Plasma Homocysteine Concentration, Statin Therapy, and the Risk of First Acute Coronary Events
Background— Elevated homocysteine levels are associated with increased coronary risk, and it has been suggested that homocysteine screening may provide a method to identify high-risk patients for aggressive primary prevention.
Methods and Results— Homocysteine was measured at baseline and after 1 year among 5569 participants in the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS), a randomized trial of lovastatin in the primary prevention of acute coronary events. The effects of homocysteine, LDL cholesterol, and lovastatin on risk were assessed over 5.2 years of trial follow-up. Median baseline homocysteine levels were significantly higher among study participants who subsequently had acute coronary events compared with those who did not (12.1 versus 10.9 μmol/L, P<0.001). The relative risks of future events from lowest (referent) to highest quartile of homocysteine were 1.0, 1.6, 1.6, and 2.2 (P<0.001). These effects were similar among those allocated to lovastatin and those allocated to placebo and were modestly attenuated after adjustment for other traditional risk factors. As predicted, the subgroup of participants with elevated LDL cholesterol and elevated homocysteine levels were at high risk and benefited greatly from statin therapy (relative risk, 0.46; 95% CI, 0.29 to 0.75; number needed to treat=26). However, in contrast to findings in this trial for C-reactive protein, homocysteine evaluation did not help to define low LDL subgroups with different responses to lovastatin therapy.
Conclusions— Although homocysteine predicted future coronary events in AFCAPS/TexCAPS, we found little evidence that homocysteine evaluation provided an improved method to target statin therapy among those with low-to-normal LDL cholesterol levels.
Received November 29, 2001; revision received February 5, 2002; accepted February 5, 2002.
Inexpensive methods to distinguish high-risk from low-risk patients may allow better targeting of statin therapy in the primary prevention of cardiovascular disease, particularly for individuals with normal-to-low levels of LDL cholesterol who may be at high risk yet are outside guidelines for therapeutic intervention.1 For example, in a recent analysis of the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS), we observed that individuals with LDL cholesterol below the study median of 149 mg/dL who had elevated levels of C-reactive protein (CRP) were not only at high risk for future vascular events but also benefited greatly from statin therapy.2 This observation is intriguing because statins have also been shown to significantly reduce plasma CRP levels.3,4⇓ However, it is unknown whether this effect is unique to CRP or whether other novel risk factors might also provide an improved method to target statin therapy in the setting of primary prevention.
In several epidemiological studies, elevated homocysteine levels have been associated with increased risk of first5–8⇓⇓⇓ as well as recurrent myocardial infarction.9–11⇓⇓ Partly on the basis of these observations, homocysteine screening also has been advocated as a method to improve detection of high-risk patients. However, no data are available addressing whether homocysteine evaluation might, like CRP, provide a method to better target statin therapy among those with low-to-normal LDL cholesterol levels.
We measured homocysteine levels at baseline and after 1 year of follow-up among 5569 participants enrolled in AFCAPS/TexCAPS, a randomized, double-blind, placebo-controlled trial of lovastatin in the primary prevention of cardiovascular events conducted among men and women with average cholesterol levels and below-average HDL levels.12 Conducted at the Lackland Air Force Base and the University of North Texas Health Science Center, AFCAPS/TexCAPS enrolled men 45 to 73 years of age and postmenopausal women 55 to 73 years of age. Individuals with uncontrolled hypertension, secondary hyperlipidemia, diabetes requiring insulin, or a body mass 50% greater than desirable were excluded.
Participants who provided written informed consent, met all entrance criteria, and completed a 12-week American Heart Association Step I diet run-in were randomly assigned to 20 mg of lovastatin per day or matching placebo. Participants in the lovastatin group were titrated in a double-blind manner to 40 mg of lovastatin per day if their LDL cholesterol levels were >110 mg/dL at the 3-month visit. Participants were followed over an average period of 5.2 years for the occurrence of first acute coronary events, prospectively defined as fatal or nonfatal myocardial infarction, unstable angina, or sudden cardiac death. It was previously reported that allocation to lovastatin, as compared with placebo, was associated with a 37% reduction in this primary clinical end point (relative risk [RR], 0.63; 95% CI, 0.50 to 0.79; P<0.001).12
The ImX Homocysteine assay (Abbott Laboratories) was used to determine total plasma homocysteine levels in blood obtained at randomization and at 1 year.13 Lipid levels were measured in a laboratory accredited through the Centers for Disease Control and Prevention Lipid Standardization Program. In total, 5569 (84%) of 6605 of the participants enrolled in AFCAPS/TexCAPS had blood available for analysis and underwent successful evaluation for both homocysteine and lipid levels. Lipid levels observed in this group are virtually identical to that observed in the AFCAPS/TexCAPS cohort as a whole.
