Primary Coronary Angioplasty vs Thrombolysis for the Management

ORIGINAL CONTRIBUTION

Primary Coronary Angioplasty vsThrombolysis for the Management of AcuteMyocardial Infarction in Elderly PatientsAlan K. Berger, MDKevin A. Schulman, MDBernard J. Gersh, MB, ChB, DPhilSarmad Pirzada, MD, MPHJeffrey A. Breall, MD, PhDAyah E. Johnson, PhDNathan R. Every, MD, MPH

ACUTE MYOCARDIAL INFARC-tion (AMI) is the leadingcause of death in elderly pa-tients. The choice of an op-

timal management strategy for pa-tients with AMI has been addressed inmultiple clinical trials and summa-rized in a meta-analysis of trials com-paring thrombolysis with placebo. Themeta-analysis found a significant ben-efit of thrombolysis in patients youngerthan 75 years, but only a trend towarddecrease in mortality rates in patientsaged 75 years or older.1 Supporters ofprimary percutaneous transluminalcoronary angioplasty (PTCA) empha-size the procedure’s higher early pat-ency rate, lower rates of death and re-current reinfarction, and markedlyreduced rate of stroke. A recent meta-analysis of randomized clinical trialsthat compared thrombolytic therapywith primary PTCA suggests that PTCAdecreases short-term mortality and theincidence of recurrent infarction.2 Ob-servational studies in unselected pa-tients have demonstrated similar out-

AuthorAffiliations: InstituteforCardiovascularSciences,DivisionofCardiology(DrsBerger,Schulman,Gersh,andBreall),andtheClinicalEconomicsResearchUnit,Depart-mentofMedicine (DrsBerger, Schulman,and Johnson),GeorgetownUniversityMedicalCenter,Washington,DC;Delmarva Foundation for Medical Care Inc, Easton, Md(DrBerger); andtheNorthwestHealthServicesResearchandDevelopmentFieldProgram,VeteransAffairsPugetSoundHealthcareSystemandCardiovascularOutcomesResearchCenterand theDivisionofCardiology,Univer-sity of Washington, Seattle (Drs Pirzada and Every).

Dr Berger is currently a fellow in the Section of Cardio-vascularMedicine,Departmentof InternalMedicine,YaleUniversity School of Medicine, New Haven, Conn.Financial Disclosure: Dr Every has a research contractwith Genentech Inc, manufacturer of a tissue plasmino-gen activator.Corresponding Author and Reprints: Kevin A. Schul-man, MD, Clinical Economics Research Unit, George-town University Medical Center, 2233 Wisconsin AveNW, Suite 440, Washington, DC 20007 (e-mail:[email protected]).

See also Patient Page.

Context Despite evidence from randomized trials that, compared with early throm-bolysis, primary percutaneous transluminal coronary angioplasty (PTCA) after acutemyocardial infarction (AMI) reduces mortality in middle-aged adults, whether elderlypatients with AMI are more likely to benefit from PTCA or early thrombolysis is notknown.

Objective To determine survival after primary PTCA vs thrombolysis in elderly pa-tients.

Design The Cooperative Cardiovascular Project, a retrospective cohort study usingdata from medical charts and administrative files.

Setting Acute care hospitals in the United States.

Patients A total of 20 683 Medicare beneficiaries, who arrived within 12 hours ofthe onset of symptoms, were admitted between January 1994 and February 1996 witha principal discharge diagnosis of AMI, and were eligible for reperfusion therapy.

Main Outcome Measures Thirty-day and 1-year survival.

Results A total of 80 356 eligible patients had an AMI at hospital arrival and met theinclusion criteria, of whom 23.2% received thrombolysis and 2.5% underwent pri-mary PTCA within 6 hours of hospital arrival. Patients undergoing primary PTCA hadlower 30-day (8.7% vs 11.9%, P = .001) and 1-year mortality (14.4% vs 17.6%,P = .001). After adjusting for baseline cardiac risk factors and admission and hospitalcharacteristics, primary PTCA was associated with improved 30-day (hazard ratio [HR]of death, 0.74; 95% confidence interval [CI], 0.63-0.88) and 1-year (HR, 0.88; 95%CI, 0.73-0.94) survival. The benefits of primary coronary angioplasty persisted whenstratified by hospitals’ AMI volume and the presence of on-site angiography. In pa-tients classified as ideal for reperfusion therapy, the mortality benefit of primary PTCAwas not significant at 1-year follow-up (HR, 0.92; 95% CI, 0.78-1.08).

Conclusion In elderly patients who present with AMI, primary PTCA is associatedwith modestly lower short- and long-term mortality rates. In the subgroup of patientswho were classified as ideal for reperfusion therapy, the observed benefit of primaryPTCA was no longer significant.JAMA. 1999;282:341-348 www.jama.com

©1999 American Medical Association. All rights reserved. JAMA, July 28, 1999—Vol 281, No. 4 341

Downloaded From: on 03/22/2018

comes in patients undergoing primaryPTCA compared with those receivingthrombolysis.3,4

The results of randomized clinical tri-als may be difficult to extrapolate to el-derly patients, who are more likely tohave extensive coronary artery dis-ease, additional risk factors, and othercomorbid conditions that may influ-ence decisions about the appropriatereperfusion strategy. In the absence ofconclusive evidence from randomizedtrials, data from observational studiesmay be used to assess treatment out-comes in clinical practice. Observa-tional studies reflect general practice inthe community and general commu-nity standards rather than care pro-vided in highly specialized centers. Inlight of these issues, we compared theclinical outcomes of elderly patients withAMI who were treated with eitherthrombolytic therapy or primary PTCA.

