Introduction: Gastrointestinal bleeding (GIB) occurs in 20-40% of patients with continuous flow left ventricular assist device (CF-LVAD) during the first years of follow up and is most commonly due to arteriovenous malformations (AVM). Although the precise pathophysiology of AVM formation during CF-LVAD support has not been established, it is postulated that hypoperfusion in the splanchnic vascular territory and resultant hypoxia may activate the angiogenesis signaling cascade via the HIF1α/angiopoietin-2 pathway. Interestingly, digoxin is a potent inhibitor of HIF1α synthesis.
Hypothesis: We hypothesize that digoxin therapy can be associated with a decrease incidence of GI and AVM bleeding in CF-LVADs patients.
Methods: Charts of all adult patients implanted with CF-LVAD between February 2006 and February 2017 were reviewed with emphasis on occurrence and etiology of GIB. Kaplan Meier (KM) and logistic regression analysis were used to assess the frequency of overall GI and AVM bleeding and their association with digoxin therapy.
Results: Sixty-two of 204 patients (30%) experienced a GIB and 20 of 62 (32%) bleeds were due to AVM. Both overall frequency of GIB (18% vs 37% of patients, p < 0.05) and AVM bleeding (3% vs 14% of patients, p < 0.05) were lower in the 73 patients receiving digoxin compared to the 131 patients not receiving digoxin during an average observation time of 516 and 446 days, respectively. Survival free of GIB was significantly higher in digoxin treated patients in KM analysis (Fig1). Multivariate logistic regression analysis revealed that digoxin therapy was independently associated with a reduced risk for overall GIB (odds ratio 0.35, 95% confidence interval 0.17-0.75) as well as AVM bleeding (odds ratio 0.19, 95% confidence interval 0.35-0.52).
Conclusions: Use of digoxin was associated with a significant reduction in overall GIB as well as AVM bleeding. Prospective studies are needed to validate this finding and its mechanistic underpinnings.