Safety and cost analysis of early discharge following percutaneous coronary intervention for acute coronary syndrome in patients with diabetes mellitus.

Introduction

Diabetes mellitus (DM) has become a twenty-first century pandemic, with a marked rise in prevalence in developed and developing countries.[1] A recent survey of the worldwide prevalence of DM reported a rate of 6.4% in 2010, representing 285 million adults, with an expected rise to 7.7% by 2030 (439 million adults).[2] In a landmark nationwide study in China in 2010, the overall prevalence of DM was reported as 11.6% in the general Chinese adult population.[3,4] DM is a multisystem disorder and a recognized risk factor for coronary artery disease (CAD). Complications associated with DM represent a significant health burden in China.[5,6]

The number of patients requiring percutaneous coronary intervention (PCI) for CAD[7,8] is increasing year by year, and the health economics of this procedure represent an important issue. Hospitalization days and costs are the key indicators of hospital medical burden. Controlling the rapid increase of hospital expenses is of significance to patients, families, governments, and medical insurance companies. Previous studies have shown that for certain types of coronary heart disease and in some Chinese hospitals, there is a trend towards declining hospitalization time,[9] while other hospital-based single-center studies have shown that the cost of hospitalization for coronary heart disease has increased.[10-12] Among these somewhat contradictory findings, however, there remains a lack of research on the relative safe hospitalization time of PCI for patients with DM and coronary heart disease in a representative large sample population, as well as the relationship with hospitalization costs.

In this study, 474 patients with DM who underwent PCI for the treatment of ACS from 2012 to 2015 were retrospectively evaluated at 30 days of follow-up after discharge to determine the cost and number of days of hospitalization. We sought to clarify whether early discharge following PCI in this patient population is safe compared with ordinary discharge, and whether early discharge can reduce costs and thus contribute to the recognized benefits of the PCI procedure.

Methods

Study population

The hospital electronic case system was used to identify hospitalized patients with DM who underwent PCI at Beijing An Zhen Hospital from January 2012 to June 2015.

Inclusion criteria: (1) age ≥45 years; (2) PCI; (3) successfully placed with DES; (4) with unstable angina pectoris (UAP), non-ST-segment elevation myocardial infarction (NSTEMI), and ST-segment elevation myocardial infarction (STEMI) as revascularization indications; (5) patients with a history of diabetes, insulin or hypoglycemic treatment at admission, and new-onset diabetes during hospitalization; (6) obtaining informed consent.

Exclusion criteria: (1) malignant tumor; (2) liver and kidney disease; (3) infection; (4) blood and immune system diseases; (5) type 1 diabetes; (6) rheumatic valvular heart disease and cardiomyopathy history; (7) Patients with incomplete data, patients who missed their visit within 1 year after receiving PCI.

Information collection and indicator definition

A retrospective analysis of prospective data was performed. Demographic, clinical, laboratory, coronary angiography, and coronary stenting data were retrieved for all enrolled patients, as well as records of medication administered during discharge and follow-up. Specifically, the following information was collected: sex, age, weight, height, blood pressure, symptoms at admission, cardiac function, and biochemical indicators. Medical history of diabetes, hypertension, dyslipidemia, smoking status, diabetes history, receipt of insulin treatment, non-fatal myocardial infarction (MI), non-fatal stroke, medications, and fasting plasma biochemical indicators before and after revascularization were also collected.

Diagnosis of DM was based on the 2014 American Diabetes Association (ADA) guidelines,[13] and dyslipidemia assessment was based on the 2013 diagnostic criteria of the European Regulations on Dyslipidemia.[14] Diagnosis of hypertension was based on the American Diabetes Prevention and Treatment Guidelines (JNC7),[15] and patients with a history of essential hypertension were diagnosed with hypertension if they were using antihypertensive drugs at enrollment. Patients who had smoked for more than 1 year before admission or who had stopped smoking for less than 1 year were classified as smokers.