METHODSStudy Sample

The Cooperative Cardiovascular Project(CCP) was initiated by the Health CareFinancing Administration as an ongo-ing national program to improve thequality of care for Medicare beneficia-ries with AMI.5,6 In this study, we useda database developed from the CCP thatincludes detailed clinical data on pa-tients with a principal discharge diag-nosis of AMI (International Classifica-tion of Diseases, 9th Revision, ClinicalModification code 410) between Janu-ary 1994 and February 1996.

We restricted the study population topatients presenting within 12 hours ofsymptom onset with evidence of AMI atthe time of hospital arrival. We definedAMI as either an elevation of the cre-atine kinase-MB fraction level (.5%),a 2-fold elevation in creatine kinase level,or diagnostic electrocardiogram changes(ST elevation or new Q waves). Pa-tients younger than 65 years were ex-cluded. Patients presenting with cardio-genic shock were excluded based on thepreferential use of primary PTCA in thishigh-risk population.7 We restricted thestudy group further to patients with-out contraindications to thrombolysis

(history of bleeding disorder, documen-tation of prior internal bleeding, activebleeding on arrival, recent trauma or sur-gery, and cardiopulmonary resuscita-tion on arrival).

The CCP data collection process fo-cuses on patients’ hospitalizations anddoes not extract information from otherinstitutions’ records when a patient istransferred. Therefore, transfers fromother hospitals were excluded from theprimary analysis. Patients transferredto another facility were included be-cause their initial diagnostic and thera-peutic strategy could be identified. Theanalysis was further limited to the firstAMI admission for any given indi-vidual in the CCP cohort (TABLE 1).

Data CollectionThe data elements we analyzed havebeen reported previously and in-cluded 140 clinical variables for eachpatient with AMI.8 Data abstraction in-cluded patient demographics, past car-diac and noncardiac history, admis-sion characteristics, diagnostic testresults, and information on in-hospital events and procedures. Docu-mentation of PTCA required that acoronary intervention was attempted;cardiac catheterizations without asso-ciated coronary interventions were ex-cluded. The time from onset of symp-toms to hospital arrival was categorizedas fewer than 6 hours, 6 to 12 hours,or longer than 12 hours. In-hospitalevents, including when thrombolysiswas initiated and when PTCA wasstarted, were obtained directly from thenursing and procedure records.

Hospital characteristics, includingnumber of hospital beds, annual vol-ume of AMI cases, and the capability toperform on-site PTCA, were obtained bylinking the CCP data set with informa-tion obtained from the American Hos-pital Association.9 We categorized eachhospital based on the annual number ofAMI patients; a cutoff of 150 AMI ad-missions per year (representing the 50thpercentile) separated high– and low AMI–volume facilities. Similarly, a cutoff of 250beds (50th percentile) was used to sepa-rate large from small hospitals.

For missing data elements, we usedindicator variables in our analysis. Cat-egorical variables, when undocu-mented, were considered to be absentfrom the patient’s history.

The primary clinical end points of theanalysis were 30-day and 1-year mor-tality rates, both of which were ascer-tained from the Medicare enrollment da-tabase. Secondary end points includedpost-AMI angina, in-hospital reinfarc-tion, heart failure, hemorrhage (ie, gas-trointestinal, genitourinary, and pul-monary hemorrhage), intracerebralhemorrhage, ischemic or hemorrhagicstroke, and performance of cardiac cath-eterization, coronary angioplasty, andcoronary artery bypass surgery.

Statistical AnalysisThe study sample was divided into 2 co-horts. The cohort receiving thrombo-lytic therapy included patients who re-ceived tissue plasminogen activator orstreptokinase within 6 hours of arrivalat the hospital. The cohort undergoingprimary PTCA consisted of patients whohad not undergone prior thrombolysisand received PTCA within 6 hours ofhospital arrival. For the 1.5% of pa-tients who underwent both thromboly-sis and PTCA, cohort assignment wasbased on the first strategy initiated. The6-hour cutoff was designed to select pa-tients who would derive the maximalbenefit from either reperfusion strat-egy and to distinguish primary PTCAcases from subsequent urgent or elec-tive PTCAs.

Categorical characteristics were com-pared using the x2 test and expressedas proportions. Differences in continu-ous variables were compared using thet test. The use of the Medicare enroll-ment database to establish the time ofdeath has been validated in the litera-ture.10 Age-adjusted mortality rates forboth sexes were calculated by the di-rect method using 1990 census data.11

Overall survival was determined foreach reperfusion strategy by the Kaplan-Meier method.12 Homogeneity of thesurvival curves was tested with both thelog-rank test and the Wilcoxon ranksum test.

