RESUMO
INTRODUCTION: Postpancreatectomy haemorrhage (PPH) is considered to be the most severe specific postoperative complication following pancreatic resections and its treatment is difficult and requires coordinated interdisciplinary collaboration. PPH causes 11-38% of all post-pancreatectomy deaths. The aim of this study was to determine the prevalence of PPH in a set of patients operated on within the last 10 years, and to analyze the diagnostic methods, treatment modalities and the outcomes. METHODS: A retrospective analysis of patients undergoing pancreatic resections between 2006 and 2015. Clinically relevant PPH (types B and C) were the subject of interest. The onset, location and severity of PPH were analysed. Other factors analysed included operation diagnosis of PPH, diagnostic methods along with signs of sentinel bleeding, treatment options undertaken including the number of transfusions. 30-day, 90-day and in-hospital mortality, as well as the length of hospital stay and readmission rate were calculated. A descriptive statistical method was used. RESULTS: A total of 449 patients were operated on. Pancreatoduodenectomy (DPE) or pylorus-preserving pancreatoduodenectomy (PPPD) was done in 76.4%, left sided pancreatectomy (LPE) in 19.8% and total pancreatectomy (TPE) in 3.8%. 190 of the patients (42.3%) were women and 259 (57.7%) men, with the mean age of 61.5±11.1 years. A total of 23 (5.1%) PPH cases were identified, 21 (4.7%) were clinically relevant. Eight patients (35%) developed early PPH with direct reoperation, late PPH was seen in 14 patients after DPE and in one after LPE. Sentinel bleeding was present in 53.3% of late PPH cases. CT/CTA was performed in four patients with subsequent DSA performed in three. DSA identified a gastroduodenal artery stump pseudoaneurysm in one patient, which was resolved using a stent. Surgical intervention for late PPH was required in 10 patients in total, six of whom needed direct surgery due to the rapid development of circulatory instability and 3 due to inconclusive radiological management. One patient needed surgical drainage of both an abscess and haematoma. In two patients the origin of bleeding was due to a gastric ulcer, which was proven and solved endoscopically and 2 patients required conservative treatment only. The specific mortality for PPH was 17.4%. In the group of patients that suffered with any PPH following DPE and PPDPE the mortality rate was 22.2%, and 28.6% for late PPH. If late PPH developed coincidentally with postoperative pancreatic fistula (POPF), the mortality was 44%. In the early PPH group, an average of 10.1±2.5 transfusion units (TUs) were used with an average length of hospital stay 17.5±4.8 days and zero mortality in comparison to an average of 11.7±10 TUs and 29.9±14.6 days in hospital and 26.6% mortality in the late PPH group. CONCLUSION: PPH is a severe complication, which has a high mortality rate. It also often coincidentally develops with POPFs. Early clinical diagnosis with identification of its cause plays a key role in management. The use of interventional radiology in the treatment of PPH has begun to dominate other treatment modalities due to a very high success rate, and close collaboration with interventional radiologists is necessary in order to reduce the rate of surgical intervention required in PPH. KEY WORDS: haemorrhage - pancreas - resection - complications - mortality.