Risk Factors for Postoperative Complications After Pancreatic Head Resection: Multivariate Analysis of 608 Consecutive Operations
Tobias Keck*1, Frank Makowiec1, Eva Fischer1, Ulrich Adam2,1, Ulrich T. Hopt1
1Dept. of Surgery, University of Freiburg, Freiburg, Germany; 2Dept. of Surgery, Vivantes-Humboldt-Klinikum, Berlin, Germany
Mortality after pancreatic head resections (PHR) has been decreased clearly below five percent. However, complications are still frequent and influence the postoperative course and costs. The aim of our analyses was to evaluate morbidity after more than 600 PHR and to assess risk factors for various complications.
Methods: From 7/1994 to 2007 608 consecutive PHR were performed by five responsible surgeons in one team. The first 317 PHR were performed until 2001 in department A, the other 291 (after the team had moved end of 2001) since 2002 in department B. The documentation of the perioperative outcome was performed prospectively. Indications for surgery were periampullary or other malignancies (52%), chronic pancreatitis (43%) or others (5%). The techniques of PHR were: classical Whipple-procedure 14%, pylorus-preserving PHR 66%, duodenum-preserving PHR (DPPHR) 18% and total pancreatectomy 2%. Forty percent of the patients preoperatively had undergone biliary drainage (PBD). After examining 16 potential risk factors for complications univariately, multivariate analyses were performed using binary logistic regression.
Results: Any complication (AnyCompl) occurred in 49% of the patients, surgery-related complications (SurgCompl) in 33% and infectious complications (InfCompl) in 17%. Mortality was 2.4%. Multivariately significant risk factos (M-RF) for AnyCompl were the type of surgery (PD more complications than DPPHR), duration of surgery > 7 Std., absence of PBD, extended resections (including adjacent organs), PHR performed in department B and a preoperatively increased creatinine (above upper normal value; 10% of the patients). M-RF forSurgCompl were the abscence of a PBD, a BMI > 25, PHR performed in department B and again an increased creatinine-level. M-RF for InfCompl were again the type of PHR, the abscence of a PBD, a BMI > 25, surgery performed in department B and an increased creatinine. None of the other laboratory values assessed (including bilirubine) showed any correlation with complication rates.Mortality decreased from 2.8% in department A to 1.8% in department B. The only risk factor for mortality was a preoperative diabetes (p<0.05), although this parameter did not show any correlation with the different complication types.
Conclusions: Obesity, impaired renal function and the absence of a PBD are relevant risk factors for complications which may be corrected preoperatively and/or considered in the perioperative management of patients undergoing PHR. Centre experience leads to decreased mortality even though complication rates increased. This phenomenon might reflect a more aggressive management of those complications.