Behind the Jump in LVAD Thrombosis Rates: Did Pump Therapy Suffer From Its Own Success?

December 13, 2013

CLEVELAND, OH — About two years ago, thrombi started to grow more aggressively within left ventricular assist devices (LVADs) and their inflow and outflow ports, and they were forming sooner after pump implantation than had been seen in prior years—or so it seemed anecdotally at several major LVAD centers. Anecdotes prompted further scrutiny of the data, which established that rates of so-called pump thrombosis had jumped over several years from 1%–2% to 6%–8%[1,2].

The bump in pump thrombosis, a familiar and sometimes daunting potential LVAD complication, was confirmed in at least two recent reports that came to similar conclusions; one based on the experience at three major LVAD centers[1] and the other on data from the INTERMACS registry[2], both limited to patients with the HeartMate2 (Thoratec) axial-flow pumps.

HeartMate 2 [Source: Thoratec]

As expressed by Dr Randall C Starling (Cleveland Clinic, OH), lead author of the three-center report, who specifically described the experience at his own center, the analysis "confirmed that we were seeing more pump thrombosis, that it was occurring earlier, and that there was a specific time frame when this change occurred, which was in the year 2011."

Although the INTERMACS analysis is more in-depth in some ways, Starling said to heartwire , the two reports corroborate each other. The three-center analysis saw a 4.7% incidence of confirmed pump thrombosis at six months, and in INTERMACS, the end point—defined a bit differently as thrombosis-related death or pump replacement—was about 6%, in both cases up sharply from a few years earlier. "That looks pretty consistent to me."

Dr James K Kirklin

Also, Starling et al saw the median time from pump implantation to onset of pump thrombosis at the three centers tighten from 18.6 months before March 2011 to only 2.7 months later on. The group's experience, covering 72 cases of pump thrombosis in 837 patients, was published about two weeks ago in the New England Journal of Medicine (NEJM) and covered by heartwire at the time.

In the INTERMACS analysis, published online at about the same time in the Journal of Heart and Lung Transplantation (JHLT), with lead author Dr James K Kirklin (University of Alabama at Birmingham), "definite or probable" cases of pump thrombosis showed an early peak one to two months after implantation, then a decline extending to about six months, before starting a gradual, continuous climb.

"Dramatic Shift to Early Pump Thrombosis"

"There's no question that the 2011–2012 era saw a sudden spike in HeartMate 2 pump thrombosis," JHLT editor Dr Mandeep Mehra (Brigham and Women's Hospital, Boston, MA) said in an interview. The Kirklin et al and Starling et al reports confirm the spike as well as "a dramatic shift to early pump thrombosis," he observed for heartwire . "Pump thrombosis in the old days used to occur after a long time."

Dr Mandeep Mehra

Mehra is lead author on a perspective[3] accompanying the Kirklin et al report and a raft of other articles published online and destined for a special issue of JHLT. In it, he and his coauthors survey possible causes of the rise in early pump-thrombosis rates, possible risk markers, management particulars for patients with affected devices, and potential strategies for prevention.

One point made in the perspective: although the two reports showing early bumps in pump thrombosis focused exclusively on the Thoratec HeartMate 2, thrombosis is one part of the principal triad of complications, along with bleeding and infection, associated with all LVADs. That includes the Thoratec device, available in the US for bridging to transplantation (BTT) since 2008 and for destination therapy (DT) in transplant-ineligible patients since 2010. There is far more patient experience with that device than with the HVAD (HeartWare), which currently isn't available for DT and was approved for BTT here only in November 2012.

"Stroke and pump thrombosis are intrinsic to all blood pumps; they are not device-specific," Starling agreed.

Dr Peter Eckman

Dr Peter Eckman (University of Minnesota, Fairview), who isn't connected with the NEJM report or JHLT series, said "many of us in the field have felt like the rate of device thrombosis has been higher over the past couple of years, so I think what they are reporting is concordant with what we're seeing," although the jump in rates reported by Starling et al seems a little high. On the other hand, he added, data from the University of Minnesota experience hasn't yet been as closely analyzed.

"The full implications aren't going to be clear for some time, but [the reports] very nicely put a real-world estimate on what we've been seeing, and there has been a change," he told heartwire . "I think the questions now are, why is there a change and what do we do about it?"

A rise in early thrombosis risk "I don't think is going to significantly blunt the use of these devices in general," according to Eckman. Patients who would be implanted with them are still facing the prospect of death from heart failure. "The complications of these devices remain their Achilles' heel: whether infection, whether bleeding, whether device thrombosis," he said. But LVADs, along with the far less certain option of transplantation, are usually the patients' best hope for treatment.

LDH: How Useful a Biomarker?

The Starling et al report also saw that levels of lactate dehydrogenase (LDH) often rose after pump implantation, from a mean of 540 IU/L to 1490 IU/L "in the six weeks leading up to confirmed pump thrombosis. This finding was consistent across institutions."

