NEJM CARDIOVASCULAR DISEASE ARTICLES - Part 9 ppsx

methods We conducted a multicenter trial in the United Kingdom involving 427 patients with ST-segment elevation myocardial infarction in whom reperfusion failed to occur less than 50 per

n engl j med 353;18 www.nejm.org november 3, 2005

d y s p n e a a n d c a r d i a c p r o g n o s i s

1897

patient or identified by the referring physician),

and differences in the variables used for

propensi-ty analysis, such as our inclusion of covariates

based on myocardial-perfusion SPECT that reflect

the percentage of ischemic and scarred

myocardi-um In addition, experience has demonstrated that

differences in “pretest referral biases” (i.e.,

differ-ences in the clinical characteristics of referral

pop-ulations)24,25 can markedly influence the perceived

prognostic accuracy of clinical variables among

studies Accordingly, there is a need to assess the

extent to which the prognostic significance of

dys-pnea is influenced by pretest referral bias across

various patient populations

Our study has a number of limitations

Ventric-ular function was not assessed in all the patients,

since gated myocardial-perfusion SPECT, required

for its assessment, was not routinely performed

until 1995 We used only a single dichotomous

ques-tion concerning dyspnea, which did not grade the

severity or precipitants of the symptom By

com-parison, the American Thoracic Society uses a

five-point scale for dyspnea.26 Paradoxically, this

limi-tation underscores the strength of our data, since

dichotomously evaluated test variables generally

convey less inherent information than variables that

are classified in more strata.27 Since we only coded

dyspnea among patients without chest pain, we

could not evaluate the potential interaction between

dyspnea and symptoms of chest pain We also did

not evaluate the reproducibility of the self-reported

symptoms Historical or testing information

regard-ing lung disease would have been useful In addition,

since our study patients represent a referral

popu-lation for myocardial-perfusion SPECT, caution

should be exercised in extrapolating our findings

to the general population

The most important limitation of our study is

that, because dyspnea is closely associated with a

variety of both cardiovascular and

noncardiovascu-lar disorders, it may not have been possible to

ac-count for all of the important resulting

interac-tions However, given the fact that the association

between dyspnea and the outcome persisted after

extensive assessment of the effect of other factors,

it is an important observation that dyspnea as a

pre-senting symptom in patients undergoing

nonin-vasive testing is associated with an increased risk

of death from any cause and from cardiac causes,

perhaps for other reasons in addition to those

com-monly recognized

In our population, asymptomatic patients out dyspnea had a rate of adverse events that wassimilar to the rate among those with chest pain Sim-ilar findings have been noted by Christopher Jones

with-et al.18 These observations may be due in part tothe tendency to designate as “asymptomatic” pa-tients with known or suspected cardiac disease whohave symptoms other than chest pain that have beennoted to be associated with an increased incidence

of adverse events, such as a sense of exhaustion,28difficulty in relaxing,29 depressive symptoms,30 andsleeplessness.31 For instance, in a follow-up of 5053male college alumni, those responding “frequent-ly” to the question “how often do you experience ex-haustion (except after exercise)” had twice the rate

of death from cardiac causes as did other dents over a 12-year follow-up.28 Given our findingsregarding dyspnea, these other somatic symptomsmay also deserve further study relative to their prog-nostic significance in cardiac populations

respon-Our results indicate that dyspnea is an tant symptom among patients with suspected andknown coronary artery disease and imply that whendyspnea is present, the likelihood of death fromcardiac causes and from any cause is increased

impor-On the basis of our results, it may be appropriate

to include an evaluation of dyspnea in the clinicalassessment of patients referred for cardiac stresstesting It may also be appropriate to include anevaluation of dyspnea in future efforts to refine al-gorithms (such as the Duke Treadmill Score) thatare used to assess the prognosis of coronary arterydisease

Presented in part at the annual American Heart Association entific Sessions, New Orleans, November 7–10, 2004.

Sci-Dr Rozanski reports having received lecture fees from Myers Squibb and Pfizer Dr Hachamovitch reports having served as

Bristol-a consultBristol-ant to King PhBristol-armBristol-aceuticBristol-als, Bristol-Myers Squibb MedicBristol-al Imaging, and Fujisawa Healthcare and having received lecture fees from Bristol-Myers Squibb Medical Imaging and Fujisawa Healthcare.

Dr Germano reports having received lecture fees from Bristol-Myers Squibb Dr Berman reports having received grant support from Bris- tol-Myers Squibb Medical Imaging and Medtronic and lecture fees from Fujisawa Healthcare and Bristol-Myers Squibb Medical Imag- ing The software used to measure ejection fractions and volumes

from its licensing A minority portion of those royalties is shared by Drs Berman and Germano Dr Abidov was a Save-A-Heart Founda- tion Research Fellow in Cardiac Imaging at the Cedars–Sinai Medi- cal Center during the data collection and analysis.

We are indebted to the nurse practitioners, nuclear technicians, members of the Artificial Intelligence in Medicine group, research coordinators, and follow-up team in the Cardiac Imaging Depart- ment, Cedars–Sinai Medical Center; to Dr Xingping Kang for tech- nical assistance in the preparation and submission of the manu- script; and to Mrs Heidi Gransar for statistical assistance.

n engl j med 353;18 www.nejm.org november 3 , 20051898

d y s p n e a a n d c a r d i a c p r o g n o s i s

r e f e r e n c e s

and exercise test scores to assess all-cause mortality in unselected patients presenting for exercise testing with symptoms of sus- pected coronary artery disease J Am Coll Cardiol 2003;42:842-50.

JS, Pollock BH, Swan HJ Computer-assisted diagnosis in the noninvasive evaluation of patients with suspected coronary artery dis- ease J Am Coll Cardiol 1983;1:444-55.

an-ginal symptoms to lung mechanics during myocardial ischemia Circulation 1972;46:

863-9.

between dyspnea and chest pain ischemic heart disease Acta Med Scand Suppl 1981;

644:16-8.

Pennert K Risk factors for angina pectoris in

a population study of Swedish men J Chronic Dis 1987;40:265-75.

Multivariate analysis of risk factors for onary heart disease Circulation 1973;48:

cor-950-8.

lung function and the risk of heart attack.

Eur Heart J 1988;9:1215-22.

probability as an aid in the clinical sis of coronary-artery disease N Engl J Med 1979;300:1350-8.

Separate acquisition rest thallium-201/stress technetium-99m sestamibi dual-isotope my- ocardial perfusion single-photon emission computed tomography: a clinical validation study J Am Coll Cardiol 1993;22:1455-64.

et al Prognostic impact of hemodynamic sponse to adenosine in patients older than age 55 years undergoing vasodilator stress myocardial perfusion study Circulation 2003;

re-107:2894-9.

Hacha-movitch R, Cohen I, Berman DS The mental prognostic value of percentage of heart rate reserve achieved over myocardial

incre-perfusion single-photon emission

comput-ed tomography in the prcomput-ediction of cardiac death and all-cause mortality: superiority over 85% of maximal age-predicted heart rate J Am Coll Cardiol 2004;44:423-30.

Automatic quantification of ejection fraction from gated myocardial perfusion SPECT.

J Nucl Med 1995;36:2138-47.

JD, Cohen I, Berman DS Comparison of the short-term survival benefit associated with revascularization compared with medi- cal therapy in patients with no prior coro- nary artery disease undergoing stress myo- cardial perfusion single photon emission computed tomography Circulation 2003;

107:2900-7.

of experiments New York: Wiley, 1984.

life-tables J R Stat Soc [B] 1972;34:187-202.

estimated propensity scores: relating theory

to practice Biometrics 1996;52:249-64.

Blackstone EH, Lauer MS Aspirin use and all-cause mortality among patients being evaluated for known or suspected coronary artery disease: a propensity analysis JAMA 2001;286:1187-94.

Black-stone EH, Lauer MS Prognostic tance of presenting symptoms in patients undergoing exercise testing for evaluation

impor-of known or suspected coronary disease Am

J Med 2004;117:380-9.

Prev-alence, clinical features and prognosis of diastolic heart failure: an epidemiologic per- spective J Am Coll Cardiol 1995;26:1565-74.

diastolic dysfunction as a cause of tive heart failure: mechanisms and manage- ment Ann Intern Med 1992;117:502-10.

G, Pillar G, Diamond GA Peripheral

arteri-al responses to treadmill exercise among healthy subjects and atherosclerotic patients.

Circulation 2001;103:2084-9.

et al Usefulness of finger blood flow during exercise as a marker of functionally signifi- cant coronary heart disease Am J Cardiol 2002;90:756-9.

and depressive illness Brain Behav Immun 2002;16:513-24.

Forrester JS, Morris D, Swan HJ The ing specificity of exercise radionuclide ven- triculography N Engl J Med 1983;309:518- 22.

Berman DS, Morris D, Swan HJ Alternative referent standards for cardiac normality: implications for diagnostic testing Ann In- tern Med 1984;101:164-71.

and management: a consensus statement: American Thoracic Society Am J Respir Crit Care Med 1999;159:321-40.

et al Application of information theory to clinical diagnostic testing: the electrocar- diographic stress test Circulation 1981;63: 915-21.

Paffen-barger RS, Lee IM Sense of exhaustion and coronary heart disease among college alum-

ni Am J Cardiol 1999;84:1401-5.

social inequalities as associated with the risk

of ischaemic heart disease a result of chosocial working conditions? Atheroscle- rosis 1993;101:165-75.

KW, Saab PG, Kubzansky L The ogy, pathophysiology, and management of psychosocial risk factors in cardiac practice: the emerging field of behavioral cardiology.

epidemiol-J Am Coll Cardiol 2005;45:637-51.

