This article is more than 5 years old. Information may no longer be current.
Immunomodulators Azathioprine/6-Mercaptopurine and Methotrexate
Click Here to Manage Email Alerts
Treatment aims in the inflammatory bowel diseases are evolving beyond the induction and maintenance of corticosteroid-free remission. More recently, it has become recognized that by treating with immunomodulatory and biologic therapies earlier in the disease course and by achieving mucosal healing, the natural history of these diseases may be modified in a positive way. Despite the clear benefits of early and appropriate introduction of these therapies, the potential for risks, some of them serious, must be understood by the clinician and discussed with patients. This excerpt reviews the relevant adverse events associated with the use of the immunomodulatory agents azathioprine, 6-mercaptopurine and methotrexate.
Azathioprine and 6-Mercaptopurine
The thiopurine immunomodulators azathioprine (AZA) and 6-mercaptopurine (6-MP) have become the mainstay of maintenance medical therapy for many patients with Crohn’s disease (CD) or ulcerative colitis (UC). In CD, meta-analyses support the use of these agents as induction and maintenance therapies and to prevent postoperative disease recurrence. In UC, the evidence of benefit is less robust, but meta-analyses demonstrate thiopurine effectiveness as maintenance therapy, with less benefit as induction agents. A previous retrospective referral center study in CD failed to demonstrate a reduction in surgical rates despite an increased use of thiopurines in recent decades. However, a more recent prospective study of almost 400 patients in Spain demonstrated thiopurines are effective in reducing hospitalizations and surgeries in both CD and UC. Similarly, a recent population-based study from Wales demonstrated a reduction in surgical rates in patients with CD on thiopurine therapy. These improved outcomes likely reflect the trend toward the earlier introduction of immunomodulators at a time when patients’ disease phenotype is believed to be more inflammatory, before the occurrence of fibrostenotic or perforating disease complications. In each patient receiving thiopurines, such benefits must be balanced against the not-infrequent potential for adverse effects, so that a favorable benefit-to-risk ratio is maintained.
Adverse Events
Adverse events (AEs) during thiopurine therapy are not infrequent, and therefore careful monitoring is standard practice for all patients who receive these agents. Meta-analyses of induction therapies in CD reveal an odds ratio (OR) of an AE in patients who receive thiopurines of 3.44 (95% confidence interval [CI] 1.52 to 7.77). AEs to thiopurines can be classified as being allergic/idiosyncratic (eg, pancreatitis, flu-like syndromes) or dose-dependent (eg, myelotoxicity, hepatic transaminitis, malignancies including lymphoproliferative disease). In a New Zealand study of 216 patients with inflammatory bowel disease (IBD), 25.9% of them developed an AE that resulted in drug cessation, the most common being allergic-type reactions, hepatitis, nausea and vomiting, myelotoxicity and pancreatitis.
The most important potential AE associated with any medication is the risk of mortality, but data regarding thiopurines and mortality are reassuring, especially in comparison to the risks with corticosteroids. A large case-control study of almost 14,500 patients demonstrated an association between mortality and current corticosteroid use but not with current thiopurine use. Similarly, recent registry data have demonstrated an increased mortality in CD associated with prednisone use but not with thiopurine exposure.
Allergic AEs to Thiopurines
The incidence of acute pancreatitis in patients with IBD is 1% to 6%, with the most common cause being thiopurines, which carry a risk of pancreatitis of approximately 3%. The incidence appears higher in CD than UC and has variably been reported to be more common in women. An association with pancreatitis and a deficiency in inosine triphosphate pyrophosphatase (ITPase) due to the 94C>A polymorphism has been demonstrated in some but not all studies. Acute pancreatitis usually occurs within a few weeks of commencing AZA or 6-MP. Clinically, it is usually mild, resolves promptly on stopping the thiopurine, and inpatient admission is rarely required. Although there are case reports of successfully treating with 6-MP after an acute pancreatitis episode during AZA therapy, it is generally understood that, when drug related, there is 100% cross-reactivity for drug-related pancreatitis between the thiopurines, and therefore rechallenge is not recommended.
An idiosyncratic early adverse reaction, consisting of constitutional flu-like symptoms, joint pains, rashes and fevers, occurs not infrequently in patients exposed to thiopurines. The mechanisms of this reaction are poorly understood, although an association with the ITPase 94C>A polymorphism has again been postulated. Rates of recurrence with 6-MP after hypersensitivity reactions to AZA are at least 50%, but unlike pancreatitis, rechallenge may be attempted (at a smaller initial dose) in appropriately counseled patients.
An idiopathic cholestatic hepatitis associated with thiopurines is rare and usually occurs within 2 weeks to 3 weeks of commencing thiopurines. Clinically, it is characterized by nausea, abdominal pain and jaundice. Biochemically, increased bilirubin and alkaline phosphatase are seen along with moderate elevations of aminotransferases. Histologically, cholestasis and parenchymal cell necrosis are observed. Upon recognition and thiopurine withdrawal, the syndrome resolves clinically within weeks, although biochemical abnormalities may be present for several months. Given the potential for hepatocyte injury, switching from AZA to 6-MP should be done cautiously, if at all, after this AE.
