Author + information
- Received April 13, 2001
- Revision received July 10, 2001
- Accepted July 23, 2001
- Published online November 1, 2001.
- Gavin J Blake, MB, MSc, MRCPI*,†,‡,§,
- Nisha Dada, BS∥,
- Jonathan C Fox, MD, FACC¶,
- JoAnn E Manson, MD, DrPH† and
- Paul M Ridker, MD, MPH, FACC*,*,†,‡,§ ()
- ↵*Reprint requests and correspondence: Dr. Paul M. Ridker, Center for Cardiovascular Disease Prevention, Brigham and Women’s Hospital, 900 Commonwealth Avenue East, Boston, Massachusetts 02215 USA
We sought to determine prospectively whether lipoprotein-associated phospholipase A2(Lp-PLA2) was a predictor of future cardiovascular risk in women.
Inflammatory markers may help predict cardiovascular risk. Lp-PLA2levels have recently been hypothesized to be an independent predictor of cardiovascular risk in hypercholesterolemic men.
We conducted a prospective, nested case-control study among 28,263 apparently healthy middle-aged women to assess the risk of death from coronary heart disease, non-fatal myocardial infarction, and stroke associated with baseline levels of Lp-PLA2over a mean follow-up of three years.
In univariate analysis, mean levels of Lp-PLA2correlated strongly with low-density lipoprotein cholesterol (r = 0.51; p = 0.0001), were lower among women currently using hormone replacement therapy (mean 0.98 mg/l vs. 1.23 mg/l; p = 0.0001) and were significantly higher at baseline among cases (n = 123) than controls (n = 123) (mean 1.20 mg/l vs. 1.05 mg/l; p = 0.016). However, the predictive value of Lp-PLA2was markedly attenuated after adjustment for these and other cardiovascular risk factors. Specifically, the multivariate relative risks of future cardiovascular events for women in the lowest (referent) to highest quartiles of Lp-PLA2were 1.00, 0.75, 0.64 and 1.17, respectively (all p values non-significant). In contrast, the adjusted relative risks of future cardiovascular events for each increasing quartile of C-reactive protein (another marker of low-grade inflammation) were 1.00, 1.78, 2.02 and 4.66, respectively (p-value for trend = 0.002). Inclusion of Lp-PLA2levels did not significantly attenuate this latter observation.
In contrast to prior data among hyperlipidemic men, the current data suggest that Lp-PLA2is not a strong predictor of future cardiovascular risk among unselected women.
Inflammatory processes play a fundamental role in the pathogenesis of atherosclerosis (1), and several inflammatory biomarkers including high-sensitivity C-reactive protein (CRP), interleukin-6 and soluble intercellular adhesion molecule-1 have been shown to predict future vascular risk (2–11). Recently, lipoprotein-associated phospholipase A2(Lp-PLA2) has been proposed as an inflammatory marker of cardiovascular risk (12–14). This enzyme circulates in the blood in association with low-density lipoprotein (LDL) cholesterol. The synthesis of this enzyme is regulated by inflammatory cytokines (15). Lp-PLA2may contribute directly to atherogenesis, by hydrolyzing oxidized phospholipids into pro-atherogenic fragments and by generating lysolecithin, which has pro-inflammatory properties (16). Alternatively, Lp-PLA2may have antithrombotic effects by hydrolyzing platelet-activating factor (14,17).
Recently, Packard and colleagues reported in a cohort of hypercholesterolemic men that increasing plasma levels of Lp-PLA2were a strong predictor of risk for incident coronary heart disease (CHD) (13). We sought to test this hypothesis prospectively in a lower-risk population of women.
The Women’s Health Study (WHS) is an ongoing randomized, double-blind, placebo-controlled trial of aspirin and vitamin E being conducted among women age 45 years and above with no history of cardiovascular disease or cancer (18). Blood samples were collected in EDTA-containing tubes from 28,263 (71%) of these women at baseline and stored in liquid nitrogen until analysis.
Questionnaires are sent to WHS participants to elicit information on cardiovascular risk factors and incident cardiovascular events. For this analysis, cases were defined as study participants who provided a baseline blood sample and who subsequently had a cardiovascular event as defined by death due to CHD, non-fatal myocardial infarction (MI) or stroke. The mean follow-up period was three years.
For all cases of MI, stroke or death due to CHD, hospital records were obtained and reviewed. Myocardial infarction was confirmed if symptoms met World Health Organization criteria and if the event was associated with diagnostic electrocardiographic changes or elevated cardiac enzymes. Reported stroke was confirmed if the patient had a new neurologic event persisting for more than 24 h; computed tomography scans or magnetic resonance images were available for the majority of women who developed stroke. Death due to CHD was confirmed by review of autopsy reports, death certificates, medical records and circumstances of death.
