Background: The predictive risk factors of clinically relevant pancreatic fistula (CR-PF) following distal pancreatectomy (DP) remain to be identified.
Methods: This is a retrospective cohort analysis of a single-institution database of patients undergoing DP, taking into account usual demographic, operative, and pathologic variables and visceral fat area (VFA), total muscle area (TMA), and surface muscle index (SMI) measured on preoperative CT scan. The primary end point was CR-PF. All variables associated with a p value < 0.05 on univariate analysis were included in a logistic regression model for multivariate analysis.
Results: From 2012 to 2016, 208 patients operated by 4 pancreatic surgeons underwent DP including 32 (15%) who developed CR-PF. Risk factors of CR-PF on univariate analysis were: BMI ≥ 25 kg/m2 (p = 0.050), VFA ≥ 92 cm2 (p = 0.006), laparotomy (p = 0.023), main pancreatic duct dilatation (p = 0.035), open passive drainage (versus closed suction drainage) (p = 0.001), and blood loss ≥ 225 ml (p = 0.001). Sarcopenia did not influence the risk of CR-PF (p = 0.076). On multivariate analysis, VFA ≥ 92 cm2 (OR 3.14; IC 95% (1.18-8.31), p = 0.022), blood loss ≥ 225 ml (OR: 2.72; IC 95% (1.06-6.96), p = 0.037), and open passive drainage (OR 3.72; IC 95% (1.40-9.87) p = 0.008) were three independent predictive factors of CR-PF. A CR-PF risk score was developed, predicting a 0% risk of CR-PF when no risk factors were present and a 39% risk when the 3 risk factors were present.
Conclusions: Visceral obesity, blood loss ≥ 225 ml and open passive drainage significantly increase the risk of CR-PF following DP.
Keywords: Abdominal drainage; Distal pancreatectomy; Postoperative pancreatic fistula; Visceral obesity.