The metabolism of azathioprine and mercaptopurine is complex [2]. Azathioprine is nonenzymatically converted to mercaptopurine after oral administration and absorption. Both azathioprine and mercaptopurine are prodrugs. Mercaptopurine can then be metabolized to an active metabolite, thioguanine. Thioguanine is incorporated into ribonucleotides, thereby exerting an antiproliferative effect on mitotically active lymphocyte populations [3]. Thiopurine methyltransferase (TPMT) metabolizes mercaptopurine into an inactive 6-methylmercaptopurine. Therefore, reduction in TPMT activity predisposes to the development of adverse effects such as bone marrow suppression due to preferential metabolism of mercaptopurine to thioguanine nucleotides. Azathioprine and mercaptopurine also may possess direct anti-inflammatory properties by inhibiting a cytotoxic T cell and natural killer cell function and inducing apoptosis of T cells. Although it has been speculated that azathioprine may possess immunosuppressive and metabolic benefits beyond that of mercaptopurine, these drugs are used interchangeably in clinical practice [4].

At present, thiopurine treatment is recommended in steroid-dependent and steroid-refractory UC patients [5]. For an arbitrary but practical reason, the “European Crohn’s and Colitis Organization” proposes several scenarios where thiopurine treatment in UC should be recommended [5]: (a) patients who have a severe relapse, (b) those who require two or more corticosteroid courses within a 12-month period, (c) those whose disease relapses as the steroid dose is reduced below an arbitrary 15 mg, and (d) those whose disease relapses within 3 months of stopping steroids.

The choice of azathioprine and mercaptopurine dosage is generally based on the weight of the patient, with the intention of achieving the highest therapeutic efficacy and, at the same time, reducing the incidence of adverse effects. Several clinical trials have shown that the adequate azathioprine dosage in Crohn’s disease patients is 2–3 mg/kg/day. Azathioprine is 55 % mercaptopurine by molecular weight, and 88 % of azathioprine is converted to mercaptopurine. When changing from mercaptopurine to azathioprine, a conversion factor of 2.07 can be used [6]; thus, the equivalent dosage of mercaptopurine is approximately 1.5 mg/kg/day [7].

Strategies for initiating treatment with thiopurines vary and range from slow titration to immediately starting at the full weight-based dose [7]. A theoretical rationale behind slow titration is to carefully monitor for clinical signs of toxicity. However, dose-dependent toxicities (such as hepatitis and delayed myelotoxicity) are unlikely until a significant cumulative dose has been given. On the other hand, idiosyncratic reactions such as pancreatitis, fever, rash, nausea/vomiting, diarrhea, and arthralgias, which are dose independent, would not be avoided by giving lower doses of the drug. Therefore, slow titration may further delay an already lengthy period before therapeutic effects are seen [7].

In conclusion, as long as TPMT activity is normal, treatment can be started at an adequate dosage (i.e., 2–3 mg/kg/day for azathioprine and 1.5 mg/kg/day for mercaptopurine) with monitoring of clinical side effects, biweekly blood count, and liver function tests for 2 months and then every 3–6 months for the entire duration of the treatment [7].

Unfortunately, more than one-third of IBD patients have to discontinue thiopurine therapy during the course of the disease, the main reason being the occurrence of intolerable adverse events, which are reported in 10–30 % of the IBD patients using thiopurines. The side effects of thiopurines can be divided into dose-independent and pharmacologically explainable dose-dependent events. Among the dose-independent events, idiosyncratic or allergic reactions are rash, fever, arthralgias, pancreatitis, and hepatitis. The dose-dependent toxicity of thiopurines may largely be explained by the complex metabolism of thiopurines, which results in a number of potentially effective or toxic metabolites. Hepatotoxicity and myelotoxicity are usually considered dose-dependent reactions [20, 21].

Nausea is usually the most frequent thiopurine-related adverse event. Although it is not a life-threatening adverse effect, it severely limits treatment with thiopurines, as more than 80 % of patients with nausea have to discontinue the treatment with these drugs. Infection is a relatively common indirect toxicity, being observed in approximately 7 % of patients [4]. In addition to bacterial infections, viral infections are associated with use of azathioprine and mercaptopurine. The herpesviruses, specifically Epstein-Barr virus, cytomegalovirus, varicella-zoster virus, and herpes simplex virus, have all been reported to cause some rare but serious complications in IBD patients receiving azathioprine/mercaptopurine. Most herpesvirus infections are probably unrecognized and manifest as self-limited viral syndromes, but life-threatening complications such as disseminated varicella-zoster, pneumonitis, and viral-mediated hemophagocytic syndrome have been reported [4].

