Usefulness of salicylate and thiopurine coprescription in steroid-dependent ulcerative colitis and withdrawal strategies

Abstract

5-aminosalicylic acid (5-ASA) and thiopurines (azathioprine and mercaptopurine) are the most common drugs used as a maintenance treatment for ulcerative colitis. A considerable proportion of these patients develop corticosteroid dependency, and thiopurines are the standard treatment in this scenario. Dual prescriptions of both thiopurines and 5-ASA are common practice in steroid-dependent ulcerative colitis, in an attempt to optimize the efficacy of therapy. On the one hand, the potential protective role of 5-ASA against colorectal cancer argues in favour of prescription of both medications. The possible synergism between the two drugs, because of the inhibition of thiopurine methyltransferase (TPMT) enzyme activity by 5-ASA, has been postulated as another justification for dual prescription. However, existing evidence does not support that this combined strategy is superior to monotherapy with thiopurines. On the other hand, in patients showing prolonged disease remission, the possibility of discontinuing maintenance treatment can be considered on an individualized basis. The high frequency of relapses after thiopurine withdrawal should always be taken into account, but the potential adverse effects of the medication also need to be considered. A properly indicated treatment with thiopurines may need to be continued for life in many patients.

Introduction

Ulcerative colitis (UC) is an inflammatory bowel disease characterized by the alternation of symptomatic periods and periods of disease remission. After achieving remission of a flare up as a result of medical treatment, the objective is to prevent UC relapse. Various potentially useful drugs are available for this purpose, and the choice depends on the characteristics of the disease: its extent, flare-up response to treatment, and the behaviour after corticosteroid withdrawal, with the possible appearance of corticosteroid dependency. The drugs used as maintenance treatment for UC are fundamentally 5-aminosalicylic acid (5-ASA) and the thiopurinic immunosuppressors [Sands, 2006; Sutherland and MacDonald, 2006; Hanauer, 2004].

For over 50 years, 5-ASA, sulfasalazine and mesalazine have been the drugs most widely used to treat UC, starting with the introduction of sulfasalazine by Svartz in 1942 [Svartz, 1942], and these are the substances normally used as maintenance therapy in patients with UC [Travis et al. 2008; Sutherland and MacDonald, 2006; Gisbert et al. 2002]. The thiopurines, azathioprine (AZA) and mercaptopurine (MP), represent the treatment of choice in UC patients who develop corticosteroid dependency [Stange et al. 2008], and these drugs offer high efficacy in this clinical scenario [Gisbert et al. 2009; Travis et al. 2008; Ardizzone et al. 2006; Ghosh et al. 2006; Sands, 2006; Lopez-Sanroman et al. 2004; Fraser et al. 2002]. A considerable proportion of patients who develop corticosteroid dependency have previously been subjected to maintenance therapy with salicylates. When the course of the disease requires taking a further step up the therapeutic ladder with the administration of thiopurines, doubts may arise as to whether or not 5-ASA should be maintained. Even in the absence of 5-ASA therapy, the question may arise as to whether it is worth adding the latter to AZA/MP treatment, in an attempt to optimize UC therapy. The metabolism of AZA/MP is well known, and involves a series of enzyme reactions with methylation by thiopurine methyltransferase (TPMT), oxidation mediated by xanthine oxidase to yield thiouric acid, or catabolism through the mediation of hypoxanthine— guanine—phosphoribosyl transferase, which leads to the production of 6-thioguanine nucleotides, which are fundamentally responsible for the activity of these drug substances. One of the most widely studied of these implicated enzymes is TPMT [Gisbert et al. 2007a], the activity of which is influenced by both genetic polymorphisms and different drugs, possibly including 5-ASA. In view of the above considerations, a review is made of the existing evidence on the usefulness of thiopurine and 5-ASA coprescription.

