Relationship of Oxidized Phospholipids on Apolipoprotein B-100 to Cardiovascular Outcomes in Patients Treated with Intensive versus Moderate Atorvastatin Therapy: Treating to New Targets Trial

Abstract

BACKGROUND

Oxidized phospholipids on apolipoprotein B-100 (OxPL-apoB) is a biomarker of increased risk for major adverse cardiovascular events (MACE) in community cohorts, but its role in patients with stable coronary heart disease (CHD) is unknown.

OBJECTIVES

We sought to examine the relationship of these oxidative biomarkers to cardiovascular outcomes in patients with established CHD.

METHODS

In a random sample from the Treating to New Targets trial, OxPL-apoB levels were measured in 1,503 patients at randomization (after an 8-week run-in period on atorvastatin 10 mg) and 1 year after being randomized to atorvastatin 10 mg or 80 mg. We examined the association between baseline levels of OxPL-apoB and MACE, defined as death from CHD, nonfatal myocardial infarction (MI), resuscitation after cardiac arrest, and fatal/ nonfatal stroke, as well as the effect of statin therapy on OxPL-apoB levels and MACE.

RESULTS

Patients with events (n = 156) had higher randomization levels of OxPL-apoB than those without events (p = 0.025). For the overall cohort, randomization levels of OxPL-apoB predicted subsequent MACE (hazard ratio [HR]: 1.21; 95% confidence interval [CI]: 1.04 to 1.41; p = 0.018) per doubling and tertile 3 versus tertile 1 (HR: 1.69; CI: 1.14 to 2.49; p = 0.01) after multivariate adjustment for age, sex, body mass index (BMI), among others, and treatment assignment. The HR per doubling and by tertiles of OxPL-apoB remained significant in the low-dose but not the high-dose atorvastatin group.

CONCLUSIONS

Elevated OxPL-apoB levels predict secondary MACE in patients with stable CHD, a risk that is mitigated by atorvastatin 80 mg.

Atherosclerosis is a chronic inflammatory disease whose pathogenesis is strongly mediated by dyslipidemia and other risk factors activating both innate and adaptive immune systems to generate chronic pro-inflammatory responses (1). Lipoprotein oxidation generates so-called “oxidation-specific epitopes” (OSE) that are an important, if not obligatory, intermediary between dyslipidemia, inflammation, and initiation and progression of atherosclerotic lesions. Clinical events ultimately result from the interaction of lipid accumulation, OSE generation, activation of inflammatory processes, recruitment of inflammatory cells to the vessel wall, endothelial dysfunction, platelet activation, and thrombosis.

These OSEs represent a major class of atherosclerosis-relevant antigens that define oxidative modifications on low-density lipoprotein cholesterol (LDL-C) particles, apoptotic cells, and proteins in the vessel wall (2). They are present both in the lipid phase as oxidized lipids and as oxidized lipids covalently bound to apolipoprotein B-100 (apoB) or other proteins. OSEs share molecular or immunological identity with danger-associated molecular patterns (known as DAMPs) on apoptotic cells and oxidized LDL-C, as well as pathogen-associated molecular patterns (or PAMPs) on microbial pathogens. Oxidation-specific epitopes are recognized by a common set of innate pattern recognition receptors, which have been evolutionary selected to protect against the pro-inflammatory properties of OSEs.

Phosphocholine (PC)-containing oxidized phospholipids (OxPL) are well studied OSEs that are highly immunogenic, pro-inflammatory, and present in atherosclerotic lesions of animals and humans, particularly in pathologically defined vulnerable and disrupted plaques (3). OxPLs are important contributors to early and late events in atherogenesis by activating pro-inflammatory genes in endothelial cells and macrophages (4), leading to inflammatory cascades in the vessel wall (5). Lipoprotein(a) [Lp(a)] is the major lipoprotein carrier of OxPL, which may impart pro-inflammatory properties (6,7). OxPL can be measured on apoB-100 lipoproteins (OxPL-apoB; reported as absolute units of phosphocholine-containing oxidized phospholipids [PC-OxPL] per unit apoB) in plasma with the murine natural antibody E06. OxPL-apoB levels have been shown to correlate with the presence of anatomical cardiovascular disease (CVD) (8) and events in clinical studies in community-based settings in patients generally without prior coronary heart disease (CHD) (9,10). The ability of OxPL-apoB to predict risk of angiographically-determined coronary artery disease (CAD) and CVD events is strongly potentiated by interleukin-1 genotypes that are associated with genetically-elevated predisposition to inflammation (11).

