Coronary Artery Surgery Study (CASS): randomized trial...

PATHOPHYSIOLOGY AND NATURAL HISTORYCORONARY ARTERY DISEASE

Coronary Artery Surgery Study (CASS): a

randomized trial of coronary artery bypass surgery

Survival data

CASS PRINCIPAL INVESTIGATORS AND THEIR AsSOCIATES*

ABSTRACT CASS includes a multicenter patient registry and a randomized controlled clinical trial.It is designed to assess the effect of coronary artery bypass surgery on mortality and selected nonfatalend points. From August 1975 to May 1979, 780 patients with stable ischemic heart disease were

randomly assigned to receive surgical (n = 390) or nonsurgical (n = 390) treatment and were followedthrough April 15, 1983. At 5 years, the average annual mortality rate in patients assigned to surgicaltreatment was 1.1%. The annual mortality rate in those receiving medical therapy was 1.6%. Annualmortality rates in patients with single-, double-, and triple-vessel disease who were in the surgicalgroup were 0.7%, 1.0%, and 1.5%; the corresponding rates in patients in the medical group were 1.4%,1.2%, and 2.1%. The differences were not statistically significant. Nearly 75% of the patients had entryejection fractions of at least 0.50. The annual mortality rates in patients in the surgical group in thissubgroup with single-, double-, and triple-vessel disease were 0.8%, 0.8%, and 1.2% and correspond-ing rates in the medical group were 1.1%, 0.6%, and 1.2%. The annual rate of bypass surgery in

patients who were initially assigned to receive medical treatment was 4.7%. The excellent survivalrates observed both in CASS patients assigned to receive medical and those assigned to receive surgicaltherapy and the similarity of survival rates in the two groups of patients in this randomized trial lead tothe conclusion that patients similar to those enrolled in this trial can safely defer bypass surgery untilsymptoms worsen to the point that surgical palliation is required.Circulation 68, No. 5, 939-950, 1983.

CORONARY ARTERY DISEASE is a major cause ofmorbidity and mortality in the United States. In spiteof encouraging decreases in age-adjusted mortalityduring the past decade,' in 1980 coronary artery dis-ease was the cause of 566,000 deaths.2 Over 683,000patients were admitted to hospitals in the United Statesthat year with acute myocardial infarction.3 An esti-mated 5.4 million individuals have been diagnosed ashaving coronary artery disease as manifest by chronicangina or healed myocardial infarction.t

Medical and surgical therapies for patients withcoronary artery disease have changed considerablyover the past 20 years. Medical management now in-cludes aggressive modification of risk factors, liberaluse of nitrates, administration of /3-adrenoceptor-

This collaborative clinical trial was supported by the National Heart,Lung and Blood Institute.

Address for reprints: Lloyd Fisher, Ph.D., CASS Coordinating Cen-ter, University of Washington, 1107 N.E. 45th St., Room 530, Seattle,WA 98105.

Received June 2, 1983; revision accepted July 21, 1983.*Contributing investigators and participating clinical centers are list-

ed before references.tThe Health Interview Survey, National Center for Health Statistics:

Unpublished data.

Vol. 68, No. 5, November 1983

blocking drugs, and more recently use of calcium an-tagonists. Current modes of medical therapy for coro-nary artery disease may favorably influence survival inselected patients. For example, results of three recentcontrolled trials support the premise that long-termadministration of fl-blocking drugs reduces mortalityin survivors of acute myocardial infarction.t

Surgical treatment of patients with coronary arterydisease has evolved from ineffective procedures suchas thoracic sympathectomy, epicardial abrasion, andinternal mammary artery ligation to the currently fa-vored procedure, coronary artery bypass grafting withthe use of either reversed saphenous vein or internalmammary artery grafts. Experienced surgical teamsperform this procedure with low risk to the patient.Numerous reports attest to the fact that complete orpartial relief of angina is accomplished in 60% to 90%of patients in the months and years after bypass.' 8

Coronary artery bypass surgery is often the treatmentof choice for patients who have poorly controlled angi-na and coronary arterial anatomy making revasculari-zation possible.9 The marked relief of angina aftercoronary bypass, the low operative risk, the increas-

939

by guest on June 23, 2018http://circ.ahajournals.org/

Dow

nloaded from

http://circ.ahajournals.org/

CASS INVESTIGATORS

ingly widespread availability of experienced surgicalteams, and the belief of many physicians that the pro-cedure may reduce long-term mortality have resultedin a rapid increase in the number of bypass proceduresperformed in the United States, from an estimated24,000 in 1971 to an estimated 159,000 in 1981.*

In 1973 The National Heart, Lung and Blood Insti-tute organized a patient registry and a randomized trialdesigned to compare results of medical and surgicaltherapy in patients with coronary artery disease. Thegoal of the randomized trial is to rigorously evaluatethe effect of assigned treatment on total mortality inwell-defined subsets of patients with coronary arterydisease. In this first end point report from the random-ized trial, survival in patients assigned to surgical ther-apy is compared with survival in patients assigned toreceive medical management.

