Author + information
- Received July 12, 2017
- Accepted July 17, 2017
- Published online September 11, 2017.
- Daniele Pastori, MDa,
- Cristina Nocella, MSca,b,
- Alessio Farcomeni, PhDc,
- Simona Bartimoccia, MSca,
- Maria Santulli, MDd,
- Fortunata Vasaturo, MDe,
- Roberto Carnevale, PhDa,b,
- Danilo Menichelli, MDa,
- Francesco Violi, MDa,
- Pasquale Pignatelli, PhDa,∗ (, )
- ATHERO-AF Study Group
- aI Clinica Medica, Department of Internal Medicine and Medical Specialties, Sapienza University of Rome, Rome, Italy
- bDepartment of Medical-Surgical Sciences and Biotechnologies, Sapienza University of Rome, Latina, Italy
- cDepartment of Public Health and Infectious Diseases, Sapienza University of Rome, Rome, Italy
- dDepartment of Experimental Medicine, Sapienza University of Rome, Rome, Italy
- eAzienda Ospedaliera Universitaria Policlinico Umberto I, Rome, Italy
- ↵∗Address for correspondence:
Dr. Pasquale Pignatelli, I Clinica Medica, Department of Internal Medicine and Medical Specialties, Sapienza University of Rome, Viale del Policlinico 155, Rome, 00161, Italy.
Background Soluble proprotein convertase subtilisin/kexin type 9 (PCSK9) has been shown to be predictive of cardiovascular events (CVEs) in patients who are at high cardiovascular risk. No data on the effect of PCSK9 levels in patients with atrial fibrillation (AF) are available.
Objectives This study investigated the association between PCSK9 and CVEs in AF as well as the relationship between PCSK9 and urinary 11-dehydro-thromboxane B2 (11-dh-TxB2), a marker of platelet activation.
Methods We conducted a prospective, single-center cohort study, including 907 patients with AF treated with vitamin K antagonists (3,865 patient-years), to assess CVEs, including fatal and nonfatal myocardial infarction, ischemic stroke, and cardiovascular death. At admission, plasma PCSK9 and urinary 11-dh-TxB2 (n = 852) were measured. The population was divided into tertiles of PCSK9 for the analysis.
Results The mean age of patients was 73.5 ± 8.2 years, and 43.0% were women. At follow-up, 179 CVEs (4.6%/year) occurred: 43 (15.3%), 49 (15.5%), and 87 (28.0%) in the first, second, and third tertiles of PCSK9, respectively (log-rank test p = 0.009). Patients with CVEs had higher median PCSK9 compared with those without (1,500 pg/ml [IQR: 1,000 to 2,300 pg/ml] vs. 1,200 pg/ml [IQR: 827 to 1,807 pg/ml], respectively; p < 0.001). Multivariable Cox regression analysis showed that the third versus the first tertile of PCSK9 (hazard ratio: 1.640; 95% confidence interval: 1.117 to 2.407; p = 0.012), female sex, age, diabetes, smoking, heart failure, previous cerebrovascular and cardiac events, digoxin use, and total cholesterol to high-density lipoprotein cholesterol ratio were associated with CVEs. In 682 patients not treated with antiplatelet therapy, circulating PCSK9 and 11-dh-TxB2 were significantly correlated (Spearman's rho: 0.665; p < 0.001).
Conclusions Plasma PCSK9 levels are associated with an increased risk of CVEs in patients with AF. The direct correlation between PCSK9 and 11-dh-TxB2 suggests PCSK9 as a mechanism potentially implicated in platelet activation.
Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with a high risk for thromboembolic stroke (1). In addition to thromboembolism, patients with AF are at high risk of cardiovascular complications, such as myocardial infarction (MI), with a rate ranging from 0.5% to 4%/year (2). This increased risk of cardiac complications is partially explained by the simultaneous presence in patients with AF of several atherosclerotic risk factors, such as hypertension, diabetes, and peripheral artery disease. A common mechanism linking all of these factors is the alteration of platelet function, which has been found to be up-regulated in these settings (2). In particular, arterial hypertension is the most commonly represented risk factor in AF and it might be associated with increased platelet activation (3); however, available data are equivocal and the issue requires further investigation (4).
The role of platelet activation was recently investigated in a large cohort of patients with AF where urinary excretion of 11-dehydro-thromboxane B2 (11-dh-TxB2), a marker of platelet activation, was directly correlated to the incidence of cardiovascular events (CVEs) (5).
