This cohort study retrospectively reviewed data available from the SHINANO registry (Shinshu prospective multi-center study of elderly patients with CAD undergoing percutaneous coronary intervention (PCI)) obtained between August 2012 and July 2013. A detailed summary of the methods and design of this registry has been published previously [10]. Briefly, the SHINANO registry is a prospective, multi-center observational registry of patients with any CAD diagnosis, including stable angina, ST-segment elevation MI, non-ST-segment elevation MI, and unstable angina, undergoing PCI at 16 collaborating hospitals located in the Nagano prefecture, Japan. As it is an all-comer registry, there are no exclusion criteria. The institutional review board approved the protocol, which was registered at the University Hospital Medical Information Network (UMIN000010070), and informed consent was obtained from each patient before enrollment. This study was performed in accordance with the Declaration of Helsinki.

Among the 1923 patients registered in the SHINANO registry, we screened 254 patients with a history of HF. After excluding patients with a history of coronary artery bypass grafting (CABG), as well as those with missing data concerning left ventricular ejection fraction (LVEF) or no SYNTAX score, we enrolled 200 patients with prior HF into the final study. All patients were prospectively followed for 12 months from the date of the PCI procedure. The study endpoint was the composite of major adverse cardiovascular events (MACE), including all-cause death, MI, stroke, and hospitalization for worsening HF using a time-to-first-event analysis.

MI was diagnosed according to the American College of Cardiology/American Heart Association (ACC/AHA) guidelines [11]. Stroke was defined as an ischemic cerebrovascular event that persisted for ≥24 h and was diagnosed by a neurologist [12]. Prior HF was based on a previous diagnosis of HF according to the Framingham criteria [13] or a history of hospitalization for worsening HF. Body mass index was defined as weight in kilograms divided by the square of height in meters. Patients with systolic blood pressure >140 mmHg and/or diastolic pressure >90 mmHg and those taking anti-hypertensive agents were considered to have hypertension. Dyslipidemia was defined as a serum total cholesterol concentration ≥220 mg/dL, low-density lipoprotein cholesterol ≥140 mg/dL, or the need for treatment with lipid-lowering agents. Diabetes mellitus was defined as hemoglobin (Hb) A1c ≥6.5 %, random plasma glucose ≥220 mg/dL, or a clinical history of oral hypoglycemic agent and/or insulin use. Patients were considered smokers if they were current smokers. CAD was defined as >50 % stenosis in a coronary vessel on angiography, history of CABG or PCI, or history of MI. Multi-vessel disease was defined as the presence of a ≥75 % lesion in ≥2 major coronary arteries. The SYNTAX score was calculated as previously described [9]. Complete revascularization was considered to have occurred when all stenotic main vessels and all side branches greater than 2 mm in diameter were revascularized [14]. Estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet and Renal Disease equation with coefficients modified for Japanese patients [15]. Chronic kidney disease was defined as an eGFR <60 ml/min/1.73 m². LVEF was calculated using the Teichholz method in patients without regional wall motion abnormality or the biplane Simpson’s method in those with regional wall motion abnormality [16, 17]. All PCI procedures and selection of medical treatments after PCI were at the discretion of the treating physician.

Continuous variables were summarized as mean ± standard deviation if normally distributed or as median and interquartile range otherwise. Normality was evaluated using the Shapiro–Wilk W test. Comparisons of baseline categorical data between the two groups were analyzed using two-sided Chi-squared tests, whereas differences between continuous variables were analyzed using an unpaired t test or the Mann–Whitney U test. The optimal cutoff value for MACE prediction was chosen as the value that maximized sensitivity and specificity on the receiver operating characteristics (ROC) curve. Kaplan–Meier curves were constructed from the date of the PCI procedure to the MACE and were compared using the log-rank test. Cox proportional hazards regression analysis was performed to identify the MACE predictors, using variables that included clinical characteristics and risk factors. Multivariate analysis was performed using all variables with a P value <0.1 in the univariate analysis. A P value <0.05 was considered to indicate statistical significance. All analyses were performed using SPSS version 22.0 (SPSS, Chicago, IL, USA).