MethodsDetails of CASS design, study protocol, procedure for ran-

domization, quality-control measures, and the baseline charac-teristics of the participating patients have been previously pub-lished. '0 CASS consists of a patient registry and a randomizedtrial. All patients undergoing coronary arteriography for sus-pected or proven coronary artery disease at 15 clinical sites wereasked to participate in the registry and 94% agreed. Patientsentering the randomized trial were drawn from the registry at the11 sites that participated in the trial; four CASS sites participat-ed in the registry only.

All registry patients at the 11 randomizing sites were screenedfor potential participation in the trial. Patients qualifying onclinical and angiographic grounds were randomly assigned toreceive medical or surgical therapy after informed consent hadbeen obtained. Clinical criteria for inclusion included age of 65years or less, angina that was Canadian Cardiovascular Societyclass I or II" with or without a history of myocardial infarction,or well-documented myocardial infarction more than 3 weeksbefore randomization. Clinical criteria for exclusion were priorcoronary bypass surgery, unstable or progressive angina, angi-na more severe than class II (angina occurring after climbingone flight of stairs or walking two blocks is class III), congestiveheart failure (New York Heart Association class III or IV), acoexisting illness that would increase the likelihood of deathwithin 5 years, and a variety of practical exclusions that mightlimit active participation during follow-up such as geographicinaccessibility, psychological problems, or language barriers. 10

Coronary arteriograms and left ventriculograms were ob-tained and analyzed according to a common protocol. Left ven-triculographic examination was performed in the right anterioroblique projection. Both ejection fraction and subjective assess-ment of left ventricular wall motion were recorded, as were theseverity and morphologic characteristics of all significant coro-nary lesions.'0 Angiographic requirements for participation inthe trial included the presence of significant operable coronaryartery disease, defined as either 70% or greater reduction in theinternal diameter of the right, left anterior descending, or leftcircumflex coronary artery, or 50% or greater reduction in theinternal diameter of the left main coronary artery. Patients with70% or greater luminal diameter reduction of the left main

*National Center for Health Statistics, Hospital Discharge Summary:Unpublished data.

coronary artery were excluded, as were those with a recognizedejection fraction of less than 0.35 and those likely to requireadditional procedures such as valve replacement or resection ofa ventricular aneurysm.

Eligible consenting patients were randomly assigned togroups by means of telephone contact with the CASS coordinat-ing center, during which the CASS Principal Investigator or hisor her designate certified the patient's eligibility, and specifiedcertain features required for stratified randomization. Randomi-zation was stratified by clinical site, number of diseased vessels,and ejection fraction within three different clinical subgroups.Patients with angina and ejection fractions of at least 0.50 wererandomized within group A, those with angina and ejectionfractions less than 0.50 were randomized within group B. andthose free of angina after well-documented myocardial infarc-tion were randomized within group C.The surgical team at each site operated on patients assigned to

receive surgical therapy as soon as the elective surgery schedulewould permit. Early postoperative angiographic examinationwas performed in a sample of patients receiving surgical ther-apy. A concerted effort was made by CASS investigators toensure optimal surgical and medical therapy for patients partici-pating in this trial.

The details of medical and surgical management have beenpublished. 10 All patients, including those assigned to the surgi-cal group, received medical therapy. For example, surgicallytreated patients were advised to modify risk factors, and thosewho developed recurrent or worsening angina during follow-upwere usually given nitrates and 3-blocking drugs. Conversely,medically treated patients whose angina worsened during fol-low-up in spite of aggressive medical therapy were not deniedbypass surgery.

Patients were examined at 6 month intervals for the durationof the follow-up period. Details concerning events such asdeath, hospitalization, new or recurrent cardiovascular symp-toms such as angina, dyspnea, and fatigue, and details concern-ing new or continuing drug therapy were routinely recorded at 6month intervals. Other data were collected less frequently, in-cluding (1) a resting electrocardiogram at 6 month intervals for2 years and yearly thereafter, (2) maximal treadmill exercise testresults at 6, 18, and 60 months, and (3) repeat coronary arterio-grams in a small number of patients at 18 months and in allwilling patients 60 months after enrollment.

