Apolipoprotein-B, Low-Density Lipoprotein Cholesterol, and the Long-Term Risk of Coronary Heart Disease in Men
2006; Elsevier BV; Volume: 97; Issue: 7 Linguagem: Inglês
10.1016/j.amjcard.2005.10.060
ISSN1879-1913
AutoresAnnie St‐Pierre, Bernard Cantin, Gilles R. Dagenais, Jean–Pierre Després, Benoı̂t Lamarche,
Tópico(s)Cancer, Lipids, and Metabolism
ResumoWe examined whether plasma apolipoprotein-B (apo-B) levels add further information on the risk of coronary heart disease (CHD) after taking into account low-density lipoprotein (LDL) cholesterol concentrations and other traditional risk factors. Among 2,072 CHD-free men from the Québec Cardiovascular Study at entry and followed for 13 years, 230 had a first CHD event (CHD death or nonfatal myocardial infarction). Increased apo-B (tertile 1 vs 3) levels were associated with a significant increased risk of CHD after adjustment for nonlipid and lipid risk factors other than LDL cholesterol levels (relative risk 1.89, 95% confidence interval 1.31 to 2.73). High plasma LDL cholesterol concentrations (tertile 1 vs 3) were also associated with an increased risk of CHD independently of nonlipid and lipid risk factors (relative risk 2.02, 95% confidence interval 1.44 to 2.84). However, apo-B levels modulated to a significant extent the risk of CHD associated with increased concentrations of LDL cholesterol (≥4.3 mmol/L). For instance, among men with high LDL cholesterol levels, those with an apo-B level <128 mg/dl were not at increased risk for CHD (relative risk 1.53, 95% confidence interval 0.89 to 2.62). In contrast, high levels of apo-B and LDL cholesterol were associated with a significant twofold increased risk of CHD (p <0.001). Receiver-operating curve analysis also indicated that plasma apo-B levels improved the ability to discriminate incident CHD cases among patients with high LDL cholesterol levels compared with a model based on LDL cholesterol levels (p = 0.04). In conclusion, plasma apo-B levels modulated the risk of CHD associated with LDL cholesterol over a 13-year follow-up. We examined whether plasma apolipoprotein-B (apo-B) levels add further information on the risk of coronary heart disease (CHD) after taking into account low-density lipoprotein (LDL) cholesterol concentrations and other traditional risk factors. Among 2,072 CHD-free men from the Québec Cardiovascular Study at entry and followed for 13 years, 230 had a first CHD event (CHD death or nonfatal myocardial infarction). Increased apo-B (tertile 1 vs 3) levels were associated with a significant increased risk of CHD after adjustment for nonlipid and lipid risk factors other than LDL cholesterol levels (relative risk 1.89, 95% confidence interval 1.31 to 2.73). High plasma LDL cholesterol concentrations (tertile 1 vs 3) were also associated with an increased risk of CHD independently of nonlipid and lipid risk factors (relative risk 2.02, 95% confidence interval 1.44 to 2.84). However, apo-B levels modulated to a significant extent the risk of CHD associated with increased concentrations of LDL cholesterol (≥4.3 mmol/L). For instance, among men with high LDL cholesterol levels, those with an apo-B level <128 mg/dl were not at increased risk for CHD (relative risk 1.53, 95% confidence interval 0.89 to 2.62). In contrast, high levels of apo-B and LDL cholesterol were associated with a significant twofold increased risk of CHD (p 20 cigarettes/day vs others), medication use (presence vs absence) at baseline, plasma log-transformed triglyceride levels, and high-density lipoprotein cholesterol were included as potential confounders. Receiver-operating characteristic curves were used to compare the additional value of apo-B and LDL cholesterol in discriminating subjects who had versus those who did not have CHD during follow-up. The resulting receiver-operating characteristic curves were compared with the likelihood ratio test and by a test developed by Hanley and McNeil.6Hanley J.A. McNeil B.J. The meaning and use of the area under a receiver operating characteristic (ROC) curve.Radiology. 