Starling said at his center they routinely follow LDH levels in their LVAD patients and take note if they start to rise. "It definitely prompts more investigation, and if the patient is showing signs and symptoms of heart failure and has an elevated LDH, we're highly suspicious that pump thrombosis could be at play."

In the INTERMACS analysis, an LDH rise from 500 IU/L to 1000 IU/L within one month of implantation more than tripled the risk of pump replacement or death due to thrombus (p<0.0001).

But there are difficulties with the appealing idea of using elevated LDH as a predictor of pump thrombosis, as a prelude to some kind of intervention. As Mehra pointed out, LDH assays aren't standardized, "meaning different institutions can have different cutoffs." Moreover, the cutoff LDH concentrations that might call for a management response are different for the HeartMate 2 and HeartWare HVAD, and any LDH elevation isn't necessarily due to a device complication—there can be many causes. "It's not as simple as just measuring LDH and monitoring it. You have to know which device you're using, your institution's assay, and what you consider to be a significant rise."

"Almost Impossible to Point to One Cause"

Dr Randall C Starling

So why are there more and sooner pump thromboses? "I think we've addressed the epidemiology, but the 'why' question, that's still out there," said Starling. Possible causes range widely but break down into patient-related factors, aspects of patient management, and device-related issues. "We think any or all of those categories, singly or combined, could be contributing."

A lot of factors can worsen thrombosis risk, so "it's almost impossible to point to one cause," observed Mehra. Some of the possible reasons stem indirectly from approval and commercialization of the devices and their broader use.

"When the experience was limited to clinical trials, the pump thrombosis rate stayed pretty low—then it was mostly bridge to transplantation." Then, starting around 2010, LVAD case volumes escalated as more patients were implanted for DT. Such transplant-ineligible patients tended to be older, sicker, and with multiple comorbidities—especially diabetes, which predisposes to pump thrombosis, Mehra said.

Also, starting in about 2009, the LVAD community directed their attention at ways to attenuate bleeding complications, especially hemorrhagic stroke and gastrointestinal bleeding. Based on new evidence in the literature that it would be safe to slightly lower anticoagulation levels, he said, the community did just that.

At about the same time, "there was a very clear management trend where we began to lower pump speeds to try to facilitate aortic-valve opening," which seemed to be suffering from the pumps' continuous blood flow. The slower speeds were to allow natural ventricular contractions a greater contribution to flow, but it's very possible they gave blood traversing the pump more opportunity to clot.

It's also been established that the angulation of the HeartMate 2's inflow cannula can affect the risk of inside thrombus formation, and that can be determined by a number of factors, some related to surgical technique and others to a benefit of pump therapy. The angle can shift or otherwise be less than optimal if the preperitoneal pocket isn't deep enough; also, Mehra noted, the left ventricle decompresses and shrinks from the continuous LV unloading, likely changing the cannula angle.

Watching the RPMs

Starling sees the process of pump thrombosis as a continuum of conditions, from minimal thrombus deposition on the device that shows up as a bump in LDH but no functional impact to widespread thrombosis that clogs the pump mechanism. For conditions between the extremes, anticoagulation might be intensified, or a "ramp" study that assesses hemodynamics at progressively greater pump speeds may or may not unmask abnormal device function, for which heparin may or may not be necessary. With severe enough pump impairment, the pump is replaced or the heart transplanted.

Starling said that before becoming suspicious that pump thrombosis was increasing, his center and, he believes, other centers had "relaxed" their procedure for maintaining implanted patients on heparin until achieving a therapeutic INR on warfarin. "As soon as we had a sense that there was more pump thrombosis, the two things that we did were one, bridging with heparin meticulously, and two, consistently running the INRs at 2 or greater," he said. Also, they started paying closer attention to pump speed, ensuring it was maintained at >9000 RPM, on Thoratec's recommendation.

And there may be other issues to address. "Maybe the problem is the antiplatelet regimen, or [the patients] are aspirin hyporesponsive. Maybe they need  [clopidogrel]. Things like this have not been addressed in the literature at all," according to Mehra.

Eckman pointed to the enormous, longstanding research effort aimed at finding the best anticoagulation and antiplatelet regimens for coronary interventions; with LVADs, the challenge is greater. "Dealing with such a small population, with so many comorbidities, it's been hard for us to really pin down an optimal anticoagulation strategy and, more important, [a way to] personalize it to minimize the risk in any individual patient."

Starling discloses receiving grant support for his institution from Thoratec and HeartWare, receiving an honorarium for his institution from Thoratec, and receiving travel-related reimbursement for HeartWare. Mehra reports consulting for Thoratec; disclosures for the coauthors are listed in the paper. Kirklin is the principal investigator for INTERMACS, which is funded by the National Heart, Lung, and Blood Institute; disclosures for the coauthors are listed in the paper. Eckman discloses receiving honoraria and having previously received grant support from Thoratec and HeartWare.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.