Related articles on this website: In the Current Literature section: Statin Therapy, Lipid Levels, C-Reactive Protein and the Survival of Patients With Angiographically Severe Coronary Artery Disease Horne BD, Muhlestein JB, Carlquist JF, et al. J Am Coll Cardiol. 2000;36:1774-1780. In the Slide Library section: Lp(a): An Independent CHD Risk Factor in Men of the Framingham Offspring Cohort Homocysteine: Role in Atherogenesis

Several studies have provided evidence that assessment of emerging risk factors—those beyond established lipid parameters—may be useful in determining heart disease risk in certain individuals. The Third Adult Treatment Panel (ATP III) of the National Cholesterol Education Program (NCEP)—the new national guidelines for the detection, evaluation, and treatment of high blood cholesterol—proposes that some of these novel risk factors may be used to guide the intensity of risk-reduction therapy.
    This article focuses on the usefulness of non–high-density lipoprotein cholesterol (non–HDL-C), C-reactive protein (CRP), lipoprotein(a) (Lp[a]), and homocysteine (Hcy) as predictors of heart disease risk and mortality and as guides to the effectiveness of risk-reduction therapy.

Non–High-Density Lipoprotein Cholesterol
Non–HDL-C is defined as the difference between total cholesterol (TC) and HDL-C. It contains all known and potentially atherogenic lipid particles, including LDL-C, Lp(a), intermediate-density lipoprotein cholesterol (IDL-C), and very-low-density lipoprotein cholesterol (VLDL-C) remnants. (See sidebar on page 2 for non–HDL-C goals for three categories of heart-disease risk as defined by ATP III.)
    ATP III recommends that all adults aged 20 or older receive a full lipid profile (TC, HDL-C, LDL-C, and triglycerides [TG]) every 5 years. However, is it possible that non–HDL-C, which requires measurement of TC and HDL-C only, might be just as useful in predicting heart disease-related mortality? And, if non–HDL-C is used as an outcome measure in patients who are being treated for hypercholesterolemia, how effective are lipid-modifying therapies such as statins in lowering it?
LRC Study. Data from the Lipid Research Clinics (LRC) Program Follow-Up Study were analyzed to determine whether non–HDL-C, particularly when compared with LDL-C, is useful in predicting cardiovascular disease (CVD) mortality.¹ Researchers obtained information on 2,406 men and 2,056 women aged 40 to 64 at study entry, with no clinically evident CVD at baseline. Enrollment had taken place from 1972 through 1976, and mortality had been determined through 1995 (average of 19 years). Approximately 40% of those selected for this analysis had elevated lipid levels or used lipid-lowering medications. During follow-up, 234 CVD deaths had occurred in men and 113 in women. In both sexes, the risk for CVD death correlated positively with baseline non–HDL-C level. For example, compared with men whose non–HDL-C levels were <160 mg/dL, those whose levels ranged from 190 to 219 mg/dL had a 43% higher risk for CVD death (relative risk [RR], 1.43). In men with non–HDL-C >220 mg/dL, the RR for CVD death was 2.14. Likewise, compared with women whose non–HDL-C levels were <160 mg/dL, those whose levels ranged from 190 to 219 mg/dL had a 61% higher risk for CVD death (RR, 1.61). In women with non–HDL-C >220 mg/dL, the RR for CVD death was 2.43. However, each of these results had a wide confidence interval (CI), particularly in women.
    Baseline LDL-C levels in men also correlated positively with CVD mortality. Among women, no significant correlation was found between baseline LDL-C and subsequent CVD death. For both men and women, an increased risk for CVD death was inversely related to HDL-C level, with considerably narrower CIs than those related to the data for non–HDL-C.
    In men, non–HDL-C and HDL-C were equally good predictors of CVD mortality, whereas LDL-C was less predictive. In women, HDL-C was the best predictor of CVD death, non–HDL-C the second best, and LDL-C the poorest.
    The authors concluded that non–HDL-C is better than LDL-C in predicting CVD death. One reason may be that non–HDL-C includes all of the potentially atherogenic lipid fractions. Non–HDL-C may be particularly useful in certain patient subgroups, such as those with type 2 diabetes and hypertriglyceridemia.
ACCESS. Clinical trial data from the Atorvastatin Comparative Cholesterol Efficacy and Safety Study (ACCESS) Group indicate that apolipoprotein (apo) B is superior to LDL-C in predicting coronary heart disease (CHD) risk because it reflects the number of both LDL-C particles and TG-rich particles (eg, VLDL, IDL). As the screening measurement of apo B is not yet routinely available, non–HDL-C is considered the best surrogate measure of apo B and, according to the authors, may be more useful than LDL-C for assessing risk and as a therapeutic target. The ACCESS Group evaluated the effects of five HMG-CoA reductase inhibitors (statins) on lipid and apolipoprotein levels in 3,916 patients with hypercholesterolemia.² Subjects were randomized to receive open-label atorvastatin, fluvastatin, lovastatin, pravastatin, or simvastatin for 54 weeks; dosages were titrated up to the maximum allowable in order to meet NCEP LDL-C targets. Non–HDL-C targets were based on values corresponding to each NCEP LDL-C goal.
    Non–HDL-C and apo B were strongly correlated at baseline and at week 54 in patients overall and across CHD risk categories. The correlation between LDL-C and apo B was weaker, particularly in patients with CHD. All five statins lowered non–HDL-C; after 6 weeks and 54 weeks, atorvastatin had the greatest effect. Fewer patients reached non–HDL-C targets than LDL-C targets for each statin studied and across all risk classification strata; the difference was particularly pronounced in patients with CHD. In each risk category, atorvastatin recipients were the most likely to reach non–HDL-C and LDL-C targets. The authors concluded that reaching non–HDL-C targets would require treating patients more aggressively (particularly those at higher risk for CHD) than is the current practice with LDL-C targets.

