Prevalence and in-hospital outcomes of diabetes among patients with acute coronary syndrome in China: findings from the Improving Care for Cardiovascular Disease in China-Acute Coronary Syndrome Project

The CCC-ACS project, a nationwide registry and quality improvement study with an ongoing database focusing on quality of ACS care, was launched in 2014 as a collaborative initiative of the American Heart Association and the Chinese Society of Cardiology. In Phases I and II of the project, only the tertiary hospitals were included, 150 centers representing the diversity of care for ACS in tertiary hospitals across China. Since July 2017, Phase III of the project has extended into secondary hospitals. The data for this study are based on Phases I and II of the project. Details of the design and methodology of the CCC project have been published [8]. A standard web-based data collection platform (Oracle Clinical Remote Data Capture, Oracle) was used in this study. Trained data abstractors in the participating hospitals reported the required data, which they abstracted from the patients’ original medical records. Eligible patients were consecutively reported to the CCC-ACS database for each month before the middle of the following month. Third-party clinical research associates performed quality audits to ensure that cases were reported consecutively rather than selectively. In addition, about 5% of reported cases were randomly selected, and the reported information was compared with the original medical records as a quality assessment and a method to promote accuracy and completeness of the reported data. According to the quality audit reports, the data in this study were appropriately reported with an accuracy rate greater than 95%.

Based on principal discharge diagnosis, 63,641 inpatients with ACS were registered between November 2014 and June 2017 from 150 hospitals. Of these, 63,450 inpatients were included in this study after excluding 191 (0.3%) patients with incomplete demographic information. The flow chart for study population recruitment can be found in Additional file 1: Figure S1.

Diabetes was defined according to one of the following criteria: (1) a self-reported diabetes which was previously diagnosed by physicians or use of glucose-lowering drugs before hospitalization; (2) diabetes listed in the medical records as the secondary discharge diagnosis; (3) glycated hemoglobin A1c (HbA1c) concentration ≥ 6.5%.

Possible diabetes was defined in ACS patients with level of fasting blood glucose (FBG) ≥ 7.0 mmol/L but without measurement of HbA1c, as we could not distinguish between undiagnosed diabetes and stress hyperglycemia in this group of patients by the results of FBG alone.

The outcomes of this study included all-cause deaths and MACCEs that occurred during hospitalization. MACCEs were defined as a combination of cardiac death, recurrent myocardial infarction, stent thrombosis, and stroke. All of these outcomes were diagnosed by doctors during patients’ hospitalization and recorded in medical records.

Hypertension was defined as having a history of hypertension, receiving antihypertensive therapy, or systolic blood pressure (SBP) ≥ 140 mmHg or diastolic blood pressure (DBP) ≥ 90 mmHg at admission. Elevated low-density lipoprotein cholesterol (LDL-C) was defined as serum LDL-C ≥ 1.8 mmol/L (70 mg/dL). Low high-density lipoprotein cholesterol (HDL-C) was defined as serum HDL-C < 1.0 mmol/L (40 mg/dL). Elevated triglyceride (TG) was defined as serum TG ≥ 2.3 mmol/L (200 mg/dL). Current smoking was defined as smoking in the preceding 1 year according to the medical records of the patients. Estimated glomerular filtration rate (eGFR) was calculated by the equation developed by the Chronic Kidney Disease Epidemiology Collaboration [9]. A history of coronary heart disease (CHD) was specified if the patients had a clinical history of myocardial infarction or underwent PCI or coronary artery bypass grafting (CABG) before the current hospitalization. Other clinical history of diseases, including cerebrovascular disease, heart failure, peripheral artery disease (PAD), atrial fibrillation, and renal failure was defined according to the notes on original medical records. Heart failure, cardiac arrest, and cardiac shock occurring within 24 h of the current admission were defined as a severe clinical condition. The definition of fivefold elevated myocardial injury markers was elevation of cardiac injury marker beyond fivefold the upper reference limit [5]. In addition, we evaluated the risk of in-hospital death using the Global Registry of Acute Coronary Events (GRACE) score whereby patients with a score greater than 140 were classified as high risk [10]. Subtypes of ACS were defined based on the principal discharge diagnosis of medical records. Patients with a diagnosis of non-STEMI and unstable angina were classified as NSTE-ACS. Cardiologists diagnosed patients based on guidelines of STEMI and NSTE-ACS issued by the Chinese Society of Cardiology [11, 12]. The diagnostic criteria included symptoms of chest pain, results of ECG, and biomarkers of myocardial injury.

