J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7

ª 2 0 1 7 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N

P U B L I S H E D B Y E L S E V I E R

I S S N 1 9 3 6 - 8 7 9 8 / $ 3 6 . 0 0

h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j c i n . 2 0 1 6 . 1 1 . 0 2 8

Direct Admission Versus InterhospitalTransfer for Primary PercutaneousCoronary Intervention in ST-SegmentElevation Myocardial Infarction

Damian Kawecki, MD, PHD,a Marek Gierlotka, MD, PHD,b,c Beata Morawiec, MD, PHD,a Michał Hawranek, MD, PHD,b,c

Mateusz Tajstra, MD, PHD,b,c Michał Skrzypek, PHD,c,d Wojciech Wojakowski,e Lech Polo�nski,b,c

Ewa Nowalany-Kozielska,a Mariusz Gasiorb,c

ABSTRACT

Fro

Sil

Un

Dis

Po

Th

Ma

OBJECTIVES This study sought to assess the influence of direct admission versus transfer via regional hospital to a

percutaneous coronary intervention (PCI) center on time delays and 12-month mortality in ST-segment elevation

myocardial infarction (STEMI) patients from a real-life perspective.

BACKGROUND Reduction of delays to reperfusion is crucial in a STEMI system of care. However, it is still debated

whether direct admission to a PCI center is superior to interhospital transfer in terms of long-term prognosis. The authors

hypothesized that compared with interhospital transfer, direct admission shortens the total ischemic time, limits the loss

of left ventricular systolic function, and finally, reduces 12-month mortality.

METHODS Prospective nationwide registry data of STEMI patients admitted to PCI centers within 12 h of symptom

onset and treated with PCI between 2006 and 2013 were analyzed. Patients admitted directly were compared with

patients transferred to a PCI center via a regional non–PCI-capable facility in terms of time delays, left ventricular ejection

fraction (LVEF), and 12-month mortality. Data were adjusted using propensity-matched and multivariate Cox analyses.

RESULTS Of the 70,093 patients eligible for analysis, 39,144 (56%) were admitted directly to a PCI center. Direct

admission was associated with a shorter median symptoms-to-admission time (by 44 min; p < 0.001) and total ischemic

time (228 vs. 270 min; p < 0.001), higher LVEF (47.5% vs. 46.3%; p < 0.001), and lower propensity-matched 12-month

mortality (9.6% vs. 10.4%; p < 0.001). In propensity-matched multivariate Cox analysis, direct admission (hazard ratio

[HR]: 1.06, 95% confidence interval [CI]: 1.01 to 1.11) and shorter symptoms-to-admission time (HR: 1.03; 95% CI:

1.01 to 1.06) were significant predictors of lower 12-month mortality.

CONCLUSIONS In a large, community-based cohort of patients with STEMI treated by PCI, direct admission to a

primary PCI center was associated with lower 12-month mortality and should be preferred to transfer via

a regional non–PCI-capable facility. (J Am Coll Cardiol Intv 2017;-:-–-) © 2017 by the American College of Cardiology

Foundation.

S T-segment elevation myocardial infarction(STEMI), the most unfavorable type of acutecoronary syndrome, is associated with high

risk of death and development of heart failure.

m the a2nd Department of Cardiology, School of Medicine with the Div

esia, Katowice, Poland; b3rd Department of Cardiology, School of Medicin

iversity of Silesia, Katowice, Poland; cDepartment of Science, Training and

eases, Zabrze, Poland; dDepartment of Biostatistics, School of Public Heal

land; and the e3rd Department of Cardiology, School of Medicine in Katow

e authors have reported that they have no relationships relevant to the c

nuscript received October 24, 2016; accepted November 17, 2016.

Its prognosis depends on the infarct size, which iscorrelated with the time from the onset of symptomsto reperfusion. Consequently, constant efforts areexerted worldwide to maximally reduce the ischemic

ision of Dentistry in Zabrze, Medical University of

e with the Division of Dentistry in Zabrze, Medical

New Medical Technologies, Silesian Center for Heart

th in Bytom, Medical University of Silesia, Katowice,

ice, Medical University of Silesia, Katowice, Poland.

ontents of this paper to disclose.

ABBR EV I A T I ON S

AND ACRONYMS

ACS = acute coronary

syndrome

AMI = acute myocardial

infarction

ECG = electrocardiogram

IQR = interquartile range

LVEF = left ventricular ejection

fraction

PCI = percutaneous coronary

intervention

STEMI = ST-segment elevation

myocardial infarction

TIMI = Thrombolysis In

Myocardial Infarction

Kawecki et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7

Direct Admission vs. Transfer for PCI in STEMI - 2 0 1 7 :- –-

2

time and initiate appropriate treatment asearly as possible. The main focus in thissetting has been shifted from in-hospital topre-hospital delays (1,2). Primary percuta-neous coronary intervention (PCI) is thetreatment with the highest level of recom-mendation for STEMI. However, despite theincreasing number of PCI centers, regional,non–PCI-capable hospitals are often involvedin the system of care. The influence of inter-hospital transfer and related time delays onoutcomes in STEMI have been the focus ofattention in Europe and the United States(3–5). Many doubts have arisen concerninginterhospital transfers (6) and consequently,numerous national initiatives that address

monitoring and reducing transfer-related delayshave been introduced (7–11). Recently released resultsof the largest national program of STEMI care provethe efficacy of improving coordinated managementprotocols as a strategy to increase the proportion ofpatients treated within the preferred limits of time(12). However, efforts to constantly reduce delays toprimary PCI centers remain a matter of debate (13),despite results showing that early versus late admin-istration of thrombolysis results in a reduction inmortality (14). Although interhospital transfer ac-counts for a substantial postponement of reperfusionand subsequent myocardial damage (6,15), several tri-als showed no difference in the clinical outcomes ofpatients transferred between hospitals or admitteddirectly for PCI (4,16–18).

