Clinical Trial Evidence for Treatment of Dyslipidemia in the Prevention of Cardiovascular Disease
Thomas A. Pearson
Randomized trials with well-defined clinical end points provide the strongest proof that a purported risk factor is in fact causal for a disease. The role of randomized clinical trials (RCTs) in support of the “cholesterol hypothesis” best illustrates this. Initial evidence from early pathology studies (cholesterol crystals in atherosclerotic plaques) and animal feeding studies provided the impetus to measure cholesterol, triglycerides, and the lipoprotein subfractions in blood and relate their levels to coronary disease in case-control and prospective epidemiologic studies. Yet, this considerable evidence was insufficient to justify a National Cholesterol Education Program (NCEP) that would set down guidelines for the screening, diagnosis, and treatment of lipid disorders, until the release of a definitive RCT, the Lipid Research Clinics Coronary Primary Prevention Trial (LRCCPPT) in 1984 (1). The RCT has established itself as the definition of efficacy and safety for lipid-lowering pharmacotherapy and the basis of evidence-based clinical lipidology.
The subsequent 25 years have provided an enormous collection of evidence from experiments involving human subjects favoring the lowering of low-density lipoprotein cholesterol (LDL-C) levels (Table 5.1) (1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25). The development of evidence might be viewed as a series of landmark trials creating the rationale for progressive changes in guidelines. The early RCTs used older, lower-potency drugs with placebo controls, typically demonstrating 15% to 20% reduction in clinical coronary events, requiring high-risk subjects, careful management of run-in periods, low dropout rates, and large sample sizes to have adequate statistical power. Their aggregate evidence, especially that of the LRC-CPPT, supported the launch of the NCEP and its Adult Treatment Panel (ATP) I in 1988.
The next series of trials, 14 or so in number, used progression of atherosclerotic stenosis, as defined by highly sensitive quantitative coronary arteriography, to test a wide array of lipid-lowering interventions (see also Table 5.2). Since these studies were largely performed in clinical cardiology inpatient units, they engaged cardiologists in terms that they could relate to, namely the extent of arteriographically defined coronary disease. This body of evidence convinced these opinion leaders of the power of lipid lowering as a primary therapy for coronary atherosclerotic disease, rather than as an adjunct to secondary preventive drugs or revascularization. The relationship between end-treatment LDL-C level and reduction in atherosclerotic progression was used to define the LDL-C goal of 100 mg/dL or less as a therapeutic target (25).
The next series of landmark RCTs involved definition of safety and efficacy for new, more powerful LDL-lowering drugs, the hydroxymethylglutaryl-CoA (HMG-CoA) reductase inhibitors (statins), especially the “natural” statins first developed for human testing. These trials involved large numbers of subjects followed for 4 to 6 years with adequate power and greater reductions in LDL-C (30% to 40%) than previously possible. Most were performed in patients with coronary disease (4S (11), CARE (13), LIPID (14)), or persons without disease but at relatively high coronary risk (WOSCOPS) (12). The exception was AFCAPS/TexCAPS (15) which studied a rather lower-risk group. The study results were striking for their consistency and magnitude (24% to 42%) of reduction of clinical end points.
Since 2001, two or three types of RCTs have enriched the literature. First, a number of trials have broadened the type of patients studied in trials. The Heart Protection Study (16), for example, enrolled patients with a range of vascular diseases (coronary disease, stroke), as well as diabetes. It randomized over 20,000 subjects to simvastatin (40 mg/day) versus placebo and allowed a large number of subgroups defined by LDL level, gender, age, type of vascular disease or diabetes, etc. to be examined. Other trials specifically studied persons with hypertension (ASCOT-LLA (17), ALLHAT-LLA (18)), or elderly persons (PROSPER (19)). Second, recent trials have not been able to use placebo controls, on ethical grounds, and have therefore compared two different statins (e.g., PROVE-IT (20), IDEAL (22)), different doses of the same drug (e.g., Treating to New Targets (TNT) (21)), or monotherapy versus an additional agent (e.g., ENHANCE (26)). These trials have led to a reconsideration of target LDL-C levels in publications updating the NCEP ATP III in 2004 (27) and an American Heart Association and American College of Cardiology (AHA/ACC) update in 2006 (28).