As most of the patients with possible diabetes could be undiagnosed or were at high risk of developing diabetes, and needed the same care as diabetic patients during hospitalization [2‐5, 13], for the purposes of this study we combined diabetes and possible diabetes for analysis. Prevalence of diabetes/possible diabetes and its 95% confidence intervals (CIs) were estimated in the overall study population and in subgroups by sex, age groups, and CHD history. The characteristics, in-hospital treatments, and in-hospital outcomes of these patients were described and compared according to diabetic status in ACS patients. Continuous variables with normal distribution were shown as mean (standard deviation [SD]) and differences between groups were compared using t-tests; continuous variables with skewed distribution were shown as median (interquartile range [IQR]) and compared using the Mann–Whitney U test; and categorical variables were presented as the number (percentage) and compared using chi-square test. Logistic multivariable regression analysis was carried out to examine the association between diabetes/possible diabetes and in-hospital outcomes. Univariate analysis was performed first, followed by multivariate-adjusted analysis. The candidate adjusted factors are confounding factors that either have been included in the risk assessment or have been reported more than once with an effect on death or MACCE, including baseline characteristics, risk factors, medical history, clinical conditions at admission, and treatment during hospitalization, i.e., age (continuous), sex (male/female), current smoking (yes/no), SBP levels (continuous), heart rate (continuous), cardiac arrest at admission (yes/no), Killip class at admission (class I/II–III/IV), history of CHD (yes/no), cerebrovascular disease (yes/no), PAD (yes/no), heart failure (yes/no), renal failure (yes/no), eGFR (continuous), administration of dual antiplatelet therapy (yes/no), anticoagulant therapy (yes/no), statins (yes/no), β-blockers (yes/no), and angiotensin-converting enzyme inhibitors (ACEIs)/angiotensin-receptor blockers (ARBs) (yes/no) during hospitalization, PCI treatment (yes/no), fivefold elevated myocardial injury markers (yes/no), type of ACS (STEMI/NSTE-ACS), and whether patients were transferred from another hospital before the current hospitalization (yes/no). After forward stepwise selection with entry and exit criteria both set at the p = 0.15 level, the variables listed in the legend of Table 4 were eventually included in the multivariable adjusted logistic model of all-cause death and MACCE, respectively. Given the differences in pathologies, management, and prognosis of STEMI and NSTE-ACS, we performed the above analyses in these two subtypes of ACS patients.

Since some ACS patients with FBG ≥ 7.0 mmol/L but with HbA1c < 6.5% were classified as patients without diabetes, who could mostly be diagnosed with stress hyperglycemia and associated with increased risk of death and MACCE, we conducted a sensitivity analysis by excluding these patients and recalculated the risk of diabetes/possible diabetes.

Subgroup analysis, including age, sex, Killip class, eGFR, GRACE score, PCI treatment, types of ACS, and whether the patient was transferred before the current hospitalization, was performed by using important characteristics in a multivariable adjusted logistic regression model. Odds ratios (ORs) between subgroups were compared using a Z-test [14].

In addition, we conducted a propensity-score-matched analysis to further confirm the association between diabetes/possible diabetes and in-hospital outcomes. First, a propensity score of having diabetes/possible diabetes was calculated by a logistic regression model with the variables age, sex, SBP levels, heart rate, LDL-C, HDL-C, TG, eGFR, Killip class at admission, history of myocardial infarction, PCI, CABG, cerebrovascular disease, heart failure, PAD, atrial fibrillation, renal failure, and type of ACS. Patients with and without diabetes/possible diabetes were then matched at a 1:1 ratio by propensity score using nearest-neighbor matching without replacement, with a caliper of 0.02. The absolute standardized differences of variables included for the calculation of propensity score were compared before and after propensity-score matching. Standardized differences < 10.0% for these included variables indicated a relatively small imbalance. The baseline characteristics and in-hospital management between the two propensity-score-matched subsets were re-compared. As some characteristics did not exactly match between the two groups even after the propensity-score matching, multivariable logistic regression was further performed to compare the risk by adjusting factors eventually included in the whole study population by stepwise selection.

The average age of the 63,450 ACS patients, 25.1% of whom were female, was 62.9 (± 12.4) years. Among these patients, a total of 23,880 (37.6%) had diabetes/possible diabetes (Table 1), including 29.7% diabetes and 7.9% possible diabetes. Both STEMI and NSTE-ACS patients had a high prevalence of diabetes/possible diabetes, but prevalence was slightly higher in patients with NSTE-ACS than in those with STEMI (39.0% versus 36.8%). Women had a higher proportion of diabetes/possible diabetes than men (45.0% versus 35.2%). The prevalence of diabetes/possible diabetes increased significantly with age. However, even in patients younger than 45 years, 26.9% of them had diabetes/possible diabetes. Patients with a history of CHD had a higher prevalence of diabetes/possible diabetes than those without CHD history (45.9% versus 36.6%).

