Elsevier

Atherosclerosis

Volume 301, May 2020, Pages 30-36
Atherosclerosis

Association of NPAC score with survival after acute myocardial infarction

https://doi.org/10.1016/j.atherosclerosis.2020.03.004Get rights and content

Highlights

  • The NPAC (neutrophil-to-lymphocyte ratio, peripheral vascular disease, age and creatinine) score was developed.

  • This score was significantly associated with survival in acute myocardial infarction patients.

  • The use of a neural network improved the precision of this prediction model.

  • Deep learning algorithms can facilitate clinical decision making.

Abstract

Background and aims

Risk stratification in acute myocardial infarction (AMI) is important for guiding clinical management. Current risk scores are mostly derived from clinical trials with stringent patient selection. We aimed to establish and evaluate a composite scoring system to improve short-term mortality classification after index episodes of AMI, independent of electrocardiography (ECG) pattern, in a large real-world cohort.

Methods

Using electronic health records, patients admitted to our regional teaching hospital (derivation cohort, n = 1747) and an independent tertiary care center (validation cohort, n = 1276), with index acute myocardial infarction between January 2013 and December 2017, as confirmed by principal diagnosis and laboratory findings, were identified retrospectively.

Results

Univariate logistic regression was used as the primary model to identify potential contributors to mortality. Stepwise forward likelihood ratio logistic regression revealed that neutrophil-to-lymphocyte ratio, peripheral vascular disease, age, and serum creatinine (NPAC) were significant for 90-day mortality (Hosmer- Lemeshow test, p = 0.21). Each component of the NPAC score was weighted by beta-coefficients in multivariate analysis. The C-statistic of the NPAC score was 0.75, which was higher than the conventional Charlson's score (C-statistic = 0.63). Judicious application of a deep learning model to our dataset improved the accuracy of classification with a C-statistic of 0.81.

Conclusions

The NPAC score comprises four items from routine laboratory parameters to basic clinical information and can facilitate early identification of cases at risk of short-term mortality following index myocardial infarction. Deep learning model can serve as a gatekeeper to facilitate clinical decision-making.

Introduction

Cardiovascular diseases pose a significant disease burden worldwide and are associated with considerable years of life lost, particularly in low- and middle-income countries [1,2]. Given similar care settings and implementation of international guidelines, the long-term mortality following the first episode of acute myocardial infarction (AMI) was more than three-fold higher in developing countries as compared to European countries [3].

Infarction of the myocardium is characterized by a fluctuation in plasma leukocyte levels as part of the acute phase response [4]. Inflammation is key to its pathogenesis, placing an onus on the discovery and subsequent implementation of inflammatory biomarkers for risk stratification in clinical settings [5]. Following AMI, the absolute neutrophil count rises while the circulating lymphocyte number decreases, leading to an elevation in the plasma neutrophil-to-lymphocyte ratio (NLR) [6]. Activated neutrophils invade the infarcted region and release various mediators that perpetuate the inflammatory process [4,7]. This causes an increase in infarct size and consequently worsens the extent of myocardial ischemia [6], tissue injury and plaque damage [8]. By contrast, the ensuing reduction in plasma lymphocyte count following AMI occurs secondarily to enhanced apoptosis [9], and is probably associated with poorer prognosis in acute ST-elevation MI (STEMI) [10]. To date, a significant body of research has focused on STEMI given the high mortality and the extensive areas of myocardial involvement. However, it is recognized that a subset of non-STEMI patients remain at a high risk of mortality. Yet, risk stratification in NSTEMI can be challenging. A number of risk scores have demonstrated proven utility, but clinicians may be reluctant to use these during clinical practice at bedside because of the time-consuming nature to calculate them [11]. Moreover, recent research has devised increasingly complex scores incorporating different novel biomarkers [12], which may not be readily available in all centers.

With a greater understanding of the mechanistic roles of both neutrophils and lymphocytes in AMI patients, NLR has recently emerged as a cost-effective biomarker for clinical outcomes in several cardiovascular diseases. NLR is a readily accessible biomarker for ongoing inflammation [13]. However, the discriminatory power of neutrophil-to-lymphocyte counts alone for mortality remains sub-optimal [14], leading to the search of its combination with other hematological markers for risk stratification. While increasing the number of variables into a composite scoring system maximizes its predictability, it compromises the simplicity for clinical use.

Our study aim was to determine independent parameters including clinical risk factors and hematological parameters for short-term mortality in incident cases of acute myocardial infarction. A small set of significant parameters were identified for model simplicity. We developed a scoring system based on adjusted odds ratio for easy clinical use. A deep learning model was also developed as a gatekeeper to provide more accurate survival classification to facilitate clinical decision-making.

