Search in the Abstract Database

Search Abstracts 2016

* = Presenting author

DOP074 Incidence and prediction of fistula formation in England

K. Sahnan*1, A. Askari1, S. Adegbola*1, P. Tozer1, R. Phillips1, A. Hart2, O. Faiz1

1St Mark’s Hospital, Colorectal Surgery, London, United Kingdom, 2St Mark’s Hospital, IBD/Physiology, London, United Kingdom


Anorectal sepsis accounts for both abscess and fistula and contributes a significant caseload in general surgical departments. It is poorly defined in respect to its epidemiology and the risk factors that predict whether an abscess will progress to a fistula. Risk stratification of fistulation is important and should guide clinical management from the first presentation with an abscess.


The Hospital Episode Statistics (HES) dataset from April 1997 to March 2012 was analysed. First presentations with abscess were used to establish an incidence of new abscess admissions in England. Of these patients, a rate of progression to fistula was derived, as well as independent predictive factors for this rate, using Cox regression analysis.


A total of 165 176 patients presented with an abscess (first admission) from April 1997–2012. The age standardised incidence rate for new abscess was 20.2/100,000, and the majority of these were male (70%). The overall rate of fistula formation was 17%. Subclassification revealed the extent IBD influences fistula formation; idiopathic (16%); UC (27%); and Crohn’s disease (47%). Of note, the majority of the fistulae (67%) occurred within the first 12 months, and the rate was once again most pronounced in the Crohn’s disease cohort. Of those patients with Crohn’s disease who developed fistula, 84% of these occurred within the first 12 months, at a median of 5 months.

Figure 1. Fistula formation for IBD and non-IBD cohort.

Cox regression analysis revealed independent predictors of fistula formation were Crohn’s disease (HR 3.51, CI 3.38–3.63); UC (HR 1.82, CI: 1.69–1.96, p < 0.001); being middle aged (41–60 years old HR 1.85, CI: 1.76–1.95, p < 0.001); an abscess in an intersphincteric (HR 1.53, CI: 1.38–1.70, p < 0.001); or an ischiorectal location (HR 1.48, CI: 1.43–1.53, p < 0.001); and being of female gender (HR 1.18, CI: 1.15– 1.21, p < 0.001).

Table 1 Cox regression analysis of fistula formation


This is the single largest study looking at the incidence of abscess to fistula progression. Nearly half of all Crohn’s abscesses progressed to fistula, and of those that did progress, 84% did so within the first year. Crohn’s disease was also the single largest independent predictor of fistula formation.