Electrocardiograms were assessed at the CASS central elec-trocardiographic laboratory.'0 Electrocardiograms demonstrat-ing the interval development of significant new Q waves wereindependently read by members of a committee of expert elec-trocardiographers unaware of clinical details; disagreementswere adjudicated by the committee as a whole. Perioperativeinfarction was diagnosed, by means of serial electrocardio-grams, when the review committee agreed that a new major Qwave had appeared during the perioperative interval.The first patient was randomly assigned in August 1975, the

last in May 1979. Vital status was known on 779 of the partici-pating patients as of April 15, 1983; one surgically treatedpatient was lost to follow-up after 41/2 years. Thus, all patientshave been followed for 46 months, 69% have been followed for60 months, 40% for 72 months, and 11% for 84 months.Group differences at baseline were assessed by the chi-square

test for discrete variables; the two sample t test is used forcontinuous variables.'2 The log-rank statistic, computed fromall of the follow-up data, was used to assess the statisticalsignificance of the observed differences between survivalcurves for the medical and surgical groups.'3 This analysisincludes all randomly assigned patients. All data on events wereincluded in the calculations according to the original randomiza-tion group regardless of subsequent therapy patients may have

940 CIRCULATION


Dow

nloaded from


PATHOPHYSIOLOGY AND NATURAL HISTORY-CORONARY ARTERY DISEASE

received. Survival curves for all patients extend to 6 years,while subgroup survival curves include data from 5 years offollow-up. The small numbers of patients followed for longerintervals make estimates beyond these cutoff points quite unsta-ble. The annual mortality and crossover rates discussed beloware derived from 5 year follow-up data. For example, in figure6, A, the 5 year mortality estimate for patients assigned toreceive medical treatment is 5% (or 1% annually). In all surviv-al curves, the error bars represent one-half of 97.5% confidenceintervals for the cumulative percent survival at each year calcu-lated with the Greenwood standard error. 14 The p values cited ineach figure are two-sided and uncorrected for multiple compari-sons.

ResultsSource of patients. The CASS registry includes a total

of 24,959 patients who, during the interval from July1974 to May 31, 1979, underwent coronary arterio-graphic examination at the 15 participating CASS reg-istry sites because of suspected coronary artery dis-ease. Of this group 23% were enrolled at four CASSsites that did not participate in the randomized studyand 10% were studied during a pilot phase before thestart of the randomized trial in August 1975. Thus,16,626 registry patients were screened to determinetheir eligibility for entry into the randomized trial.

CASS REGISTRY AT RANDOMIZING SITESN = 16,626

INCLUDING PRIOR CABG, CHF, AGE > 65

N=780 1,319

RANDOMIZED RANDOMIZABLE

FIGURE 1. Allocation of patients in CASS registry at randomizingsites. Reasons for exclusion of patients from study. Width is proportion-al to the number of patients in each category. LMCA = left maincoronary artery; CABG = coronary artery bypass graft; CHF = con-

gestive heart failure.

Patients were frequently excluded for more than onereason. Exclusion factors are arbitrarily ordered in fig-ure 1 to form mutually exclusive categories. The firstcategory in this hierarchy includes individuals withnormal coronary arteries and those with less than 70%luminal diameter reduction (normal or minimal dis-ease), the next includes those with no operable vessels,and the others those with severe angina, those withsevere left main coronary stenosis, those excluded byprotocol definition, and finally those meeting protocolrequirements. The shaded portion of the figure illus-trates the percentage of patients in whom therapeuticassignment to surgery or medicine was generally un-contested. With the exception of the 12.7% eligible forrandomization, the percentages in each exclusion cate-gory in figure 1 would change with a different hierar-chy of exclusions. For example, the 36.5% of patientsexcluded because of class III or IV angina are patientswith at least one diseased vessel and at least one oper-able vessel. Therefore, this category does not includeall CASS registry patients with class III and IV angina.The total number of registry patients with 70% orgreater reduction in luminal diameter of the left maincoronary artery was 714 of 16,626 or 4.3%, but a valueof 1.5% is obtained when the chosen hierarchy ofexclusions is used.

