New advances in preventive cardiology: A critical appraisal of 2018 cholesterol guidelines
Editor’s Note: This is a commentary on the 2018 cholesterol guidelines by Rhanderson Cardoso, MD; Roger S. Blumenthal, MD; and Seth S. Martin, MD, MHS.
The American Heart Association and the American College of Cardiology, in collaboration with multiple other societies, recently published two highly significant documents in the field of preventive cardiology: Use of Risk Assessment Tools to Guide Decision-Making in the Primary Prevention of Atherosclerotic Cardiovascular Disease, henceforth called Risk Assessment Statement, and the 2018 Guideline on the Management of Blood Cholesterol, henceforth called Cholesterol Guidelines.
These recommendations have been much anticipated in light of the breadth of data published in the field over recent years. These documents update previous statements and guidelines, such as the 2013 ACC/AHA Guideline on Treatment of Blood Cholesterol, the 2016 ACC Expert Consensus Decision Pathway on the Role of Non-Statin Therapies for Low Density Lipoprotein-Cholesterol Lowering and its 2017 Focused Update.
Herein, we discuss the new guidelines with special focus on four objectives: to examine the key changes from previous recommendations; to summarize the evidence from the primary studies that led to these changes; highlight recommendations that were maintained and strengthened; and to discuss points where the guideline recommendations could be further improved.
LDL thresholds emphasized
Epidemiological and Mendelian randomization studies and clinical trials clearly show that LDL has a causal effect in atherosclerosis. Several dozen trials in the past 30 years have consolidated statins as the mainstay of lipid-lowering therapies due to their efficacy, safety and proven reduction in CV endpoints, including mortality. A landmark meta-analysis from the Cholesterol Treatment Trialists’ Collaboration showed that each reduction in LDL of 39 mg/dL (1 mmol/L) in statin trials was associated with a 22% RR reduction in major adverse CV events, without any evidence of threshold within the range of studied cholesterol levels.
Despite evidence for better atherosclerotic CVD outcomes with lower levels of LDL, the 2013 ACC/AHA Guideline on the Treatment of Blood Cholesterol did not highlight specific LDL numbers for intensification of therapy. There was a focus on maximizing statin therapy, in addition to lifestyle, and assessing the percentage reduction in LDL as a marker of adherence to therapy. It was felt that LDL was the primary target of therapy, but evidence was insufficient to support goals/thresholds to guide use of nonstatins due to an absence of randomized data showing a reduction in hard endpoints with addition of nonstatins to statin therapy.
That has since changed. In 2015, results from the IMPROVE-IT study were published. Approximately 18,000 patients with a recent ACS and baseline LDL between 50 mg/dL and 125 mg/dL were randomly assigned to ezetimibe 10 mg daily or placebo, in addition to simvastatin 40 mg daily. The ezetimibe group’s LDL was 16 mg/dL lower than the placebo group, yielding a 7% relative reduction in major adverse CV events during a median 6-year follow-up. This effect was fully consistent with the benefit of statin therapy, on the basis of risk reduction per mg/dL of LDL lowering.
In 2017, a second drug class demonstrated CV outcomes benefit when added to statin therapy. The FOURIER trial reported outcomes in more than 27,000 patients with atherosclerotic CVD and LDL of at least 70 mg/dL (1.8 mmol/L) on statin therapy, who were randomly assigned to injections of evolocumab (Repatha, Amgen) or placebo. Patients receiving the PCSK9 inhibitor achieved an LDL that was 56 mg/dL lower than the placebo group, resulting in a 15% RR reduction in the primary major adverse CV event endpoint during a 2-year follow-up.
The ODYSSEY Outcomes trial of another PCSK9 inhibitor reported similar results. Nearly 19,000 patients with an ACS in the preceding year and LDL of at least 70 mg/dL despite high-intensity or maximally tolerated statin therapy were randomly assigned to injections of alirocumab (Praluent, Sanofi/Regeneron) or placebo. The investigators specifically targeted LDL levels of 25 mg/dL to 50 mg/dL and reported a 50% to 60% reduction in LDL in the alirocumab group vs. placebo. A 15% RR reduction in the primary major adverse CV event endpoint was observed during a median follow-up of 2.8 years. Interestingly, with slightly longer follow-up, a mortality benefit also seemed to emerge, with a 15% RR reduction favoring alirocumab.
