Author + information
- Received November 7, 2005
- Revision received January 11, 2006
- Accepted January 16, 2006
- Published online June 6, 2006.
- Markus Exner, MD⁎,
- Erich Minar, MD†,
- Wolfgang Mlekusch, MD†,
- Schila Sabeti, MD†,
- Jasmin Amighi, MD†,
- Wolfgang Lalouschek, MD‡,
- Gerald Maurer, MD§,
- Christian Bieglmayer, PhD⁎,
- Heidi Kieweg⁎,
- Oswald Wagner, MD⁎ and
- Martin Schillinger, MD†,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Martin Schillinger, Division of Angiology, Department of Internal Medicine II, Medical University Vienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria.
Objectives We investigated the effect of myeloperoxidase (MPO) on progression of carotid stenosis in states of high and low high-density lipoprotein-cholesterol (HDL-C) and low-density lipoprotein-cholesterol (LDL-C) levels.
Background Myeloperoxidase is pivotally involved in the pathogenesis of atherosclerosis. In vitro data suggest that MPO exerts deleterious effects via oxidative modulation of lipoproteins.
Methods We prospectively studied 1,019 of 1,268 consecutive patients who were asymptomatic with respect to carotid artery disease. Patients underwent serial carotid ultrasound investigations at baseline and after a follow-up interval of median 7.5 months (range 6 to 9 months), categorizing carotid arteries as 0% to 29%, 30% to 49%, 50% to 69%, 70% to 89%, or 90% to 99% stenosed or occluded. The MPO, HDL-C, and LDL-C levels were measured at baseline, grouped by medians, and correlated with progression of carotid atherosclerosis.
Results Progression of carotid atherosclerosis was found in 100 of 1,019 patients (9.8%). Myeloperoxidase (p = 0.014) but not HDL-C (p = 0.95) or LDL-C (p = 0.30) were associated with progressive disease. However, MPO ≥310 ng/ml was significantly associated with progressive disease (adjusted odds ratio [OR] 2.57, 95% confidence interval [CI] 1.39 to 4.75) only in patients with HDL-C levels <49 mg/dl. Otherwise, in patients with higher HDL-C levels (≥49 mg/dl), MPO ≥310 ng/ml did not predict disease progression (adjusted OR 1.42, 95% CI 0.72 to 2.78). No interaction of MPO with LDL-C was observed.
Conclusions Myeloperoxidase was associated with progression of carotid atherosclerosis in patients with HDL cholesterol levels below 49 mg/dl.
Recent evidence suggests that myeloperoxidase (MPO), a hemoprotein with mainly microbicidal properties, is fundamentally involved in the development of atherosclerotic lesions (1). Myeloperoxidase seems to play a potentially important role in promoting the progression of pre-existent atherosclerosis, thus triggering the occurrence of acute vascular events. Leukocyte- and blood-MPO levels are significantly higher in patients with coronary artery disease compared with healthy control subjects (2), and elevated MPO levels in patients with acute chest pain and acute coronary syndromes identify individuals with increased risk for an adverse cardiovascular outcome (2–4).
Imbalance between oxidative proatherogenic factors and antiatherogenic mechanisms is a key feature in the pathogenesis of atherosclerosis (5,6). Myeloperoxidase has been detected in human atherosclerotic lesions (7) and is essentially involved in the catalytic consumption of nitric oxide, impairing endothelium-dependent vasorelaxation (8–10). Myeloperoxidase directly activates metalloproteinases and facilitates destabilization and rupture of the vulnerable plaque (11,12). However, a major part of MPO’s deleterious effects on the vasculature seems to be mediated via oxidative modification of lipoproteins. In vitro studies demonstrated that MPO catalyzes oxidative modification of high-density lipoprotein (HDL) (13) leading to the development of dysfunctional proinflammatory and proatherogenic HDL (14,15), a finding which recently has been confirmed in vivo in human atherosclerotic tissue (16). Myeloperoxidase also catalyzes low-density lipoprotein (LDL) modification, thus facilitating its uptake by macrophages (17–22). Macrophages thereby convert into foam cells, which are characteristic of early atherosclerotic lesion formation.