On the other hand, in patients with steroid-dependent UC showing prolonged remission, the doubt may arise in both the patient and prescribing physician as to whether treatment can be suspended in view of the stability of the disease. However, in order to be able to make a decision, due evaluation is required of the potential risks and benefits of suspending treatment. On the one hand, suspension can give rise to undertreated disease and the risk of reactivation, while on the other hand the indefinite maintenance of treatment increases the possibility of drug adverse effects [de Jong et al. 2003]. The present study thus reviews the existing evidence to help take decisions in this clinical scenario.

Frequency of thiopurine and 5-ASA coprescription, and its bases

In some series, the frequency of AZA and 5-ASA coprescription in patients with steroid-dependent UC exceeds 70% [Actis et al. 2009]. In a survey conducted in the United Kingdom involving 97 gastroenterologists (38% response rate with respect to the total physicians surveyed), 79% of those interviewed claimed to apply coprescription. In addition, 20% of their UC patients received combined treatment with AZA and 5-ASA, and 94% of the physicians recommended life-long 5-ASA use [Mayberry et al. 2009]. The possible existence of synergism between the two drug substances, as a result of the inhibition of TPMT enzyme activity by 5-ASA, described in in vitro studies [Xin et al. 2005; Szumlanski and Weinshilboum, 1995], has been postulated as a justification of coprescription. The mentioned enzyme is of particular interest, owing to its intervention in several steps of thiopurine metabolism. In this context, it has been suggested that in patients with high TPMT enzyme levels (i.e. hypermethylators), the administration of 5-ASA may help to optimize thiopurine efficacy by directing metabolism of the drug towards the production of 6-thioguanine nucleotides (active thiopurine metabolites), and avoiding the accumulation of potentially hepatotoxic methylated derivatives [Lowry et al. 2001].

However, a considerable number of in vivo studies have been unable to demonstrate significant changes in TPMT levels associated with 5-ASA therapy [Lowry et al. 2001] or with the discontinuation of such therapy [Dewit et al. 2002]. Xin and colleagues have reported that of the different 5-ASA drugs, only sulfasalazine exhibits in vivo TPMT inhibiting potential, although the authors did not analyse the impact of this fact upon treatment efficacy [Xin et al. 2005]. The technical procedure used to measure erythrocyte TPMT activity could explain the absence of changes in TPMT activity with 5-ASA use reported by in vivo studies. In effect, during preparation of the erythrocyte lysates, all metabolites are washed out. Accordingly, if 5-ASA metabolites are removed during TPMT analysis, this might lead to normalization of inhibition of TPMT activity [Gilissen et al. 2005].

In a recent study comparing a group of patients receiving cotreatment versus another administered AZA monotherapy, Daperno and colleagues observed no differences in TPMT activity between the two groups, and no relationship was found between the 5-ASA dose used and TPMT activity. Likewise, no differences were found in 6-thioguanine and 6-methylmer-captopurine levels between the two patient groups [Daperno et al. 2009]. These results differ from those recorded in other intervention studies in which the addition of 5-ASA to stable thiopurine doses was found to significantly increase the 6-thioguanine levels [de Boer et al. 2007; Gilissen et al. 2005]. Daperno and colleagues postulated that after early interference, adaptation occurs, leading to minor interference between the two drugs [Daperno et al. 2009]. This hypothesis would be supported by the findings of a study in patients with surgically treated Crohn's disease, in which both AZA and mesalazine did not alter the TPMT levels after 1 year of treatment [Dilger et al. 2007].

Lastly, it could be speculated that the interaction between 5-ASA and thiopurines would have a greater potential impact upon patients with low TPMT activity, increasing the risk of bone marrow toxicity, although this would have to be confirmed by adequately designed prospective studies.