Most recently, increasingly low-risk subjects are being studied for evidence of safety and efficacy. AFCAPS/TexCAPS (15) provided initial evidence for large relative risk reductions (RRRs) in events despite low absolute risk. The JUPITER study (23) in patients with lower LDL-C levels but abovemedian high-sensitivity C-reactive protein (hs-CRP) levels also demonstrated a large RRR despite the relatively low risk of the study population. It is anticipated that any update of the NCEP ATP will address these and other recent developments.
Few specialties of medicine have as rich an RCT evidence base as does clinical lipidology. The goal of this chapter is to review the design and conduct of randomized clinical trials of lipid-modifying interventions and to identify the key lessons that have been learned from them. RCTs that were solely performed
in patients with diabetes or in whom there was a significant detailed analysis of subsets of diabetic patients are discussed in detail in Chapter 10.
in patients with diabetes or in whom there was a significant detailed analysis of subsets of diabetic patients are discussed in detail in Chapter 10.
TABLE 5.1 DEVELOPMENT OF ADULT TREATMENT PANEL GUIDELINES AND SOURCES OF EVIDENCE TO SUPPORT THEM: A HISTORICAL PERSPECTIVE | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
THEORETICAL CONSIDERATIONS: CONCEPTUAL ISSUES IN RANDOMIZED TRIALS OF LIPID-MODIFYING THERAPIES
The design of a randomized trial usually entails the identification and recruitment of subjects at risk for the study end point, random allocation to treatment arms, interventions differing between the treatment arms, follow-up for preordained and standardly defined end points, and prespecified statistical analyses of differences in rates of end points between the arms.
The defining feature of an RCT trial is the random allocation of subjects to different treatments, thereby specifying the level of “exposure” and assuming that systematic differences between exposed and unexposed groups do not occur other than by the treatments. Nonrandom self-selection of persons into exposure groups has been frequently observed in observational studies, confounding the results. This has not apparently been as important in lipid-lowering therapy, but has been a major lesson from other RCTs in preventive cardiology. For example, observational studies of estrogen replacement therapy (in women) or antioxidant vitamins had demonstrated extensive observational evidence in favor of their use, only to have RCT solidly refute any treatment benefit and occasionally suggest potential harm.
The primary aim of randomized trials is comparability of subjects between treatment arms, achieved through success in the random allocation of subjects. Often, run-in phases and multiple prerandomization visits select subjects who are exceptionally compliant and committed to the RCT. Successful randomization is usually demonstrated in the first table of results in the publication of an RCT, not only describing the subjects but also comparing characteristics of potential importance between
treatment arms. Nonrandom allocation should be obvious with numerous imbalances unlikely due to chance. Occasional imbalances may occur by chance and can be adjusted for in the analysis. Probably the most important role of randomization is to balance the allocation of important determinants of end points that are not known or measured. While comparability is the primary concern, some attention should be given to generalizability of the results. Do the results of the trial have relevance to other groups of patients? For example, a recent RCT, the ENHANCE study (26), randomized middle-aged patients with familial hypercholesterolemia (FH) on statins to ezetimibe or placebo. The interpretation of that negative trial has unjustifiably generalized this lack of benefit to all hyperlipidemic patients, whose lipid disorders may be very different in terms of severity and duration from those of patients with lifelong FH.
treatment arms. Nonrandom allocation should be obvious with numerous imbalances unlikely due to chance. Occasional imbalances may occur by chance and can be adjusted for in the analysis. Probably the most important role of randomization is to balance the allocation of important determinants of end points that are not known or measured. While comparability is the primary concern, some attention should be given to generalizability of the results. Do the results of the trial have relevance to other groups of patients? For example, a recent RCT, the ENHANCE study (26), randomized middle-aged patients with familial hypercholesterolemia (FH) on statins to ezetimibe or placebo. The interpretation of that negative trial has unjustifiably generalized this lack of benefit to all hyperlipidemic patients, whose lipid disorders may be very different in terms of severity and duration from those of patients with lifelong FH.
Related posts:
Disorders of Hypertriglyceridemia
Dyslipidemia and Atherosclerosis in Women
Lipid, Apolipoprotein, and Lipoprotein Metabolism: Implications for the Diagnosis and Treatment of Dyslipidemia
Disorders of Hypertriglyceridemia
Dyslipidemia in the Elderly: Special Insights to Inform the Management Strategy
Trans Fatty Acids, Dyslipidemia, and Cardiovascular Disease
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