Akerstedt T, Orth-Gomer K Poor sleep creases the prospective risk for recurrent events in middle-aged women with coro- nary disease: the Stockholm Female Coro- nary Risk Study J Psychosom Res 2003;54: 121-7.

in-Copyright © 2005 Massachusetts Medical Society.

for the REACT Trial Investigators*

From the Department of Cardiology,

Uni-versity Hospitals of Leicester, Leicester

(A.H.G., A.S.-L., S.H.); the Departments

of Health Sciences (K.R.A.) and

Cardio-vascular Sciences (S.E.S.), University of

Leicester, Leicester; the Department of

Cardiology, Royal Infirmary Edinburgh,

Edinburgh (N.G.U.); Sussex Cardiac

Cen-tre, Royal Sussex County Hospital,

Brigh-ton (A de Belder); the Department of

Cardiology, North Staffordshire

Hospi-tal, Stoke-on-Trent (J.D.); the

Depart-ment of Cardiology, Heartlands Hospital,

Birmingham (M.P.); the Department of

Cardiology, John Radcliffe Hospital,

Ox-ford (A Banning); the Department of

Cardiology, Bristol Royal Infirmary,

Bris-tol (A Baumbach); the Department of

Cardiology, Walsgrave Hospital,

Coven-try (M.F.S.); the Department of

Cardiolo-gy, Papworth Hospital, Cambridge (P.S.);

Wessex Cardiac Unit, Southampton

Gen-eral Hospital, Southampton (K.D.D.);

Trent Cardiac Centre, Nottingham City

Hospital, Nottingham (R.A.H.); the

De-partment of Cardiology, Western

Infir-mary, Glasgow (K.G.O.); and the

Depart-ment of Cardiovascular Medicine,

Queens Medical Centre, Nottingham

(R.W.) — all in the United Kingdom

Ad-dress reprint requests to Dr Gershlick at

the University Hospitals of Leicester,

Groby St., Leicester LE3 9QP, United

Kingdom, or at agershlick@aol.com.

*The participants in the Rescue

Angio-plasty versus Conservative Treatment or

Repeat Thrombolysis (REACT) trial are

listed in the Appendix.

methods

We conducted a multicenter trial in the United Kingdom involving 427 patients with ST-segment elevation myocardial infarction in whom reperfusion failed to occur (less than 50 percent ST-segment resolution) within 90 minutes after thrombolytic treatment The patients were randomly assigned to repeated thrombolysis (142 pa-tients), conservative treatment (141 patients), or rescue PCI (144 patients) The pri-mary end point was a composite of death, reinfarction, stroke, or severe heart failure within six months

results

The rate of event-free survival among patients treated with rescue PCI was 84.6 percent, as compared with 70.1 percent among those receiving conservative therapy and 68.7 percent among those undergoing repeated thrombolysis (overall P = 0.004) The adjusted hazard ratio for the occurrence of the primary end point for repeated thrombolysis versus conservative therapy was 1.09 (95 percent confidence interval, 0.71 to 1.67; P = 0.69), as compared with adjusted hazard ratios of 0.43 (95 percent confidence interval, 0.26 to 0.72; P = 0.001) for rescue PCI versus repeated throm-bolysis and 0.47 (95 percent confidence interval, 0.28 to 0.79; P = 0.004) for rescue PCI versus conservative therapy There were no significant differences in mortality from all causes Nonfatal bleeding, mostly at the sheath-insertion site, was more com-mon with rescue PCI At six months, 86.2 percent of the rescue-PCI group were free from revascularization, as compared with 77.6 percent of the conservative-therapy group and 74.4 percent of the repeated-thrombolysis group (overall P = 0.05)

conclusions

Event-free survival after failed thrombolytic therapy was significantly higher with rescue PCI than with repeated thrombolysis or conservative treatment Rescue PCI should be considered for patients in whom reperfusion fails to occur after throm-bolytic therapy

n engl j med 353;26 www.nejm.org december 29, 2005 2759

Patients who have an open infarct

-related artery after acute myocardial

infarc-tion with ST-segment elevainfarc-tion have better

clinical outcomes than patients without an open

artery.1-4

Although primary percutaneous coronary

intervention (primary PCI) is a proven therapeutic

approach in this setting5,6

and is used increasingly, intravenous thrombolysis remains the first-line

therapy in 30 to 70 percent of cases worldwide.7,8

However, thrombolysis results in a grade 3 flow,

according to the Thrombolysis in Myocardial

In-farction (TIMI) classification system, in only 60

percent of patients, even with current

fibrin-spe-cific agents.9

To date, it has been unclear how best

to treat the remaining patients, in whom

throm-bolysis has failed Some physicians, particularly

those at hospitals without interventional facilities,

treat such patients conservatively.10

Others believe that a second dose of a thrombolytic agent may

be beneficial.11

Many advocate emergency PCI

(rescue PCI) on the basis of small trials that have

suggested a benefit of this intervention.12,13

The Rescue Angioplasty versus Conservative Treatment

or Repeat Thrombolysis (REACT) trial was

under-taken to establish which of these three options

achieves the best clinical outcome among patients

in whom thrombolysis has failed

m e t h o d s

We conducted a multicenter, randomized,

paral-lel-group trial that was approved by United

King-dom national and local ethics committees and

fulfilled the conditions of the Declaration of

Hel-sinki The trial was funded by the British Heart

Foundation; Roche Pharmaceuticals provided

re-teplase for repeated thrombolysis (its use was

op-tional for physician investigators) The sponsors

had no role in study design, data collection, or study

analysis or in the writing of this report

patients

Between December 1999 and March 2004, trial

candidates were evaluated at 35 centers (which

joined the study on a rolling basis over three years);

19 of the centers had on-site angiographic facilities

Adults 21 to 85 years of age were eligible for

in-clusion if they had received any licensed

throm-bolytic agent for myocardial infarction with

ST-segment elevation within 6 hours of the onset of

chest pain and if reperfusion had then failed to

occur, as judged by the predetermined 90-minute

electrocardiographic criterion (less than 50 cent resolution in the lead with previous maximal ST-segment elevation) The inclusion and exclusion criteria are listed in Table 1 A screening log of potential subjects was kept through November

per-2002 to catalogue patients who did or did not ticipate in the trial; however, this log was not maintained after November 2002 because of fund-ing constraints The trial subjects were enrolled after giving written informed consent

par-randomization

Patients were randomly assigned by a 24-hour computer-generated random-allocation system to undergo repeated thrombolysis, conservative treat-ment, or rescue PCI Patients assigned to repeated thrombolysis received a fibrin-specific thrombo-lytic agent (alteplase or reteplase, according to the physician’s choice) and intravenous heparin, ac-cording to standard clinical practice Low-molec-ular-weight heparin was not used in the first 24 hours Patients assigned to the conservative-ther-apy group received standard medical therapy for myocardial infarction without thrombolysis or PCI

To ensure a standardized group, conservative apy included intravenous heparin for 24 hours, irrespective of the first thrombolytic agent Hepa-rin administration in the repeated-thrombolysis and conservative-therapy groups was titrated to an activated partial-thromboplastin time ratio of 1.5

ther-to 2.5 Patients assigned ther-to rescue PCI went coronary angiography, proceeding to an-gioplasty if required (i.e., if the patient had less than TIMI grade 3 flow and more than 50 percent stenosis in the infarct-related artery) Adjunctive strategies (e.g., stenting or glycoprotein IIb/IIIa receptor inhibition) were used at the discretion

under-of the interventionist Crossover among the three treatment groups was discouraged but was al-lowed if a patient had ongoing or further chest pain associated with ST-segment re-elevation or new elevation in at least two contiguous leads or had cardiogenic shock

data collection

Clinical examination, electrocardiography, tologic measurements, and biochemical tests (in-cluding measurement of cardiac biomarkers) were performed on all patients 4 hours after the initia-tion of the randomly assigned therapy (to account for the potential time delay to rescue PCI), at 12 and 24 hours after randomization, and at dis-

hema-n ehema-ngl j med 353;26 www.nejm.org december 29, 2005

2760

Table 1 Criteria for Inclusion and Exclusion and Definitions of Trial End Points.

