Elsevier

Thrombosis Research

Volume 130, Issue 4, October 2012, Pages 607-611
Thrombosis Research

Regular Article
Eosinophil count predicts mortality following percutaneous coronary intervention

https://doi.org/10.1016/j.thromres.2012.05.033Get rights and content

Abstract

Introduction

Several inflammatory markers have been shown to be independent predictors for both the development of clinically significant atherosclerosis and for adverse outcome in patients with symptomatic coronary artery disease (CAD). We investigated the prognostic role of eosinophil count in low to intermediate risk patients with CAD.

Methods

We studied 909 patients admitted for elective or urgent percutaneous coronary intervention (PCI) from April 2002 to December 2004, and measured pre-procedural total and differential white blood cell (WBC) counts. Inter-tertile WBC differences in short (6 months) and long term (up to 74 months) mortality were analysed after adjusting for differences in baseline characteristics.

Results

Over a median period of 54 months (inter-quartile range 47–65), a total of 138 deaths (15.2%) occurred, of which 24 were in the first 6 months of follow-up. Cox regression analysis showed that high pre-procedural eosinophil count (top tertile) was associated with improved outcome within the first 6 months (OR = 0.23 [0.06–0.84]; p = 0.03) but after this period there was an increased risk of mortality (OR = 2.21, [1.26–3.88]; p = 0.006).

Conclusions

Eosinophil count is a novel biomarker for risk stratification of CAD patients, which was associated initially with reduced mortality, but after 6 months with increased mortality.

Introduction

Risk stratification in symptomatic coronary artery disease (CAD) remains difficult despite well-validated traditional cardiovascular risk factors. Several inflammatory markers have been shown to be independent predictors for both the development of clinically significant atherosclerosis and for adverse outcome in patients with symptomatic CAD [1], [2], [3]. Of the inflammatory markers associated with adverse outcome, high sensitivity C-reactive protein (hs-CRP) has been extensively investigated. However, hs-CRP remains a research tool and is not routinely used clinically in risk stratification in part due to its high cost. In contrast, total white blood cell (WBC) count, a simple marker of inflammation, is associated with long term mortality following percutaneous coronary intervention (PCI) in unstable angina, non-ST elevation myocardial infarction (MI) and ST elevation MI [4], [5], [6].

Total WBC count comprises several cell-types implicated in the development, progression and instability of atherosclerotic plaque. The role of monocytes and T-cell lymphocytes in the formation of atherosclerotic plaques has been well described [7], [8]. Neutrophils are associated with a number of pathological processes including disruption of atherosclerotic plaques, increased neutrophil-platelet adhesion leading to vascular plugging [9], [10], and neutrophil count has been shown to be an independent risk factor for the development of symptomatic CAD [11]. Furthermore, a relative neutrophilia or lymphopenia is associated with increased risk of long term mortality in symptomatic CAD [12]. Indeed, Duffy et al. showed a greater neutrophil to lymphocyte (N/L) ratio to be associated with long-term mortality following revascularisation with PCI [13]. Several prospective studies have also reported an association of eosinophil count with the incidence of CAD [14], [15]. Patients with unstable angina have higher eosinophils counts than healthy controls [16].

In this study, we hypothesised that pre-procedural eosinophil count is a marker of risk for all-cause mortality following PCI. We tested this hypothesis in a large prospective study of patients undergoing PCI at our large tertiary referral teaching hospital centre.

A total of 1425 consecutive patients with pre-procedural WBC count data underwent PCI between 1st April 2002 and 31st December 2004 at a large tertiary referral centre (City Hospital, Birmingham, UK). To ensure that this study would be focused on low to intermediate risk patients undergoing PCI, the following exclusion criteria were applied: (i) PCI for an acute ST-elevation MI (n = 172), (ii) significant haemodynamic instability pre-PCI (i.e. patients requiring intra-aortic balloon pump support or with impending cardiogenic shock) (n = 36), (iii) PCI for chronic total occlusions or saphenous vein graft intervention (n = 137) and (iv) PCI of a restenosis lesion (n = 64), (v) failure of PCI procedure (n = 107).