E06 binds to the PC headgroup of a variety of OxPL, but does not recognize phospholipids on unoxidized phospholipids or non-PC-oxidized phospholipids (12). Clinical outcome studies using alternative methods to measure OxPL, such as tandem liquid chromatography/mass spectrometry, have not been performed due to the laborious need to quantitate the multiple OxPL species and the lack of appropriate standards. Our group recently documented the presence of OSE in progressive coronary and carotid atheromata (3) and specific E06 and non-E06 detectable OxPL and oxidized cholesteryl esters by tandem liquid-chromatography/mass spectrometry in debris captured by distal protection devices in patients undergoing coronary or peripheral procedures, documenting their relevance to clinical disease (13,14).

The relationship of these oxidative biomarkers to cardiovascular outcomes in patients with established CHD is not defined. The aim of this study, therefore, was to assess the effect of low- versus high-dose statin therapy on OxPL-apoB levels in such patients and their relationship to subsequent major adverse cardiac events (MACE).

METHODS

STUDY DESIGN

This substudy is based on a random sample from the TNT (Treating to New Targets) study population, a randomized trial that compared the efficacy of high- (80 mg) versus low- (10 mg) dose atorvastatin for secondary prevention of CHD (15). Importantly for this substudy, blood samples were obtained at randomization, which occurred after an 8-week run-in period on atorvastatin 10 mg, and again 1 year after being randomized to 10 mg or 80 mg atorvastatin. This substudy consisted of 1,503 patients, of whom 1,347 did not experience an event and 156 had a documented MACE during the 4.9 years of study follow-up. MACE was defined as CVD death, nonfatal/ nonprocedure-related MI, resuscitated cardiac arrest, and fatal or nonfatal stroke.

OxPL-apoB levels were measured in a chemiluminescent immunoassay using the murine monoclonal antibody E06 that recognizes the PC group on oxidized but not on native phospholipids (Taleb et al. [8] and references therein). In prior studies, this variable was expressed as OxPL/apoB, reflecting the fact that this measure quantitates the number of OxPL moles per unit mass of apolipoprotein B-100 present on microtiter well plates (and not the level in the circulation). The nomenclature is now changed to OxPL-apoB to minimize confusion that this measure represents a ratio of OxPL divided by plasma levels of apoB. Details of the OxPL-apoB assay are provided in the Online Appendix. The Lp(a) levels were measured by a latex immonoturbidimetric assay (Wako Chemicals USA, Inc., Richmond, Virginia) (16).

STATISTICAL ANALYSIS

Patient demographics and disease characteristics at randomization were compared between treatment groups using a chi-square test for categorical variables and a Wilcoxon rank sum test for continuous variables. Similar comparisons were conducted between those who did and those who did not experience an event. Results in OxPL-apoB were also confirmed in a Student t test after a log transformation. The distribution of OxPL-apoB was highly skewed, but the normality was greatly improved after the log transformation. Changes from randomization (i.e., after 8 weeks of open-label run-in treatment with atorvastatin 10 mg) to year 1 in OxPL-apoB were tested with a Wilcoxon signed rank test, and compared with a Wilcoxon rank sum test between treatment groups and between patients with and without a clinical outcome. Spearman correlation was calculated to assess association between randomization and year 1 OxPL-apoB as well as between OxPL-apoB and lipid and other nonlipid biomarkers. The association between on-treatment OxPL-apoB (at time of randomization and at 1 year) and a MACE endpoint were assessed in a Cox proportional hazards model after adjustment for age, sex, and treatment effect; using time to cardiovascular endpoint as the dependent variable; and for all patients and for patients within each treatment group. First, OxPL-apoB was used as a continuous variable, with log₂ transformation. Then, OxPL-apoB was ranked into tertiles and expressed as 3 dummy indicator variables using the 1^st tertile as the control. Treatment by OxPL-apoB interaction was examined separately in the same model with the addition of interaction term. The proportional hazards assumption was examined and confirmed not violated by looking at the supremum test and the cumulative martingale residuals. All analyses were performed using SAS software, version 9.