1982; 143: 29-36PubMed Google Scholar, 7Hanley J.A. McNeil B.J. A method of comparing the areas under receiver operating characteristic curves derived from the same cases.Radiology. 1983; 148: 839-843PubMed Google ScholarResultsAnthropometric and metabolic characteristics of men at the 1985 baseline evaluation are presented in Table 1. Plasma LDL cholesterol levels correlated positively with plasma apo-B concentrations (r = 0.78, p <0.001).Table 1Baseline characteristics of the 1,842 men without coronary heart disease and the 230 men with a first coronary heart disease event during the 13-year follow-upVariablesCHDPercent Differencep ValueNo (n = 1,842)Yes (n = 230)Age (yrs)56.3 ± 6.959.1 ± 7.45.0%<0.001Body mass index (kg/m2)26.1 ± 3.726.8 ± 4.32.7%0.02Systolic blood pressure (mmHg)130 ± 17139 ± 196.9%<0.001Type 2 diabetes mellitus80 (4.3%)22 (9.6%)5.3%<0.001Smokers412 (22.4%)67 (29.1%)6.7%0.02Total cholesterolmg/dl221 ± 39232 ± 435.3%<0.001mmol/L5.7 ± 1.06.0 ± 1.1LDL cholesterolmg/dl151 ± 35163 ± 397.7%<0.001mmol/L3.9 ± 0.94.2 ± 1.0HDL cholesterolmg/dl40.2 ± 10.138.7 ± 9.7−3.8%0.05mmol/L1.04 ± 0.261.00 ± 0.25Cholesterol/HDL cholesterol5.8 ± 1.66.4 ± 2.110.3%<0.001Apo-B (mg/dl)116 ± 30127 ± 359.5%<0.001Triglyceridesmg/dl142 ± 133151 ± 1336.3%0.004mmol/L1.6 ± 1.51.7 ± 1.5LDL %<255Å39.5 ± 20.243.0 ± 19.23.5%0.01Values are means ± SD or numbers of patients (percentages). Plasma triglyceride levels are presented as geometric mean.HDL = high-density lipoprotein; LDL %<255Å = proportion of total LDL with a diameter <25.5 nm. Open table in a new tab Figure 1 shows the relative risk of CHD that was computed for each tertile of apo-B and LDL cholesterol using the low risk tertile as reference (relative risk 1.0). Increased plasma apo-B levels (tertile 3 vs 1) were associated with an univariate 90% increased risk of CHD (relative risk 1.91, 95% confidence interval 1.38 to 2.63, p <0.001). The association between apo-B levels and risk of CHD was not attenuated after further adjustment for nonlipid risk factors (age, body mass index, systolic blood pressure, type 2 diabetes, smoking habits, and medication use at baseline) and lipid risk factors (high-density lipoprotein cholesterol and log-transformed triglyceride levels; relative risk 1.89, 95% confidence interval 1.31 to 2.73, p <0.001). High plasma LDL cholesterol levels (tertile 3 vs 1) were also associated with a 2.0-fold greater risk of CHD after adjustment for nonlipid and lipid risk factors (relative risk 2.02, 95% confidence interval 1.44 to 2.84, p <0.001).Figure 2 shows the combined effect of concomitant variations in apo-B and LDL cholesterol levels on the relative risk of CHD. Apo-B levels modulated, to a significant extent, the risk of CHD associated with variations in LDL cholesterol concentrations. Among men with high LDL cholesterol levels (top tertile, LDL cholesterol level ≥4.3 mmol/L), patients with an apo-B level <128 mg/dl were not at increased risk for CHD (relative risk 1.53, p = 0.13). In contrast, the cumulative presence of increased apo-B (top tertile, apo-B level ≥128 mg/dl) and LDL cholesterol levels resulted in a significant 2.0-fold increase in risk of CHD (p <0.001).Figure 2Synergistic effect of LDL cholesterol (LDL-C) levels and apo-B concentrations on the 13-year relative risk of CHD. Groups with concomitant low apo-B and low LDL-C levels were used as reference. Relative risks (RRs) were adjusted for age, body mass index, systolic blood pressure, type 2 diabetes, smoking habits, and medication use at baseline.View Large Image Figure ViewerDownload (PPT)The extent to which apo-B and LDL cholesterol concentrations improved our ability to discriminate incident cases of CHD from healthy patients compared with a series of traditional risk factors was also investigated