 35% with normal TPMT genotypes, p = 0.0003). Gastric intolerance was the commonest reason for withdrawal among heterozygotes and was commoner than in those with homozygous normal genotype (37% vs. 7% of patients, p<0.001), as was bone marrow toxicity (26% vs. 0.5%, p<0.01). Gastric intolerance (particularly nausea) occurred early (i.e. within 6 weeks of starting treatment) and only those remaining on azathioprine beyond this went on to develop bone marrow toxicity (which did not manifest until 12 weeks or later).15

TESTING FOR TPMT

It is possible to measure a patient's TPMT enzyme activity. Laboratory tests for activity measure enzyme function in red blood cells, which reflects activity in other tissues and cells.16^,17 In patients who have had a blood transfusion in the previous 3 months, tests for TPMT activity can be unreliable because of TPMT activity in the received blood.18 It is also possible to check whether the patient is heterozygous or homozygous for normal or mutant alleles (genotype testing). Tests for TPMT activity are available at two UK hospitals; the Purine Research Laboratory at Guy's and St. Thomas' Hospital, and at the Pathology Department at City Hospital, Birmingham. The NHS is charged around £30 for a test.19 Both laboratories can also test for genotype. However, genotype testing tends not to be routinely available outside research settings.4^,20

Potential benefits of testing

In theory, measurement of TPMT activity before starting therapy could identify patients with low or undetectable TPMT activity and so prevent them from receiving azathioprine and developing potentially fatal bone marrow suppression. Relying on blood counts alone in such patients could mean that bone marrow suppression has already occurred before blood cell counts have dropped, while marrow stores of white cells are exhausted.1 TPMT activity testing could also help identify patients with intermediate TPMT activity who might have toxicity with standard doses of azathioprine, but might tolerate and respond to lower doses. Finally, TPMT testing could help identify patients with high TPMT activity in whom azathioprine is less effective and who could benefit from earlier use of higher doses of the drug or alternative strategies, to avoid long periods of non-response.4

Does testing prevent toxicity?

One study has assessed whether knowledge of TPMT genotype might predict azathioprine toxicity.21 It involved 67 patients with rheumatic disease whose TPMT genotype was determined and who were prescribed azathioprine (2–3 mg/kg daily) as second-line therapy. The patients' clinicians were unaware of the TPMT genotype results. The primary endpoint was discontinuation of azathioprine therapy because of toxicity. Six of 67 patients (9%) were heterozygous (i.e. one normal and one mutant allele). Treatment was discontinued in 5 of these 6 patients within 1 month of starting because of low leucocyte counts. The sixth patient did not adhere to treatment. Patients with a normal genotype received therapy for longer than heterozygous patients (median 39 weeks vs. 2 weeks, p = 0.018). There were no unwanted haematological effects in patients with normal genotype. The investigators suggested that TPMT genotype testing “is a quick way to identify patients at risk for acute toxicity from azathioprine”.

Target trial

The Department of Health in England has funded a randomised controlled trial with the aim of assessing the clinical value of TPMT genotyping and enzyme testing before starting azathioprine: the TPMT Azathioprine Response to Genotyping and Enzyme Testing (TARGET) trial.22 The primary outcome measure is neutropenia (neutrophil count falling below 1.0×10⁹/L) in the first 4 months of azathioprine maintenance treatment. The trial has been completed, but it is not expected to be published in the near future.

Does testing help tailor dose?

One study has provided limited evidence on whether using a graduated dose determined by body weight and prior measurement of TPMT activity is safe and effective.23 It was a double-blind randomised placebo-controlled trial of azathioprine in 63 patients with moderate-to-severe atopic eczema. Overall, 5 patients in the azathioprine group and 2 in the placebo group had heterozygous TPMT genotype. The primary outcome measure was change in disease activity from baseline to 12 weeks, assessed with the “six area six sign atopic dermatitis” (SASSAD) score. At the end of the trial, there had been a significantly greater reduction in mean disease activity with azathioprine (baseline SASSAD score not stated; a reduction of 12.0 units in the azathioprine group vs. 6.6 units in the placebo group, difference 5.4 units, 95% CI 1.4 to 9.3). None of the few patients with intermediate TPMT activity (2.5–7.5 nmol/hour per mL red blood cells) developed bone marrow suppression, and efficacy was maintained in all of them. Two patients with normal TPMT activity (>7.5 nmol/hour per mL red blood cells) developed mild neutropenia. The trial's authors considered that TPMT activity-based dosing seemed to reduce predicted toxicity while maintaining drug efficacy.