We hypothesized that an interaction of MPO with lipoproteins (HDL and LDL) in the circulation may be associated with progression of atherosclerotic disease assuming that MPO may differentially exert its deleterious effects depending on patients’ HDL and LDL levels. Therefore, the aim of the present study was to investigate the association between progression of the degree of carotid stenosis measured by high-resolution duplex ultrasound and systemic MPO serum levels in states of low and high HDL-cholesterol (HDL-C) and LDL-cholesterol (LDL-C) levels.
We prospectively enrolled 1,268 consecutive Caucasian patients in the Inflammation in the Carotid Arteries Risk for Atherosclerosis Study (ICARAS) protocol. Analysis of MPO as a candidate predictive biomarker was prespecified in the original ICARAS protocol. Study design, patient characteristics, and the association between C-reactive protein (CRP) and serum amyloid A (SAA) with progression of carotid atherosclerosis have been reported (23). Briefly, patients without recent symptoms of carotid artery disease, defined as the absence of transient ischemic attacks, amaurosis fugax, or stroke 12 months before inclusion, assessed by a neurologist, underwent serial carotid duplex ultrasound investigations at baseline and after a six- to nine-month interval (median 7.5 months). Primary study end point was uni- or bilateral progression of carotid atherosclerosis in the extracranial internal carotid arteries (ICA), measured by duplex ultrasound. In the current study, we aimed to directly evaluate progression of atherosclerotic stenosis instead of measuring surrogate markers of carotid disease (such as intima-media thickness); therefore, changes in the degree of stenosis were used as a measure of plaque progression.
For the current analysis, we included 1,019 of 1,268 patients (80%) in whom MPO and HDL cholesterol levels could be obtained at the baseline visit. The study complied with the Declaration of Helsinki and was approved by institutional ethics committee. All patients provided written informed consent.
We used the following categories to quantify the degree of ICA stenosis at baseline and follow-up: 0% to 29% (carotid plaques), 30% to 49% (advanced plaques), 50% to 69% (moderate stenosis), 70% to 89% (high-grade stenosis), 90 to 99% (subocclusive stenosis), and occlusion (Table 1).Progression of carotid disease was defined as an increase of the degree of stenosis by at least one category. Progression of stenosis in either one or both ICAs was considered to be indicative of progressive disease. A secondary objective was the change of the peak systolic velocity in the ICA (PSVICA) from baseline to follow-up, reflecting a continuous surrogate marker for the extent of progressive carotid disease.
The agreement of duplex ultrasound with respect to North American Symptomatic Carotid Endarterectomy Trial angiographic criteria was assessed previously in our duplex laboratory in an independent cohort including 1,006 carotid arteries (24) Positive and negative predictive values ranged from 70% to 98%, and interobserver agreement with respect to the degree of stenosis was excellent (kappa 0.83, 95% confidence interval [CI] 0.79 to 0.88). Experienced medical technicians under supervision by one of the authors did all duplex recordings, and baseline and follow-up investigators for the same patient varied randomly and were blinded for the baseline ultrasound findings. Two independent investigators (kappa 0.85, 95% CI 0.80 to 0.89) determined progression of carotid atherosclerosis based on the recordings from baseline and follow-up duplex investigations. We censored duplex follow-up for ischemic neurologic events and carotid revascularization.
Clinical and laboratory data
After patient identification at the ultrasound laboratory, medical history and data from physical examination were recorded. We screened for infectious or inflammatory diseases by evaluating patients’ clinical history and current symptoms. Clinical suspicion for infectious or inflammatory diseases prompted further specific investigations according to clinical judgment. Criteria for definitions of risk factors and comorbidities correspond to current recommendations and are listed in the ICARAS protocol (23).