Therefore, although statins remain the mainstay of lipid-lowering pharmacotherapy, the arsenal of pharmacotherapy for treatment of blood cholesterol has expanded to include ezetimibe and PCSK9 inhibitors. Importantly, recent randomized trials have provided evidence not only for the efficacy of such nonstatin therapies and aggressive LDL lowering, but also for the safety of very low levels of LDL. The average LDL level in IMPROVE-IT, FOURIER and ODYSSEY Outcomes in the active therapy arms were 54 mg/dL, 30 mg/dL and 53 mg/dL, respectively, without any signal toward increased adverse events related to very low levels of LDL, including new-onset diabetes and neurocognitive events. Furthermore, a post hoc analysis of FOURIER showed no significant association between achieved levels of LDL and safety outcomes, including in those patients among the lower decile, with LDL less than 20 mg/dL. Although definitive conclusions about the safety of these new therapies should await long-term results, thus far randomized data have attested to the safety and effectiveness of aggressive lipid lowering in secondary prevention patients.
Based on the aforementioned trials, it is appropriate to intensify lipid-lowering pharmacotherapy in high-risk patients with LDL of at least 70 mg/dL, as these were the inclusion criteria in those trials. The new guidelines encourage a stepwise approach starting with ezetimibe and then next considering a PCSK9 inhibitor if LDL remains at least 70 mg/dL. We feel that clinicians should also consider the patient’s current LDL level and the desired level, along with drug costs and oral vs. injectable administration. Ezetimibe is expected to reduce LDL levels by about 20%, whereas currently approved doses of PCSK9 inhibitors lower LDL threefold more, by about 60%.
Focus on lifestyle interventions
Emphasis on lifestyle interventions for lipid management and global CV health is expanded in the current guidelines. Smoking cessation, weight reduction, a healthy dietary pattern and participation in regular physical activity are advised for all patients. In patients with previous atherosclerotic CVD events or high-risk primary prevention, more intensive lifestyle efforts are advised, in addition with the preventive pharmacotherapy. Clinicians should share with patients and their families Life’s Simple 7 from the AHA — control BP, cholesterol levels and blood glucose, get active, eat better, lose weight and stop smoking. There are a number of patient-friendly resources available at http://www.heart.org/en/healthy-living/healthy-lifestyle/my-life-check--lifes-simple-7 on how to achieve these goals.
Emphasis on risk assessment, including lifetime risk
A key concept in preventive cardiology is the influence of baseline risk on the expected magnitude of benefit from prevention therapies. Assuming a similar relative reduction in risk from a given intervention, such as LDL lowering or BP control, groups with a higher absolute risk at baseline will derive a larger absolute risk reduction and lower number needed to treat as compared with populations with lower baseline risk. Therefore, risk assessment is critical in deciding intensity of prevention strategies.
Patients with previous atherosclerotic CVD events (ie, secondary prevention) generally fall under the highest-risk categories, for whom the most aggressive prevention therapies should be allocated. Yet, the secondary prevention population is not entirely homogeneous. In a post hoc analysis of IMPROVE-IT, patients with diabetes and those with a high-risk score for secondary prevention (eg, smoking, peripheral artery disease, chronic kidney disease) benefited the most from ezetimibe. Similarly, subgroup analyses from FOURIER identified a higher magnitude of benefit in patients at highest risk, such as those with PAD, recent MI, multiple prior MIs or residual multivessel CAD. Enlightened by these analyses, the new guidelines identify a very high-risk group as those with recent ACS, a history of MI or ischemic stroke, or symptomatic PAD.