Study design

This was a retrospective observational cohort study. The primary data source was the local coronary artery disease database in the hospital electronic case system, which contains information on every PCI procedure performed at our hospital, Including prospectively entered baseline demographics, procedural details, and clinical presentation.

Outcome measures

The main outcome measure was all-cause mortality at 30 days and at 1 year of post-discharge follow-up. Discharge minor events including bleeding, infection, and fever were collected, as was the incidence of major adverse cardiac and cerebrovascular events (MACCE) observed at 30 days and at 1 year of follow-up. Clinical follow-up data were obtained by independent follow-up teams during telephone and outpatient visits.

Length of hospital stay

Patients were classified by length of hospital stay into (1) an early discharge group, consisting of patients discharged from hospital within 6 days following PCI; and (2) an ordinary discharge group, consisting of all other patients.

Statistical analysis

Continuous variable distributions were assessed for normality using the Kolmogorov–Smirnov test. Continuous variable data with a normal distribution were expressed as the mean (standard deviation), and data with skewed distributions were expressed as medians (interquartile range [IQR]). Differences between normally distributed continuous variables were analyzed using an unpaired test. The Mann–Whitney U test was used for continuous variables with a skewed distribution. Groups were compared using a chi–squared test. Multiple binary logistic regression analyses were performed to identify independent variables associated with early discharge. The cumulative incidence of cardiovascular events was calculated using the Kaplan–Meier method and compared using a log-rank test. A two-sided value of p<0.05 was considered significant. All analyses were performed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, IL, USA).

Ethics approval and consent to participate

Ethical approval was obtained from the Ethics Committee of Beijing An Zhen Hospital Capital Medical University, and all participants provided signed informed consent prior to participation.

Results

Patient disposition

All 474 patients (60.0±8.0 years old, 68.6% male) who underwent PCI at our center and survived to hospital discharge were included in the study. There were 192 patients (40.5%) in the early discharge group and 282 patients (59.5%) in the regular discharge group. The median and IQR of length of hospital stay was 6 (3–6) days (mean of 4.5 days) in the early discharge group compared with 9 (6–11) days (mean of 8.5 days) in the ordinary discharge group.

There were significant differences in hospitalization costs, left ventricular ejection fraction (LVEF), high density lipoprotein (HDL), troponin I (TNI), lesion length ≥20 mm, previous angina, double vessel, pre-PCI, and number of stents observed between the two groups (p<0.05). There was no significant difference between the groups in age, sex, body mass index (BMI), blood pressure, blood glucose, other biochemical indicators, hemoglobin A1c (HbA1c), smoking status, previous medical history, medication intake, and ACS composition ratio, the difference between early discharge and ordinary discharge. Furthermore, differences in chronic total occlusion (CTO), left main (LM) involvement, diameter of stent, in-stent restenosis, pre-CAD, pre-thrombolytic, history of hypertension, and pre-CABG were not statistically significant between the two groups. Differences in minor events at discharge and in MACCE at 30 days and 1 year of follow-up were not statistically significant between the two groups. Baseline demographics and procedure-related characteristics are shown in Tables 1 and 2, respectively.

Table 1.

