Ritu Singh, MD1, Neil Sharma, MD1, Waseem Amjad, MD2; 1Parkview Health, Fort Wayne, IN; 2Albany Medical Center, Albany, NY
Introduction: Venous thromboembolism (VTE) is common in patients with gastrointestinal tract cancers, and particularly in pancreaticobiliary (PB) cancers where the incidence is 10-25%. National inpatient sample (NIS) database is by far the largest database of hospitalized patients. We aim to determine in-hospital outcomes in PB cancer patients admitted for VTE. Methods: It is a retrospective cohort study utilizing the combined 2016 and 2017 national inpatient database. Patients discharged with the principle diagnosis of VTE in PB cancers were included. Main outcomes were in-hospital mortality, length of stay (LoS) and total cost of hospitalization in patients with PB cancers. Results: Total number of included patients with VTE in PB cancers was 3015 and the number of all VTE patients was 248,890. Mean age of patients in the PB group was comparable to all VTE patients (66.28 years versus 63.03). In-hospital VTE related mortality in PB group was 2.65% (95% confidence interval, CI 1.63-4.28) compared to 1.49% in all VTE patients (95% CI 1.39-1.60). Mean LoS and total cost of hospitalization was comparable in both the groups (LoS 6 days [95% CI 4.08-7.91] versus 6.97 days [95% CI 6.56-7.38], total charge 50,209 (95% CI 39,019-61,399) versus 47,852 [95% CI 46,915-48,799]). None of the 110 patients who underwent intervention (thrombolysis or thrombectomy) for VTE died in the PB cancer group while 7.31% (95% CI 5.79-9.20) died among all VTE patients who underwent thrombolysis or thrombectomy. There was no difference in mean LoS between those with or without intervention, however, total cost of hospitalization was significantly higher in those undergoing intervention. Discussion: VTE related in-hospital mortality is significantly higher in patients with PB cancers compared to all VTE patients of comparable age. There was reduced mortality in patients undergoing intervention for VTE in PB cancers, however, this interpretation is limited by a small number of cases in the intervention subgroup. The major limitation of the study is retrospective study design and use of administrative database.
Disclosures: Ritu Singh indicated no relevant financial relationships. Neil Sharma indicated no relevant financial relationships. Waseem Amjad indicated no relevant financial relationships.