This article is more than 5 years old. Information may no longer be current.
Do not go with the flow: Stemming the tide of GI bleeding with continuous-flow VADs
Click Here to Manage Email Alerts
Continuous-flow ventricular assist devices have become a life-sustaining treatment option for patients with advanced HF during the past decade. One of the most common and significant complications after continuous-flow VAD implantation is gastrointestinal bleeding, which occurs in 20% of patients with continuous-flow VADs, with recurrent bleeding in 35%.
The pathogenesis of gastrointestinal (GI) bleeding in continuous-flow VADs is multifactorial (see Table 1 below). Although all patients with continuous-flow VADs require thromboprophylaxis with aspirin and warfarin, antithrombotic therapy alone does not account for the high rate of GI bleeding. First-generation VADs delivered pulsatile-flow and had a lower incidence of GI bleeding; however, newer-generation devices have demonstrated improved durability and mortality. The continuous-flow physiology causes acquired von Willebrand disease, platelet dysfunction and angiodysplasia, which leads to arteriovenous malformations.
High shear stress from the pump motor induce conformational changes in von Willebrand factor, causing cleavage of von Willebrand factor by the plasma metalloprotease ADAMTS13 and loss of high-molecular-weight multimers leading to impaired von Willebrand factor-platelet binding.
Furthermore, an imbalance in angiopoietin (Ang)-1 and Ang-2 may lead to angiodysplasia and subsequent arteriovenous malformation. Ang-1 promotes normal vessel growth, whereas Ang-2 works synergistically with VEGF to promote abnormal vessel growth. Patients with continuous-flow VADs have significantly higher levels of circulating Ang-2 and increased expression of VEGF.
Initial management
Initial management of a patient with a continuous-flow CAD and GI bleeding consists of temporary de-escalation of antithrombotic therapy and administration of a proton pump inhibitor. Once bleeding is controlled, antithrombotic therapy should be resumed.
Although the majority of GI bleeds are due to arteriovenous malformations, ulcer-related bleeds account for approximately 25% of all continuous-flow VAD GI bleeding. Therefore, a proton pump inhibitor is warranted until the source of bleeding is confirmed, and can be continued if an ulcer is identified on endoscopy. Additionally, studies have shown the use of selective serotonin reuptake inhibitors decrease platelet aggregation and increase gastric acid secretion, which can both potentiate the risk for GI bleeding. Given this information, in patients on selective serotonin reuptake inhibitors, carefully consider the risks and benefits of continued use. Once the patient is stable, initial secondary prevention should consist of an ACE inhibitor or angiotensin receptor blocker, which decreases the expression of Ang-2, transforming growth factor-beta and VEGF. When ACE inhibitor or angiotensin receptor blocker therapy fails for secondary prevention, the options in Table 2 can be considered.
Treatment with octreotide
Octreotide is a synthetic somatostatin analogue that decreases splanchnic blood flow, improves platelet aggregation, increases vascular resistance and inhibits angiogenesis. Octreotide can be administered as a three-times-daily subcutaneous injection, continuous infusion or monthly intramuscular depot injection.
A retrospective, single-center analysis by Aggarwal and colleagues evaluated 101 patients who were implanted with a HeartMate II device (Abbott) from January 2005 to August 2011. Of these 101 patients, 23 experienced a GI bleed and 13 of the patients with GI bleeding received octreotide treatment. The investigators reported no difference in the amount of packed red blood cells transfused, re-bleeding rates, length of hospital stay or all-cause mortality in 13 patients treated with octreotide compared with those who did not receive octreotide. All patients received either a continuous infusion or subcutaneous injection of octreotide. The dose and route of administration for each patient was not described in the study. Another study by Shah and colleagues compared 51 patients with HeartMate II devices who were discharged from the hospital with subcutaneous or long-acting release (LAR) octreotide injection (Sandostatin LAR, Novartis) vs. a historical cohort of 240 patients. Patients in the octreotide group (38% LAR depot, 62% subcutaneous injections) were less likely to have a re-bleeding event at 6 months compared with the historical control group (P = .02).
Although the data generally favor the use of octreotide over placebo, the evidence is very limited regarding the choice between subcutaneous and LAR octreotide. The results of two small studies demonstrate a promising role for the LAR formulation of octreotide in this patient population. The cost of the LAR depot may be prohibitive for some patients; hence, determining adequate insurance coverage or utilizing a patient assistance program is necessary before prescribing.
