Hospital-level balloon tamponade use is associated with increased mortality for all patients presenting with acute variceal haemorrhage

Abstract

Background & aims: Balloon tamponade (BT) can bridge patients to salvage therapy for uncontrollable acute variceal haemorrhage (AVH). However, data are limited regarding the reasons for, rate of and outcomes associated with Balloon tamponade use.

Methods: First, we performed an single-centre cohort study of all patients (N = 139) with oesophageal acute variceal haemorrhage from 01/2009 to 10/2015. Associations between Balloon tamponade use and adherence to four quality metrics (endoscopy within 12 hours, band-ligation, pre-endoscopy antibiotics and octreotide) were evaluated. Second, we analysed the National Inpatient Sample (2005-2011) to determine the association between in-hospital mortality for patients and their hospital's Balloon tamponade-utilization to acute variceal haemorrhage volume ratio.

Results: In the national cohort, 5.5% of 140 521 acute variceal haemorrhage admissions required Balloon tamponade utilization. Adjusting for patient- and hospital-level confounders, the rate of Balloon tamponade use per acute variceal haemorrhage managed at any given hospital was associated with increased mortality for all-comers with acute variceal haemorrhage. Compared to the lowest tertile, acute variceal haemorrhage admissions in the highest Balloon tamponade utilizers were associated with increased mortality of (OR1.17 95%CI (1.01-1.37). In the single-centre cohort, 14 (10.1%) patients required Balloon tamponade. Balloon tamponade utilization was significantly associated with alcohol abuse (50.4% vs 21.4%, P = .04), hepatocellular carcinoma (35.7% vs 8.8%, P = .01), higher median model for end-stage liver disease (MELD) score (26.3vs15.5, P = .002) and active bleeding during endoscopy (64.3% vs 27.5%, P = .01). Failure to provide all quality metrics was associated with a higher model for end-stage liver disease-adjusted risk of Balloon tamponade use: OR 16.7 95% CI(4.17-100.0, P < .0001).

Conclusion: Balloon tamponade use is associated with severity of bleeding but may also implicate deficits in processes of care. Even for patients who did not need Balloon tamponade, presentation to hospitals with high Balloon tamponade utilization increases their odds of dying from acute variceal haemorrhage.

Keywords: cirrhosis; endoscopy; liver disease; model for end-stage liver disease score; portal hypertension.