Results after Definitive Surgical Treatment in Patients with Enteroatmospheric Fistula.

2018 
: As enteroatmospheric fistulas (EAF) lack healthy overlying tissue, spontaneous healing is very unlikely. Our aim was to identify risk factors for recurrence and mortality after definitive surgical treatment for EAF. Sixty-two consecutive patients with a diagnosis of EAF were submitted to definitive surgical repair (fistula resection and primary anastomosis) during a 6-year period. Several patient, disease, and operative variables were assessed as risk factors associated to our endpoints: recurrence and mortality. All patients were followed-up until hospital discharge or death. Univariate and multivariate analysis were performed. There were 24 females and 38 males with a median age of 53 years (interquartile ranges 43-63). EAF recurred in 23 patients. Univariate analysis identified several risk factors for recurrence which included performing more than one anastomosis (20 vs 52%, P = 0.013), failure of achieving total abdominal closure (16 vs 47%, P = 0.025), intraoperative hemorrhage >400 cc (28 vs 65%, P = 0.007), presence of multiple fistulas (25 vs 61%, P = 0.008), and preoperative C-reactive protein >0.5 mg/dL (54 vs 82%, P = 0.029). The latter two remained significant after multivariate analysis. Final EAF closure was attained in 47 patients (76%) and 8 more (13%) had a low-output (<50 mL/day) enterocutaneous fistula. Timing of surgery was not related to fistula recurrence. Eight patients died (13%), and fistula recurrence was the only risk factor found related to mortality both through univariate (26 vs 5%, P = 0.043) and after multivariate analysis. EAF management represents a rather challenging problem. Timing for surgical treatment is controversial and is based mostly on patient status and surgeon's criteria. Recurrence is associated to EAF characteristics and an inflammatory state; it was also the only factor associated to mortality.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    31
    References
    3
    Citations
    NaN
    KQI
    []