Baseline characteristics of enrolled patients

Age (years)	60±8
Men/women	325/149
Height (cm)	167±23
BMI Median (Q1,Q3)	26.6(24.4,28.8)
LOS Median (Q1,Q3)	6(5,8)
Expense in hospital (RMB)	59012±26980
HbA1C (%)	7.3±2.4
UAP/NSTEMI/STEMI	389/19/66
Thrombosis history	45
Pectoris history	407
Pre-PCI history	98
Pre-CABG	16
Hypertension history	352
SBP (mmHg)	130±18
DBP (mmHg)	76±10
Anti-hypertension	262
Arrhythmia history	15
Heart failure	2
CVD history	71
Peripheral disease	15
Ex-smoker	169
HR (time/minute)	65±19
ALT median (Q1,Q3)	22(14.5,32)
AST median (Q1,Q3)	20(16, 25)
Cr median (Q1,Q3)	97(79, 115)
UA	290±134
Glucose (mmol/L)	7.8±5.9
TG (mmol/L)	2.04±1.68
CRP median (Q1,Q3)	1.50(0.67, 3.66)
cTnI median (Q1,Q3)	0.01(0, 0.07)
Pro-BNP median (Q1,Q3)	27(10, 62)
WBC (G/L)	6.78±2.13
RBC (T/L)	4.54±2.05
HGB (g/L)	136±32
PLT (G/L)	206±66
Discharge minor events (n,% )	30(6.3)
30-day events (n,%)	10(2.1)
TRI (n, %)	57(12.0)
Triple vessels (n,%)	103(21.7)
Double vessels (n,%)	169(35.7)
CTO (n,%)	89(18.8)

Values are mean±SD or %

BMI, body mass index; LOS, length of stay; HbAC1, hemoglobin A1c; UAP, unstable angina pectoris; NSTEMI, non-ST elevation myocardial infarction; STEMI, ST elevation myocardial infarction; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft; SBP, systolic pressure; DBP, diastolic blood pressure; CVD, cardiovascular disease; HR, heart rate; ALT, alanine aminotransferase; AST, aspartate aminotransferase; Cr, creatinine; UA, uric acid; TG, triglyceride; CRP, c-reactive protein; cTnI, sensitive cardiac troponin T; BNP, brain natriuretic peptide; WBC, white blood cell; RBS, red blood cell; HGB, hemoglobin; PLT, platelet; TRI, transfemoral coronary intervention; CTO, chronic total occlusion

Table 2.

Clinical comparison between the early discharge group and ordinary discharge group.

	Early discharge (<6 days n=192)	Ordinary discharge (>=6 days n=282)	P-value
Age (years)	60±8.4	60±8.2	0.91
Men/women	131/61	194/88	0.88
BMI	26±6.3	27±7.5	0.21
Cost (RMB)	50983±16780	65750±14950	<0.0001
HbAC1 (%)	7.2±1.3	7.5±1.4	0.024
UAP	158/192	231/282	0.94
NSTEMI	7/192	12/282
STEMI	27/192	39/282
SBP (mmHg)	129±15	131±17	0.45
DSP (mmHg)	76±10	76±9	0.79
LVEDD (mm)	39±18	42±15	0.08
LVESD (mm)	24±14	26±12	0.10
LVEF (%)	64±6	62±7	0.002
HR beats/minute	64±20	65±18	0.56
Cr (µmol/L)	101±36	138±42	0.18
UA (µmol/L)	290±129	289±130	0.90
HDL (mmol/L)	0.99±0.27	0.92±0.27	0.012
TG (mmol/L)	1.94±1.67	2.11±1.71	0.26
TC (mmol/L)	4.12±0.97	4.01±1.04	0.54
LDL (mmol/L)	2.38±0.70	2.37±0.86	0.93
CRP (mg/L)	1.25(0.54,2.80)	1.73(0.76,4.41)	0.013
cTnI (µg/L)	0.01(0,0.04)	0.03(0,0.09)	0.04
Pro-BNP (pg/mL)	23(8,61)	29(11,62)	0.06
Glucose (mmol/L)	8.0±7.4	7.6±4.9	0.47
1 year FU events (n,%)	18(9.38)	30(10.64)	0.65
30 days events (n,% )	6(3.13)	4(1.42)	0.21
Discharge minor event (n,% )	10(5.21)	20(7.09)	0.40
CTO	31(16.15)	58(20.57)	0.21
LMS	4(2.08)	12(4.26)	0.18
Lesion length >=20 mm	6(3.13)	22(7.80)	0.027
Double vessel	80(41.67)	89(31.56)	0.025
Ex-smoker	63(32.8)	106(37.59)	0.28
Denovol (y/n)	26/166	32/250	0.47
In-stent restenosis (y/n)	10/182	14/268	0.90
CABG (y/n)	1/192	3/279	0.51
Heparin (y/n)	90/102	146/136	0.29
Pre-CAD	171/192	251/282	0.98
Pre-thrombolytic	4/192	10/282	0.34
Previous angina	174/192	233/282	0.012
Pre-PCI	49/192	49/282	0.032
History of hypertension	145/192	207/282	0.60
Pre-CABG	6/192	9/282	0.96
Taking antihypertensive drugs	113/192	149/282	0.19
Arrhythmia	4/192	11/282	0.26
HF	0/192	2/282	0.15
Previous cerebrovascular disease	28/192	43/282	0.84
Previous peripheral blood vessels	6/192	9/282	0.97
ALT (U/L)	28±26	28±27	0.84
AST (U/L)	23±17	23±20	0.76
Number of stents	1.78±1.03	2.26±1.50	0.0001
Diameter of stent	2.90±0.43	2.80±0.45	0.11
Hyperlipidemia	37/192	50/282	0.67