For each case who provided a baseline blood sample, one control subject matched for age (within one year) and smoking status (current, former or never) was selected from among the remaining study participants who remained free of cardiovascular events and who had also provided a blood sample at baseline. Using these criteria, 123 cases and 123 controls were selected. The cases were composed of 63 women who had a non-fatal MI, 49 women who had a stroke and 11 women who died from CHD.
Baseline plasma samples were thawed and assayed for Lp-PLA2with an enzyme-linked immunoassay as previously described (13). Samples were captured with a monoclonal antibody against Lp-PLA2and the enzyme identified with a second monoclonal antibody labeled with biotin and a streptavidin-alkaline phosphatase conjugate. The standard was purified recombinant Lp-PLA2and there was no cross-reactivity with other phospholipase A2enzymes. C-reactive protein assays were performed using a latex-enhanced immunonephelometric assay on a BN II analyzer (Dade Behring, Newark, Delaware) (19). Samples were handled in a blinded fashion throughout the study.
The Student ttest was used to evaluate differences in means. Because the distribution of CRP was skewed, differences in medians were tested with the rank-sum test. The chi-square statistic was used to compare proportions. Analysis of trends was used to test for any evidence of association between increasing levels of each plasma marker and the risk of future cardiovascular events, after the sample was divided into quartiles according to the distribution of each marker. Pearson’s coefficient was used to assess the correlation between Lp-PLA2and LDL cholesterol, high-density lipoprotein (HDL) cholesterol and body mass index.
Logistic regression models were used to calculate relative risks and 95% confidence intervals (CIs). In addition to accounting for the variables used for matching (age and smoking status), these models adjusted for random assignment to aspirin or vitamin E. Further analyses were performed that included LDL and HDL cholesterol, body mass index, a history of hypertension, a history of diabetes, a parental history of MI, frequency of exercise and current use of hormone replacement therapy. All p values were two-tailed.
The baseline clinical characteristics of the study population are shown in Table 1. As expected, the women who had cardiovascular events had a higher mean body mass index, and were more likely to have a history of hypertension, a history of diabetes, or a parental history of early MI (before the age of 60 years) than women free from cardiovascular events. Because of matching, cases and controls were almost identical with respect to age and smoking status.
In univariate analyses, baseline levels of Lp-PLA2were higher among cases than controls (mean 1.20 mg/l vs. 1.05 mg/l; p = 0.016) (Table 1). However, levels of Lp-PLA2were also highly correlated with LDL cholesterol (r = 0.51, p = 0.0001), as well as body mass index (r = 0.22, p = 0.0007) and HDL cholesterol (r = −0.34, p = 0.0001). Thus as shown in Table 2, after adjustment for these and other cardiovascular risk factors, there was little evidence of association between Lp-PLA2and future cardiovascular risk. Specifically, the adjusted relative risks from lowest (referent) to highest quartiles of Lp-PLA2at baseline were 1.00, 0.75, 0.64 and 1.17 (p-value for trend = 0.8). In comparison, the relative risks associated with each increasing quartile of CRP, after adjustment for both lipid levels and other traditional risk factors, were 1.00, 1.78, 2.02 and 4.66 respectively (p-value for trend = 0.002) (Fig. 1). As shown in Table 3, the adjusted risk of future cardiovascular events increased by 62% for each quartile increase in CRP (p = 0.002), whereas the risk increased by 5% for each quartile of Lp-PLA2, an effect that was not statistically significant (p = 0.8). These results were essentially unchanged when both Lp-PLA2and CRP were included in the same model.
To determine whether there was a threshold effect above which Lp-PLA2levels conferred an increased risk, we performed a post-hoc analysis in which we calculated the relative risk of future cardiovascular events at different cutoffs for Lp-PLA2ranging from the 50th to the 95th percentile. In this post-hoc analysis, women with levels above the 95th percentile had a significantly increased risk in univariate analyses (relative risk = 2.67, 95% CI 1.06 to 6.69; p = 0.04). Again, however, even among this small subgroup of women with the very highest baseline levels of Lp-PLA2, this effect was attenuated in multivariate analysis (relative risk = 2.08, 95% CI 0.69 to 6.30; p = 0.2).
In subgroup analyses restricted to stroke as a separate endpoint, for each quartile increase of Lp-PLA2the crude relative risk of stroke increased by 6% (95% CI −21% to +41%; p = 0.7), whereas for each quartile increase of CRP the crude relative risk of stroke increased by 78% (95% CI +28% to +248%; p = 0.0006).