using receiver-operating characteristic curves (Table 2). The area under the receiver-operating characteristic curve based on the combination of traditional risk factors was 68.9% (model 1). Apo-B (area under the receiver-operating characteristic curve 70.3%) and LDL cholesterol (area under the receiver-operating characteristic curve 70.7%), defined as continuous variables, added discriminating power to the model of traditional risk factors (p <0.001). There was no difference in the ability to predict CHD events between the 2 models when incorporating traditional risk factors and apo-B or LDL cholesterol (p = 0.3) when computing analyses using all subjects. Removing age from these multivariate models did not alter these results. However, a model based on plasma apo-B versus LDL cholesterol levels improved the ability to discriminate incident cases of CHD from noncases in patients with high LDL cholesterol levels (areas under the receiver-operating characteristic curve 58.0% vs 53.5%, p = 0.04), which is concordant with data presented in Figure 2.Table 2Area under the receiver-operating characteristics curvesModel 1⁎Model 1 included age, body mass index, systolic blood pressure, smoking, type 2 diabetes, medication at baseline, high-density lipoprotein cholesterol, and log-transformed triglycerides.Model 2Model 3Base + Apo-BBase + LDL CholesterolAUROC‡The AUROC curves for these models were compared using a test developed by Hanley and McNeil.6,768.9%70.3% (p <0.001)†The p value reflects the incremental benefit of adding apo-B (as a continuous variable) or LDL cholesterol levels to the traditional model of risk factors (model 1) in discriminating CHD cases from noncases and was obtained using the likelihood ratio test in the logistic regression model.70.7% (p <0.001)p = 0.3†The p value reflects the incremental benefit of adding apo-B (as a continuous variable) or LDL cholesterol levels to the traditional model of risk factors (model 1) in discriminating CHD cases from noncases and was obtained using the likelihood ratio test in the logistic regression model. Model 1 included age, body mass index, systolic blood pressure, smoking, type 2 diabetes, medication at baseline, high-density lipoprotein cholesterol, and log-transformed triglycerides.† The p value reflects the incremental benefit of adding apo-B (as a continuous variable) or LDL cholesterol levels to the traditional model of risk factors (model 1) in discriminating CHD cases from noncases and was obtained using the likelihood ratio test in the logistic regression model.‡ The AUROC curves for these models were compared using a test developed by Hanley and McNeil.6Hanley J.A. McNeil B.J. The meaning and use of the area under a receiver operating characteristic (ROC) curve.Radiology. 1982; 143: 29-36PubMed Google Scholar, 7Hanley J.A. McNeil B.J. A method of comparing the areas under receiver operating characteristic curves derived from the same cases.Radiology. 1983; 148: 839-843PubMed Google Scholar Open table in a new tab As presented in Figure 3, plasma apo-B levels modulated, to a significant extent, the relative risk of CHD associated with an increased Framingham score. Among men with a high Framingham score (≥9, the top tertile in this cohort corresponds to a 10-year absolute CHD risk ≥22%), patients with apo-B levels below the third tertile (<128 mg/dl) had a 4.0-fold increased risk for CHD (relative risk 3.99, p <0.001). In contrast, the cumulative presence of increased apo-B levels and a Framingham score ≥9 resulted in a synergistic increased risk of CHD (relative risk 5.86, p <0.001). Concordant with these data, we observed that apo-B, defined as a continuous variable, added discriminating power to the use of the Framingham score, defined also as a continuous variable (areas under the receiver-operating characteristic curve 69.5% vs 69.3%, p = 0.03).Figure 3Synergistic effect of the risk associated with the