Prudence is never excessive: Caution warranted in 5-aminosalicylic acid derivatives withdrawal for ulcerative colitis

Abstract

This invited commentary discusses the recent study by Atay et al, which investigated relapse rates following the spontaneous withdrawal of maintenance 5-aminosalicylates in ulcerative colitis. The discussion focuses, in this patient setting, on the possible reasons that might prompt clinicians to pursue such exit strategies, and on the importance of exercising caution in these decisions, given the extremely narrow subsets of patients for whom international guidelines allow any degree of leeway.

Key Words: Ulcerative colitis; Inflammatory bowel disease; 5-aminosalicylic acid derivatives; Withdrawal; Exit strategy

Core Tip: The discontinuation of 5-aminosalicylates in ulcerative colitis should be considered only in highly selected, low-risk patients. Notwithstanding this requirement, however, the evidence generated to date does not yet permit the formulation of a robust and definitive operational pathway. Strict monitoring and adherence to European Crohn’s and Colitis Organisation guidelines remain essential to balance the potential benefits against the risks of relapse and colorectal cancer.

TO THE EDITOR

We read with great interest the retrospective study by Atay et al[1], in which the authors selected, from their retrospective records, a subset of ulcerative colitis (UC) patients in clinical remission who had spontaneously discontinued maintenance 5-aminosalicylic acid derivatives (5-ASA) monotherapy due to poor adherence, intending to identify how many of them subsequently developed a relapse and after what interval. With a median retrospective follow-up of approximately three years after 5-ASA discontinuation, the authors observed in this subset of patients (thirty out of the 880 initially included in the analysis) a gradual decrease in the remission rate (89% at the first year, 73% at the second, and finally 64% at the third year).

Therapeutic adherence in UC is undoubtedly a matter of utmost importance, considering that it is estimated that just over 30% of patients forget doses of 5-ASA, thereby compromising therapeutic efficacy, and that not all modifiable factors on which to intervene to restore full adherence have yet been identified[2]. Psychotherapy has also been postulated as a possible tool for this purpose[3]. In addition, the coronavirus disease 2019 pandemic has further exacerbated this problem, as shown by our data collected during the peak years of the pandemic in Italy from outpatients with inflammatory bowel disease[4]. Notwithstanding what is currently available in UC, it is beyond doubt that the studies produced over the past decades have provided high-certainty evidence that 5-ASA is markedly superior to placebo in maintaining clinical remission, with a sustainable safety profile[5]. Moreover, and even more importantly, these agents, at a dosage of at least 1.2 g/day, exert a chemopreventive effect against colorectal cancer and colorectal dysplasia - events to which these patients, particularly those with chronically active and/or long-standing disease, are especially vulnerable[6].

Regardless of these premises, the prospect of having robust stopping rules on which to base exit strategies in patients who respond well to the prescribed treatments and can maintain a solid, long-lasting remission with maintenance doses of 5-ASA is clearly of interest, not least to avoid overtreatment. On the other hand, however, there remains the concept of the chronicity of the disease, the difficulty of having reliable tools to predict relapses over time, and the theoretical possibility of having to face potential acute severe UC that the complete absence of treatment might trigger. It is so a real challenge, when dealing with chronic diseases with a not entirely predictable course that we have managed to control with effective induction therapy and subsequent maintenance therapy with a favourable safety profile such as 5-ASA, to weigh in the long term, on the scales of the risk-benefit profile, the primary need of primum non nocere against the underlying suspicion of possible, even severe, potential disease exacerbations.

To this end, we know that the European Crohn’s and Colitis Organisation (ECCO) produced a topical review in 2018 precisely for this purpose[7]. In our view, this review highlighted several concepts that we consider essential when approaching treatment withdrawal: (1) The patient should be reassessed to ensure objectively demonstrable remission based on biochemical, endoscopic, and clinical data; (2) Patient preference is a key pillar in the decision-making process regarding treatment discontinuation; and (3) The patient in whom such a choice is made must be optimally monitored to ensure that any relapses are not left undiagnosed.

