P334 Evolution of a long-term follow-up cohort of Crohn’s disease with complex perianal fistula: from antibiotic to combined AZA and anti-TNF based treatment ending up clinical and radiological healing with or without stoma
Y. Erzin*, K. Ercaliskan, I. Hatemi, K. Atay, S. Bozcan, N. Demir, A. F. Celik
Istanbul University Cerrahpasa Medical School, Gastroenterology, Istanbul, Turkey
The present study aimed to determine the evolution of complex perianal fistula (Cpfis) through clinical and radiological response rates under optimal medical treatment (MedTx) along with influence of stoma.
In this retrospective assessment of prospectively gathered data, evolution of CD patients (pts) with Cpfis between 1999 and 2015 were reviewed from chart records. Optimal triple MedTx starting with double antibiotics and adding azathioprine (AZA) and anti-TNFs at once keeping the same regimen at least for 3 months (mo) were the pre-determined parameters of follow-up. When intolerance was the case, double or single agent Tx was the alternative. Abscesses were treated with drainage and seton left 3 to 6 mo when there was no recurrence, and if there was long-term remaining seton was indicated. Tx success was stratified as complete cessation of discharge or additional closure of external orifice and, ultimately, radiological disappearance of the tract assessed by MRI. In case of MedTx failure, stoma was made with pts agreement keeping the same Tx. Each patient’s last fistula status was determined as the primary endpoint together with luminal activity. Sub-analysis of stoma patients was done.
Table 1 shows univariate analysis between groups. There were 66 Cpfis/825 (8%) CD patients. F/M and smoking ratios were same in each group. Fistula follow-up time was 48.50 ± 32.83 mo. Further, 29/66 (44%) patients had one fistula and the remaining had multiple. In addition, 44 patients (66%) were complicated by abscesses, and loose seton was applied to 38/44(86%), and 22 of 38 (58%) developed abscesses under Seton. Stoma was performed in 16 (24%) patients as a last hope, and Cpfis closure was achieved in only 8 (50%) between 2–11 mo. However, adjustment of potential relevant factors between patients with or without stoma but with the same MedTx did not show any significant healing under stoma. At the last visit 26/66 (40%) were in remission, but only 16/66 (24%) achieved radiological healing, showing no correlation with luminal activity(r = 0.12; p = 0.45). Reopening after complete clinical closure was 27% of all patients in remission, but in half of them, it occurred after radiological healing. An age-sex adjusted Cox regression analysis disclosed total anti-TNF Tx and total AZA Tx time as the only independent predictors of Tx response (p = 0.001 for both).
Even under long-term combined treatment, clinical and radiological closure of Cpfis occurred in less than half and one-quarter of CD pts, respectively, showing no correlation with luminal activity. Reopening of Cpfis in at least in 25% radiologically confirmed cases suggests that even radiological closure is not a reliable marker of Cpfis healing. Stoma does not seem to have any influence on Cpfis closure.