Discussion
Current evidence advocating the use of thiopurines as prophylaxis against post-surgical recurrence is poor. There is marked variation in study design, medication dosage, drug used, sample size, duration of follow-up, outcome measures and definition of control groups leading to uncertainty regarding treatment effect and indication. This heterogeneity underpins the lack of clinical confidence and robust guidelines in thiopurine use after intestinal resection. Of the four randomised trials, only three are published in full, and only two are properly blinded. It is not surprising therefore that current guidelines differ in their recommendation for use and this ultimately manifests in differing prescribing practices, with clinicians opting to use experience to guide treatment as opposed to strong trial evidence.
Both the CESAME group and meta-analyses of previous cohort data have described increased risk of lymphoma in IBD patients, although the absolute risk remains small and causality remains unclear. Given this increasing appreciation of long-term risk of thiopurine-associated neoplasia, it becomes increasingly pertinent to justify the use of toxic agents to our patients based on methodologically sound data. This is of particular relevance, if proposed use is as prophylaxis compared to treating active disease, in which their efficacy is well established.
None of the four randomised clinical trials or longest retrospective and prospective observational studies have consistently shown risk factors that predict clinical, endoscopic or surgical recurrence. Only smoking has strong evidence in predicting post-operative recurrence with a 2.5 fold increased risk of re-operation compared to nonsmokers in meta-analysis. Identification of a high-risk phenotype within individual patients would be valuable to address the uncertainty clinicians' face in preventing disease recurrence and targeting treatment effectively. However, putative risk factors thus far have provided only a broad insight at best and predicting recurrence in an individual remains poor.
Evidence for anti-TNF use is limited to one randomised trial with no guidelines advocating its use routinely as primary prophylaxis. However, the limited data suggest a strong treatment effect that warrants assessment in large randomised trials. Regueiro and colleagues however considered the benefit so stark that they consider IFX use routinely post-operatively in those with ileal penetrating disease or with multiple previous resections.
In summary, when counselling our patients regarding the merits of immunosupression post-intestinal resection, we must offset the risks of prophylactic treatment vs. benefits of long-term mucosal healing. The current evidence to aid us in this assessment is heterogeneous, and the final results from ongoing trials in UK and Australia keenly awaited.