Short-term anti-TNF therapy with surgical closure versus anti-TNF therapy in the treatment of perianal fistulas in Crohn's disease (PISA-II): a patient preference randomised trial

Background: Guidelines on Crohn's perianal fistulas recommend anti-tumour necrosis factor (TNF) treatment and suggest considering surgical closure for patients with surgically amenable disease. However, long-term outcomes following these two strategies have not been directly compared. The aim of this study was to assess radiological healing in patients who received short-term anti-TNF treatment and surgical closure compared with those who received anti-TNF treatment alone. Methods: The PISA-II trial was a multicentre, patient preference study done in nine hospitals in the Netherlands and one hospital in Italy. Adult patients with Crohn's disease and an active high perianal fistula with a single internal opening were eligible for inclusion. After counselling, patients with no treatment preference were randomly assigned (1:1) using random block randomisation (block sizes of six without statification), to 4-month anti-TNF therapy and surgical closure or anti-TNF therapy for 1 year, after seton insertion. Patients with a treatment preference received their preferred therapy. The primary outcome was radiological healing assessed by MRI at 18 months, defined as a complete fibrotic tract or a MAGNIFI-CD (Magnetic Resonance Index for Fistula Imaging in Crohn's Disease) score of 0, assessed according to the intention-to-treat principle. Secondary outcomes included clinical closure, number of patients undergoing surgical reintervention and number of reinterventions, recurrences, and impact on quality of life measured by the Perianal Disease Activity Index (PDAI). Analyses were performed on an intention-to-treat basis and additionally an as-treated analysis for radiological healing and clinical closure. This study was registered at the Dutch Trial Registry, NL7625, and with EudraCT, 2018-002064-15, and is closed to accrual due to completion. Findings: Between Sept 14, 2013, and Dec 7, 2019, 94 patients were enrolled onto the trial, of whom 32 (34%) were randomly assigned and 62 (66%) chose a specific treatment. 38 (40%) patients were assigned to the surgical closure group and 56 (60%) patients to the anti-TNF group. At 18 months, radiological healing was significantly more common in the surgical closure group (12 [32%] patients) than in the anti-TNF group (five [9%] patients; p=0·005). By contrast, clinical closure was not significantly different between the two treatment groups (26 [68%] patients in the surgical closure group vs 29 [52%] patients in the anti-TNF group; p=0·076). Significantly fewer patients required a reintervention in the surgical closure group than in the anti-TNF therapy group (five [13%] patients in the surgical closure group, median one reintervention [IQR one to three] vs 24 [43%] patients in the anti-TNF group, median two reinterventions [one to two]; p=0·005). Among patients who reached clinical closure during follow-up, four (14%) of 29 in the surgical closure group and five (16%) of 31 in the anti-TNF therapy group had a recurrence, which occurred only in patients without radiological healing. PDAI was significantly lower in the surgical closure group than in the anti-TNF group after 18 months (p=0·031). Adverse events and serious adverse events were similar in both treatment groups and mostly entailed reinterventions. Ten (11%) patients had side-effects associated with anti-TNF treatment. Two serious adverse events unrelated to study treatment occurred (appendicitis and myocardial infarction). One patient died from a tongue base carcinoma, unrelated to study treatment. Interpretation: Short-term anti-TNF treatment combined with surgical closure induces long-term MRI healing more frequently than anti-TNF therapy in patients with Crohn's perianal fistulas. These data suggest that patients with Crohn's perianal fistula amenable for surgical closure should be counselled for this therapeutic approach. Funding: Netherlands Organisation for Health Research and Development and Broad Medical Research Program.