Author + information
- Received June 7, 2000
- Revision received August 23, 2000
- Accepted October 4, 2000
- Published online February 1, 2001.
- Arnold von Eckardstein, MD∗,* (, )
- Helmut Schulte, PhD†,
- Paul Cullen, MD, FRCP∗ and
- Gerd Assmann, MD, FRCP∗,†
- ↵*Reprint requests and correspondence: Dr. Arnold von Eckardstein, Institut für Klinische Chemie und Laboratoriumsmedizin, Zentrallaboratorium, Westfälische Wilhelms-Universität Münster, Albert-Schweitzer-Strasse 33, D-48129 Münster, Germany
This prospective population study was conducted to assess the role of elevated lipoprotein(a) [Lp(a)] as a coronary risk factor.
The role of elevated Lp(a) as a risk factor for coronary heart disease is controversial. In addition, little attention has been paid to the interaction of Lp(a) with other risk factors.
A total of 788 male participants of the Prospective Cardiovascular Münster (PROCAM) study aged 35 to 65 years were followed for 10 years. Both Lp(a) and traditional cardiovascular risk factors (e.g., age, low density lipoprotein [LDL] cholesterol, high density lipoprotein [HDL] cholesterol, triglycerides, systolic blood pressure, cigarette smoking, diabetes mellitus, angina pectoris, and family history of myocardial infarction) were evaluated in 44 men who suffered from myocardial infarction, and in 744 men who survived without major coronary events or stroke. A multiple logistic function algorithm was used to estimate global cardiovascular risk by the combined effects of traditional risk factors.
Overall, the risk of a coronary event in men with an Lp(a) ≥0.2 g/liter was 2.7 times that of men with lower levels (95% confidence interval [CI]: 1.4 to 5.2). This increase in risk was most prominent in men with LDL cholesterol level ≥4.1 mmol/liter (relative risk [RR]: 2.6; 95% CI: 1.2 to 5.7), with HDL cholesterol ≤0.9 mmol/liter (RR 8.3; 95% CI: 2.0 to 35.5), with hypertension (RR 3.2; 95% CI: 1.4 to 7.2), or within the two highest global risk quintiles (relative risk: 2.7; 95% CI: 1.3 to 5.7).
Lp(a) increases the coronary risk, especially in men with high LDL cholesterol, low HDL cholesterol, hypertension and/or high global cardiovascular risk.
Coronary heart disease (CHD) is multifactorial in origin. In addition to the traditional risk factors, which include age, male gender, smoking, diabetes mellitus, dyslipidemia, and hypertension, a series of novel risk factors have been identified in prospective population studies—for example, lipoprotein(a), homocysteine, fibrinogen or C-reactive protein (1,2). Because cardiovascular risk factors interact with one another, an individual’s risk of coronary events such as myocardial infarction or coronary death can best be predicted by combining the information from single risk factors using mathematical algorithms (1,3,4). Several algorithms have been deduced using “traditional” risk factors, but the interaction of novel risk factors with traditional risk factors or with global risk estimates has been little investigated.
One important novel risk factor is lipoprotein(a) (Lp[a]), a cholesteryl ester- and apolipoprotein (apo) B-containing particle, which differs from low density lipoprotein (LDL) by the additional presence of a glycoprotein termed apo(a), which is homologous to plasminogen (5,6). In vitro and in vivo, Lp(a) has been found to exert a broad variety of pro-atherogenic and pro-thrombotic properties (5,6). As the clinical consequence, Lp(a) has been identified in most but not all prospective studies as a risk factor for coronary events (7–16).
In the Prospective Cardiovascular Münster (PROCAM) study, 10-year follow-up data have now become available on a cohort of 5,333 men aged 35 to 65 years at recruitment. In a subgroup of 820 men, Lp(a) was measured in freshly isolated serum. We have now analyzed these data to investigate the interactions between traditional risk factors alone and between these risk factors and Lp(a) in predicting the occurrence of fatal or nonfatal myocardial infarction or sudden coronary death.
Description of the PROCAM study
The design of the PROCAM study has been described in detail previously (10,17). The protocol has been approved by the Ethics Committee of the Westphalian Wilhelms-University. The study has been performed in accordance with the standards of the Helsinki Declaration. In brief, some 34,000 16- to 65-year-old employees of 52 companies and public authorities were recruited between 1979 and 1985. Examinations were carried out during paid working hours and included case history using standardized questionnaires, measurement of body weight and height as well as blood pressure, a resting electrocardiogram and the collection of a blood sample after a 12-h fast for the determination of >20 laboratory parameters. All findings were reported to the participant’s general practitioner. The investigators neither carried out nor arranged for any intervention (10,17). Lp(a) was measured in a subgroup of 1,989 men and 995 women who were recruited into the PROCAM study in 1982 and 1983.