Considering this discrepancy in outcomes, weaimed to assess the influence of direct admissionversus transfer via regional hospital to a primary PCIcenter on time delays, early outcomes and 12-monthmortality from a real-life perspective in a large pop-ulation of STEMI patients treated by PCI procedures.We hypothesized that compared with interhospitaltransfer, direct admission to PCI centers would sub-stantially shorten the total ischemic time and, as aconsequence, limit the loss of left ventricular systolicfunction and reduce 12-month mortality.

METHODS

THE PL-ACS REGISTRY. We analyzed the data ofSTEMI patients from the PL-ACS registry (Polish reg-istry of Acute Coronary Syndromes), which is anongoing, prospective, observational, nationwide reg-istry of consecutive patients hospitalized with acutecoronary syndromes (ACS) in Poland. It was a jointinitiative of the Silesian Center for Heart Diseasesand the Polish Ministry of Health. The registry’s

methodology and analyses for STEMI patients con-cerning time delays have been previously published(19–21). In brief, all admitted patients with suspectedACS were screened for their eligibility to enter theregistry, but they were not enrolled until ACS wasconfirmed. STEMI was defined as the presence ofboth: 1) ST-segment changes on electrocardiogram(ECG) consistent with an infarction, that is, anelevation of $2 mm in the contiguous chest leadsand/or ST-segment elevation of $1 mm in 2 or morestandard leads or new left bundle branch block; and2) positive cardiac necrosis markers. The skilledphysicians who were in charge of each individualpatient collected the data. Internal checks for missingor conflicting data and values markedly out ofthe expected range were implemented using theregistry software. In the Silesian Centre for HeartDiseases data management and analysis center,further checks were applied if necessary. If thepatient was transferred via regional hospital to aprimary PCI center and both hospitalizations werereported, they were combined into 1 record for thefinal data management.

CURRENT ANALYSIS. The sample for the currentanalysis was patients hospitalized between January 1,2006, and December 31, 2013. The inclusion criteriawere STEMI presentation on admission and PCI as thefinal method of treatment. The exclusion criteriawere non–ST-segment elevation ACS presentation onadmission, time from onset of symptoms to admis-sion to PCI center longer than 12 h or missing data onthe time delays. The following time points (exact dateand time) were available in the registry: onset ofsymptoms, admission to the PCI center, and firstballoon inflation. For each patient, the time of theonset of symptoms was reported to the registry by thetreating physician who obtained detailed informationbased on clinical interview with the patient, patient’sfamily, or medical staff members present at the firstmedical contact in case of communication problemswith the patient. According to this, the following timedelays were calculated: onset of symptoms toadmission, admission to PCI, and onset of symptomsto PCI. According to the protocol, left ventricularejection fraction (LVEF) was defined as the latestavailable measure, preferably just before discharge,from echocardiography. The vital status at 12 monthswas obtained from the official mortality records fromthe government databases and was available for allpatients with an exact date of death.

Patients were stratified and analyzed in 2 groups:those directly admitted to a PCI center and thosetransferred from another non–PCI-capable facility to a

TABLE 1 Baseline and Clinical Characteristics of Patients Admitted Directly and Transferred

Entire Cohort Matched Cohort

Direct Admission(n ¼ 39,144)

Interhospital Transfer(n ¼ 30,947) p Value

Direct Admission(n ¼ 29,624)

Interhospital Transfer(n ¼ 29,624) p Value

Baseline characteristics—risk factors

Age (yrs) 63.6 � 11.9 63.5 � 11.9 0.18 63.5 � 11.9 63.6 � 11.9 0.44

Age $75 yrs 8,313 (21.2) 6,433 (20.8) 0.15 6,238 (21.1) 6,224 (21.0) 0.89

Females 12,105 (30.9) 9,730 (31.4) 0.15 9,143 (30.9) 9,277 (31.3) 0.23

Smoking (current or past) 23,401 (59.8) 18,207 (58.8) 0.011 17,447 (58.9) 17,321 (58.5) 0.29

Hypertension 24,891 (63.6) 20,007 (64.7) 0.004 19,012 (64.2) 19,228 (64.9) 0.064

Diabetes mellitus 8,376 (21.4) 7,182 (23.2) <0.001 6,806 (23.0) 6,938 (23.4) 0.20

Obesity (BMI >30 kg/m2) 6,741 (17.2) 5,634 (18.2) <0.001 5,334 (18.0) 5,491 (18.5) 0.095