In the same way, it has been suggested that there is an increased risk for the development of some malignancies in IBD patients under thiopurine therapy. The relationship between thiopurines and development of cancer, especially hematological malignancies such as lymphomas, remains a controversial topic. A meta-analysis of the risk of malignancy associated with the use of immunosuppressive drugs suggested that the administration of immunosuppressive drugs in IBD patients probably does not confer a significantly increased risk of malignancy compared with patients with IBD who are not receiving these agents [22]. There was not a significant difference when the authors analyzed the length of exposure to immunosuppressants or whether the patients had Crohn’s disease or UC.

The issue of the relationship between lymphoma and IBD is complex due to the effects caused by the disease per se and by the disease activity and because different IBD therapies clearly overlap. A meta-analysis of Kandiel et al. identified six cohort studies with azathioprine or mercaptopurine exposure that have been specifically designed to evaluate cancer as adverse outcome [23]. The total number of observed cases was 11, with a pooled relative risk of 4.18. Recently, results from the very large French population-based CESAME study suggest a doubling of the risk of lymphoma in patients with IBD, with the majority of cases occurring in association with immunosuppressive therapy [24]. However, because these data were obtained from observational studies, it is not possible to exclude the possibility that disease severity is a confounding factor. Another recent meta-analysis by Kotlyar et al. found that there is a higher risk for lymphoma in patients using azathioprine or 6-MP [25]. The overall Standard Incidence Ratio (SIR) for lymphoma was 4.49 (95 % CI, 2.81–7.17), ranging from 2.43 (95 % CI, 1.50–3.92) in eight population studies to 9.16 (95 % CI, 5.03–16.7) in ten referral studies. Population studies demonstrated an increased risk among current users (SIR = 5.71; 95 % CI, 3.72–10.1), but not in former users of azathioprine or 6-mercaptopurine (SIR = 1.42; 95 % CI, 0.86–2.34). The level of risk became significant after 1 year of exposure to azathioprine or 6-mercaptopurine. Also, men have greater risk than women (RR = 2.05; P < .05); both sexes were at increased risk for lymphoma (SIR for men = 3.60; 95 % CI, 2.68–4.83; SIR for women = 1.76, 95 % CI, 1.08–2.87). Also, there was found to be an age-dependent risk for lymphoma with patients younger than 30 years having the highest RR (SIR = 6.99; CI, 2.99–16.4); younger men had the highest risk. The absolute risk was highest in patients older than 50 years (1:377 cases per patient-year).

As an overall conclusion, the consensus about the relationship between immunosuppressants and lymphoma is that the risk is of small magnitude and, in any case, the beneficial effects exerted by these drugs on IBD patient outcomes would clearly outweigh the risk caused by the drug itself.

In addition to lymphoma, there has been a concern that azathioprine may be related to other malignancies. However, this association remains controversial. Azathioprine and mercaptopurine do not increase the risk of colorectal cancer; moreover, it seems to have a protector effect by controlling mucosal inflammation [26]. On the other hand, an increased risk of nonmelanoma skin cancer is well recognized in the immunosuppressed transplant population as well, and it has also been reported in IBD [27, 28]. A recent meta-analysis has reinforced this association. However, this finding has not been confirmed by all authors [29].

Determining TPMT genotype or enzyme activity phenotype prior to initiating azathioprine therapy has the potential to reduce myelosuppression by 25–50 %. Approximately, 90 % of the population has a wild-type genotype with a normal activity of the enzyme; about 11 % of the population has intermediate activity and 0.3 % of the population has low activity [30, 31]. However, some patients will still experience myelosuppression despite a normal TPMT activity, and thus, all patients undergoing thiopurine treatment will need regular complete blood count monitoring during follow-up [32]. Nevertheless, determination of TPMT activity prior to the administration of these drugs would allow for identification of those patients who should avoid thiopurine treatment due to a very high risk of severe myelotoxicity. This has been proven to be a cost-benefit strategy [33, 34].

Inosine triphosphate pyrophosphatase (ITPA) is another enzyme involved in the metabolism of thiopurines. The deficient function of this enzyme leads to an abnormal accumulation of potential toxic metabolites. Indeed, some authors have reported that ITPA polymorphisms are associated with allergic reactions to thiopurines. However, this association has not been confirmed by all authors. Until further studies confirm the utility of ITPA testing, its use cannot be recommended in clinical practice [4].