Framingham risk score and apo-B concentrations on the 13-year relative risk of CHD. Groups with concomitant low apo-B levels (below the third tertile of apo-B levels) and low Framingham score (risk score <6, corresponding to the first tertile and to a 10-year CHD risk ≤11%) were used as reference. Abbreviations as in Figure 1.View Large Image Figure ViewerDownload (PPT)DiscussionAlthough plasma apo-B levels show a highly significant correlation with LDL cholesterol concentrations, we previously showed that there may be an important degree of discordance between apo-B and LDL cholesterol levels.8Sniderman A.D. St Pierre A.C. Cantin B. Dagenais G.R. Despres J.P. Lamarche B. Concordance/discordance between plasma apolipoprotein B levels and the cholesterol indexes of atherosclerotic risk.Am J Cardiol. 2003; 91: 1173-1177Abstract Full Text Full Text PDF PubMed Scopus (165) Google Scholar By examining the long-term 13-year follow-up data from the Québec Cardiovascular Study, we have significantly extended our previous investigation of the association between apo-B levels and CHD risk. Our data first indicated that high plasma apo-B and LDL cholesterol concentrations were equally and independently associated with an increased long-term relative risk of CHD. Our data also showed that plasma apo-B levels improved the ability to discriminate incident CHD cases compared with a model based on LDL cholesterol levels in patients with high LDL cholesterol levels and with a model based on the Framingham score.We have shown that high apo-B and LDL cholesterol levels (top vs low tertiles) are associated with an approximately twofold increase in the long-term relative risk of CHD after adjustment for nonlipid and lipid variables. These results agree with data from Jiang et al,9Jiang R. Schulze M.B. Li T. Rifai N. Stampfer M.J. Rimm E.B. Hu F.B. Non-HDL cholesterol and apolipoprotein B predict cardiovascular disease events among men with type 2 diabetes.Diabetes Care. 2004; 27: 1991-1997Crossref PubMed Scopus (201) Google Scholar who associated increased apo-B and LDL cholesterol levels with an almost equivalent increase in the relative risk of CHD after multivariate adjustment for age, body mass index, and other lifestyle risk factors in 746 diabetic men who were prospectively followed for 6 years. Other studies have indicated that increased apo-B levels may be associated with a greater risk of CHD compared with high LDL cholesterol concentrations.10Moss A.J. Goldstein R.E. Marder V.J. Sparks C.E. Oakes D. Greenberg H. Weiss H.J. Zareba W. Brown M.W. Liang C.S. et al.Thrombogenic factors and recurrent coronary events.Circulation. 1999; 99: 2517-2522Crossref PubMed Scopus (262) Google Scholar, 11Talmud P.J. Hawe E. Miller G.J. Humphries S.E. Nonfasting apolipoprotein B and triglyceride levels as a useful predictor of coronary heart disease risk in middle-aged UK men.Arterioscler Thromb Vasc Biol. 2002; 22: 1918-1923Crossref PubMed Scopus (214) Google Scholar Most previous studies were case versus control by design or prospective population-based studies conducted over relatively short follow-up periods, which may explain the apparent discrepancy between these studies and our results. Our data suggest that the apparent superiority of apo-B over LDL cholesterol levels may be attenuated over a longer follow-up period of observation.Most studies that reported that apo-B is a better predictor of CHD risk than other cholesterol indexes reached their conclusions based on their analyses of relative risk. Relative risk analyses and discrimination analyses based on receiver-operating characteristic curves represent distinct statistical approaches. The fundamental application of receiver-operating characteristic analyses is to measure the performance of diagnostic tests, or of a series of risk factors, in their ability to properly classify patients into groups that have or do not have a target disorder.Receiver-operating characteristic curve analysis, in the present study, showed that plasma apo-B levels improved the ability to discriminate incident CHD cases from CHD-free men compared with LDL cholesterol in men with high LDL cholesterol levels. This is partly consistent with results from Walldius et al12Walldius G. Jungner I. Holme I. Aastveit A.H. Kolar W. Steiner E. High apolipoprotein B, low apolipoprotein A-I, and improvement in the prediction of fatal myocardial infarction (AMORIS study) a prospective study.Lancet. 