乌鸦传媒

This article is a collaboration between 乌鸦传媒 and:

A new risk score found that older age and surgeon experience are two of several significant factors that could help triage patients receiving a left ventricular assist device (LVAD).

Using the four-component HeartMate Risk Score (HMRS), Jennifer Cowger, MD, of the University of Michigan Cardiovascular Center in Ann Arbor, and colleagues found that a high score correlated with nonsurvival at 90 days.

The multivariable analysis identified advanced age, elevated serum creatinine, low albumin, and elevated international normalized ratio (INR) as significant predictors of death, according to the study in the online Journal of the American College of Cardiology.

Centers that performed fewer than 15 implants during the HeartMate II clinical trials also were significantly associated with higher mortality.

The current methods to gauge risk are the Destination Therapy Risk Score (DTRS) and INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) profiles.

DTRS comprises eight variables including platelet count, pulmonary arterial pressure, INR, serum albumin, and and lack of intravenous inotropic support.

INTERMACS profiles are derived from a large registry and include seven metrics such as critical cardiogenic shock, progressive decline on inotropic support, resting symptoms, and home on oral therapy.

The problem with these two scores, Cowger and colleagues said, is that they were devised with older technology in mind, specifically the pulsatile LVAD, as opposed to today's continuous flow devices.

"[T]he utility of both tools for assessing LVAD candidate risk in the contemporary LVAD era is in question, and a reassessment and revision of LVAD candidate risk prediction is warranted," they wrote.

Wayne Levy, MD, of the University of Washington in Seattle, wrote in an accompanying editorial, that a "validated multivariate risk score would be extremely helpful in selection of patients at such high risk that they are unlikely to survive even with a LVAD, so-called futile implants."

To develop and validate the HeartMate Risk Score (HMRS) researchers used data from two different HeartMate II trials: one where the device was used as a bridge-to-transplant, the other for destination therapy. The cohort from both trials comprised 1,122 patients with end-stage heart failure.

Researchers then assigned patients to either the derivation (583) or the validation (539) cohorts. Baseline demographics of the two groups were similar and each group had similar duration of LVAD support: 347 versus 332 days. There were about 23% women in each group.

The 14% of patients in the derivation cohort who died within 90 days of the operation tended to be older, have poorer renal function, a higher INR, and greater degrees of hypoalbuminemia.

Those four factors remained significant after the multivariable analysis and were used to developed the HMRS.

Both cohorts had similar median HMRS (1.75 and 1.71, derivation and validation, respectively). Survivors had lower HMRS scores than patients who died (1.68 versus 2.37, P<0.001).

Researchers also calculated cutoffs to divide the risk into low (<1.58), medium (≥1.58 to ≤2.48), and high (>2.48).

Mortality strongly correlated with the medium- and high-risk groups: 4% low, 16% medium, and 29% high. The difference between the medium/high groups and the low group was significant. Those figures for the validation group were 8%, 11%, and 25%, respectively. Only the difference between the low and high groups met significance.

The investigators found the same pattern of greater survival associated with lower risk when they followed patients out to 1 year: 83%, 72%, and 58% for patients in the low, medium, and high groups, respectively.

There was good discrimination of risk at 90 days and thereafter, the authors noted, particularly for those at high risk.

However, Levy pointed out that the area under the curve (AUC) for the derivation group was high at 0.77 but fell to 0.64 in the validation cohort. Whereas the derivation group had a seven-fold difference in risk between the low- and high-risk groups, the validation difference dropped to threefold.

But Cowger and colleagues concluded that compared with other risk stratification scores, the HMRS is easy to calculate with only the four variables of age, serum albumin, creatinine, and INR. In addition, these factors are all "routinely obtained and reproducible clinical data points."

The authors noted several limitations to the study including the fact that the patients as a group are not representative of nonclinical trial patients, so further validation needs to be done in a broader population. In addition, researchers did not include the effect of comorbidities on outcomes in the longer term analysis.

From the American Heart Association: