Adjustment of the GRACE score by HemoglobinA1c enables a more accurate prediction of long-term major adverse cardiac events in acute coronary syndrome without diabetes undergoing percutaneous coronary intervention

Background The Global Registry of Acute Coronary Events (GRACE) risk score is widely recommended for risk assessment in patients with acute coronary syndrome (ACS). Chronic hyperglycemia [hemoglobinA1c (HbA1c)] can independently predict major adverse cardiac events (MACEs) in patients with ACS. We investigated whether the prediction of MACEs with the GRACE score could be improved with the addition of HbA1c content in ACS patients without diabetes mellitus (DM) undergoing percutaneous coronary intervention (PCI). Methods We enrolled 549 ACS patients without DM who underwent PCI. The GRACE score and HbA1c content were determined on admission. Correlation was analyzed by Spearman’s rank correlation. Cumulative MACE curve was calculated using the Kaplan–Meier method. Multivariate Cox regression was used to identify predictors of MACEs. Additionally, the predictive value of HbA1c content alone and combined with GRACE score was estimated by the area under the receiver-operating characteristic curve (AUC), continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Results During a median of 42.3 months (interquartile range 39.3–44.2 months), 16 (2.9 %) were lost to follow-up, and patients experienced 69 (12.9 %) MACEs: 51 (9.6 %) all-cause deaths and 18 (3.4 %) nonfatal myocardial infarction cases. The GRACE score was positively associated with HbA1c content. Multivariate Cox analysis showed that both GRACE score and HbA1c content were independent predictors of MACEs (hazard ratio 1.030; 95 % CI 1.020–1.040; p < 0.001; 3.530; 95 % CI 1.927–6.466; p < 0.001, respectively). Furthermore, Kaplan–Meier analysis demonstrated increased risk of MACEs with increasing HbA1c content (log-rank 33.906, p < 0.001). Adjustment of the GRACE risk estimate by HbA1c improved the predictive value of the GRACE score [increase in AUC from 0.75 for the GRACE score to 0.80 for the GRACE score plus HbA1c, p = 0.012; IDI = 0.055, p < 0.001; NRI (>0) = 0.70, p < 0.001]. Conclusions HbA1c content is positively associated with GRACE risk score and their combination further improved the risk stratification for ACS patients without DM undergoing PCI. Electronic supplementary material The online version of this article (doi:10.1186/s12933-015-0274-4) contains supplementary material, which is available to authorized users.


Background
Patients with acute coronary syndrome (ACS) are diverse in terms of clinical presentation and risk of death or disability. Accurate management decisions with comprehensive evaluation may improve the outcomes of patients at high risk. To identify high-risk patients, current guidelines recommend a standardized approach involving validated scoring systems such as the Global Registry of Acute Coronary Events (GRACE) score [1][2][3]. Although the GRACE risk score has been validated, the score does not include measurement of important biomarkers. Therefore, whether combining other biomarkers with the GRACE score can provide a more accurate risk estimation in ACS needs to be explored.
Long-term glycometabolic disorder implies high risk for cardiovascular disease [4,5]. Glycosylated hemoglobin (HbA1c) is a well-known biomarker of long-term glycometabolic state and is minimally affected by stress during ACS. Previous research found elevated HbA1c content related to increased risk of cardiovascular events [6][7][8].
In the present study, we investigated the predictive value of HbA1c content and GRACE score individually for major adverse cardiac events (MACEs) in patients with ACS but without diabetes mellitus (DM) undergoing PCI and the potential incremental prognostic value of HbA1c content added to GRACE score.

Study cohort
We performed a single-center, observational study of consecutive non-diabetes patients with ACS performed PCI in the First Affiliated Hospital of Medical College of Xi'an Jiaotong University from December 2010 to December 2011, which included unstable angina, non-ST-segment elevation myocardial infarction (NSTEMI), and ST-segment elevation MI (STEMI).
They all performed PCI using standard techniques after Qualitative and quantitative coronary angiographic analyses. All procedural decisions, including device selection and adjunctive pharmacotherapy, were made at the discretion of experienced interventional cardiologists according to 2007 focused update of the ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention [9]. The diagnostic criteria of DM were: HbA1c ≥6.5 %, FPG ≥7.0 mmol/L, 2-h PG ≥11.1 mmol/L according to 2010 ADA Diagnosis and classification of diabetes mellitus [10].
Exclusion criteria were history of DM, treatment with diabetes drugs, HbA1c content ≥6.5 % on admission, no treatment with PCI, advanced liver disease, renal failure, cancer, valvular heart disease, stroke, idiopathic dilated or hypertrophic cardiomyopathy, peripheral arterial disease, pregnancy, receiving anti-inflammatory drugs, acute or chronic infections or autoimmune disease, and blood or thyroid disease.
The study complied with the Declaration of Helsinki and was approved by the ethics committee of the hospital. Written informed consent was obtained from all patients.