Cox regression analysis was used to test for association between increasing levels of homocysteine at baseline and the risk of future cardiovascular events after dividing the study sample into quartiles of homocysteine. Analyses were performed for the study group as a whole and after stratification by random assignment to either lovastatin or to placebo. Adjusted risk estimates were obtained from analyses that additionally controlled for age, sex, smoking status, hypertension, parental history of coronary disease, lipid levels, and CRP, all of which have previously been shown to be independent predictors in this cohort.2,14⇓ The dose-response relation between homocysteine and coronary events was estimated by generalized additive logistic regression performed in SPLUS, a method that allows for graphical representation of the association between homocysteine concentration and the log odds of coronary risk15; because this method is sensitive to outlying values, we excluded from this analysis those individuals with baseline homocysteine levels in the upper or lower 2.5% of the distribution. The absolute and percent change in homocysteine associated with the use of lovastatin at the end of 1 year of therapy was additionally computed and compared with the absolute and percent change in homocysteine observed among those allocated to placebo.
To assess the utility of homocysteine evaluation as a potential tool to target statin therapy among those with low as compared with high lipid levels, we divided the study sample on an a priori basis into 4 groups: (1) those with LDL cholesterol above the study median (>149 mg/dL) and homocysteine above the study median (>11.27 μmol/L) (n=1380), (2) those with LDL cholesterol above the study median and homocysteine below the study median (n= 1385), (3) those with LDL cholesterol below the study median and homocysteine above the study median (n=1384), and (4) those with LDL cholesterol below the study median and homocysteine below the study median (n=1375). We then computed RR reductions associated with lovastatin as compared with placebo therapy in each of these groups as well as the number of persons needed to treat (NNT) for 5 years to avoid 1 cardiovascular event.
Baseline homocysteine values in AFCAPS/TexCAPS were similar to those observed in prior studies. Specifically, the median homocysteine level at study entry was 11.27 μmol/L, the 90th percentile cutpoint was 16.03 μmol/L, and the interquartile ranges for homocysteine were <9.34, 9.34 to 11.27, 11.28 to 13.52, and >13.52 μmol/L. As would be expected, a small number of individuals had baseline homocysteine levels >40 μmol/L, and 3 participants had values >80 μmol/L.
Median homocysteine levels at baseline were higher among the 207 participants who subsequently had a first acute coronary event during follow-up as compared with the 5362 study participants who did not (12.1 versus 10.9 μmol/L, P<0.001). Coronary event rates increased with baseline homocysteine levels such that the RRs from lowest (referent) to highest quartiles of baseline homocysteine were 1.0, 1.6, 1.6, and 2.2 (P=0.001) (Table 1). The computed odds ratios for coronary risk increased across the spectrum of homocysteine levels (Figure 1). These effects were similar in stratified analyses according to lovastatin or placebo allocation. The effect of baseline homocysteine level on vascular risk in AFCAPS/TexCAPS was attenuated but remained significant after additional adjustment for age, sex, marital status, hypertension, parental history of coronary disease, smoking status, lipid levels, and CRP levels. In fully adjusted analyses, the risk of developing acute coronary events increased 15% for each quartile increase in baseline homocysteine (95% CI, 5% to 31%).
At the end of 1 year of treatment, median homocysteine levels declined in 52% of the placebo group and in 58% of the lovastatin group. The absolute magnitude of this effect was small but somewhat greater among those allocated to lovastatin (−3.7%) as compared with those allocated to placebo (−1.8%) (Table 2).
