Improving Prediction of Ischemic Cardiovascular Disease in the General Population Using Apolipoprotein B The Copenhagen City Heart Study

Transcription

1 Improving Prediction of Ischemic Cardiovascular Disease in the General Population Using Apolipoprotein B The Copenhagen City Heart Study Marianne Benn, Børge G. Nordestgaard, Gorm Boje Jensen, Anne Tybjærg-Hansen Downloaded from by guest on November 9, 2017 Background Apolipoprotein B (apob) levels predict fatal myocardial infarction. Whether apob also predicts nonfatal ischemic cardiovascular events is unclear. We tested the following hypotheses: apob predicts ischemic cardiovascular events, and apob is a better predictor of ischemic cardiovascular events than low-density lipoprotein cholesterol. Methods and Results We studied 9231 asymptomatic women and men from the Danish general population followed prospectively for 8 years and observed the following incident events: ischemic heart disease 591, myocardial infarction 278, ischemic cerebrovascular disease 313, ischemic stroke 229, and any ischemic cardiovascular event 807. Women with apob in the upper versus the lower tertile had hazard ratios for ischemic heart disease of 1.8 (1.2 to 2.5), for myocardial infarction 2.6 (1.4 to 4.7), and for any ischemic cardiovascular event 1.8 (1.3 to 2.3), and men had hazard ratios for ischemic heart disease of 1.9 (1.5 to 2.6), for myocardial infarction 2.4 (1.5 to 3.6), and for any ischemic cardiovascular event 1.6 (1.3 to 2.1). Women had similar hazard ratios for ischemic cerebrovascular disease and ischemic stroke. ApoB had a higher predictive ability than low-density lipoprotein cholesterol in the prediction of ischemic heart disease, myocardial infarction, any ischemic cardiovascular event, and any nonfatal ischemic cardiovascular event in both genders (P 0.03 to 0.001). Finally, in smokers older than 60 years with systolic blood pressure 160 mm Hg, apob contributed 11% in women and 15% in men to the increase in absolute 10-year risk from the lower to the upper apob tertile. Conclusion ApoB predicts ischemic cardiovascular events in both genders, and is better than low-density lipoprotein cholesterol in this respect. We suggest that prediction of future ischemic cardiovascular events could be improved by measuring apob. (Arterioscler Thromb Vasc Biol. 2007;27: ) Key Words: apolipoproteins atherosclerosis cardiovascular diseases lipids lipoproteins Ischemic cardiovascular disease, that is ischemic heart disease (IHD) and ischemic cerebrovascular disease (ICVD), is one of the major causes of hospitalization and death in affluent societies. Because elevated low-density lipoprotein (LDL) cholesterol levels cause atherosclerosis and thus ischemic cardiovascular disease, levels of LDL cholesterol are used for screening to identify individuals at risk of this disease. However, the protein component of LDL particles, apolipoprotein B (apob), has also been shown to predict fatal myocardial infarction in both genders and nonfatal coronary heart disease in men. 1 7 These findings suggest that apob could be a better predictor of fatal and non-fatal cardiovascular events than. A likely explanation for this is that while is an estimate of the mass of cholesterol in the LDL fraction only, the value for apob is a measurement of the total number of atherogenic particles (including LDL, intermediate density lipoprotein, very low density lipoprotein, chylomicrons, and chylomicron remnants), because each of these contain only a single molecule of apob. Because risk of atherosclerosis appears to be more directly related to the number of circulating atherogenic particles that enter the arterial wall than to the concentrations of cholesterol in these fractions, this suggests that apob would be a better predictor of risk than the concentrations of cholesterol in the LDL fraction. Thus, whether apob is a better predictor than of IHD, myocardial infarction, ICVD, and ischemic stroke in both genders should be further explored. We tested the hypothesis that apob predicts IHD, myocardial infarction, ICVD, and ischemic stroke in women and men from the general population. Furthermore, we compared apob and as predictors of ischemic cardiovascular disease. Finally, we evaluated the absolute 10-year risk of any ischemic cardiovascular event by tertiles of apob Original received September 23, 2006; final version accepted November 30, From the Department of Clinical Biochemistry (M.B., A.T.-H.), Rigshospitalet, Copenhagen University Hospital; Department of Clinical Biochemistry (M.B., B.G.N.), Herlev University Hospital; The Copenhagen City Heart Study (B.G.N., G.B.J., A.T.-H.), Bispebjerg University Hospital, University of Copenhagen, Denmark. Correspondence to Anne Tybjærg-Hansen, MD, DMSc, Chief Physician and Associate Professor, Department of Clinical Biochemistry, KB3011, Rigshospitalet, Copenhagen University Hospital, Blegdamsvej 9, DK-2100 Copenhagen Ø, Denmark. at-h@rh.dk 2007 American Heart Association, Inc. Arterioscler Thromb Vasc Biol. is available at DOI: /01.ATV e 661