PRIMARY CORONARY ANGIOPLASTY VS THROMBOLYSIS

342 JAMA, July 28, 1999—Vol 281, No. 4 ©1999 American Medical Association. All rights reserved.


Multivariate correlates of 30-day sur-vival were analyzed with a Cox model.13

Candidate predictor variables for theanalysis were those variables shown tobe important predictors of mortality inpublishedstudiesandthosevariables thatdiffered between patients receiving the2treatments:cardiacriskfactors(age,sex,diabetes,hypertension,andtobaccouse),race, functional status (emphysema,impaired mobility, dementia, and priorstroke), cardiac history (prior myocar-dial infarction, prior heart failure, andprior PTCA or coronary artery bypasssurgery), presenting features (cardiacarrest, hypotension, bradycardia, andheart failure), electrocardiographic fea-tures (Qwave,STelevation, rightand leftbundle-branch block, atrial fibrillation,and anterior location of infarction),admission serum creatinine levels, coro-nary artery bypass surgery during theindex hospitalization, and hospital char-acteristics (numberofbeds, annualnum-ber of patients treated with AMI, andcapability to perform PTCA).14 All can-didatepredictorvariables thatwere iden-tified as important predictors of mortal-ity by univariate analysis (P,.10) wereincluded in the final multivariate Coxmodel.13 Wereport thehazardratio (HR)and 95% confidence interval (CI) asso-ciatedwithprimaryPTCAvs thromboly-sis for the general CCP cohort at 30 daysand 1 year. The HR and 95% CIs for theprimary reperfusion strategy were thenstratified by patient category (older than

75 years, sex, presence of hypertension,diabetes, or congestive heart failure onarrival, and the location of the infarct)by means of the 30-day Cox models. Thestatistical analyses were performed withthe SAS 6.12 statistical analysis soft-ware package.14

Using institutional characteristics tostratify the population, we further vali-dated our findings by performing a se-

ries of subgroup analyses. We first de-fined an ideal subgroup by restrictingour cohort to patients with either ST el-evation or left bundle-branch block onarrival and who presented to the hos-pital within 6 hours of symptom on-set. To evaluate the method of reper-fusion in institutions with minimal andextensive experience treating AMI, werepeated our analyses after classifying

Table 1. Derivation of Study Sample FromCooperative Cardiovascular Project Cohort*

Original CCP cohort 234 769Exclusion criteria

Transfer from other hospital 42 177Repeat admission in CCP 23 773Age <65 y 17 593Cardiogenic shock on arrival 5967Chest pain not reported 49 649Duration of symptoms unknown 26 376Symptom onset exceeding 12 h 35 226Contraindication to thrombolysis 43 234

Reperfusion-eligible 80 356Reperfusion therapy

delayed more than 6 h59 673

Study sample 20 683Percutaneous transluminal

coronary angioplasty2038

Thrombolysis 18 645

*CCP indicates Cooperative Cardiovascular Project. Alldata are presented as number of patients. More than 1criterion could apply to a single patient.

Table 2. Baseline Clinical and Hospital Characteristics of Patients Receiving Early ReperfusionTherapy*

ThrombolysisGroup

(n = 18 645)

PrimaryPTCA Group

(n = 2038)P

Value

DemographicsAge, mean (SD), y 73.2 (6.0) 73.3 (6.0) .11

Median age (25th, 75th percentiles), y 72 (68, 77) 72 (68, 77) NA

Women 7907 (42.4) 798 (39.2) .005

White race 17 280 (92.7) 1911 (93.8) .07

Risk factorsHypertension 10 407 (55.8) 1168 (57.3) .20

Diabetes 4252 (22.8) 451 (22.1) .49

Smoker 3735 (20.0) 382 (18.7) .17

Noncardiac historyPrior stroke 1230 (6.6) 186 (9.1) .001

COPD 2931 (15.7) 291 (14.3) .09

Dementia 323 (1.7) 24 (1.2) .06

Limited mobility 1594 (8.6) 195 (9.6) .12

Cardiac historyPrior MI 3961 (21.2) 418 (20.5) .44

Prior heart failure 1181 (6.3) 111 (5.4) .12

Prior coronary angioplasty 1296 (7.0) 274 (13.4) .001

Prior bypass surgery 1535 (8.2) 236 (11.6) .001

Clinical presentationSymptom onset <6 h 17 141 (91.9) 1783 (87.5) .001

Mean (SD) time to treatmentafter arrival, min

68.3 (54.4) 142.6 (68.1) <.001

Hypotension (SBP <90 mm Hg) 315 (1.7) 33 (1.7) .83

Initial BP data missing 335 (1.8) 45 (2.2) .41

Bradycardia (HR <60/min) 3343 (17.9) 364 (17.9) .92

Initial HR data missing 25 (0.1) 1 (0.05) .59

Heart failure on examination 2278 (12.2) 205 (10.1) .004

Serum creatinine, mean (SD)µmol/L 106 (44) 106 (70)

.64mg/dL 1.2 (0.5) 1.2 (0.8)

Electrocardiographic featuresST elevation 12 688 (68.0) 1256 (61.6) .001

Q-wave MI 5640 (30.2) 479 (23.5) .001

Anterior location 9994 (53.6) 1083 (53.1) .69

LBBB 325 (1.7) 50 (2.4) .02

Hospital characteristicsHigh annual MI volume (>150/y) 8889 (47.7) 1676 (82.2) .001

Large hospital (>250 beds) 8563 (45.9) 1529 (75.0) .001

PTCA service available 6805 (36.5) 2038 (100) .001

*Data presented as number (percentage) unless otherwise indicated. PTCA indicates percutaneous transluminal coro-nary angioplasty; NA, not applicable; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; SBP,systolic blood pressure; BP, blood pressure; HR, heart rate; and LBBB, left bundle-branch block.




hospitals as either low– or high AMI–volume facilities. In the final sub-group analysis, we evaluated the ef-fect of the treating hospital on the out-come of reperfusion therapy. Westratified hospitals according to their ca-pability to perform PTCA and com-pared the mortality rates of institu-tions that made both reperfusionmodalities available.

Finally, we compared overall mor-tality rates between patients receivingreperfusion therapy and those not re-ceiving reperfusion therapy using thex2 test.