Table 3 presents results of the efficacy analyses for lovastatin in the AFCAPS/TexCAPS trial stratified by LDL cholesterol and homocysteine values above or below the study medians. As shown, the highest-risk subgroup included those with elevated LDL and elevated homocysteine levels, a subgroup in which lovastatin was clearly effective (RR, 0.46; 95% CI, 0.29 to 0.75; NNT=26). However, as also shown, homocysteine evaluation had no impact on discriminating lovastatin efficacy among those with below-median LDL cholesterol levels, the critical group of study participants currently outside guidelines for statin therapy. Specifically, the placebo event rate in the below-median LDL/below-median homocysteine subgroup (0.033 events per 5 years) was similar to the placebo event rate in the below-median LDL/above-median homocysteine subgroup (0.038 events per 5 years). Furthermore, estimates of lovastatin efficacy in the two low LDL subgroups were similar, such that the NNT among those with increased homocysteine levels (NNT=130) was not significantly different from that observed among those with lower homocysteine levels (NNT=104) or from the low LDL participants taken as a whole (NNT=113). As shown in Figure 2, this result for homocysteine was in marked contrast to our prior observation in this trial, in which evaluation for the inflammatory biomarker CRP led to the definition of low LDL subgroups with very different event rates as well as very different efficacies associated with lovastatin therapy.
Among participants in AFCAPS/TexCAPS, baseline homocysteine levels were an independent predictor of first acute cardiovascular events. Furthermore, participants in the highest-risk subgroup defined on an a priori basis were those with elevated LDL cholesterol and elevated homocysteine levels. However, despite these observations, we found little evidence in this randomized trial setting that homocysteine evaluation provided an improved method to target statin therapy among those with below-median LDL cholesterol levels, a critical group currently outside treatment guidelines for statin therapy.
The present data for those with low-to-normal cholesterol levels contrast with recent observations made in this trial with regard to another novel biomarker of risk, CRP. As shown in Figure 2, stratification of participants with low LDL in AFCAPS/TexCAPS according to CRP levels led to a distinction between groups at high and low risk and between groups with and without efficacy of statin therapy in primary prevention. That this effect was not observed among these same study participants for homocysteine adds important information to our understanding of mechanisms of statin effect because they suggest that potential interactions between HMG CoA reductase inhibitors and CRP are not due solely to the ability of CRP to determine a high-risk subset. Rather, these null data for homocysteine provide indirect evidence that the interaction between statins and CRP has a biologically mediated mechanism, which is consistent with recent laboratory observations regarding inflammation and statins16 and with the fact that statin-induced changes in CRP are substantially greater than the statin-associated changes in homocysteine observed in the present study.3,4⇓
Our data describing a direct positive relation between homocysteine and subsequent vascular events within the AFCAPSA/TexCAPS cohort are consistent with most but not all prior studies of homocysteine and vascular risk.5–11,17⇓⇓⇓⇓⇓⇓⇓ Particular strengths of the present analysis that we believe increase its validity and generalizability include our large sample size, the use of a validated commercial assay for homocysteine assessment, and clear end point definitions and complete cohort ascertainment.
There are several possible explanations for our finding of a small reduction in homocysteine concentration in association with lovastatin use. First, because homocysteine levels declined significantly between baseline and 1 year in both the placebo and lovastatin groups, it is probable that at least part of this effect is due simply to regression to the mean. It is also possible that the reductions in homocysteine seen across the study groups reflect an overall cohort effect, perhaps caused by increased usage of folic acid–containing multivitamins during the study follow-up period. Although multivitamin use was not assessed at baseline or during follow-up in AFCAPS/TexCAPS, the likelihood of a major shift within 1 year seems unlikely, particularly because fortification of the US food supply did not occur until well after the samples were collected.18 In any event, the absolute difference in homocysteine levels associated with lovastatin use was very small and of uncertain biological significance.
In sum, the present data confirm in a large-scale population of apparently healthy American men and women that homocysteine levels are an independent predictor of future vascular risk. Indeed, participants in the AFCAPS/TexCAPS trial with elevated LDL cholesterol and elevated homocysteine levels were found to be at particularly high vascular risk and benefited greatly from lovastatin therapy. Data from this trial do not, however, support the use of homocysteine screening as a method to better target statin therapy among individuals with low-to-normal levels of LDL cholesterol, subgroups with potentially high risk who currently are outside treatment guidelines for statin therapy.1
This study was supported by grants from the National Heart, Lung, and Blood Institute (HL-58755, HL-63293) and the Leducq Foundation, Paris, France. Dr Ridker is additionally supported by an Established Investigator Award from the American Heart Association and a Doris Duke Distinguished Clinical Scientist Award from the Doris Duke Charitable Foundation.