2 662 Arterioscler Thromb Vasc Biol. March 2007 Downloaded from by guest on November 9, 2017 and as well as by gender, smoking status, age, and systolic blood pressure. For these purposes we studied 9231 asymptomatic women and men from the Danish general population followed prospectively for 8 years. Subjects and Methods Participants The Copenhagen City Heart Study is a prospective cardiovascular study of the Danish general population. An almost equal number of women (55%) and men were stratified into 10-year age groups from 20 to 80 years and older; 99% were white and of Danish descent. At the third examination, 1991 to 1994, participated. In the present study, we included 9231 individuals from the 1991 to 1994 examination who had apob, total cholesterol,, high-density lipoprotein (HDL) cholesterol, and apolipoprotein A1 (apoa1) determined and who were free of ischemic cardiovascular disease at baseline. Endpoints used in the study were IHD, myocardial infarction, ICVD, ischemic stroke, or any ischemic cardiovascular event; myocardial infarction was a subgroup of IHD and ischemic stroke was a subgroup of ICVD. In participants dying within 3 months after an event, the event was regarded as fatal (99% died within 1 week). Participants were followed from baseline, defined as the date of participation in the 1991 to 1994 examination, until the occurrence of IHD, myocardial infarction, ICVD, ischemic stroke, or any ischemic cardiovascular event or end of 2000, whichever came first. Participants were followed using their unique Central Person Register number and follow-up was 100%. Information on diagnosis of IHD, myocardial infarction, ICVD, and ischemic stroke (WHO International Classification of Diseases, 8 th edition codes 410 to 414, 410, 432 to 435, and 432 to 434, respectively; and 10 th edition codes I20 to I25, I21 to I22, I63 to I64, and I63, respectively) was collected and verified until December 31, 2000 by reviewing all hospital admissions and diagnoses entered in the Danish National Hospital Discharge Register, and all causes of death entered in the Danish National Register of Causes of Death, and medical records from hospitals and general practitioners. IHD was determined on the basis of previous myocardial infarction or characteristic symptoms of stable or unstable angina pectoris based on location, character, and duration of pain, and the relation of pain to exercise. A diagnosis of myocardial infarction required the presence of at least 2 of the following criteria: characteristic chest pain, elevated cardiac enzymes, or electrocardiographic changes indicative of myocardial infarction. An experienced neurologist reviewed all potential cases of ICVD, that is, ischemic stroke, transient ischemic attack (focal neurological symptoms lasting 24 hour), or amaurosis fugax (transient blindness on one eye only). Possible ischemic stroke events were validated using the WHO definition of stroke: an acute disturbance of focal or global cerebral function with symptoms lasting 24 hour or leading to death with presumably no other reason than of vascular origin. To distinguish between infarction, intracerebral hemorrhages and subarachnoid hemorrhages, either a CT or an MRI scan, autopsy, spinal fluid examination, or surgical description was necessary. Subjects receiving lipid-lowering medication at baseline were excluded (n 92). Hypertension was defined as use of antihypertensive medication, a systolic blood pressure 140 mm Hg, or a diastolic blood pressure 90 mm Hg at inclusion into the study. Diabetes mellitus was defined as self-reported disease or a nonfasting plasma glucose 11.0 mmol/l at baseline. Smokers were active smokers. The study was approved by institutional review boards and a Danish ethical committee (No /91, Copenhagen and Frederiksberg committee), and was conducted according to the Declaration of Helsinki. Participants gave written informed consent. Analyses Turbidimetric and colorimetric assays were used on fresh samples to measure plasma levels of apob and apoa1 (nonstandardized assays), total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides. 8 was calculated when plasma triglycerides were 5 mmol/l (97.4% of individuals) using the equation of Friedewald. Non-HDL cholesterol was total cholesterol HDL cholesterol. Statistical Analysis Data were analyzed using SPSS and Stata. Two-sided P 0.05 was considered significant. Student t-test or Pearson 2 test was used in 2-group comparisons. Gender- and age-specific tertiles of apob, total cholesterol,, apob/apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol were constructed. Pearson product-moment correlation coefficient, r, was used to assess the relationship between apob and. Cumulative incidences were plotted using Kaplan-Meier curves and differences between tertiles of apob examined using log-rank tests. Cox proportional hazards regression models with age as a covariate, with age and as covariates, or multifactorially adjusted for age, HDL cholesterol, triglycerides, body mass index, hypertension, diabetes mellitus, and smoking status, were used to estimate hazard ratios for ischemic cardiovascular events by tertiles of apob, total cholesterol,, apob/apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol. In addition, Cox regression was performed with age as time scale, which means that differences in age are automatically adjusted for, as well as with gender-specific tertiles of apob, total cholesterol,, apob/apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol within 10-year age groups. The predictive ability of apob, total cholesterol, apob/apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol in the prediction of ischemic cardiovascular events was estimated as the area under the curve from receiver operating characteristics curves. Equality of the area under the curves on the receiver operating characteristics plot was tested using the algorithm suggested by DeLong. The estimated absolute risk of any ischemic cardiovascular event (IHD and ICVD combined) was calculated using the regression coefficients from a Poisson regression model by tertiles of apob and, respectively, age in 3 groups ( 40, 40 to 60, and 60 years), smoking status (nonsmoker/smoker), and systolic blood pressure in 3 groups ( 140, 140 to 160, and 160 mm Hg) for women and men separately. Absolute risks are presented as estimated incidence rates (number of events/10 years) in percent. To compare and apob for use in a risk prediction score, all individuals in the population who later developed IHD were first assigned a baseline 10-year risk value using the Framingham coronary heart disease risk prediction score sheet based LDL cholesterol levels. 9 This score sheet assigns points for age, LDL cholesterol, HDL cholesterol, blood pressure, diabetes, and smoking status. The points are added up for each individual and converted into a 10-year risk estimate. Subsequently, the same individuals were scored now using exactly the same score system, but substituting the score with an apob score. To compare the 2 tests, values of the cut points used in the Framingham risk prediction score sheet were converted to a percentile value in the Copenhagen City Heart Study population and the corresponding value for apob was used as the apob cut point (group 1: LDL 2.59 mmol/l percentile apob 63 mg/dl; group 2: LDL 2.60 to 3.36 mmol/l to percentile apob 64 to 78 mg/dl; group 3: LDL 3.37 to 4.14 mmol/l to percentile apob 79 to 94 mg/dl; group 4: LDL 4.15 to 4.91 mmol/ L to percentile apob 95 to 110 mg/dl; group 5: LDL 4.92 mmol/l percentile apob 111 mg/dl). Results Baseline characteristics of individuals from the general population by diagnostic status at the end of follow-up are shown (supplemental Table I, available online at org). During 8 years of follow-up ( person-years), we observed the following number of incident events: IHD, 591;

3 Benn et al ApoB and Ischemic Cardiovascular Disease 663 Downloaded from by guest on November 9, 2017 Figure 1. Cumulative incidence of ischemic heart disease, myocardial infarction, ischemic cerebrovascular disease, ischemic stroke, and any ischemic cardiovascular event as a function of years of follow-up (left panel) and age (age-adjusted) (right panel), and by tertiles of apolipoprotein B (apob) in women and men from the general population. 1 (green) lower tertile of apob: women 0 to 74 mg/dl; men 0 to 75 mg/dl. 2 (orange) middle tertile of apob: women 75 to 94 mg/dl; men 76 to 94 mg/dl. 3 (red) upper tertile of apob: women 95 to 242 mg/dl; men 95 to 204 mg/dl. myocardial infarction, 278; ICVD, 313; ischemic stroke, 229; and any ischemic cardiovascular event, 807. Of the latter, 25% in women and 28% in men were fatal ischemic cardiovascular events. ApoB levels increased as a function of age in 10-year age groups for both women and men (supplemental Figure I). Levels of apob were positively correlated with those of (women: r , P 0.000; men: r , P 0.000). Risk of Ischemic Cardiovascular Disease by Tertiles of ApoB Plots of the cumulative incidence of IHD, myocardial infarction, ICVD, and ischemic stroke, and any ischemic cardiovascular event as a function of follow-up time and apob in tertiles showed a stepwise increase in cumulative incidence from the lower tertile through the middle to the upper tertile for all end-points except for ICVD and ischemic stroke in men (Figure 1, left panel). In the latter 2 groups, the cumulative incidence for the lower and middle tertiles of apob were similar, whereas in the upper tertile the cumulative incidence appeared to be increased (ICVD: P 0.01; ischemic stroke: P 0.10). At the age of 70 years, on Kaplan-Meier curves using age as time scale, the difference in cumulative incidence of any ischemic cardiovascular event between the lower and upper tertile of apob was 10% in both genders (Figure 1, right panel). Plots of the cumulative incidence of any ischemic cardiovascular event as a function of follow-up time and apob in tertiles, and stratified on gender and age 60 or 60 years, also showed an increased cumulative incidence in the upper tertile for both age groups; however, the cumulative incidence in the middle tertile was not significantly increased compared with the lower tertile in individuals older than 60 years (supplemental Figure II).

