Prognostic value of the SYNTAX score on myocardial injury and salvage in STEMI patients after primary percutaneous coronary intervention: a single-center retrospective observational study

This single-center retrospective observational study enrolled STEMI patients who underwent primary PCI at Cangzhou central hospital (tertiary care hospital) between October 2018 and September 2020. Inclusion criteria were as follows: (1) electrocardiography (ECG) features consistent with acute STEMI in accordance to Fourth Universal Definition of Myocardial Infarction [1]; (2) time duration within 12 h from typical chest pain to primary PCI; (3) CMR within 3–10 days from symptom onset. In addition, patients were excluded for the following reasons: (1) old myocardial infarction history; (2) poor image quality; (3) declined CMR for personal reasons. The present study was approved by the ethical committee of Cangzhou Central Hospital (2020-286-01). The requirement for informed consent was waived by the ethical committee of Cangzhou Central Hospital due to the retrospective nature of the study.

Primary PCI was performed based on current guidelines [21]. Patients were administered a loading dosage of aspirin 300 mg (Bayer, 100 mg) and clopidogrel 300 mg (Plavix, 75 mg) or ticagrelor 180 mg (BRILINTA, 90 mg) before coronary angiography and PCI was performed via radial access using a 6-F arterial sheath. After access was established, 2500 U of unfractionated heparin was injected through the sheath. Coronary angiography was performed using a 6-F IL 3.5 multifunctional guiding catheter (Terumo, Tokyo, Japan). ECG and invasive blood pressure monitoring were continuously performed during the surgery. For PCI, additional heparin (total 100 U/kg, including the previous dose) was routinely administered through the arterial sheath. The guiding catheter was deployed at the ostium of the coronary artery for diagnostic angiography. A revascularization strategy (balloon angioplasty only or stenting) was chosen based on the surgeon’s decision. After the procedure, patients were administered a maintenance dosage of aspirin 100 mg daily and ticagrelor 90 mg twice a day or clopidogrel 75 mg once a day. Additional medications, such as statins, β-receptor blockers, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, glycoprotein IIb/IIIa inhibitors, and low-molecular-weight heparin, were administered based on the recommended guidelines unless contraindicated. Pre- and post-PCI antegrade flow in the infarct-related artery was characterized using the TIMI system.

The SS for each patient was calculated retrospectively, based on scoring for all coronary lesions with diameter stenosis > 50%, and vessels with diameter > 1.5 mm, using the SS algorithm, as recommended [14, 16]. The calculation was done using an openly accessible web-based score calculator (http://​www.​syntaxscore.​com). Based on the previous reports [22, 23] SS was categorized as low (≤ 22), intermediate (22 < SYNTAX score ≤ 32), or high (≥ 33), and patients were divided into low SS and mediate-high SS groups. The SYNTAX score was assessed visually by two experienced interventional cardiologists trained to perform SYNTAX score assessments and blinded to the treatment and CMR data. Every significant inter-observer difference required a recalculation.

Blood samples were taken immediately after hospitalization, and echocardiography was performed on 5 (3–7) days from symptom onset. CMR was performed on 7.15 (5.26–8.24) days from symptom onset using a 3.0-T scanner (GE Discovery MR750w; GE Healthcare, Milwaukee, WI, USA) with electrocardiographic-gated image acquisition, as described in previous studies [24, 25]. MRI parameters were measured on short-axis images covering the entire left ventricle (8-/0-mm slice thickness/slice gap) with the following sequences: a steady-state free precession (SSFP) cine sequence to determine the left ventricular (LV) function, mass and volume, and a short-tau inversion recovery T2-weighted (T2-STIR) sequence to determine the area at risk (AAR) of myocardial infarction. Late gadolinium enhancement (LGE) images were acquired approximately 10 ~ 15 min after the intravenous administration of gadolinium-based contrast medium (0.2 mmol/kg, Magnevist, gadopentetate dimeglumine injection, Bayer) to determine the IS.

Analysis was performed using dedicated software (cmr42 version 5.11.3, Circle Cardiovascular Imaging, Calgary, Alberta, Canada). Images were anonymized, batched, and analyzed in a blinded fashion by two experienced operators. The AAR was defined as a high-signal myocardial edema mass/LV mass ratio. The IS was defined as the hyperenhanced myocardium on the LGE images and is expressed as the infarcted LV mass/LV mass ratio. MVO was defined as dark areas surrounded by hyperenhanced myocardium on the LGE images. The presence of IMH was defined as hypointense areas within the brighter edematous zone on T2-STIR images. The papillary muscles were included in the LV cavity volume. The regions of interest for the volumes of AAR, IMH, IS, and MVO were created by manually drawing the lesion contours. In contrast, the LV volume was calculated by the semiautomated drawing of endocardial and epicardial contours on each slice for the whole LV myocardium. Myocardial edema was described as areas with a signal intensity > 5 standard deviations (SD) that of remote normal myocardium. The IS was calculated using the > 5 SD method. Discordant cases were reviewed and reconciled with superior imaging specialists. The myocardial salvage index (MSI) was defined as (AAR − IS)/AAR × 100.

For echocardiographic evaluation, Xcelera R4.1 (Philips Medical Systems) was used. Echocardiographic morphological and functional parameters were measured according to the definitions and recommendations from the available American Society of Echocardiography guidelines [26].