In primary prevention, the guidelines continue to support use of validated clinical scores for risk estimation. Besides their value in guiding intensity of prevention therapies, evidence suggests that utilizing risk scores may reduce CV risk factors and increase the use of preventive medications. In particular, the 2018 ACC/AHA Cholesterol Guidelines endorse the U.S.-derived Pooled Cohort Equations (PCE) to estimate 10-year risk for atherosclerotic CVD events (coronary or stroke death, nonfatal MI or nonfatal stroke). They acknowledge limitations of the PCE, such as overestimation of risk in groups with 10-year predicted risk more than 10%, higher socioeconomic status or those receiving consistent preventive care and screening. On the contrary, risk can be underestimated in those with lower socioeconomic status or with chronic inflammatory conditions, such as HIV and rheumatologic diseases, which are not accounted for in the PCE.
The 2018 ACC/AHA Risk Assessment Statement highlights the importance of lifetime risk assessment in younger individuals (younger than 50 years), in whom the 10-year risk may be low due to age, but the lifetime risk may be elevated due to comorbidities. Acknowledging elevated long-term risk may be important in promoting lifestyle changes and risk-modifying behavior. In selected patients, it may also guide pharmacotherapy. Lipids mediate atherosclerotic risk according to the lifetime cumulative exposure to atherogenic lipoproteins, which can be estimated by multiplying a person’s age by their average LDL concentration. Thus, young patients with high LDL levels or high long-term risk likely benefit from lipid-lowering therapies to prevent atherosclerosis.
Increased accuracy in estimation of LDL
In the era of aggressive lipid-lowering therapies, with very low LDL targets in the secondary prevention population, it is important to have reliable and accurate assessments of LDL levels. Unfortunately, direct measurement of LDL by ultracentrifugation is not feasible in clinical practice due to its high cost, cumbersome technique and time requirement. Historically, LDL has been estimated by the Friedewald equation. Created in 1972 based on a population of 448 patients with primary dyslipidemias and fasting triglycerides less than 400 mg/dL, the Friedewald equation assumes a fixed ratio of 5:1 between triglyceride levels and VLDL. Estimated VLDL levels are then subtracted along with measured HDL from measured total cholesterol to yield estimated LDL levels.
The Friedewald equation, unfortunately, is prone to error due to several reasons. First and foremost, the ratio of triglyceride to VLDL is not fixed and can vary substantially. Second, ever increasing conditions such as obesity, diabetes and metabolic syndrome increase the ratio of triglycerides to VLDL, thus underestimating LDL levels according to the Friedewald equation. Nonfasting states and any cause of hypertriglyceridemia also alter the ratio of triglycerides to VLDL, leading to inaccuracies in LDL estimation. Of note, the Friedewald equation is particularly prone to underestimation of LDL levels in the lower range of LDL, precisely where accuracy is most important in the modern era of aggressive lipid-lowering therapies. Nearly one-quarter of patients with Friedewald-estimated LDL less than 70 mg/dL have a value of at least 70 mg/dL by ultracentrifugation.
In light of these concerns, a novel, more accurate method for LDL estimation was developed for clinical use. The Martin-Hopkins equation replaces the triglyceride to VLDL fixed ratio of 5 in the Friedewald equation with an adjustable factor, ranging from 3.1 to 11.9. This new equation has been validated and shown to be superior to Friedewald in several worldwide cohorts. It has gained recent widespread acceptance by several laboratories, such as Quest Diagnostics, as standard of care for LDL estimation.
Clinician-patient risk discussion, CAC as decision aid
The 2018 ACC/AHA Cholesterol Guidelines and Risk Assessment Statement underscores the importance of an informed clinician-patient discussion in CV prevention, particularly in primary prevention. Physicians should share with their primary prevention patients their 10-year predicted atherosclerotic CVD, as well as the lifetime risk, when appropriate. Risk assessment should be the starting point to frame a conversation about CV risk, lifestyle changes and pharmacotherapies, if indicated. The guidelines expand the tool kit for shared decision-making, highlighting a number of simple risk-enhancing factors that favor statin therapy when the decision is otherwise uncertain.