4 664 Arterioscler Thromb Vasc Biol. March 2007 TABLE 1. Risk of Ischemic Heart Disease, Myocardial Infarction, Ischemic Cerebrovascular Disease, Ischemic Stroke, and Any Ischemic Cardiovascular Event by Tertiles of ApoB During 8 Years of Follow-up in the General Population Women Men Downloaded from by guest on November 9, 2017 ApoB (mg/dl) Ischemic heart disease Incidence per Person-Years Age-adjusted hazard ratio (95%CI) Age- and LDL Cholesterol- Adjusted Hazard Ratio (95% CI) By Cox regression adjusted for age only, women and men with apob in the upper versus lower tertile had an increased risk of IHD, myocardial infarction, and any ischemic cardiovascular event with hazard ratios ranging from 1.6 (95% CI:1.3 to 2.1) for any ischemic cardiovascular event in men to 2.6 (1.4 to 4.7) for myocardial infarction in women (Table 1). These hazard ratios were similar, after multifactorial adjustment for total cholesterol,, HDL cholesterol, triglycerides, body mass index, hypertension, diabetes mellitus, and smoking. Women, but not men, with apob in the upper tertile also had an increased risk of ICVD and ischemic stroke, which became more pronounced after adjustment for other cardiovascular risk factors, with hazard ratios of 2.4 (1.3 to 4.4/4.5) for both ICVD and ischemic stroke (Table 1). Cox regression using age as time scale, or using apob tertiles determined within 10-year age groups in each gender, gave similar results (data not shown). Comparison of ApoB and LDL Cholesterol as Predictors of Ischemic Cardiovascular Events In women, the upper tertiles of apob, apob/apoa1, and total cholesterol/hdl cholesterol versus the lower tertiles were associated with an increased risk of both IHD and myocardial infarction, whereas and non-hdl cholesterol were associated with an increased risk of myocardial infarction only (Figure 2). In men, the upper tertiles of apob, Multi-factorially Adjusted Hazard Ratio* (95% CI) ApoB (mg/dl) Incidence per Person-Years Age-adjusted Hazard Ratio (95% CI) Age- and LDL Cholesteroladjusted Hazard Ratio (95% CI) Multi-factorially Adjusted Hazard Ratio* (95% CI) Lower tertile Middle tertile ( ) 1.4 ( ) 1.1 ( ) ( ) 1.1 ( ) 1.1 ( ) Upper tertile ( ) 2.2 ( ) 1.4 ( ) ( ) 1.8 ( ) 1.7 ( ) Myocardial infarction Lower tertile Middle tertile ( ) 1.3 ( ) 1.3 ( ) ( ) 1.6 ( ) 1.4 ( ) Upper tertile ( ) 2.0 ( ) 2.0 ( ) ( ) 2.9 ( ) 1.9 ( ) Ischemic cerebrovascular disease Lower tertile Middle tertile ( ) 1.0 ( ) 1.4 ( ) ( ) 0.9 ( ) 0.8 ( ) Upper tertile ( ) 1.6 ( ) 2.4 ( ) ( ) 1.5 ( ) 1.2 ( ) Ischemic stroke Lower tertile Middle tertile ( ) 0.9 ( ) 1.4 ( ) ( ) 1.0 ( ) 0.8 ( ) Upper tertile ( ) 1.3 ( ) 2.4 ( ) ( ) 1.6 ( ) 1.2 ( ) Any ischemic cardiovascular event Lower tertile Middle tertile ( ) 1.2 ( ) 1.1 ( ) ( ) 1.0 ( ) 1.0 ( ) Upper tertile ( ) 1.8 ( ) 1.5 ( ) ( ) 1.6 ( ) 1.4 ( ) *Adjusted for age, total cholesterol,, high density lipoprotein cholesterol, triglycerides, body mass index, hypertension, diabetes mellitus, and smoking status by Cox regression. total cholesterol,, apob/apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol were all associated with increased risks of both IHD and myocardial infarction with the largest hazard ratios for apob. In women only, the upper tertiles of apob, total cholesterol, LDL cholesterol, and non-hdl cholesterol were also associated with an increased risk of ICVD and ischemic stroke (borderline for and non-hdl cholesterol). Finally, the upper tertiles of all 6 predictors were associated with increased risks of any ischemic cardiovascular event in both genders (Figure 2). When distinguishing between nonfatal and fatal ischemic cardiovascular events, only the upper tertile of apob versus the lower tertile predicted an increased risk of nonfatal events in both genders, whereas the upper tertiles of apob, total cholesterol,, apob/ apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol predicted risk of fatal events in men only (data not shown). ApoB had a higher predictive ability than in the prediction of IHD, myocardial infarction, any ischemic cardiovascular event, and any nonfatal ischemic cardiovascular event in both genders (Table 2), and similar trends were found for apob versus total cholesterol. ApoB was also a more reliable risk predictor than apob/apoa1, total cholesterol/hdl cholesterol in the prediction of ICVD and ische-

5 Benn et al ApoB and Ischemic Cardiovascular Disease 665 Downloaded from by guest on November 9, 2017 Figure 2. Risk of ischemic heart disease, myocardial infarction, ischemic cerebrovascular disease, ischemic stroke, and any ischemic cardiovascular event by tertiles of apob, total cholesterol,, apob/apolipoprotein A1, total cholesterol/hdl cholesterol, and non-hdl cholesterol in women and men from the general population. CI indicates confidence interval.

