How Low to Go With Glucose, Cholesterol, and Blood Pressure in Primary Prevention of CVD

Kimberly N. Hong, MD, MHSA; Valentin Fuster, MD; Robert S. Rosenson, MD; Clive Rosendorff, MD, PHD; Deepak L. Bhatt, MD, MPH


J Am Coll Cardiol. 2017;70(17):2171-21-85. 

In This Article


Atherosclerosis begins with the deposition of lipids in the vascular wall and is mediated by disturbances in cholesterol homeostasis.[42] Prior to the current AHA/ACC 2013 guidelines, treatments had focused on target cholesterol for determining the initiation of statins and other nonstatin lipid-modifying therapies.[43] The new guidelines ushered in a paradigm shift that used CVD risk, instead of LDL-C levels, as a guide to treatment. Part of this was inspired by evidence that the atherosclerotic process began in infancy and possibly even earlier, with genetic factors influencing the development of CAD. For instance, 1 study that used intracoronary ultrasound on donor hearts found the presence of plaque in 17% of hearts between the ages of 13 to 19 years and in 60% of hearts between the ages of 30 to 39 years.[2] This change in guidelines significantly expanded the indication for statins, with 1 study estimating an increase from 43 million individuals (37.5% of the U.S. adult population) to 56 million (48.6%). Appropriately, a majority of these newly eligible individuals met the primary prevention criteria (10 million).[44,45]

When Should Statin Therapy be Initiated?

The Cardiovascular Risk in Young Finns study followed individuals from age 3 to 18 years for 27 years and found that physical inactivity and reduced fruit intake correlated with accelerated carotid intima-media thickness. Furthermore, correction of these risk factors during childhood attenuated the risk for progression during adulthood.[2] The Bogalusa Heart Study similarly followed individuals from childhood to adulthood.[46,47] Autopsy data on study participants who died from non-CVD deaths found an increasing prevalence of coronary fatty streaks with age (50% of patients age 2 to 15 years and 85% of patients age 21 to 39 years). Antemortem body mass index, SBP, total cholesterol, and LDL-C were correlated with the presence of coronary fatty streaks.[46] A follow-up survey conducted in 2000 to 2001 on nondeceased participants (mean age 31.9 years) included carotid intima-media thickness by ultrasound and found that higher childhood measurements of non–high-density lipoprotein cholesterol, LDL-C, and apoB were associated with increased carotid intima-media thickness.[47] The CARDIA (Coronary Artery Risk Development in Young Adults Study), on the other hand, suggested a role for more aggressive lipid-lowering therapy. This was a longitudinal study that followed patients age 18 to 30 years for either 15 or 20 years until a coronary artery calcium scan after the age of 35 years. Interestingly, coronary artery calcium scores >0 were associated with LDL-C >100 mg/dl prior to the age of 35 years, even after risk adjustment.[48] These studies suggest that in childhood to early adulthood, lifestyle changes should be implemented. However, cholesterol-lowering therapies can be considered if LDL-C levels remain elevated after institution of these interventions.

Although there have been pharmacological trials looking at the primary prevention population, the majority of them have focused on higher-risk populations. Two trials that included low-risk populations were the MEGA (Management of Elevated Cholesterol in the Primary Prevention Group of Adult Japanese) and JUPITER (Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin) trials. MEGA compared incidence of CAD in individuals age 40 to 70 years with a total cholesterol between 220 and 270 mg/dl who were treated with diet alone versus diet plus pravastatin. Incidence of CAD was significantly lower in the statin treatment group.[49] The JUPITER trial included men age ≥50 years and women age ≥60 years with an LDL-C <130 mg/dl and a high-sensitivity CRP ≥2.0 mg/dl, of whom 50% had a Framingham score ≤10%.[50] These individuals were randomized to rosuvastatin versus placebo, with the primary endpoint being a composite of MI, stroke, arterial revascularization, hospitalization for unstable angina, or CVD death. Similar to MEGA, the JUPITER trial showed a reduction in the primary endpoint in the statin treatment group.

More recently, HOPE-3 (Heart Outcomes Prevention Evaluation–3) was a 2 × 2 factorial design trial that assessed treatment with a candesartan and hydrochlorothiazide combination pill and rosuvastatin 10 mg compared with placebo in individuals with intermediate CVD risk. Inclusion criteria for this trial were men age ≥55 years and women age ≥65 years with at least 1 of the following CVD risk factors: elevated waist-to-hip ratio, low high-density lipoprotein cholesterol, current or recent tobacco use, dysglycemia, and family history of premature CAD; women age ≥60 years with 2 of the aforementioned cardiovascular risk factors were also included. Pre-diabetic patients were included.[51,52] Of note, the trial protocol only excluded "documented clinically manifest" CVD and did not require either invasive or noninvasive confirmation. To assess tolerability to treatment and avoid drop-outs and adherence issues, there was a 4-week run-in phase for the medications, and only those without significant side effects underwent randomization. The primary outcomes were a composite that included CVD death, nonfatal MI, or nonfatal stroke, and the prior composite endpoint with the addition of resuscitated cardiac arrest, heart failure, or revascularization. Secondary outcomes included DM, cognitive function, and erectile dysfunction. Safety outcomes included cancer, myopathy, rhabdomyolysis, and hospitalization.[52]

