Gastrointestinal bleeding with the HeartMate II left ventricular assist device Jeffrey A. Morgan, MD,a Gaetano Paone, MD,a Hassan W. Nemeh, MD,a Scott E. Henry, MD,a Rosan Patel, BA,a Jessica Vavra, BA,a Celeste T. Williams, MD,b David E. Lanfear, MD,b Cristina Tita, MD,b and Robert J. Brewer, MDa From the Divisions of aCardiothoracic Surgery and bCardiovascular Medicine, Heart and Vascular Institute, Henry Ford Hospital, Detroit, Michigan.
KEYWORDS: heart failure; left ventricular assist device; LVAD; gastrointestinal; bleeding
BACKGROUND: Continuous flow left ventricular assist devices (CF-LVADs) have yielded improved outcomes compared with pulsatile flow devices for patients on long-term support. However, significant rates of gastrointestinal bleeding (GIB) have been observed during CF-LVAD support. METHODS: From March 2006 through March 2011, 86 patients with chronic heart failure underwent implantation of a CF-LVAD (HeartMate II; Thoratec Corp., Pleasanton, CA). Records were reviewed to determine the prevalence of post-implant GIB, location of the bleeding site and associated morbidity and mortality. Uni- and multivariate analyses were conducted to identify independent predictors of GIB. RESULTS: GIB occurred in 19 patients (22.1%) with a duration of support that ranged from 5 to 456 days. Sources of GIB included small bowel and rectum in 6 patients each, large bowel in 2 patients and stomach in 1 patient. No definite source was identified in 4 patients. There were no deaths referable to GIB. Recurrent GIB occurred in 4 patients. History of a GIB prior to LVAD implant was the only variable significantly different between patients with and without post-implant GIB (21.1% vs 10.4%, p ⫽ 0.016), and was the only independent predictor of GIB (OR ⫽ 2.24, 95% CI 2.121 to 2.435, p ⫽ 0.004). CONCLUSIONS: Gastrointestinal bleeding is a frequent source of morbidity for patients on HeartMate II LVAD support but does not significantly impact survival. As implantation of CF-LVADs with non-pulsatile flow gains popularity for both bridge-to-transplant and destination therapy, a better understanding of the pathophysiology of GIB in these patients will be needed for minimizing this complication. J Heart Lung Transplant 2012;31:715– 8 © 2012 International Society for Heart and Lung Transplantation. All rights reserved.
Left ventricular assist devices (LVADs) have become the standard of care for selected patients with end-stage heart failure refractory to maximal medical therapy. Over the last several years, there has been a transition from pulsatile flow (PF) pump designs to continuous flow (CF) pumps. Among the different devices, the HeartMate II (HM II) has demonstrated safety and efficacy in providing circulatory support for long periods of time with improved device durability and increased long-term survival for both bridge-to-transplant (BTT) and destination therapy (DT) patients.1–5 However, despite the encouraging results, an increased incidence Reprint requests: Jeffrey A. Morgan, MD, Division of Cardiothoracic Surgery, Henry Ford Hospital, K-14, Room 1439, 2799 West Grand Boulevard, Detroit, MI 48202. Telephone: 313-916-2695. Fax: 313-916-2687. E-mail address: [email protected]
of gastrointestinal bleeding (GIB) has been noted with CFLVAD support compared with older PF pumps, probably due to alterations in vascular endothelial physiology and hematologic imbalances.6 –15 In this study, we review our experience with patients on long-term HM II LVAD support to determine the prevalence and outcome of GIB.
Methods From March 2006 through March 2011, 86 patients with chronic heart failure underwent implantation of a HM II LVAD as a BTT (n ⫽ 54) or as DT (n ⫽ 32). The clinical records were retrospectively reviewed to identify those patients whose post-implant course was complicated by GIB, which we defined as melena, hematochezia or hematemesis along with a drop in hemoglobin requiring transfusion. A variety of
1053-2498/$ -see front matter © 2012 International Society for Heart and Lung Transplantation. All rights reserved. doi:10.1016/j.healun.2012.02.015
The Journal of Heart and Lung Transplantation, Vol 31, No 7, July 2012 The mean hematocrit of patients with GIB at the time of presentation was 25.4 ⫾ 5.0, platelet count was 245 ⫾ 108 and INR was 2.4 ⫾ 0.8. The latter was in the expected range of our institutional anti-coagulation guidelines for HM II patients, which is 1.8 to 2.5. There were no mortalities as a result of GIB. Patients were transfused an average of 3.8 ⫾ 1.3 units (U) of packed red blood cells (PRBCs). Three patients were transfused an average of 0.9 ⫾ 0.4 U of platelets. In addition, 15 patients received an average of 1.9 ⫾ 0.4 U of fresh-frozen plasma (FFP). No surgical interventions were required. The median length of stay associated with the bleeding episode was 10 days.