6 666 Arterioscler Thromb Vasc Biol. March 2007 TABLE 2. Predictive Ability of ApoB, Total Cholesterol, LDL Cholesterol, ApoB/ApoA1, Total Cholesterol/HDL Cholesterol, and Non-HDL Cholesterol in the Prediction of Ischemic Cardiovascular Disease in the General Population Using Receiver Operating Characteristics Curves Ischemic Heart Disease Myocardial Infarction Ischemic Cerebrovascular Disease Ischemic Stroke Any Ischemic Cardiovascular Event Any Ischemic Cardiovascular Event Nonfatal Any Ischemic Cardiovascular Event Fatal AUC (95%CI) P AUC (95%CI) P AUC (95%CI) P AUC (95%CI) P AUC (95%CI) P AUC (95%CI) P AUC (95%CI) P Downloaded from by guest on November 9, 2017 Women ApoB 0.66 ( 0.69) 0.70 ( 0.75) ( ) ( 0.70) 0.66 ( 0.69) 0.66 ( 0.69) (0.54 ) Total cholesterol 0.64 ( 0.67) ( ) ( 0.71) ( ) 0.93 ( ) 0.20 ( ) 0.26 ( ) 0.89 ( 0.67) ( ) ( 0.68) 0.29 ( 0.68) ( ) ( ) 0.01 ( ) 0.47 ApoB/apoA1 ( ) ( ) 0.75 ( 0.64) ( 0.64) ( ) ( 0.67) 0.02 ( ) 0.74 Total cholesterol/hdl cholesterol 0.64 ( ) 0.67 ( 0.72) 0.29 (0.54 ) (0.54 ) ( 0.66) 0.04 ( 0.66) 0.02 ( ) 0.97 Non-HDL cholesterol ( ) ( ) 0.44 ( ) 0.43 ( 0.69) ( 0.68) 0.54 ( ) 0.34 ( ) 0.45 Men ApoB 0.62 ( ) 0.62 ( 0.66) 0.56 ( ) ( ) ( ) 0.61 ( 0.64) ( ) Total cholesterol ( ) 0.07 ( ) 0.19 ( ) 0.85 ( ) 0.68 ( ) 0.15 ( 0.62) 0.15 ( ) 0.82 (0.56 ) 0.01 ( ) 0.03 (0.50 ) (0.48 ) 0.17 ( 0.61) (0.56 ) ( ) ApoB/apoA1 ( 0.64) ( 0.66) ( )* ( )* 0.96 ( 0.62) 0.49 ( ) 0.27 (0.54 ) 0.75 Total cholesterol/hdl cholesterol ( ) ( 0.66) ( ) ( ) 0.86 ( ) 0.16 ( ) 0.18 (0.52 ) 0.75 Non-HDL cholesterol 0.61 ( 0.64) ( 0.66) 0.73 ( ) ( )* 0.91 ( ) ( 0.64) 0.56 ( ) 0.72 AUC indicates area under the curve from receiver operating characteristics curves; CI, confidence interval; LDL, low density lipoprotein. An AUC 0.5 indicates increased predictive ability (null hypothesis: true area 0.5). P 0.001, P 0.01, and *P 0.05 testing the predictive ability of apob, total cholesterol, or, respectively, against the null hypothesis. P for the comparison of AUC between apob and total cholesterol,, apob/apoa1, total cholesterol/hdl cholesterol, and non-hdl cholesterol. mic stroke in women, and of any ischemic cardiovascular event in both genders (Table 2). Comparing the predictive ability of apob between strata of age, level, body mass index, hypertension, diabetes, and smoking showed that apob was a better predictor in both women and men younger than 60 years compared with those older than 60 (P and P 0.01, respectively), and in women with a body mass index 25 kg/m 2 (P 0.001) and without hypertension (P 0.006; supplemental Table II). There was no interaction between levels of LDL cholesterol and apob in the prediction of any ischemic cardiovascular event in either gender, suggesting that level of did not affect the predictive value of apob as reported by others 1 (supplemental Table II). Maximizing both sensitivity and specificity for prediction of any ischemic cardiovascular event, identified a cut-point for levels of apob of 90 mg/dl (sensitivity and specificity both 61%) and a cut-point for levels of of 3.9 mmol/l (sensitivity and specificity both 60%) (Figure 3). Absolute 10-Year Risk of Any Ischemic Cardiovascular Event The lowest absolute 10-year risks of any ischemic cardiovascular event were 0.3% and 0.6%, respectively, in nonsmoking women and men younger than 40 years of age with normal blood pressure and in the lowest apob tertile (Figure 3). Absolute risks increased with smoking, increasing age, blood pressure, and apob levels. The highest absolute 10-year risks of any ischemic cardiovascular event were 36% and 55% in smoking women and men older than 60 years of age, with a systolic blood pressure 160 mm Hg, and an apob level in the upper tertile; incorporating instead of apob, the equivalent highest absolute 10-year risks were 34% and 53% (Figure 3, lower right panel in both genders). Within this group of individuals (smokers, 60 years, and systolic blood pressure 160 mm Hg), apob contributed 11% and 15% in women and men, respectively, to the increase in 10-year risk between the upper and lower tertiles of apob. ApoB Versus LDL Cholesterol by Framingham Coronary Heart Disease Risk Prediction Score Scoring all individuals who later developed ischemic heart disease using the Framingham coronary heart disease risk prediction score sheet with either or apob, showed similar risk scores in 83% of women and 68% of men (Figure 4). However, in 10% of women and 20% of men apob assigned a higher risk score than, whereas assigned a higher score than apob in 7% of women and 12% of men. Individuals in whom apob assigned a higher risk score than had higher levels of triglycerides and body mass index and reduced LDL cholesterol (supplemental Table III). Discussion In this study of the general population, upper versus lower tertile of apob predicted an increased risk of IHD, myocardial infarction, and any ischemic cardiovascular event in both women and men, as well as an increased risk of ICVD and ischemic stroke in women. These risk estimates only changed slightly when adjusting for, indicating that apob had predictive value beyond. Furthermore, apob had a higher ability than in predicting IHD, myocardial infarction, and any ischemic cardiovascular event in both genders. Because of the overlap in levels of both apob and in individuals with and without ischemic cardiovascular events, it was not possible to select clinically relevant cut-points to separate between these 2 diagnostic groups using levels of apob or