Mounting evidence suggests that proprotein convertase subtilisin/kexin type 9 (PCSK9) has adverse effects on the cardiovascular system through several pathways, including pro-inflammatory low-density lipoprotein (LDL) cholesterol oxidation and direct modification of coronary plaque composition (6,7). Recently, circulating PCSK9 was implicated in platelet activation independent of its effect on lipid metabolism (7). Thus, patients with acute coronary syndrome (ACS) were found to have significantly higher concentration of circulating PCSK9 compared with control subjects (8), and increased PCSK9 levels were associated with higher platelet reactivity in patients with ACS undergoing percutaneous coronary intervention (PCI) (9).
The interplay between soluble PCSK9 and CVEs in the general population remains unclear, given contrasting results (10–12); noteworthy, plasma concentration of PCSK9 has never been investigated in the setting of AF.
Thus, the aims of this study were to investigate: 1) the association between circulating PCSK9 and incidence of CVEs in patients with AF; and 2) the relationship between soluble PCSK9 and urinary excretion of 11-dh-TxB2.
We conducted a prospective, single-center cohort study of patients with nonvalvular AF referred from February 2008 to December 2016 to the Athero-thrombosis Centre of the Department of Internal Medicine and Medical Specialties, Sapienza-University of Rome, for monitoring and management of antithrombotic therapies. All patients presenting with nonvalvular AF who were >18 years of age were eligible for the study. All patients were treated with vitamin K antagonists after appropriate thrombotic risk stratification (13) (international normalized ratio target: 2.5). Exclusion criteria were prosthetic heart valves or the presence of any severe valvulopathies, severe cognitive impairment, chronic infections (human immunodeficiency virus infection, hepatitis C virus, hepatitis B virus), or systemic autoimmune disease. Subjects were also excluded from the study if they had active cancer or liver insufficiency (e.g., cirrhosis). At study entry, medical history was recorded for each patient. Cardiovascular risk factors, such as arterial hypertension (14), diabetes (15), and heart failure (16), were defined according to international guidelines.
At baseline, a lipid profile was obtained, including total cholesterol, high-density lipoprotein (HDL), and triglycerides. LDL cholesterol was calculated by the validated Friedewald formula, and very low-density lipoprotein cholesterol as triglycerides/5. The total cholesterol–HDL ratio was also calculated.
The primary outcome of the study was a combined endpoint of CVEs, including fatal or nonfatal MI and ischemic stroke, cardiac revascularization (stent or coronary artery bypass surgery), cardiovascular death, and transient ischemic attack. The classification and adjudication of CVEs have been previously described (5). All patients provided written informed consent at baseline. The study protocol was approved by the local ethical board of Sapienza-University of Rome (no. 1306/2007) and was conducted according to principles of the Declaration of Helsinki.
Blood samples obtained from patients after supine rest for at least 10 min were taken into tubes with 3.8% sodium citrate and centrifuged at 300 g for 10 min to obtain supernatant, then immediately stored at −80°C until use. Plasma levels of PCSK9 were measured by a commercial enzyme-linked immunoadsorbent assay. Plasma samples were diluted 1:10 in diluent buffer. Data are expressed as pg/ml, and the minimal detectable dose of PCSK9 was <10 pg/ml in human plasma. The intra-assay and interassay coefficients of variance were 5.8% and 6.9%, respectively.
Urinary samples were immediately stored at −80°C until use. The excretion of urinary 11-dh-TxB2 was measured by an enzyme-linked immunosorbent assay commercial kit. Data are expressed as ng/mg creatinine. Intra-assay and interassay coefficients of variation were 4.0% and 3.6%, respectively.
Categorical variables were reported as counts (percentage), and Pearson chi-square test was used to compare proportions. Distribution of continuous variables was assessed by the Kolmogorov-Smirnov test.
Normally-distributed continuous variables were expressed as mean ± SD. Bivariate analysis was performed with the Student t test and Pearson's linear correlation. Non-normally distributed continuous variables were expressed as median and interquartile range (IQR) (25th and 75th), and analyzed by appropriate nonparametric tests (Mann-Whitney U test and Spearman's rank correlation coefficient, rS). Group comparisons were made with analysis of variance or Kruskal-Wallis test, as appropriate. Continuous variables with non-normal distribution were log-transformed for multivariate analysis.
For the analysis, we divided the cohort into tertiles of PCSK9 values. The cumulative risk for CVEs was estimated using a Kaplan-Meier method. The survival curves were then formally compared using the log-rank test. Cox proportional hazards analysis was used to calculate the adjusted relative hazard ratio (HR) of CVEs by each clinical variable. After testing for collinearity, the following variables were used as covariates in the Cox multivariable regression model: persistent/permanent (vs. paroxysmal) AF; female sex; age (as continuous variable); diabetes (yes or no); current smoking (yes or no); heart failure (yes or no); previous cerebrovascular events (yes or no); previous cardiac events (yes or no); antiplatelet agent (yes or no); angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (yes or no); beta-blocker (yes or no); digoxin (yes or no); calcium-channel blockers (yes or no); amiodarone (yes or no); second tertile of PCSK9 (vs. first); third tertile of PCSK9 (vs. first); body mass index (as continuous variable); and total cholesterol–HDL ratio. We analyzed the use of antihypertensive drugs rather than hypertension for the model. A stepwise linear regression analysis was performed to assess determinants of 11-dh-TxB2 using the same covariates mentioned in the previous text. Values of 11-dh-TxB2 were log-transformed for the linear regression analysis. All tests were 2-tailed, and analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, Illinois). Only p values < 0.05 were considered statistically significant.
Sample size was planned assuming an HR of 1.67 when comparing the third to the first 2 tertiles, for an overall event rate of 20%. We consequently calculated that a sample size of 902 patients would give 90% power for the primary endpoint at a pre-fixed Type I error rate of 5%.
Of the initial 1,138 patients referred, 150 patients did not meet entry criteria and were excluded, 26 patients refused to participate in the follow-up study, and 55 patients were lost to follow-up (28 were diagnosed with cancer and 27 underwent cardiac valve surgery or organ transplantation) and data were censored. Thus, the final study cohort was composed of 907 patients with AF. Median follow-up was 40.5 months (IQR: 20.9 to 67.6 months; minimum 3 months, maximum 100 months), yielding 3,865 patient-years of observation. Characteristics of patients are reported in Table 1. Mean age was 73.5 ± 8.2 years, and 43.0% were women.
Patient characteristics according to tertiles of PCSK9 are reported in Table 2. We found a significant difference in the distribution of persistent or permanent AF, beta-blockers, and amiodarone use among tertiles of PCSK9 (Table 2). PCSK9 levels were not correlated with lipid profile in patients treated or not treated with statins (Table 3).
The total number of CVEs registered during the follow-up was 179 (4.6%/year). Of these, there were 39 fatal or nonfatal MIs, 20 cardiac revascularizations, 72 cardiovascular deaths, and 48 cerebrovascular events (41 fatal or nonfatal ischemic strokes and 7 transient ischemic attacks). Patients with CVEs were older and had a higher prevalence of diabetes, heart failure, and previous ischemic events; moreover, they had a higher CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke or transient ischemic attack, vascular disease, age 65 to 74 years, female) score, lower HDL levels, and a higher total cholesterol–HDL ratio (Table 1).
Of the 179 CVEs, 43 (15.3%), 49 (15.5%), and 87 (28.0%) were in the first, second, and third tertile of PCSK9, respectively (log-rank p = 0.009) (Figure 1). Higher median PCSK9 was detected in patients with compared to those without CVEs (1,500 pg/ml [IQR: 1,000 to 2,300 pg/ml] vs. 1,200 pg/ml [IQR: 827 to 1,807 pg/ml]; p < 0.001).
Multivariable Cox regression analysis showed that female sex, age, diabetes, smoking, previous cerebrovascular and cardiac events, digoxin use, being in the third tertile of PCSK9 (vs. the first), and total cholesterol–HDL ratio were all significantly associated with CVEs (Table 4).
Similar findings were observed if PCSK9 was used as a continuous variable (log PCSK9 HR: 1.406; 95% confidence interval [CI]: 1.086 to 1.819; p = 0.010) (Online Table 1). Furthermore, the association between PCSK9 and CVEs remained significant after excluding patients with previous coronary or cerebrovascular events (Online Table 2).
PCSK9 and 11-dh-TxB2
To investigate the potential mechanism accounting for the increased risk of CVEs, we measured the urinary excretion of 11-dh-TxB2 in 852 patients. For this analysis, we excluded 170 patients treated with antiplatelet drugs: 106 received aspirin and 64 other antiplatelet drugs (25 ticlopidine, 19 clopidogrel, 16 lysine acetylsalicylic acid, and 4 indobufen). Patients on aspirin showed a significantly lower urinary excretion of 11-dh-TxB2 than those not taking aspirin (100 ng/mg creatinine [IQR: 50 to 160 ng/mg creatinine] vs. 120 ng/mg creatinine [IQR: 82 to 190 ng/mg creatinine]; p = 0.026).