Figure 1 Comparison of Kaplan–Meier survival between patients with and without GIB. Red line: GIB; blue line: no GIB (p ⫽ 0.55, log-rank test). localization and treatment options were employed in consultation with our gastroenterology colleagues. Variables included duration of device support at the time of GIB; laboratory values, including hematocrit, platelets and international normalized ratio (INR) at the time of GIB; and the type and amount of blood product transfusion that patients received. We identified predictors of GIB and analyzed its effect on survival. The anti-coagulation regimen after HM II LVAD implantation generally consisted of aspirin 81 mg/day and warfarin with an INR target of 1.5 to 2.5. Anti-coagulation management subsequent to bleeding episodes was reviewed to determine whether changes to the anti-coagulation protocol resulted in increased device thrombosis or other thromboembolic complications. The procedures followed were in accordance with institutional guidelines, and this review was performed with institutional review board (IRB) approval and the consent of participants.
Statistical analysis Data were represented as frequency distributions and percentages. Values of continuous variables were expressed as a mean ⫾ standard deviation (SD). Continuous variables were compared using independent samples t-tests, and categorical variables were compared by chi-square tests. Using uni- and multivariate Cox proportional hazard models, patients with GIB were compared to patients without GIB with respect to clinical demographic variables to assess for possible independent predictors of GIB. For all analyses, p ⬍ 0.05 was considered statistically significant. Kaplan–Meier analysis was used to calculate survival along with a log-rank p-value when comparing groups. Actuarial survival rates at 1, 3 and 5 years post-implant were calculated by constructing life-tables. All data were analyzed utilizing SPSS, version 11.5 (SPSS, Inc., Chicago, IL).
Results There were 19 (22.1%) patients whose post-implant course was complicated by GIB, which included 14 BTT patients and 5 DT patients. Duration of support at the time of bleeding ranged from 5 to 456 days with a median of 176 days. GIB was localized to the stomach in 1 patient, the small intestine in 6 patients, the large intestine in 2 patients and the rectum in 6 patients. No definite source was identified in 4 patients.
Mid-term survival of patients with GIB Overall Kaplan–Meier survival was 85% at both 12 and 24 months for GIB patients, and 88% and 83% at 12 and 24 months for non-GIB patients, respectively (p ⫽ 0.655; Figure 1). Of the 14 BTT patients with GIB, 8 have been transplanted, 5 continue on CF-LVAD support, and 1 died of causes unrelated to GIB. Of the 5 DT patients with GIB, 3 are on ongoing support and 2 died at 1,062 days and 1,299 days of support from causes other than GIB.
Comparison of patients with GIB vs patients without GIB A comparison of demographic variables between patients with and without GIB is presented in Table 1. The only significant difference between the groups was a history of pre-LVAD GIB (21.1% in patients with GIB vs 10.4% in patients without GIB; p ⫽ 0.016). Likewise, uni- and multivariate analysis identified a history of previous GIB as the only independent predictor of post-implant GIB (odds ratio
Table 1 Comparison of Patients With GIB vs Patients Without GIB Variable Age (years) Male Caucasian African American Etiology of heart failure Ischemic cardiomyopathy Non-ischemic dilated cardiomyopathy Diabetes mellitus Hypertension Creatinine (mg/dl) Peptic ulcer disease Body mass index Body surface area History of pre-VAD GIB
GIB (n ⫽ 19)
No GIB (n ⫽ 67)
51.3 ⫾ 13.9 12 (63.2%) 12 (63.2%) 7 (36.8%)
53.2 ⫾ 10.9 50 (74.6%) 37 (55.2%) 29 (43.3%)
0.548 0.475 0.252
9 (47.4%) 8 (42.1%) 1.8 ⫾ 1.3 1 (5.2%) 26.5 ⫾ 4.9 1.91 ⫾ 0.25 4 (21.1%)
32 (47.8%) 28 (41.8%) 1.4 ⫾ 0.6 3 (4.4%) 29.1 ⫾ 5.5 2.03 ⫾ 0.30 7 (10.4%)
0.299 0.681 0.104 0.568 0.077 0.133 0.016
Morgan et al.