Efficacy of thiopurine and 5-ASA coprescription

The interaction between 5-ASA and thiopurines could have an influence on the management of patients with inflammatory bowel disease in a number of aspects: AZA/MP dose, efficacy and potential adverse effects. There are contradictory data and opinions on the possible usefulness of coprescription. Some investigators suggest that coprescription is useful and increases the efficacy of AZA/MP and the levels of 6-thioguanine nucleotides (active thiopurine metabolites) [Stocco et al. 2008; de Boer et al. 2007; Hande et al. 2006; Dewit et al. 2002]. On the other hand, the use of 5-ASA as colon cancer chemo-prophylaxis would be another element in favour of coprescription in patients with steroid-dependent UC [Rubin et al. 2008; Van Staa et al. 2005; Velayos et al. 2005]. Patients with long-evolving UC are at greater risk of developing colon cancer than the general population, with a prevalence of 3.7% [Eaden et al. 2001]. 5-ASA treatment appears to reduce this risk by 40–75%, with a number needed to treat (NNT) of 7 patients after 30 years [Van Staa et al. 2005; Eaden et al. 2000]. Daily sulfasalazine doses of over 2g, or mesalazine doses above 1.2g, appear to offer increased preventive benefit [Velayos et al. 2005]. The chemoprophylactic action of 5-ASA appears to be dependent upon binding of the drug to a potent colonocyte differentiating factor [Dubuquoy et al. 2006]. This potential beneficial effect of 5-ASA in reducing colon cancer risk at an acceptable cost could be important in justifying the administration of maintenance therapy for years, apart from the inherent therapeutic effect obtained in terms of the prevention of disease relapse, and would support 5-ASA and AZA/MP coprescription [Andrews et al. 2009].

In contrast, other authors have reported that the addition of 5-ASA does not increase the efficacy of AZA/MP in steroid-dependent UC. Indeed, it may even increase toxicity, by directing drug metabolism towards the production of 6-thioguanine nucleotides, thus increasing the risk of bone marrow toxicity [Lowry et al. 2001] and the like-liness of having to suspend AZA/MP, which is the main source of therapeutic effects in patients with steroid-dependent UC. In this sense, Shah and colleagues have described the appearance of more adverse effects in patients subjected to coprescription than in those administered thiopurine monotherapy [Shah et al. 2008]. On the other hand, the fact of administering a larger number of drugs in the context of prolonged maintenance therapy could adversely affect patient adherence, this being a key factor for the success of a chronic disorder such as UC [Kane, 2006; López San Román et al. 2005; Kane et al. 2003].

A recent systematic review of the efficacy of thiopurine and 5-ASA coprescription in UC concluded that the existing evidence obtained from clinical trials is insufficient to demonstrate whether simultaneous AZA and 5-ASA therapy improves the results of AZA in monotherapy [Andrews et al. 2009]. The conclusions of the mentioned review point to AZA and 5-ASA as compatible treatments without a significant increase in adverse effects, although monitoring of the possible development of bone marrow toxicity is recommended if 5-ASA is added to treatment, or if the dose of this drug is increased [Andrews et al. 2009]. In this context, the guidelines of the European Crohn's and Colitis Organization (ECCO), in the section on thiopurine treatment for UC, indicate that the addition or maintenance of oral mesalazine can be recommended, with special monitoring of possible marrow toxicity, although this recommendation is based on low levels of evidence (level of evidence 5, degree of recommendation D) [Travis et al. 2008].

What happens on suspending AZA/MP in remitting steroid-dependent UC?

UC patients receiving treatment with AZA/MP often ask for how long such treatment is needed. Can the medication be suspended at some point? No definitive answer to this question can be offered. The ECCO guidelines on UC [Travis et al. 2008] regarding the duration of thiopurine treatment indicate that because of the lack of evidence, no recommendation can be made, although the prolonged use of these drugs can be considered if necessary (level of evidence 4, degree of recommendation D).