The same panel, moreover, within the framework of this topical review, concluded that although the optimal duration of maintenance treatment with 5-ASA in patients with UC is far from being fully clarified, such treatment should not be discontinued even during remission, precisely because the majority of the available high-quality studies (i.e., randomised controlled trials) assess clinical remission generally between 6 and 12 months, suggesting that this option could be considered only in a few highly selected groups of patients[7]. In general, reference was made to patients with minimal disease extent as a baseline characteristic (e.g., proctitis), sustained remission over several years with no significant history of more than one disease flare, and no previous need for steroids[7]. Without taking into account, in our view, a parameter such as mucosal healing, which should perhaps be given the highest consideration, given its association with long-term clinical remission as well as with a drastic reduction in the risk of colectomy[8]. All this is to avoid hasty decisions in patients who may have achieved clinical-endoscopic remission but not deep remission, including histological remission, with residual histological activity that could subsequently act as a trigger for future relapse. Deep remission, moreover, is a concept widely debated in the literature, but it is generally defined as the concurrent presence of clinical, biochemical, endoscopic, and histological remission[9,10].

The study under discussion in this invited letter and, more generally, this debate is more readily applicable because it addresses populations of patients with UC receiving 5-ASA monotherapy, that is, patients who have not developed the need for biologics, experienced multiple therapeutic failures, or faced the risk of surgery with a history of acute severe UC, in which cases prudence dictates entirely different approaches. This retrospective study[1] includes not only those categories of patients so strongly endorsed by ECCO, but also patients with extensive colitis (10% of the 30 who discontinued 5-ASA due to non-adherence), 10% of whom were steroid-dependent, and 6.7% and 10%, respectively, had already been treated with infliximab and thiopurines. In our view, this makes the study population a heterogeneous mix of very different patients within a minimal sample size, without even considering that 46.7% of these patients presented extraintestinal manifestations. The latter point, however, should lead us to assume the absence of treatments for such manifestations, considering that the authors had excluded patients with this condition.

Having a three-year follow-up is certainly a merit; however, the unfortunately minimal sample size strictly precludes the detection of moderate differences from a statistical point of view, thus undermining the statistical power of the entire model. Moreover, including patients who discontinued treatment due to non-adherence rather than because of an informed physician-patient decision introduces an inevitable behavioural bias with a double edge. For instance, non-adherent individuals might have a less aggressive disease course in the first place, or they might be patients with a denialist attitude towards their illness and therefore not entirely truthful in patient-reporting their symptoms. Indeed, it is known that there is a certain percentage of patients with UC who nevertheless have a very mild disease history, with even long-lasting spontaneous remissions[11]. Furthermore, these patients had a very long pre-discontinuation remission (median over ten years), which indirectly confirms the benefits of 5-ASA, given the subsequently high relapse rates observed, even in a group of only thirty patients. In our view, relapse was underestimated, since it was defined according to both clinical and endoscopic criteria; in real-world practice, for reasons of waiting lists and economic constraints, endoscopic monitoring cannot always be so timely, and conversely, had the authors relied solely on clinical data, the relapse rates might have been even higher.

In addition, a follow-up of only three years nullifies any oncological considerations. What can we draw from this for clinical practice, and how can we silence that inner voice reminding us of the risk of dysplasia and cancer in these patients - especially those with long-term disease - when the follow-up is so short? Do we truly feel safe? This is particularly relevant given that a non-negligible proportion of the thirty patients had extensive colitis without considering that dysplasia is often challenging to detect in patients with UC, particularly in long-standing disease[12]. The authors also emphasise that 75% of patients (from among the thirty included) who had relapsed eventually regained remission with 5-ASA monotherapy alone; however, two questions remain, which become even more pressing when considering the extremely small sample size. This would mean that almost 30% of patients abandoned a decade-long remission on 5-ASA, subsequently developed a relapse, and required an escalation to something more (possibly steroids), with greater morbidity and increased risks of treatment-related adverse events. Secondly, for the 75% of patients who regained remission, we lack long-term follow-up to determine whether this remission will be as durable as the one from which they had started. These were all consequences to which the patients voluntarily exposed themselves through poor adherence, and not because of medical decision-making; however, remission had been identified through medical visits, and therefore, this consideration may be extrapolated to similar cases that could prompt a deliberate interruption initiated by the physician.