GI Bleeding with HeartMate II
[OR] ⫽ 2.24, 95% confidence interval [CI] 2.12 to 2.44, p ⫽ 0.004).
Management of anti-coagulation after GIB Warfarin and aspirin were stopped in all patients at the time of GIB. After cessation of GIB, aspirin (81 mg/day) was restarted in all patients. Warfarin was restarted in 14 patients with a lower INR target of 1.5 to 2.0, and 5 patients were maintained on aspirin alone without warfarin. This latter group on aspirin only included 4 BTT patients and 1 DT patient. The DT patient has been on LVAD support for 252 days since being taken off warfarin. The decision to restart or hold off on restarting warfarin was somewhat subjective and was dependent on the magnitude of the patient’s GI bleed, required number of units of PRBCs received by the patient, and the endoscopic appearance of the culprit lesion when/if identified on endoscopy. On this modified anti-coagulation protocol, there were no strokes, device thromboses or malfunctions.
Recurrence of GIB Recurrent GIB occurred in 4 patients after anti-coagulation was restarted. This occurred in 3 patients on aspirin and low-dose warfarin as well as in 1 patient on aspirin only. Among the 3 patients on both aspirin and warfarin at the time of their recurrent GIB episode, the mean INR was 1.8 ⫾ 0.4. Recurrences occurred 10 days, 32 days, 66 days and 95 days after starting patients on their modified anticoagulation regimens. All recurrences were from the same location as the original bleeding source. These included 1 from the colon, 2 from the small bowel and 1 from the rectum. The colon and rectal lesions were previously cauterized during the initial bleeding episode and were cauterized again after recurrent bleeding. The 2 small bowel foci of bleeding were not intervened upon endoscopically during the initial bleed or after the recurrent episode of bleeding. No surgical interventions were required. A second recurrence of GIB occurred in 2 patients, again from the same original source. Both of these patients were on low-dose warfarin and aspirin at the time of their second recurrence. Both patients were taken off of warfarin and maintained only on aspirin and have not had an additional recurrence. The 8 GIB patients who were transplanted have had no recurrent episodes of GIB after transplantation.
Discussion With the introduction of CF-LVADs, outcomes for patients on mechanical circulatory support have improved significantly.1–5 However, continuous flow pumps have been associated with an increased incidence of GI bleeding.6,16 The 22.1% incidence of GIB reported in this series of patients with HM II LVADs is substantially higher than what has been reported for the HM I extended vented electric (XVE) device.6,9,10 This finding has been corrobo-
717 rated in other series, including the study by Stern et al, who reported a 40% incidence of GIB in HM II recipients versus 0% for patients with a HM I XVE.6 Similar to what we found in our series, Demirozu and colleagues reported an incidence of GIB in 19% of HM II recipients.16 Several possible explanations have been offered for the increased incidence of GIB associated with CF pumps.16 –23 In patients with CF-LVADs the left ventricle is relatively decompressed, with minimal opening of the aortic valve. This creates flow patterns similar to those observed with aortic stenosis (AS), a condition that has been associated with GIB.6,16,24 –28 Heyde et al suggested24 that the abnormal pulse wave associated with AS may cause distention of the sub-mucosal venous plexuses of the GI tract and eventually lead to angiodysplasia, arteriovenous malformations and bleeding. Another possible mechanism of GIB in these patients is acquired von Willebrand disease Type 2A, which has been demonstrated in CF-LVAD patients secondary to shear stress on red blood cells.19,21,22 Some investigators have reported improvement in GIB by decreasing CF-LVAD flow and allowing the heart to eject through the aortic valve,16 thus maintaining some degree of pulsatility. Cessation of GIB has also been reported in CF-LVAD patients after cardiac transplantation,6 and this has been our experience as well, with no recurrences of GIB observed in 8 patients with GIB who underwent cardiac transplantation. In several patients, we have been able to localize and successfully treat actively bleeding lesions and we routinely attempt to identify the specific site of bleeding. Upper and lower endoscopy are the initial examinations to localize the bleeding, followed by small bowel capsule endoscopy, which is performed when the upper and lower endoscopy fails to localize the source of bleeding.29 –32 Enteroscopy has been used to reach more challenging small bowel lesions. Endoscopic and interventional techniques have included cauterization, injection, clipping of a visible vessel, and mesenteric angiography with coil embolization. The management of anti-coagulation in these patients must balance the concern for continued or recurrent bleeding with the risk of pump thrombosis or thromboembolic stroke. There is a growing literature on the relative safety of maintaining selected HM II LVAD patients on decreased anti-coagulation.33–35 Because the only independent predictor of GIB in our series was a pre-implant history of GIB, we now routinely initiate warfarin at the lower INR target in these patients after CF-LVAD implant. Additional follow-up will be needed to determine the safety of this approach. This study was a retrospective review and its limitations include the potential inaccuracy of data retrieved from medical records. In addition, the number of patients in this study was relatively small, thus limiting the statistical power of the analysis and conclusions. Further studies with more patients and longer follow-up will be useful. It is possible that if a larger patient population could be studied, additional pre-operative patient-specific factors that predict post-operative GI bleeding may be identified. Finally, a
The Journal of Heart and Lung Transplantation, Vol 31, No 7, July 2012
more extensive analysis of this subject should include preoperative von Willebrand factor levels as well as vascular morphology.