Inclusion criteria

Acute myocardial infarction with ST-segment elevation of more than 0.1 mV in at least two contiguous leads, excluding V1

Aspirin and thrombolysis administered within 6 hours of onset of symptoms

Age 21 to 85 years

Ability to give informed consent

At 90 minutes (±15 minutes) after the beginning of initial thrombolytic therapy, electrocardiogram shows failed thrombolytic therapy — i.e., less than 50% resolution of the ST segment in the lead showing the greatest ST-segment elevation measured from the baseline (isoelec- tric line) to 80 msec beyond the J point, with or without chest pain

Rescue angioplasty, if assigned, can be performed within 12 hours of the onset of pain

Exclusion criteria

Probable inability to gain femoral access for intervention (e.g., severe peripheral vascular disease)

Left bundle-branch block

Life expectancy less than 6 months owing to noncardiac cause

Previous inclusion in this trial at any time, or in any other clinical trial during the previous month

Contraindication to thrombolysis (e.g., cardiopulmonary resuscitation after first thrombolytic treatment)

Hemoglobin greater than 1.5 g/dl below normal range within previous 6 hours

Platelet count below normal range within previous 6 hours

For patients 75 years of age or older: systolic blood pressure above 200 mm Hg, diastolic blood pressure above 100 mm Hg, or both at any time during the current episode of pain, even if successfully reduced by therapy

For patients less than 75 years of age: after prescription of first thrombolytic therapy, systolic blood pressure above 200 mm Hg, diastolic blood pressure above 100 mm Hg, or both on more than one occasion

Estimated body weight less than 65 kg

Cardiogenic shock, either in the opinion of the investigator or defined as persistent (lasting more than 30 minutes) systolic hypotension (less than 90 mm Hg) with oliguria and autonomic activation, with or without pulmonary edema despite appropriate volume replacement, and considered to be due to ventricular dysfunction rather than to any other cause

Administration of low-molecular-weight heparin within the previous 12 hours

Definitions of trial end points

Reinfarction

During index admission: further chest pain lasting more than 30 minutes and accompanied by new electrocardiographic changes (new

Q waves above 0.04 second or ST-segment elevation above 0.1 mV in two leads for more than 30 minutes), further enzyme rise,

Severe heart failure

Early heart failure: any new-onset cardiogenic shock or heart failure with pulmonary edema that is resistant to medical therapy and that occurs during the index admission and after randomization

Late heart failure: admission to hospital for treatment of heart failure (New York Heart Association class III or IV)

n engl j med 353;26 www.nejm.org december 29, 2005 2761

charge, with clinical follow-up at 1, 6, and 12

months The components of the primary end

point were continuously documented More than

90 percent of study data were subjected to source

validation according to strictly controlled

cri-teria

end points

The primary end point was a composite of major

adverse cardiac and cerebrovascular events at six

months, including death, recurrent myocardial

in-farction, cerebrovascular event, and severe heart

failure The secondary end points included the

components of the primary end point, as well as

bleeding and revascularization Events were

adju-dicated by an independent end-point committee,

whose members were blinded to treatment

assign-ment Quality-of-life and resource-use data were

collected at follow-up Definitions of all end points

are given in Table 1

power and sample size

On the basis of the limited evidence available at

the time of study design (1998),12 the steering

committee estimated that the rate of the primary

composite end point in the conservative-therapy

group would approach 20 percent and

hypothe-sized a 40 percent relative reduction in this rate

in the rescue-PCI group; thus, it was calculated

that 1200 patients would be required (80 percent

power, α = 0.05) In December 2001, the members

of the steering committee and the data and safety

monitoring committee (who did not have access

to the trial data) examined new published

evi-dence suggesting that the rate of death or

recur-rent myocardial infarction would be 29 percent

with conservative therapy, 26.5 percent with

re-peated thrombolysis, and 15 percent with rescue

PCI.11,13-15 Because the rates of heart failure and

cerebrovascular events were inconsistently

report-ed in those studies, the power of our study was

re-calculated on the basis of assumed rates of death

and recurrent myocardial infarction alone It was

determined that a sample size of 156 patients in

each group would provide 80 percent power

(α = 0.05) to detect the same 40 percent relative

re-duction in the composite end point that was

pre-viously hypothesized It was assumed that heart

failure and cerebrovascular events would be likely

to increase rather than reduce such power in the

final analysis

During 2003 and 2004, enrollment in the trial

began to decline The precise reason for this cline is uncertain, because the screening log was not maintained after November 2002 (as noted above) However, other ongoing clinical trials, as well as the introduction of the new thrombolytic agent tenecteplase (and the concomitant unli-censed use of low-molecular-weight heparin), lim-ited the number of suitable candidates for partici-pation Because of declining trial recruitment and

de-a finite funding period, the steering committee terminated enrollment in the trial in March 2004

statistical analysis

All analyses were performed on an treat basis Process times are reported as medi-ans with interquartile ranges and compared with use of the Kruskal–Wallis test The proportions of subjects in each of the groups who reached any end point during the six months were compared with use of either the chi-square test or Fisher’s exact test, as appropriate Survival and event-free survival were plotted as Kaplan–Meier curves, and the log-rank test was used to compare them Haz-ard ratios with 95 percent confidence intervals were calculated for all pairwise comparisons Cox proportional-hazards regression models were used

intention-to-to investigate the potential influence of all line covariates on treatment effects Covariates were selected for a final model by a forward vari-able-selection procedure The assumption of pro-portional hazards was assessed both graphically, with the use of log–log survivor plots, and by adding associated time-dependent covariates to the model.16 There was no evidence that the as-sumption of proportional hazards was violated in any of the results presented here No formal ad-justment for multiple testing was undertaken, but the P values were interpreted cautiously All statis-tical analyses were performed with SAS software, version 8.2 (SAS Institute)

base-r e s u l t s

At the termination of the trial, 435 patients had been enrolled and randomly assigned to one of the three treatment groups Of these, six withdrew consent (one each in the groups assigned to re-peated thrombolysis and rescue PCI and four in the group assigned to conservative therapy), and another two were excluded (one each in the re-peated-thrombolysis and rescue-PCI groups) be-cause they had inappropriately undergone random-

n engl j med 353;26 www.nejm.org december 29, 2005

2762

ization before giving consent, which they declined

to do The data for 427 patients are therefore sented Of these, 142 were assigned to repeated thrombolysis, 141 to conservative therapy, and 144

pre-to rescue PCI (Table 2)

The trial screening log, which was maintained until November 2002, included 713 patients who did not undergo randomization (as compared with

304 patients who had undergone randomization

by that date) Of those who did not undergo domization, most were excluded on the basis of clinical criteria, including delayed presentation (beyond six hours) (24 percent), advanced age (21.4

ran-percent), and severe hypertension (13.6 percent) Only 4.2 percent were excluded on the basis of the judgment of the patient’s physician

baseline characteristics

The baseline characteristics were similar in all groups (Table 2) There was no difference among the groups in the median time from the onset of pain to the first (nontrial) thrombolytic treatment (P = 0.73) The median time from presentation until the first thrombolytic treatment (“door-to-needle time”) was 27 minutes (interquartile range,

16 to 43)

Table 2 Baseline Characteristics of Enrolled Patients.

All Patients (N=427)

Repeated Thrombolysis (N=142)

Conservative Therapy (N=141)

Rescue PCI (N=144) Age — yr

Medical history — no (%)

Percutaneous coronary vention

First thrombolytic therapy —

no (%)

Time to first thrombolytic therapy (min)

* Data were missing for one patient.

n engl j med 353;26 www.nejm.org december 29, 2005 2763

actual treatment received

Eighteen patients (4.2 percent) did not receive their

randomly assigned treatment Among the patients

who were assigned to rescue PCI, 14 received

con-servative therapy and 2 received repeated

throm-bolysis; among the patients who were assigned

to repeated thrombolysis, 1 received conservative

therapy and 1 received rescue PCI The results of

the analysis according to the intention-to-treat

principle were unchanged when the data were

analyzed according to actual treatment received

rescue pci

Of the 144 patients assigned to rescue PCI, 88

(61.1 percent) were recruited from hospitals with

interventional capabilities The median transfer

time for patients from hospitals without

interven-tional capabilities was 85 minutes (interquartile

range, 55 to 120) Sixteen patients in this group crossed from their assigned therapy, and 128 pro-ceeded to angiography, 13 of whom did not re-quire angioplasty because of patent vessels Of the remaining 115 patients, only 9 were deemed

to have had an unsuccessful rescue-PCI dure; in 6 of these patients the artery was deemed not amenable to PCI, in one instance affecting

proce-1 patient there was a technical failure of x-ray equipment, and in 2 patients the attempts to open the artery were unsuccessful

Rescue PCI was commenced (i.e., the wire crossed the lesion) a median of 414 minutes af-ter the onset of pain (interquartile range, 350 to 505) Stents were deployed in 68.5 percent of pa-tients, and a glycoprotein IIb/IIIa receptor inhibitor (abciximab) was administered in 43.4 percent For patients assigned to rescue PCI rather than re-

Table 3 End-Point Events Occurring within Six Months of Treatment.*

Repeated Thrombolysis (N = 142)

Conservative Therapy (N = 141)

Rescue PCI (N = 144)

Primary end-point events (predetermined

hierarchical analysis)

Death from any cause — no (% of patients) 18 (12.7) 18 (12.8) 9 (6.2) 0.12

Death from cardiac causes — no (% of

Severe heart failure — no (% of patients) 10 (7.0) 11 (7.8) 7 (4.9) 0.58

Composite primary end point —

* PCI denotes percutaneous coronary intervention, and CABG coronary-artery bypass grafting The proportions of

sub-jects in each of the groups who reached any end point during the six months were compared by either the chi-square

test or Fisher’s exact test, as appropriate.