In a simple Student t test with an observed standard deviation of 4.16 nmol/l PC-OxPL, a sample size of 1,500 has 80% power to detect a difference of 0.15 SD (0.62 nmol/l) between the 2 treatment groups, or a difference of 0.24 SD (1.00 nmol/l) between patients who did and did not experience a primary endpoint event during the study, given the event rate of 10%. Alternatively, since PC-OxPL is log-transformed before analyses, on a log₂ scale, the SD is 1.043. Using a Student t test based on a log-transformed data, a sample size of 1,500 has 80% power to detect an 11.5% difference between the 2 treatment groups, or 19.1% difference between patients who did and did not experience a primary endpoint event during the study, given the event rate of 10%.

RESULTS

The characteristics of patients in the main study and in the biomarker substudy at the time of randomization are shown in Table 1. In the biomarker study 1,503 patients were included, of whom 754 were treated with atorvastatin 10 mg and 749 with atorvastatin 80 mg. Of these subjects, 156 had a MACE and 1,347 had no event. Patients with MACE were older and characterized by more hypertension, diabetes, current smokers, and lower high-density lipoprotein cholesterol (HDL-C), but similar LDL-C and Lp(a) levels. At randomization, OxPL-apoB levels were not significantly different between atorvastatin 10 mg and atorvastatin 80 mg. However, patients with events had higher levels (median interquartile range [Q1, Q3]) of OxPL-apoB at randomization than those without events 4.4 (2.7, 7.9) versus 3.7 (2.3, 6.8) nmol/l PC-OxPL (p = 0.025; Table 1). These findings on OxPL-apoB were confirmed in a Student t test based on a log-transformed OxPL-apoB. Patients with events had 14.5% higher OxPL-apoB level, or 0.195 higher on a log₂ scale, than those without events (p = 0.027). No significant difference was observed between treatment groups.

TABLE 1.

Patient Characteristics at Randomization^*

	Main Study		Biomarker Substudy
Characteristic	By Treatment		All Patients	By Treatment		By Event
	ATV 10 mg (n = 5,006)	ATV 80 mg (n = 4,995)	(n = 1,503)	ATV 10 mg (n = 754)	ATV 80 mg (n = 749)	With Event (n = 156)	Without Event (n = 1,347)	p Value†
Age, yrs	60.9 (8.8)	61.2 (8.8)	61.5 (8.7)	61.3 (8.5)	61.8 (8.9)	62.8 (8.5)	61.4 (8.7)	0.034
Male, %	80.8	81.2	78.8	80.1	79.7	83.3	78.3	0.14
Risk factor
Current smoker, %	13.4	13.4	12.4	14.6	11.2	18.6	11.7	0.034
Hypertension, %	54.4	53.9	57.7	58.4	55.9	69.2	56.4	0.002
Diabetes, %	15.0	15.0	16.6	16.9	16.3	26.3	15.5	0.0006
Lipids
LDL cholesterol, mg/dl	98 (18)	98 (17)	98 (18)	98 (18)	97 (18)	98 (18)	98 (18)	0.91
Total cholesterol, mg/dl	175 (24)	175 (24)	175 (24)	175 (23)	175 (25)	175 (23)	175 (24)	0.96
Triglycerides, mg/dl	134 (101, 184)	135 (102, 183)	138 (106, 189)	142 (106, 191)	136 (106, 185)	139 (110, 200)	138 (105, 187)	0.27
HDL cholesterol, mg/dl	47 (11)	47 (11)	47 (11)	46 (11)	47 (11)	44 (10)	47 (11)	0.0007
Lp(a), mg/dl	--	--	15 (5, 39.5)	15 (5, 42)	14 (5, 37)	20 (5, 47.5)	14 (5, 38)	0.104
OxPL-apoB, nmol/l	--	--	3.8 (2.4, 6.9)	3.8 (2.4, 7.0)	3.7 (2.3, 6.9)	4.4 (2.7, 7.9)	3.7 (2.3, 6.8)	0.025

Values are mean (SD) or median (interquartile range) unless otherwise indicated.