7 Benn et al ApoB and Ischemic Cardiovascular Disease 667 Downloaded from by guest on November 9, 2017 Figure 3. Absolute 10-year risk of any ischemic cardiovascular event by smoking status, age, systolic blood pressure, and tertiles of apob (apob) (2 left panels) and (2 right panels) for women and men, separately. Left panels: 1 (green) lower tertile of apob: women 0 to 74 mg/dl; men 0 to 75 mg/dl. 2 (orange) middle tertile of apob: women 75 to 94 mg/dl; men 76 to 94 mg/dl. 3 (red) upper tertile of apob: women 95 to 242 mg/dl; men 95 to 204 mg/dl. Right panels: 1 (green) lower tertile of : women 0 to 3.1 mmol/l; men 0 to 3.1 mmol/l. 2 (orange) middle tertile of : women 3.2 to 4.2 mmol/l; men 3.2 to 4.0 mmol/l. 3 (red) upper tertile of : women 4.3 to 12.6 mmol/l; men 4.1 to 8.7 mmol/l.

8 668 Arterioscler Thromb Vasc Biol. March 2007 Downloaded from by guest on November 9, 2017 Figure 4. ApoB vs using the Framingham coronary heart disease risk prediction score sheet. Number of individuals who later developed IHD with different baseline 10-year risk scores using and apob are tabulated against each other on, respectively, the vertical and horizontal axes. alone, emphasizing that apob and LDL cholesterol values alone should not be used as risk predictors in the general population, but should be combined with information on other risk factors to enable risk stratification. Several case-control studies have shown that levels of apob are more strongly associated with the presence of IHD than levels of total or. 10,11 Among prospective studies, the AMORIS study demonstrated that apob was a predictor of fatal myocardial infarction even at low LDL cholesterol levels, whereas alone was a marginally significant predictor in men, but not in women. 1 Similar results were reported the Quebec Cardiovascular Study using coronary events as end-point, 2 in The Northwick Park Heart Study, 3 for apob compared with non-hdl cholesterol in prediction of ischemic cardiovascular disease in the Health Professionals Follow-up Study, 4 for apob alone and apob/apoa1 ratio in prediction of coronary heart disease in the Nurses Health Study, 5 and for apob as a predictor of risk of recurrent coronary events. 6 In the Women s Health Study apob/apoa1 ratio was shown to be only marginally

9 Benn et al ApoB and Ischemic Cardiovascular Disease 669 Downloaded from by guest on November 9, 2017 better than total cholesterol/hdl cholesterol ratio in prediction of cardiovascular. 7 Novel findings in the present study of a general population sample include that apob level is a predictor of an increased risk of: (1) ischemic heart disease and myocardial infarction not only in men, but also in women; (2) ICVD and ischemic stroke in women; (3) any ischemic cardiovascular event and any nonfatal ischemic cardiovascular event in both genders; and (4) apob has a higher predictive ability than in the prediction of IHD, myocardial infarction, any ischemic cardiovascular event and any nonfatal ischemic cardiovascular event in both women and men. Taken together, this suggests that prediction of future ischemic cardiovascular events in the general population could be improved in both women and men by measuring apob. It could be argued that measurement of non-hdl cholesterol could be equally good at predicting ischemic cardiovascular disease compared with apob, because non-hdl cholesterol measure the cholesterol content of all apobcontaining lipoproteins. In accordance with this idea, we found that the predictive ability using receiver operating characteristics curves for non-hdl cholesterol was similar to that for apob for all ischemic cardiovascular disease endpoints studied in both women and men. From a methodological point of view, the measurement of apob has been standardized by the International Federation of Clinical Chemistry, 12 it has been automated, and fasting samples are not required. The interindividual biological coefficient of variation is lower for apob than for LDL cholesterol (CV within: 6.9% versus 8.3%; westgard.com/biodatabase1.htm), making repeated measurements of apob more reliable in the single patient. Measurement of non-hdl cholesterol is even less reliable, because this measurement depends on variation in measurement of both total cholesterol (CV within: 6.0%) and HDL cholesterol (7.1%). Although there is abundant evidence that the risk of both coronary and peripheral atherosclerotic cardiovascular disease is directly related to plasma cholesterol levels, the relationship between levels of apob, total cholesterol, and and ICVD and ischemic stroke is not so well-established ; however, recent data from the AMORIS study have shown an association of increased apob/apoa1 ratio and increased risk of fatal ischemic stroke in both genders. 16 In this study we demonstrate that the upper tertiles compared with the lower tertiles of apob, total cholesterol,, and non-hdl cholesterol are all associated with risk of ICVD and apob and total cholesterol with ischemic stroke in women, but not in men. Furthermore, apob, total cholesterol,, and non-hdl cholesterol all have significant ability in the prediction of ICVD and ischemic stroke in women. Information on the absolute 10-year risk of any cardiovascular event offers a tool useful in the primary prevention of such events, and can be used to identify patients with asymptomatic or subclinical disease who could potentially benefit from intensive primary prevention efforts. Neither the American Heart Association s recommendations for risk assessment in asymptomatic individuals based on the Framingham Study, 9 nor the recommendations of the First Joint Task Force of the European Societies on Coronary Prevention based on the SCORE project 17 have included apob in their risk assessments. The data from the AMORIS study, 1 The Quebec Cardiovascular Study, 2 The Northwick Park Heart Study, 3 the Women s Health Study, 7 the Nurses Health Study, 5 and the present study suggests that apob can contribute further information to better-define cardiovascular disease risk, not only in men but also in women. A likely explanation for this is that while is an estimate of the mass of cholesterol in the LDL fraction only, the value for apob is a measurement of the total number of atherogenic particles in plasma, including LDL, intermediate density lipoprotein, very-low-density lipoprotein particles, chylomicrons, and chylomicron remnants. A limitation of our study is that apob was measured at baseline only and not by a standardized assay. It would be interesting to determine whether repeated measures of apob values over time, ie, as area under curves, could contribute even more information on cardiovascular disease risk. Another limitation is that our risk estimates are obtained from a high-risk white population. Using the estimates of absolute 10-year risks in the individual is therefore limited to areas of the world with similar prevalence of risk factors and underlying rates of cardiovascular events. A minor drawback is that we only have information about use of statins at baseline and not during follow-up. Because both apob and are evaluated in the same population and statin treatment would approximately reduce both to the same degree, this would not alter the main conclusion of the article. In conclusion, apob predicts ischemic cardiovascular events in both genders, and is better than in this respect. For this reason, we suggest that prediction of future ischemic cardiovascular events could be improved by measuring apob. Sources of Funding This work was supported by The Danish Heart Foundation, The Danish Medical Research Council, Chief Physician Johan Boserup and Lise Boserup s Fund, Ingeborg and Leo Dannin s Grant, and the Research Fund at Rigshospitalet, Copenhagen University Hospital. None. Disclosures References 1. Walldius G, Jungner I, Holme I, Aastveit AH, 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: Lamarche B, Moorjani S, Lupien PJ, Cantin B, Bernard PM, Dagenais GR, Despres JP. Apolipoprotein A-I and B levels and the risk of ischemic heart disease during a five-year follow-up of men in the Quebec cardiovascular study. Circulation. 1996;94: Talmud PJ, Hawe E, Miller GJ, Humphries SE. Nonfasting apolipoprotein B and triglyceride levels as a useful predictor of coronary heart disease risk in middle-aged UK men. Arterioscler Thromb Vasc Biology. 2002;22: Pischon T, Girman CJ, Sacks FM, Rifai N, Stampfer MJ, Rimm EB. Non-high-density lipoprotein cholesterol and apolipoprotein B in the prediction of coronary heart disease in men. Circulation. 2005;112: Shai I, Rimm EB, Hankinson SE, Curhan G, Manson JE, Rifai N, Stampfer MJ, Ma J. Multivariate Assessment of Lipid Parameters as