2001; 358: 2026-2033Abstract Full Text Full Text PDF PubMed Scopus (1044) Google Scholar who showed that apo-B had a higher sensitivity and specificity than did LDL cholesterol in predicting future coronary death in men and women in the Apolipoprotein-related Mortality Risk Study (AMORIS).12Walldius G. Jungner I. Holme I. Aastveit A.H. Kolar W. Steiner E. High apolipoprotein B, low apolipoprotein A-I, and improvement in the prediction of fatal myocardial infarction (AMORIS study) a prospective study.Lancet. 2001; 358: 2026-2033Abstract Full Text Full Text PDF PubMed Scopus (1044) Google Scholar In contrast, the 10-year Göttingen Risk Incidence and Prevalence Study (GRIPS) showed by using receiver-operating characteristic curve analyses that LDL cholesterol provides more sensitivity in predicting myocardial infarction at all given levels of specificity (or vice versa) than does apo-B in men.13Cremer P. Nagel D. Mann H. Labrot B. Muller-Berninger R. Elster H. Seidel D. Ten-year follow-up results from the Goettingen Risk, Incidence and Prevalence Study (GRIPS). I. Risk factors for myocardial infarction in a cohort of 5790 men.Atherosclerosis. 1997; 129: 221-230Abstract Full Text Full Text PDF PubMed Scopus (114) Google Scholar In the Interheart Study,14Yusuf S. Hawken S. Ounpuu S. Dans T. Avezum A. Lanas F. McQueen M. Budaj A. Pais P. Varigos J. Lisheng L. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study) case-control study.Lancet. 2004; 364: 937-952Abstract Full Text Full Text PDF PubMed Scopus (8269) Google Scholar the comparative analysis of apo-B and LDL cholesterol levels in predicting future CHD events is not likely to ever be undertaken because measurements were performed on nonfasting blood samples, thus partly invalidating the calculation of LDL cholesterol levels. The small number of studies that has investigated this issue specifically and differences in statistical procedures and in duration of follow-up may explain the discrepancies between studies. To further test the hypothesis that apo-B levels may add to the predictive value of LDL cholesterol, apo-B levels were integrated into a model of risk based on the Framingham score, which by definition included LDL cholesterol levels. Data showed that apo-B significantly increased the discriminating value of the model based on the Framingham risk score (p = 0.03). These results do not necessarily imply that apo-B should replace LDL cholesterol in the Framingham score. Rather, they suggest that patients at high risk for CHD may benefit from having a combined measurement of apo-B and cholesterol index, such as LDL cholesterol, to better estimate their long-term risk of CHD.Non–high-density lipoprotein cholesterol has been proposed as a good surrogate for apo-B levels. However, a previous report from our group has demonstrated that there is considerable discordance between apo-B and non–high-density lipoprotein cholesterol levels; we have also shown that those 2 risk factors could identify 2 distinct high-risk groups of patients.8Sniderman A.D. St Pierre A.C. Cantin B. Dagenais G.R. Despres J.P. Lamarche B. Concordance/discordance between plasma apolipoprotein B levels and the cholesterol indexes of atherosclerotic risk.Am J Cardiol. 