^*At the time of randomization, all participants had received atorvastatin 10 mg for 8 weeks.

^†p value for patients who experienced an event versus those who did not in the biomarker subgroup.

ATV = atorvastatin; HDL = high-density lipoprotein; LDL = low-density lipoprotein; Lp(a) = lipoprotein(a); OxPL-apoB = oxidized phospholipids on apolipoprotein B-100.

In the overall cohort, the level of OxPL-apoB at randomization (after the 8-week run-in period with atorvastatin 10 mg) predicted recurrent events. The HR associated with doubling of the OxPL-apoB concentration was 1.20 (95% CI: 1.03 to 1.39; p = 0.018) after adjusting for age, sex, and treatment (Table 2). The difference in the atorvastatin 10 mg group remained significant (HR: 1.23; 95% CI: 1.00 to 1.51; p = 0.05), whereas in the atorvastatin 80 mg group it lost significance (HR: 1.17; 95% CI: 0.94 to 1.46; p = 0.16). Treatment interaction by OxPL-apoB levels at randomization did not reach statistical significance, likely due to the fact that the test of interaction often requires a much larger sample size. The 1-year OxPL-apoB levels and the change from randomization to year 1 in OxPL-apoB levels were not associated with higher risk of recurrent events for the overall, atorvastatin 10 mg, or atorvastatin 80 mg groups.

TABLE 2.

Association Between Randomization, Year 1, Change from Randomization to Year 1 in OxPL-apoB and Clinical Outcomes^*

	All Patients			Atorvastatin 10 mg			Atorvastatin 80 mg
	HR†	95% CI	p Value	HR†	95% CI	p Value	HR†	95% CI	p Value
Randomization	1.20	1.03, 1.39	0.018	1.23	0.99, 1.51	0.053	1.17	0.94, 1.46	0.156
Year 1	1.15	0.96, 1.38	0.143	1.22	0.95, 1.58	0.127	1.06	0.81, 1.38	0.670
Change from randomization to year 1	0.93	0.74, 1.17	0.523	0.95	0.68, 1.31	0.736	0.90	0.65, 1.23	0.503

^*At the time of randomization, all participants had received 10mg atorvastatin for 8 weeks.

^†Hazard ratio (HR) associated with doubling the concentration and adjusting for age, sex, and treatment effect. Treatment interaction by individual biomarker is not significant for all biomarkers analyzed.

CI = confidence interval; other abbreviations as in Table 1.

The findings between randomization levels of OxPL-apoB and subsequent MACE remained consistent after further adjustment for all traditional risk factors including age, sex, BMI, diabetes mellitus, systolic blood pressure (SBP), LDL-C, HDL-C, ApoB, and treatment (HR: 1.21; 95% CI: 1.04 to 1.41; p = 0.018) per doubling of OxPL-apoB concentration in the total population. This remained significant in the atorvastatin 10 mg group (HR: 1.28; 95% CI: 1.03 to 1.58; p = 0.028), but not the atorvastatin 80 mg group (HR: 1.16; 95% CI: 0.93 to 1.44; p = 0.19)).

Evaluating groups by tertiles (Figure 1), tertile 3 was associated with a higher risk of MACE compared to the first tertile in the overall group (HR: 1.69; 95% CI: 1.14 to 2.49; p = 0.01) and in the atorvastatin 10 mg group (HR: 2.08; 95% CI: 1.20 to 3.61; p = 0.01). This was not significant in the atorvastatin 80 mg group (HR: 1.40; 95% CI: 0.80 to 2.46; p = 0.24). The treatment by OxPL-apoB interaction was not significant (p = 0.37).

FIGURE 1. Association Between OxPL-apoB Levels at Randomization and Subsequent Major CV Events.