Values are %, mean±SD, or median (interquartile range)

BMI, body mass index; HbAC1, hemoglobin A1c; UAP, unstable angina pectoris; NSTEMI, non-ST elevation myocardial infarction; STEMI, ST elevation myocardial infarction; SBP, systolic pressure; DBP, diastolic blood pressure; LVEDD, left ventricular end diastolic diameter; LVESD, left ventricular end-systolic diameter; LVEF, left ventricular ejection fraction; HR, heart rate; Cr, creatinine; UA, uric acid; HDL, high density lipoprotein; TG, triglyceride, TC, total cholesterol; LDL, low-density lipoprotein; CRP, c-reactive protein; cTnI, sensitive cardiac troponin; CTO, chronic total occlusion; LMS, left main stem; BNP, brain natriuretic peptide; FU, follow-up; CABG, coronary artery bypass graft; CAD, coronary artery disease; PCI, percutaneous coronary intervention; HF, heart failure; ALT, alanine aminotransferase; AST, aspartate aminotransferase

Predictors of early discharge

In a multiple logistic regression model, the following variables were tested for their effect on time to discharge: age, sex, BMI, HbA1c, glucose, pre-PCI, hypertension, pre-CABG, hyperlipidemia, ex-smoker, HDL, C-reactive protein (CRP), hemoglobin (HGB), CTO, LVEF, TNI, and lesion length ≥20 mm, previous angina, double vessel, pre-PCI, and number of stents. These variables were selected as they are recognized as clinically important in defining low-risk patients and were found to have a significant impact on early discharge in univariate analysis. HDL, TNI, number of stents, and previous angina were found to be independent predictors of early discharge, and their odds ratios (OR) in multiple logistic regression analysis are shown in Table 3.

Table 3.

Multivariate logistic analysis for predictors of early discharge

Variable	Multivariable OR (95%CI)	P-value
HDL-C	3.30 (1.42-7.90)	0.0036
TNI	0.60 (0.32-0.92)	0.0073
Number of stents	0.70 (0.59-0.83)	<0.001
Previous angina	2.27 (1.22-4.40)	0.0091

HDL-C, high density lipoprotein cholesterol; TNI, troponin I

Mortality outcomes

At discharge, minor events were reported in 10 patients in the early discharge group and 20 patients in the ordinary discharge group. At the 30-day follow-up, MACCE events were reported in 6 patients in the early discharge group and 4 patients in the ordinary discharge group. At the 1-year follow-up, events were reported in 18 patients in the early discharge group and in 30 patients in the ordinary discharge group. The incidence of MACCE was not significantly increased in the early discharge groups at either 30 days or 1 year, and there were no significant differences in the incidence of MACCE between the two groups.

Impact of the hospital stay on outcomes

Figure 1 shows the impact of the hospital stay on outcomes of patients undergoing PCI in the early and ordinary discharge groups. Kaplan–Meier analysis showed a similar prognosis between the groups at 30 days (log rank P=0.19).