Risk factor assessment
Glucose, total cholesterol, triglycerides, and high density lipoprotein (HDL) cholesterol were measured using the Hitachi 737 autoanalyzer, enzymatic assays and (for HDL cholesterol) a precipitation method from Roche Diagnostics (Mannheim, Germany). Lp(a) was analyzed in freshly isolated sera by electroimmunodiffusion using the IgG-fraction of specific rabbit anti-Lp(a)-antisera from Behringwerke (Marburg, Germany) and standards from Immuno (Vienna, Austria) (18). The interassay imprecision for Lp(a) measurement was below 10%. Low density lipoprotein cholesterol was calculated either by the original Friedewald formula (19)or by a modification of the Friedewald formula where the estimated cholesterol content of Lp(a) (i.e., 0.0078·Lp(a) in mmol/liter) was subtracted from LDL cholesterol (20).
Anyone smoking cigarettes within the last 12 months was considered a current smoker. A proband was considered as hypertensive when he knew this diagnosis and was treated with anti-hypertensive drugs or, in accordance with the definition by the Joint National Committee (21), when systolic and diastolic blood pressures were ≥140 mm Hg and/or ≥90 mm Hg, respectively. A subject was defined as affected by diabetes mellitus if this diagnosis was known to the patient or, according to the American Diabetes Association definitions, if fasting glucose in the serum was ≥7 mmol/liter (22). A family history was considered as positive of CHD when at least one first- or second-degree relative had suffered from myocardial infarction before the age of 60 years (17).
Questionnaires were sent to the participants every two years to determine the occurrence of myocardial infarction, stroke or death. The response rate to these questionnaires was 96% after an average of two reminders per participant by mail and telephone if necessary. Hospital and physician records of diseased or deceased probands, death certificates as well as eyewitness accounts of deaths were obtained and examined by a Critical Event Committee (members: Prof. Dr. U. Gleichmann, Bad Oeynhausen; Prof. Dr. K. Kochsiek, Würzburg; and Prof. Dr. E. Köhler, Bad Salzuflen) to verify the diagnosis or cause of death. Nonfatal myocardial infarction, fatal myocardial infarction and sudden cardiac death were defined as major coronary events (MCEs) and considered as end points (10).
An explorative analysis was performed using the statistical package for the social sciences (SPSS-X) (23). Comparisons between groups were done with the Mann-Whitney U-test for continuous variables and with the chi-square test for discrete variables. Odds ratios (ORs) were calculated using the category with the lowest coronary event rate as the baseline. The ORs of elevated Lp(a) in the presence or absence of additional risk factors were compared by logistic regression analysis (24).
The simultaneous contributions of several factors to the risk of MCEs were analyzed by the use of a multiple logistic function (MLF) model (10,23). This model was originally derived from data of a population of 5,333 men aged 35 to 65 years, without a prior history of myocardial infarction or stroke, who were followed for 10 years. Of this group, 345 men experienced a coronary event; 4,626 men survived the 10 years after initial examination without definite coronary events or stroke. Age, blood pressure and laboratory data were used as continuous variables, implying a dose-response relationship to MCEs. Forward and backward selection was used to build up the logistic regression model. Both procedures were modified so that at each point of the selection process the partial significance of each term included in, or excluded from, the model was reviewed. An independent parameter was defined to be associated with coronary events if the initial probability and the partial probability value in the presence of other variables was less than 0.05. Maximum likelihood statistics were used for the selection process. As a last step, interactions between the variables remaining in the final model were tested.
The importance of Lp(a) as a cardiovascular risk factor was investigated in 820 men aged between 35 and 65 years at recruitment who were followed up for 10 years. In this subgroup, 32 men suffered from definite nonfatal myocardial infarctions and 8 others from fatal myocardial infarction. In eight other men, CHD was diagnosed but no MCE occurred. Four sudden cardiac deaths were also recorded. Eighteen men died from causes other than CHD. Eight fatal and nonfatal strokes were recorded. Seven hundred and forty-four men survived the 10 years after initial examination without nonfatal myocardial infarction or stroke.
At recruitment the 44 men with MCE+ had significantly higher levels of cholesterol, LDL cholesterol, triglycerides, and blood glucose, and lower levels of HDL cholesterol than did the 744 men without coronary events or stroke (MCE−). Systolic blood pressure was also increased in this group (Table 1).