RESULTSBaseline Characteristics

From the CCP database, 80 356 pa-tients had an AMI at the time of hos-pital arrival and met the inclusion cri-teria of the study. Table 1 shows the

derivation of the study sample. In thiscohort, 18 645 patients (23.2%) re-ceived thrombolysis and 2038 pa-tients (2.5%) underwent PTCA withinthe 6-hour time frame. Among the pa-tients receiving thrombolysis, 76.5% re-ceived tissue plasminogen activator,22.6% received streptokinase, 0.7% re-ceived anistreplase, and 0.2% receivedurokinase. The remaining 59 673 pa-tients (74.2%) did not receive reperfu-sion therapy during the first 6 hours af-ter hospital arrival, and 54 989 (68.4%)did not receive PTCA or thrombolytictherapy at any time.

TABLE 2 shows a comparison of thebaseline characteristics of patients whoreceived early reperfusion therapy bythrombolysis or PTCA group. The mean(SD) age of the cohort was 73.2 (6.0)years, and 42.1% of the patients were fe-male. There was no significant differ-

ence in the distribution of age, sex, orrace between the 2 groups. The preva-lence of cardiac risk factors and non-cardiac conditions was similar be-tween the 2 groups; however, priorPTCA, prior stroke, and prior coro-nary artery bypass surgery were moreprevalent among patients undergoingPTCA.

Patients undergoing primary PTCApresented longer after symptom onsetthan patients receiving thrombolysis.Furthermore, more time lapsed beforeinitiation of treatment for patients un-dergoing primary PTCA (mean [SD],142.6 [68] minutes; median, 129 min-utes) than for patients receiving throm-bolysis (mean [SD], 68.3 [ 54.4] min-utes; median, 52 minutes). Both groupshad a similar clinical presentation, al-though patients undergoing primaryPTCA were less likely to have evi-dence of congestive heart failure on ex-amination. Fewer patients undergo-ing primary PTCA had ST elevation(61.6% vs 68.0%, P = .001) and Q waves(23.5% vs 30.2%, P = .001). Patients un-dergoing primary PTCA were morelikely to have left bundle-branch block,although the difference between the 2groups was not clinically significant.

Morbidity and MortalityTABLE 3 displays the 30-day and 1-yearmortality rates for both groups. Com-pared with patients receiving throm-bolysis, patients undergoing primaryPTCA had lower 30-day (8.7% vs11.9%, P = .001) and 1-year mortalityrates (14.4% vs 17.6%, P = .001).FIGURE 1 shows 30-day and 1-year mor-tality rates, stratified by sex. After ad-justing for age, primary PTCA was as-sociated with lower mortality rates forboth sexes. Kaplan-Meier survivalcurves for both reperfusion strategiesare shown in FIGURE 2. Mortality re-mained significantly lower for patientsundergoing primary PTCA for the fol-low-up period of 18 months (P,.001).

Patients undergoing primary PTCAhad a lower prevalence of cerebral hem-orrhage and postinfarction angina anda higher prevalence of minor and ma-jor bleeding events during hospitaliza-

Figure 1. Age-Adjusted 30-Day and 1-Year Mortality Rates Stratified by Sex and ReperfusionMethod

25

5

20

10

0

15

Men

8.6

30-Day Mortality 1-Year MortalityWomenMenWomen

Mor

talit

y, %

Primary PTCA Thrombolysis

10.5P = .02

16.9

21.0

16.5

13.815.4

10.4

P<.001P = .01

P = .02

After adjusting for age, both men and women experienced a lower mortality rate with percutaneous translu-minal coronary angioplasty (PTCA) vs early thrombolysis. The reduced mortality with a strategy of primaryPTCA extended to 1 year after the acute myocardial infarction.

Table 3. Crude and Cox-Adjusted 30-Day and 1-Year Mortality*

ThrombolysisGroup

(n = 18 645)

PrimaryPTCA Group

(n = 2038)P

Value

30-Day mortalityCrude, No. (%) 2227 (11.9) 178 (8.7)

.001Adjusted rate (95% CI) 1.00 0.74 (0.63-0.88)

1-Year mortalityCrude, No. (%) 3278 (17.6) 293 (14.4)

.001Adjusted rate (95% CI) 1.00 0.83 (0.73-0.94)

*CI indicates confidence interval; PTCA, percutaneous transluminal coronary angioplasty. Model adjusted for age, race,sex, tobacco use, diabetes, hypertension, dementia, chronic obstructive pulmonary disease, impaired mobility, priorstroke, prior myocardial infarction (MI), history of congestive heart failure, prior coronary artery bypass graft (CABG),prior PTCA, congestive heart failure on arrival, systolic blood pressure on arrival, characteristics of electrocardio-gram on arrival to the hospital (Q wave, atrial fibrillation, left bundle-branch block, right bundle-branch block, ST el-evation), anterior myocardial infarction, serum creatinine, CABG during hospitalization, PTCA services available onsite, annual MI volume of hospital, and number of hospital beds.




tion (TABLE 4). Of the 18 645 patientswho were initially treated with throm-bolytic therapy, 39.2% eventually un-derwent cardiac catheterization and12.4% underwent PTCA. A greater pro-portion of patients in the primary PTCAgroup had coronary artery bypass sur-gery during hospitalization (10.4% vs6.3%, P = .001).