Categorical data are presented as numbers and percentages, and continuous data are presented as the mean (standard deviation) or median (interquartile range). The Shapiro–Wilk test was applied to assess for data normality. Continuous variables with normal distributions were compared using the t-test. Continuous parameters that were not normally distributed were compared using the Mann–Whitney test. Categorical variables were compared using the chi-squared test (Fisher’s exact test when the expected value was < 5). Spearman’s correlation coefficient was used to analyze associations between SS and CMR or echocardiographic parameters. Independent predictors of high IS and low MSI were determined in a multivariable binary logistic regression model using the backward stepwise method adjusted for all baseline variables found significant (p < 0.05) in the univariable logistic regression model. Receiver operating characteristic (ROC) curves were generated to determine the usefulness of SS to discriminate high IS and low MSI. All statistical analyses were performed using SPSS version 20 (IBM, Armonk, NY, USA) and graphs were generated using GraphPad Prism version 8.0 (GraphPad Software, La Jolla, CA, USA). P-value of < 0.05 was regarded as statistically significant.

A total of 215 AMI patients who underwent primary PCI within 12 h from symptom onset were retrospectively included. Out of the 39 with no CMR data, 24 with old myocardial infarction, and three with poor image quality were excluded. Finally, 149 patients (mean age 59.89 ± 10.86 years, 72.5% male) were enrolled for analysis (Fig. 1 and Table 1). Examples of obtained arteriographic results are illustrated in Fig. 2. According to SS (17[9–25]), patients were divided into low SS (≤ 22, n = 96) and mediate-high SS (> 22, n = 53) groups. Differences in baseline characteristics between the two groups are shown in Table 1. Compared with the low SS group, patients in the mediate-high SS group were older (58.32 ± 11.47 vs. 62.72 ± 9.09, p = 0.011), and had a lower BMI (26.25 ± 3.62 vs. 24.92 ± 2.86, p = 0.023). The prevalence of multivessel disease (p < 0.001), initial TIMI thrombus grade 4/5 (p = 0.002), initial TIMI flow grade 0/1 (p = 0.008), and APOB (p = 0.038) were higher but creatinine clearance rate (p = 0.016) was lower in the mediate-high SS group. There were no differences between groups concerning medical history, culprit lesion, mode of reperfusion, and medications after surgery (all p > 0.05).

All transthoracic echocardiography (TTE) parameters were comparable between the two groups (Table 2). Whereas CMR showed a lower left ventricular ejection fraction (LVEF, 52.86 ± 13.45 vs. 45.65 ± 12.01%, p = 0.001), MSI (30.30[19.23–45.41] vs. 24.52[13.92–35.69] %, p = 0.016) and higher left ventricular end-systolic volume (LVESV, 51.28[38.23,72.17] vs. 63.96[51.34–80.38] mL, p = 0.029), and IS (33.93 ± 11.94 vs. 38.06 ± 9.40%LV, p = 0.029) in mediate-high SS group patients when compared with those in low SS group. There was no significant difference between the two groups in terms of MVO, IMH and AAR (Table 2 and Fig. 3).

Correlations between baseline variables (SS, CK, initial TIMI flow, GRACE score, CRUSADE score) and CMR parameters by Spearman correlation analysis are presented in Table 3. Among them, SS (r = 0.292, p < 0.001) and CK (r = 0.328, p < 0.001) showed relative stronger correlation with IS (%LV) than other variables. Also, SS (r = -0.314, p < 0.001) and initial TIMI flow (r = 0.216, p = 0.011) were more correlated with MSI. Therefore, SS had a positive correlation with IS (%LV) and a negative correlation with MSI (Fig. 4).

Univariable analysis revealed that nine variables (all p < 0.05) were significantly associated with an increased risk of high IS (≥ mean 35.43), including initial heart rate, pro-BNP, CK, CKMB, LVEF-TTE, left ventricular end-diastolic diameter (LVEDD), infarct-related artery LAD, SS > 22 (as a categorical variable) and SS (as a continuous variable); LVEF-TTE, baseline TIMI flow 0/1 and SS (not as a categorical variable) were significant predictors for risk of low MSI (≤ median 28.01) (Table 4). After adjustment for other potential confounders including initial heart rate, pro-BNP, CK-MB, LVEDD-TTE, infarct-related artery LAD, multivariable analysis showed that LVEF-TTE (OR = 0.913, 95%CI [0.865–0.962], p = 0.001) and SS > 22 (OR = 2.245, 95%CI [1.002–5.053], p = 0.048) were significant for high IS in model 1; in model 2, LVEF-TTE (OR = 0.914, 95%CI [0.866–0.965], p = 0.001) and SS (as continuous variable) (OR = 1.053, 95%CI [1.014–1.095], p = 0.008) were also the independent predictors of high IS; for low MSI in model 3, only LVEF-TTE (OR = 0.924, 95%CI [0.881–0.970], p = 0.001) was the independent predictor. Among the three models, the SS had higher odd ratio than LVEF-TTE (Fig. 5).

The ROC analysis results of SS predicting high IS (≥ mean 35.43) and low MSI (≤ median 28.01) are shown in Fig. 6, with AUC of 0.664 (0.577–0.751, p = 0.001), 0.610 (0.516–0.700, p = 0.026), respectively. At the optimal cutoff point of SS (10.8 for high IS, 10.3 for low MSI), the sensitivity and specificity were similar for high IS diagnosis (81.8%, 45.8%) and for low MSI (81.2%, 40.6%).