Abstract
Evaluation of: Heslop CL, Frohlich JJ, Hill JS. Myeloperoxidase and C-reactive protein have combined utility for long-term prediction of cardiovascular mortality after coronary angiography. J. Am. Coll. Cardiol. 55(11), 1102–1109 (2010).
Identifying people at high risk of cardiovascular events is the cornerstone of cardiovascular disease prevention and a major challenge for healthcare worldwide. Recently, both inflammatory and oxidative markers have been shown to improve cardiovascular risk prediction models in a wide range of patients. Here, we evaluate a recent publication investigating the value of inflammatory and oxidative markers for the prediction of cardiovascular mortality in patients with stable coronary artery disease. This study shows that the use of multiple markers may increase the predictive power of traditional risk models. The findings are discussed in the context of cardiovascular risk prediction in general.
In this article, we shall discuss the importance and goals of primary and secondary prevention of atherosclerotic complications and evaluate a recent paper examining the combined value of C-reactive protein (CRP) and myeloperoxidase (MPO) in predicting cardiovascular mortality among patients with stable coronary artery disease (CAD) Citation[1]. Identifying individuals at high risk of cardiovascular disease is the cornerstone of cardiovascular disease prevention and a major public health challenge worldwide. The Framingham risk score is one of the best validated risk algorithms currently available and has shown excellent discriminative power in predicting cardiovascular events using traditional risk factors including sex, age, diabetes mellitus, systolic blood pressure, smoking, and LDL- and HDL-cholesterol. Apart from these established risk factors, our more recently acquired understanding of the pathogenesis of atherosclerosis and cardiovascular disease suggests that inflammation Citation[2] and oxidative stress also play a crucial role in the development of atherosclerosis and its clinical manifestations, such as myocardial infarction and stroke. However, these recent advances have not been translated into routine clinical practice, where the mainstay of treatment is still lipid-targeted therapy. Biomarker risk prediction is rapidly evolving, and recently, numerous reports on the relationship between various inflammatory and oxidative biomarkers have shown them to be associated with cardiovascular risk, both in apparently healthy individuals and in patients with prevalent coronary heart disease. Among the range of biomarkers proposed for diagnostic use, oxidized LDL (oxLDL), MPO, lipoprotein-associated phospholipase A2, pentraxin-3, cytokines such as IL-6, proteases such as matrix metalloproteinase-9, and CRP have received considerable attention Citation[2–5].
Myeloperoxidase is a heme-containing peroxidase abundantly expressed in neutrophil leukocytes. Enzymatically active MPO is a key contributor to the oxygen-dependent microbicidal activity of phagocytes. In addition, excessive generation of MPO-derived oxidants has been linked to tissue damage in many diseases, especially those characterized by acute or chronic inflammation. Furthermore, MPO affects various processes involved in cell signaling and cell–cell interactions, and is, as such, capable of modulating inflammatory responses. Recent epidemiological studies have shown associations between MPO and the risk of cardiovascular disease, both in apparently healthy people and in patients with prevalent CAD Citation[6]. In addition, MPO activity can result in nitration of protein tyrosine residues, yielding nitrotyrosine, an independent risk factor of CAD Citation[7,8].
