Methods

The study population comprised all consecutive patients (n = 11,441) who underwent PCI at our institution at the 2 hospitals of the Rabin Medical Center in Israel from April 2004 to September 2013. Data collection was approved by the hospital ethics committee in compliance with the Declaration of Helsinki. As we have previously reported, all data regarding the index and subsequent procedures, as well as clinical and echocardiographic data, were extracted from the patients’ electronic medical records. Demographic data and death dates were obtained from the medical centers’ demographic information system, which is linked to the State of Israel Ministry of Interior data system and the Clalit Health Organization data warehouse. The accuracy of the mortality data were verified with the Israel Central Bureau of Statistics. All data regarding previous and subsequent hospitalizations, including all International Classification of Diseases, Ninth Revision, diagnoses, were retrieved from the medical centers’ data warehouse. Laboratory data were retrieved from the medical centers’ central laboratory database. Definitions regarding ST-elevation myocardial infarction (STEMI) were obtained from the Rabin Medical Center interventional cardiology database, which records detailed data regarding all patients with STEMIs. All follow-up data were collected up to September 30, 2013.

An initial list of candidate variables to influence outcome was chosen a priori, selected and based on relevance, as identified in previous research, or by the researchers’ clinical experience. Baseline demographic, clinical and angiographic parameters were then examined by univariate logistic regression analysis for their relation to the study outcome of interest: death or myocardial infarction (MI) during follow-up.

On the basis of previous studies, the left ventricular ejection fraction, creatinine clearance (CrCl), and hemoglobin level were all dichotomized and treated as binary variables. All patients with moderate or worse left ventricular dysfunction (LVD) defined in the echocardiographic record were flagged as “moderate to severe LVD.” All PCIs for acute/recent MI or acute coronary syndromes, defined according to the indication as noted on the electronic record, were flagged as “urgent PCI.” Primary PCI for STEMI was defined by the prerequisites for inclusion in the STEMI registry: within 12 hours of symptoms, without previous thrombolysis. All patients who arrived at the PCI laboratory after resuscitation or with cardiogenic shock were flagged as “critical state.” Chronic kidney injury (CKI), based on the Kidney Disease Outcomes Quality Initiative criteria, was defined as estimated CrCl ≤60 ml/min (calculated using the Modification of Diet in Renal Disease [MDRD] equation). Anemia, based on World Health Organization criteria, was defined as baseline hemoglobin concentration <13 g/dl in men and <12 g/dl in women. To assess coronary artery disease complexity, the number of vessels with coronary disease was determined by analyzing the diagnostic catheterization report. Significant disease was considered when >50% stenosis was noted. Treated territories were defined by analyzing the angioplasty report. For each territory, PCI sites were counted (e.g., proximal and mid left anterior descending coronary artery, first or second diagonal branch) and a simple score of sites and territories, termed “complexity,” was defined, which reflects the number of discrete lesions treated per territory. Whenever treatment involved ≥1 ostial or proximal main vessel (left anterior descending, circumflex, or right coronary artery) or the left main coronary artery, the procedure was flagged as “proximal main vessel.” Total stent length was calculated and filed for each procedure. All patients with any drug-eluting stent (DES) implanted in the index PCI were included in the DES group regardless of use of any additional bare-metal stent.

Univariate analysis was used to identify which of the candidate variables had a statistical association with the composite end point of death or MI during follow-up and with all-cause mortality (e.g., p <0.05). On the basis of this univariate analysis, we selected the most clinically meaningful variables as potential candidates for inclusion in the multivariate analysis. Multivariate logistic regression was then performed to identify independent predictors of outcomes and calculate their individual hazard ratios (i.e., effect sizes). Baseline parameters were compared between groups using Student’s t test for continuous variables and the chi-square test for categorical variables. Statistical analyses were performed using IBM SPSS version 20 (IBM, Armonk, New York). All tests were 2 tailed, and p values <0.05 were considered significant. All data processing and statistical analyses were performed by the main investigator.

Results

Patients’ demographic, clinical, and angiographic features at the index PCI are listed in Table 1 . Echocardiographic data before the PCI date was available for 6,979 of the patients (61%). Preprocedural serum creatinine and hemoglobin levels were not available for 2,288 of the patients (20%). For all other variables, <4% of values were missing. Mean follow-up time was 5.5 years (range 3 months to 9.5 years), during which a total of 1,846 deaths (16%) and 450 MIs (3.9%) occurred.

Statistically significant univariate outcome predictors are listed in Table 2 . In the multivariate analysis ( Tables 3 and 4 ), in addition to well-known predictors of death or nonfatal MI such as age (adjusted hazard ratio [HR] 1.031 for death or MI and for all-cause mortality, p <0.001), female gender (HR 1.23 for death or MI and for all-cause mortality, p <0.001), diabetes mellitus (HR 1.28 for death or MI, HR 1.26 for all-cause mortality, p <0.001) and an urgent setting (HR 1.23 for death or MI, p <0.001; HR 1.09 for all-cause mortality, p = 0.115), we particularly noted previous anemia (HR 1.55 for death or MI, HR 1.75 for all-cause mortality, p <0.001), previous CKI (HR 1.93 for death or MI, HR 2.13 for all-cause mortality, p <0.001), and previous moderate to severe LVD (HR 2.29 for death or MI, HR 2.50 for all-cause mortality, p <0.001) as independent predictors. With the exception of urgent PCI and DES refraining, all outcome predictors were found to anticipate death better than the composite outcome of death and MI. Urgent PCI did not significantly predict future death risk (p = 0.115). Previous coronary artery bypass surgery and coronary artery disease complexity were not associated with higher risk. The use of a DES provided significantly better long-term outcomes (HR 0.70 for death or MI, HR 0.71 for all-cause mortality, p <0.001).

Table 3

Variables associated with Death or Myocardial Infarction in the multivariable analysis

Variable	Hazard-ratio	95% CI	p Value
Age years	1.03	1.03-1.04	<0.001
Women	1.23	1.09-1.34	<0.001
Urgent percutaneous coronary intervention ∗	1.23	1.08-1.31	<0.001
Diabetes-mellitus	1.28	1.17-1.41	<0.001
Anemia †	1.55	1.39-1.71	<0.001
Chronic kidney injury ‡	1.93	1.78-2.16	<0.001
Moderate-severe left ventricular dysfunction	2.29	2.13-2.58	<0.001
Bare metal stent	1.43	1.32-1.55	<0.001

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Eleven-Year Prognostic Value of Dobutamine Stress 99m Tc-Sestamibi Myocardial Perfusion Imaging in Patients With Limited Exercise Capacity
2. Usefulness of Sonication of Cardiovascular Implantable Electronic Devices to Enhance Microbial Detection
3. Usefulness of Thrombocytopenia at Admission as a Prognostic Marker in Native Valve Left-Sided Infective Endocarditis
4. Measuring the Base of the Triangle is a Useful Tool to Differentiate Between Electrocardiographic Patterns of Healthy Athletes and Brugada Syndrome

Stay updated, free articles. Join our Telegram channel

Join