2003; 91: 1173-1177Abstract Full Text Full Text PDF PubMed Scopus (165) Google ScholarLarge-scale trials with clinical end points have established that statin therapy decreases the risk of CHD events in primary and secondary prevention settings. However, analysis of data of the statin group in the Long-term Intervention with Pravastatin in the Ischemic Disease (LIPID) trial showed that levels of apo-B after therapy were a better predictor of CHD risk compared with LDL cholesterol.15Sniderman A.D. Furberg C.D. Keech A. Roeters van Lennep J.E. Frohlich J. Jungner I. Walldius G. Apolipoproteins versus lipids as indices of coronary risk and as targets for statin treatment.Lancet. 2003; 361: 777-780Abstract Full Text Full Text PDF PubMed Scopus (406) Google Scholar The persisting increased levels of atherogenic particles that were not captured by measurement of LDL cholesterol alone suggests that even what has been considered aggressive lipid-lowering therapy may not be aggressive enough to optimize risk decrease in some patients.16Ballantyne C.M. Achieving greater reductions in cardiovascular risk lessons from statin therapy on risk measures and risk reduction.Am Heart J. 2004; 148: S3-S8Abstract Full Text Full Text PDF PubMed Scopus (14) Google Scholar This suggests that therapy that focuses not only on LDL cholesterol levels but also on apo-B concentrations may be useful to more adequately decrease the risk of CHD in the general population. To our knowledge, only a few long-term prospective studies have investigated and compared the coronary heart disease (CHD) risk predictive value of low-density lipoprotein (LDL) cholesterol and apolipoprotein-B (apo-B) levels. The present study investigated the extent to which plasma apo-B levels may add further information to the long-term risk of CHD as assessed from LDL cholesterol levels by using data from 2,072 CHD-free men of the Québec Cardiovascular Study who were followed for 13 years. MethodsStudy population and follow-upThe Québec Cardiovascular Study has been previously described.1St-Pierre A.C. Ruel I.L. Cantin B. Dagenais G.R. Després J.P. Lamarche B. Comparison of various electrophoretic characteristics of LDL particles and their relationship to the risk of ischemic heart disease.Circulation. 2001; 104: 2295-2299Crossref PubMed Scopus (224) Google Scholar Briefly, among the 4,635 participants (35 to 64 years of age) who were evaluated in 1974, 2,552 (55.1%) had an electrocardiogram and plasma lipid and lipoprotein levels determined in 1985. Among these 2,552 patients, 102 volunteers and 265 patients with previous CHD before 1985 were excluded from the present prospective analyses. Diagnosis of diabetes was considered in men who self-reported the disease. From 1990 to 19911St-Pierre A.C. Ruel I.L. Cantin B. Dagenais G.R. Després J.P. Lamarche B. Comparison of various electrophoretic characteristics of LDL particles and their relationship to the risk of ischemic heart disease.Circulation. 2001; 104: 2295-2299Crossref PubMed Scopus (224) Google Scholar, 2Lamarche B. Despres J.P. Moorjani S. Cantin B. Dagenais G.R. Lupien P.J. Prevalence of dyslipidemic phenotypes in ischemic heart disease (prospective results from the Quebec Cardiovascular Study).Am J Cardiol. 1995; 75: 1189-1195Abstract Full Text PDF PubMed Scopus (143) Google Scholar and in 1998,3St-Pierre A.C. Cantin B. Dagenais G.R. Mauriege P. Bernard P.M. Després J.P. Lamarche B. LDL subfractions and the long-term risk of ischemic heart disease in men 13-year follow-up data from the Québec Cardiovascular Study.Arterioscler Thromb Vasc Biol. 2005; 25: 553-559Crossref PubMed Scopus (347) Google Scholar participants were contacted by mail and invited to complete a short questionnaire that provided information on smoking habits, medication use, history of cardiovascular diseases, and type 2 diabetes. For those who reported such diseases and those who died, hospital charts were reviewed by study cardiologists. For those who died outside the hospital, death certificates were obtained and the family was questioned on t
Referência(s)