Meeting News CoveragePerspective

Elevated central venous pressure after LVAD implantation increased GI bleeding risk

LAS VEGAS — Patients with elevated central venous pressure shortly after implantation with a continuous-flow left ventricular assist device were at increased risk for gastrointestinal bleeding in a study presented at the Heart Failure Society of America Annual Scientific Meeting.

Researchers conducted a retrospective, single-institution, two-arm cohort study in which they assessed 209 patients implanted with HeartMate II LVADs (Thoratec Corp.) at a single institution between August 2009 and October 2013.

The first arm of the study focused on GI bleeding and pulsatility index (PI), and included 185 HeartMate II recipients. Participants with a mean PI above the median for this group (n=90) were classified as having high PI, whereas those with a PI below the median (n=95) were identified as having low PI. The second arm focused on bleeding and central venous pressure (CVP) and included 173 HeartMate II recipients. Mean CVP at 2 and 3 days post-procedure was obtained for each patient; those with a mean CVP above the median for this group (n=86) were considered as having high CVP, those with CVP below the median (n=87) were identified as having low CVP.

The primary outcome for both arms of the trial was freedom from GI bleeding.

After a mean of 9 months of follow-up, no significant difference in incidence of GI bleeding was observed according to PI in the first arm of the trial. Patients with low PI experienced bleeding in 31% of cases vs. 29% in the high PI group (P=.6925). Secondary analysis in which patients were divided into tertiles according to PI yielded similar results.

The researchers noted a correlation between PI and pulse pressure at 2 and 3 days post-procedure (r=0.4511), suggesting that PI is a “modest, yet imperfect” indicator of pulsatility, Christopher T. Sparrow, MD, of Washington University in St. Louis, said during his presentation here.

In the second arm, after a mean follow-up of 12 months, patients with high CVP experienced bleeding in 40% of cases vs. 31% in the low CVP group (P=.0378). Dividing the group into tertiles according to CVP also indicated significantly increased rates of GI bleeding with increased CVP.

Sparrow said the differences between groups in the CVP arm were not apparent until after 90 days, suggesting that perioperative CVP may be a marker for long-term risk, rather than short-term, perioperative risk. This allows an opportunity to alter antithrombotic therapy early after implantation as needed to decrease future bleeding risk. Also, CVP measurement at different time points may be similarly useful for the prediction of GI bleeding risk, and that subsequent, separate analyses have supported this possibility.

“We believe prospective studies should be aimed at early identification of risk factors for bleeding, ideally in the preoperative or early postoperative setting, to allow alterations in antithrombotic therapy and potentially further patient education regarding the risks and benefits of LVAD support,” Sparrow said.

For more information:

Sparrow CT. Abstract #005. Jay N. Cohn New Investigator Physiology/Clinical Award. Presented at: the Heart Failure Society of American Annual Scientific Meeting; Sept. 14-17, 2014; Las Vegas.

Disclosure: Sparrow reported no relevant financial disclosures. Other investigators reported numerous financial disclosures.

LAS VEGAS — Patients with elevated central venous pressure shortly after implantation with a continuous-flow left ventricular assist device were at increased risk for gastrointestinal bleeding in a study presented at the Heart Failure Society of America Annual Scientific Meeting.

Researchers conducted a retrospective, single-institution, two-arm cohort study in which they assessed 209 patients implanted with HeartMate II LVADs (Thoratec Corp.) at a single institution between August 2009 and October 2013.

The first arm of the study focused on GI bleeding and pulsatility index (PI), and included 185 HeartMate II recipients. Participants with a mean PI above the median for this group (n=90) were classified as having high PI, whereas those with a PI below the median (n=95) were identified as having low PI. The second arm focused on bleeding and central venous pressure (CVP) and included 173 HeartMate II recipients. Mean CVP at 2 and 3 days post-procedure was obtained for each patient; those with a mean CVP above the median for this group (n=86) were considered as having high CVP, those with CVP below the median (n=87) were identified as having low CVP.

The primary outcome for both arms of the trial was freedom from GI bleeding.

After a mean of 9 months of follow-up, no significant difference in incidence of GI bleeding was observed according to PI in the first arm of the trial. Patients with low PI experienced bleeding in 31% of cases vs. 29% in the high PI group (P=.6925). Secondary analysis in which patients were divided into tertiles according to PI yielded similar results.

The researchers noted a correlation between PI and pulse pressure at 2 and 3 days post-procedure (r=0.4511), suggesting that PI is a “modest, yet imperfect” indicator of pulsatility, Christopher T. Sparrow, MD, of Washington University in St. Louis, said during his presentation here.

In the second arm, after a mean follow-up of 12 months, patients with high CVP experienced bleeding in 40% of cases vs. 31% in the low CVP group (P=.0378). Dividing the group into tertiles according to CVP also indicated significantly increased rates of GI bleeding with increased CVP.

Sparrow said the differences between groups in the CVP arm were not apparent until after 90 days, suggesting that perioperative CVP may be a marker for long-term risk, rather than short-term, perioperative risk. This allows an opportunity to alter antithrombotic therapy early after implantation as needed to decrease future bleeding risk. Also, CVP measurement at different time points may be similarly useful for the prediction of GI bleeding risk, and that subsequent, separate analyses have supported this possibility.

“We believe prospective studies should be aimed at early identification of risk factors for bleeding, ideally in the preoperative or early postoperative setting, to allow alterations in antithrombotic therapy and potentially further patient education regarding the risks and benefits of LVAD support,” Sparrow said.

For more information:

Sparrow CT. Abstract #005. Jay N. Cohn New Investigator Physiology/Clinical Award. Presented at: the Heart Failure Society of American Annual Scientific Meeting; Sept. 14-17, 2014; Las Vegas.

Disclosure: Sparrow reported no relevant financial disclosures. Other investigators reported numerous financial disclosures.

    Perspective

    We really don’t know what causes GI bleeding, so [the researchers] are trying to find clues to why patients would bleed when they’re having an LVAD; and one, of course, would be inadequate perfusion of the intestinal wall. That’s a legitimate hypothesis.

    They looked at PI as evidence for perfusion, and venous pressure, which also would inhibit perfusion. They looked at systemic, not portal venous pressure, which is a little removed from the true intestinal circulation, and found that venous pressure was, in fact, a predictor.

    There are many possible explanations for this: It could be that people with high venous pressure have more bleeding because it impairs their clotting mechanisms; it could be that they’re just sicker people, and sicker people have more bleeding. It may be that they’re bleeding from some other cause that isn’t related. There are multiple factors playing a role, and they tried to sort them out. In that way, I think this is actually a unique situation. I don’t think it’s been well studied before, so we’re getting some new insight into a complication of a device.

    However, this is just scratching the surface. We first have to demonstrate the site of the bleeding, which [the investigators] weren’t able to do, and document it if possible. What is the nature of the bleeding? What causes it? Is it a vascular rupture? Is it necrosis of tissue? Then, seek out markers, and of course the hope would be that you’d find evidence of something you could prevent or treat, so you could protect people from the bleeding. I think we’re a long way from getting an answer to that, but it’s a fertile area for more investigation.

    • Jay N. Cohn, MD
    • Cardiology Today Editorial Board member

    Disclosures: Cohn reports no relevant financial disclosures.

    See more from Heart Failure Society of America Annual Scientific Meeting