Demographic and clinical data
Main demographic data, cardiovascular risk factors and cardiovascular drugs received were obtained from medical records. Current smokers were defined as having smoked more than 100 cigarettes during their lifetime and still smoking in the past 30 days. Hypertension was defined as resting blood pressure ≥140/90 mmHg at two different visits or receiving hypertension drugs. Previous MI was based on a history of acute MI (AMI) or with signs of an infarction outside the area of the index infarction.

Blood samples
Peripheral blood was sampled from patients in a fasting state the morning following the admission day. Venous plasma concentrations of glucose, lipids, lipoproteins, serum creatinine, N-terminal pro-B-type natriuretic peptide (NT-proBNP), white blood cell, platelet count (PLT), Neutrophile count, Monocyte count and HbA1c content (normal values 4-6 %) were determined in the biochemistry department using standard biochemical techniques for the hospital. We calculated estimated glomerular filtration rate as (ml min −1 1.73 m −2 ) = 194 × (serum creatinine) −1.094 × (age) −0.287 (× 0.739 if female).
They all performed Echocardiography for left ventricle function through left ventricle ejection fraction (LVEF).

Calculation of GRACE risk score
The GRACE risk prediction tool was previously described [11]. The score is derived from several variables (age, heart rate, systolic blood pressure, creatinine level, congestive heart failure, in-hospital percutaneous coronary intervention, in-hospital coronary aortic bypass grafting, history of MI, ST-segment depression, and elevated cardiac enzyme/marker levels) and calculated for each patient. The GRACE risk score was originally designed to predict post-discharge 6 month mortality and had been shown to provide good discrimination of mortality up to 4 years after an ischemic event [12][13][14].

Outcomes and follow-up
All-cause death and nonfatal MI were defined as MACEs. All patients were followed up by interview or telephone in our hospital, and the end of follow-up was the date of the first MACE occurrence obtained by reviewing hospital records. Some patients were followed up until December 2014.

Statistical analysis
Data were analyzed by use of SPSS 19.0 for Windows (SPSS Inc., Chicago, IL, USA). Continuous variables were expressed as mean ± SD. Categorical variables were expressed as frequency and percentage. The Kolmogorov-Smirnov test was used to assess normal distribution of quantitative variables. Independent samples t test was used to compare two groups, and categorical variables were compared by Chi square test. One-way ANOVA was used to compare multiple groups. To limit the influence of extreme observations, NT-proBNP was natural logarithmically transformed to obtain Ln NT-proBNP. Correlation was analyzed by Spearman's rank correlation. Univariate and multivariate survival analyses involved Cox regression analysis. To assess the prognostic value of HbA1c content, Kaplan-Meier survival curves were used. Additionally, the incremental predictive value resulting from adding HbA1c variable to GRACE risk score was analyzed in the validation set using several measures of improvement in discrimination: increase in the area under the receiver-operating characteristic (ROC) curve (AUC), as well as integrated discrimination improvement (IDI), and continuous net reclassification improvement (NRI). DeLong's test was used to compare the AUC from each of models [15], which were analyzed by use of Med-Calc Version 11.4.2.0. The IDI was equal to the increase in discrimination slope defined as the mean difference in predicted risks between those with and without events. The continuous NRI was a non-parametric analogue of the IDI and equals twice the difference in probabilities of upward reclassification for events minus for non-events [16], which were analyzed by use of SAS 9.2 (SAS Institute Inc., Cary, NC, USA).
All probability values were two-tailed. P < 0.05 was considered statistically significant.
All patients were segregated into three groups by tertiles of baseline HbA1c content (≤5.4 %, 5.5-5.8 %, 5.9-6.4 %). Baseline characteristics were shown in Additional file 1: Table S1: Patients in the higher HbA1c levels more often had a prior history of PCI, were lower ejection fraction, eGFR level, as well as higher in Monocyte count and LnNT-ProBNP.

Clinical characteristics of patients with and without MACEs
Patients with MACEs were elderly, had more frequent prior history of hypertension, prior history of MI, prior history of PCI and were higher LnNT-ProBNP, PLT, monocyte count, GRACE score and HbA1c content as well as lower ejection fraction compared to patients without MACEs (Table 1). We carried out the correlation analysis between the GRACE risk score and HbA1c levels as continuous variables and showed that GRACE score was positively correlated with HbA1c content (R = 0.192, p < 0.001).

Effect of HbA1c content and GRACE score combined on MACE occurrence
Since both HbA1c content and GRACE score were independent risk factors of a MACE, we assessed the effect of their combination on predicting long-term risk of MACE occurrence. The AUC increased from 0.75 (95 % CI 0.69-0.82) for GRACE score alone to 0.80 (95 % CI 0.75-0.85) for GRACE score adjustment by HbA1c content (difference in the AUCs, 0.05; z value 2.521, p = 0.012) (Fig. 2). Addition of HbA1c content improved GRACE score alone model discrimination, which was confirmed by the IDI and the continuous, category-free NRI (>0). The IDI for HbA1c content was 0.055 (95 % CI 0.035-0.075, P < 0.001), suggesting further average separation of events from non-events by the HbA1c; the NRI (>0) for HbA1c content was 0.70, (95 % CI 0.47-0.94, P < 0.001), with events contributing 0.42 and non-events 0.28 (Table 4), showing that the HbA1c content led to a significant net reclassification of patients , risk in the appropriate directions.