Cox Proportional Hazards ModelsAfter adjustment for baseline character-istics (Table 2), the modality of reper-fusion remained an independent pre-dictor of mortality. Compared with earlythrombolysis, primary PTCA reduced30-day mortality (HR, 0.74; 95% CI,0.63-0.88) (FIGURE 3) and 1-year mor-tality rates (HR, 0.88; 95% CI, 0.73-0.94). Stratifying the groups by age, sex,hypertension, diabetes, prior heart fail-ure, and the location of myocardial in-jury did not change the observed differ-ence in mortality. The benefit of primaryPTCA remained significant in each of thesubgroups except diabetic patients, pa-tients with prior congestive heart fail-ure, and individuals without anteriorlocation of injury.

Subgroup AnalysisOf the 80 356 patients eligible for reper-fusion therapy, 28 955 presented to thehospital within 6 hours of symptom on-set and with ST elevation or left bundle-branch block on the initial electrocar-diogram. Among these “ideal” patients,12 941 (44.7%) received primary PTCAor thrombolysis within 6 hours of hos-pital arrival. A strategy of primary PTCAwas associated with a lower 30-daymortality rate than early thromboly-sis, although the difference was not sta-tistically significant (10.1% vs 12.0%,P = .06). In a Cox proportional hazardmodel, primary PTCA was associatedwith a lower point estimate for 30-daymortality (HR, 0.84; 95% CI, 0.68-1.03). The difference in survival be-tween primary PTCA and thromboly-sis diminished at 1 year (16.2% vs 17.8%;P = .18; HR, 0.92; 95% CI, 0.78-1.08).

Because the volume of patientstreated for AMI at an institution hasbeen associated with outcome, we per-

formed a stratified analysis based onwhether a patient was treated at a low-volume hospital (,150 AMI cases peryear) or high-volume hospital ($150AMI cases per year) (FIGURE 4). In thisanalysis, primary PTCA continued tobe associated with lower short- andlong-term mortality rates, regardless ofthe volume of patients with AMI. Inaddition, patients treated at high-volume institutions experienced lower30-day and 1-year mortality.

In a stratified analysis comparingreperfusion therapies at hospitals withand without on-site angiography, pa-tients who received thrombolysis at fa-cilities that did not offer PTCA expe-rienced the highest mortality rate;patients who received thrombolysis atfacilities that did offer PTCA had bet-ter survival rates (FIGURE 5). Patientstreated with primary PTCA had lower

30-day and 1-year mortality rates thaneither group treated with thromboly-sis. Similar trends were noted in the“ideal” subgroup of patients, althoughthe differences in mortality rates werenot statistically significantly different(FIGURE 6).

Finally, we compared mortality ratesfor patients receiving early reperfu-sion therapy with those who receivednone. Patients not receiving therapy hadhigher 30-day (17.2% vs 11.8%,P = .001) and 1-year mortality rates(33.1% vs 17.6%, P = .001).

COMMENTReperfusion therapy (primary PTCA orearly thrombolysis) has been estab-lished as a pivotal therapeutic strategyin the management of patients withAMI.16 Although randomized clinicaltrials have attempted to identify the op-

Figure 2. Kaplan-Meier Survival Curves for Patients Treated With Primary PercutaneousTransluminal Coronary Angioplasty (PTCA) vs Thrombolysis

100

60

40

80

Time Since Hospital Admission, mo

Thrombolysis

Primary PTCA

Sur

viva

l, %

0 6 12 18

P<.001

Survival estimates are illustrated as points and 95% confidence intervals are shown with vertical crossbars.Each vertical crossbar represents an event. The curves were statistically different by both the log-rank (P,.001)and Wilcoxon rank sum (P,.001) tests.

Table 4. In-Hospital Events for Patients Undergoing Thrombolysis or Primary PercutaneousTransluminal Coronary Angioplasty (PTCA)*

ThrombolysisGroup

(n = 18 645)

PrimaryPTCA Group

(n = 2038)P

Value

Post-MI angina 5513 (29.6) 489 (24.0) .001

Reinfarction 992 (5.3) 81 (4.0) .009

Congestive heart failure 5218 (28.0) 548 (26.9) .29

Cerebral hemorrhage 271 (1.4) 4 (0.2) .001

Stroke 554 (3.0) 43 (2.1) .03

Hemorrhage 4003 (21.5) 583 (28.6) .001

Cardiac catheterization 7304 (39.2) 2038 (100.0) .001

Coronary angioplasty 2311 (12.4) 2038 (100.0) .001

Coronary artery bypass surgery 1173 (6.3) 211 (10.4) .001

*Data are presented as number (percentage). MI indicates myocardial infarction.




timal reperfusion strategy, no single trialhas had statistical power to detect a dif-ference in mortality between these 2therapies, particularly in elderly pa-tients. A recent meta-analysis pooleddata on 1290 patients from 10 random-ized trials that compared primary PTCAwith thrombolysis and demonstrated a34% risk reduction in short-term mor-tality with an interventional strategy of

PTCA.2 The homogeneity of the popu-lation (resulting from exclusion of el-derly patients in several trials) may pre-vent the generalization of these resultsto older patients. The Primary Angio-plasty in Myocardial Infarction (PAMI)trial detected a trend toward a reduc-tion in short-term mortality (5.7% vs15.0%, P = .07) among patients olderthan 65 years who were treated with pri-

mary PTCA.17 The Global Use of Strat-egies to Open Occluded Coronary Ar-teries-IIB study (GUSTO-IIB), thelargest trial to date, also identified atrend toward lower 30-day mortalitywith primary PTCA compared withthrombolysis in patients aged 70 yearsand older.18 No randomized clinical trialto date has shown a statistically signifi-cant benefit of primary PTCA overthrombolysis in the elderly.