One of the most welcomed changes from the 2018 ACC/AHA Cholesterol Guidelines and Risk Assessment Statement is the recommendation for selective use for coronary artery calcium (CAC) for further risk stratification in primary prevention patients at intermediate risk. This recommendation is in sharp contrast to the recently published U.S. Preventive Services Task Force on Cardiovascular Risk Assessment with Nontraditional Risk Factors, which concluded that the current evidence is insufficient to assess the balance of benefits and harms of adding CAC score to traditional CV risk assessment.
We and others disagree with the USPSTF conclusion. As a direct measure of subclinical atherosclerosis, CAC is a powerful tool in risk prediction, with clear ability to improve atherosclerotic CVD risk prediction beyond traditional risk factors. Improved risk prediction may not be warranted in a large group of patients who may be at extremes of risk based on risk factor profile alone. Nevertheless, several groups may benefit from CAC for further refinement of risk category.
The first of these groups is the population at intermediate- or borderline-risk patients, as recommended in the 2018 ACC/AHA Guidelines, where intermediate and borderline risk were defined as a predicted 10-year atherosclerotic CVD risk between 7.5% to 20% and 5% to 7.5%, respectively. This is in line with a 2017 expert consensus statement from the Society of Cardiovascular Computed Tomography. Several studies have attested the power of CAC to redefine risk in this population. Khurram Nasir, MD, MPH, and colleagues showed that among nearly 2,000 patients from the MESA study with a baseline PCE-predicted 10-year atherosclerotic CVD risk of 5% to 20%, patients with a CAC greater than 0 Agatston units (AU) had an event rate that was twofold to fivefold higher than those with zero CAC.
Approximately 50% of those patients had a CAC of zero. Those patients had a very low event rate, below currently accepted thresholds for statin therapy in primary prevention. Other studies have confirmed the prognostic “power of zero” CAC to predict a low rate of atherosclerotic CVD events. Michael J. Blaha, MD, MPH, and colleagues showed that a CAC of zero resulted in the greatest downward shift in atherosclerotic CVD risk as compared with 11 other clinical, imaging and biomarker-related nontraditional risk factors.
The 2018 ACC/AHA Cholesterol Guidelines specifically added that clinicians and patients may consider avoiding or postponing lipid-lowering pharmacotherapy in individuals with no CAC, given the low event rates in this population. Indeed, in a study of patients without CAC at Walter Reed Army Medical Center, investigators found that when compared with patients without statin exposure, those with statin therapy had similar incidence of major adverse CV events in a follow-up of nearly 10 years (HR = 1).
CAC scoring may also be particularly helpful in the population where risk is driven by chronological age, if age is discordant from the remainder of the risk factor profile. This occurs in young patients with a high risk-factor burden, where short-term risk is low due to chronological age, and in elderly patients without other traditional risk factors, where risk is predicted as high primarily due to age. In a study of nearly 7,000 asymptomatic adults aged 45 to 84 years, approximately 1 in 7 patients aged 45 to 54 years had CAC greater than 0, whereas 1 in 7 individuals aged 75 to 84 years had CAC of 0. Furthermore, a CAC of 0 in the elderly population aged 75 to 84 years predicted a low incidence of CHD events (1.5 per 1,000 person-years), which was lower than the risk in the population aged 45 to 54 years with CAC of 1 to 100 (3.8 per 1,000 person-years) or CAC greater than 100 (21.1 per 1,000 person-years). Thus, once CAC is known, chronological age is less important in determining atherosclerotic CVD risk.
The 2018 ACC/AHA Cholesterol Guidelines continue with the recommendation for statin therapy in adults with diabetes older than 40 years with no previous atherosclerotic CVD events. Although certainly appropriate for this high-risk population as a whole, a fairly large subset of these patients is at lower risk, and they can be identified by CAC testing. Thus, a reasonable alternative would be to perform CAC testing in those without a high burden of other risk factors, particularly in statin-reluctant individuals. Among 881 asymptomatic individuals with diabetes from the MESA cohort, 37% had a CAC of 0. This was only slightly lower than the 55% rate of CAC of 0 in the population without diabetes. More importantly, the event rate in the diabetic cohort with CAC of 0 was exceedingly low, at 3.7 per 1,000 person-years. Therefore, it would be reasonable to consider CAC scoring for this population for improved risk assessment in statin-reluctant persons who have had diabetes for less than 5 years, and withholding lipid pharmacotherapy if no other indications, with reassessment of CAC in 5 years.