Dose-dependent AEs
Myelotoxicity is the most important potential AE associated with thiopurines. A recent meta-analysis reported a cumulative incidence of myelotoxicity in patients who received thiopurines of 7%, with an incidence of 3% per patient-year. The cumulative incidence of severe myelotoxicity (defined as a neutrophil count of less than 0.5 × 109/L) is 1.1%, with an annual incidence rate of 0.9%. Myelotoxicity with thiopurines has been reported to occur in as little as days after starting, to as long as 27 years after the commencement of therapy. Of those who develop myelotoxicity, 6.5% of patients had an associated infectious complication.
The seemingly unpredictable nature of this complication supports the need for ongoing hematological monitoring for the duration of therapy with these agents. Myelotoxicity appears to occur, at least in part, due to the accumulation of high intracellular levels of the active metabolite 6-thioguanine nucleotide (6-TGN). The production of this metabolite is at least partially dependent on the activity of thiopurine methyltransferase (TPMT). TPMT activity is determined by a genetic polymorphism that has a trimodal distribution within the population. Eighty-nine percent of people are homozygous high, with normal enzyme activity; 11% are heterozygotes, with intermediate activity; and 0.3% are homozygous low, with negligible enzyme activity. TPMT heterozygotes and homozygous low patients are at particular risk of myelotoxicity. Therefore, it is recommended heterozygotes should receive one-third to one-half the usual thiopurine starting dose and homozygous low patients should not receive thiopurine therapy.
A recent meta-analysis of patients of all TPMT genotypes receiving thiopurines revealed an OR of 2.93 for the presence of TPMT polymorphisms and any adverse drug reaction, and 5.93 for polymorphisms and myelotoxicity. Despite the importance of this pharmacogenomic relationship, most of patients who develope myelotoxicity have normal TPMT activity, outlining the importance of regular laboratory monitoring for the duration of thiopurine therapy in all patients. A suggested schedule is to monitor the white blood cell count weekly for 4 weeks, alternate weekly for 4 weeks, then at 12 weeks and every 3 months thereafter, including testing 2 weeks after any thiopurine dose escalation.
More recently, there has been an appreciation of the risk of infections and vaccine-preventable illnesses with immunosuppression. Therefore, it is recommended to ensure patients are adequately vaccinated according to routine immunization schedules for the general population in order to avoid vaccine-preventable infections prior to the commencement of thiopurines or any other immunomodulatory therapy. In particular, in patients with IBD, vaccination is recommended for the following fivevaccines: hepatitis B virus in seronegative patients; human papillomavirus in young women influenza virus annually (including H1N1); Strep pneumonia 5 yearly; and varicella in patients without a history or serologic evidence of chickenpox or varicella zoster infection. All of these are attenuated vaccines except varicella, which is a live vaccine, and should not be given to patients already on immunomodulatory agents.
Hepatotoxicity is a well-recognized AE of thiopurines; however, the prevalence varies greatly among published studies. A recent systematic review of 34 retrospective studies showed a low prevalence of hepatotoxicity of 3.3%, whereas in a prospective study of 161 patients, the rate of hepatotoxicity was as high as 23%. Some of these differences may be due to the variety of definitions of hepatotoxicity employed — for example, abnormalities of liver function tests (LFTs) are common, while true “hepatotoxicity” is rare. In addition, hepatotoxicity appears to be less common with 6-MP compared to AZA. Cases of thiopurine hepatotoxicity can be grouped into three types — an idiopathic cholestatic hepatitis (as discussed above), a dose-dependent transaminitis, and hepatotoxicity due to endothelial cell injury (nodular regenerative hyperplasia [NRH] and veno-occlusive disease).
The dose-dependent transaminitis seen with dose escalation of thiopurines is the most common hepatic AE and may be asymptomatic or may cause nausea, headaches, and malaise. It may resolve spontaneously with time, meaning that when mild (transaminase levels two to three times normal), and in an asymptomatic patient, it may be simply monitored. Alternatively, when aminotransferase elevations are moderate, or in symptomatic patients, it will reverse with dose reduction or cessation without long-term clinical sequelae. A 50% dose reduction is initially suggested, after which if LFTs normalize, cautious dose re-escalation can often be achieved successfully. Anecdotal experience and small retrospective case series have shown more than 50% of patients intolerant of AZA may tolerate 6-MP, including those patients with AZA-induced hepatotoxicity. Although unproven, it has been postulated that the intolerance, including hepatotoxicity associated with AZA, may be due to derivatives produced from the imidazole side chain of AZA. Given the equal efficacy of AZA and 6-MP, a trial of 6-MP in patients with AZA-induced hepatotoxicity is recommended. The dose-dependent hepatitis has been variably associated with elevated levels of the end-metabolite 6-methylmercaptopurine (6-MMP), as reported in the initial pediatric studies in which 6-MMP levels of greater than 5700 pmol/8 × 108 red blood cells count were associated with biochemical hepatitis. These findings have been replicated in some, but not all, studies, and the sensitivity and specificity of 6-MMP for drug-induced hepatotoxicity is low, meaning that dose reduction due to high 6-MMP levels is only recommended for patients with elevated aminotransferases. More recently, it has become recognized that approximately 15% of patients preferentially metabolize thiopurines to produce 6-MMP instead of 6-TGN, predisposing them to hepatotoxicity and therapeutic inefficacy. Small, retrospective, single-center studies have demonstrated that the addition of low doses of adjunctive allopurinol, in combination with thiopurine dose reduction, can be used to reverse this inefficacious metabolic profile and the associated hepatitis.