Based on ACC/AHA guidelines low risk patient were those with stable angina undergoing elective procedures and, stable patients presenting with acute coronary syndrome who required urgent procedures were defined as intermediate risk [17], [18]. Following these exclusion criteria, the remaining 909 patients were included in the study.

Patient demographics, pre-intervention clinical status and procedural details were obtained from the City Hospital British Cardiovascular Interventional Society (BCIS) database [19]. Clinical characteristics recorded included age, gender, history of diabetes mellitus, hypertension, smoking, peripheral vascular disease, myocardial infarction and stroke. Patients with diabetes mellitus and hypertension were classified using documented previous clinical diagnosis. Patients who had never smoked or stopped greater than 10 years ago were classified as non-smokers.

Pre-procedural WBC counts (within 24 hours of the procedure) were measured using flow cytometry by an automated system (Advia120, Bayer Diagnostics, USA). These indices were obtained from histograms of two-dimensional light scatter signals that were converted into WBC count and sub-types. The intra- and inter-assay coefficients of variation were < 3% and < 7% respectively. Patients were divided into tertiles based on their pre-procedural differential WBC counts (first tertile was lower and the 3rd tertile was higher differential WBC count) in order to facilitate categorization of patients into low, intermediate, and high-risk groups. Reference ranges for differential WBC counts are shown in Table 1.

Angiographic details recorded included number of diseased vessels and the target artery. Multi-vessel disease was defined as a > 50% lesion in ≥ 2 major coronary arteries (left anterior descending, circumflex, or right coronary) and/or a left main coronary artery lesion. American College of Cardiology (ACC) grading score was used to assess coronary artery plaques (A or B1 vs. B2 or C) and was used as a marker of target lesion complexity. Stent size and type was also recorded. All-cause mortality from index PCI to May 2008 was determined from the Office of National Statistics Central Registry, UK [20]. Local ethics approval was obtained for this study from the South West Birmingham Ethics Committee Board.

Categorical variables are expressed as frequencies (percent), and continuous variables are expressed as mean ± standard deviation or as median (interquartile range) for continuous variable with a non-normal distribution. For categorical variables, differences between groups were assessed using the Pearson chi-square test or 2-tailed Fisher's exact test. Odds ratios (OR) are reported with 95% confidence intervals (CI). Continuous variables with a normal distribution were analyzed using Student t test, and variables with a non-normal distribution were analyzed using Wilcoxon rank test.

Using a stepwise approach, Cox regression analysis of mortality predictors was performed after adjusting for age, male gender, diabetes, hypertension, peripheral vascular disease, prior MI, stroke, family history of CAD, multi-vessel disease, ACC grade, procedural urgency, drug eluting stent and tertiles of pre-procedural total WBC, neutrophil, lymphocyte, eosinophil, basophil count and N/L ratio. Variables were entered into the regression model based on univariate association (p < 0.1) with the dependent variable. Kaplan-Meier survival analysis was used to compare differences in mortality between tertiles of pre-procedural WBC differential count using the log-rank test.

To evaluate predictors of early and late mortality, a 'landmark analysis' using the landmark time point of 6 months following the index procedure was performed [21], [22]. The time point was set at 6 months, as during the study period patients at our institution were routinely prescribed dual anti-platelet therapy (clopidogrel in addition to aspirin) for up to 6 months following PCI. In the landmark analysis, patients who were event-free (no death) at 6 months were stratified into tertiles of pre-procedural total and differential WBC counts. Outcome for these groups was evaluated up to 73 months after the index PCI procedure. Statistical analysis was completed using SPSS 14.0 (SPSS Inc., Chicago, Illinois) for Windows. A p-value of ≤ 0.05 was considered statistically significant.

The mean age of the study population was 64 ± 11 years, 650 (71.6%) were male, 212 (23.4%) had diabetes mellitus, 308 (33.9%) were current smokers and 453 (49.9%) had prior history of MI. Five hundred and sixty procedures (61.6%) were performed during an acute admission. Drug eluting stents were placed in 177 cases (19.5%).