During the recruitment period, 2099 (12.7%) CASSregistry patients met the criteria for participation in thisstudy; 780 patients, 37% of those eligible, were ran-domly assigned to receive medical or surgical therapy.Three hundred and ninety patients were assigned toeach group and 514 were group A patients, 106 groupB patients, and 160 group C patients. As illustrated infigure 1, the remaining 1319 patients were eligible forthis study but did not participate because of physicianor patient refusal. These randomizable patients havebeen followed as part of the CASS registry; they havesimilar baseline characteristics and subsequent mortal-ity. A comparison of randomized and randomizablepatients is the subject of a separate report.'5

Baseline characteristics. Randomization created bal-anced treatment groups with respect to important prog-nostic characteristics, as depicted in tables 1 and 2.The two groups of patients were similar with respect toage, sex, work status, severity of angina, history ofmyocardial infarction, hypertension, congestive heartfailure, and use of medication. The groups were wellmatched for electrocardiographic, arteriographic, andventriculographic characteristics. There were no statis-tically significant differences between patients as-

signed to receive surgical and those assigned to receivemedical therapy with respect to the baseline character-

941Vol. 68, No. 5, November 1983


Dow

nloaded from


CASS INVESTIGATORS

TABLE 1Clinical description of medical and surgical groups at entryA

Medical Surgicalgroup group Total

Variable (n = 390) (n = 390) (n = 780)

Mean age

SDSex

MaleFemale

RaceWhite

Work status

Full-timePart-timeRetired or quitOtherB

AnginaNo anginaClass IClass II

Class III and IVNonexertional

HistoryPrior MIHypertensionCongestive failureDiabetes mellitusStrokePeripheral artery disease

Cigarette use

Present smokerFormer smokerNever smoked

Medication at entry

NitroglycerinLong-acting nitrates,8-blockersAntiarrhythmics

50.97.5

90.010.0

51.57.3

90.59.5

51.27.4

90.39.7

98.7 98.0 98.3

64.14.6

22.88.5

21.512.162.30.04.1

62.629.42.38.12.19.3

71.04.619.54.9

22.116.955.60.05.4

57.232.73.99.32.17.3

67.64.6

21.26.7

21.814.559.00.04.7

59.931.13.18.72.18.3

40.8 38.7 39.743.6 44.4 44.015.6 16.9 16.3

54.944.642.69.0

56.247.344.19.7

55.546.043.39.4

AValues, except those for age, are percentages.

BUnemployed, temporary sick leave, disabled secondary to a noncar-

diac illness.

istics listed in tables 1 and 2. At entry, 59% of patientsparticipating in this trial had class II angina, 60% had a

history of myocardial infarction, 29% showed electro-cardiographic evidence of prior myocardial infarction,40% smoked cigarettes, and 43% were taking ,3-adre-noceptor-blocking drugs. In group C patients, theaverage time from documented myocardial infarctionto randomization was 286 days; 22% of group C pa-

tients were randomized within 6 weeks of infarction.The percentages of patients with single-, double-,

and triple-vessel disease, applying the definition of atleast 70% luminal reduction, was 27%, 40%, and33%, respectively. In all CASS analyses this 70% defi-nition is used, but in two other large randomized stud-

942

ies a definition of 50% luminal diameter reductionwas used. Applying the 50% definition to the random-ly assigned patients in this study changes the percent-ages to 18%, 32%, and 50% for single-, double-, andtriple-vessel disease, respectively.

Compliance to therapeutic assignment. Of 390 subjectsrandomly allocated to operative treatment, two diedbefore their scheduled date of surgery; 41 (11%) ini-tially refused surgery and thus "crossed" to medicaltherapy. During the follow-up interval, 10 of thesepatients subsequently had bypass surgery so that only31 (8%) noncompliant surgical group patients re-mained. The likelihood of a patient refusing surgery isinversely related to the angiographically determinedseverity of coronary disease. Thus, 15% of patientswith single-vessel disease assigned to receive surgicaltherapy declined operation, whereas only 3% of thosewith triple-vessel disease refused operation (p = .02).

