For Non-HDL Cholesterol, “Lower is Better” but “Lower for Longer” May be Best

Cardiovascular disease (CVD) begins early in life, and efforts to target disease modification at earlier vulnerable stages in its pathogenesis are evolving. Abnormal blood cholesterol levels are an important determinant of atherosclerotic CVD¹, with a wealth of data supporting the use of lipid-lowering therapy for CVD prevention and treatment². Primary prevention trials of lipid lowering therapy have focused primarily on middle-aged or older patients, given the need to demonstrate efficacy over a limited time horizon², and societal CVD prevention guidelines do not offer specific treatment recommendations for persons younger than 40 years in the absence of severely elevated low density lipoprotein cholesterol (LDL-C)³. Certainly, LDL-C below treatment thresholds in younger individuals may be an important harbinger of future subclinical CVD⁴, supporting ongoing investigations of initiating risk factor modification efforts and pharmacologic therapy earlier in life.

In this issue of Circulation Research, Park et al. begin to evaluate this gap in CVD prevention by investigating the risk of myocardial infarction (MI) and stroke associated with abnormal lipid values in ≈1.9 million individuals 20–39 years of age from a nationwide population-based cohort in South Korea⁵. Modestly abnormal lipid values were associated with higher risk for MI over a median follow up of 5.2 years, with greatest risk residing in the highest decile for total cholesterol (≥223.4 mg/dL), LDL-C (≥139.5 mg/dL) and triglycerides (≥200.1 mg/dL). For high-density lipoprotein cholesterol (HDL-C), significantly lower MI risk was observed for all values above the first decile (>42 mg/dL). While HDL-C and triglycerides remained significantly associated with stroke risk, LDL-C and total cholesterol were not. In addition, the authors investigated measures of visit-to-visit variability in lipid values, finding that higher variability of LDL-C was paradoxically associated with lower MI risk, while higher variability of HDL-C and triglycerides were associated with increased risk. The overall incidence rate in this sample was low (0.18 – 0.35 events per 1000 person-years), reflecting the fact that MI is a fortunately rare event in individuals younger than 40 years, though lifetime risk is substantial (7.7–43.6% of individuals experiencing a first CVD event over 30 years in another study⁶). Collectively, despite initially low absolute risk, the study by Park et al. suggests that the short-term risk of abnormal lipid values in individuals <40 years may not be negligible, suggesting the potential to extend lipid-lowering therapy to select individuals under 40 years as a testable strategy for CVD prevention.

The concept that lifetime risk of disease originates with risk factor burden in young adulthood is not new in CVD prevention, and low risk at a shorter time horizon may not exclude an individual from high long-term (lifetime) risk (a “risk-treatment paradox”)⁷. Indeed, genetic susceptibility to higher LDL-C is related to long-term risk of CVD¹, intimating potential benefits from earlier intervention. In this context, the findings of Park et al. complement other recently published reports demonstrating that non-HDL cholesterol in this age group is associated with long-term (i.e., 15–30 year) CVD risk^{6, 8}. Leveraging observations from ≈100,000 individuals younger than 45 years old across 44 population-based cohorts, Brunner et al. reported that non-HDL cholesterol and LDL-C levels were associated with a higher 30-year CVD risk⁶. In the Brunner study, the hazard ratio for CVD was highest in younger relative to older individuals. While there are several potential explanations for this finding, it is likely that abnormal lipid values detected earlier in life reflect both a higher lifetime burden of pro-atherogenic lipoproteins and a propensity toward dyslipidemia over a lifetime⁹.

In addition, Park et al. found that low HDL-C and high triglycerides were associated with MI and stroke risk, consistent with observations that young people with MI are more likely to have low HDL-C than high LDL-C¹⁰. Given evidence that low HDL-C may not be causal in CVD development¹¹, it is possible that its association with CVD reflects underlying cardiometabolic risk related to diabetes, obesity, physical inactivity, and other unmeasured metabolic risk factors that correlate with reduced HDL-C. Apart from statin (or other pharmacologic) therapy, this observation supports ongoing societal prevention efforts targeting behavioral modifications to improve cardiometabolic lifestyle factors in this age group as an important public health intervention to reduce CVD risk.

Several other observations from Park et al. merit mention. The suggestion that abnormal lipid values may be associated with future MI to a greater degree than stroke is consistent with prior observations in older individuals¹² and is interesting. Ischemic stroke prior to age 60 is rare, suggesting that a longer period of exposure to pro-atherogenic dyslipidemia may be required for stroke or that ischemic stroke risk is largely dependent on other primary stroke risk factors (such as hypertension and cumulative blood pressure exposure), which accumulate with age. In addition, Park et al. did not find visit-to-visit lipid variability to be associated with increased risk, divergent from prior reports¹³. Higher LDL-C variability was actually inversely associated with MI risk. By contrast, in older individuals, visit-to-visit variability in lipid measures does appear to increase risk of CVD and lower cognitive performance¹³. Perhaps cumulative burden of abnormal lipids may be more relevant, such that a certain burden of atherosclerosis is required before lipid variability is important? Further studies are necessarily to support these observations.

A few features of the Park study should be kept in mind while interpreting its key findings. Events were adjudicated based on ICD-10 codes using administrative-level data and incorrect ascertainment of events is therefore possible. Moreover, clinical information regarding MI presentation and treatment are not reported. It is therefore unknown whether MIs in individuals younger than 40 years exposed to modestly high cholesterol are pathophysiologically distinct from those that occur later in life. There was a disproportionate amount of men versus women in this sample as women receiving prenatal care were not included. The study was performed in a single country and may not be generalizable to other countries or racial/ethnic groups. This is especially relevant due to previous observations that the prevalence and risk associated with lipid abnormalities in Asian populations may differ from non-Asians¹⁴.

Despite these limitations, how do we move forward? A natural conclusion of this work is in support of randomized controlled trials of lipid-lowering therapy extended to a younger age range, though such studies face major challenges. Although relative risk reduction of lipid lowering is likely to be similar as in other groups, because short to intermediate term event rates are low in this population, massive sample sizes and/or long-term follow-up would be required to demonstrate efficacy for clinical endpoint reduction. One possible solution could be innovative trial designs leveraging electronic health records and administrative data (as Park et al. have done). In addition, the age at which to begin therapy—and what therapy to select (e.g., PCSK9 inhibition versus statin therapy)—remain important considerations. Ultimately, even if efficacy could be demonstrated, traditional approaches to assessment of cost effectiveness might not reach acceptable thresholds given long term disutility of daily or even monthly medication and very long timeframe for accruing benefit. Adherence over a long time-horizon may also be limiting, though innovations in lipid-lowering therapy may address this¹⁵. Finally, potential consequences of long-term exposure to pharmacologic therapy remain an important consideration to guide patient-centered care. Certainly, behavioral interventions (e.g., physical activity, smoking cessation and diet modification) targeting obesity as one prime impetus of pro-atherogenic dyslipidemia also remain at the foundation of any approach.

Collectively, the study by Park and colleagues, taken together with other recent reports, provide additional evidence in a large population supporting associations of abnormal lipid values in young adulthood with near-term risk for MI (with low absolute risk) and long-term risk for CVD (with high absolute risk). With increasing awareness that CVD is a “life-course” disease warranting life-long prevention efforts, optimal methods to identify and modulate CVD risk in early adulthood, when it is perhaps most modifiable, may be an important next chapter in the saga of CVD prevention.