Prior AMI 4,414 (11.3) 3,751 (12.1) <0.001 3,554 (12.0) 3,700 (12.5) 0.067

Prior PCI 2,786 (7.1) 2,103 (6.8) 0.097 2,074 (7.0) 2,070 (7.0) 0.95

Prior CABG 759 (1.9) 669 (2.2) 0.038 631 (2.1) 657 (2.2) 0.46

Baseline characteristic—clinical presentation

Anterior STEMI 15,509 (39.6) 12,730 (41.1) <0.001 12,057 (40.7) 12,084 (40.8) 0.82

Pre-hospital cardiac arrest 869 (2.2) 808 (2.6) <0.001 712 (2.4) 767 (2.6) 0.15

Killip class 3 599 (1.5) 559 (1.8) 0.004 516 (1.7) 546 (1.8) 0.35

Killip class 4 1,768 (4.5) 1,547 (5.0) 0.003 1,425 (4.8) 1,514 (5.1) 0.092

Baseline ECG findings

Sinus rhythm 36,518 (93.5) 28,965 (93.8) 0.11 27,681 (93.4) 27,749 (93.7) 0.26

Atrial fibrillation 1,609 (4.1) 1,318 (4.3) 0.33 1,242 (4.2) 1,276 (4.3) 0.49

Pacing 99 (0.25) 79 (0.26) 0.95 74 (0.25) 74 (0.25) 1.0

RBBB 1,010 (2.6) 836 (2.7) 0.32 778 (2.6) 802 (2.7) 0.54

LBBB 629 (1.6) 613 (2.0) <0.001 541 (1.8) 602 (2.0) 0.068

Time intervals (in minutes)

Onset of symptoms to admission 166 (105–283) 210 (134–345) <0.001 168 (105–285) 211 (133–345) <0.001

Admission to PCI 45 (30–72) 41 (30–63) <0.001 45 (30–74) 41 (30–63) <0.001

Onset of symptoms to PCI 230 (155–380) 270 (180–420) <0.001 235 (158–385) 270 (180–420) <0.001

Values are mean � SD, n (%), or median (interquartile range).

AMI ¼ acute myocardial infarction; CABG ¼ coronary artery bypass grafting; ECG ¼ electrocardiogram; LBBB ¼ left bundle branch block; PCI ¼ percutaneous coronaryintervention; RBBB ¼ right bundle branch block; STEMI ¼ ST-segment elevation myocardial infarction.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7 Kawecki et al.- 2 0 1 7 :- –- Direct Admission vs. Transfer for PCI in STEMI

3

PCI center. Directly admitted patients were thosereaching a PCI center directly from the place of onsetof symptoms by means of an emergency medicalsystem or reaching a PCI center by themselves. Directtransfer to PCI center and interhospital transfer wereidentified according to a dedicated field in the regis-try dataset.

The main outcome measure considered was12-month mortality. Secondary outcomes measureswere pre-hospital, in-hospital, and total time delays;LVEF; and in-hospital and 30-day mortality.

STATISTICS. The data are presented as counts andpercentages, mean � SD, or medians and interquartileranges (IQRs) and were compared between the groupswith the Student t test, the Mann-Whitney U test, orthe chi-square test with respect to data type and itsdistribution. Long-term survival was analyzed withthe Kaplan-Meier method and compared with the log-rank test. Multivariate logistic regression and Coxproportional hazard regression models were used to

analyze the influence of admission strategy (direct ortransfer) on in-hospital and 12-month mortality. Twomodels were defined. The baseline model (baselineand clinical characteristics on admission) includedthe following variables: type of admission (direct ortransfer), year of admission, age, sex, smoking, hy-pertension, diabetes, obesity, prior acute myocardialinfarction (AMI), prior PCI, prior coronary arterybypass grafting, anterior STEMI (versus other ECGlocalizations), pre-hospital cardiac arrest, Killip class3, Killip class 4 on admission, and initial ECG pattern(sinus rhythm, atrial fibrillation, pacing, and right orleft bundle branch block). The second model includedall variables from the baseline model plus the timefrom the onset of symptoms to admission to a PCIcenter.

Furthermore, for better control of selection andclinical bias, a propensity score method was used tomatch the groups for the year of admission andall baseline characteristics and parameters listedin Table 1. The package MatchIt, available in R

FIGURE 1 Patient Flow

PCI ¼ percutaneous coronary intervention; STEMI ¼ ST-segment elevation myocardial

infarction.

Kawecki et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7

Direct Admission vs. Transfer for PCI in STEMI - 2 0 1 7 :- –-

4

(R Foundation for Statistical Computing, Vienna,Austria), was used for the nearest neighbor matchingmethod with the distance option set to “logit” andcaliper to 0.1. All analyses were then recalculated forthe propensity-matched cohorts.

A p value <0.05 was considered the level of sig-nificance for all analyses. Statistics were calculatedwith STATISTICA 10 (Statsoft, Tulsa, Oklahoma), IBMSPSS Statistics version 22 (SPSS, Chicago, Illinois),and R statistical package, version 2.15.3.