† P = 0.05 by the log-rank test.

n engl j med 353;26 www.nejm.org december 29, 2005

2764

peated thrombolysis, the median additional delay

in the time to the assigned treatment was 84 utes (4.6 hours for rescue PCI vs 3.2 hours for repeated thrombolysis)

min-primary end point

All components of the primary end point were corded for 406 subjects (95.1 percent) Mortality status was confirmed for the remaining 21 sub-jects (4.9 percent): 6 each in the repeated-throm-bolysis and conservative-therapy groups and 9 in the rescue PCI-group Data on these subjects were included in the analyses as censored observations, with a median study period of 105 days (range,

re-5 to 177)

In the rescue-PCI group, 15.3 percent of the patients reached at least one component of the primary end point, as compared with 31.0 percent

in the repeated-thrombolysis group and 29.8 cent in the conservative-therapy group (overall

per-P = 0.003) (Table 3) The rate of event-free survival (Fig 1) was 84.6 percent in the rescue-PCI group,

as compared with 70.1 percent in the therapy group and 68.7 percent in the repeated-thrombolysis group (overall P = 0.004) Among patients assigned to rescue PCI, there was no significant difference in event rates between those who were transferred for intervention (16.4 per-cent) and those who were recruited in hospitals with on-site facilities for intervention (14.6 per-

conservative-cent, P = 0.80), and logistic-regression analysis cated that the time to repeated PCI (up to 12 hours) had no significant effect on outcome Although the numbers are very small, the incidence of the primary end point was much higher among those who underwent unsuccessful rescue PCI (5 of

indi-9 patients [55.6 percent]) than among those who underwent successful rescue PCI (12 of 106 pa-tients [11.3 percent], P = 0.007)

Age and infarct site were the only baseline acteristics that were identified as predictors of the primary end point by multivariate analysis Ad-justed pairwise hazard ratios (Fig 2) confirmed a statistically significant benefit of rescue PCI as compared with conservative therapy (hazard ra-tio, 0.47; 95 percent confidence interval, 0.28 to 0.79; P = 0.004) and repeated thrombolysis (haz-ard ratio, 0.43; 95 percent confidence interval, 0.26

char-to 0.72; P = 0.001) There was no significant ence in benefit between repeated thrombolysis and conservative therapy (hazard ratio, 1.09; 95 percent confidence interval, 0.71 to 1.67; P = 0.69)

differ-components of the primary end point

There was a trend toward lower mortality at six months in the rescue-PCI group (6.2 percent) than

in either the repeated-thrombolysis group (12.7 percent) or the conservative-therapy group (12.8 percent, P = 0.12 for both comparisons) (Table 3) When the rescue-PCI group was compared with the two other groups combined, this difference was statistically significant (hazard ratio, 0.48;

95 percent confidence interval, 0.23 to 0.99; P<0.05) Multivariate analysis identified age and diabetes

as significant predictors of death, and the adjusted hazard ratios significantly favored rescue PCI: the hazard ratio for rescue PCI as compared with re-peated thrombolysis was 0.42 (95 percent confi-dence interval, 0.19 to 0.94; P<0.04), and the hazard ratio for rescue PCI as compared with conserva-tive therapy was 0.42 (95 percent confidence in-terval, 0.19 to 0.94; P<0.04) The trial was not pow-ered to detect a difference in mortality alone.There were no significant differences in the rates of cerebrovascular events or severe heart fail-ure among the three treatment groups (Table 3) However, the rate of recurrent myocardial infarc-tion was significantly lower in the rescue-PCI group (2.1 percent) than in the repeated-throm-bolysis group (10.6 percent) or the conservative-therapy group (8.5 percent); the hazard ratio for rescue PCI as compared with repeated thromboly-

Days after Randomization

No of Event-free Patients

95

95 116

96

96 117

99

97 118

99

98 120

101

99 122

105

102 124

106

104 127

110

109 129 P=0.004

Figure 1 Kaplan–Meier Estimates of the Cumulative Rate of the Composite

Primary End Point (Death, Recurrent Myocardial Infarction, Severe Heart

Failure, or Cerebrovascular Event) within Six Months.

PCI denotes percutaneous coronary intervention, and CI confidence interval.

n engl j med 353;26 www.nejm.org december 29, 2005 2765

sis was 0.23 (95 percent confidence interval, 0.09

to 0.62; P = 0.004), and the hazard ratio for rescue

PCI as compared with conservative therapy was

0.33 (95 percent confidence interval, 0.12 to 0.93;

P = 0.04)

bleeding complications

Bleeding events were defined according to a

mod-ified TIMI classification (Table 1).17 There were no

significant differences among the groups in

ma-jor bleeding events (Table 3) However, there was

a tendency toward higher mortality from major

bleeding episodes in the repeated-thrombolysis

group (four deaths from hemopericardium and

one death from intracranial hemorrhage) and the

conservative-therapy group (one death from

he-mothorax and two deaths from intracranial

hem-orrhage) than in the rescue-PCI group, in which

there were no deaths associated with bleeding

events Minor bleeding episodes were significantly

more frequent in the rescue-PCI group (P<0.001);

minor bleeding occurred at the access site in 28

patients, 5 of whom required blood transfusion

Among the patients in the rescue-PCI group who

had bleeding events, 69 percent had received

ab-ciximab, as compared with 43 percent of all

pa-tients in this group (P = 0.17) There were no

sig-nificant differences among the groups in the

incidence of bleeding episodes characterized by

a fall in hemoglobin without an identified

bleed-ing site

revascularization

Revascularization rates tended to be lower in the

rescue-PCI group (Table 3) At six months, 86.2

percent of the patients in the rescue-PCI group

were free from revascularization, as compared

with 77.6 percent of those undergoing

conserva-tive therapy and 74.4 percent of those undergoing

repeated thrombolysis (overall P = 0.05 by the

log-rank test) The unadjusted hazard ratio for

revas-cularization was 0.50 (95 percent confidence

in-terval, 0.29 to 0.88; P<0.02) for rescue PCI as

compared with repeated thrombolysis and 0.58

(95 percent confidence interval, 0.33 to 1.04;

P<0.07) for rescue PCI as compared with

conser-vative therapy There was no difference between

the two groups not assigned to rescue PCI

(haz-ard ratio for repeated thrombolysis as compared

with conservative therapy, 1.17; 95 percent

confi-dence interval, 0.71 to 1.92; P = 0.56)

d i s c u s s i o n

Our study compared three therapeutic options ter failed thrombolytic therapy We found that res-cue PCI was superior to either conservative care

af-or repeated thrombolysis, even though a tial proportion of patients treated with rescue PCI were transferred from hospitals without interven-tional facilities, and there was a median addition-

substan-al time delay of 84 minutes until treatment with rescue PCI in comparison with repeated throm-bolysis A trend toward a higher frequency of fatal bleeding was noted in both the conservative-treat-ment group and the repeated-thrombolysis group, but given the small number of cases reported, no firm conclusions can be drawn from these data

The higher rates of nonfatal bleeding in the cue-PCI group may be due to the use of glycopro-tein IIb/IIIa receptor inhibitors

Previous evidence supporting the use of cue PCI is limited, and current guidelines rec-ommend it only for certain high-risk subgroups

res-of patients.18,19 Rescue PCI has been reported to lower the rate of recurrent myocardial infarction, reduce the incidence of early severe heart failure, and improve one-year survival.12,15 However, the sample sizes in most studies have been small;

moreover, failed rescue PCI has been associated with a high incidence of adverse outcomes (ap-proximately 30 percent),14,20 a result that could

HR, 1.09 95% CI, 0.71–1.67

HR, 0.43 95% CI, 0.26–0.72

HR, 0.47 95% CI, 0.28–0.79

Rescue PCI (n=144)

vs conservative treatment (n=141)

Figure 2 Adjusted Hazard Ratios for the Occurrence of the Composite Primary End Point (Death, Recurrent Myocardial Infarction, Severe Heart Failure, or Cerebrovascular Accident) among the Trial Groups.

HR denotes hazard ratio, CI confidence interval, and PCI percutaneous onary intervention

cor-n ecor-ngl j med 353;26 www.nejm.org december 29, 2005

2766

reduce the overall benefit of the technique.21,22The recent use of stents and glycoprotein IIb/IIIa receptor inhibitors may have improved outcomes

in comparison with those in studies performed in the mid-1990s

The findings of our study favoring the use of rescue PCI contradict those of the recent Middles-brough Early Vascularization to Limit Infarction (MERLIN) trial,23 which found a significant re-duction in revascularization rates only There are

a number of important differences between the two trials The MERLIN trial was a locally con-fined study, whereas ours was a national multi-center trial In the MERLIN trial, the first throm-bolytic agent was more often streptokinase (96 percent, vs 59 percent in our trial), and eligibility was determined on the basis of electrocardiogra-phy at 60 minutes, rather than 90 This strategy may have reduced the rates of the end points in the conservative-treatment group, since some pa-tients treated with streptokinase probably under-went perfusion at 60 to 90 minutes (as suggested

by the fact that 40 percent of the patients in the rescue-PCI group had TIMI grade 3 flow ac-cording to angiography before intervention) The MERLIN trial also showed lower rates of stent-ing and of glycoprotein IIb/IIIa inhibitor use, which may have contributed to a higher reinfarc-tion rate in the rescue-PCI group For reasons that remain unexplained, the mortality in the rescue-PCI group was unusually high in the MERLIN trial,20 as was the rate of cerebrovascular events

in this group (4.6 percent) In addition, despite the absence of a group randomly assigned to re-peated thrombolytic treatment, 11.7 percent of the conservatively treated patients in the MERLIN trial underwent repeated thrombolysis, further con-founding the results

The optimal approach for detecting the ure of thrombolytic therapy has been the subject

fail-of much debate.24,25 Historically, entry into studies

of rescue PCI has been determined by

angiograph-ic findings,13,26 whereas in clinical practice, failure

of reperfusion is generally detected by clinical, noninvasive markers There is evidence that the ratios of biochemical markers (including creatine kinase MB fraction, troponin, and myoglobin mass) measured before and 60 minutes after the administration of thrombolytic therapy have good predictive value,27,28 with low ratios correlating with poor patency However, differential degrees

of ST-segment resolution also correlate well with TIMI flow grade29-31 and predict longer-term out-come.32 The value of ongoing pain as a sensitive marker of nonreperfusion is questionable, given its low specificity33 and the routine use of analge-sia Although certain markers (e.g., myoglobin) may be considered the most sensitive for detecting failed thrombolytic therapy, these were not widely available in the clinical setting when our trial was designed Therefore, an ST-segment resolution of