Hazard ratios versus tertile 1 were adjusted for age, sex, body mass index, diabetes mellitus, systolic blood pressure, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, apoB, and difference between treatment groups. Within each panel, the p values refer to comparisons of tertile 2 or 3 to tertile 1. ATV = atorvastatin; OxPL-apoB = oxidized phospholipids on apolipoprotein B-100.

To test the robustness of our results, we also performed a logistic regression analysis looking at the proportion of patients within the total study population who experienced a MACE during the study. After adjusting for age, sex, and treatment group, OxPL-apoB significantly predicted MACE as expected. The odds ratio (OR) associated with doubling the concentration of OxPL-apoB was 1.21 (95% CI: 1.03 to 1.42; p = 0.021). A receiving operating characteristics (ROC) analysis was also performed, showing that the area under the ROC curve increased from 0.5735 to 0.5868 by adding OxPL-apoB to the model. However, this increase of 0.0133 was not statistically significant (95% CI: −0.0175 to 0.0440; p = 0.39). The incremental area under ROC remained nonsignificant when it was added to the model of other traditional risk factor variables, including baseline levels of LDL-C, HDL-C, apoB, SBP, BMI, and diabetes mellitus. Nevertheless, this nonsignificance of incremental area under the ROC curve does not negate the potential prognostic value of OxPL-apoB for subsequent MACE since the well-established traditional risk factors also failed to show a significant incremental area under ROC to predict MACE when analyzed in the same model. However, like OxPL-apoB, traditional risk factors including age, BMI, diabetes mellitus, SBP, and HDL-C were all significant predictors in both the Cox proportional hazard and the logistic regression analyses, but failed to achieve significant incremental area under ROC. Further research is needed to validate the utility of the ROC analytical approach in such circumstances. Likely due to their strong correlation (see below), neither OxPL-apoB nor Lp(a) were significant predictors when both were included in the Cox proportional hazard analysis. We had previously reported the prognostic value of OxPL-apoB in primary prevention settings by showing that it reclassifies ~30% of patients from the general community using c-statistics and net reclassification measures (17).

In the overall cohort, the median percent change in OxPL-apoB levels from randomization to 1 year was 3.5% (interquartile range [IQR]: −28.9% to +45.1%; Figure 2A). In the atorvastatin 10 mg group, there was no significant change in the median percent change (0.5%; IQR: −30.8% to +36.7%; p = 0.061). However, in the atorvastatin 80 mg group, there was a significant increase in the median percent change at 8.0% (−IQR: −26.9% to +53.3%; p<0.0001). This treatment group difference was statistically significant (p = 0.0089). Similar findings were present in the changes of absolute levels of OxPL-apoB (Figure 2A).

FIGURE 2. Changes in OxPL-apoB Levels from Randomization to Year 1.

Both percent and absolute changes in OxPL-apoB levels are shown according to (A) treatment group and (B) whether the patient experienced a major cardiovascular (CV) event. The p values are from comparisons using the 2-sided Wilcoxon rank sum test. Other abbreviations as in Figure 1.

In patients with recurrent events, the median percent change in OxPL-apoB levels from randomization to 1 year was −7.6% (IQR: −35.5% to +49.5%; p = 0.34; Figure 2B). In patients who did not experience an event, there was a significant increase in the median change at 5.0% (IQR: −27.7% to +43.8%; p <0.0001). However, the differences between patients who did and did not experience an event were not significant (p = 0.24). Similar findings were present in the changes of absolute levels of OxPL-apoB (Figure 2B).

Spearman correlation analyses were carried out between randomization and 1-year OxPL-apoB as well as with Lp(a), which have been previously shown to correlate with each other (18). OxPL-apoB at randomization correlated with 1-year OxPL-apoB (r = 0.65; p < 0.0001) and Lp(a) at randomization correlated with 1-year Lp(a) (r = 0.94; p < 0.0001). OxPL-apoB correlated with Lp(a) at randomization (r = 0.73; p < 0.0001) and 1-year Lp(a) (r = 0.71; p < 0.0001) (Figure 3). Weak to absent correlations were noted with lipid and laboratory variables in Table 1 (data not shown).