The 44 participants who suffered an MCE had a significantly higher median level of Lp(a) (0.09 g/liter, range: 0 to 3.57 g/liter; p < 0.05) than did the 744 men without MCE (0.04 g/liter, range: 0 to 0.98 g/liter). The risk of MCE did not increase significantly within the lowest four quintiles of Lp(a) but was significantly elevated in the highest quintile of Lp(a)—that is, in men with an Lp(a) >0.17 g/liter (Fig. 1). This cutoff is close to the risk threshold value of 0.20 g/liter, which we have previously found in a case-control study (19). The RR of coronary events in men with Lp(a) ≥0.2 g/liter amounted to 2.7 (95% confidence interval: 1.4 to 5.2).
The effect of elevated Lp(a) on coronary risk was modulated by other risk factors (Table 2). The combination of Lp(a) ≥0.2 g/liter with elevated LDL cholesterol, low HDL cholesterol or hypertension further increased the risk of coronary events. In men with an LDL cholesterol <4.1 mmol/liter, HDL cholesterol <0.9 mmol/liter, or without hypertension, the presence of an elevated Lp(a) did not further increase the risk of an MCE. After correction of LDL cholesterol for Lp(a), which contributes to the LDL cholesterol estimated by the Friedewald formula, high Lp(a) increased cardiovascular risk in men with both normal and low LDL cholesterol (Table 2). Surprisingly, Lp(a) levels ≥0.2 g/liter increased the risk of coronary events in normotriglyceridemic men and in nonsmokers but not in hypertriglyceridemic men or smokers (Table 2). The interaction of elevated Lp(a) with low triglycerides also remained stable after lowering of the triglyceride cutoff to 1.7 mmol/liter (not shown). Elevated Lp(a) increased coronary risk in men without angina pectoris, without diabetes mellitus or without a family history of premature myocardial infarction. The case numbers were too small to address the question on the role of elevated Lp(a) in men with angina pectoris, diabetes mellitus or with a family history of premature myocardial infarction.
As the result of these interactions, Lp(a) ≥0.2 g/liter only increased the risk of coronary events in those men, whose coronary risk was in the top two quintiles as estimated by MLF but not in those men whose coronary risk was within the lowest three quintiles of global risk as estimated by MLF (Fig. 2). It is, however, important to note that all seven men who had an estimated risk within the lowest three quintiles and who suffered from a coronary event during follow-up presented with Lp(a) levels below 0.2 g/liter (Fig. 2). Other risk factors did not differ between men with high global risk/Lp(a) <0.2 g/liter and men with high global risk/Lp(a) ≥0.2 g/liter (Table 3).
This prospective, population-based 10-year follow-up study of 820 middle-aged men confirms and strengthens the importance of Lp(a) as a risk factor of coronary events (10). To our knowledge, the PROCAM study is the only prospective population study where Lp(a) was measured in fresh serum samples. Most prospective data on the role of Lp(a) as a cardiovascular risk factor have been obtained from population and nested case-control studies where Lp(a) was measured in serum or plasma samples that have been stored for several years (7–9,11–16). Because freezing and thawing of serum interferes with the accurate measurement of Lp(a) (25,26), this interference may be responsible for the negative outcome of some prospective studies (13–16,24).
In this PROCAM subpopulation the average risk of a 35- to 60-year-old man to suffer from a coronary event was 56 per 1,000 per 10 years. This incidence rate was significantly surpassed in men with an Lp(a) level within the highest quintile—that is, >0.17 g/liter (Fig. 1). This threshold value is close to the cutoff of 0.2 g/liter, which we previously found in a large case-control study (18). This cutoff was therefore also used to study the interaction of elevated Lp(a) with other risk factors in the determination of coronary risk (Table 2).
In agreement with data from several prospective and case-control studies, elevated Lp(a) aggravated the coronary risk in men with elevated LDL cholesterol but not in men with LDL cholesterol levels <4.1 mmol/liter (8,9,11,15,20,27,28)(Table 2). However, after adjustment of LDL cholesterol for the estimated content of cholesterol in Lp(a), this interaction disappeared. This finding is in contrast to that of others (20,29)but suggests that some of the apparant aggravating effect of elevated LDL cholesterol is explained by Lp(a) itself and therefore underscores the importance of Lp(a) as an independent risk factor.
As expected and in agreement with data from the Quebec Cardiovascular Study (15)and the National Heart, Lung and Blood Institute (NHLBI) Family Heart Study (29), elevated Lp(a) increased coronary risk more profoundly in men with hypertension or low HDL cholesterol as compared to men with normal blood pressure or HDL cholesterol. By contrast and also in light of data from the NHLBI study (29), elevated Lp(a) had statistically significant associations with coronary events in nonsmokers and in men with normal triglycerides, but not in smokers or hypertriglyceridemic subjects. These interactions must be reproduced in other studies before one speculates on the metabolic basis of these unexpected findings. Elevated Lp(a) increased the coronary risk in patients without diabetes mellitus or without family history of premature myocardial infarction. Because of small case numbers it is not reasonable to make conclusions from our study on the role of elevated Lp(a) on cardiovascular risk in subgroups of patients with angina pectoris, diabetes mellitus or a family history positive for myocardial infarction (Table 2). Previous studies are controversial as to whether Lp(a) increases cardiovascular risk in diabetes mellitus (30,31).