50 percent was considered the most reliable sible entry criterion, and this cutoff was deemed likely to pick up most reperfusion failures, with a low rate of false positives for patent arteries.30,32,34Although the trial was terminated early, ter-mination occurred before the investigators were unblinded to the data and was necessary, given the falling recruitment rates and the finite fund-ing period for the study In the absence of a full registry, we cannot exclude some element of selec-tion bias in the population enrolled However, all consecutive patients at each site in whom throm-bolytic therapy had failed were evaluated, and the baseline characteristics as recorded in the screen-ing log until November 2002 do not suggest such bias The great majority of patients, according to this record, were excluded for predefined clinical reasons, with only 4.2 percent being excluded by choice of the patient’s physician

pos-In conclusion, the trial found that rescue PCI after failed thrombolytic treatment was associated with a statistically significant reduction in the in-cidence of major adverse cardiac and cerebrovas-cular events, as compared with either repeated thrombolysis or conservative management These results indicate that rescue PCI, with transfer to

a tertiary site if required, should be considered for patients in whom thrombolysis for myocardial in-farction with ST-segment elevation fails to achieve reperfusion

Dr Gershlick reports having served as a consultant to and ceived lecture fees from Cordis, Boston Scientific, and Medtronic and being currently in receipt of research grant support from Medtronic; Dr Baumbach, having served as a consultant to Boston Scientific; Dr Schofield, having served as a consultant to Cordis and Guidant; and Dr Dawkins, having served as a consul- tant to Eli Lilly, Boston Scientific, Guidant, and Conor Medsystems, having received lecture fees from Eli Lilly, Boston Scientific, and Guidant, and having appeared as an expert wit- ness for Boston Scientific No other potential conflict of interest relevant to this article was reported.

re-We are indebted to Leslie Shortt for data entry, and to the nursing and medical staff of the cardiac care units and catheter laboratories at all sites.

n engl j med 353;26 www.nejm.org december 29, 2005 2767

appendix

The members of the REACT trial were as follows: Steering Committee — A.H Gershlick (principal investigator), M de Belder, H

Swanton, R Wilcox, K Abrams, D de Bono (deceased), and A Stephens-Lloyd; Data and Safety Committee — K Fox, J Birkhead, and

M Bland; End-Point Committee — J Hampton and S Davies; Trial coordinators — A Stephens-Lloyd and S Hughes; Statisticians — S

Stevens, K Abrams, and A Skene; Economic evaluation — M Buxton and M Dyer; Investigators, interventionists, and study-site

coor-dinators — Glenfield Hospital, Leicester: J.D Skehan, J Kovac, N.J Samani, and P.J.B Hubner; Leicester Royal Infirmary, Leicester: I Squire, L

Shipley, and E Parker; Leicester General Hospital, Leicester: I Hudson, R Pathmanathan, and K Fairbrother; Royal Infirmary of Edinburgh,

Ed-inburgh: N.G Uren, D.E Newby, P Bloomfield, N.A Boon, A.D Flapan, L Flint, M O’Donnell, and L Cameron; Royal Sussex County

Hospital, Brighton: A de Belder, S Holmberg, D Hildick-Smith, N Cooter, and L Bennett; North Staffordshire Hospital, Stoke-on-Trent: J

Davis, J Nolan, J Creamer, D O’Gorman, J Machin, and C Butler; Heartlands Hospital, Birmingham: M Pitt, P Ludman, G Murray, J

Beattie, S Eccleshall, N El-Gaylani, J Pitt, and J Hulse; Hull Royal Infirmary, Hull: M Norell, F Alamgir, J Caplin, G Kaye, A Clark,

M Nasir, J Bristow, A Fussey, E Owen, and A Baksh; John Radcliffe Hospital, Oxford: A Banning, N Alp, H McCullough, N Meldrum,

C Hamer, and R Douthwaite; Bristol Royal Infirmary, Bristol: A Baumbach, G Dalton, and K Carson; Walsgrave Hospital, Coventry: M.F

Shiu, H Singh, M Been, P Glennon, S Constantinides, and L Gill; Hairmyres Hospital, East Kilbride: K Oldroyd, B.D Vallance, J Young,

and G Moreland; Papworth Hospital, Cambridge: P Schofield, M.C Petch, L.M Shapiro, S.C Clarke, H Millington, A Emerton, and C

Rhydwen; Southampton General Hospital, Southampton: K Dawkins, I Simpson, H Gray, A Calver, J Morgan, Z Nicholas, and S Kitt;

Addenbrookes Hospital, Cambridge: P Weissberg, S Blackwood; Nottingham City Hospital, Nottingham: R Henderson, D Falcon-Lang, and M

Marriott; University Hospital of Wales, Cardiff: W Penny and L Davies; Royal Devon and Exeter Hospital, Exeter: B Smith, M Gandhi, J Dean,

C Rinaldi, A Renouf, D Taylor, and J Hunt; Royal Surrey County Hospital, Guildford: E Leatham, R Mitra, S Green, and M Eaton; Wishaw

General Hospital, Wishaw, Larnarkshire: M Malekian and A Sloey; Derby Royal Infirmary, Derby: M Millar-Craig and A Joy; St Thomas’ Hospital,

London: S Redwood, S Patel, and S Hogun; Epsom General Hospital, Epsom: S Odemuyiwa, A Redwood, and C Cooper; Southmead

Hospi-tal, Bristol: P Walker, K Potts, and D Foster-Hargreaves; West Suffolk HospiHospi-tal, Bury-St-Edmunds: E Lee and S Reader; Western Infirmary,

Glasgow: K Oldroyd, S Robb, W Hillis, and J Kelly; Crosshouse Hospital, Kilmarnock: D O’Neill and O El-Wassief; Manchester Heart Centre,

Manchester: N Curzen, F Fath-Ordoubadi, L Neyses, and H Iles-Smith; Royal Alexandra Hospital, Paisley: S Hood, I Findlay, and J

Dou-gall; Hemel Hempstead General Hospital, Hemel Hempstead: D Hackett and L Birkhead; Kent and Sussex Hospital, Tunbridge Wells: C Lawson and

J Highland; Middlesex Hospital, London: H Swanton and E Firmin; St Mary’s Hospital, London: R Foale, J Mayet, S Smart, and J

Var-ghese.

references

The GUSTO Angiographic

Investiga-tors The effects of tissue plasminogen

acti-vator, streptokinase, or both on

coronary-artery patency, ventricular function, and

survival after acute myocardial infarction

N Engl J Med 1993;329:1615-22 [Erratum,

N Engl J Med 1994;330:516.]

Puma JA, Sketch MH Jr, Thompson

TD, et al Support for the open-artery

hy-pothesis in survivors of acute myocardial

infarction: analysis of 11,228 patients

treated with thrombolytic therapy Am J

Cardiol 1999;83:482-7.

Fath-Ordoubadi F, Huehns TY,

Al-Mohammad A, Beatt KJ Significance of

the Thrombolysis in Myocardial

Infarc-tion scoring system in assessing

infarct-related artery reperfusion and mortality

rates after acute myocardial infarction Am

Heart J 1997;134:62-8.

French JK, Hyde TA, Patel H, et al

Survival 12 years after randomization to

streptokinase: the influence of

thromboly-sis in myocardial infarction flow at three

to four weeks J Am Coll Cardiol 1999;

34:62-9.

Budaj A, Brieger D, Steg PG, et al

Global patterns of use of antithrombotic

and antiplatelet therapies in patients with

acute coronary syndromes: insights from

the Global Registry of Acute Coronary

Events (GRACE) Am Heart J

2003;146:999-1006.

Keeley EC, Boura JA, Grines CL

Pri-mary angioplasty versus intravenous

throm-bolytic therapy for acute myocardial

infarc-tion: a quantitative review of 23 randomised

opportuni-The Fibrinolytic opportuni-Therapy Trialists’ (FTT) Collaborative Group Indications for fibri- nolytic therapy in suspected acute myo- cardial infarction: collaborative overview

of early mortality and major morbidity results from all randomised trials of more than 1000 patients Lancet 1994;343:311-

22 [Erratum, Lancet 1994;343:742.]

Cannon CP, Gibson CM, McCabe CH,

et al TNK-tissue plasminogen activator compared with front-loaded alteplase in acute myocardial infarction: results of the TIMI 10B trial Circulation 1998;98:2805- 14.

Prendergast BD, Shandall A, ter MB What do we do when thromboly- sis fails? A United Kingdom survey Int J Cardiol 1997;61:39-42.

Buchal-Sarullo FM, Americo L, Di Pasquale P, Castello A, Mauri F Efficacy of rescue thrombolysis in patients with acute myo- cardial infarction: preliminary findings

Cardiovasc Drugs Ther 2000;14:83-9.

Ellis SG, da Silva ER, Heyndrickx G, et

al Randomized comparison of rescue gioplasty with conservative management

an-of patients with early failure an-of bolysis for acute anterior myocardial in- farction Circulation 1994;90:2280-4.

throm-Ross AM, Lundergan CF, Rohrbeck

SC, et al Rescue angioplasty after failed

out-Sutton AG, Campbell PG, Grech ED,

et al Failure of thrombolysis: experience with a policy of early angiography and rescue angioplasty for electrocardiograph-

ic evidence of failed thrombolysis Heart 2000;84:197-204.

Ellis SG, Da Silva ER, Spaulding CM, Nobuyoshi M, Weiner B, Talley JD Review

of immediate angioplasty after lytic therapy for acute myocardial infarc- tion: insights from the RESCUE I, RESCUE

fibrino-II, and other contemporary clinical riences Am Heart J 2000;139:1046-53.

expe-Parmar MKB, Machin D Survival analysis: a practical approach Chichester, England: John Wiley, 1995.