Although data from randomizedclinical trials suggest a potential ben-efit of primary PTCA in the elderly, itmay not be appropriate to generalizethese results to the community at large.A comparison of primary PTCA withalteplase showed only a slightly lowermortality rate in patients older than 75years who were treated with primaryPTCA (14.4% vs 16.5%, P = .13).4 In-stitutions in which PTCA is per-formed less often have had worse mor-tality rates in randomized clinical trials18

and in observational studies.12,19,20

In this study, we evaluated reperfu-sion strategies in elderly patients whopresented to hospitals with AMI as cap-tured in the CCP database. The sampleof patients receiving reperfusion therapyin the CCP is larger than the com-bined populations of patients older than65 years from the previous random-ized clinical and community trials. Fur-thermore, in contrast to prior studiesthat evaluated 30-day survival rates, theMedicare enrollment database per-mits analysis of long-term survival rates.

Patients who underwent primaryPTCA, in comparison to early throm-bolysis, had lower 30-day and 1-yearmortality rates after adjustment forbaseline characteristics. The lower mor-tality rates associated with primaryPTCA were observed in each of our sub-group analyses, in both men andwomen, and independent of hospitals’volume of AMI cases, a key indicatorof hospital-level quality of care.21 There-fore, the survival advantage associatedwith primary PTCA was not solely dueto its performance in hospitals with ex-perience in the management of AMI. Wefurther explored whether our study find-ings were biased by consideration of con-

Figure 3. Adjusted Hazard Ratio for 30-Day Mortality for Subgroups of Patients TreatedWith Either Primary Percutaneous Transluminal Coronary Angioplasty (PTCA) or EarlyThrombolysis

MenWomen

HypertensionNo Hypertension

DiabeticNondiabetic

Prior CHFNo prior CHF

Anterior MINonanterior MI

Overall

Age <75 yAge ≥75 y

Hazard Ratio for 30-Day Mortality

Favors PTCA Favors Early Thrombolysis

1.500.50 1.251.000.750.25

Analyses were adjusted for demographics, risk factors, noncardiac and cardiac history, clinical presentation,electrocardiographic features, and hospital characteristics. The benefit of primary PTCA on 30-day mortalityrates varied by subgroup. A benefit of PTCA was evident in patients aged 65 to 75 years and patients olderthan 75 years, and the benefit was significant in women, nondiabetics, patients without prior congestive heartfailure (CHF), and patients with an anterior myocardial infarction (MI).

Figure 4. 30-Day and 1-Year Mortality Rates Stratified by Reperfusion Method and HospitalVolume of Acute Myocardial Infarctions

25

5

20

10

0

15

Low AMI–Volume

11.6

30-Day Mortality 1-Year MortalityHigh AMI–VolumeLow AMI–VolumeHigh AMI–Volume

Mor

talit

y, %

Primary PTCA Thrombolysis

12.4P = .67 14.0

16.618.5

16.0

11.5

8.1

P = .001

P = .24P = .008

Patients undergoing primary percutaneous transluminal coronary angioplasty (PTCA), in comparison with earlythrombolysis, had a reduced mortality in both high– and low AMI–volume facilities. Hospitals that treated 150or more cases of AMI per year were designated high AMI–volume, whereas institutions that treated fewerthan 150 cases of AMI per year were designated low AMI–volume.




traindications to treatment in our pri-mary analysis by assessing the outcomesfor patients who were considered idealcandidates for treatment. Results in thissubsample were in the same direction asour main analysis, although the benefitof primary PTCA strategy was less andno longer significant.

Our findings support the subgroupanalyses from the PAMI and GUSTO-IIB randomized clinical trials and theSecond National Registry of Myocar-dial Infarction (NRMI-2) observa-tional study. These analyses demon-strated a trend toward decreasedmortality in elderly patients who un-derwent primary PTCA.4,17,18 Further-more, the data suggest that elderly pa-tients (a group with an inherentlyincreased risk of mortality) may achievea greater benefit with coronary inter-vention compared with the generalpopulation.

The observed lower mortality rate inpatients undergoing primary PTCAcould be a reflection of the higher com-plication rate observed in patients re-ceiving thrombolytic therapy. Patientsreceiving thrombolysis, in addition tohaving a higher prevalence of cerebralhemorrhage, were more likely to de-velop postinfarction angina and conges-tive heart failure. Bleeding from pulmo-

nary and gastrointestinal sources mayhave also contributed to the higher mor-tality rate among patients receivingthrombolysis.

The observed lower mortality rate inthe PTCA group also could be due toearlier presentation in patients treatedwith primary PTCA, but our data in-dicate that a greater proportion of thepatients who received thrombolysis pre-sented to the hospital within 6 hoursof symptom onset. Furthermore, themean time to treatment after hospitalarrival was 74.3 minutes shorter in thethrombolysis group. The more rapid de-livery of thrombolysis compared withprimary PTCA is consistent with pub-lished trials.2

Reperfusion therapy was used rela-tively infrequently among elderly pa-tients who experienced ST elevation orleft bundle-branch block at the time ofAMI. Reperfusion therapy (primaryPTCA or early thrombolysis) has beenestablished as a fundamental guide-line in the management of patients withAMI.22 Less than half of the elderly pa-tients ideal for reperfusion therapy re-ceived primary PTCA or thrombolysiswithin 6 hours of hospital arrival. Whilethis likely reflects the high prevalenceof contraindications in the elderly popu-lation not ideal for reperfusion therapy,

it may also signal the underutilizationof this important therapy.23 Com-pared with patients who received earlyreperfusion therapy, patients not re-ceiving therapy had higher 30-day and1-year mortality rates. While an argu-ment can be made for primary PTCAover thrombolysis in the elderly, moreattention needs to be focused on theearly recognition of AMI in this popu-lation and the rapid delivery of eitherreperfusion therapy.