Antecubital venous blood samples were obtained at baseline visits. Measurements included HDL, LDL, and total cholesterol, glycated hemoglobin A1, high-sensitivity CRP (hs-CRP), SAA, fibrinogen, and complete blood count. Serum MPO at baseline was measured by enzyme-linked immunosorbent assay (Immundiagnostik, Bensheim, Germany) according to the manufacturer’s instructions (detection level 1.6 ng/ml, coefficient of variation 6.5%). Treating physicians and ultrasonographers were blinded for all laboratory values.
Pharmacotherapy of patients with evidence of atherosclerosis followed a standard protocol: Patients received antithrombotic therapy with either 100 mg acetylic salicylic acid or 75 mg clopidogrel once daily. Patients with hyperlipidemia (LDL-C >130 mg/dl) received inhibitors of the 3-hydroxy-3-methylglutaryl coenzyme A reductase (statins); statin therapy was intended not to be de novo or changed during the study period.
Continuous data are presented as median and interquartile range (IQR, range from the 25th to the 75th percentile) or total range. Discrete data are given as counts and percentages. We used chi-square tests, Mann-Whitney Utests, and Spearman correlation coefficients for univariate analyses as appropriate. Assessment of an interaction between MPO and lipoproteins (HDL and LDL) as a study objective was prespecified before starting the laboratory and biometric analysis on MPO. Multivariable logistic regression analysis was applied to assess the joint effects of MPO and HDL-C or LDL-C on progression of the disease, adjusting for potential confounders. The MPO and HDL-C or LDL-C levels were divided into quartiles before being entered jointly into the models to obtain clinically useful measures for the effect sizes. Interaction between the parameters was assessed by multiplicative interaction terms and log likelihood ratio tests. To account for an interaction between MPO and HDL-C or LDL-C, we grouped patients in a predefined approach according to low and high MPO and lipoprotein levels using the respective medians as cut-off values, resulting in four similarly sized groups. Results of the logistic regression models are presented as the odds ratio (OR) and the 95% CI. The linearity of the logic assumption was checked for continuous predictor variables, and an analysis of residuals was performed. Regression diagnostics and overall model fit were performed according to standard procedures (25). A two-sided p value of <0.05 was considered to be statistically significant. Calculations were performed with Stata release 8.0 (Stata Corp., College Station, Texas) and SPSS for Windows version 10.0 (SPSS Inc., Chicago, Illinois).
We analyzed 1,019 of 1,268 patients (80%) of the ICARAS trial cohort with complete baseline and follow-up data and available serum samples for determination of MPO levels. Median age of these patients was 69 years and 62% were males. Demographic data and clinical characteristics are given in Table 2.These variables were statistically not significantly different to the complete study population of 1,268 patients enrolled in the ICARAS protocol.
In 249 patients (20%) serum samples for MPO measurement were not available. Baseline characteristics and proportion of patients with progressive disease within these 249 patients were comparable with the entire study population of 1,268 patients (data not shown).
The MPO levels at baseline were median 310 ng/ml (IQR 220 to 416 ng/ml), and HDL-C and LDL-C levels at baseline were median 49 mg/dl (IQR 41 to 61 ng/ml) and 118 mg/dl (IQR 93 to 147 ng/ml), respectively. Myeloperoxidase and HDL-C were weakly negatively correlated (r = −0.083; p = 0.013). Myeloperoxidase showed no correlation with LDL-C (r = 0.02; p = 0.52). Patients with a history of stroke tended to have higher MPO levels (median 468 ng/ml, IQR 348 to 585 vs. median 425 ng/ml, IQR 300 to 583; p = 0.055). There were no significant associations between MPO levels and symptoms of coronary artery disease (p = 0.55), history of myocardial infarction (p = 0.49), or symptoms of peripheral artery disease (p = 0.96).