Bovill EG, Tracy RP, Knatterud GL, et

al Hemorrhagic events during therapy with recombinant tissue plasminogen ac- tivator, heparin, and aspirin for unstable angina (Thrombolysis in Myocardial Is- chemia, phase IIIB trial) Am J Cardiol 1997;79:391-6.

Antman EM, Anbe DT, Armstrong PW,

et al ACC/AHA guidelines for the ment of patients with ST-elevation myo- cardial infarction: a report of the Ameri- can College of Cardiology/American Heart Association Task Force on Practice Guide- lines (Committee to Revise the 1999 Guide- lines for the Management of Patients with Acute Myocardial Infarction) J Am Coll Cardiol 2004;44:E1-E211.

manage-Van de Werf F, Ardissino D, Betriu A,

et al Management of acute myocardial

n engl j med 353;26 www.nejm.org december 29, 2005

2768

infarction in patients presenting with segment elevation: The Task Force on the Management of Acute Myocardial Infarc- tion of the European Society of Cardiolo-

ST-gy Eur Heart J 2003;24:28-66.

Bar FW, Ophnis TJ, Frederiks J, et al

Rescue PTCA following failed sis and primary PTCA: a retrospective study

thromboly-of angiographic and clinical outcome

J Thromb Thrombolysis 1997;4:281-8.

Gibson CM, Cannon CP, Greene RM,

et al Rescue angioplasty in the bolysis in Myocardial Infarction (TIMI) 4 trial Am J Cardiol 1997;80:21-6.

Throm-McKendall GR, Forman S, Sopko G, Braunwald E, Williams DO Value of res- cue percutaneous transluminal coronary angioplasty following unsuccessful throm- bolytic therapy in patients with acute myo- cardial infarction Am J Cardiol 1995;76:

1108-11.

Sutton AG, Campbell PG, Graham R,

et al A randomized trial of rescue plasty versus a conservative approach for failed fibrinolysis in ST-segment elevation myocardial infarction: the Middlesbrough Early Revascularization to Limit INfarction (MERLIN) trial J Am Coll Cardiol 2004;

angio-44:287-96.

de Lemos JA, Morrow DA, Gibson CM,

et al Early noninvasive detection of failed epicardial reperfusion after fibrinolytic therapy Am J Cardiol 2001;88:353-8.

el Deeb F, Ciampricotti R, el Gamal

M, Michels R, Bonnier H, Van Gelder B

Value of immediate angioplasty after travenous streptokinase in acute myocar- dial infarction Am Heart J 1990;119:786- 91.

in-Stewart JT, French JK, Theroux P, et al

Early noninvasive identification of failed reperfusion after intravenous thrombolytic therapy in acute myocardial infarction J Am Coll Cardiol 1998;31:1499-505.

Tanasijevic MJ, Cannon CP, Antman

EM, et al Myoglobin, creatine-kinase-MB and cardiac troponin-I 60-minute ratios predict infarct-related artery patency after thrombolysis for acute myocardial infarc- tion: results from the Thrombolysis in Myo- cardial Infarction study (TIMI) 10B J Am Coll Cardiol 1999;34:739-47.

Fernandez AR, Sequeira RF, Chakko

S, et al ST segment tracking for rapid termination of patency of the infarct-related artery in acute myocardial infarction J Am Coll Cardiol 1995;26:675-83.

de-Zeymer U, Schroder R, Tebbe U, hoek GP, Wegscheider K, Neuhaus KL

Mol-Non-invasive detection of early infarct sel patency by resolution of ST-segment el-

al Eur Heart J 2001;22:769-75.

Gibson CM, Karha J, Giugliano RP, et

al Association of the timing of ST-segment resolution with TIMI myocardial perfusion grade in acute myocardial infarction Am Heart J 2004;147:847-52.

Schroder R, Wegscheider K, Schroder

K, Dissmann R, Meyer-Sabellek W Extent

of early ST segment elevation resolution:

a strong predictor of outcome in patients with acute myocardial infarction and a sen- sitive measure to compare thrombolytic regimens: a substudy of the International Joint Efficacy Comparison of Thrombolyt- ics (INJECT) trial J Am Coll Cardiol 1995; 26:1657-64.

Christenson RH, Ohman EM, Topol

EJ, et al Assessment of coronary sion after thrombolysis with a model com- bining myoglobin, creatine kinase-MB, and clinical variables Circulation 1997;96: 1776-82.

reperfu-Andrews J, Straznicky IT, French JK,

et al ST-segment recovery adds to the sessment of TIMI 2 and 3 flow in predict- ing infarct wall motion after thrombolytic therapy Circulation 2000;101:2138-43.

as-Copyright © 2005 Massachusetts Medical Society.

31.

32.

33.

34.

CLINICAL TRIAL REGISTRATION

The Journal encourages investigators to register their clinical trials

in a public trials registry The members of the International Committee

of Medical Journal Editors plan to consider clinical trials for publication only if they have been registered (see N Engl J Med 2004;351:1250-1)

The National Library of Medicine’s www.clinicaltrials.gov is a free registry, open to all investigators, that meets the committee’s requirements.

n engl j med 357;5 www.nejm.org august 2, 2007

PERSPECTIVE

439

Only 30 or 40 years ago,

rheumatic fever was a

com-mon topic in the Journal A PubMed

search for articles on rheumatic

fever published between 1967 and

1976 returned 55 New England

Jour-nal of Medicine articles — fewer

than for endocarditis (77) but

more than for stroke and

syphi-lis (24 entries each) A similar

PubMed search for the decade

1997 through 2006 yielded just

eight entries for rheumatic fever

This trend holds for all

Medline-indexed journals: an average of

516 articles on rheumatic fever per

year from 1967 through 1976,

but only 172 per year from 1997

through 2006 Most observers

would probably consider this

de-crease to be a reasonable

reflec-tion of the waning incidence of

the disease After all, in the

mid-20th century, children with

rheu-matic fever occupied many of the

beds in pediatric wards in

indus-trialized countries — indeed,

en-tire hospitals were dedicated to

the treatment of, and tion from, rheumatic fever But

rehabilita-in the latter half of the 20th century, rheumatic fever receded

as an important health problem

in almost all wealthy countries

Today, most physicians in these countries are unlikely ever to see

a case of acute rheumatic fever, and their experience with rheu-matic heart disease will be lim-ited to heart-valve lesions in older patients who had rheumatic fever

in their youth

The reality, however, is that the decrease in publications reflects only the waning burden of disease among the less than 20% of the world’s population living in high-income countries For everyone else, rheumatic fever and rheumat-

ic heart disease are bigger lems than ever It was estimated recently that worldwide 15.6 mil-lion people have rheumatic heart disease and that there are 470,000 new cases of rheumatic fever and 233,000 deaths attributable to

prob-rheumatic fever or prob-rheumatic heart disease each year.1 These are con-servative estimates — the actual figures are likely to be substan-tially higher Almost all these cas-

es and deaths occur in developing countries

How did rheumatic fever come rare in wealthy countries? Medical science can take some of the credit, thanks largely to the use of penicillin for primary pre-vention, but most of the reduction

be-is attributable to improved living conditions, which have resulted in less overcrowding and better hy-giene, with consequent reductions

in transmission of group A tococci In other words, rheumatic fever is a disease of poverty That

strep-it is in many ways the epstrep-itome of diseases of poverty and social in-justice is exemplified by the situ-ations in Australia and New Zea-land In these countries, which boast living standards that are among the best in the world, there are indigenous populations,

Rheumatic Heart Disease in Developing Countries

Focus on Research

Rheumatic Heart Disease in Developing Countries

Jonathan R Carapetis, Ph.D., F.R.A.C.P

Related article, page 470

an impressive list of unexpected

associations between genes or

chromosomal regions and a broad

range of diseases There have

been few, if any, similar bursts

of discovery in the history of

med-ical research Relatively

conven-tional statistical techniques are

adequate for the analysis and

in-terpretation of these initial

stud-ies But as we delve further into

the genome in the search for

net-works of interacting gene

vari-ants and interactions between

these networks and tal factors,5 much more sophis-ticated methods of statistical analysis are likely to be required

environmen-Dr Hunter is a professor of epidemiology at the Harvard School of Public Health, Bos-

ton, a statistical consultant to the Journal,

and codirector of the National Cancer stitute’s Cancer Genetic Markers of Sus- ceptibility project Dr Kraft is an assistant professor of epidemiology and biostatis- tics at the Harvard School of Public Health, Boston.

In-This article (10.1056/NEJMp078120) was published at www.nejm.org on July 18, 2007

Christensen K, Murray JC What wide association studies can do for medi- cine N Engl J Med 2007;356:1094-7 Witte JS Multiple prostate cancer risk vari- ants on 8q24 Nat Genet 2007;39:579-80 Wacholder S, Chanock S, Garcia-Closas

genome-M, El Ghormli L, Rothman N Assessing the probability that a positive report is false: an approach for molecular epidemiology stud- ies J Natl Cancer Inst 2004;96:434-42 NCI-NHGRI Working Group on Replica- tion in Association Studies, Chanock S, Mani- olo T, et al Replicating genotype-phenotype associations Nature 2007;447:655-60 Thomas DC, Clayton DG Betting odds and genetic associations J Natl Cancer Inst 2004;96:421-3.