Disclosure statement The authors have no conflicts of interest to disclose. This study was supported by the Henry Ford Hospital. These findings were presented at the 31st annual meeting of the International Society for Heart and Lung Transplantation, April 2011, San Diego, California.
References 1. Miller LW, Pagani FD, Russell SD, et al. Use of a continuous-flow device in patients awaiting heart transplantation. N Engl J Med 2007; 357:885-96. 2. Pagani FD, Miller LW, Russell SD, et al. Extended mechanical circulatory support with a continuous-flow rotary left ventricular assist device. J Am Coll Cardiol 2009;54:312-21. 3. Slaughter MS, Rogers JG, Milano CA, et al. Advanced heart failure treated with continuous-flow left ventricular assist device. N Engl J Med 2009;361:2241-51. 4. Slaughter MS, Pagani FD, Rogers JG, et al. Clinical management of continuous-flow left ventricular assist devices in advanced heart failure. J Heart Lung Transplant 2010;29(suppl):S1-39. 5. John R, Kamdar F, Liao K, et al. Improved survival and decreasing incidence of adverse events with the HeartMate II left ventricular assist device as bridge-to-transplant therapy. Ann Thorac Surg 2008;86: 1227-34. 6. Stern DR, Kazam J, Edwards P, et al. Increased incidence of gastrointestinal bleeding following implantation of the HeartMate II LVAD. J Card Surg 2010;25:352-6. 7. Frazier OH, Myers TJ, Westaby S, et al. Clinical experience with an implantable, intracardiac, continuous flow circulatory support device: physiologic implications and their relationship to patient selection. Ann Thorac Surg 2004;77:133-42. 8. Letsou GV, Shah N, Gregoric ID, et al. Gastrointestinal bleeding from arteriovenous malformations in patients supported by the Jarvik 2000 axial-flow left ventricular assist device. J Heart Lung Transplant 2005; 24:105-9. 9. Crow S, John R, Boyle A, et al. Gastrointestinal bleeding rates in recipients of nonpulsatile and pulsatile left ventricular assist devices. J Thorac Cardiovasc Surg 2009;137:208-15. 10. Strauch JT, Spielvogel D, Haldenwang PL, et al. Recent improvements in outcome with the Novacor left ventricular assist device. J Heart Lung Transplant 2003;22:674-80. 11. Dagenais F, Portner PM, Robbins RC, et al. The Novacor left ventricular assist system: clinical experience from the Novacor registry. J Card Surg 2001;16:267-71. 12. John R, Lee S. The biological basis of thrombosis and bleeding in patients with ventricular assist devices. J Cardiovasc Transl Res 2009; 2:63-70. 13. John R, Boyle A, Pagani F, et al. Physiologic and pathologic changes in patients with continuous-flow ventricular assist devices. J Cardiovasc Transl Res 2009;2:154-8. 14. Slaughter MS. Hematologic effects of continuous flow left ventricular assist devices. J Cardiovasc Transl Res 2010;3:618-24. 15. Stein LH, Elefteriades JA. Can the occurrence of gastrointestinal bleeding in nonpulsatile left ventricular assist device patients provide clues for the reversal of arteriosclerosis? Cardiol Clin 2011;29:641-5.