Compared with ACS patients without diabetes, patients with diabetes/possible diabetes had a higher frequency of previous diseases and major cardiovascular risk factors (Table 2). Of these ACS patients with diabetes/possible diabetes, 21.5% had previously diagnosed CVD and 51.2% had three or more other cardiovascular risk factors, including hypertension, different types of dyslipidemia, and smoking.

ACS patients with diabetes/possible diabetes also had more severe clinical conditions than those without diabetes at admission, with a higher frequency of heart failure (11.9% versus 7.2%), cardiac shock (3.7% versus 2.7%), and cardiac arrest (2.2% versus 1.8%). In addition, the proportion of high-risk patients based on GRACE scores was also significantly higher in patients with diabetes/possible diabetes in comparison with non-diabetic patients (39.2% versus 31.2%). Similar results were observed in subtypes of ACS patients with and without diabetes/possible diabetes.

We compared the treatments for ACS between patients with and without diabetes/possible diabetes with regard to STEMI and NSTE-ACS (Table 3). Most of the received treatments, including PCI, antiplatelet therapy, anticoagulant therapy, statins, and β-blockers, were comparable between patients with and without diabetes/possible diabetes. We did not observe a higher rate of CABG in ACS patients with diabetes in this study.

The in-hospital outcomes were compared between ACS patients with and without diabetes/possible diabetes (Fig. 1 and Additional file 1: Table S1). Higher rates of all-cause death and MACCE were observed in all ACS patients with diabetes/possible diabetes as well as in the subtypes of ACS. In univariate logistic regression analysis, a significantly higher risk of all-cause death and MACCE was observed in patients with diabetes/possible diabetes (Table 4). The independent association was further evaluated using multivariable analyses (Table 4 and Additional file 1: Tables S2 and S3). After multivariable adjustment, diabetes/possible diabetes was associated with a twofold increased risk of all-cause death (OR, 2.04 [95% CI 1.78–2.33]) and a 1.5-fold increased risk of MACCE (OR, 1.54 [95% CI 1.39–1.72]).

We then conducted a sensitivity analysis to evaluate the risk of diabetes/possible diabetes. After excluding patients with possible stress hyperglycemia (n = 2465) in patients without diabetes, diabetes/possible diabetes was still associated with an increased risk of in-hospital all-cause death (OR, 2.22 [95% CI 1.92–2.56]) and MACCE (OR, 1.63 [95% CI 1.46–1.83]).

Subgroup analyses were performed based on important baseline characteristics. Diabetes/possible diabetes was associated with increased risk of all-cause death and MACCE in all subgroups (Fig. 2).

In addition, we conducted a propensity-score-matched analysis to further confirm the association between diabetes/possible diabetes and in-hospital outcomes. After propensity-score matching, 19,315 ACS patients with diabetes/possible diabetes were matched with 19,315 patients without diabetes (patients with possible stress hyperglycemia were excluded before matching). After matching, the standardized differences were less than 10.0% for all variables included for the calculation of propensity score, indicating that ACS patients with and without diabetes/possible diabetes were well matched (Additional file 1: Figure S2). The characteristics and in-hospital treatment between these two groups were re-compared, whereby most of the characteristics were comparable (Additional file 1: Table S4). The rates of all-cause death and MACCE remained higher in patients with diabetes/possible diabetes, and an excess risk of in-hospital outcomes independently associated with diabetes/possible diabetes was also found (all-cause death: OR, 2.21 [95% CI 1.83–2.66]; MACCE: OR, 1.58 [95% CI 1.38–1.82]) (Fig. 1 and Table 4).

We found that 1 in 3 male ACS patients and 2 in 5 female ACS patients had diabetes/possible diabetes. With the rapid increase in prevalence of diabetes in China, the proportion of diabetes in the ACS patients will continue to rise [15]. The China Heart Study published in 2006 reported that 37.4% of patients with acute coronary artery disease were diagnosed with diabetes by medical history and FBG [7], and 17.4% of these patients were further diagnosed with diabetes by oral glucose tolerance test (OGTT). In this situation, the prevalence of diabetes in ACS patients may be higher than the current prevalence reported herein, as some patients with diabetes may not have been identified because OGTT currently is not applied in the routine clinical workup to assess the diabetic status of patients. These findings indicate that cardiologists in China have to manage a large proportion of ACS patients with diabetes in their clinical care.