An additional instance where CAC testing may be particularly insightful is in patients with a low to intermediate burden of traditional risk factors and comorbid inflammatory disease states, such as lupus, psoriasis rheumatoid arthritis or HIV/AIDS. These conditions are stated in the 2018 Risk Assessment Statement as risk-enhancing factors, although notably the writing group acknowledges it is difficult to determine to what extent risk is elevated in this population. CAC can help the clinician in this regard.
In a study including 129 psoriatic patients with a mean age of 51 years, and average 10-year Framingham Risk Score of 9%, 58% had no CAC. Similarly, among patients with rheumatoid arthritis aged 46 to 60 years, without known atherosclerotic CVD, 60% have a CAC of 0 AU. In participants of the Swiss HIV cohort study, aged at least 45 years, no documented atherosclerotic CVD, and a median Framingham Risk Score of 9.4%, 47% had no CAC. Admittedly, more long-term outcome data in these special groups are warranted before definitive recommendations can be made with regard to lipid-lowering medications; however, these studies clearly show that it is possible to risk-stratify such groups beyond the binary presence or absence of these higher-risk conditions.
Finally, we highlight a paradigm shift in the new guidelines toward lower CAC thresholds to initiate statin therapy, in addition to lifestyle-related interventions. In the 2013 ACC/AHA Guideline on the Treatment of Blood Cholesterol, statin therapy received a class IIb recommendation in individuals who did not otherwise meet criteria for moderate- or high-intensity statin, and had a CAC of at least 300 AU or at least 75th percentile for age/sex/race/ethnicity. The 2018 Risk Assessment Statement advises statin therapy in patients with a CAC of at least 100 AU or the 75th percentile for age/sex/race/ethnicity, because these patients have event rates in a range where the benefits of statins clearly outweigh the risks of therapy. In the CARDIA study, even among patients aged 32 to 46 years, CAC of at least 100 AU was associated with a mortality of 22.4 per 100 participants during 12.5 years of follow-up, which was nearly 10-fold higher than the population with CAC of 0 AU. Similarly, Nasir and colleagues reported an atherosclerotic CVD rate threefold to eightfold higher in patients with CAC greater than 100 AU as compared with CAC of 0 AU.
We have highlighted important aspects of the recently published 2018 ACC/AHA Risk Assessment Statement and Cholesterol Guidelines. These documents represent an important advance in the field of preventive cardiology, incorporating new evidence for LDL thresholds to guide use of evidence-based nonstatin therapies in the highest risk patients, and a more prominent role assigned to CAC testing for risk stratification and guidance of treatment strategies in primary prevention.
Baigent C, et al. Lancet. 2010;doi:10.1016/S0140-6736(10)61350-5.
Blaha MJ, et al. Circulation. 2016;doi:10.1161/CIRCULATIONAHA.115.018524.
Blaha M, et al. JACC Cardiovasc Imaging. 2009;doi:10.1016/j.jcmg.2009.03.009.
Bonaca MP, et al. Circulation. 2018;doi:10.1161/CIRCULATIONAHA.117.032235.
Budoff MJ, et al. Am Heart J. 2009;doi: 10.1016/j.ahj.2009.08.007.
Cannon CP, et al. N Engl J Med. 2015;doi:10.1056/NEJMoa1410489.
Carr JJ, et al. JAMA Cardiol. 2017;doi:10.1001/jamacardio.2016.5493.
Chen Y, et al. Lipids Health Dis. 2010;doi:10.1186/1476-511X-9-52.
Curry SJ, et al. JAMA. 2018;doi:10.1001/jama.2018.8359.
Elias-Smale SE, et al. Eur Heart J. 2011;doi:10.1093/eurheartj/ehr125.
Ference BA,et al. J Am Coll Cardiol. 2018;doi:10.1016/j.jacc.2018.06.046.
Friedewald WT, et al. Clin Chem. 1972;18:499-502.