RESULTS

Of the 132,715 STEMI patients from the PL-ACS reg-istry, 70,093 were included in the analysis on thebasis of the inclusion and exclusion criteria (Figure 1).From the studied population, 39,144 (55.8%) patientswere admitted directly to a primary PCI center, and30,947 (44.2%) were transferred to a primary PCI

center via a regional non–PCI-capable facility. Theproportion of direct admissions increased signifi-cantly over time and reached 68% in 2013 (Figure 2).There were no sex or age differences between thegroups. Patients transferred from non–PCI-capablefacilities had higher rates of cardiovascular risk fac-tors (hypertension, diabetes, obesity), prior AMI, andworse clinical presentations, with more frequent pre-hospital cardiac arrest, anterior AMI, and Killip clas-ses 3 or 4 on admission. Despite being statisticallysignificant, these differences in the clinical andbaseline characteristics were small in absolutenumbers (Table 1). Transferred patients wereadmitted to primary PCI centers with significantlylonger delays (absolute difference of median onset ofsymptoms to admission time was 44 min; p < 0.001);however, they had a shorter median admission-to-PCItime than patients admitted directly, by 4 min(Table 1, Figure 3). This resulted in a significantlylonger median time from the onset of symptoms toPCI (270 min [IQR]: 180 to 420 min] compared with228 min [IQR: 156 to 378 min] in the direct admissiongroup; p < 0.001). Among patients admitted directly,13.6% had PCI performed in <2 h from the symptomonset, compared with 9.0% of patients transferredfrom regional hospitals.

Data on the major medications used during thehospital stay and PCI procedures are shown inTable 2. Glycoprotein IIb/IIIa receptor inhibitors weremore frequently used in patients admitted directly,whereas P2Y12 receptor antagonists, beta-blockers,angiotensin-converting enzyme inhibitors and sta-tins were administered more often to transferredpatients. The need for diuretic administration in theacute phase was similar in both groups (18% of pa-tients; p ¼ 0.62). Total or subtotal occlusion of theinfarct-related artery on initial coronary angiogramwas present more frequently among patientsadmitted directly; however, the reperfusion successmeasured by final Thrombolysis In MyocardialInfarction (TIMI) flow grade 3 was achieved withequal frequency in both groups and was approxi-mately 92%.

LVEF was significantly higher in patients admitteddirectly than in transferred patients (47.5 � 10.2% vs.46.3 � 10.4%, respectively; p < 0.001). Comparison ofin-hospital adverse events revealed higher rates oftarget vessel revascularization in transferred pa-tients, with similar incidence of stroke, majorbleeding complications, and resuscitated cardiac ar-rest in both groups (Table 3). In-hospital mortalitywas slightly higher in patients admitted directly (by0.3%). Later, after 30 days, 6- and 12-month mortalitywas higher among transferred patients with an

FIGURE 2 Direct Admission and Interhospital Transfer Rates in STEMI

Direct admission and interhospital transfer rates in ST-segment elevation myocardial infarction (STEMI) from 2006 to 2013 in the entire

cohort (A) and a propensity-matched cohort (B).

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7 Kawecki et al.- 2 0 1 7 :- –- Direct Admission vs. Transfer for PCI in STEMI

5

absolute difference at 1 year of 0.9% in favor of directadmissions (Table 3, Figure 4).

In multivariate models regarding the in-hospitalperiod, interhospital transfer was significantly asso-ciated with lower mortality after adjustment forbaseline and clinical characteristics, and for the onsetof symptoms to admission time (Table 4). Conversely,in both multivariate models of 12-month mortality, itwas a significant predictor of higher mortality.

In a 1:1 match, 76% of patients admitted directly(29,624 from 39,144 patients) were matched with 96%of patients transferred (29,624 from 30,947 patients).The baseline characteristics in the matched cohortwere well-balanced between the groups admitteddirectly and transferred, with no statistically signifi-cant differences (Table 1, Figure 2). The in-hospitaland 12-month outcomes and the multivariatemodels for the matched cohort are presented inTables 3 and 4 and Figures 3 and 4. The results ach-ieved for the matched cohort were similar to thosedescribed for the entire cohort of patients.

DISCUSSION

The main finding of this study from the real-liferegistry is a 12-month mortality benefit after directadmission to PCI center as compared with inter-hospital transfer in an unselected cohort of STEMIpatients undergoing primary PCI. Interhospitaltransfer to PCI was identified as a predictor of higher12-month mortality after adjustment for baseline

characteristics and time from onset of symptoms toreperfusion in the entire cohort and additionally afterpropensity-score matching. By contrast, directadmission to a primary PCI center was associated withshorter time delays to PCI, higher LVEF, and lowerrates of 1-year mortality. Longer time to reperfusionwas previously described as a risk factor for adverseoutcome independent of admission strategy (22). Wehave shown that both factors, interhospital transferand longer time to reperfusion, might contributeindependently and additively to higher mortality,most probably via diminished preservation of leftventricular systolic function.

The contrary results of equal rates of mortality inpatients admitted directly and transferred were re-ported previously in randomized controlled trials(4,15–18). It was postulated that no difference inoutcome between patients admitted directly ortransferred could be explained by late presentation,exceeding the optimal 60 to 120 min from onset ofsymptoms, when time to reperfusion is critical foroutcome (15,23). During this time window, the mor-tality benefit decreased with time (24). In the PL-ASCregistry, time delays to primary PCI are still prolongeddespite a constant increase in the density of catheterlaboratories in the STEMI network, which doubledbetween 2006 and 2013 from 1.7 to 3.9 per millioninhabitants in a country with a population ofapproximately 38 million. As a result, in more than90% of cases, the distance to the nearest primaryPCI center is <90 km, and 748 primary PCI procedures

FIGURE 3 Cumulative Distribution of the Onset of Symptoms to Admission, Admission to PCI, and Onset of Symptoms to PCI Times

Cumulative distribution of the onset of symptoms to admission, admission to percutaneous coronary intervention (PCI), and onset of symptoms to PCI times in the entire

cohort (A, C, and E, respectively) and the propensity-matched cohort (B, D, and F, respectively). The numbers show medians and interquartile ranges.