FIGURE 3. Spearman Correlation Analysis of OxPL-apoB and Lp(a).

Correlation between OxPL-apoB and lipoprotein(a) (Lp[a]) is shows (A) at randomization according to treatment; (B) at randomization according to whether the patient experienced a major CV event; (C) at 1 year according to treatment; and (D) at 1 year according to whether the patient experienced a major CV event. Other abbreviations as in Figures 1 and 2.

DISCUSSION

This TNT trial substudy demonstrates that elevated levels of OxPL-apoB predict secondary MACE in subjects with stable CHD. This was shown despite the fact that under this protocol there was no placebo arm and all patients were treated with atorvastatin 10 mg for 8 weeks prior to randomization, which has known effects on OxPL-apoB (8). Importantly, it suggests that high-dose treatment with atorvastatin mitigates the risk conferred by OxPL-apoB. This study expands the database of OxPL-apoB predicting CVD and clinical outcomes in primary prevention settings to statin-treated patients with established CHD in the context of a contemporary randomized clinical trial (Central Illustration).

CENTRAL ILLUSTRATION. Oxidized Phospholipids, Statin Therapy, and CV Outcomes.

Upper panel: The bar graph depicts the hazard or odds ratio of oxidized phospholipids on apolipoprotein B-100 (OxPL-apoB) for the primary endpoint in each study indicated, comparing the highest to the lowest tertile or quartile. These studies included subjects without or with prior cardiovascular disease (CVD), defined as myocardial infarction, stroke, or peripheral arterial disease The data are derived from prior publications including the Bruneck Study (17), EPIC (European Prospective Investigation of Cancer)-Norfolk Trial (9), Nurses’ Health Study (NHS) and Health Professionals Follow-Up Study (HPFS) (38), Mayo Study (11) and Treating to New Targets (TNT) Trial. The Mayo Study (11) reflect odds ratio for angiographically-determined coronary artery disease.

Lower panel: The illustration depicts the risk of CVD mediated by increased circulating levels of OxPL-apoB in the context of intensity of statin therapy in patients with subclinical atherosclerosis, angiographically-determined CAD, and in established CAD by clinical criteria. As the intensity of statin therapy increases, the risk of new events (denoted a ruptured plaque on the cartoon) mediated by OxPL-apoB decreases but is not completely abrogated.

It is well accepted that innate and adaptive immune mechanisms modulate atherosclerosis. Immune responses to OSEs, such as OxPLs, play a central role in the development of atherosclerosis (2). OxPLs are pro-atherogenic and upregulate pro-inflammatory molecules in macrophages and endothelial cells (5). OxPL are present on Lp(a), both in the lipid phase and covalently bound to apolipoprotein (a) (apo(a), and the kringle IV type 10 segment of apo(a) strongly influences their binding (6,19). OxPLs mediate macrophage apoptosis, a key component of plaque vulnerability, in endoplasmic reticulum-stressed macrophages by signaling through the CD36/Toll-like receptor-2 pathway (20). Additionally, monocyte chemoattractant protein-1 (MCP-1) binds to OxPL on Lp(a) in human plasma, and once Lp(a) is retained in the vessel wall, MCP-1 retains its ability to recruit monocytes, thus promoting enhanced trafficking into the vessel wall (21). Finally, apo(a) and OxPL epitopes are significantly enriched in pathologically-defined human vulnerable plaques and clinically-relevant lesions that require intervention, consistent with a clinically relevant pathological role in the expression of CVD and events (3,13,14,22).

An extensive literature demonstrates that elevated plasma levels of OxPL-apoB predict the presence and progression of coronary, carotid, and femoral/peripheral arterial disease; predict death, MI, and stroke in subjects from the general community; and reclassify ~30% of patients into different risk categories (11). Lp(a) is also the major lipoprotein carrier of OxPLs, which may impart pro-inflammatory properties (6,7). The risk of Lp(a) is strongly associated with its content of OxPLs, which in turn are greatly enriched in small apo(a) isoforms that are associated with the highest plasma Lp(a) levels (18). Indeed, elevated plasma levels of OxPL-apoB are associated with acute coronary syndromes and increase post percutaneous coronary interventions (23,24).

