An abrupt and unexpected increase in pump thrombosis has occurred since March 2011 in patients with advanced heart failure who received the HeartMate II left ventricular assist device, according to a report published online Nov. 27 in the New England Journal of Medicine.
Those who develop LVAD thrombosis are at substantial risk of death and morbidity unless the device is replaced or cardiac transplantation is performed, said Dr. Randall C. Starling of the Cleveland Clinic and his associates.
The reason for the sudden rise in pump thrombosis is not yet known, but it may be related to excess "deposition of material (fibrin and denatured protein) in proximity to the inflow bearing" of the device, which has been observed by clinicians in at least three cardiac centers.
"Our data suggest that the HeartMate II may have a more narrow tolerance with respect to thrombus formation than originally understood, and may be vulnerable to the timing and intensity of anticoagulation," Dr. Starling and his colleagues noted.
The HeartMate II, made by Thoratec, is a small axial-flow LVAD that "rapidly became integral in the treatment of patients with advanced heart failure" after it was introduced. Early trials and postmarketing approval studies reported a 2%-4% incidence of thrombosis. But a quality review at the Cleveland Clinic showed a sudden, marked increase in that rate beginning in March 2011.
Researchers at the Cleveland Clinic then pooled their data on the HeartMate II (from October 2004 to February 2013) with data on the device from two other high-volume cardiology centers: Washington University Barnes-Jewish Hospital, St. Louis (January 2004 to April 2013) and Duke University Medical Center, Durham, N.C. (May 2005 to May 2013). They now report on their findings from 895 HeartMate II LVADs implanted in 837 patients during that time.
"The HeartMate II may have a more narrow tolerance with respect to thrombus formation than originally understood, and may be vulnerable to the timing and intensity of anticoagulation..."
The mean patient age was 55 years, and 21% of patients were women.
The primary endpoint of the study was confirmed LVAD thrombosis, defined as a finding at pump replacement, urgent transplantation, or autopsy of a thrombus on the blood-contacting surface of the device, its inflow cannula, or its outflow conduit. A total of 72 such thromboses were found in 66 patients; in addition, 36 cases of suspected thromboses occurred.
The rate of confirmed pump thrombosis was initially stable at roughly 2.2% for several years, then rose steeply after March 2011 to 8.4%. This pattern was seen at all three medical centers and across all the cardiac surgeons at those centers.
The median time interval from implantation to pump thrombosis had been 18.6 months before March 2011, but that rate plummeted to 2.7 months during and after March 2011, Dr. Starling and his associates reported (N. Engl. J. Med. 2013 Nov. 27 [doi: 10.1056/NEJMoa1313385]).
The investigators also reviewed serial measurements of lactate dehydrogenase (LDH) that had been taken in recipients of 568 of the devices. Elevated LDH indicates hemolysis that may herald impending thrombosis.
"The graph of the occurrence of LDH levels above 1000 IU per liter ... was nearly superimposable on the graph of the occurrence of confirmed pump thrombosis. The LDH data showed a parallel tendency for elevations to occur early after implantation beginning in approximately March 2011," they said.
The LDH level typically rose from an average of 540 IU/L to an average of 1,490 IU/L during the weeks leading up to LVAD thrombosis at all three medical centers. This confirms that LDH level is a useful clinical biomarker of hemolysis associated with LVAD thrombosis, the researchers said.
Of the 72 pump thromboses, 11 were managed by heart transplantation and 21 by pump replacement. Of the remaining 40 thromboses, 2 were managed by LVAD removal because left ventricular function had improved, and 38 were managed medically, with augmentation of anticoagulation therapy and thrombolytic agents. This stabilized the clinical course in some patients, but others elected withdrawal of care, citing "futility."
Mortality was 48% among the patients who could not be managed by heart transplantation or pump replacement, and this elevated rate was consistent across the three medical centers.
The reasons for this abrupt, dramatic rise in thromboses associated with the HeartMate II are not known. There were no changes in LVAD implantation techniques at any of the three medical centers during the course of this study, "and we could find no association between pump thrombosis and the surgeon performing the implantation." The design of the HeartMate II was changed during this interval – "modification of the outflow graft and bend relief, inflow conduit, and software" was introduced – but these changes have not been linked to pump thrombosis as yet.