Treatment with thalidomide
Thalidomide (Thalomid, Celgene) is an anti-angiogenic compound that inhibits VEGF and basic fibroblast growth factor. In 2015, Draper and colleagues published a treatment protocol in a case series of eight patients. The investigators aimed to evaluate thalidomide for GI bleeding in patients with continuous-flow VADs. Five patients had no recurrent bleeding, two had a reduction in bleeding and one patient died of sepsis shortly after thalidomide initiation. Three patients developed symptomatic neuropathy resulting in dose reduction or thalidomide discontinuation. In 2017, Seng and colleagues published a retrospective review of 11 patients with LVADs prescribed thalidomide for refractory GI bleeding. All patients were started on 50 mg thalidomide at bedtime; two patients required dose reduction due to adverse effects. Overall, 10 patients had resolution of GI bleeding. Median duration of therapy was 98 days. Before prescribing thalidomide, it is important to carefully screen patients for a history of thromboembolic events, symptomatic autonomic neuropathy, peripheral neuropathy and bradycardia. Prescribers and pharmacies must enroll in the THALOMID REMS program due to medication-associated embryo-fetal toxicities. It is important to highlight the FDA boxed warning for thromboembolism with this medication. In light of this warning, thalidomide should not be given to patients with a history of device-related thrombosis, and antithrombotic therapy should not be discontinued in patients with continuous-flow VADs who are prescribed thalidomide.
Salvage therapies
When the above measures fail in preventing further GI bleeding, there are limited data for alternate therapeutic options for salvage therapy.
Three salvage options are doxycycline, desmopressin and danazol. Doxycycline is an inhibitor of ADAMTS13, which leads to decreased cleavage of von Willebrand factor; however, the safety of doxycycline at doses high enough to achieve adequate concentrations to inhibit ADAMTS13 remains to be elucidated. Desmopressin activates the V2 receptor, which leads to a rise in intracellular concentrations of cyclic adenosine monophosphate, inducing exocytosis of von Willebrand factor from endothelial cells. The data to support the use of inhaled desmopressin in continuous-flow VAD-related GI bleeding are limited to one case report with positive results. If this therapy is chosen in a patient with refractory bleeding, insurance coverage should be confirmed, given the cost of therapy. Danazol is an androgen that may increase factor VIII levels and shorten the activated thromboplastin time, leading to a reduction in bleeding. One small cohort study of 19 patients with continuous-flow VADs who experienced one or more episodes of GI bleeding or were at high risk for recurrent bleeding received danazol 200 mg to 400 mg per day for a mean of 1 year and showed a significant reduction in the average number of GI bleeding-related hospitalizations per month without concerning changes in liver function tests.
Available options
In summary, when agents such as ACE inhibitors or angiotensin receptor blockers fail as secondary prevention for continuous-flow VAD GI bleeding, octreotide is a reasonable second-line option. Thalidomide also may be considered, but clinical judgment and caution is warranted due to the high risk for thrombosis and neuropathies. More data are needed regarding the salvage therapies doxycycline, desmopressin and danazol.
- For more information:
- Audrey J. Littlefield, PharmD, BCPS, BCCCP, is clinical pharmacy manager of cardiothoracic intensive care at NewYork-Presbyterian/Weill Cornell Medical Center. Alana M. Ciolek, PharmD, BCPS, is clinical pharmacy manager of cardiothoracic intensive care at NewYork-Presbyterian/Columbia University Irving Medical Center. Sarah A. Spinler, PharmD, FCCP, FAHA, FASHP, AACC, BCPS (AQ Cardiology), is the Cardiology Today Pharmacology Consult column editor. She is professor and chair of the department of pharmacy practice, School of Pharmacy and Pharmaceutical Sciences, Binghamton University. She can be reached at sspinler@binghamton.edu.
- References:
- Aggarwal A, et al. Ann Thorac Surg. 2012;doi:10.1016/j.athoracsur.2012.02.035.
- Bartoli CR, et al. JACC Heart Fail. 2015;doi:10.1016/j.jchf.2015.06.016.
- Crow S, et al. J Thorac Cardiovasc Surg. 2009;doi:10.1016/j.jtcvs.2008.07.032.
- Draper KV, et al. Gastrointest Endosc. 2014;doi:10.1016/j.gie.2014.03.040.
- Draper K, et al. J Heart Lung Transplant. 2015;doi:10.1016/j.healun.2014.09.013.
- Hollis IB, et al. ASAIO J. 2017;doi:10.1097/MAT.0000000000000433.
- Jiang HY, et al. Clin Gastroenterol Hepatol. 2015;doi:10.1016/j.cgh.2014.06.021.
- Malhotra R, et al. ASAIO J. 2017;doi:10.1097/MAT.0000000000000491.
- Schettle S, et al. J Heart Lung Transplant. 2018;doi:10.1016/j.healun.2018.04.013.
- Seng BJJ, et al. Int J Artif Organs. 2017;doi:10.5301/ijao.5000620.
- Shah KB, et al. J Heart Lung Transplant. 2017;doi:10.1016/j.healun.2017.01.321.
- Sieg AC, et al. Pharmacotherapy. 2017;doi:10.1002/phar.2016.
- Smallfield GB, et al. J Heart Lung Transplant. 2016;doi:10.1016/j.healun.2016.
01.694.
Disclosures: The authors report no relevant financial disclosures.