A recent multicentre retrospective study offers interesting information for attempting to offer answers to our patients [Cassinotti et al. 2009]. In the mentioned study, Cassinotti and colleagues presented a series of 127 patients in which treatment was suspended upon decision by either the patient or the physician (3/4 of all cases), or as a result of intolerance. Approximately 75% of the subjects had been in disease remission without flare ups, while the rest had shown some degree of activity during treatment with AZA. As an interesting observation related the previous section of our study, the authors found cotreatment with 5-ASA to be the only predictor of sustained disease remission during treatment with AZA. Following AZA suspension, 5-ASA was prescribed in most of the patients, although 21% were left without maintenance therapy. Clinical relapse during follow up was both frequent and early: in one third of the patients after 12 months, one half after 2 years, and in two thirds after 5 years [Cassinotti et al. 2009], these figures being globally similar to those reported for Crohn's disease [Treton et al. 2009]. The relapse-predicting factors after AZA suspension were: (1) treatment duration, with less-frequent relapses among subjects who had received therapy for over 2 years; (2) sustained remission during treatment, with a 59% relapse rate among patients showing sustained remission, versus 91% in those who had not achieved sustained remission; and (3) the extent of UC, with an increased relapse risk in extensive disease than in patients with left colitis. However, the possible influence of the extent of the disease must be viewed with caution, since approximately one half of all UC patients experience changes in disease extent over time [Farmer et al. 1993]. In the study published by Cassinotti and colleagues, colectomy was required in up to 10% of the patients (4% after 1 year and 9% after 3 years) in which AZA was suspended, such surgery in turn involving a nonnegligible degree of morbidity—mortality [Cassinotti et al. 2009]. Although the global data are similar to those published by Treton and colleagues in luminal Crohn's disease [Treton et al. 2009], on analysing the specific data corresponding to the patients with a favourable course and who remained in remission during treatment with AZA, it was seen that AZA suspension could be decided with good results in up to two thirds of the patients with Crohn's disease versus only one third of the patients with UC. Therefore, it seems that AZA suspension implies a poorer course in UC than in Crohn's disease.

The only randomized trial on AZA withdrawal in 79 patients with steroid-dependent UC involving a follow up of 1 year reported a 36% relapse rate in the group maintaining active treatment (AZA), versus 59% in the placebo group (p = 0.03). The differences were evident on analysing the subgroup of patients with prolonged disease remission (31% versus 61%; p < 0.01) [Hawthorne et al. 1992]. Likewise, data are available on the experience of a single centre over a period of 30 years in 222 patients, although this series combines patients with UC and Crohn's disease. In this study, the relapse rate was 37% after 1 year, 54% after 2 years, and 75% after 5 years, with no influence on the part of the duration of AZA therapy [Fraser et al. 2002].

A European multicentre but retrospective study offers interesting information, considering the large number of patients involved: 818 patients with Crohn's disease and 358 with UC. The patients were stratified according to the duration of AZA therapy (less than 3 years, 3–4 years and over 4 years). The reasons for suspending AZA were physician decision (50%), patient decision (25%) and toxicity or other reasons in the rest of the cases. The results, and thus the recommendations of the authors, differed for Crohn's disease and UC [Holtmann et al. 2006]. In the case of Crohn's disease, AZA suspension after 3–4 years did not lead to disease reactivation or to an increase in corticosteroid requirements, although when therapy was prolonged for over 4 years a significant decrease was noted in the number of flare ups and in corticosteroid requirements. As a result, in asymptomatic patients with Crohn's disease with no need for corticosteroid therapy after 3–4 years of AZA treatment, the authors suggested that the latter medication could be suspended, while in patients with clinical disease activity or who required corticosteroids at some point during treatment with AZA, the latter should be maintained. In contrast, in patients with UC, treatment for over 4 years significantly reduced the corticosteroid requirements and number of flare ups, and when AZA was suspended, the observed tendency was towards an increase in flare ups and corticosteroid use. The authors therefore considered it to be of benefit to continue AZA for more than 4 years in UC patients [Holtmann et al. 2006].