During the follow-up period of median 7.5 months (range 6 to 9 months), 100 of 1,019 patients (9.8%) showed an increase in the degree of carotid stenosis by at least one category. Of these, 42 patients (4.1%) showed progression of a left ICA stenosis, 52 patients (4.1%) of a right ICA stenosis, and 6 patients (0.6%) in both ICAs. According to the baseline degree of stenosis, 20 (4.0%) of 498 patients with carotid plaques, 11 (11.1%) of 99 with advanced plaques, 36 (16.0%) of 225 patients with moderate stenosis, and 33 (17.6%) of 188 patients with high-grade stenosis showed a progression (p < 0.001). In the entire group of 1,019 patients the change in PSVICAas a continuous surrogate marker for the degree of carotid stenosis from baseline to follow-up was median 0.01 m/s (IQR −0.01 to 0.11 m/s); in the 100 of 1,019 patients with a progressive lesion the increase in PSVICAwas median 0.54 m/s (IQR 0.34 to 1.03 m/s). Eight patients (0.8%) developed a de novo occlusion of a carotid artery; all of these patients had an ipsilateral subocclusive stenosis (90% to 99%) at baseline. None of the patients underwent a carotid revascularization procedure during the follow-up period, 10 patients (1%) had ipsilateral neurologic events, and none of the patients showed regression of the disease by one or more category.
MPO, lipoproteins, and disease progression
The MPO level at baseline was associated with progressive disease during follow-up: patients with a progressive stenosis by at least one category had significantly higher baseline MPO levels (median 369 ng/ml, IQR 264 to 445 ng/ml) compared with patients with stable disease (median MPO 304 ng/ml, IQR 216 to 409 ng/ml; p = 0.014). The HDL-C (p = 0.95) and LDL-C (p = 0.30) baseline levels were not significantly associated with disease progression. Consistent findings were observed when considering the change of PSVICAfrom baseline to follow-up as a continuous measure of progressive carotid disease: MPO (r = 0.083; p = 0.008), but not HDL-C (r = −0.27; p = 0.39) or LDL-C (r = 0.001; p = 0.98), was significantly correlated to the change in PSVICA.
Analyzing MPO and HDL-C jointly, it seemed that particularly patients with MPO ≥310 ng/ml and simultaneously HDL <49 mg/dl frequently showed progression of the degree of stenosis (Fig. 1).Formal statistical testing using multiplicative interaction terms and log likelihood ratio tests confirmed a significant interaction between these parameters: MPO ≥310 ng/ml was significantly associated with progressive disease (adjusted OR 2.57, 95% CI 1.39 to 4.75; p = 0.003) only in patients with HDL-C levels <49 mg/dl (Table 3).Otherwise, in patients with higher HDL-C levels (≥49 mg/dl), MPO ≥310 ng/ml was not significantly associated with an increased risk (adjusted OR 1.42, 95% CI 0.72 to 2.78; p = 0.31). Figure 2displays adjusted ORs for progression of disease according to baseline MPO (ptrend= 0.029) and HDL-C (ptrend= 0.47) levels separately and when combining these parameters.
Analyzing MPO and LDL-C jointly revealed no significant interaction between high and low MPO (by the median of 310 ng/ml), high and low LDL cholesterol levels (by the median of 118 mg/dl), and progression of the disease. Compared with patients with low MPO and simultaneously low LDL-C (reference group), the adjusted hazard ratio for progressive disease of patients with MPO ≥310 ng/ml and LDL-C ≥118 mg/dl was 1.72 (95% CI 1.01 to 3.08), of patients with MPO ≥310 ng/ml and LDL-C <118 mg/dl it was 1.77 (95% CI 1.01 to 3.08), and of patients with MPO <310 ng/ml and LDL-C ≥118 mg/dl it was 0.81 (95% CI 0.41 to 1.61), indicating that MPO ≥310 ng/ml was associated with disease progression irrespective of the LDL-C level.