Kawecki et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7

Direct Admission vs. Transfer for PCI in STEMI - 2 0 1 7 :- –-

6

TABLE 2 In-Hospital Drug Therapy and Reperfusion/Procedure Data for Patients Admitted Directly or Transferred

Entire Cohort Matched Cohort

Direct Admission(n ¼ 39,144)

Interhospital Transfer(n ¼ 30,947) p Value

Direct Admission(n ¼ 29,624)

Interhospital Transfer(n ¼ 29,624) p Value

In-hospital drug therapy

P2Y12 receptor antagonist 37,736 (96.4) 30,087 (97.2) <0.001 28,401 (95.9) 28,848 (97.4) <0.001

GP IIb/IIIa receptor inhibitor 13,735 (35.1) 9,277 (30.0) <0.001 10,186 (34.4) 8,926 (30.1) <0.001

Beta-blocker 28,761 (73.5) 23,883 (77.2) <0.001 22,050 (74.4) 22,812 (77.0) <0.001

Statin 31,332 (80.0) 25,838 (83.5) <0.001 23,965 (80.9) 24,708 (83.4) <0.001

ACE inhibitor 27,304 (69.8) 22,708 (73.4) <0.001 21,124 (71.3) 21,683 (73.2) <0.001

Diuretics 7,019 (17.9) 5,505 (17.8) 0.62 5,481 (18.5) 5,296 (17.9) 0.049

Reperfusion/procedure data

Thrombolysis 204 (0.52) 190 (0.61) 0.10 162 (0.55) 181 (0.61) 0.30

Stent implantation 36,545 (93.4) 28,983 (93.7) 0.12 27,588 (93.1) 27,745 (93.7) 0.009

Pre-PCI TIMI flow grade 0 or 1 31,391 (81.0) 23,498 (76.3) <0.001 23,850 (81.1) 22,527 (76.3) <0.001

Post-PCI TIMI flow grade 3 35,618 (91.8) 28,337 (91.9) 0.74 26,923 (91.5) 27,127 (91.9) 0.091

CABG 45 (0.11) 73 (0.24) <0.001 44 (0.15) 68 (0.23) 0.023

IABP 511 (1.3) 515 (1.7) <0.001 408 (1.4) 499 (1.7) 0.002

Values are n (%).

ACE ¼ angiotensin-converting enzyme; GP ¼ glycoprotein; IABP ¼ intra-aortic balloon pump; TIMI ¼ Thrombolysis In Myocardial Infarction; other abbreviations as in Table 1.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7 Kawecki et al.- 2 0 1 7 :- –- Direct Admission vs. Transfer for PCI in STEMI

7

per million inhabitants were performed in Poland in2013. Regardless of the continuous optimization ofinvasive treatment and the high incidence of AMI(25), transport to a primary PCI center remains one ofthe major phases and concerns of STEMI care becausethe majority of patients in both groups in the currentanalysis were treated outside the timespans requiredfor the greatest impact of reperfusion on myocardialsalvage. However, the adjusted data indicate thateven over these timespans, a shorter time to

TABLE 3 In-Hospital Outcomes and 1-Year Mortality for Patients Adm

Entire Cohort

Direct Admission(n ¼ 39,144)

Interhospital Trans(n ¼ 30,947)

In-hospital outcomes

LVEF (%) 47.5 � 10.2 46.3 � 10.4

Resuscitated cardiac arrest 1,519 (3.9) 1,154 (3.7)

Target vessel revascularization 293 (0.75) 286 (0.92)

Stroke 117 (0.30) 87 (0.28)

Bleeding 544 (1.4) 432 (1.4)

Death 1,734 (4.4) 1,273 (4.1)

Mortality

30 days 2,215 (5.7) 1,909 (6.2)

6 months 3,108 (7.9) 2,704 (8.7)

12 months 3,696 (9.4) 3,177 (10.3)

Post-discharge mortality

30 days 714 (1.9) 876 (3.0)

6 months 1,571 (4.2) 1,662 (5.6)

12 months 2,159 (5.8) 2,135 (7.2)

Values are mean � SD n (%).

LVEF ¼ left ventricular ejection fraction.

reperfusion results in better outcomes, according tothe hypothesis of LVEF preservation with the short-ening of total ischemic time.

