By Michael H. Crawford, MD, Editor
Synopsis: A pooled cohort from the Multi-Ethnic Study of Atherosclerosis and the United Kingdom Biobank study has shown that, overall, the American Heart Association PREVENT risk scores performed well at predicting atherosclerotic cardiovascular disease risk even in those with high lipoprotein(a) levels, but considering lipoprotein(a) in those with low PREVENT scores may help make therapeutic decisions in these individuals.
Source: Bhatia HS, Ambrosio M, Razavi AC, et al. AHA PREVENT equations and lipoprotein(a) for cardiovascular disease risk: Insights from MESA and the UK Biobank. JAMA Cardiol. 2025; Jun 4. doi:10.1001/jamacardio.2025.1603. [Online ahead of print].
Although known to increase the risk of atherosclerotic cardiovascular disease (ASCVD), elevated lipoprotein(a) (Lp[a]) is not part of the American Heart Association (AHA) predicting risk of cardiovascular event (PREVENT) equations. Thus, this study evaluated the performance of the PREVENT risk scores in subjects with elevated Lp(a) levels (> 50 mg/dL) in the Multi-Ethnic Study of Atherosclerosis (MESA) and the United Kingdom Biobank (UKB) populations.
The 6,670 MESA subjects recruited at six centers in the United States were free of known ASCVD at entry in 2000-2002 and followed through 2018. The UKB population of 308,113 subjects without ASCVD were recruited between 2006 and 2010. For both populations, the risk of coronary heart disease (CHD), ASCVD, heart failure (HF), and total cardiovascular disease (CVD) was determined. The subjects were risk-stratified as low (< 5%), borderline (5% to < 7.5%), intermediate (7.5% to < 20%), and high (≥ 20%) risk for each outcome.
Ten-year observed event rates were collected and the association with elevated Lp(a) was analyzed. Slightly more than half the subjects in both populations were women. The mean age of the MESA subjects was somewhat higher than that of the UKB population (62 vs. 56 years). In the pooled cohort from both studies, subjects with elevated Lp(a) had higher 10-year ASCVD event rates (hazard ratio = 1.3; 95% confidence interval, 1.22-1.38). Similar results were seen for CHD, HF, and total CVD. Adding elevated Lp(a) levels to PREVENT modestly enhanced the 10-year risk of ASCVD prediction, with the greatest effect in those at low risk. The authors concluded that elevated Lp(a) levels may improve personalized risk assessment, especially in low-risk individuals, but overall, the PREVENT equation performed well, even in those with elevated Lp(a) levels.
Commentary
The pooled cohort equation (PCE), which is 10 years old, is the current standard for estimating the risk of ASCVD. Adding Lp(a) levels improves the PCE’s risk prediction modestly, mainly in the low to borderline 10-year risk individuals (< 7.5%), but Lp(a) currently is not included, perhaps because it is being supplanted by the PREVENT equation, which also does not include Lp(a).
PREVENT has more clinical information variables than the PCE but does not include race/ethnicity because these are believed to be social constructs, not biologic ones. PCE does include separate scores for Black and white subjects. Lp(a) is known to vary by race/ethnicity, which is one reason why the effect of Lp(a) levels on risk is germane to PREVENT.
The authors defend the omission of Lp(a) in PREVENT because it is not measured frequently, and it currently is not a therapeutic target. Also, in this study, PREVENT performed well in those with or without Lp(a). On the other hand, Lp(a) was most strongly associated with CHD in low-risk individuals. Compared to the PCE, PREVENT results in fewer subjects who meet the criteria for statin therapy. Perhaps Lp(a) would help fill in this gap in statin therapy.
There are limitations to the Bhatia et al study. The focus was on 10-year outcomes, whereas MESA has data for almost 30 years now. The PREVENT equations include statin therapy, which few of the subjects studied were receiving. Also, there were few high-risk patients in either MESA or UKB. Risk levels for ASCVD are not established for PREVENT, so the PCE ones were used in the Bhatia study. In addition, the UKB population probably is not equivalent to the U.S. population. Finally, the UKB study measured Lp(a) in nanomoles/L, which were converted to mg/dL to combine them with the MESA Lp(a) levels, but this conversion is known to be imperfect and may have distorted the results.
Lp(a) levels now are included in some biomarker risk factor blood assay packages, which, in addition to standard lipid levels, include Lp(a), lipoprotein B, high-sensitivity C-reactive protein, and other factors such as measures of low-density lipoprotein particle size (e.g., Cardio IQ). As these enhanced biologic tests become more frequently used and perhaps in combination with coronary calcium scores, there will be an opportunity to fine-tune management of the lower-risk patients by the PREVENT equations.
Michael H. Crawford, MD, is Professor Emeritus of Medicine and Consulting Cardiologist, UCSF Health, San Francisco.