Several therapeutic interventions increase OxPL-apoB levels, including low-fat diets, garlic supplements, and most statins, as noted here in the TNT Trial; we emphasize that these are absolute levels and not a ratio of OxPL to apoB levels measured in plasma. The initial hypothesis was that OxPL-apoB and Lp(a) levels would increase during hypercholesterolemia and atherosclerosis progression and would decrease during atherosclerosis regression in animals. Similarly, it was thought that OxPL-apoB and Lp(a) levels would decrease during therapies that caused regression in animal models or during low-fat diets in humans. Surprisingly, and paradoxically, we have observed the opposite to date, namely, increases in plasma OxPL-apoB in New Zealand white rabbits and cynomolgus monkeys on regression diets (25). Similarly, we observed increases in both OxPL-apoB and its carrier Lp(a) in humans on low-fat diets (26,27) or Step II American Heart Association diets (28), or following statin therapy (28–33) or administration of aged garlic supplements in subjects with CAD on a background of statin therapy (34,35).

These data demonstrate that the increase in OxPL-apoB and Lp(a) that occurred in response to various interventions was associated with regression of atherosclerosis in experimental models and with improvement in vascular function and reduced rate of progression of coronary calcification in humans (34,35). Although the mechanisms of the increase in OxPL-apoB in plasma in response to effective therapy are not understood yet, immunostaining of atherosclerotic lesions for their content of OxPL at baseline and after dietary-induced lesion regression in New Zealand white rabbits and cynomolgus monkeys showed loss of OxPL in atherosclerotic plaques concomitantly with an increase in OxPL-apoB levels in the plasma. This is consistent with an efflux of OxPL from the vessel wall to the circulation, where it was bound to apoB-containing lipoproteins such as Lp(a), and suggest the hypothesis of a potential stabilization or regression of lesions (25). Presumably, if one followed such subjects for sufficient periods of time in the future, the OxPL-apoB levels would revert to lower levels, thus explaining the ability of baseline OxPL-apoB levels to predict cardiovascular outcomes. A second possibility is that OxPL-apoB may increase along with Lp(a) increases induced by statin therapy (8). Interestingly, if one directly reduces Lp(a) in plasma, as has been shown recently with antisense therapy to apo(a) in both transgenic-Lp(a) mice (36) and humans (37), then OxPL-apoB levels decrease substantially and in proportion to the decrease in the carrier Lp(a).

STUDY LIMITATIONS

This is a TNT substudy and the number of events was modest (n = 156) compared to the overall trial (n = 982) (15). Blood samples were not available prior to randomization and it is likely that 8 weeks of atorvastatin therapy affected the levels of OxPL-apoB, which may have attenuated its full predictive power.

CONCLUSIONS

OxPL-apoB levels are elevated in patients with subsequent MACE and these levels also predict secondary MACE in subjects with stable CHD, a risk that is mitigated by high-dose (80 mg) atorvastatin. Elevated levels of OxPL-apoB appear to be a clinically informative biomarker in predicting MACE in secondary care prevention settings.

PERSPECTIVES.

Competency in Medical Knowledge

Patients with cardiovascular disease who have elevated levels of pro-inflammatory oxidized phospholipids on apoB-containing lipoproteins are a high risk of ischemic events. High-dose statin therapy, but not low-dose therapy, reduces the risk mediated by oxidized phospholipids.

Translational Outlook

Future studies should assess the value of measuring oxidized phospholipid levels on apoB-containing lipoproteins to risk-stratify patients for ischemic events and guide statin dosing.

ABBREVIATIONS

CHD

coronary heart disease

LDL-C

low-density lipoprotein cholesterol

Lp(a)

lipoprotein(a)

MACE

major adverse cardiac events

OSE

oxidation-specific epitope

OxPL-apoB

oxidized phospholipids on apolipoprotein B-100

PC-OxPL

phosphocholine-containing oxidized phospholipids