Combination of the risk factor information by a mathematical algorithm (10)allowed us to define two high-risk quintiles of men in which no fewer than 83% of all coronary events occurred. Lp(a) improved prediction of coronary events, especially in men with high (i.e., 5th MLF quintile) and moderately increased (i.e., 4th quintile of MLF) global risk of coronary events but does not do so in men with low estimated global coronary risk (1st to 3rd quntiles of MLF) (Fig. 2). This finding is in agreement with those of the NHLBI Family Heart Study, where, however, only the number of risk factors was evaluated, but not as in our study where the severity of multiple risk factors was evaluated (29). These observations may also explain why Lp(a) was found to be a risk factor of coronary events in those studies where participants were at high risk for coronary events (e.g., the Lipid Research Clinics Prevention Trial or the 4S study; [8,11]) but not studies such as the Physician’s Health study, the participants of which were at a generally low cardiovascular risk (14).
Finally, in this prospective population study of middle-aged men, elevated Lp(a) evolved as an important independent coronary risk factor that aggravates the coronary risk exerted by elevated LDL cholesterol (as estimated by the Friedewald formula), low HDL cholesterol, hypertension or the combined effects of multiple risk factors. The Lp(a) levels are predominantly determined by genetic variation of the apo(a) gene (5,6). Except for sex steroids, thyroid hormone and growth hormone, no drugs are known to lower Lp(a) levels (5,6). To date it is only the HERS study (32), which demonstrated an association between the lowering of Lp(a) by the use of hormone replacement therapy and the reduction of second coronary event rates in postmenopausal women. Therefore and because Lp(a) increases the risk of coronary events strongly depending on the presence of additional coronary risk factors, it is imperative to strictly control additional risk factors in individuals with elevated Lp(a). In agreement with this concept, lowering of LDL cholesterol with colestipol or simvastatin was previously found to reduce coronary events in individuals with elevated Lp(a) (11,20).
☆ This PROCAM study was supported by Bundesministerium für Forschung und Technologie, Ministerium für Wissenschaft und Forschung NRW, Deutsche Forschungsgemeinschaft, and Landesversicherungsanstalt (LVA) Westfalen.
- body mass index
- coronary heart disease
- high density lipoprotein
- low density lipoprotein
- major coronary event
- multiple logistic function
- odds ratio
- PROCAM study
- Prospective Cardiovascular Münster study
- Received June 7, 2000.
- Revision received August 23, 2000.
- Accepted October 4, 2000.
- American College of Cardiology
- International Task Force for the Prevention of Coronary Heart Disease
- Assmann G.,
- Carmena R.,
- Cullen P.,
- et al.
- Grundy S.M.,
- Pasternak R.,
- Greenland P.,
- Smith S.,
- Fuster V.
- Rosengren A.,
- Wilhelmsen L.,
- Eriksson E.,
- Risberg B.,
- Wedel H.
- Cremer P.,
- Nagel D.,
- Mann H.,
- Muche R.,
- Elster H.,
- Seidel D.
- Stubbs P.,
- Seed M.,
- Lane D.,
- Collinson P.,
- Kendall F.,
- Noble M.
- Cantin B.,
- Gagnon F.,
- Moorjani S.,
- et al.
- West of Scotland Coronary Prevention study group,
- Gaw A.,
- Brown E.A.,
- Docherty G.,
- Ford I.
- Assmann G.,
- Schulte H.
- Sandkamp M.,
- Funke H.,
- Schulte H.,
- Köhler E.,
- Assmann G.
- Friedewald W.T.,
- Levy J.,
- Fredrickson D.S.
- ↵The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of high blood pressure. Arch Intern Med 1997;157:2413–46.
- Expert Committee on the Diagnosis and Classification of Diabetes mellitus
- ↵Nie NH. SPSSX User’s Guide. New York: McGraw-Hill, 1983.
- Hosmer D.W.,
- Lemeshow S.
- Sgoutas D.S.,
- Turen T.
- Kronenberg F.,
- Trankwalder E.,
- Dieplinger H.,
- Utermann G.
- Mohan V.,
- Deepa R.,
- Haranath S.P.,
- et al.