Two uncontrolled trials in patients with inflammatory bowel disease provide contrasting views on whether dose tailoring according to TPMT genotype or activity is useful.24^,25 One was a study involving 77 consecutive patients started on azathioprine or mercaptopurine at doses individualised and subsequently adjusted according to blood thioguanine nucleotides concentrations and clinical response to treatment, including haematological monitoring.24 Both TPMT activity and genotype were tested, allowing identification of a wild-type group and a heterozygous group. Although initial doses for these groups were similar, there was a twofold difference in the doses between the groups at 9 months (1.8 mg/kg daily in the wild-type group vs. 0.9 mg/kg daily in the heterozygous group, p = 0.0006). Based on the trial findings, the investigators estimated that the azathioprine maintenance dose for those with wild-type genotype had to be around three times that for heterozygous people to produce similar therapeutic concentrations of 6-thioguanine nucleotides. The other study involved 131 patients whose azathioprine or mercaptopurine dose was based on previously measured TPMT activity.25 In all, 7% of patients had intermediate TPMT levels (5–13.7 U/mL) and 93% had high levels (>13.8 U/mL). Bone marrow toxicity devloped in 3 of the 122 patients with high TPMT activity, and 1 of the 9 with intermediate activity. The investigators concluded that the results could not confirm that dosing based on TPMT activity helped to prevent bone marrow toxicity.

EVIDENCE ON COST-EFFECTIVENESS

In theory, determining TPMT activity or genotype could prevent some instances of bone marrow suppression and associated costs. A systematic review identified studies providing data on costs of treating azathioprine-induced adverse drug reactions.26 Most of the seven studies found were limited by the fact that they used theoretical modelling with a high rate of approximation of costs, or included very few reports of unwanted effects with azathioprine. From these data, and information on the cost of TPMT genotyping, the review's authors estimated that the cost to prevent one case of neutropenia using such testing is €5,300.

WHAT HAPPENS IN PRACTICE?

Both the BNF and SPC mention that patients with inherited deficiency of TPMT may be unusually sensitive to the myelosuppressive effect of azathioprine, but neither explicitly recommends testing TPMT activity before starting therapy. Even so, there has been a considerable uptake in TPMT testing in UK practice over the past 5 years.27 There is a marked difference between specialties in the use of the enzyme-activity testing before starting azathioprine therapy: for example, in a UK survey of three groups of NHS specialists, 94% of dermatologists, 60% of gastroenterologists and 47% of rheumatologists reported using testing in this way.20

The differences between specialities appear to reflect the various specialist guideline recommendations. UK dermatology guidelines suggest that the case for testing is clear-cut, stating that: “Pretreatment TPMT testing should be performed in all patients prescribed azathioprine for treatment of dermatological conditions”.4 For those patients with intermediate TPMT activity in whom a trial of azathioprine is deemed appropriate, these guidelines recommend using a low dose (e.g. 0.5–1 mg/kg daily) and taking “extra care with haematological surveillance”.4 Neurology guidelines for the management of patients with myasthenia gravis state that “Those patients who lack enzyme [TPMT] activity are bound to develop myelosuppression. If the assay is available then it is sensible to use it”.28 In contrast, the British Society of Gastroenterology guidelines state “It cannot yet be recommended as a prerequisite to therapy”.9 Similarly, European gastroenterology guidelines state that no recommendation can be made about routine measurement of TPMT activity or genotype prior to starting therapy.29 British Society for Rheumatology guidelines on the monitoring of disease-modifying drugs do not make any recommendations on TPMT testing in patients given azathioprine.30

CONCLUSION

Thiopurine methyltransferase (TPMT) is an enzyme involved in the metabolism of the immunosuppressant drugs azathioprine and mercaptopurine. Up to 6 in every 1,000 people have inherited deficiency of this enzyme and so are at risk of potentially fatal bone marrow suppression from these drugs. A further 10% of people have lower than normal TPMT and are likelier than people with normal activity to develop bone marrow suppression. Testing TPMT activity before starting azathioprine can identify people who are TPMT deficient and who should not receive the drug. It is also potentially useful in identifying those who have intermediate TPMT activity (who would probably benefit from a lower dose), and those who have high TPMT activity (who may need a higher dose). However, bone marrow suppression can still occur in people with normal TPMT activity. Also, it is unclear whether TPMT testing before therapy helps prevent bone marrow suppression; how it compares with careful monitoring of full blood counts during therapy; or whether it helps with dose-tailoring. A recently completed but unpublished trial might provide answers to these questions. However, pending these data, we believe it is a sensible precaution to test all patients starting on azathioprine for TPMT enzyme activity.