The measurement of serum thiopurine metabolite levels has been proposed by some authors as a useful tool to optimize treatment with azathioprine and mercaptopurine [35, 36]. However, the utility of this strategy has been debated in the literature and even referred to as the “metabolite controversy.” In 2000, Dubinsky et al. showed that, in children, higher thioguanine levels corresponded to a higher frequency of response [37]. In fact, 65 % of patients with thioguanine levels in the therapeutic range had a beneficial response as opposed to the 27 % with suboptimal levels [37]. Similar results have been reported by others, but this has not been consistent among all groups [38–40]. As an example, a recent large multicenter trial did not support the determination of thioguanine levels to predict treatment outcomes, and no useful serum metabolites threshold value to adjust the drug’s dose was identified [40].

In summary, although thioguanine is an important metabolite associated with both the efficacy and toxicity of azathioprine/mercaptopurine, other metabolites are likely to also play an essential role. At the current time, we believe that the data are insufficient to support routine monitoring of mercaptopurine and thiopurine metabolites.

As thiopurine therapy is associated with a wide range of adverse events, more data are required to determine the optimal duration of therapy, particularly for patients in remission. Data regarding the long-term efficacy of thiopurines and about IBD patient outcomes after the cessation of these drugs are scarce. Several trials have shown that, irrespective of the duration of remission, withdrawing thiopurine therapy increases the risk of relapse in Crohn’s disease patients [41].

In the case of UC, Holtman et al. performed a multicenter retrospective study aiming to evaluate the long-term efficacy of thiopurine therapy [42]. Data from 358 UC patients were analyzed according to the duration of the treatment (less than 3 years, 3–4 years, and longer than 4 years). In this study, the risk of relapse and the need for steroid treatment were significantly lower after initiating thiopurine treatment. The authors found that discontinuation of the thiopurines after 3 years of treatment was associated with a high risk of relapse. Therefore, authors concluded that treatment with these drugs should be maintained for at least 4 years [42].

These benefits of the long-term treatment with thiopurines have been confirmed by other authors. In this respect, in a study performed on 622 IBD patients (346 UC), it was found that the beneficial effect of azathioprine remains at least after 5 years of treatment [43]. In the same way, Chebli et al. evaluated the efficacy of azathioprine on a cohort of 42 UC patients who had been on this drug for at least 3 years. They found that the remission rate and the steroid-sparing effect were maintained at the end of follow-up [43].

The consequences of the discontinuation of thiopurine treatment in UC patients were evaluated by Hawthorne et al. [44]. Seventy-nine UC patients who had been taking azathioprine for at least 6 months were included. Patients were randomized to receive azathioprine or placebo for 12 months. This study showed that the protective effect of azathioprine in the maintenance of remission in UC lasts for at least 2 years in patients who have achieved remission while taking the drug. Moreover, discontinuation of the treatment led to a double-risk of relapse, when compared with patients who maintained treatment (36 % in the azathioprine group vs. 56 % in the placebo group).

Similar results were reported by Cassinotti et al. [45]. These authors included 127 patients who were in steroid-free remission at the time of azathioprine withdrawal, and they were followed-up for a median of 55 months or until relapse. Sixty-seven percent of patients relapsed at a median of 12 months after withdrawal of the drug. Several predictive factors for relapse were identified in this study: lack of sustained remission during azathioprine maintenance, extensive colitis, and treatment duration, with a higher risk of relapse among those patients who had received short treatments (3–6 months) compared with those who had been under thiopurines for longer than 48 months.

All these data suggest that the efficacy of azathioprine in UC patients remains in long-term treatment and that withdrawal of the drug in patients who were in remission is associated with a high risk of relapse. Therefore, in the same sense as in transplant patients, thiopurine therapy should probably be indefinitely maintained once the remission is reached in UC patients.

In conclusion, thiopurine withdrawal trials have shown that, irrespective of the duration of remission, withdrawing thiopurine therapy increases the risk of relapse both in CD and UC. Given these results, continuation may be favorable in the majority of patients. Nevertheless, there remain a minority who needlessly continue thiopurine therapy and are exposed to the associated risks. Accordingly, the identification of patients who, despite cessation of thiopurine therapy, will be at a low risk of relapse is of particular interest.