The thiopurine hepatotoxicities associated with endothelial cell injury (NRH and veno-occlusive disease) are important, as these can be complicated by portal hypertension and potentially life-threatening sequelae. NRH has been associated most frequently with 6-thioguanine (6-TG) use but is also recognized to occur less commonly with AZA/6-MP, and even in thiopurine-naive patients with IBD. In the initial cohort of 111 patients who received 6-TG at high doses (80 mg daily) for a median of 9.1 months, NRH was demonstrated on liver biopsy in 76% of patients with laboratory abnormalities (elevated liver enzymes or thrombocytopenia) and 33% of patients with normal laboratory values. This finding led to the widespread recommendation that 6-TG should not be used in the treatment of patients with IBD. More recent European data have shown lower doses of 6-TG (20 mg daily) may be used safely in patients with IBD, although further data are required before this can be more widely recommended. NRH has also been associated with AZA/6-MP use, especially in men and in patients with fibrostenosing small bowel disease or prior extensive small bowel resections. In one multicenter study of 37 patients with NRH, the median time to diagnosis of NRH after starting AZA was 48 months (range 6 month to 187 months), and the cumulative incidence rate was 1.25% at 10 years (0.29 to 2.21).
After a median follow-up of 16 months, 14 patients developed complications of portal hypertension. Thrombocytopenia, with or without abnormal LFTs, should alert clinicians to the rare but potentially serious possibility of NRH in patients who received thiopurines. Interestingly, histologic evidence of sinusoidal pathologies, including NRH, has been documented in thiopurine-naive patients with IBD.
In a study of 83 liver biopsy specimens from thiopurine-naive patients with IBD (51 CD), sinusoidal dilation was noted in 34% and NRH in 6%, indicating the increased frequency of NRH in IBD may not all be thiopurine related. Regardless of the etiology, the natural history of NRH appears variable, with some case studies showing complete resolution and others characterized by persistence of portal hypertensive sequelae after an associated thiopurine was discontinued.
As with all immunomodulatory agents, the thiopurines may be associated with an increase in the long-term risk for malignancies and are classified as carcinogens by regulatory agencies. The mechanisms for the risk of carcinogenesis may include the disruption in DNA structure, replication, and repair that occurs with the incorporation of 6-TGN into DNA, and the increased cellular sensitivity to ultraviolet radiation that occurs after thiopurine exposure. In IBD populations, thiopurines have been associated with an increased risk of lymphomas and nonmelanoma skin cancers. Data supporting an association between thiopurines and lymphoma from both referral center and population-based studies have been variable. Confounding factors, including the disease itself and the underlying disease activity, mean that distinction of lymphoma causality between the underlying disease or the thiopurine has been difficult. Recent well-designed, large meta-analyses; case-control studies; and observational studies suggest the increased risk of lymphoma associated with thiopurine exposure is real. A meta-analysis of six studies totaling 3,891 patients with IBD, yielding 11 lymphomas, revealed a pooled relative risk of 4.18 for lymphoma among patients who took thiopurines compared to patients who did not receive these agents. A recent case-control study of 15,471 IBD patients from the United Kingdom General Practice Research Database did not show an increased risk of developing any cancer in patients who had filled prescriptions for AZA. However, patients who had never used AZA were shown to have an increased prevalence of lymphoma. More recently, a large French prospective observational study of 19,486 patients with IBD calculated a multivariate-adjusted HR of 5.28 for patients currently receiving thiopurines, which reduced to baseline following treatment discontinuation.
Excerpted from Regueiro MD, Swoger JM, eds. Clinical Challenges and Complications of IBD (pp 287-302) © 2013 SLACK Incorporated.
- References:
- Doherty G, et al. Cochrane Database Syst Rev. 2009;(4):CD006873.
- Gisbert JP, et al. Aliment Pharmacol Ther. 2008;28(2):228-238.
- Gisbert JP, et al. Aliment Pharmacol Ther. 2009;30(2):126-137.
- Peyrin-Biroulet L, Deltenre P, Ardizzone S, et al. Am J Gastroenterol. 2009;104(8):2089-2096.
- Prefontaine E, et al. Cochrane Database Syst Rev. 2009;(1):CD000067.