Copyright © 2007 Massachusetts Medical Society

n engl j med 357;5 www.nejm.org august 2, 2007

440

many of whose members live in

poverty, with documented rates

of rheumatic fever and

rheumat-ic heart disease that are among

the highest in the world.1 Among

aboriginal people of northern

Aus-tralia, for example, acute

rheu-matic fever develops in 0.2 to

0.5% of school-age children each

year, and more than 2% of

peo-ple of all ages have rheumatic

heart disease

An unfortunate consequence

of the decline in rheumatic fever

in industrialized countries has

been a parallel reduction in

re-lated research Indeed, although

our understanding of the

patho-genesis of this mysterious

dis-ease has improved somewhat, the

only advances that have

substan-tially altered the management or

prevention of rheumatic fever

dur-ing the past 40 or 50 years have

occurred in the medical and

sur-gical treatment of severe

rheu-matic heart disease — treatment

that is largely palliative and

nei-ther accessible nor affordable to

the majority of affected patients

The mainstays of the control of rheumatic fever remain treatment

of group A streptococcal gitis with penicillin (primary pro-phylaxis) and administration of penicillin G benzathine injections every 3 to 4 weeks for many years

pharyn-in people with a history of matic fever to prevent recurrent episodes (secondary prophylaxis)

rheu-Both strategies are based on ings from seminal studies in the United States published in the 1950s.2,3

find-The available and potential trol measures for rheumatic fe-ver and rheumatic heart disease are summarized in the diagram

con-Of these, only one — secondary prophylaxis — has been proved

to be cost-effective and practical even in the poorest countries For more than 20 years, the World Health Organization has recom-mended secondary prophylaxis, most effectively delivered within

a coordinated program using a registry of patients, as the first

priority for the control of matic heart disease.4 Yet most developing countries still do not have effective secondary-prophy-laxis programs

rheu-How to ensure that secondary prophylaxis is delivered to those who need it is one of several crit-ical questions related to the imple-mentation of current knowledge about the control of rheumatic fe-ver and rheumatic heart disease Other relevant questions include how to identify people with mild rheumatic heart disease so that they may be offered secondary prophylaxis earlier, whether pri-mary prophylaxis can be a practi-cal and cost-effective public health measure in developing countries, and how to ensure that limited health care funds are spent most effectively — which may entail shifting some funding from the provision of expensive cardiac sur-gery for severe rheumatic heart disease to the development of ro-bust secondary-prophylaxis pro-grams

Other key issues revolve around the need to develop new approach-

es to primary prevention, ularly a vaccine that protects against rheumatic fever A num-ber of vaccines are in development, and a safe and effective vaccine may well be available within one

partic-or two decades However, ence with other relatively recent vaccines, including conjugate

experi-pneumococcal and Haemophilus

in-fluenzae type B vaccines, suggests

that there may be many barriers

to the funding, acceptance, and use of new vaccines in the places that need them most Potential alternative strategies, including controlling streptococcal skin in-fections, are intriguing but of un-proven benefit.5

How will these issues be

ad-Rheumatic Heart Disease in Developing Countries

33p9

Rheumatic heart disease Acute rheumatic fever Group A streptococcal infection

Causal Pathway Preventive Measures

Cardiac failure endocarditisStroke, Death

AUTHOR:

FIGURE:

4-C H/T

RETAKE

SIZE

ICM CASE

H/T Combo

Revised

AUTHOR, PLEASE NOTE:

Figure has been redrawn and type has been reset.

Please check carefully.

REG F

Enon

1st 2nd 3rd

Hygiene Primary prevention Sore-throat treatment Vaccine (unavailable) Control of skin infections (unproved)

Secondary prevention Secondary prophylaxis

“Tertiary” prevention Medication for heart failure Valve surgery

Anticoagulation

Potential Preventive Measures for Rheumatic Fever and Rheumatic Heart Disease.

“Sore-throat treatment” refers to primary prophylaxis — that is, diagnosis and

treat-ment of group A streptococcal pharyngitis.

n engl j med 357;5 www.nejm.org august 2, 2007

PERSPECTIVE

441

dressed? Although some basic

re-search on pathogenesis and the

development of early-stage

vac-cine candidates can take place in

laboratories anywhere in the

world, the clinical,

epidemiolog-ic, and public health studies

re-quire access to populations with

high rates of disease In recent

years, many such studies have been

conducted in Australia, New

Zea-land, and the Rocky Mountain

region of the United States But

applied research of relevance to

developing countries should take

place in developing countries —

a proposition that presents many

obvious challenges Even if

bar-riers caused by poor education,

the absence of a skilled

work-force, limited finances, inadequate

technology, and remoteness of the

populations that are at the

high-est risk of disease can be

over-come, the burden of rheumatic

fever and rheumatic heart disease

is often either unappreciated or

dwarfed by epidemics of human immunodeficiency virus, malaria, tuberculosis, and pneumonia

Marijon and colleagues are to

be applauded for the results of the study reported in this issue of

the Journal (pages 470–476) The

study represents a partnership among researchers in Mozam-bique, Cambodia, France, and Australia It tackles an important practical issue: whether and how

to conduct screening for

rheumat-ic heart disease among school-age children in developing countries

Also, it presents a compelling gument for the use of echocar-diographic screening (see image)

ar-The counterargument is that the use of such expensive technology

is neither feasible nor affordable

in the countries with the highest disease burden Yet if clinical di-agnosis had been relied on, ap-proximately 90% of echocardio-graphically detected cases would have been missed It is not ac-

ceptable to leave these cases diagnosed and these children at risk for recurrence of rheumatic fever simply because echocardio-graphic screening is seen as an inappropriate use of modern tech-nology in developing countries Instead, further research is need-

un-ed to define models of diographic screening that are practical, affordable, and widely applicable

echocar-Marijon et al found that 2 to 3% of school-age children in Cambodia and Mozambique have rheumatic heart disease, almost all of it previously undiagnosed

We know that this represents the tip of the iceberg: cases in chil-dren 5 to 14 years of age are like-

ly to represent only 15 to 20% of all cases in the population.1 These data confirm that rheumatic fe-ver and rheumatic heart disease are of sufficient importance to warrant the urgent attention of the international public health and research communities

Dr Carapetis is the director of the Menzies School of Health Research, Charles Darwin University, Casuarina, Northern Territory, Australia.

Carapetis JR, Steer AC, Mulholland EK, Weber M The global burden of group A streptococcal diseases Lancet Infect Dis 2005;5:685-94.

Stollerman GH, Rusoff JH, Hirschfeld I Prophylaxis against group A streptococci in rheumatic fever: the use of single monthly injections of benzathine penicillin G N Engl

Rheumatic fever and rheumatic heart ease: report of a WHO expert consultation World Health Organ Tech Rep Ser 2004;923: 1-122.

dis-McDonald M, Currie BJ, Carapetis JR Acute rheumatic fever: a chink in the chain that links the heart to the throat? Lancet In- fect Dis 2004;4:240-5.

Copyright © 2007 Massachusetts Medical Society.

JOB: ISSUE:

4-C H/T

RETAKE 1st 2nd

SIZE

ICM CASE

H/T Combo

Revised

AUTHOR, PLEASE NOTE:

Figure has been redrawn and type has been reset.

Please check carefully.

Still Image from a Two-Dimensional Echocardiogram in a Patient with Moderate

Mitral Regurgitation Due to Rheumatic Heart Disease.

An apical four-chamber view is shown, with color Doppler imaging illustrating a

regurgitant-flow signal that extends along the lateral wall of the left atrium The

blue-and-yellow mosaic pattern indicates the regurgitant jet (arrow) (Courtesy of Dr

Andrew Steer, University of Melbourne.)

n engl j med 353;7 www.nejm.org august 18, 2005

Paul D Frederick, M.P.H., M.B.A., Jerome L Abramson, Ph.D., Hal V Barron, M.D., Ajay Manhapra, M.D., Susmita Mallik, M.D., and Harlan M Krumholz, M.D., for the National Registry

of Myocardial Infarction Investigators

From the Department of Medicine, sion of Cardiology (V.V., N.K.W., J.L.A.) and Division of General Medicine (S.M.), Em- ory University School of Medicine; and the Department of Epidemiology, Rollins School of Public Health, Emory University (V.V.), Atlanta; the Section of Cardiovascu- lar Medicine, Department of Medicine (S.S.R., H.M.K.), the Division of Health Policy and Administration, Department of Epidemiology and Public Health (H.M.K.), and the Robert Wood Johnson Clinical Scholars Program (H.M.K.) at Yale Uni- versity School of Medicine and Yale–New Haven Hospital Center for Outcomes Re- search and Evaluation — both in New Ha- ven, Conn.; the Ovation Research Group, Seattle (P.D.F.); Genentech, South San Francisco, Calif (H.V.B.); and Hackley Hos- pital, Spring Lake, Mich (A.M.) Address reprint requests to Dr Vaccarino at the De- partment of Medicine, Division of Cardiol- ogy, Emory University School of Medicine,