Giugliano RP, et al. Lancet. 2017;doi:10.1016/S0140-6736(17)32290-0.
Giugliano RP, et al. Circulation. 2018;doi:10.1161/CIRCULATIONAHA.117.030950.
Hecht H, et al. J Cardiovasc Comput Tomogr. 2017;doi:10.1016/j.jcct.2017.02.010.
Karmali KN, et al. Cochrane Database Syst Rev. 2017;doi:10.1002/14651858.CD006887.pub4.
Karpouzas GA, et al. Rheumatology (Oxford). 2018;doi:10.1093/rheumatology/key057.
Lee J, et al. PLoS One. 2016;doi:10.1371/journal.pone.0148147.
Mansouri B, et al. JAMA Dermatol. 2016;doi:10.1001/jamadermatol.2016.2907.
Martin SS, et al. J Am Coll Cardiol. 2013;doi:10.1016/j.jacc.2013.01.079.
Martin SS, et al. JAMA. 2013;doi:10.1001/jama.2013.280532.
Martin SS, et al. JAMA Cardiol. 2018;doi:10.1001/jamacardio.2018.1533.
Mitchell J, et al. J Am Coll Cardiol. 2018;doi:10.1016/j.jacc.2018.09.051.
Nasir K, et al. J Am Coll Cardiol. 2015;doi:10.1016/j.jacc.2015.07.066.
Paixao AR, et al. JACC Cardiovasc Imaging. 2015;doi:10.1016/j.jcmg.2015.06.015.
Palmer MK, et al. Clin Biochem. 2018;doi:10.1016/j.clinbiochem.2018.10.011.
Polonsky TS, et al. JAMA Cardiol. 2018;doi:10.1001/jamacardio.2018.2199.
Rim JH, et al. Medicine (Baltimore). 2016;doi:10.1097/MD.0000000000003230.
Sabatine MS, et al. Circulation. 2018;doi:10.1161/CIRCULATIONAHA.118.034309.
Sabatine MS, et al. N Engl J Med. 2017;doi:10.1056/NEJMoa1615664.
Schwartz GG, et al. N Engl J Med. 2018;doi:10.1056/NEJMoa1801174.
Silverman MG, et al. Eur Heart J. 2014;doi:10.1093/eurheartj/eht508.
Stone NJ, et al. J Am Coll Cardiol. 2014;doi:10.1016/j.jacc.2013.11.002.
Tarr PE, et al. Eur Heart J. 2018;doi:10.1093/eurheartj/ehy163.
Tota-Maharaj R, et al. Mayo Clin Proc. 2014;doi:10.1016/j.mayocp.2014.05.017.
Whelton SP, et al. J Clin Lipidol. 2017;doi:10.1016/j.jacl.2017.05.005.
For more information:
Rhanderson Cardoso, MD, is a cardiology fellow at Johns Hopkins University School of Medicine. Roger S. Blumenthal, MD, is director of the Johns Hopkins Ciccarone Center for the Prevention of Heart Disease and professor of medicine at Johns Hopkins University School of Medicine. He is also the editor of the Prevention section of Cardiology Today. Seth S. Martin, MD, MHS, is director of the Advanced Lipid Disorders Program of the Johns Hopkins Ciccarone Center for the Prevention of Heart Disease and assistant professor of medicine at Johns Hopkins University School of Medicine. He is also a member of the Cardiology Today Editorial Board. The authors can be reached at Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Division of Cardiology, Johns Hopkins University School of Medicine, 600 N. Wolfe St., Halsted 560, Baltimore, MD 21827.
Disclosures: Cardoso and Blumenthal report no relevant financial disclosures. Martin reports he receives personal fees for serving on scientific advisory boards for Akcea Therapeutics, Amgen, Esperion, Novo Nordisk, Quest Diagnostics and Sanofi/Regeneron; he receives grant support from the Aetna Foundation, the American Heart Association, Apple, Google, the Maryland Innovation Initiative, the NIH, Nokia, the PJ Schafer Cardiovascular Research Fund and the David and June Trone Family Foundation; and he is listed as a co-inventor on a patent application pending on a method of LDL estimation.