MPO and inflammation
The MPO was weakly correlated to hs-CRP (r = 0.18; p < 0.001), SAA (r = 0.10; p = 0.002), fibrinogen (r = 0.12, p < 0.001), and leukocyte counts (r = 0.21; p < 0.001). Additionally adjusting the multivariable model for the individual’s levels of hs-CRP (or alternatively SAA/fibrinogen/leucocytes) slightly attenuated the association between MPO and progression of disease: adjusted ORs for disease progression for increasing quartiles of MPO were 1.08 (95% CI 0.55 to 2.14), 1.58 (95% CI 0.84 to 2.99), and 2.06 (95% CI 1.18 to 3.81) and for increasing quartiles of hs-CRP were 1.54 (95% CI 0.66 to 3.59), 1.88 (95% CI 0.82 to 4.31), and 2.60 (95% CI 1.18 to 6.72) compared with the respective lowest quartiles, suggesting that MPO predicts progressive disease independently of hs-CRP.
Similarly, adjusting the interaction model of MPO and HDL-C for hs-CRP levels (quartiles) revealed only a modest effect modification. Elevated MPO still was significantly associated with progressive disease (adjusted OR 2.14, 95% CI 1.15 to 4.08; p = 0.017) only in patients with HDL-C levels <49 mg/dl; otherwise, in patients with HDL levels ≥49 mg/dl, MPO ≥310 ng/ml was not significantly associated with an increased risk (adjusted OR 1.05, 95% CI 0.57 to 1.92; p = 0.88).
We found that MPO is significantly associated with progression of atherosclerotic disease, particularly in patients with HDL-C levels <49 mg/dl, whereas in patients with HDL-C ≥49 mg/dl MPO did not predict disease progression. This effect was independent from other markers of inflammation.
There is substantial evidence that oxidative processes are key features in the pathogenesis of atherosclerosis. Several studies demonstrated a potential link between the hemoprotein MPO and the development of coronary artery disease, mainly through its role in lipid peroxidation, nitric oxide consumption, and inflammatory mechanisms leading to endothelial dysfunction and plaque destabilization (8,11,21). Consistently, subjects with MPO deficiency seem cardioprotected (26), and individuals with polymorphisms for MPO linked to decreased expression similarly appear to have fewer cardiac events or evidence of cardiac disease (27–29). Furthermore, expression of human MPO by macrophages was recently reported to promote atherosclerosis in mice (22). Systemic levels of MPO are correlated to atherosclerotic burden and the incidence of cardiac events in patients with chest pain, and they predict outcome of patients with acute coronary syndromes (3,4). However, these observations relied on clinical surrogate markers of progressive atherosclerotic disease. Our present observation directly demonstrates an association between MPO and progression of the disease. This effect was independent of the level of systemic inflammation as measured by acute phase parameters and might indicate the underlying activation of leukocytes as a major contributor to progressive atherosclerotic disease (12,30,31). However, further studies are needed to determine whether this reflects a causal relationship and whether MPO specifically exerts proatherogenic effects.
An inverse correlation between HDL-C levels and the incidence of cardiovascular events has been demonstrated in numerous epidemiologic studies (32,33). The beneficial cardioprotective properties of HDL are attributed to its role in reverse cholesterol transport and its antioxidant and anti-inflammatory functions (34,35). Recent data suggest a deleterious effect of MPO on HDL, leading to dysfunctional HDL and impaired reverse cholesterol transport (14–16). Myeloperoxidase directly binds to apolipoprotein A1, the major protein component of HDL, thereby facilitating its oxidative modification (15). The results of the present study corroborate these in vitro findings and extend them to the clinical setting of carotid atherosclerosis: Patients with high baseline MPO levels showed an increased risk for progression of carotid stenosis only when they simultaneously had low HDL-C levels. Low HDL-C, in states of high MPO may serve to exaggerate the proatherogenic effects of MPO and thus uncover a severe atherosclerotic phenotype, suggesting that this interaction is of direct biologic impact and seems to be titrated by the amount of HDL-C available. No such interaction was observed for LDL-C, because high MPO was associated with disease progression irrespective of patients’ LDL-C levels.