10 670 Arterioscler Thromb Vasc Biol. March 2007 Downloaded from by guest on November 9, 2017 Predictors of Coronary Heart Disease Among Postmenopausal Women: Potential Implications for Clinical Guidelines. Circulation. 2004;110: Moss AJ, Goldstein RE, Marder VJ, Sparks CE, Oakes D, Greenberg H, Weiss HJ, Zareba W, Brown MW, Liang CS, Lichstein E, Little WC, Gillespie JA, Van Voorhees L, Krone RJ, Bodenheimer MM, Hochman J, Dwyer EM, Arora R, Marcus FI, Watelet LFM, Case RB. Thrombogenic factors and recurrent coronary events. Circulation. 1999;99: Ridker PM, Rifai N, Cook NR, Bradwin G, Buring JE. Non-HDL Cholesterol, Apolipoproteins A-I and B100, Standard Lipid Measures, Lipid Ratios, and CRP as Risk Factors for Cardiovascular Disease in Women. JAMA. 2005;294: Tybjærg-Hansen A, Steffensen R, Meinertz H, Schnohr P, Nordestgaard BG. Association of mutations in the apolipoprotein B gene with hypercholesterolemia and the risk of ischemic heart disease. N Engl J Med. 1998;338: Wilson PWF, D Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97: Sniderman AD, Furberg CD, Keech A, van Lennep JR, Frohlich J, Jungner I, Walldius G. Apolipoproteins versus lipids as indices of coronary risk and as targets for statin treatment. Lancet. 2003;361: Rader DJ, Hoeg JM, Brewer HB. Quantitation of Plasma Apolipoproteins in the Primary and Secondary Prevention of Coronary-Artery Disease. Ann Int Medicine. 1994;120: Marcovina SM, Albers JJ, Dati F, Ledue TB, Ritchie RF. International Federation of Clinical Chemistry standardization project for measurements of apolipoproteins A-I and B. Clin Chem. 1991;37: Shahar E, Chambless LE, Rosamond WD, Boland LL, Ballantyne CM, McGovern PG, Sharrett AR. Plasma lipid profile and incident ischemic stroke: the Atherosclerosis Risk in Communities (ARIC) study. Stroke. 2003;34: Qureshi AI, Giles WH, Croft JB, Guterman LR, Hopkins LN. Apolipoproteins A-1 and B and the likelihood of non-fatal stroke and myocardial infarction data from The Third National Health and Nutrition Examination Survey. Med Sci Monit. 2002;8:CR311 CR Chien KL, Sung FC, Hsu HC, Su TC, Lin RS, Lee YT. Apolipoprotein A-I and B and stroke events in a community-based cohort in Taiwan: report of the Chin-Shan Community Cardiovascular Study. Stroke. 2002; 33: Walldius G, Aastveit AH, Jungner I. Stroke mortality and the apob/ apoa-i ratio: results of the AMORIS prospective study. J Intern Med. 2006;259: Pyorala K, Debacker G, Graham I, Poolewilson P, Wood D. Prevention of Coronary Heart-Disease in Clinical-Practice - Recommendations of the Task-Force of the European-Society of Cardiology, European Atherosclerosis Society and European-Society of Hypertension. Eur Heart J. 1994; 15:

11 Downloaded from by guest on November 9, 2017 Improving Prediction of Ischemic Cardiovascular Disease in the General Population Using Apolipoprotein B: The Copenhagen City Heart Study Marianne Benn, Børge G. Nordestgaard, Gorm Boje Jensen and Anne Tybjærg-Hansen Arterioscler Thromb Vasc Biol. 2007;27: ; originally published online December 14, 2006; doi: /01.ATV e Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2006 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online at:

12 Supplementary Table I. Baseline characteristics of participants from the general population by diagnostic status at the end of follow-up. No event Ischemic heart disease Myocardial infarction Ischemic cerebrovascular disease Ischemic stroke Any ischemic cardiovascular event Women Number of individuals Fatal events (n(%)) 0(0%) 74(28%) 34(32%) 34(22%) 28(24%) 94(25%) Age (years) 57.2± ±0.5*** 69.5±0.8*** 71.3±0.7*** 71.9±0.7*** 70.4±0.4*** Apolipoprotein B (mg/dl) 84.8± ±1.4*** 104.2±2.3*** 97.7±1.7*** 97.7±2.0*** 98.3±1.2*** Total cholesterol (mmol/l) 6.2± ±0.1*** 7.2±0.1*** 7.0±0.1*** 7.0±0.1*** 6.9±0.1*** (mmol/l) 3.7± ±0.1*** 4.6±0.1*** 4.4±0.1*** 4.4±0.1*** 4.4±0.1*** HDL cholesterol (mmol/l) 1.7± ±0.0** 1.6±0.0** 1.8± ± ±0.0* Apolipoprotein B/apolipoprotein A1 ± ±0.0*** 0.74±0.0*** ±0.0** 0.64±0.0*** 0.68±0.0*** Total cholesterol/hdl cholesterol 3.9± ±0.1*** 4.9±0.2*** 4.2±0.1** 4.2±0.1** 4.5±0.1*** Non-HDL cholesterol (mmol/l) 4.5± ±0.1*** 5.6±0.1*** 5.2±0.1*** 5.0±0.1*** 5.2±0.1*** Triglycerides (mmol/l) 1.6± ±0.1*** 2.2±0.1*** 1.8±0.1* 1.9±0.1*** 2.0±0.1*** Body mass index (kg/m 2 ) 25.1± ±0.3*** 26.7±0.5*** 26.0±0.4* 25.7± ±0.2*** Hypertension (n(%)) 2018(46%) 198(75%)*** 82(77%)*** 133(85%)*** 102(87%)*** 289(77%)*** Diabetes mellitus (n(%)) 102(2.3%) 21(7.9%)*** 9(8.9%)*** 12(7.7%)*** 10(8.3%)*** 26(6.9%)*** Smokers (n(%)) 1997(46%) 127(48%) 60(57%)*** 75(48%) 56(48%) 180(48%)