The lower in-hospital mortality of patients trans-ferred via regional hospital to a PCI center requires abrief comment. This outcome measure was studiedpreviously in regard to interhospital transfer and wasfound to be equal in patients transferred andadmitted directly (22,26). In our cohort, patientstransferred to a PCI center had better in-hospital

itted Directly or Transferred

Matched Cohort

ferp Value

Direct Admission(n ¼ 29,624)

Interhospital Transfer(n ¼ 29,624) p Value

<0.001 47.4 � 10.2 46.2 � 10.4 <0.001

0.30 1,221 (4.1) 1,115 (3.8) 0.025

0.011 223 (0.75) 271 (0.91) 0.030

0.66 88 (0.30) 85 (0.29) 0.82

0.94 416 (1.4) 404 (1.4) 0.67

0.04 1,363 (4.6) 1,231 (4.2) 0.008

0.004 1,714 (5.8) 1,848 (6.2) 0.021

<0.001 2,382 (8.0) 2,620 (8.8) <0.001

<0.001 2,832 (9.6) 3,071 (10.4) <0.001

<0.001 536 (1.9) 848 (3.0) <0.001

<0.001 1,178 (4.2) 1,611 (5.7) <0.001

<0.001 1,628 (5.8) 2,062 (7.3) <0.001

FIGURE 4 12-Month Mortality

Kaplan-Meier analysis of 12-month mortality in the entire cohort (A) and the propensity-matched cohort (B).

Kawecki et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7

Direct Admission vs. Transfer for PCI in STEMI - 2 0 1 7 :- –-

8

survival despite a worse baseline clinical profile andalso after adjustments. Of note is the lower rate ofinitial TIMI flow grade 0 or 1 among transferred pa-tients; however, PCI was equally successful in bothgroups, as measured by the final TIMI flow grade 3.This paradoxical finding of lower in-hospital mortal-ity in transferred patients could be affected by thebias caused by the longer delay of reaching the PCIhospital, causing a higher proportion of patientsto die before PCI could be performed comparedwith patients admitted directly. As a result, thetransferred group may be positively pre-selected.Nevertheless, the rates of mortality were inverted in

TABLE 4 Multivariate Analyses of In-Hospital and 12-Month Mortalit

Odds Rat

In-hospital mortality

Adjusted in baseline model*

Interhospital transfer (vs. direct admission) 0.85 (0

Baseline model þ onset of symptoms to admission time

Interhospital transfer (vs. direct admission) 0.83 (0

Onset of symptoms to admission time (per every 3 h) 1.10 (1

12-month mortality

Adjusted in baseline model

Interhospital transfer (vs. direct admission) 1.08 (1

Baseline model þ onset of symptoms to admission time

Interhospital transfer (vs. direct admission) 1.07 (1

Onset of symptoms to admission time (per every 3 h) 1.04 (1

*Baseline model of multivariate Cox proportional hazard regression included parameters:years more), female, smoking, hypertension, diabetes mellitus, obesity, prior myocardigrafting, anterior ST-segment elevation myocardial infarction, pre-hospital cardiac arrest,pacing, right and left bundle branch block.

the post-discharge period, with a higher mortality intransferred patients at 30 days and at 12 months.

Another general point emerging from the study isthe high proportion of direct admissions, exceeding50%, compared with approximately one-third re-ported in previous trials and registries (6,15,16).Starting with a 70% rate of fibrinolysis in the late1990s (27), consistent efforts to increase the quality ofSTEMI care in Poland, increase in the density of PCI-capable centers, well-established education of inva-sive cardiologists, subsequent increase in PCI centersworking in a 24/7 system, and finally introduction ofpre-hospital ECG transfer directly from an ambulance

y

Entire Cohort Matched Cohort

io (95% CI) p Value Odds Ratio (95% CI) p Value

.78–0.92) <0.001 0.84 (0.77–0.92) <0.001

.76–0.91) <0.001 0.82 (0.75–0.90) <0.001

.05–1.15) <0.001 1.08 (1.03–1.14) <0.001

.03–1.13) 0.002 1.06 (1.01–1.12) 0.016

.02–1.12) 0.006 1.06 (1.01–1.11) 0.031

.01–1.06) 0.007 1.03 (1.01–1.06) 0.039

interhospital transfer (vs. direct admission), year of admission (per 1 more), age (per 5al infarction, prior percutaneous coronary intervention, prior coronary artery bypassKillip class 3 on admission, Killip class 4 on admission, sinus rhythm, atrial fibrillation,

PERSPECTIVES

WHAT IS KNOWN? Reduction of delays to reperfusion is

crucial in a ST-segment elevation myocardial infarction (STEMI)

system of care. However, it is still debated whether direct

admission to a percutaneous coronary intervention (PCI) center is

superior to interhospital transfer in terms of long-term

prognosis.

WHAT IS NEW? Direct admission to a primary PCI center seems

to be better than involvement of a regional hospital and subse-

quent transfer for primary PCI and should be considered the

preferred strategy in all STEMI cases.

WHAT IS NEXT? Transfer for primary PCI from regional

hospitals cannot be eliminated from the STEMI system of care in

the near future. Therefore, further analyses focused on patients

being transferred should address the opportunities for maximal

shortening of delays at the level of regional hospitals and

transport services.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7 Kawecki et al.- 2 0 1 7 :- –- Direct Admission vs. Transfer for PCI in STEMI

9

to a PCI center, resulted in the current marginal ratesof thrombolysis and high rates of primary PCI. Mainlythe latter (ECG transfer) is responsible for high per-centage of direct admissions to PCI canters in ourcountry. As a result, the crude rates of 30-day and12-month mortality in our registry are comparable toor even lower than those reported world-wide, eitherin real-life studies or randomized controlled trials,taking into account that high-risk patients were notexcluded from our study (6,16,28,29).