However, there remains some doubt about whether our cardiologists are fully prepared to manage this group of patients. In this study, we found that 68.2% of patients (with both measurement of FBG and HbA1c) with FBG ≥ 7.0 mmol/L could be diagnosed with diabetes by HbA1c, which meant that about 70% of patients with possible diabetes could be diagnosed with diabetes with HbA1c tests; however, 57.0% patients did not receive a test for HbA1c during hospitalization. Therefore, for a considerable number of patients the best opportunity to identify and treat their previously undiagnosed diabetes might have been missed, particularly for those patients with little or no routine health care before the occurrence of ACS events. Effectively identifying these patients during hospitalization is thus the first key step in cardiologists’ management strategy. In addition, diabetes also has a great impact on the prognosis of various diseases, and long-term monitoring is necessary [16‐18].

Our study showed that ACS patients with diabetes/possible diabetes had a substantially high risk for in-hospital outcomes compared with patients without diabetes, namely a twofold increased risk of all-cause death and a 1.5-fold increased risk of MACCE. A recently published systematic review and meta-analysis provided a summarized excess risk of early mortality from diabetes status in patients with myocardial infarction/ACS based on 86 studies published from 1970 to 2011 [19]. Here it was reported that diabetes was associated with a 1.7-fold higher risk of early mortality and that the relative risk of early death associated with diabetes did not change over time [19]. Compared with previous studies in Chinese ACS patients, the rates of all-cause death and MACCE during hospitalization have been significantly decreased in our study [20‐22]. These findings might suggest that the advancements in the management of ACS patients during the last decades have improved the prognosis of ACS patients but have not led to a reduction of the risk gap between diabetic and non-diabetic patients.

However, one point worth noting is that most of the previous studies did not address the problem of undiagnosed diabetes and stress hyperglycemia [23, 24], which has been defined as possible diabetes in our study. Researchers compared patients with history (previously diagnosed) of diabetes and those without history of diabetes, which included all patients without diabetes, with undiagnosed diabetes, or with stress hyperglycemia. These analyses may underestimate the relative risks of diabetes given the increased risk of the reference group. In our study, we classified patients with undiagnosed diabetes and stress hyperglycemia (FBG ≥ 7 mmol/L) as possible diabetes as they did not have an HbA1c result, and who were associated with a threefold increased risk of all-cause death compared with those without diabetes. Therefore, all ACS patients with FBG ≥ 7 mmol/L or with diabetes should raise major concern in clinical practice in light of their extremely high risk. Relative hyperglycemia, a new concept, reported to associated with complications following an acute myocardial infarction [25], also need to be concerned.

The reasons for the excess risk of all-cause death and MACCE in ACS patients with diabetes/possible diabetes could be partially be explained [26‐28], but some reasons are unexplained based on current analysis as the information on anti-diabetic treatment was not available to our study. In our study, we observed that the in-hospital management for ACS was similar between patients with and without diabetes. However, anti-diabetic therapy in the acute phase is also very important for the prognosis of ACS patients with diabetes, and inappropriate hypoglycemic treatment could significantly increase the risk of death [29]. The guidelines have given clear anti-diabetic drug recommendations for patients with both CVD and diabetes [30, 31], and an increasing number of studies have found that newer types of anti-diabetic drugs have a beneficial effect on lowering both blood glucose levels and risks of CVD, but conflicting results still exist [32‐36]. In addition, the combined use of anti-diabetic drugs on cardiovascular events should also be concerned [37]. Future studies should take this information into consideration.

Whether in the American, European, or other countries of the world, ACS patients with diabetes is common (usually greater than one-third of patients) and associated with a higher risk of death and other adverse events [2‐5, 38]. Although studies have reported that the cardiovascular outcomes of diabetes have been improved in recent years, number of people with diabetes still rises, the absolute burden of CVD will still be high [39]. Effective strategies to better manage the risk of these ACS patients with diabetes and improve their prognosis has always been the focus but also a challenge for cardiologists worldwide. In 2013, the European Society of Cardiology in collaboration with the European Association for the Study of Diabetes developed the second guideline for diabetes, pre-diabetes, and cardiovascular diseases, which calls for physicians in the fields of cardiovascular medicine and diabetes to join forces to research and manage these conditions, given the close relationship between CVD and diabetes [30]. In 2015, the Chinese Society of Cardiology in collaboration with other societies also issued a guideline on the management of abnormal glucose metabolism and CVD [40]. Following the efforts of both cardiologists and diabetologists, the risk of adverse events for ACS patients with diabetes/possible diabetes is expected to decrease [41].