13 Men Number of individuals Fatal events (n(%)) 0(0%) 101(31%) 60(35%) 30(19%) 20(18%) 121(28%) Age (years) 54.4± ±0.5*** 67.2±0.7*** 68.5±0.8*** 69.1±0.8*** 67.6±0.5*** Apolipoprotein B (mg/dl) 84.7± ±1.2*** 94.2±1.5*** 90.7±1.8** 90.3±1.9** 92.5±1.0*** Total cholesterol (mmol/l) 5.9± ±0.1*** 6.3±0.1*** 6.1±0.1** 6.1± ±0.1*** (mmol/l) 3.6± ±0.1*** 4.0±0.1*** 3.8±0.1* 3.8± ±0.1*** HDL cholesterol (mmol/l) 1.4± ±0.0* 1.3±0.0** 1.4± ± ±0.0 Apolipoprotein B/apolipoprotein A1 0.68± ±0.0*** 0.77±0.0*** 0.73± ±0.0** 0.75±0.0*** Total cholesterol/hdl cholesterol 4.6± ±0.1*** 5.3±0.1*** 4.9± ± ±0.1*** Non-HDL cholesterol (mmol/l) 4.5± ±0.1*** 5.0±0.1*** 4.8±0.1* 4.7±0.1* 4.8±0.1*** Triglycerides (mmol/l) 2.1± ±0.1* 2.5±0.2*** 2.1± ± ±0.1* Body mass index (kg/m 2 ) 25.9± ±0.2*** 26.9±0.3** 26.5± ± ±0.2*** Hypertension (n(%)) 1730(54%) 252(77%)*** 136(79%)*** 122(78%)*** 87(78%)*** 328(76%)*** Diabetes mellitus (n(%)) 144(4.5%) 29(8.8%)*** 18(10.2%)*** 19(12.1%)*** 12(10.5%)*** 47(10.8%)*** Smokers (n(%)) 1663(52%) 173(53%) 112(65%)*** 82(52%) 59(53%) 229(53%) Continuous variables are mean±standard error of the mean. LDL= low density lipoprotein. HDL= high density lipoprotein. Ischemic heart disease, myocardial infarction, ischemic cerebrovascular disease, and ischemic stroke. ***p<0.001, **p<0.01, and *p<0.05 by Student s t-test or Pearson s chi-square test comparing individuals with and without ischemic cardiovascular events.

14 Supplementary Table II. Predictive ability of apolipoprotein B, total cholesterol,, apolipoprotein B/apolipoprotein A1, total cholesterol/hdl cholesterol, and non-hdl cholesterol as predictors of any ischemic cardiovascular event in strata of age, body mass index, hypertension, diabetes, and smoking in women and men. Women Men AUC (95% CI) P-value within strata* AUC (95% CI) P-value within strata* P-value between strata AUC (95% CI) P-value within strata* AUC (95% CI) P-value within strata* P-value between strata Age <60 years n=2, years n=2,468 <60 years n=1, years n=1,566 Apolipoprotein B 0.72 (-0.79) 0.56 (0.52-) (-0.70) 0.56 (0.52-) 0.01 Total cholesterol 0.66 (-0.74) (0.51-) (-0.68) 0.19 (0.51-) (-0.74) (0.51-) (0.54-) (0.51-) Apolipoprotein B/apolipoprotein A ( ) 0.24 ( ) (-0.67) 0.10 ( ) Total cholesterol/hdl cholesterol 0.67 (-0.74) 0.21 (0.53-) 0.01 (-0.64) 0.03 ( ) Non-HDL cholesterol 0.68 ( ) 0.02 (0.52-) (-0.68) 0.08 (0.53-) <3.7mmol/L n=2, mmol/L n=2,460 <3.6mmol/L n=1, mmol/L n=1,893

15 Apolipoprotein B (-0.70) ( ) 0.11 ( ) (-0.62) 0.67 Total cholesterol 0.61 ( ) 0.11 (-) 0.56 ( ) (0.52-) (0.54-) 0.00 (-) (0.50-) (0.50-) Apolipoprotein B/apolipoprotein A1 (0.54-) 0.03 ( ) ( ) 0.69 ( ) Total cholesterol/hdl cholesterol ( ) 0.04 ( ) 0.09 ( ) 0.87 (0.51-) Non-HDL cholesterol 0.64 (-0.68) 0.46 (-0.64) ( ) 0.93 (0.53-) Body mass index <25 kg/m 2 n=2, kg/m 2 n=2,083 <25 kg/m 2 n=3, kg/m 2 n=1,959 Apolipoprotein B 0.69 (-0.73) (-) ( ) ( ) 0.15 Total cholesterol 0.69 (-0.73) 0.84 ( ) (-0.70) 0.28 (0.51-) ( ) (0.52-) ( ) (0.50-) Apolipoprotein B/apolipoprotein A1 (-0.69) (-) ( ) 0.04 ( ) Total cholesterol/hdl cholesterol ( ) 0.02 ( ) (0.52-) (0.52-) Non-HDL cholesterol 0.69 (-0.73) 0.48 ( ) ( ) 0.98 (0.52-)

