As described in our previous study, platelet-rich cryoprecipitate and thrombin were obtained from 300-400ml whole blood of each patient enrolled in the autologous PRFG(platelet-rich fibrin glue) group and then frozen at -20°C for storage. Prior to application, frozen cryoprecipitate and thrombin stored were thawed in a 37°C water bath. Aminomethylbenzoic Acid (1ml: 1mg, Sigma-Aldrich, St Louis, MO) was added into the cryoprecipitate in the volume ratio of 1:10.

The commercial fibrin sealants used in the study were purchased from Guangzhou Bioseal Biotech Co. Ltd, Guangzhou, China. Upon application, two components, fibrinogen and thrombin, were placed separately in two syringes which were joined by a Y-shaped connector. The trunk of the Y-shaped connecter was connected to a single lumen catheter.

A forward-viewing fistula-fiberscope (EndoView, Outai Medical Equipment, Shanghai, China), with 15cm length and 5mm width of fiber optical wire, was inserted into the fistula tract to accomplish endoscopic visualization. Briefly, the fistula-fiberscope assisted procedure was carried out percutaneously, allowing the exposure of internal hole and the whole tracts, followed by insertion of this catheter with distal mixing device. Upon injection, two components, fibrinogen and thrombin, were mixed together in the mixing reservoir and yield a gelatin-like glue. This procedure was repeated in patients up to 3 times within 14-day FG treatment period.

Closure time (defined as the interval between the day of enrollment and day of fistula closure), Closure rate up to 14 days, Closure rate up to 180 days, Fistula recurrence rate

Incidence of adverse events and severe adverse events up to 180 days (defined as an event that was fatal or life-threatening, led to additional hospitalization or disability, or required an intervention to prevent one of these outcomes)

Hospital cost upon enrollment; Hospital cost during entire hospital stay; Cost between fistula onset and final outcome