Lp-PLA2levels were higher among women currently not taking hormone replacement therapy (n = 141) than among women taking hormone replacement therapy (n = 105) (1.23 mg/l ± vs. 0.98 mg/l ± 0.47; p = 0.0001). This finding is consistent with data from animal models suggesting that estrogen decreases plasma Lp-PLA2activity (20,21). Thus we examined for evidence of effect modification according to current use of hormone replacement therapy. In analyses restricted to women not currently taking hormone replacement therapy, the unadjusted relative risk of future cardiovascular events increased by 29% with each quartile increase of Lp-PLA2(95% CI −4% to +72%; p = 0.09). This effect was attenuated in adjusted analyses, where the relative risk increased by 7% for each quartile increase of Lp-PLA2(95% CI −29% to +62%; p = 0.7). Among women currently taking hormone replacement therapy, the relative risk of future cardiovascular events for each quartile increase of Lp-PLA2increased by 10% (95% CI −21% to +52%; p = 0.6) in crude analysis and 3% (95% CI −37% to +66%; p = 0.9) in adjusted analysis.
In this prospective study of healthy middle-aged women, plasma levels of Lp-PLA2were higher among women who subsequently developed cardiovascular events than in those who remained free of cardiovascular events. However, this effect was minimal and no longer statistically significant in analyses adjusting for traditional cardiovascular risk factors. In contrast, among the same women another marker of inflammation, CRP, was a significant predictor in both univariate and multivariate analyses.
Our univariate results are consistent with the finding of Packard and colleagues (13)that Lp-PLA2levels are somewhat higher at baseline among patients who subsequently have cardiovascular events than in those who remain free of cardiovascular disease. These data also confirm prior observations that Lp-PLA2levels are highly correlated with LDL cholesterol levels (13). However, the current data do not find an important role for Lp-PLA2as an independent predictor of future cardiovascular risk in women.
There are several potential explanations for these observed differences. First, our study evaluated women rather than men, and thus it is theoretically possible that gender differences exist for Lp-PLA2. The use of hormone replacement therapy may have affected our results, although prevalence of hormone replacement therapy did not differ between cases and controls, we controlled for use of hormone replacement in our adjusted analysis, and no significant difference was observed in analyses stratified by hormone replacement therapy use.
Second, in contrast to the hypothesis-generating study from Packard and colleagues that evaluated hypercholesterolemic individuals (13), our study evaluated a much broader cohort with lipid levels representative of the general population. Thus, as both our study and that of Packard demonstrate strong correlation between Lp-PLA2and LDL (as well as inverse correlation with HDL), it is possible that any true predictive value of Lp-PLA2may be limited to those with overt hyperlipidemia. Indeed, the strong correlation between Lp-PLA2and LDL likely explains why the distribution of Lp-PLA2values in our study is lower than that observed in the Packard data despite using an identical assay.
Third, our study involved a smaller sample size than the West of Scotland report; nonetheless previous studies of similar size to ours have found other novel plasma markers of inflammation to be significant predictors of future cardiovascular risk (10,22). Furthermore, in the present study, baseline levels of CRP were a significant predictor of future risk, indicating that this sample was of adequate size for another marker of inflammation.
Finally, the randomized use of aspirin in the present study, or pravastatin in the West of Scotland study, may have affected the results. However, our analyses were adjusted for randomized use of aspirin and vitamin E, so we do not believe this had an important effect on our results.
In the current data, women with the very highest levels of Lp-PLA2at baseline (>95th percentile) did appear to have an increased risk of future cardiovascular events. Thus, although these data do not confirm a role for Lp-PLA2as a potential screening test for atherosclerotic risk in a general population, they do support further research regarding this unique phospholipase in sub-populations selected for more traditional risk factors.
☆ Supported by grants from the National Heart, Lung, and Blood Institute (HL58755 and HL43851). Dr. Ridker is also the recipient of an Established Investigator Award from the American Heart Association and a Doris Duke Clinical Scientist Award from the Doris Duke Charitable Foundation.
- confidence interval
- coronary heart disease
- C-reactive protein
- high-density lipoprotein
- low-density lipoprotein
- lipoprotein-associated phospholipase A2
- myocardial infarction
- Women’s Health Study
- Received April 13, 2001.
- Revision received July 10, 2001.
- Accepted July 23, 2001.
- American College of Cardiology
- Kuller L.H,
- Tracy R.P,
- Shaten J,
- Meilahn E.N
- Tracy R.P,
- Lemaitre R.N,
- Psaty B.M,
- et al.
- Koenig W,
- Sund M,
- Frohlich M,
- et al.
- Ridker P.M,
- Rifai N,
- Stampfer M.J,
- Hennekens C.H
- Hwang S.J,
- Ballantyne C.M,
- Sharrett A.R,
- et al.
- Ridker P.M,
- Glynn R.J,
- Hennekens C.H
- Cao Y,
- Stafforini D.M,
- Zimmerman G.A,
- McIntyre T.M,
- Prescott S.M
- MacPhee C.H,
- Moores K.E,
- Boyd H.F,
- et al.
- Buring J.E,
- Hennekens C.H
- Rifai N,
- Tracy R.P,
- Ridker P.M
- Yasuda K,
- Takashima M,
- Sawaragi I
- Ridker P.M,
- Buring J.E,
- Rifai N