In the 682 participants not treated with antiplatelet therapy, circulating PCSK9 and 11-dh-TxB2 were significantly correlated (rS = 0.665; p < 0.001). We found a significant increase of urinary 11-dh-TxB2 among groups of PCSK9: 70 ng/mg creatinine (IQR: 50 to 100 ng/mg creatinine) in the first tertile, 135 ng/mg creatinine (IQR: 100 to 179 ng/mg creatinine) in the second tertile, and 200 ng/mg creatinine (IQR: 110 to 374 ng/mg creatinine) in the third tertile (p < 0.001).
At follow-up, patients with CVEs showed higher levels of urinary 11-dh-TxB2 compared with those without CVEs (Table 1). A stepwise multivariable linear regression analysis showed that age and log PCSK9 were the main determinants of urinary log-11-dh-TxB2 (Table 5).
This prospective study demonstrated that high circulating levels of PCSK9 are predictive of CVEs in patients with AF. The significant association between PCSK9 concentration and 11-dh-TxB2 suggested that PCSK9 might contribute to CVEs by increasing, at least in part, platelet activation (Central Illustration).
The secreted protein PCSK9 represents a new target for lowering plasma LDL cholesterol and preventing cardiovascular disease. The relationship between circulating PCSK9 and incident cardiovascular events in the general population was recently investigated in 3 studies, which provided divergent results (10–12). Leander et al. (10), for example, studied a large population of 4,232 younger subjects (mean age 60 years) in a primary prevention setting with 15 years of follow-up. They found a direct correlation between PCSK9 plasma concentration and incident CVEs. In particular, PCSK9 concentration in the highest compared with the lowest quartile was associated with an HR of 1.69, which remained significant after adjustment for typical cardiometabolic risk factors. Conversely, Ridker et al. (11) published a report in which plasma PCSK9 was examined in relation to CVEs in a relatively small, nested, case-control study, including 716 women: 358 with cardiovascular disease and 358 control subjects (free from events at 17-year follow-up). The authors did not observe any association between circulating PCSK9 concentration and CVEs. A third study by Zhu et al. (12) investigated the relationship between PCSK9 concentrations and measures of vascular health, subclinical atherosclerosis, and adverse CVEs, in a cohort of 1,527 middle-aged men enrolled in the FATE (Firefighters and Their Endothelium) study, who were free from vascular disease at enrollment and followed for a mean of 7.2 years. Although PCSK9 correlated with LDL cholesterol, insulin, and triglycerides, it was not associated with either vascular indexes or CVEs. Of note, all 3 studies were performed in relatively young, low-risk subjects, and used different methods to measure PCSK9, making a direct comparison of the results difficult.
The role of PCSK9 also has been investigated in acute cardiovascular settings, such as in patients undergoing PCI. Thus, the LURIC (Ludwigshafen Risk and Cardiovascular Health) study, a prospective evaluation of 2,139 patients referred for coronary angiography, found that PCSK9 was not predictive of cardiovascular mortality (17). Similar results were obtained in patients affected by ACS, in whom PCSK9 plasma levels were associated with inflammatory markers such as C-reactive protein, but did not predict mortality at 1 year (18). On the contrary, a prospective study in patients with ACS undergoing PCI showed that baseline levels of PCSK9 were independently associated with increased ischemic cardiac outcomes at 1-year follow-up (9).
In our study, we investigated PCSK9 levels in patients with AF, who are characterized by an increased risk of cardiovascular and cerebrovascular complications. We found that patients with elevated PCSK9, especially those in the upper tertile of PCSK9 (>1,600 pg/ml), were at high risk for CVEs, after adjustment for common cardiovascular risk factors. The predictive value of PCSK9 was maintained after adjusting for statin use or lipid profile.