Methods & results
The study under evaluation reports a post hoc analysis from a cohort of 1117 consecutive patients (797 men and 320 women) who underwent coronary angiography at two teaching hospitals in Vancouver, Canada. Indications for angiography included stable angina, previous myocardial infarction, aortic stenosis and/or regurgitation, and mitral regurgitation. Patients with unstable angina or myocardial infarction within the preceding 2 months were excluded. Lesions visualized in major epicardial vessels were assessed semiquantitatively for percent stenosis, rounded to the nearest 10%. Presence of CAD was defined by the presence of any lesion of 20% or greater stenosis and severe CAD was defined by any lesion of 50% or greater stenosis. Baseline characteristics were obtained by a questionnaire, measurements, or from patients’ medical charts. Fasting blood samples were collected before angiography and used for the biomarker measurements. CRP, MPO and nitrotyrosine concentrations were measured by validated assays. Mortality data to the end of 2007 were obtained using patient identifiers linked to the British Columbia Vital Statistics Agency mortality database. Cardiovascular deaths were defined by the WHO International Classification of Disease 10th Revision mortality codes I20–I25 and I60–I69; 117 cardiovascular deaths and 257 deaths were recorded over a median follow-up of 12.9 years. The risk of cardiovascular mortality was increased among patients in the highest tertiles for MPO (p-value between highest and lowest tertiles = 0.007) and nitrotyrosine (p-value between highest and lowest tertiles = 0.029). After adjustment for age, sex, total/HDL-cholesterol ratio, BMI, smoking, diabetes mellitus, hypertension, the presence of 50% or more stenosis and CRP, patients in the top MPO tertile had a hazard ratio (HR) of 1.75 (95% CI: 1.16–3.10; p = 0.011) compared with those in the bottom tertile. The equivalent HR for nitrotyrosine was 1.47 (95% CI: 0.89–2.50; p = 0.10). Upon additional adjustment for CAD severity and left ventricular ejection fraction (n = 415), the risk associated with the highest MPO tertile remained elevated (HR: 2.88; 95% CI: 1.18–7.03; p = 0.02).
Next, an oxidative stress score was calculated based on the combined presence of elevated levels of MPO, nitrotyrosine, oxLDL and antioxidant capacity (AOC). Patients with three or four elevated markers had a 2.8-fold increased risk of cardiovascular mortality (95% CI: 1.5–5.4; p = 0.001), but this effect was attenuated by MPO, suggesting that these effects were predominantly MPO-mediated. However, MPO combined with CRP predicted cardiovascular mortality risk such that patients with both biomarkers in the highest tertile had a 5.3-fold increased risk of cardiovascular mortality compared with patients with both markers in the lowest tertiles (95% CI: 1.9–14.9; p = 0.002), which remained significant after multivariate adjustment. Finally, a multivariate adjusted model including CAD severity showed that the area under the receiver-operating characteristic curve improved from 0.715 to 0.781 upon adding MPO and CRP (p = 0.004). Reclassification analyses showed that, compared with a model based on established risk factors, adding MPO resulted in 14.4% improvement in the net reclassification index (p = 0.003). Compared with a model based on established risk factors and CRP, adding MPO resulted in an improvement of 9.6% (p = 0.05).
Discussion & significance
This is the first post hoc analysis showing an association between MPO and cardiovascular mortality in patients with stable CAD. Although several markers of oxidative stress improved the prediction of cardiovascular mortality in this analysis, the use of MPO resulted in the strongest improvement. Interestingly, Stefanescu et al. could not find an independent predictive role for MPO among 382 patients with stable CAD. A proportional hazard model that corrected for multiple traditional risk factors showed that the highest MPO tertile compared with the first and second tertiles was associated with a HR of 1.06 (95% CI: 0.71–1.59; p = 0.77) for mortality Citation[9]. In the article under review, the combined use of MPO and CRP had independent value in improving cardiovascular risk prediction.
Expert commentary & conclusion
Association studies are of great value to identify novel risk factors and to improve existing risk algorithms. However, the design and statistical analysis of this study have some important drawbacks, which should be taken into account when interpreting the results. First, survival analyses virtually always yield a higher C-statistic when assessing the predictive accuracy in the dataset used to develop the model Citation[10]. This should be kept in mind when interpreting the conclusions from the survival studies. Second, it was the authors’ objective to evaluate “whether plasma oxidative stress biomarkers, separately and in combination, predict future fatal cardiovascular events in a large, prospective cohort of nonemergent coronary angiography patients”. Their main observation, however, suggests that MPO combined with inflammatory biomarker CRP had the greatest predictive value for cardiovascular disease, indicating that other markers were included in their analysis as well. Together with the number of biomarkers and covariates, including but not limited to MPO, left ventricular ejection fraction, use of ACE inhibitors, nitrotyrosine, AOC, ox-LDL, severity of CAD, combined oxidative risk scores containing four biomarkers, and combined testing of CRP and MPO, this already suggests that numerous tests have been performed. Furthermore, despite the fact that overall mortality is mentioned in the methods section, no results are presented in the paper. Thus, we can only speculate how many biomarkers were tested to elucidate the risk associated with overall mortality. From the 78 presented p-values in the paper, we could only find six nonsignificant values, from which three are presented in the baseline characteristics. If we make a conservative estimate that 100 statistical tests were performed, multiple testing correction according to the Bonferroni method would possibly result in other conclusions than the ones presented in the paper. Although the addition of biomarkers to current risk algorithms holds promise for clinical practice, a thorough analysis with external validation should be performed, investigating large cohorts including patients from the entire risk spectrum. Only then can subsequent steps can be taken to incorporate biomarkers into daily clinical treatment decisions to improve outcome in the primary prevention setting and for patients with established stable CAD.