Divi-1256 Briarcliff Rd., Suite 1N, Atlanta, GA

manage-m e t h o d s

With the use of data from the National Registry of Myocardial Infarction, we examinedsex and racial differences in the treatment of patients who were deemed to be “ideal can-didates” for particular treatments and in deaths among 598,911 patients hospitalizedwith myocardial infarction between 1994 and 2002

r e s u l t s

In the unadjusted analysis, sex and racial differences were observed for rates of fusion therapy (for white men, white women, black men, and black women: 86.5, 83.3,80.4, and 77.8 percent, respectively; P<0.001), use of aspirin (84.4, 78.7, 83.7, and78.4 percent, respectively; P<0.001), use of beta-blockers (66.6, 62.9, 67.8, and 64.5percent; P<0.001), and coronary angiography (69.1, 55.9, 64.0, and 55.0 percent;

reper-P<0.001) After multivariable adjustment, racial and sex differences persisted for rates

of reperfusion therapy (risk ratio for white women, black men, and black women: 0.97,0.91, and 0.89, respectively, as compared with white men) and coronary angiography(relative risk, 0.91, 0.82, and 0.76) but were attenuated for the use of aspirin (risk ratio,0.97, 0.98, and 0.94) and beta-blockers (risk ratio, 0.98, 1.00, and 0.96); all risks wereunchanged over time Adjusted in-hospital mortality was similar among white women(risk ratio, 1.05; 95 percent confidence interval, 1.03 to 1.07) and black men (risk ratio,0.95; 95 percent confidence interval, 0.89 to 1.00), as compared with white men, butwas higher among black women (risk ratio, 1.11; 95 percent confidence interval, 1.06

to 1.16) and was unchanged over time

c o n c l u s i o n s

Rates of reperfusion therapy, coronary angiography, and in-hospital death after cardial infarction, but not the use of aspirin and beta-blockers, vary according to raceand sex, with no evidence that the differences have narrowed in recent years

myo-a b s t r myo-a c t

n engl j med 353;7 www.nejm.org august 18 , 2005

The n e w e n g l a n d j o u r n a l of m e d i c i n e

672

n recent years, attention has beenfocused on variations in the treatment of coro-nary heart disease that are related to the sex andrace of the patient Landmark studies in the late1980s and early 1990s reported differences in treat-ment according to sex and race.1-4 In the past de-cade, other investigations have described a generallyconsistent pattern of less intensive treatment ofacute myocardial infarction in women, as comparedwith men,5-11 and in blacks, as compared withwhites,8,9,12-17 across a variety of settings Efforts

to remedy racial and sex differences in health careuse have received prominent attention, including arecent Institute of Medicine report18 and the PublicHealth Service’s Healthy People 2010 initiative.19Although sex and racial differences in the treat-ment of coronary heart disease have been docu-mented for more than a decade, little is knownabout whether these differences have persisted inmore recent years We assessed temporal trends insex and racial differences in the use of guideline-based management for patients hospitalized withacute myocardial infarction

p a t i e n t s

Since July 1, 1990, hospitals participating in theNational Registry of Myocardial Infarction (NRMI)have enrolled consecutive patients with myocar-dial infarction, as previously described.20 BecauseNRMI-1 (July 1990 through May 1994) collectedlittle information on patients’ characteristics, werestricted our analysis to the 1,724,984 patientsfrom 1917 hospitals who were enrolled in NRMI-2(June 1994 through March 1998), NRMI-3 (April

1998 through June 2000), and NRMI-4 (July 2000through May 2002) We excluded 12,132 patientswith erroneous discharge dates and 381,018 pa-tients who were transferred from another acute carehospital because their early treatments were notdocumented We also excluded 131,474 patientswho survived less than 24 hours because of insuffi-cient time to begin treatments; 40,881 patients ofunknown age, sex, race, or survival status; 60,689patients whose race was not recorded as white orblack; and 55,316 patients with missing data formodel covariables We restricted our analysis to 658hospitals (out of 1917 hospitals) participating inNRMI for the full study period, resulting in a finalsample of 598,911 patients NRMI data collection

has previously been validated by comparison withthe Cooperative Cardiovascular Project.21 This pro-tocol was deemed exempt from review by the insti-tutional review board at Emory University

t r e a t m e n t o f m y o c a r d i a l i n f a r c t i o n

Patients were evaluated for the use of treatmentsrecommended by the American College of Cardi-ology–American Heart Association (ACC–AHA)guidelines for the treatment of myocardial infarc-tion since 1990.22-24 These included acute reperfu-sion therapy for patients with ST-segment elevationwithin 24 hours of admission, the administration

of aspirin and beta-blockers within 24 hours of mission, and coronary angiography during hospi-talization As secondary treatment end points, weexamined the frequency of coronary-artery bypassgraft (CABG) surgery and percutaneous translumi-nal coronary angioplasty (PTCA) (except for primaryPTCA, which was included in our definition of re-perfusion therapy) during hospitalization

ad-To exclude racial or sex variations in treatmentthat may reflect differences in the proportion of pa-tients for whom treatment is considered appropri-ate, we identified subgroups of patients who wereideally suited for each management strategy — inother words, patients with the strongest indica-tions for treatment (ACC–AHA class I) and withoutmajor contraindications, according to guidelinespublished in 1990,22 1996,23 and 1999.24 Whenvariations were present in the three sets of guide-lines, the 1996 guidelines were followed, since theyare similar to the 1999 guidelines and were pub-lished closest to the beginning of our observationperiod

To avoid bias in regard to the availability of vices, rates of coronary angiography were calculatedamong patients admitted to facilities with full ca-pability of performing invasive cardiovascular pro-cedures Rates of use of CABG and PTCA were cal-culated among patients admitted to these facilitieswho were “ideal candidates” for coronary angiog-raphy and who underwent angiography Becauseinformation was lacking on angiographic findings,

ser-we ser-were not able to define further patient eligibilityfor revascularization The only contraindication tothe use of aspirin in the initial management ofmyocardial infarction is true allergy to salicylates,which is uncommon and was not recorded in NRMI.Therefore, no ideal-candidate subgroup was creat-

ed for aspirin

i

m e t h o d s

n engl j med 353;7 www.nejm.org august 18, 2005

s e x a n d r a c i a l d i f f e r e n c e s i n t r e a t m e n t o f m y o c a r d i a l i n f a r c t i o n

673

i n - h o s p i t a l m o r t a l i t y

We examined trends in hospital mortality

accord-ing to sex and race This analysis was restricted to

patients who were not transferred to another acute

care hospital, since the survival status of transferred

patients in the second hospital was unknown

s t a t i s t i c a l a n a l y s i s

We categorized patients into four groups according

to race and sex: white men, white women, black

men, and black women Sex and racial differences

in demographic and clinical factors and in the

char-acteristics of hospitals were assessed over the full

study period and stratified according to year of

treat-ment (with a year defined as the period from June

through May) We calculated crude rates of

treat-ment and in-hospital mortality for the selected

sub-groups of ideal-candidate patients in the four

groups

We used logistic-regression models to derive

the likelihood of treatment and death for the four

groups.25 We tested whether differences in the use

of treatments according to sex and race changed

over time by including a three-way interaction term

reflecting the sex and race of patients and the year

Three consecutive models were constructed for

each end point Model 1 included sex, race, year,

and all two-way and three-way interaction terms

among sex, race, and year; model 2 expanded the

data in model 1 to include other demographic and

clinical factors; and model 3 expanded the data in

model 2 to include characteristics of the hospitals

To assess whether the clustering of patients within

hospitals affected our results, analyses were

repeat-ed with the use of generalizrepeat-ed-estimating-equation

models The results were similar and are not

report-ed All analyses were performed using SAS software

(version 8.2)

c h a r a c t e r i s t i c s o f p a t i e n t s a n d h o s p i t a l s

The mean age of patients did not change

substan-tially over time, but the prevalence of most

coro-nary risk factors increased in all subgroups (Table

1), whereas there was a decline in the proportion of

patients with ST-segment elevation or Q waves on

initial electrocardiography The four subgroups

showed similar time trends in most factors, as

shown by the nonsignificant interaction among sex,

race, and year In all years combined, there were

substantial differences in many factors according

to sex and race For example, women in both racialgroups were older than men, whereas blacks in bothsex groups were younger than whites As comparedwith white men, fewer female and black patientshad ST-segment elevation or Q waves on initial elec-trocardiography, but women and blacks had morerisk factors, a higher Killip class, and a longer delay

to reach the hospital As compared with whites,black patients tended to be hospitalized more often

in facilities that were used for teaching, were ated with medical schools, were located in urbanareas, and had equipment for performing cardio-vascular procedures

affili-i d e a l c a n d affili-i d a t e s f o r t r e a t m e n t s

a n d p r o c e d u r e s

The proportion of patients qualifying as ideal didates for reperfusion and the administration ofbeta-blockers was 50 percent or less and declinedover time in all groups At each time point, womenand blacks were less likely than white men to be ide-

can-al candidates (Fig 1) Approximately 10 percent

of patients were classified as ideal candidates forcoronary angiography This percentage was simi-lar in all sex and racial groups and fairly constantover time

t r e a t m e n t s a n d p r o c e d u r e s

a m o n g i d e a l c a n d i d a t e s

In the unadjusted analysis, treatment rates differedaccording to sex and race, with rates highest inwhite men and lowest in black women (Table 2)

Differences were larger for rates of reperfusion apy and coronary angiography, particularly for blackwomen, but smaller for the use of aspirin and beta-blockers The use of aspirin and beta-blockers in-creased over time, whereas rates of reperfusiontherapy remained stable and those of coronary an-giography decreased slightly, with similar timetrends in the four demographic groups As a result,there was no significant variation over time in treat-ment differences according to sex or race

ther-Results that were adjusted for the tics of patients and hospitals were similar (Table 3)

characteris-Because models 2 and 3 provided almost identicalresults, only the results of model 3 (adjusted forboth patient and hospital characteristics) are pre-sented The interaction among the factors of sex,race, and year, as well as all other pairwise interac-tions, were not significant, indicating that racial and

r e s u l t s

n engl j med 353;7 www.nejm.org august 18 , 2005

n engl j med 353;7 www.nejm.org august 18, 2005

s e x a n d r a c i a l d i f f e r e n c e s i n t r e a t m e n t o f m y o c a r d i a l i n f a r c t i o n

675