A novel finding of our study was the significant association between PCSK9 and urinary excretion of 11-dh-TxB2, suggesting a potential role of platelet activation in the association between PCSK9 and vascular disease. In accordance with this, the HOPE (Heart Outcomes Prevention Evaluation) and CHARISMA (Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance) studies showed that urinary concentrations of 11-dh-TxB2 were predictive of MI or cardiovascular death in aspirin-treated patients (19,20). Although the significant correlation between PCSK9 and urinary 11-dh-TxB2 might suggest a potential role for COX1 activation, other mechanisms should be considered. Thus, PCSK9 might be implicated in platelet activation via expression of oxidized LDL receptors (21) and toll-like receptor 4 (22), which are both involved in platelet activation (23,24). However, direct evidence of the role of PCSK9 on platelet function is still lacking, and interventional trials need to be performed to clarify whether modulation of PCSK9 might also affect platelet function.
The observational design of the study did not allow us to establish a cause-effect relationship between the observed associations. Additionally, the inclusion of only Caucasian elderly patients (the mean age of patients in our study was >70 years) limited the generalizability of the results to other populations. Furthermore, even if urinary excretion of 11-dh-TxB2 is considered a validated marker of platelet activation, it reflects the activity of different enzymatic systems such as COX2, and does not fully correlate with on-treatment platelet reactivity (25,26). Hence, the current data should be considered with caution and needs to be confirmed by studies directly measuring platelet reactivity. Finally, we cannot exclude that systemic inflammation might contribute to PCSK9 circulating concentrations; this is supported by a previous study demonstrating that plasma PCSK9 levels are correlated with markers of inflammation, such as C-reactive protein and fibrinogen (27).
Circulating PCSK9 levels are associated with CVEs in patients with AF, and PCSK9-mediated platelet activation might represent a novel mechanism responsible for the increased cardiovascular risk in this population.
COMPETENCY IN MEDICAL KNOWLEDGE: Soluble PCSK9 is a risk factor for adverse cardiovascular events in AF patients independent of their lipid profile.
TRANSLATIONAL OUTLOOK: The mechanisms linking PCSK9 and urinary 11-dh-TxB2 synthesis deserve further investigation.
The ATHERO-AF Study Group: Mirella Saliola, Marco Antonio Casciaro, Domenico Ferro, Tommasa Vicario, Fabiana Albanese, Francesco Cribari, Alberto Paladino, Francesco Del Sole, Marta Novo, Vittoria Cammisotto, Paola Andreozzi, Tiziana Di Stefano, Patrizia Iannucci, and Elio Sabbatini.
For supplemental tables, please see the online version of this article.
This work was supported by a university grant from Ateneo Sapienza. The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Dr. Pastori and Ms. Nocella contributed equally to this work. Drs. Violi and Pignatelli are joint senior authors on this work.
- Abbreviations and Acronyms
- atrial fibrillation
- cardiovascular event
- high-density lipoprotein
- low-density lipoprotein
- myocardial infarction
- percutaneous coronary intervention
- proprotein convertase subtilisin/kexin type 9
- 11-dehydro-thromboxane B2
- Received July 12, 2017.
- Accepted July 17, 2017.
- 2017 American College of Cardiology Foundation
- Wolf P.A.,
- Abbott R.D.,
- Kannel W.B.
- Violi F.,
- Soliman E.Z.,
- Pignatelli P.,
- et al.
- Minuz P.,
- Patrignani P.,
- Gaino S.,
- et al.
- Varughese G.I.,
- Patel J.V.,
- Tomson J.,
- et al.
- Burchardt P.,
- Rzezniczak J.,
- Dudziak J.,
- et al.
- Navarese E.P.,
- Kolodziejczak M.,
- Winter M.P.,
- et al.
- Leander K.,
- Malarstig A.,
- Van't Hooft F.M.,
- et al.
- Zhu Y.M.,
- Anderson T.J.,
- Sikdar K.,
- et al.
- Mancia G.,
- Fagard R.,
- Narkiewicz K.,
- et al.
- Ryden L.,
- Grant P.J.,
- Anker S.D.,
- et al.
- McMurray J.J.,
- Adamopoulos S.,
- Anker S.D.,
- et al.
- Silbernagel G.,
- Scharnagl H.,
- Kleber M.E.,
- et al.
- Eikelboom J.W.,
- Hirsh J.,
- Weitz J.I.,
- et al.
- Eikelboom J.W.,
- Hankey G.J.,
- Thom J.,
- et al.
- Tang Z.H.,
- Peng J.,
- Ren Z.,
- et al.
- Carnevale R.,
- Bartimoccia S.,
- Nocella C.,
- et al.
- Nocella C.,
- Carnevale R.,
- Bartimoccia S.,
- et al.