Five-year view
Over the next 5 years, we will continue to see rapid progress in the field of cardiovascular risk modeling. New biomarkers, including inflammatory and oxidative markers, will probably prove to be useful in improving cardiovascular risk algorithms. However, real advances will not be made easily. A major limitation of the current one-measurement approach is that it provides no dynamic information for the biomarker under investigation. Most likely, intra-individual variation of biomarker concentrations will turn out to be very important in predicting cardiovascular risk. However, the design of currently available datasets does not allow the testing of this hypothesis. We will need an entirely new approach to study designs addressing this topic. In addition, new risk prediction models should be tested in multiple independent cohorts in patients from all risk categories. The subsequent implementation into daily clinical practice depends on ease of use, cost–effectiveness, measurement reproducibility and standardization, and the ability to add to the predictability of existing biomarkers and risk models.
Key issues
• Cardiovascular disease is the leading cause of death in the Western world.
• Identification of low-risk patients enables more accurate treatment allocation, leading to more cost-effective therapy.
• Identification of high-risk patients enables adequate allocation of preventive therapy, improving patient outcome and treatment success.
• Multiple biomarkers incorporated in current risk algorithms hold promise for improved risk assessment.
Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.
References
- Heslop CL, Frohlich JJ, Hill JS. Myeloperoxidase and C-reactive protein have combined utility for long-term prediction of cardiovascular mortality after coronary angiography. J. Am. Coll. Cardiol.55(11), 1102–1109 (2010).
- Libby P, Ridker PM, Hansson GK. Inflammation in atherosclerosis: from pathophysiology to practice. J. Am. Coll. Cardiol.54(23), 2129–2138 (2009).
- Munzel T, Gori T. Lipoprotein-associated phospholipase A(2), a marker of vascular inflammation and systemic vulnerability. Eur. Heart J.30(23), 2829–2831 (2009).
- De Keyzer D, Karabina SA, Wei W et al. Increased PAFAH and oxidized lipids are associated with inflammation and atherosclerosis in hypercholesterolemic pigs. Arterioscler. Thromb. Vasc. Biol.29(12), 2041–2046 (2009).
- Ridker PM, Danielson E, Fonseca FA et al. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. N. Engl. J. Med.359(21), 2195–2207 (2008).
- Schindhelm RK, van der Zwan LP, Teerlink T, Scheffer PG. Myeloperoxidase: a useful biomarker for cardiovascular disease risk stratification? Clin. Chem.55(8), 1462–1470 (2009).
- Shishehbor MH, Aviles RJ, Brennan ML et al. Association of nitrotyrosine levels with cardiovascular disease and modulation by statin therapy. JAMA289(13), 1675–1680 (2003).
- Mohiuddin I, Chai H, Lin PH et al. Nitrotyrosine and chlorotyrosine: clinical significance and biological functions in the vascular system. J. Surg. Res.133(2), 143–149 (2006).
- Stefanescu A, Braun S, Ndrepepa G et al. Prognostic value of plasma myeloperoxidase concentration in patients with stable coronary artery disease. Am. Heart J.155(2), 356–360 (2008).
- Ohman EM, Granger CB, Harrington RA, Lee KL. Risk stratification and therapeutic decision making in acute coronary syndromes. JAMA284(7), 876–878 (2000).