For patients with AMI who presentwith ST segment elevation or bundle-branch block, the choice of reperfu-sion strategies continues to be contro-versial because most US hospitals do nothave the facilities or staff to perform pri-mary PTCA in an expert fashion. Ourfindings argue that there may be a mod-est mortality benefit from the use of pri-mary PTCA in elderly patients withAMI. However, because the mortalitybenefit at 1 year was small and furtherdiminished in the “ideal” subset of pa-tients, we do not believe that our find-ings support a policy of triage of the el-derly to primary PTCA. Rather, thesefindings are most consistent with thecurrent American College of Cardiol-ogy/American Heart Association guide-line recommendation that suggests that“primary PTCA should be used as an

Figure 5. Thirty-Day and 1-Year Mortality Rates Stratifiedby Reperfusion Method and Hospital’s Availability of PercutaneousTransluminal Coronary Angioplasty (PTCA) for Entire Cohort

25

5

20

10

0

15

30-Day

8.7

1-Year

Mor

talit

y, %

PTCA Thrombolysis, PTCA Available Thrombolysis, PTCA Not Available

17.916.5

14.412.2

∗ ∗

∗∗

11.3

Mortality

In the unrestricted cohort, patients undergoing primary PTCA had a lower 30-daymortality rate than patients undergoing early thrombolysis, independent of whetherthe facility had resources to perform coronary angioplasty. A similar pattern wasobserved at 1-year follow-up. Statistical comparisons are referenced against pa-tients who underwent PTCA. Asterisk indicates P,.05.

Figure 6. Thirty-Day and 1-Year Mortality Rates Stratified byReperfusion Method and Hospital’s Availability of PercutaneousTransluminal Coronary Angioplasty (PTCA) for Ideal Subgroup

25

5

20

10

0

15

30-DayMortality

10.1

1-Year

Mor

talit

y, %

18.116.416.2

12.310.9

PTCA Thrombolysis, PTCA Available Thrombolysis, PTCA Not Available

∗

In the ideal subgroup (ie, patients with either ST elevation or left bundle-branch blockon arrival and who presented to the hospital within 6 hours of symptom onset), pa-tients undergoing primary PTCA had a significantly lower 30-day mortality rate thanpatients undergoing early thrombolysis in hospitals without resources to perform PTCA.A similar trend was observed at 1-year follow-up, although the difference was notstatistically significant (P = .12). Statistical comparisons are referenced against pa-tients who underwent PTCA. Asterisk indicates P = .04.




alternative to thrombolytic therapy onlyif performed in a timely fashion by in-dividuals skilled in the procedure andsupported by experienced personnel inhigh-volume centers.”22 Because theoverall rate of reperfusion in appropri-ate patients was low, efforts would bestbe concentrated on increasing the pro-portion of eligible patients treated witheither form of reperfusion. Choice ofreperfusion therapy should be based onthe expertise of an individual hospi-tal. In elderly patients with contrain-dications to thrombolytic therapy, astrategy of primary PTCA should beconsidered if available on site or, forcenters without on-site PTCA, the useof transfer protocols may result in agreater proportion of eligible elderly pa-tients receiving reperfusion therapy.

There are several limitations to ouranalysis. First, this was an observa-tional study based on a retrospectivechart analysis and was, therefore, sub-ject to missing data. Patients were not

assigned treatments at random, and un-measured selection factors could haveinfluenced our findings. However, thelarge number of variables available inthe CCP allowed adjustment for manypatient characteristics in multivariateanalysis and by stratification for vol-ume of AMI cases at each hospital. Also,our data were collected in a periodwhen stents were infrequently used andwhen glycoprotein IIB/IIIA inhibitorswere unavailable. While the use ofstents has reduced the rates of reste-nosis and revascularization, there is noevidence that the utilization of stentshas reduced mortality, particularly inthe setting of an AMI.23 The advent ofglycoprotein IIB/IIIA inhibitors mayhave further improved outcomes in pa-tients undergoing primary PTCA. Fi-nally, the observed benefit of primaryPTCA cannot be generalized to insti-tutions without adequate institutionalor operator volume to ensure optimalprimary PTCA results.

In summary, the use of primaryPTCA in elderly patients is associatedwith modestly lower short- and long-term mortality rates compared withthrombolysis. The observed benefit ofprimary PTCA was diminished in thesubgroup of patients who were classi-fied as ideal for reperfusion therapy.

Funding/Support: This work was supported in partby Georgetown University, Washington, DC, and bythe Delmarva Foundation for Medical Care Inc, Eas-ton, Md. The analyses upon which this publication isbased were performed under contract numbers 500-96-P623 and 500-96-P624, entitled “Utilization andQuality Control Peer Review Organization for the Stateof Maryland and the District of Columbia,” spon-sored by the Delmarva Foundation for Medical CareInc, and the Health Care Financing Administration(HCFA), US Department of Health and Human Ser-vices.Disclaimer: The content of this publication does notnecessarily reflect the views or policies of the Depart-ment of Health and Human Services, nor does men-tion of trade names, commercial products, or orga-nizations imply endorsement by the US government.The Health Care Quality Improvement Program(HCQIP), initiated by HCFA, encourages identifica-tion of quality improvement projects derived from theanalysis of patterns of care. The CCP represents aproject within the HCQIP. The authors assume full re-sponsibility for the accuracy and completeness of theideas expressed in this article.