Median (interquartile range) baseline total and differential WBC counts for the entire study population, pre-procedural total and differential WBC counts were divided into tertiles, as shown in Table 1. The clinical and procedural data by tertiles of pre-procedural total WBC count are shown in Table 2. Patients in the third tertile were younger, and more likely to be a smoker, and to have a history of previous MI. However, target lesion complexity (as judged by ACC grade) was similar across the tertiles.

In the first six months of follow up, there were 24 deaths (2.6%). Mortality during the initial 6 months, was associated with total WBC count (third vs. first tertile; OR 4.75 [1.34–16.84], p = 0.01), neutrophil count (third vs. first tertile; OR 5.71 [1.63–19.62], p = 0.002), N/L ratio (third vs. first tertile; OR 7.92 [1.80–34.75], p = 0.001), monocyte count (third vs. first tertile; OR 3.54 [0.94–12.70], p = 0.053) and eosinophil count (third vs. first tertile; OR 0.28 [0.08–0.99]; p = 0.04) on univariate analysis.

Between 6 months and end of study period, N/L ratio (third vs. first tertile; OR 2.31 [1.34–3.99]; p = 0.002), monocyte (third vs. first tertile; OR 1.69 [0.99–2.90]; p = 0.06) and eosinophil count (third vs. first tertile; OR 1.80 [1.04–2.96], p = 0.05) were associated with long-term mortality.

Over a median follow-up period of 54 months (inter-quartile range 47–65) there were a total of 138 deaths (15.2%). Univariate analysis showed an increased risk of mortality associated with total WBC count (third vs. first tertile; OR 1.65 [1.02–2.68]; p = 0.04), neutrophil count (third vs. first tertile; OR 1.98 [1.18–3.30]; p = 0.007) and monocyte count (third vs. first tertile; OR 2.03 [1.18–3.22]; p = 0.009). In contrast a reduced risk of mortality was associated with pre-procedural lymphocyte count (third vs. first tertile; OR 0.57 [0.35–0.91]; p = 0.03). Consequently, neutrophil to lymphocyte count (N/L) ratio was associated with an increased risk of mortality (third vs. first tertile; OR 2.98 [1.78–4.97]; p < 0.0001).

Over the entire study period (73 months), Cox regression analysis showed pre-procedural N/L ratio (third vs. first tertile; OR 2.21 [1.29–3.77]; p = 0.004) and monocyte count (third vs. first tertile; OR 2.02 [1.18–3.40]; p = 0.01) were both associated with increased risk of adverse outcome (Table 3). In addition to N/L ratio, eosinophil count (third vs. first tertile; OR = 0.23 [0.06–0.84], p = 0.03) was an independent predictor of short term outcome and associated with improved outcome within the first 6 months (Table 4, Fig. 1). After the initial 6 months, Cox regression analysis showed eosinophil count to be associated with an increased risk of mortality (third vs. first tertile; OR 2.21 [1.26–3.88], p = 0.006) following PCI (Table 5). Pre-procedural Basophil count was not associated with long-term mortality.

In low to intermediate risk patients with CAD, we have shown that eosinophil count in the third tertile was associated with a reduced risk of mortality at 6 months following PCI. On long-term follow-up eosinophil count in the third tertile was associated with an increased risk of mortality after the first 6 months following PCI. Neutrophil to lymphocyte ratio and monocyte count in the third tertiles were also associated with an increased risk of mortality. The association between eosinophil count and mortality has not previously been described.

An emerging role for eosinophils is regulation of the inflammatory response. PCI has consistently been shown to produce a significant inflammatory response resulting in the elevation of post-procedural IL-6 and CRP [23], [24]. Regulation of the inflammatory response may be protective by limiting additional myocardial tissue injury following PCI.

The immunosuppressive cytokines IL-10, IL-4, IL-13 and TGF-Beta1 are implicated in the resolution of acute inflammation and form part of the alternatively activated immune response [25], [26], [27]. Eosinophils are able to express IL-10, IL-4, IL-13 and are a significant source of TGF-Beta1, which suggests they may be able to modulating the acute phase and innate inflammatory response [28]. The activity of T regulatory cells is also directly affected by these cytokines, and in addition eosinophil derived indoleamine 2,3,-dioxygenase lowers the viability or activation of T cells [29]. Clarifying whether eosinophils have a regulatory role in the resolution of acute inflammation following PCI could help uncover new potential therapeutic targets.

