ABSTRACT
BACKGROUND: Gastrointestinal (GI) bleeding is a leading cause of intensive care unit (ICU) admission in pancreatic cancer patients. AIMS: To analyze causes, ICU mortality and hemostatic treatment success rates of GI bleeding in pancreatic cancer patients requiring ICU admission. METHODS: Retrospective multicenter cohort study between 2009 and 2021. Patients with a recent pancreatic resection surgery were excluded. RESULTS: Ninety-five patients were included (62 % males, 67 years-old). Fifty-one percent presented hemorrhagic shock, 41 % required mechanical ventilation. Main GI bleeding causes were gastroduodenal tumor invasion (32 %), gastroesophageal varices (21 %) and arterial aneurysm (12 %). Arterial aneurysms were more frequent in patients with previous pancreatic resection (36 % vs 2 %, p < 0.001). Hemostatic procedures included gastroduodenal endoscopy in 81 % patients and arterial embolization in 28 % patients. ICU mortality was 19 %. Multivariate analysis identified four variables associated with mortality: performance status >2 (OR 9.34, p = 0.026), mechanical ventilation (OR 14.14, p = 0.003), treatment success (OR 0.09, p = 0.010), hemorrhagic shock (OR 11.24, p = 0.010). Treatment success was 46 % and was associated with aneurysmal bleeding (OR 29.89, p = 0.005), ongoing chemotherapy (OR 0.22, p = 0.016), and prothrombin time ratio (OR 1.05, p = 0.001). CONCLUSION: In pancreatic cancer patients with severe GI bleeding, early identification of aneurysmal bleeding (particularly in case of previous resection surgery) and coagulopathy management may increase the treatment success and reduce mortality.
ABSTRACT
BACKGROUND: We sought to describe the reasons for intensive care unit (ICU) admission and outcomes of patients with pancreatic cancer requiring unplanned medical ICU admission. PATIENTS AND METHODS: Retrospective cohort study in five ICUs from 2009 to 2020. All patients with pancreatic cancer admitted to the ICU were included. Patients having undergone recent surgery were excluded (< 4 weeks). RESULTS: 269 patients were included. Tumors were mainly adenocarcinoma (90%). Main reason for admission was sepsis/septic shock (32%) with a biliary tract infection in 44 (51%) patients. Second reason for admission was gastrointestinal bleeding (28%). ICU and 3-month mortality rates were 26% and 59% respectively. Performance status 3-4 (odds ratio OR 3.58), disease status (responsive/stable -ref-, newly diagnosed OR 3.25, progressive OR 5.99), mechanical ventilation (OR 8.03), vasopressors (OR 4.19), SAPS 2 (OR 1.69) and pH (OR 0.02) were independently associated with ICU mortality. Performance status 3-4 (Hazard ratio HR 1.96) and disease status (responsive/stable -ref-, newly diagnosed HR 2.67, progressive HR 4.14) were associated with 3-month mortality. CONCLUSION: Reasons for ICU admissions of pancreatic cancer patients differ from those observed in other solid cancer. Short- and medium-term mortality are strongly influenced by performance status and disease status at ICU admission.