Undoubtedly, the high frequency of relapses after AZA/MP suspension must be taken into account by both the patient and the supervising physician, before taking the decision to suspend an effective treatment. On the downside, we have the potential adverse effects of the thiopurines, which are seen in approximately 20% of all cases [de Jong et al. 2003; Bowen and Selby, 2000], and which lead to treatment suspension in about 10% of all cases [Gisbert et al. 2007c; Lichtenstein et al. 2006]. Some of these adverse effects, such as bone marrow toxicity and liver toxicity, can manifest even years after the start of treatment [Gisbert and Gomollon, 2008; Gisbert et al. 2007b]. Another important aspect is a certain increase in the risk of developing lymphoproliferative disorders, described by a number of investigators [Kandiel et al. 2005], although globally the absolute risk is low. All of these factors should be taken into consideration when pondering the suspension of thiopurine therapy in patients with long-remitting UC.

What is the effect of suspending the aminosalicylate in steroid-dependent UC subjected to AZA/MP and 5-ASA coprescription?

5-ASA suspension in patients subjected to combined treatment with this drug and AZA/MP may give rise to a drop in 6-thioguanine levels [Gilissen et al. 2005; Dewit et al. 2002], although the true effect of such variations in thiopurine metabolite levels upon the clinical efficacy of these drug substances is not known. There have been isolated reports of patients subjected to coprescription in which the suspension of mesalazine was accompanied by UC relapse despite treatment with AZA [Actis et al. 2008; Stocco et al. 2008]. As a result, monitoring of possible relapse has been recommended if the decision is taken to suspend 5-ASA in UC patients subjected to coprescription [Andrews et al. 2009], although the supporting evidence is very limited.

On the other hand, on examining the advantages and risks of 5-ASA suspension, it must be remembered that discontinuation of the drug could lead to a loss of the protective effect of these drugs against colorectal cancer in UC patients [Rubin et al. 2008; Van Staa et al. 2005; Velayos et al. 2005]. In turn, however, such a measure could improve adherence to therapy, since the patient would be required to take fewer drugs and fewer tablets each day [Kane, 2006].

Some authors have suggested the possibility of suspending salicylate treatment in UC if the patient has been free of symptoms for a long period of time and is found to be in clinical, histological and endoscopic disease remission [Ardizzone et al. 1999]. Other experts and clinical guides recommend long-term maintenance therapy with 5-ASA in order to secure optimum treatment results, based on the efficacy of such drugs in reducing the risk of colorectal cancer [Travis et al. 2008; Sandborn, 2006; Carter et al. 2004]. Nevertheless, the lack of conclusive data from clinical trials makes it necessary to establish this decision on an individualized basis according to the characteristics of the patient and the disease. Thus, in the context of steroid-dependent UC, in coprescribed patients and on contemplating suspension of some of the drugs in response to patient request or with the purpose of improving adherence to therapy, we could consider withdrawing 5-ASA and maintaining AZA/MP, since the latter represents the active treatment reverting the situation of corticosteroid dependency.

Conclusions

The coprescription of thiopurines and 5-ASA is common practice in patients with steroid-dependent UC, although the existing evidence does not allow us to confirm that this strategy is superior to monotherapy with AZA/MP. The potential protective role of 5-ASA against colorectal cancer argues in favour of coprescription. However, the associated increase in the number of drugs administered as long-term maintenance therapy can worsen patient adherence: good adherence is a key factor for the success of treatment.

In patients showing prolonged disease remission as confirmed by the clinical, endoscopic and histological findings, the possibility of suspending maintenance treatment may be considered. Such a decision must be established on an individualized basis according to the characteristics of the patient (initial indication, adherence to therapy), the disease (severity and previous course) and patient opinion. In this context, the patient must be made aware not only of the high probability of disease relapse if AZA/MP is suspended in the context of steroid-dependent UC, but also of the potential adverse effects of the medication. However, if thiopurine treatment was well indicated, it may have to be continued for life.