In contrast to the association with short-term outcome, eosinophil count in the third tertile was associated with an increased risk of all-cause long-term mortality after the initial 6 months following PCI. This may be due to an increased risk of thrombus formation as eosinophils infiltrate the site of stent implantation and release a number of mediators which can increase platelet activation and aggregation [30]. The increased risk of thrombus formation may have initially been counteracted by dual anti-platelet therapy, which in our institution is continued for 4 to 6 months after PCI.

A relative neutrophilia or lymphopenia have both previously been associated with increased long-term mortality in symptomatic CAD [12]. Studies have also shown N/L ratio to be a stronger predictor of adverse outcome, compared to either total WBC count or differential WBC count [31], [32]. Indeed, Duffy et al. reported that patients with pre-procedural N/L ratio in the third tertile had nearly a two-fold increased risk of long-term mortality following PCI [13]. The period following PCI is associated with activation of an inflammatory response. In the activated state, neutrophils, together with platelets, form aggregates and adhere to endothelial cells leading to microvessel plugs and progression of myocardial tissue ischaemia, which may contribute to the steep increase in short-term risk of mortality [9], [10]. The long-term prognostic significance of neutrophils may be mediated through their role in atherosclerotic plaque disruption through the release of reactive species and proteolytic enzymes [33], [34], [35].

The association between monocyte count and mortality is most likely to be a reflection of the central role of monocytes in the formation and de-formation (rupture) of atherosclerotic lesions rather than biological response to PCI. Monocyte-derived plaque macrophages secrete a number of matrix-degrading enzymes which weaken the fibrous cap and so increase the risk of plaque rupture [36]. Thus, monocytes may accelerate the progression to rupture of atherosclerotic plaques present within the coronary system which were not significant at the time of the index procedure. This could explain the late acceleration of mortality risk in the Kaplan-Meier curve of long-term mortality for patients with pre-procedural monocyte count in the third tertile.

WBC subtype counts are readily available and reproducible prognostic biomarkers which could be used to improve the risk stratification of CAD patients due to undergo PCI. Current prognostic models are based on study population which exclude low risk patients despite them forming the majority of patients undergoing PCI. This study suggests that in addition to traditional clinical variables, WBC subtype counts have a significant prognostic role. Accurate risk stratification would enhance a clinician's ability to accurately assess a patient's cardiovascular risk and enable the appropriate use of therapeutic measures. Risk stratification with WBC subtype count could help further reduce medium to long-term mortality following PCI through identifying patients who require more aggressive implementation of secondary prevention measures including lifestyle and pharmacological strategies.

A pre-procedural prognostic marker could also be used to identify patients who may benefit from the use of drug eluting stents. Risk stratification of long-term mortality using pre-procedural variables such as WBC subtype count would be especially useful in patients with stable angina. These patients commonly present via their general practitioner who will not have access to the angiographic risk markers described in many of the current risk stratification models.

This was an observational study with inherent limitations and several confounding variables. Whilst a single baseline measurement of total and differential WBC count was used, it is plausible that serial measurement with assays of cellular activation may yield a different picture. It was not possible to accurately address whether pre-procedural WBC count was confounded by burden of CAD. However differential WBC counts were independent predictors of outcome after controlling for angiographic markers of CAD severity and previous MI. This would suggest that WBC count may assist clinical decision making by providing prognostic information in addition to currently used markers of CAD severity. Smoking affects WBC count with a positive dose–response relationship between the two. Our study was not powered to quantify the effect of smoking. With regards to medical therapy, statins have a proven benefit in patients with CAD, in addition to their anti-inflammatory effect. The concordance of our patient population with regular use of statins was not recorded. It was not possible to differentiate the cause of death into cardiac and non-cardiac events from the mortality data collected from the Office of National Statistics, UK [16].

This study confirmed WBC subtype counts to be independent predictors of adverse outcome following PCI. Eosinophil count is a novel biomarker for risk stratification of CAD patients at increased risk of adverse outcome. The association of eosinophils count initially with reduced mortality, but after 6 months with increased mortality is a particular interesting finding. This suggests a dynamic role for eosinophils in the natural history of CAD patients which has not been previously reported.