Consistent with a recent report (3), MPO serum levels of patients enrolled in the ICARAS protocol, all with high atherosclerotic burden, were higher than those described for healthy subjects (36) using comparable analytical methods. Because blood was obtained from a peripheral vein, and taking into account the widespread activation of leukocytes described in atherosclerotic vessels (12), the local MPO levels in the atherosclerotic carotid arteries might even be higher than the levels measured systemically. Measurement of MPO serum levels seems to add to the risk stratification of patients with advanced carotid atherosclerosis and supports the rationale that pharmacologic inhibition of MPO might be a potential future antiatherosclerotic strategy. Interestingly, MPO independently from hs-CRP predicted atherosclerosis progression and thus adds to the prognostic value of this global marker of inflammation. However, no significant interaction between MPO, hs-CRP, and HDL-C was found, suggesting that a multibiomarker approach including MPO and hs-CRP jointly did not increase the diagnostic utility.
As described for the ICARAS protocol, some limitations of the present study have to be acknowledged. The potential selection bias inherent in a consecutive series of duplex examinations cannot be ruled out, and the generalizability of the findings to younger individuals with lower atherosclerotic burden or to unselected population-based samples remains to be determined. Direct demonstration of the action of MPO on HDL-C in carotid atherosclerotic plaques was beyond the scope of this epidemiologic project, and the proof of a causal relationship between MPO, HDL-C, and progression of atherosclerosis remains to be provided by experimental in vitro and in vivo data.
Intima-media thickness (IMT) measurements were not performed in these patients, although it is a sensitive parameter for progression of early stages of atherosclerosis. Our results have to be interpreted in the context of progressive atherosclerotic stenosis, and different findings may be observed when analyzing IMT progression.
Myeloperoxidase was associated with progression of carotid atherosclerosis in patients with HDL-C levels <49 mg/dl.
The authors are indebted to Angelika Haumer, Edyta Madroszkiewicz, Irene Liegler, Irene Mlekusch, Karin Pikesch, Sonja Fasching, and Sylvia Kaiser for their excellent technical assistance.
- Abbreviations and Acronyms
- confidence interval
- C-reactive protein
- high-density lipoprotein cholesterol
- high-sensitivity C-reactive protein
- internal carotid artery
- interquartile range
- low-density lipoprotein cholesterol
- odds ratio
- serum amyloid A
- Received November 7, 2005.
- Revision received January 11, 2006.
- Accepted January 16, 2006.
- American College of Cardiology Foundation
- Klebanoff S.J.
- Baldus S.,
- Heeschen C.,
- Meinertz T.,
- et al.
- Eiserich J.P.,
- Baldus S.,
- Brennan M.L.,
- et al.
- Vita J.A.,
- Brennan M.L.,
- Gokce N.,
- et al.
- Fu X.,
- Kassim S.Y.,
- Parks W.C.,
- et al.
- Bergt C.,
- Pennathur S.,
- Fu X.,
- et al.
- Zheng L.,
- Settle M.,
- Brubaker G.,
- et al.
- Zhang R.,
- Brennan M.L.,
- Shen Z.,
- et al.
- McMillen T.S.,
- Heinecke J.W.,
- LeBoeuf R.C.
- Schillinger M.,
- Exner M.,
- Mlekusch W.,
- et al.
- Hosmer D.W.,
- Lemeshow S.
- Naruko T.,
- Ueda M.,
- Haze K.,
- et al.
- Mauriello A.,
- Sangiorgi G.,
- Fratoni S.,
- et al.
- Myers C.D.,
- Kashyap M.L.
- Calabresi L.,
- Gomaraschi M.,
- Franceschini G.