Section snippets

Study design, population and case identification

This retrospective single center observational study was conducted on consecutive admissions due to a first episode of acute myocardial infarction using the electronic health records registry database of the Prince of Wales Hospital, Hong Kong, between January 2013 and December 2017 (derivation cohort). The center is a university-affiliated teaching hospital with 173,114 discharges and registered in-hospital deaths per year collectively, and the sole trauma center in the territory where 7.3

Baseline characteristics

From 2013 through 2017, 2249 unique cases with principal diagnosis of first episode of acute myocardial infarction were identified. Of these, 122 patients (5.4%) had incomplete laboratory data, and were therefore removed from subsequent analysis. A total of 2127 patients were finally included in this study. The mean age (standard deviation) was 71.08 ± 14.04 years. The majority was male (65.5%). Of note, the proportion of elders (≥65 years) was over-represented in the female group as compared

Discussion

In this large cohort study, we developed and validated a rapid and easy-to- use risk stratification tool for acute myocardial infarction. The timing for percutaneous coronary intervention in the non-STEMI population is important [17]. As such, we developed the four-variable NPAC scoring system for short-term mortality risk stratification in patients experiencing their first episode of AMI, regardless of ST-elevation. The NPAC scoring system consists of age, history of PVD, serum creatinine and

Financial support

Gary Tse was supported by the Croucher Foundation, Hong Kong.

Author contributions

Christien KH Li: Study conception, manuscript drafting, data-analysis and data collection.

Jeffery Ho: Data collection and data analysis.

Zhongzhi Xu: Data-analysis, deep learning model and critical appraisal.

Ishan Lakhani: Data collection and data analysis.

Ying Zhi Liu: Deep learning model and data collection.

George Bazoukis: Critical appraisal and data analysis.

Tong Liu: Critical appraisal and data collection.

Wing Tak Wong: Critical appraisal and data collection.

Shuk Han Cheng: Data collection.

Declaration of competing interest

The authors declared they do not have anything to disclose regarding conflict of interest with respect to this manuscript.

References (30)

  • K.S. Reddy et al.

    Emerging epidemic of cardiovascular disease in developing countries

    Circulation

    (1998)
  • G.A. Roth et al.

    Global and regional patterns in cardiovascular mortality from 1990 to 2013

    Circulation

    (2015)
  • J. Kampfer et al.

    Long-term outcomes after acute myocardial infarction in countries with different socioeconomic environments: an international prospective cohort study

    BMJ open

    (2017)
  • J. He et al.

    Neutrophil-to-lymphocyte ratio (NLR) predicts mortality and adverse-outcomes after ST-segment elevation myocardial infarction in Chinese people

    Int. J. Clin. Exp. Pathol.

    (2014)
  • M.J. Forteza et al.

    Apoptosis and mobilization of lymphocytes to cardiac tissue is associated with myocardial infarction in a reperfused porcine model and infarct size in post-PCI patients

    Oxidative medicine and cellular longevity

    (2018)
  • Cited by (30)

    • Predicting Stroke and Mortality in Mitral Regurgitation: A Machine Learning Approach

      2023, Current Problems in Cardiology
      Citation Excerpt :

      Comprehensive medical data were accessed from the healthcare database (Clinical Management System, CMS) that is linked to a territory-wide Clinical Data Analysis and Reporting System (CDARS) with unique reference identifier for each patient. Our team and other teams have previously used this system for epidemiological studies.20-22 Clinical details including patient age, gender, blood pressure, smoking status, hypertension, diabetes mellitus, hypercholesterolemia, ischemic heart disease were extracted using CDARS.

    • Comparing the Performance of Published Risk Scores in Brugada Syndrome: A Multi-center Cohort Study

      2022, Current Problems in Cardiology
      Citation Excerpt :

      Patients were identified using Clinical Data Analysis and Reporting System, a territory-wide database that centralizes patient information from 43 local hospitals and their associated ambulatory and outpatient facilities to establish comprehensive medical data, including clinical characteristics, disease diagnosis, laboratory results, and medication prescription details. The system has been previously used by local teams in Hong Kong to conduct registry-based and population-based studies,16-18 including rare arrhythmic syndromes.19-22 Diagnosis of BrS was confirmed by G.T. and N.S.M. after reviewing the relevant case notes and ECGs of the patients based on the Expert Consensus Statement proposed in 2017.23

    View all citing articles on Scopus
    1

    These authors contributed equally to this work.

    View full text