16. Demirozu ZT, Radovancevic R, Hochman LF, et al. Arteriovenous malformation and gastrointestinal bleeding in patients with the HeartMate II left ventricular assist device. J Heart Lung Transplant 2011; 8:849-53. 17. Schaffer JM, Arnaoutakis GJ, Allen JG, et al. Bleeding complications and blood product utilization with left ventricular assist device implantation. Ann Thorac Surg 2011;91:740-9. 18. Slaughter MS, Sobieski MA, Gallagher C, et al. Fibrinolytic activation during long-term support with the HeartMate II left ventricular assist device. ASAIO J 2008;54:115-9. 19. Klovaite J, Gustafsson F, Mortensen SA, et al. Severely impaired von Willebrand factor-dependent platelet aggregation in patients with a continuous-flow left ventricular assist device (HeartMate II). J Am Coll Cardiol 2009;53:2162-7. 20. Letsou GV, Shah N, Gregoric ID, et al. Gastrointestinal bleeding from arteriovenous malformations in patients supported by the Jarvik 2000 axial-flow left ventricular assist device. J Heart Lung Transplant 2005; 24:105-9. 21. Geisen U, Heilmann C, Beyersdorf F, et al. Nonsurgical bleeding in patients with ventricular assist devices could be explained by acquired von Willebrand disease. Eur J Cardiothorac Surg 2008;33:679-84. 22. Uriel N, Pak SW, Jorde UP, et al. Acquired von Willebrand syndrome after continuous-flow mechanical device support contributes to a high prevalence of bleeding during long-term support and at the time of transplantation. J Am Coll Cardiol 2010;56:1207-13. 23. Boley SJ, Sammartano R, Adams A, et al. On the nature and etiology of vascular ectasias of the colon: degenerative lesions of aging. Gastroenterology 1977;72:650-60. 24. Heyde EC. Gastrointestinal bleeding in aortic stenosis [letter to editor]. N Engl J Med 1958;259:196. 25. Shoenfeld Y, Eldar M, Bedazovsky B, et al. Aortic stenosis associated with gastrointestinal bleeding. A survey of 612 patients. Am Heart J 1980;100:179-82. 26. Greenstein RJ, McElhinney AJ, Reuben D, et al. Colonic vascular ectasias and aortic stenosis: coincidence or causal relationship? Am J Surg 1986;151:347-51. 27. Pate GE, Mulligan A. An epidemiological study of Heyde’s syndrome: an association between aortic stenosis and gastrointestinal bleeding. J Heart Valve Dis 2004;13:713-6. 28. Vincentelli A, Susen S, Le Tourneau T, et al. Acquired von Willebrand’s syndrome in aortic stenosis. N Engl J Med 2003;349:343-9. 29. Garatti A, Bruschi G, Girelli C, et al. Small intestine capsule endoscopy in magnetic suspended axial left ventricular assist device patient. Interact Cardiovasc Thorac Surg 2006;5:1-4. 30. Seow CH, Zimmerman MJ. Capsule endoscopy in the detection of small-intestinal bleeding in patients supported by a nonpulsatile axialflow Jarvik 2000 left ventricular assist device [letter to editor]. Gastrointest Endosc 2006;63:1087. 31. Daas AY, Small MB, Pinkas H, et al. Safety of conventional and wireless capsule endoscopy in patients supported with nonpulsatile axial-flow Heart-Mate II left ventricular assist device. Gastrointest Endosc 2008;68:379-82. 32. Elmunzer BJ, Padhya KT, Lewis JJ, et al. Endoscopic findings and clinical outcomes in ventricular assist device recipients with gastrointestinal bleeding. Dig Dis Sci 2011;56:3241-6. 33. John R, Kamdar F, Liao K, et al. Low thromboembolic risk for patients with the HeartMate II left ventricular assist device. J Thorac Cardiovasc Surg 2008;136:1318-23. 34. Boyle AJ, Russell SD, Teuteberg JJ, et al. Low thromboembolism and pump thrombosis with the HeartMate II left ventricular assist device: analysis of outpatient anti-coagulation. J Heart Lung Transplant 2009; 28:881-7. 35. Slaughter MS, Naka Y, John R, et al. Postoperative heparin may not be required in transitioning patients with HeartMate II LVAD to long term warfarin therapy. J Heart Lung Transplant 2010;29:616-24.