At the same time, it is necessary to acknowledge that in Atay’s study, the majority of relapsed patients regained remission with a simple 5-ASA re-induction, and no severe downstream outcomes (hospitalisations, colectomy, serious adverse events) emerged during the observation period. These elements suggest that discontinuation of 5-ASA should not be framed as intrinsically “dangerous”, but rather as a choice that increases the risk of relapse and therefore requires careful candidate selection, informed counselling, close monitoring, and a predefined re-induction plan and, at the same time, the small sample size and the limited duration of follow-up call for caution in extending such results beyond low-risk scenarios, while current evidence continues to support maintenance therapy in the majority of patients.

Moreover, it should be considered that those patients who require steroids early at diagnosis and thus develop early steroid dependence often have a less favourable natural history of the disease and, indeed, frequently require aggressive treatment from the outset[13]. Therefore, this calls for caution when considering the discontinuation of maintenance 5-ASA therapy in those patients whose diagnosis was characterised by the early use of systemic steroids. A further point for discussion that this retrospective study recalls is the fact that almost half of the entire sample selected by the authors presented with an extraintestinal manifestation, although, from the data published by the authors, we cannot ascertain precisely which ones. Regardless of this, discontinuation of 5-ASA in this patient setting is noteworthy, since multiple extraintestinal manifestations often depend on the intestinal activity of the underlying UC, such as type 1 enteropathic arthritis or, for example, erythema nodosum, to cite just a few common examples[14].

A further question raised by this study, always bearing in mind the interesting ECCO topical review[7], concerns the lack of post-discontinuation monitoring data for 5-ASA non-adherence. It is indeed surprising that this work does not include data on fecal calprotectin, whose predictive power for long-term disease relapse when measured quarterly is well known[15]. Given the remarkably non-invasive nature of this biomarker, it would have been interesting to follow patients over time to see how it evolves. Although not fully elucidated, the factors associated with uncertainty regarding 5-ASA discontinuation, as identified in a recent review by Chapman et al[16], include being under 40 years of age, less than two years of sustained remission, a history of multiple flares, and extensive disease.

Therefore, knowledge of these factors may assist in selecting patients who are suitable candidates for 5-ASA discontinuation. However, although there is, moreover, no evidence that has rigorously evaluated fecal calprotectin as a definitive guide to stopping 5-ASA, it is recognized that it may serve as a non-invasive surrogate marker of histological remission (albeit without a well-defined cut-off owing to substantial heterogeneity between studies, with reported values ranging from 40.5 µg/g to 250 µg/g)[16]. Accordingly, regular monitoring of calprotectin (e.g., quarterly) may be a strategy to follow non-invasively those patients in whom discontinuation is attempted, capitalizing on its predictive value for relapse[17]. From a health economics perspective, moreover, in low-risk profiles, the exit strategy could reduce the direct costs associated with maintaining 5-ASA in terms of potentially unnecessary treatment days (i.e., days during which the patient remains in deep remission without actually taking 5-ASA). Ultimately, although the authors’ intentions are commendable, it would be helpful to extend this study with a longer post-5-ASA discontinuation follow-up, and it would be of interest to design it as a multicentre study to retrospectively examine data from multiple countries, consolidating international experience within a large pool of eligible patients, pending ad hoc prospective studies specifically designed for this purpose.

CONCLUSION

In conclusion, while appreciating the efforts of the authors, we believe that the small sample size and the considerations we have previously outlined lead us to maintain that the current approach suggested by the aforementioned international societies namely, to consider 5-ASA discontinuation only in very narrow subgroups of patients with an exceptionally favourable disease course and specific conditions, should not be in the least undermined by these data which, although interesting in their origin, do not possess the statistical power to justify extending this exit strategy approach to other disease phenotypes. While maintaining a cautious approach, we acknowledge a conditional option: A trial of 5-ASA withdrawal may be considered only in clearly low-risk patients who are closely monitored, with a predefined re-induction plan in place.