This trial randomized individuals between 2007 and 2010. Median follow-up was 5.6 years. At baseline, the mean age of the trial participants was 65.7 years, and 5.8% had uncomplicated DM. Drug adherence by the end of the trial was ~75%. The incidence of the primary outcome was reduced in the treatment group compared with the placebo group (Table 2). Secondary analyses, which included the individual components of the primary endpoint, revealed stroke to be a primary contributor to the effect size of treatment with rosuvastatin. There was also no change in treatment effect in subgroup analyses stratified by demographics, SBP, LDL-C, and CRP.[52]

Notably, the incidence of new-onset DM was the same in both groups. In terms of safety outcomes, although the incidences of myalgias and muscle weakness were increased in the treatment group (5.8% vs. 4.7%; p < 0.005), there was no difference in liver enzyme elevation or permanent discontinuation of the study drug because of muscle complaints including rhabdomyolysis or myopathy.[52]

Last, ECAD (Eliminate Coronary Artery Disease) is an ongoing trial designed to determine whether treatment with atorvastatin reduces CVD events in a lower-risk population than previously evaluated. The study population includes men age 35 to 50 years and women age 45 to 59 years who have no history of CVD, an LDL-C ≥70 mg/dl, and at least 1 of the following risk factors: current smoking, HTN, truncal obesity, family history of premature MI (before 60 years of age), or South Asian ethnic history.[53]

As lower-risk populations are considered for statins, the adverse effect profile needs to be considered. It must be understood that the risk–benefit ratio favoring statins decreases as the patient population moves into a lower-risk group and as the intensity of the statin increases.[54] The CTT (Cholesterol Treatment Trialists') Collaboration concludes that although there may be a slight increase in hemorrhagic strokes, myopathy and DM are the only significant adverse effects that can be reliably attributed to statin use. Furthermore, elevated transaminases, defined as ≥3× the normal limit, in isolation are not specific for significant hepatocellular injury. As a result, routine surveillance of transaminases is not recommended due to the inappropriate discontinuation of statins.[55,56] Similarly, a recent review of statin trials showed no association between statins and objective measures of cognitive decline.[57]

How Much can LDL-C Levels be Lowered?

Another controversy in the 2013 ACC/AHA cholesterol guidelines is the removal of specific LDL-C goals as targets of therapy.[43] There is no definitive evidence that there is a true lower threshold for LDL-C levels. Two trials evaluating the addition of nonstatins to statin therapy, IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial) and FOURIER (Further Cardiovascular Outcomes Research With PCSK9 Inhibition in Subjects With Elevated Risk), found that incremental decreases in LDL-C were associated with fewer cardiovascular events and no increase in serious adverse events. IMPROVE-IT randomized patients following an acute coronary syndrome event to simvastatin with or without ezetimibe, and FOURIER randomized patients who were at high CVD risk (with 81% having had a prior MI) to atorvastatin with or without evolocumab.[58,59] Although both trials targeted secondary CVD prevention cohorts, these results suggest a potential role for investigating nonstatins in primary CVD prevention in very high-risk patients. Longer-term follow-up from proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor trials, such as the FOURIER OLE, will provide more information concerning the safety of very low LDL-C levels.[60,61] From a biological perspective, neonatal levels of LDL-C at birth range from 21 to 39 mg/dl and have been postulated as a physiological lower limit.[22,62,63] Results from the PCSK9 trials, which found no difference in adverse events, including neurocognitive events, hemorrhagic strokes, or incidence of new-onset DM between patients with an LDL-C <25 and ≥25 mg/dl, suggest that there is no clinically significant lower LDL-C limit.[64–67] The main caveat to extrapolating these data is limited follow-up and differences in the patient populations. In particular, as mentioned earlier, the majority of patients enrolled in the PCSK9 inhibitor trials were secondary prevention patients who were older and who likely had reduced exposure time to low levels of LDL-C compared with a primary prevention cohort.

A meta-analysis carried out by the CTT Collaboration reported a 1% relative risk reduction for every 1.8 mg/dl of LDL-C lowered to a lower threshold of 50 mg/dl in individuals with a baseline cholesterol of >78 mg/dl.[20,22,64,68] Efforts to lower LDL-C are driven by the concept of residual risk even after optimal risk factor modification. Part of this may be some unaccounted epiphenomenon or that lifetime exposure to LDL-C needs to be considered. This perspective was introduced in Mendelian studies examining people with PCSK9 loss of function who have a disproportionately lower lifetime risk for CAD compared with those enrolled in clinical trials. A meta-analysis of these studies showed that lifelong exposure to lower levels of LDL-C conferred a 3-fold greater risk reduction than treatment with statins started later in life.[69] Current guidelines recommend a decrease in statin dose after 2 consecutive readings of LDL-C <40 mg/dl.[18] However, data from PCSK9 inhibitor trials have not shown safety concerns at LDL-C levels <25 mg/dl.[60,64–67]