A look back at a pair of large cohort studies suggests a telling relation between two distinct predictors of atherosclerotic cardiovascular disease (ASCVD) risk and may offer guidance on how to interpret them together.
Elevated levels of lipoprotein(a), or Lp(a), and high coronary artery calcium (CAC) scores were both predictive of ASCVD risk over 10 years, but independent of each other and a host of more traditional cardiovascular risk factors, for example, in the analysis of data from the MESA (Multi-Ethnic Study of Atherosclerosis) and DHS (Dallas Heart Study) longitudinal cohorts.
Notably, the risk when both Lp(a) and CAC scores were high far exceeded that associated with either marker alone. But when CAC scores were less than 100 Agatston units, predicted ASCVD risk wasn’t influenced by levels of Lp(a). Indeed, a CAC score of 0 predicted the lowest levels of ASCVD risk, even with elevated Lp(a).
That is, the findings suggest, the addition of Lp(a) makes a difference to the risk assessment only when CAC scores are high, at least 100 units, and elevated Lp(a) doesn’t mean increased ASCVD risk in the absence of coronary calcium.
“Our novel findings indicate that elevated Lp(a) drives ASCVD risk independent of the subclinical coronary atherosclerosis burden captured by CAC score,” concluded a report on the analysis, published in the Journal of the American College of Cardiology, with lead author Anurag Mehta, MD, Emory University, Atlanta.
There are no formal recommendations on how to interpret Lp(a) and CAC scores together, but the current findings “provide impetus for measuring Lp(a) in more individuals as part of the shared decision-making process,” the authors contended.
“Really, the calcium score carries the majority of the information in terms of risk, except in the highest CAC score group. That is, if you have a high Lp(a) and a high burden of calcium, your risk is significantly higher than if you just have the high calcium score and the normal Lp(a),” senior author Parag H. Joshi, MD, MHS, said in an interview.
“We thought we would see that the group with higher Lp(a) would have more events over 10 years, even among those who didn’t have coronary calcium,” said Dr. Joshi, of the University of Texas Southwestern Medical Center, Dallas. “But we really don’t see that, at least in a statistically significant way.”
A CAC score of 0 would at least support a more conservative approach in a patient with elevated Lp(a) “who is hesitant to be on a statin or to be more aggressive managing their risk,” Dr. Joshi said.
“This study should be very reassuring for a patient like that,” Ron Blankstein, MD, director of cardiac computed tomography at Brigham and Women’s Hospital, Boston, said in an interview.
“If you have a high Lp(a) and you’re concerned, I think this study really supports the role of calcium scoring for further risk assessment,” said Dr. Blankstein, who is not associated with the new report. “We often check Lp(a) in individuals who perhaps have a family history or who come to see us in a preventive cardiology clinic. If it is high and there is concern, a calcium score can be very helpful. If it’s zero, that really means a very low risk of events. And if it’s elevated, I think we’re going to be more concerned about that patient.”
The current analysis suggests “that, when a patient without clinical cardiovascular disease is identified with either CAC ≥100 or Lp(a) >50 mg/dL, the next step in the risk evaluation should be to measure either Lp(a) or CAC, respectively – if not already performed – to identify the patients at highest risk,” Sotirios Tsimikas, MD, director of vascular medicine at University of California, San Diego, wrote in an accompanying editorial.
“Both Lp(a) and CAC should be more broadly applied in clinical care settings in patients without prior ASCVD to identify those that most likely will benefit from more aggressive therapy and, in the future, from Lp(a)-lowering therapies,” he wrote.
The analyses were conducted separately on data from 4,512 initially asymptomatic patients in MESA and 2,078 from the DHS cohort, who were followed for ASCVD events an average of 13 years and 11 years, respectively. Such events included coronary heart disease–related death, nonfatal MI, and fatal or nonfatal stroke.
In the MESA cohort – 52% women, 36.8% White, 29.3% Black, 22.2% Hispanic, and 11.7% Chinese – elevated Lp(a) (quintile 5 vs. quintiles 1-4) and CAC scores of 1-99 and above 100 (both compared with 0) were each independently associated with increased risk for ASCVD events. The hazard ratio was 1.29 (P = .02) for elevated Lp(a), 1.68 (P < .01) for a CAC score of 1-99, and 2.66 (P < .01) for a CAC score of at least 100.
The corresponding HRs in the DHS cohort were 1.54 (P = .07) for Lp(a), 3.32 (P < .01) for a CAC score of 1-99, and 5.21 (P < .01) for a CAC score of at least 100.
Of note, the authors wrote, ASCVD risk among MESA participants with a CAC score of 0 was not significantly different in those with normal and elevated Lp(a).
The findings were similar in the corresponding DHS analysis, the authors noted.
When both Lp(a) and CAC scores are considered as dichotomous variables, the highest 10-year ASCVD incidence in MESA was in participants with both elevated Lp(a) (≥50 mg/dL) and a high CAC score (≥100). The lowest risk was seen when Lp(a) was normal (<50 mg/dL) and the CAC score was no more than moderately high (<100).
The results in the corresponding DHS analysis, according to the report, again mirrored those from MESA.
“This study has important implications for our patients and also potentially for future clinical trial design,” Dr. Blankstein noted. “A big part of developing a trial in this space is identifying the patients who are at higher risk,” and the current analysis supports CAC scores for identifying the highest-risk patient among those with elevated Lp(a).
Current wisdom is that, for the most part, Lp(a) levels are genetically mediated and are mostly unaffected by interventions such as diet management or exercise. It’s unknown whether reducing elevated Lp(a) levels pharmacologically will cut ASCVD risk, but there are a number of clinical trial programs currently aimed at learning just that. They include the Novartis-sponsored phase 3 HORIZON trial of the antisense agent pelacarsen (TQJ230), with an estimated enrollment of almost 7,700; a randomized, controlled dose-finding study of the small interfering RNA agent olpasiran (AMG890), with 290 patients and funded by Amgen; and an 88-patient phase 1 study of another siRNA agent, SLN360, supported by Silence Therapeutics.
Dr. Mehta reported no relevant relationships. Dr. Joshi has received grant support from Novo Nordisk and consulting income from Bayer and Regeneron; holds equity in G3 Therapeutics; and has served as site investigator for GlaxoSmithKline, Sanofi, AstraZeneca, and Novartis. Dr. Blankstein reported serving as a consultant to Amgen, Novartis, and Silence Therapeutics.
A version of this article first appeared on Medscape.com.