C-Reactive Protein
Half of all coronary events occur in persons without overt hyperlipidemia.³ As atherosclerosis is thought to involve an inflammatory process,⁴ it has been proposed that certain plasma markers of inflammation, including CRP, might be particularly useful in assessing CVD risk.^3,4 This hypothesis raises several questions, including the following: How useful is CRP in predicting risk of a CV event? Would statins, which have antiinflammatory properties, reduce CRP levels? If so, would they prevent coronary events in persons with elevated CRP, even in those without overt hyperlipidemia?
WHS. Using a prospective nested case-control design, researchers in the ongoing Women's Health Study (WHS) compared baseline levels of high-sensitivity (hs) CRP, other markers of inflammation, and lipid parameters in women who did (n=122) or did not (n=244) experience a CV event.⁴ Median levels of hs-CRP were significantly higher in the cases than in the controls (0.42 vs 0.28 mg/dL). Among 12 markers measured, CRP was the strongest univariate predictor of CVD risk; women in the highest quartile were 4.4 times as likely as those in the lowest quartile to have a CV event. Even among women whose LDL-C was below the NCEP target level for primary prevention, hs-CRP levels were independently predictive of CVD risk.
CARE Study. In a nested case-control analysis of data from the Cholesterol and Recurrent Events (CARE) study, investigators sought to determine whether long-term treatment with pravastatin altered CRP levels.⁵ They compared CRP values at baseline with those at 5 years in 472 randomly selected participants in the CARE study, all of whom had a history of myocardial infarction, had been randomized to receive pravastatin or placebo for 5 years, and had remained free of recurrent coronary events during the study period. CRP values, similar in both groups at baseline, tended to increase over time in the placebo group (median change, +4.2%), whereas they dropped significantly in the pravastatin group (median change, -17.4%). Interestingly, no significant correlation was observed between the magnitude of change in CRP and the magnitude of change in LDL-C, TC, TG, or HDL-C among patients allocated to pravastatin or placebo. Together with evidence that elevated CRP levels are associated with increased CVD risk in the presence of normal lipid levels, these results suggest that CRP is a modifiable risk factor.
PRINCE. The Pravastatin Inflammation/ CRP Evaluation (PRINCE) was conducted to determine whether pravastatin has anti-inflammatory effects, as represented by CRP reduction.⁶ This prospective study, which enrolled both primary- and secondary-prevention patients, analyzed the effects of a 24-week course of pravastatin on CRP. Participants in the randomized, double-blind, primary-prevention cohort (n=1,702) received pravastatin 40 mg daily or placebo; those in the secondary-prevention cohort (n=1,182) received open-label pravastatin 40 mg daily. In the primary-prevention trial, pravastatin led to a median 16.9% reduction in CRP levels at 24 weeks compared with placebo. The secondary-prevention cohort fared similarly; CRP reductions at 12 and 24 weeks, compared with baseline values, were 14.3% and 13.1%, respectively. Researchers found minimal correlation between CRP reductions and changes in lipid values at 24 weeks. PRINCE was not designed as an end-point trial, so the effect of CRP reduction on CV events was not evaluated.
AFCAPS/TexCAPS. The relation between CRP and CV end-point events was addressed in a follow-up to the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS).³ Investigators measured CRP levels at baseline and after 1 year in 5,742 persons (aged 45–73 years) enrolled in this randomized, double-blind, placebo-controlled, primary-prevention trial of lovastatin. In the placebo group, rates of coronary events rose proportionately with baseline CRP; risk increased by 21% with each increasing quartile (95% CI, 4% – 41%). A 1-year course of lovastatin reduced median CRP by 14.8% (95% CI, 12.5% – 17.4%); this effect, which was significant, was unrelated to lovastatin's effect on lipid levels. In terms of reducing the risk for acute coronary events, lovastatin was effective not only in subjects with high LDL-C, regardless of their CRP level, but also in those with low LDL-C and high CRP; their risk was cut almost in half. In the placebo group, coronary event rates were just as high among subjects with low lipid levels and high CRP levels as among those with hyperlipidemia regardless of CRP levels. The authors concluded that lovastatin may be effective as primary prevention in persons with relatively low lipid levels but with elevated CRP. However, these hypothesis-generating results require further evaluation in clinical trials of patients who have evidence of systemic inflammation without overt hyperlipidemia.

