Data presented at this year’s American Society of Artificial Internal Organs (ASAIO) symposium have shown that the HeartMate II continuous-flow left ventricular assist device (LVAD) increased both the incidence of gastrointestinal (GI) bleeds and stroke in patients implanted with the device. However, previous data comparing HeartMate II to the pulsatile-flow device HeartMate XVE may contradict these findings.
“Two complications commonly associated with LVAD therapy are GI bleeding and stroke,” Jeffrey A. Morgan, MD, of the Heart and Vascular Institute at the Henry Ford Hospital in Detroit, and principal investigator of the studies, told Cardiovascular Business.
Morgan, who presented these data at the ASAIO conference June 11 in Washington, D.C., and colleagues evaluated the complication rates associated with the HeartMate II device in 64 chronic heart failure (HF) patients who received a device between March 2006 and May 2010. Of the 64 patients, 48 underwent implantation of a HeartMate II LVAD device as a bridge-to-transplant operation while 16 patients received the device as a destination therapy.
The first study used GI bleeding as the primary endpoint; the second study used development of a major adverse neurologic event (ANE). All patients received anticoagulation therapy that consisted of 81 mg of aspirin and warfarin with a target international normalized ratio (INR) ratio of 2.0 to 2.5.
“What we found was a 22 percent incidence of GI bleeding and a 7.8 percent incidence of stroke for patients who received the HeartMate II device,” Morgan noted.
During the first study, Morgan and colleagues characterized the location of GI bleeding and assessed whether these events affected mortality. “While these complications did not ultimately affect mortality, they did represent a significant morbidity, which required patients to be transfused both blood as well as blood products to reverse the coagulopathy,” Morgan noted.
The researchers reported that at the index event, the INR averaged 2.18 and platelet count was 230 x 109/l. There were no significant differences in age, gender, race, etiology of HF, diabetes, chronic renal insufficiency or body mass index between patients with and without GI bleeding. In a multivariate analysis, Morgan et al found that the only independent predictor of GI bleeds was a previous history of GI bleeding. This occurred in 35.7 percent of patients with a GI bleed versus 16 percent of those without. There were no thromboembolic complications reported.
The second study found that ANEs occurred in 7.8 percent of patients and most cases were hemorrhagic stroke, Morgan noted. Patients who experienced these events were older, had a higher incidence of chronic renal insufficiency and higher INRs at the time of the index event. The researchers noted that these events were associated with a 60 percent 90-day mortality rate.
Older age, chronic renal insufficiency and an INR greater than three were independent predictors of ANEs; however, having an INR greater than three was the strongest predictor.
Morgan hypothesized that GI bleeding occured for two reasons: the development of arteriovenous malformations (AVMs) in the gut as a result of nonpulsatile, continuous-flow and high INRs. “The lack of pulsatility may be a contributing factor to GI bleeding and also lead to the development of arteriovenous malformations,” he said.
Susan Benton Russell, spokesperson for Pleasanton, Calif.-based Thoratec, maker of HeartMate II, told Cardiovascular Business that the median INR of patients who experienced a bleeding event during the trial was 2.9, what she called “high.” Additionally, she noted that in the current study patients were being managed at an INR range of 2.0-3.0, “the older recommendation,” which could have been linked to the neurological complications found in the study. However, to thwart off these complications, Morgan and colleagues suggested that the anticoagulation target be lowered to 1.8 to 2.2 rather than the previous rate of 2.0 to 3.0.
“The analysis presented at ASAIO is on 64 HeartMate II patients and demonstrates that in cases where patients experienced a gastrointestinal bleeding event that clinicians were able to make adjustment to those patients’ anticoagulation regimen to effectively manage the situation,” Benton