Discussion
Management decisions in ACS should be based on risk stratification of patients. The GRACE score provides validated prognostic information for MACEs in ACS patients [12,14]. In accordance with previous research results, the GRACE score could independently predict a MACE in our ACS patients without DM undergoing PCI. However, the AUC with the GRACE score alone was only 0.75, which might be due to some potential risk factors not fully captured by the scoring system. Growing epidemiological evidence supports that HbA1c content in the general population and in patients with and without diabetes is an independent risk factor of cardiovascular events including MI [6,8,17]. However, HbA1c content has not been considered with the GRACE scoring Data are mean ± SD or n (%) BMI body mass index, Prior MI prior myocardial infarction, Prior PCI prior percutaneous coronary intervention, SBP systolic blood pressure, DBP diastolic blood pressure, eGFR estimated glomerular filtration rate, FBS fasting blood sugar, TC total cholesterol, TG triglycerides, HDL-C high-density lipoprotein cholesterol, LDL-C low-density lipoprotein cholesterol, Apo-A1 apolipoprotein A1, Apo-B apolipoprotein B, NT-proBNP N-terminal pro-B-type natriuretic peptide, PLT platelets, LVEF left ventricle ejection fraction, WBC white blood cell count, ACEI angiotensin-converting enzyme inhibition, ARB angiotensin receptor blocker, HbA1c hemoglobin A1c, GRACE score Global Registry of Acute Coronary Events (GRACE) score

Variable All patients (n = 533) With MACE (n = 69) Without MACE (n = 464) p value
Age (   reached an impressive 0.70, which suggested that the HbA1c content led to a significant net reclassification of patients , risk in the appropriate directions. In the present study, the long-term event rate seems low compared previous researches. It may attribute to followed reasons. Firstly, the enrolled patients were non-DM, which were lower in the cardiovascular events rate than that in patients with diabetic mellitus, especially in the patients of poor nocturnal glycemic control [18]. Secondly, Recent some reports have shown ethnic differences effect the MACE occurrence rate after PCI not only between African Americans and whites [19,20],but also in Asian subpopulations, such as compared to Indian, Chinese patients have lower Major Adverse Cardiovascular Events [21][22][23]. In addition, our results was comparable compared with the MACE rate documented by Clinical Pathways for Acute Coronary Syndromes in China (CPACS) study [24,25].
Although some research had found that higher level of admission glucose predict a worse prognosis including mortality and MI in ACS patients [26][27][28], recent studies showed that the prediction value of admission glucose was not improved by combining GRACE score [29]. As different from admission glucose, HbA1c is an indicator of general glycometabolic state and is minimally affected by acute stress and also acute glucose management. Increased HbA1c content was not only showed as a indictor of complications including ACS [30] but also a predictor of long-term survival in ACS patients with and without DM [31][32][33]. Whether the HbA1c content in ACS patients without DM undergoing PCI could be an independent predictor of cardiovascular events is uncertain. In our study, HbA1c content at baseline was higher in patients with than without a MACE, and HbA1c content could independently predict a longterm MACE in ACS patients without DM undergoing PCI. Furthermore, we found that the predictive value on MACE of adding HbA1c on the top of GRACE score system.
Researches have found that C-reactive protein [34], NT-proBNP [35], erythrocyte fatty acids [36], growth differentiation factor-15 [37], cystatin C [38], dikkopf [39], RDW/PDW [40], high-sensitivity troponin [41], mean platelet volume(MPV) [42] and other factors add value to a scoring system for predicting adverse cardiovascular events after ACS. We found that GRACE score and baseline HbA1c content were positively correlated, and their combination improved the predictive value. Whether this combination can improve the outcome of ACS patients with DM needs further comprehensive investigation in clinical practice.

Limitations
The number of patients in the cohort was relatively small. Therefore, these findings need to be verified by multicenter and larger cohort studies. In addition, this study was limited to Chinese subjects, so conclusions for other ethnic groups are cautioned.

Conclusions
HbA1c content measured in ACS patients without DM undergoing PCI could predict a MACE and was positively related with GRACE score. The combination of the two factors may improve on risk stratification of ACS patients without DM undergoing PCI.

Additional file
Additional file 1: Table S1. Baseline characteristics of 533 non-diabetes mellitus (DM) patients with acute coronary syndrome (ACS) undergoing Percutaneous Coronary Intervention (PCI) by tertiles of HbA1c content.