PREVENT Risk Calculator Works Well in Patients With High and Low Lp(a)

The latest American Heart Association (AHA) calculator for assessing atherosclerotic cardiovascular disease (ASCVD) risk performs well in patients with high and low lipoprotein(a) levels, a new analysis shows.

While Lp(a) is not included in the AHA’s PREVENT equations, risk prediction was only modestly improved when investigators added it to the calculator.   

“We generally saw that people with high Lp(a) fell within their predicted risk category,” lead investigator Harpreet Bhatia, MD (University of California San Diego), told TCTMD. “If someone has high Lp(a), you can use PREVENT and generally feel pretty confident that it does a good job of predicting someone's risk category on a broad population-level at least.”

The analysis, which was published recently in JAMA Cardiology, includes more than 314,000 participants enrolled in the Multiethnic Study of Atherosclerosis (MESA) and UK Biobank databases who were categorized by their 10-year risk of ASCVD. With the PREVENT calculator, which is likely to become the new standard for risk assessment for primary prevention, Bhatia said there was an opportunity to test how well the equations performed in the setting of high Lp(a) levels.

“I see a lot of patients in clinic specifically for high Lp(a),” he said. “We are seeing more and more people getting tested and more people getting referred to us for high Lp(a). Part of the impetus for this is to understand how we should treat these people: can we use the tools we have, or the tools that are coming up, to help risk-stratify these people like we do anyone else?”

Salim Virani, MD, PhD (Aga Khan University, Karachi, Pakistan), who was involved in the development of the PREVENT risk equations, said the new study confirms that the calculator predicts risk quite accurately across the spectrum of Lp(a) levels.

“Basically, what it tells you is that traditional risk factors do matter,” he told TCTMD. “Even when you have high Lp(a) levels, diabetes matters, blood pressure matters, cholesterol matters. Whether you’re smoking or not, sex, and age matter. These are all risk factors in the PREVENT equation as well as in the pooled cohort equations.”

This doesn’t diminish the importance of Lp(a), he said. The present analysis clearly shows that Lp(a) predicts risk over and above the traditional markers and can be used to personalize treatment, particularly in those at intermediate risk, said Virani.

PREVENT and Lp(a)

The PREVENT equations modernize the American College of Cardiology (ACC) and AHA’s pooled cohort equations recommended by current guidelines to aid in the clinical decision-making process for primary prevention. PREVENT includes a spectrum of cardiovascular, kidney, and metabolic risk factors and can estimate both the 10- and 30-year risks of MI, stroke, and heart failure in patients as young as 30 years old.

Lp(a) has been shown to be an independent predictor of ASCVD risk and calcific aortic stenosis, with multiple genetic studies suggesting there is a causal relationship between high levels of Lp(a) and cardiovascular disease. In the latest US cholesterol guidelines, Lp(a) is considered a risk enhancer that can help inform treatment decisions in select intermediate-risk patients who may be on the fence about starting medical therapy, such as statins.

In MESA (mean age 62.1 years; 53% female), 20% of participants had elevated Lp(a) levels, defined as greater 125 mmol/L. In the UK Biobank cohort (mean age 56.3 years; 55% female), 11.1% had elevated Lp(a) levels. The median 10-year risks of ASCVD in the studies were 4.7% and 3.9%, respectively.

In the pooled MESA and UK Biobank cohorts, the 10-year ASCVD event rates were higher for participants with Lp(a) greater than 125 mmol/L compared with those with lower levels. Across all 10-year risk categories—those with ASCVD risks less than 5%, 5% to less than 7.5%, 7.5% to less than 20%, and 20% or greater—the event rates were within the bounds of estimated risk by the PREVENT equations; an exception was those with elevated Lp(a) and a 10-year risk of 5% to less than 7.5%.

“If you look at people's actual future risk of events in this large population, if they had elevated Lp(a), they still fell into the predicted risk categories by PREVENT,” said Bhatia. “The risk is higher if they have high Lp(a). We saw that across all risk categories. If you look at two individuals in the same risk category and one has high Lp(a) and one has lower Lp(a), the one with high Lp(a) clearly has increased risk of events over a 10-year period, but on average, people kind of were still within those predicted risk categories.”

Overall, those with Lp(a) levels greater than 125 mmol/L had a 30% higher risk of ASCVD than those with lower levels (HR 1.30; 95% CI 1.22-1.38) and a 21% higher risk of total cardiovascular disease events (HR 1.21; 95% CI 1.15-1.27), with no interaction seen by PREVENT risk level. The risk of coronary heart disease was 42% higher among those with elevated Lp(a), but the risk of heart failure was only marginally higher among those with Lp(a) levels greater than 125 mmol/L.

There was no significant improvement in the C-index, the measure of concordance between predicted and observed event rates, when Lp(a) was added to the PREVENT risk equation in the pooled cohort or in the individual MESA or UK Biobank studies. This was true when Lp(a) values were evaluated continuously or when using different thresholds.

In the overall cohort, adding elevated Lp(a) to PREVENT had a modest ability to improve predicted ASCVD risk when assessed using the continuous and categorical net reclassification improvement (NRI) measure. The greatest NRI was observed in patients at borderline risk. When modeled as a continuous variable, the greatest improvement in ASCVD risk prediction was seen in low-risk participants.

Not ‘Either-Or’

At the population level, said Bhatia, their data don’t suggest the PREVENT risk equations need to use Lp(a), but he also added that the measurement can be used in individual patients to help with risk stratification.

To TCTMD, Virani said one interesting aspect of analysis is that Lp(a) appeared to matter more in patients with a 10-year borderline risk of ASCVD, which is important as these are the patients for whom physicians might be on the fence about starting lifelong therapy. Importantly, he said, the paper should not be interpreted as “either-or” when it comes to Lp(a) or PREVENT.  

“Traditional risk factors remain important in primary care,” said Virani. “If you’ve measured Lp(a), it doesn't mean that you don't have to [use] the PREVENT risk calculator. If you’re using the 10-year risk calculation and you believe that you need to measure Lp(a), don't be afraid in doing so. Use your clinical judgement and if you see a level that's high, then think about what you're going to do with that.”

In an editorial, Donald Lloyd-Jones, MD (Boston University School of Medicine, MA), and Amit Khera, MD (University of Texas Southwestern Medical School, Dallas), say that given the lack of difference in the C-statistic and modest reclassification of risk when Lp(a) was added to PREVENT, routinely measuring Lp(a) to assess patient risk is not necessary. However, it should be measured once in a patient’s lifetime, so that clinicians can understand and personalize individual risk, which may aid in the use and intensity of preventive therapies, they argue.

Right now, there are no approved therapies for lowering Lp(a), but trials are ongoing with various drugs in development, including pelacarsen (Ionis/Novartis Pharmaceuticals), lepodisiran (Eli Lilly), olpasiran (Amgen), and zerlasiran (Silence Therapeutics). Muvalaplin (Eli Lilly), a selective, small molecule inhibitor of Lp(a), is an oral agent in development.