The major aim of all initiatives to improve STEMIcare is to maximally shorten the delay from the onsetof chest pain to reperfusion because there is a corre-lation between a longer delay to reperfusion and thesubsequent increase in infarct size (6) and mortality(30). Our study, based on this hypothesis, shows theimportance of the pre-PCI strategy in shortening theischemia time. The between-groups difference inadmission-to-PCI time was significant but accountedfor only a 4-min longer median delay in patentsadmitted directly than in those that were transferred.Therefore, the calculated difference in total ischemictime may be attributable exclusively to the pre-PCIcenter delay. Considering this, direct admission forprimary PCI seems to be the preferred strategy forreduction of total time delays in STEMI. If directadmission is unavailable, efforts should be focusedon maximal shortening of the delay related to pe-ripheral hospitals (door-in door-out time) as analternative target to reduce delays (31).

STUDY LIMITATIONS. For the current analysis, weattributed longer delays in the transferred groupsentirely to the fact that the patient was transferred,whereas patient-related delays, unavailable for theanalysis, could balance the difference in time to arrivalat the primary PCI center. Additionally, consideringtransfer from a peripheral hospital as a risk factor, wecannot provide insights into the partial influence of itscomponents in the worsening of outcomes. Informa-tion on the time of the first medical contact, time ofarrival to peripheral hospital, the door-in door-outtime, and delay related to transport service were un-available for the analysis. Finally, the data on theclinical course of patients from the pre-primary PCI

period were unavailable for the analysis and resultedin blanking of a subset of patients who died in a pe-ripheral hospital before reaching a PCI center. Thiscould result in additional selection bias in the transfergroup, making this group rather positively than nega-tively pre-selected. Therefore, it should not influencethe obtained results.

CONCLUSIONS

In a large, community-based cohort of patients withSTEMI treated by PCI, direct admission to a primaryPCI center was associated with lower 12-month mor-tality and should be preferred to transfer via aregional non–PCI-capable facility.

ADDRESS FOR CORRESPONDENCE: Dr. DamianKawecki, 2nd Department of Cardiology, School ofMedicine with the Division of Dentistry in Zabrze,Medical University of Silesia, Katowice, Poland,M. Sklodowskiej-Curie 10, 41-800 Zabrze, Poland.E-mail: d.kawecki@interia.pl.

RE F E RENCE S

1. Bates ER, JacobsAK. Time to treatment inpatientswith STEMI. New Engl J Med 2013;369:889–92.

2. Le May MR, Wells GA, So DY, et al. Reduction inmortality as a result of direct transport from thefield to a receiving center for primary percuta-neous coronary intervention. J Am Coll Cardiol2012;60:1223–30.

3. Ting HH, Rihal CS, Gersh BJ, et al. Regionalsystems of care to optimize timeliness of reper-fusion therapy for ST-elevation myocardialinfarction: the Mayo Clinic STEMI Protocol. Circu-lation 2007;116:729–36.

4. Henry TD, Sharkey SW, Burke MN, et al.A regional system to provide timely access to

percutaneous coronary intervention for ST-elevation myocardial infarction. Circulation 2007;116:721–8.

5. Grines CL, Westerhausen DR Jr., Grines LL, et al.A randomized trial of transfer for primary angio-plasty versus on-site thrombolysis in patients withhigh-risk myocardial infarction: the Air Primary

Kawecki et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . - , N O . - , 2 0 1 7

Direct Admission vs. Transfer for PCI in STEMI - 2 0 1 7 :- –-

10

Angioplasty in Myocardial Infarction study. J AmColl Cardiol 2002;39:1713–9.

6. Liebetrau C, Szardien S, Rixe J, et al. Directadmission versus transfer of AMI patients for pri-mary PCI. Clin Res Cardiol 2011;100:217–25.

7. McNamara RL, Wang YF, Herrin J, et al. Effect ofdoor-to-balloon time on mortality in patients withST-segment elevation myocardial infarction. J AmColl Cardiol 2006;47:2180–6.

8. Levine GN, Bates ER, Blankenship JC, et al. 2011ACCF/AHA/SCAI guideline for percutaneous coro-nary intervention: executive summary: a report ofthe American College of Cardiology Foundation/American Heart Association Task Force on PracticeGuidelines and the Society for CardiovascularAngiography and Interventions. J Am Coll Cardiol2016;67:1235–50.

9. Krumholz HM, Anderson JL, Bachelder BL, et al.ACC/AHA 2008 performance measures for adultswith ST-elevation and non-ST-elevation myocar-dial infarction: a report of the American College ofCardiology/American Heart Association Task Forceon Performance Measures (Writing Committee toDevelop Performance Measures for ST-Elevationand Non-ST-Elevation Myocardial Infarction).J Am Coll Cardiol 2008;52:2046–99.

10. Bradley EH,Herrin J,WangYF, et al. Strategies forreducingthedoor-to-balloontime inacutemyocardialinfarction. New Engl J Med 2006;355:2308–20.

11. Aguirre FV, Varghese JJ, Kelley MP, et al. Ruralinterhospital transfer of ST-elevation myocardialinfarction patients for percutaneous coronaryrevascularization: the Stat Heart Program. Circu-lation 2008;117:1145–52.