REFERENCES

1. Fibrinolytic Therapy Trialists’ (FTT) CollaborativeGroup. Indications for fibrinolytic therapy in sus-pected acute myocardial infarction: collaborative re-view of early mortality and major morbidity results fromall randomised trials of more that 1000 patients. Lan-cet. 1994;343:311-322.2. Weaver WD, Simes RJ, Betriu A, et al. Compari-son of primary coronary angioplasty and intravenousthrombolytic therapy for acute myocardial infarc-tion: a quantitative review. JAMA. 1997;278:2093-2098.3. Every NR, Parsons LS, Hlatky M, Martin JS, WeaverWD. A comparison of thrombolytic therapy with pri-mary coronary angioplasty for acute myocardial in-farction. N Engl J Med. 1996;335:1253-1260.4. Tiefenbrunn AJ, Chandra NC, French WJ, et al. Clini-cal experience with primary percutaneous translumi-nal coronary angioplasty compared with alteplase (re-combinant tissue-type plasminogen activator) in patientswith acute myocardial infarction: a report from the Sec-ond National Registry of Myocardial Infarction (NRMI-2). J Am Coll Cardiol. 1998;31:1240-1245.5. Vogel RA. HCFA’s Cooperative CardiovascularProject: a nationwide quality assessment of acute myo-cardial infarction. Clin Cardiol. 1994;17:354-356.6. Jencks SF, Wilensky GR. The health care quality im-provement initiative. JAMA. 1992;268:900-903.7. Hochman JS, Gersh BJ. Acute myocardial infarc-tion: complications. In: Topol EJ, ed. Textbook of Car-diovascular Medicine. Philadelphia, Pa: Lippincott-Raven; 1998:437-480.8. Marciniak TA, Ellerbeck EF, Radford MJ, et al. Im-proving the quality of care for Medicare patients with

acute myocardial infarction. JAMA. 1998;279:1351-1357.9. American Hospital Association. The Annual Sur-vey of Hospitals. Chicago, Ill: American Hospital As-sociation; 1994.10. Fleming C, Fisher ES, Chang CH, Bubolz TA,Malenka DJ. Study outcomes and hospital utilizationin the elderly: the advantages of a merged data basefor Medicare and Veterans Affairs hospitals. Med Care.1992;30:377-391.11. Kahn HA, Sempos CT. Statistical Methods in Epi-demiology. New York, NY: Oxford University Press;1989.12. Kaplan EL, Meier P. Nonparametric estimation fromincomplete observations. J Am Stat Assoc. 1958;53:457-481.13. Cox DR. Regression models and life tables. J RStat Soc B. 1972;34:187-192.14. Normand ST, Glickman ME, Sharma RG, McNeilBJ. Using admission characteristics to predict short-term mortality from myocardial infarction in elderlypatients: results from the Cooperative Cardiovascu-lar Project. JAMA. 1996; 275:1322-1328.15. SAS System for Windows [computer program].Version 6.12. Cary, NC: SAS Institute; 1996.16. Ribideiro EE, Silva LA, Carneiro R, et al. Random-ized trial of direct coronary angioplasty versus intra-venous streptokinase in acute myocardial infarction.J Am Coll Cardiol. 1993;22:376-380.17. Stone GW, Grines CL, Vlietstra R, et al. Primary an-gioplasty is the preferred therapy for women and theelderly with acute myocardial infarction: results of thePrimary Angioplasty in Myocardial Infarction (PAMI)

trial [abstract]. J Am Coll Cardiol. 1993;21:330A.18. The Global Use of Strategies to Open OccludedCoronary Arteries in Acute Coronary Syndromes(GUSTO IIb) Angioplasty Substudy Investigators. Aclinical trial comparing primary coronary angioplastywith tissue plasminogen activator for acute myocar-dial infarction. N Engl J Med. 1997;336:1621-1628.19. Ellis SG, Weintraub W, Holmes D, Shaw R, BlockPC, King SB III. Relation of operator volume and ex-perience to procedural outcome of percutaneous coro-nary revascularization at hospitals with high interven-tional volumes. Circulation. 1997;95:2479-2484.20. Jollis JG, Peterson ED, Stafford JA, DeLong ER,Muhlbaier LH, Mark DB. Relationship beween phy-sician and hospital angioplasty volume and outcomein elderly patients. Circulation. 1997;95:2485-2491.21. Thiemann DR, Oetgen WJ, Powe NR. Acute MIpatients admitted directly to cardiac centers have lowerlong-term mortality: a HCFA Cooperative Cardiovas-cular Project study [abstract]. Circulation. 1997;96(suppl 1):562.22. Ryan TJ, Anderson JL, Antman EM, et al. ACC/AHA guidelines for the management of patients withacute myocardial infarction: a report of the AmericanCollege of Cardiology/American Heart Association TaskForce on Practice Guidelines (Committee on Manage-ment of Acute Myocardial Infarction). J Am Coll Car-diol. 1996;28:1328-1348.23. Krumholz HM, Murillo J, Chen J, Radford MJ, Eller-beck EF, Wang Y. Thrombolytic therapy for the treat-ment of eligible elderly Medicare beneficiaries withacute myocardial infarction. JAMA. 1997;277:1683-1688.




Documents

Primary Coronary Angioplasty vs Thrombolysis for the Management