Figure 1.

Impact on duration of hospital stay on outcomes of patients with diabetes undergoing PCI

Discussion

The incidence of coronary heart disease is significantly increased among patients with DM, and is closely related to the severity of coronary atherosclerosis.[16-18] Because of the diffuse lesions observed in this patient population, distal lesions are more numerous, degree of atherosclerosis and rate of restenosis is higher, and the coronary diastolic reserve is poorer than in patients without DM. Thus, the number of patients undergoing PCI is significantly increased, the length of hospital stay is increased, and the total hospitalization expenses are significantly increased.[19,20]

With advancements in vascular access technology, scaffolding technology, and antiplatelet pharmacology, the discharge time after PCI has been greatly reduced.[21] Multiple clinical studies have demonstrated the safety of early and day-to-date discharge in patients with selective PCI,[22-25] and this strategy was deemed acceptable in the 2018 SCAI Expert Consensus Document on Hospitalization Time after Percutaneous Coronary Intervention (PCI) Promotion;[26] furthermore, this strategy prevents unnecessary hospitalization and is thus cost-effective for the Chinese healthcare system. These guidelines clearly state that these recommendations are intended to support reasonable clinical decisions for postoperative hospital stays in a wide range of patients undergoing PCI, rather than specifying specific observation periods for individual patients.

Time limit for early discharge

A recent large observational study of 33,920 patients from the PCI registry provided insights into post-PCI discharge patterns and related outcomes. Hospital stay (LOS) was divided into short-term (up to 3 days), medium-term (4–5 days) and long-term (>5 days). There were no significant differences in 3-day mortality or MACE between the short- and medium-term LOS groups. The study found that diabetes uncomplicated by coronary heart disease was the main factor influencing LOS, and that patients without diabetes were discharged earlier.[27] All patients in our study had DM, so the length of hospital stay in our study population was significantly longer than that in the previous study, and no patients were discharged in less than 3 days (median of 6 days). All patients in the present study had ACS, and the maximum UAP share was 82.1% and mild non-ST-segment elevation ACS (NSTEACS) was 0.4%, while STEMI patients comprised 13.6% of those undergoing direct PCI (PPCI). The degree of coronary atherosclerosis is an important consideration in patients with diabetes. Angiography showed that the incidence of multi-vessel disease, diffuse stenosis, small vessel disease, high atherosclerotic plaque load, small vascular development, poor development of collateral circulation, and serious microcirculatory dysfunction was increased in this patient population. Patients with DM have a higher clinical risk of acute ischemic and hemorrhagic events.[28] All patients in the present study had DM, and the average postoperative hospitalization stay after PCI was 6 days, longer than the SCAI expert consensus in 2018.[26] We analyzed the reasons for multiple hospitalizations, massive stent implantation, and postoperative hypercoagulability requiring 5–7 days of low-molecular-weight heparin therapy.

Early discharge predictors

Our findings showed that HDL level, TNI, number of stents, and previous angina were independent predictors of whether patients with DM could be discharged early. These factors are associated with the overall burden of DM with cardiovascular disease (CVD) and reflect the recurrence or progression of underlying CAD.

High density lipoprotein (HDL) is negatively correlated with the presence and development of CAD.[30] HDL-cholesterol is the only standardized and repeatable parameter that can be used to estimate the plasma concentrations of these lipoproteins. HDL exerts a variety of anti-atherosclerotic effects, and evidence showing that HDL has protective effects in atherosclerosis is increasing [31,32]. Previous studies have shown that DM is associated with an increased burden of CAD, and that low HDL is an independent predictor of CAD progression[33]. Our research further indicates that HDL level is associated with the severity of coronary artery lesions and early discharge from hospital after PCI.

It has been established that cardiac troponin (cTnI) significantly improves the early diagnosis of ACS [34], and that sensitive cTnI detection is of great significance for early diagnosis of ACS patients and prediction of long-term prognosis[35]. Our research supports this observation, and can predict whether patients can be discharged early, the increase in cTnI is delayed, and the patient’s condition is complicated.