Lipoprotein(a)
Many studies have shown a correlation between high levels of Lp(a), an LDL-C variant, and CHD.⁷ Lp(a) is elevated in 15% to 20% of the white population⁸; blacks are even more likely than whites to have elevated values. As with CRP, it is important to investigate whether Lp(a) is an independent predictor of CHD, including whether it might be particularly informative in blacks.
ARIC Study. The Atherosclerosis Risk in Communities (ARIC) study was conducted to determine whether CHD risk can be adequately assessed by measuring TC, LDL-C, HDL-C, and TG, or whether measurement of Lp(a), apo A-1, apo B, and/or HDL density subfractions would enhance prediction.⁹ A total of 12,339 middle-aged, CHD-free subjects were followed for 10 years. Over that time, 509 CHD events occurred in men and 216 occurred in women. Baseline levels of TC, LDL-C, TG, apo B, and Lp(a) were significantly higher, and those of HDL-C, apo A-1, HDL₂-C, and HDL₃-C were significantly lower, in men and women with subsequent CHD than in those without CHD.
    LDL-C, HDL-C, TG, Lp(a), and HDL₃-C were independently significant predictors of CHD, whereas apo B, apo A-1, and HDL₂-C were not. Overall, Lp(a) added only modest predictive value to that provided by LDL-C, HDL-C, and TG. Lp(a) associations in blacks, examined separately, were found to be somewhat less predictive. In subjects with subsequent CHD, Lp(a) values were higher than in control subjects. For black women, the difference approached statistical significance (P=0.07), but it was not significant in black men.

Homocysteine
This sulfur-containing amino acid is formed during the metabolism of methionine, and is eliminated through one of two vitamin-mediated pathways.¹⁰ Mild hyperhomocysteinemia is found in 5% to 7% of the general population.¹¹ Epidemiologic evidence shows an association between plasma Hcy and CVD, particularly in high-risk individuals, but a causal relation has not been established. Since low plasma levels of folic acid and vitamins B₆ and B₁₂ are associated with increased Hcy concentrations, it has been hypothesized that folic acid and/or vitamin B supplementation to decrease Hcy levels may reduce CVD risk.^10,12
WHS. Researchers used the database of the ongoing WHS, which has enrolled 28,263 postmenopausal women with no history of CVD at baseline, to generate this prospective case-control study (mean follow-up, 3 years).¹² A total of 122 women who subsequently experienced a CV event were matched with 244 women who remained disease-free to determine whether baseline Hcy might predict CVD risk. Baseline Hcy was found to be significantly higher in the cases than in the controls (14.1 vs 12.4 mmol/L). In addition, women with Hcy levels in the highest quartile were twice as likely as those in the lowest quartile to have a CV event. Nevertheless, the risk associated with Hcy elevation was modest and smaller than that observed with hs-CRP in this cohort.
HLTC Meta-Analysis. The Homocysteine Lowering Trialists Collaboration (HLTC) performed a meta-analysis of 12 randomized, controlled trials assessing the effects of folic-acid–based supplements, alone or with vitamin B₁₂ or B₆, on Hcy levels in a total of 1,114 subjects.¹³ Dietary folic-acid supplementation significantly (25%) reduced blood Hcy concentrations. Vitamin B₁₂ supplementation produced an additional 7% reduction, whereas vitamin B₆ did not have such an effect. It remains to be determined whether lowering Hcy levels might reduce CVD risk.

Clinical Implications
Among novel risk factors, it appears that non–HDL-C and CRP are valuable independent predictors of heart disease; the jury is still out on Lp(a) and Hcy. Clearly, more research is needed to determine not only whether these emerging risk factors are worth measuring and following, but also whether altering them affects clinical outcomes.

References