All of the patients initially enrolled in the study completed the follow-up. Using the median value of the SYNTAX score (12), patients were divided into a high-SYNTAX score group (n = 100) and a low-SYNTAX group (n = 100) (Fig. 1). During the 12-month follow-up period, adverse events were observed in 39 patients (19.5 %) and included deaths (n = 25), MIs (n = 5), strokes (n = 5), and hospitalizations for HF (n = 18) (Table 1). The baseline clinical characteristics are listed in Table 2. Patients with high SYNTAX scores did not differ from those with low scores with respect to age, sex, the prevalence of hypertension, dyslipidemia, diabetes mellitus, atrial fibrillation, chronic kidney disease, prior stroke, hemodialysis, or peripheral artery disease. Patients with high SYNTAX scores had significantly more severe HF symptoms, as estimated by New York Heart Association (NYHA) functional class, compared with those with low SYNTAX scores. As expected, the prevalence of multi-vessel disease was significantly higher in patients with high SYNTAX scores compared with those with low scores. Complete revascularization was significantly rarer in patients with high SYNTAX scores. The use of angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, beta-blockers, and statins was similar between the two groups. Hb, low-density lipoprotein cholesterol, eGFR, and HbA1c levels were similar between the two groups, whereas levels of brain natriuretic peptide tended to be higher in patients with high SYNTAX scores than in those with low scores. LVEF was significantly lower in patients with high SYNTAX scores than in those with low scores.

In the Kaplan–Meier analysis, patients with high SYNTAX scores (≥12) showed worse prognoses than those with low SYNTAX scores (<12) (26.0 vs. 13.0 %, respectively, P = 0.021) (Fig. 2). After multivariate Cox proportional hazards analysis, which included age, sex, and all variables with P < 0.1 in the univariate analysis, a high SYNTAX score predicted a poor prognosis (Table 3). Based on the SYNTAX scores in our study, the area under the ROC curve (AUC) was 0.63 (Fig. 3) and the optimal cutoff point for predicting adverse events was a SYNTAX score of 11.8 (sensitivity 68.0 %, specificity 56.0 %).

Our study demonstrated that the SYNTAX score is useful for assessing the prognosis of patients with a prior diagnosis of HF undergoing PCI. Given that prior HF patients have a high risk of recurrent cardiovascular events such as sudden death and hospitalization for worsening HF, we recommend that calculation of the SYNTAX score should be performed in prior HF patients with CAD to allow for precise risk stratification for MACE.

The major limitation of our study is that it was an observational study with a relatively small number of subjects and, therefore, the possibility of selection bias and unmeasured confounding factors might not have been completely excluded. Thus, our results should be interpreted cautiously until verified by large-scale multi-center studies. However, our study is the first to report on the utility of the SYNTAX score in prior HF patients with CAD undergoing PCI. Second, our definition of HF relied on an investigator diagnosis based on the Framingham criteria, rather than a requirement to fulfill all of the criteria recommended in the guidelines. Third, we analyzed the predictive value of SYNTAX score using an individual cutoff point and, therefore, further studies are needed to determine the optimal cutoff points in this patient population. Fourth, we did not assess biomarkers which are known predictors of cardiovascular events, despite the fact that certain novel biomarkers are associated with a higher SYNTAX score [26, 27]. Fifth, only prior HF patients diagnosed with CAD were included in this study and, therefore, our results may not apply to all HF patients. Despite these limitations, our findings provide new insight into the risk stratification for cardiovascular events in prior HF patients undergoing PCI. Furthermore, since our study was based on observational registry data from patients with CAD, we consider that our results represent a real-world unselected population of prior HF patients undergoing PCI.