12. Jollis JG, Al-Khalidi HR, Roettig ML, et al.Regional Systems of Care Demonstration Project:American Heart AssociationMission: Lifeline STEMISystems Accelerator. Circulation 2016;134:365–74.

13. MeneesDS, PetersonED,WangY, et al. Door-to-balloon time and mortality among patients under-goingprimaryPCI.NewEngl JMed2013;369:901–9.

14. Boersma E, Maas ACP, Deckers JW,Simoons ML. Early thrombolytic treatment inacute myocardial infarction: reappraisal of thegolden hour. Lancet 1996;348:771–5.

15. Wohrle J, Desaga M, Metzger C, et al. Impactof transfer for primary percutaneous coronaryintervention on survival and clinical outcomes

(from the HORIZONS-AMI trial). Am J Cardiol2010;106:1218–24.

16. Busk M, Maeng M, Rasmussen K, et al. TheDanish multicentre randomized study of fibrino-lytic therapy vs. primary angioplasty in acutemyocardial infarction (the DANAMI-2 trial):outcome after 3 years follow-up. Eur Heart J2008;29:1259–66.

17. Brodie BR, Stuckey TD, Hansen CJ, et al. Effectof treatment delay on outcomes in patients withacute myocardial infarction transferred fromcommunity hospitals for primary percutaneouscoronary intervention. Am J Cardiol 2002;89:1243–7.

18. Andersen HR, Nielsen TT, Rasmussen K, et al.A comparison of coronary angioplasty with fibri-nolytic therapy in acute myocardial infarction.New Engl J Med 2003;349:733–42.

19. Polonski L, Gasior M, Gierlotka M, et al. PolishRegistry of Acute Coronary Syndromes (PL-ACS)characteristics, treatments and outcomes of pa-tients with acute coronary syndromes in Poland.Kardiol Pol 2007;65:861–72.

20. Polonski L, Gasior M, Gierlotka M, et al. Whathas changed in the treatment of ST-segmentelevation myocardial infarction in Poland in2003-2009? Data from the Polish Registry ofAcute Coronary Syndromes (PL-ACS). Kardiol Pol2011;69:1109–18.

21. Gierlotka M, Gasior M, Wilczek K, et al.Reperfusion by primary percutaneous coronaryintervention in patients with ST-segment eleva-tion myocardial infarction within 12 to 24 hours ofthe onset of symptoms (from a prospective na-tional observational study [PL-ACS]). Am J Cardiol2011;107:501–8.

22. Manari A, Ortolani P, Guastaroba P, et al.Clinical impact of an inter-hospital transfer strat-egy in patients with ST-elevation myocardialinfarction undergoing primary angioplasty: theEmilia-Romagna ST-segment elevation acutemyocardial infarction network. Eur Heart J 2008;29:1834–42.

23. Silvain J, Vignalou JB, Beygui F, et al. Impactof transfer time on mortality in acute coronarysyndrome with ST-segment elevation treated byangioplasty. Arch Cardiovasc Dis 2012;105:639–48.

24. Gersh BJ, Stone GW, White HD, Holmes DR.Pharmacological facilitation of primary percuta-neous coronary intervention for acute myocardialinfarction: Is the slope of the curve the shape ofthe future? JAMA 2005;293:979–86.

25. Gierlotka M, Zdrojewski T, Wojtyniak B, et al.Incidence, treatment, in-hospital mortality andone-year outcomes of acute myocardial infarctionin Poland in 2009-2012—nationwide AMI-PLdatabase. Kardiol Pol 2015;73:142–58.

26. Stone GW. Angioplasty strategies in ST-segment-elevation myocardial infarction: part I:primary percutaneous coronary intervention. Cir-culation 2008;118:538–51.

27. Wodniecki J, Polo�nski L, Kawecki D,Walichiewicz P, Fudal M, Wilczewski P. Wyniki akcjipromujacej leczenie fibrynolityczne zawału sercana terenie województwa katowickiego w latach1996-1998 [Results of promoting fibrinolytictherapy for myocardial infarction in the KatowiceProvince from 1996-1998]. Pol Arch Med Wewn1999;101:233–6.

28. Nallamothu BK, Normand SLT, Wang YF, et al.Relation between door-to-balloon times andmortality after primary percutaneous coronaryintervention over time: a retrospective study.Lancet 2015;385:1114–22.

29. Widimsky P, Budesinsky T, Vorac D, et al. Longdistance transport for primary angioplasty vs im-mediate thrombolysis in acute myocardial infarc-tion: final results of the randomized nationalmulticentre trial - PRAGUE-2. Eur Heart J 2003;24:94–104.

30. Terkelsen CJ, Sorensen JT, Maeng M, et al.System delay and mortality among patients withSTEMI treated with primary percutaneous coro-nary intervention. JAMA 2010;304:763–71.

31. Dauerman HL, Bates ER, Kontos MC, et al.Nationwide analysis of patients with ST-segment-elevation myocardial infarction trans-ferred for primary percutaneous intervention:findings from the American Heart AssociationMission: Lifeline Program. Circ Cardiovasc Interv2015;8:e002450.

KEY WORDS 12-month mortality, acutemyocardial infarction, interhospital transfer,primary PCI, STEMI networks, time delays