16 Hypertension Normotensive n=4,181 Hypertensive n=540 Normotensive n=3,239 Hypertensive n=317 Apolipoprotein B 0.66 ( ) ( ) (-0.62) 0.61 ( ) 0.46 Total cholesterol 0.64 ( ) 0.36 ( ) (0.54-) 0.06 (-0.72) (-0.67) 0.02 ( ) (0.54-) 0.02 ( ) 0.44 Apolipoprotein B/apolipoprotein A1 (-0.67) ( ) ( ) (0.48-) Total cholesterol/hdl cholesterol (-0.66) (0.48-) (-0.62) (0.46-) Non-HDL cholesterol ( ) 0.68 ( ) (-0.61) 0.51 ( ) Diabetes Non-diabetics n=4,695 Diabetics n=103 Non-diabetics n=3,399 Diabetics n=141 Apolipoprotein B ( ) ( ) 0.39 (-0.64) 0.61 ( ) 0.98 Total cholesterol 0.64 ( ) ( ) (-) ( ) (-0.66) 0.01 ( ) ( ) 0.05 ( ) Apolipoprotein B/apolipoprotein A1 (-0.66) ( ) 0.98 ( ) ( )

17 Total cholesterol/hdl cholesterol 0.61 (-) 0.01 ( ) (-0.61) 0.06 ( ) Non-HDL cholesterol ( ) 0.52 ( ) 0.33 (-) 0.71 ( ) Smoking status Non-smokers n=2,563 Smokers n=2,163 Non-smokers n=1,671 Smokers n=1,945 Apolipoprotein B ( ) ( ) (0.56-) ( ) 0.82 Total cholesterol (-0.67) 0.09 ( ) (-0.62) 0.11 (-0.62) 0.11 (-0.67) 0.08 (-0.67) ( ) 0.01 (-0.62) Apolipoprotein B/apolipoprotein A1 ( ) 0.73 ( ) (0.56-) ( ) Total cholesterol/hdl cholesterol (-0.69) 0.83 ( ) ( ) 0.15 (-0.69) 0.83 Non-HDL cholesterol ( ) (-0.68) (-) 0.25 ( ) 0.96 AUC=area under the curve from receiver operating characteristics curves. CI=confidence interval. LDL=low density lipoprotein. An AUC above 0.5 indicates increased predictive ability (null hypothesis: true area=0.5) *P value for pairwise comparisons of AUC within strata, i.e. the predictive ability of, respectively, total cholesterol,, apolipoprotein B/apolipoprotein A1, total cholesterol/hdl cholesterol, and non-hdl cholesterol, compared with the predictive ability of apolipoprotein B. P value for pairwise comparisons of AUC between strata, i.e. the predictive ability of apolipoprotein B, total cholesterol,, apolipoprotein B/apolipoprotein A1, total cholesterol/hdl cholesterol, and non-hdl cholesterol in individuals <60 years of age compared with those 60 years of age.

18 Supplementary Table III. Characteristics of participants from the general population who had an ischemic cardiovascular event during followup by Framingham risk score classification. Women Men Risk score by > apolipoprotein B Risk score by = apolipoprotein B Risk score by apolipoprotein B > Risk score by LDL cholesterol > apolipoprotein B Risk score by = apolipoprotein B Risk score by apolipoprotein B > Number of individuals Age (years) 73.9± ± ± ± ± ±11.5 Apolipoprotein B (mg/dl) 81.9± ± ± ± ± ±15.7*** Total cholesterol (mmol/l) 6.6± ± ± ± ± ±0.8 (mmol/l) 4.2± ± ±0.9** 4.4± ± ±0.7** HDL cholesterol (mmol/l) 1.9± ± ± ± ± ±0.3* Triglycerides (mmol/l) 1.2± ± ±1.0*** 1.5± ± ±1.0*** Body mass index (kg/m 2 ) 24.5± ± ±5.5* 26.5± ± ±3.2* Hypertension (n(%)) 6(27%) 71(30%) 9(29%) 10(22%) 38(16%) 20(27%) Diabetes mellitus (n(%)) 1(5%) 12(5%) 5(16%) 2(4%) 16(7%) 10(14%) Smokers (n(%)) 10(45%) 115(48%) 15(48%) 22(49%) 146(62%) 42(58%) Continuous variables are mean±standard error of the mean. LDL= low density lipoprotein. HDL= high density lipoprotein. ***p<0.001, **p<0.01, and *p<0.05 by Bonferroni corrected post hoc analysis in a t-test or Pearson s chi-square test comparing individuals in whom apolipoprotein B resulted in a higher risk score compared to those in whom LDL resulted in the highest risk score.

19 Supplementary Figure I. Apolipoprotein B level as a function of age in 10-year age groups in women (upper panel) and men (lower panel) separately. Values are median (2.5 percentile to 97.5 percentile).

20 Supplementary Figure II. Cumulative incidence of any ischemic cardiovascular event as a function of years of follow-up by tertiles of apolipoprotein B in women and men below 60 years of age (upper panels, left and right), and 60 years or above (lower panels, left and right). 1 (green)=lower tertile of apolipoprotein B: women 0-74 mg/dl and men 0-75 mg/dl. 2 (orange)=middle tertile of apolipoprotein B: women mg/dl and men mg/dl. 3 (red)=upper tertile of apolipoprotein B: women mg/dl and men mg/dl.

21 Supplementary Figure III. Sensitivity and specificity (%) Apolipoprotein B (mg/dl) (mmol/l) Sensitivity (solid line) and specificity (dashed line) as a function of apolipoprotein B (top) and (bottom), respectively, projected onto histograms of the distributions of apolipoprotein B and in individuals without events (white bars) and individuals with ischemic cardiovascular disease (red bars). Maximizing both sensitivity and specificity for prediction of any ischemic cardiovascular event, identified a cut-point for levels of apob of 90mg/dL(sensitivity and specificity both 61%) and a cut-point for levels of LDL cholesterol of 3.9mmol/L(sensitivity and specificity both 60%). The consequence of selecting these two cutpoints was that 9% of any ischemic cardiovascular events could be predicted using apob >90mg/dL, while was <3.9mmol/L, and 6% of events could be predicted using >3.9mmol/L while apob was <90mg/dL. Using the clinical established cut-point of LDL=3.0mmol/L cholesterol and apob=70mg/dl(70mg/dl corresponds approximately to the same percentile as 3.0mmol/L for ), had the consequence that 6% of any ischemic cardiovascular events could be predicted using apob>70mg/dl, while was <3.0mmol/L, and 3% of events could be predicted using LDL cholesterol while apob was <70mg/dL.