  • Eosinophil count has been associated with the incidence of coronary artery disease.

  • Pre-procedural eosinophil count predicts all-cause mortality following percutaneous coronary intervention.

  • A possible role for eosinophils in regulation of acute inflammation may be a future potential therapeutic target in coronary artery disease.

None.

References (36)

  • J.L. Dinerman et al.

    Increased neutrophil elastase release in unstable angina pectoris and acute myocardial infarction

    J Am Coll Cardiol

    (1990)
  • L.M. Biasucci et al.

    Elevated levels of C-reactive protein at discharge in patients with unstable angina predict recurrent instability

    Circulation

    (1999)
  • J. Danesh et al.

    Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies

    JAMA

    (1998)
  • J.W. Yarnell et al.

    Fibrinogen, viscosity, and white blood cell count are major risk factors for ischemic heart disease. The Caerphilly and Speedwell collaborative heart disease studies

    Circulation

    (1991)
  • H.V. Barron et al.

    Association between white blood cell count, epicardial blood flow, myocardial perfusion, and clinical outcomes in the setting of acute myocardial infarction: a thrombolysis in myocardial infarction 10 substudy

    Circulation

    (2000)
  • A.C. van der Wal et al.

    Site of intimal rupture or erosion of thrombosed coronary atherosclerotic plaques is characterized by an inflammatory process irrespective of the dominant plaque morphology

    Circulation

    (1994)
  • P. Libby

    Inflammation in atherosclerosis

    Nature

    (2002)
  • R.L. Engler et al.

    Leukocyte capillary plugging in myocardial ischemia and reperfusion in the dog

    Am J Pathol

    (1983)
  • Cited by (33)

    • Eosinophil Deficiency Promotes Aberrant Repair and Adverse Remodeling Following Acute Myocardial Infarction

      2020, JACC: Basic to Translational Science
      Citation Excerpt :

      Inflammation is essential for infarct repair but when excessive or prolonged is associated with adverse cardiac remodeling (29). In other settings, eosinophil-derived proresolving lipid mediators can reduce the numbers of neutrophils in inflamed tissue (3,4) by counter-regulating neutrophil influx and stimulating macrophage phagocytosis of apoptotic neutrophils (5), thus promoting the resolution of inflammation. Neutrophil recruitment to the heart following MI was increased in eosinophil-deficient ΔdblGATA mice, and this was associated with increased expression of proinflammatory mediators, including neutrophil-recruiting chemokines.

    • Asthma Attacks and Cardiovascular Events: Earthquakes and Tsunamis

      2020, Journal of Allergy and Clinical Immunology: In Practice
    • Eosinophils count in peripheral circulation is associated with coronary artery disease

      2019, Atherosclerosis
      Citation Excerpt :

      Although there are few eosinophils present in atherosclerotic plaque, studies have shown that eosinophils may promote plaque progression in coronary arteries, leading to cardiovascular events via the release of eosinophils’ mediators [7], and are involved in several steps leading to plaque progression and acute thrombotic events [8]. Clinical studies have also found that high eosinophil levels are associated with increased mortality rates [9]. In an autopsy study assessing patients who died of acute myocardial infarctions, the eosinophil count in the inflammatory response was significantly higher in hearts that suffered cardiac ruptures than in those without cardiac ruptures [10].

    • Circulating CD14<sup>+</sup>CD16<sup>−</sup> classical monocytes do not associate with a vulnerable plaque phenotype, and do not predict secondary events in severe atherosclerotic patients

      2019, Journal of Molecular and Cellular Cardiology
      Citation Excerpt :

      However, existing data regarding the prognostic value of circulating monocytes remain inconclusive. Some studies indicate that monocytes have no prognostic value [16–19], whereas other studies suggest that high monocyte counts associate with increased risk for recurrent CVD events or mortality [20–23]. This may partly be explained by the fact that different monocyte subsets exist with distinct plaque modifying properties [3,4,24,25].

    View all citing articles on Scopus
    View full text