Dr Shih is an employee of Abbott Laboratories, provider of the homocysteine assays used in this study. Dr Ridker is a coinventor on patents filed by the Brigham and Women’s Hospital that relate to inflammatory biomarkers in cardiovacular disease.
- ↵Ridker PM. Evaluating novel cardiovascular risk factors: can we better predict heart attacks? Ann Intern Med. 1999; 130: 933–937.
- ↵Ridker PM, Rifai N, Clearfield M, et al, for the Air Force/Texas Coronary Atherosclerosis Prevention Study Investigators. Measurement of C-reactive protein for the targeting of statin therapy in the primary prevention of acute coronary events. N Engl J Med. 2001; 344: 1959–1965.
- ↵Albert MA, Danielson E, Rifai N, et al, for the PRINCE Investigators. Effect of statin therapy on C-reactive protein levels: the Pravastatin Inflammation/CRP Evaluation (PRINCE): a randomized trial and cohort study. JAMA. 2001; 286: 64–70.
- ↵Ridker PM, Rifai N, Pfeffer M, et al. Long-term effects of pravastatin on plasma concentration of C-reactive protein. Circulation. 1999; 100: 230–235.
- ↵Stampfer MJ, Malinow MR, Willett WC, et al. A prospective study of plasma homocyst(e)ine and risk of myocardial infarction in US physicians. JAMA. 1992; 268: 877–881.
- ↵Graham IM, Daly LE, Refsum H, et al. Plasma homocysteine as a risk factor for vascular disease: the European Concerted Action Project. JAMA. 1997; 277: 1775–1781.
- ↵Boushey CJ, Beresford SAA, Omenn GS, et al. A quantitative assessment of plasma homocysteine as a risk factor for vascular disease. JAMA. 1995; 274: 1049–1057.
- ↵Ridker PM, Manson JE, Buring JE, et al. Homocysteine and risk of cardiovascular disease among postmenopausal women. JAMA. 1999; 281: 1817–1821.
- ↵Nygard O, Nordrehaug JE, Refsum H, et al. Plasma homocysteine levels and mortality in patients with coronary artery disease. N Engl J Med. 1997; 337: 230–236.
- ↵Stubbs PJ, Al-Obaidi MK, Conroy RM, et al. Effect of plasma homocysteine concentration on early and late events in patients with acute coronary syndromes. Circulation. 2000; 102: 605–610.
- ↵Evans RW, Shaten J, Hempel JD, et al. Homocysteine and risk of cardiovascular disease in the Multiple Risk Factor Intervention Trial. Arterioscler Thromb Vasc Biol. 1997; 17: 1947–1953.
- ↵Downs JR, Clearfield M, Weis S, et al, for the AFCAPS/TexCAPS Research Group. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: results of AFCAPS/Tex CAPS. JAMA. 1998; 279: 1615–1622.
- ↵Shipchandler MT, Moore EG. Rapid, fully automated measurement of plasma homocyst(e)ine with the Abbott IMx analyzer. Clin Chem. 1995; 41: 991–995.
- ↵Gotto AM, Whitney E, Stein EA, et al. Relation between baseline and on treatment lipid parameters and first acute major coronary events in the Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS). Circulation. 2000; 101: 477–484.
- ↵Hastie TJ, Tibshirani RJ. Generalized Additive Models. Monographs on Statistics and Applied Probability. Series 43. London: Chapman & Hall; 1990.
- ↵Rosenson RS, Tangney CC. Antiatherothrombotic properties of statins: implications for cardiovascular event reduction. JAMA. 1998; 279: 1643–1650.
- ↵Malinow MR, Bostom AG, Krauss RM. Homocyst(e)ine, diet, and cardiovascular disease. Circulation. 1999; 99: 178–182.
- ↵Malinow MR, Duell PB, Hess DL, et al. Reduction of plasma homocysteine levels by breakfast cereal fortified with folic acid in patients with coronary heart disease. N Engl J Med. 1998; 338: 1009–1015.