In our study, high-risk patients were identified a priori as having LM involvement, CTO, double vessels, and pre-PCI. Previous studies have reported multiple stent complications in patients with diabetes and severe ACS, complicated lesions, and multiple stents.[35,36] Anticoagulation requires 5–7 days of low-molecular-weight heparin therapy, thus prolonging the hospital stay[36,37].

In patients with DM and a history of angina pectoris, coronary atherosclerosis is often more diffuse than in patients without DM, and features longer lesions, smaller vascular lumens, and larger plaques[38]. In addition, vascular endothelial function is impaired in patients with diabetes[39], as is platelet systolic function[28], and these factors can contribute to the development of ACS as well as to more complex lesions and complications, and prolonged hospital stays. Therefore, previous history of angina pectoris is an independent predictor of early discharge after PCI.

30-day assessment

The assessment of early discharge safety has been extensively studied in recent years. In a 2017 systematic review and meta-analysis of safety and cost, data from 12 randomized controlled trials in 2962 patients were analyzed[40]. At the follow-up time of 30 days in patients with stable angina, the rate of re-hospitalization of patients with ACS increased, and the early discharge duration was shorter (<72 hours), although no specific analysis of the reasons for re-hospitalization was reported.

In addition, a PCI medical cost and safety study in the United States[41] examined data from 206,869 inpatients discharged after PCI and re-admitted within 30 days of discharge. A total of 24,889 patients (12%) received treatment again within 30 days. The most common cause of re-admission was non-specific chest pain, although the majority of readmissions were the result of low-risk chest pain that did not require any intervention.

It has been established that patients with coronary heart disease and diabetes have a higher risk of coronary dissection, stent thrombosis, and in-stent restenosis when undergoing coronary revascularization, and that these events are prone to occur within 30 days after surgery[42,43]. In the present study, care was taken to assess risks and to select appropriate patients for early discharge based on HDL level, TNI level, number of interventional stents, and previous angina. No safety concerns were associated with early discharge within 6 days of surgery. There was no increase in adverse cardiovascular events during the 30-day follow-up period in either group, and no significant difference in adverse cardiovascular events between the two groups.

Cost analysis and social implications

Rapid increases in medical expenses are a major healthcare consideration, and are of significant concern to society and governments. The results of this study can support the evaluation of PCI from the perspective of health economics and medical outcomes. CVD is associated with the highest economic costs compared with other diseases[44]. According to the 2012 National Healthcare Cost and Utilization Project statistics, the mean hospital charge for a vascular or cardiac surgery or procedure in 2012 was USD78,897; cardiac revascularization cost USD149,480 and percutaneous interventions cost approximately USD70,027. CVD has higher costs than any other diagnostic group.[44] For a consecutive cohort of 5,306 patients undergoing PCI in China in 2010, total hospital costs were RMB57,900 for transradial intervention and RMB67,418 for transfemoral intervention[45]. The present study showed that the total cost of hospitalization for PCI following early discharge was significantly lower than that following ordinary discharge (RMB50,983 versus RMB65,750, a reduction of RMB14,767), highlighting the benefit of early discharge. Although we found no additional clinical benefit compared with regular discharge during the first 30 days after surgery, early discharge appears to be more cost-effective and may have a significant benefit in the care of patients with DM and resource allocation for coronary revascularization protocols in clinical settings. After PCI, the hospital can effectively reduce costs by reducing the average hospital stay and unnecessary hospitalization time while maintaining patient prognosis.

Limitations

This study had a number of limitations. First, although the data used in this study were from a cardiac center with a large number of patients undergoing PCI, this was a single-center observational study. Second, the follow-up time was relatively short and the status of patients after 3 years was unknown, so the long-term risk of revascularization and medical costs may have been underestimated in patients with DM complicated with CAD. Finally, the limited sample size means that further research is needed to confirm our findings.