TABLE 2Electrocardiographic, ventriculographic, and arteriographic de-scription of patients in medical and surgical groups at entryA

Medical Surgicalgroup group Total

Variable (n = 390) (n = 390) (n = 780)

Q Wave MI 29.4 29.0 29.2ST depression 11.2 8.5 9.8T wave inversion 38.4 36.7 37.6Normal ECG 30.8 27.1 28.9One-vessel diseaseB 27.4 27.4 27.4Two-vessel diseaseB 38.0 41.0 39.5Three-vessel diseaseB 34.6 31.5 33.1Left anterior descending

artery70% or greater lesion in

one or more segments 71.3 72.1 71.770% or greater lesion in

a proximal segment 30.3 32.8 31.5Left ventricular scoreC

5 39.0 39.1 39.16-10 43.9 41.9 42.911-15 15.3 17.5 16.416-20 1.8 1.5 1.7

Left ventricular ejectionfraction

< 50 21.0 20.0 20.5B 50 72.8 74.6 73.7Not measuredD 6.2 5.4 5.8

AValues are percentages.BLuminal diameter reduction of at least 70%.CFive ventricular segments viewed in the right anterior oblique pro-

jection. Grading scale: Normal = 1, moderate hypokinesis = 2, severehypokinesis = 3, akinesis = 4, dyskinesis = 5, aneurysmal = 6. Totalscore attained by adding values of each segment. '0 Maximum observedscore was 20.

DTechnically uninterpretable ventriculograms.

CIRCULATION


Dow

nloaded from



Patients assigned to surgical therapy who accepted thisassignment from the outset had bypass surgery a meanof 54 days after randomization (range 5 to 345 days).During the follow-up period 10 patients who were as-signed to the surgical group but did not comply initial-ly had surgery a mean 934 days after randomization.Of 390 patients assigned to receive medical therapy,

23.5% (4.7% per year) had bypass surgery by 5 yearsand thus "crossed" to surgical therapy. In most in-stances crossover was due to the development of wors-ening angina in spite of vigorous medical therapy. Theestimated annual rate of crossover from the medicalgroup to surgical therapy is directly related to the ex-tent of coronary disease noted at entry. The annualcrossover rates for those in the medical group who hadsingle-, double-, and triple-vessel disease were 2.0%,4.2%, and 7.6%, respectively. The observed differ-ences in annual crossover rates among these subgroupsof patients are unlikely to have arisen by chance (p <.001). Cumulative time to surgery in patients assignedto receive surgical therapy and in those assigned tothese medical subgroups is shown in figure 2.

Surgical therapy. In patients randomly assigned toundergo surgery, operative mortality, defined as deathoccurring in the hospital or within 30 days of surgery,was 1.4% (five of 357 patients). One operative deathoccurred at each of three CASS sites; there were twooperative deaths at another site. There were no opera-tive deaths at the seven other CASS randomizing sites.The operative mortality in those patients initially as-signed to medical therapy but subsequently undergo-ing surgery was 2.0% (two of 100 patients). Periopera-tive myocardial infarction in the patients assigned tosurgical therapy occurred in 23 of 357 or 6.4%. Opera-tive mortality and perioperative myocardial infarctionrates are for all patients followed over the entire fol-low-up interval, which ended April 15, 1983.A total of 954 distal anastomoses were constructed

in 357 patients in the surgical group, an averge of 2.7

LXJ

:DVn

0

100 '...-00

80 .. ....

60 .. ... 60

LEGEND2 VES MED ASSIGN 4040 22VES -MED ASSIGN.40.....--3 VES MED ASSIGN

_ SURGICALLY ASSIGNED20 .................. ................. ................... 20

n t i i

YEAR 0 3

FIGURE 2. Percent of randomly assigned patients in the surgical andmedical groups with single-, double-, and triple-vessel disease who hadnot yet undergone surgery at various time intervals.


per patient. In patients with single-, double-, and tri-ple-vessel disease the average number of distal anasto-moses was 1.7, 2.7,and 3.4, respectively. Graft paten-cy in 129 patients was assessed within 60 days ofsurgery and the results have been published. "0 A totalof 334 distal anastbmoses were evaluated in the 129patients; 90% of the grafts were open, 97% of patientshad at least one open graft, and in 81% of patients allgrafts were patent.

Medical therapy. A standard protocol for medicaltreatment was developed by CASS investigators andshared with physicians involved in the care of all of therandomly assigned patients.'0 This included efforts tomodify risk factors as well as the use of nitrates and /3-blocking drugs. fl-blocking drugs were used withequal frequency in both groups at entry, but were sub-sequently given more frequently to patients assigned tomedical therapy. The percentages of patients in themedical group given fl-blocking drugs at 12, 36, and60 months were 63%, 65%,and 64%; the correspond-ing percentages for those in the surgical group were22%, 28%, and 34%. These differences are statistical-ly significant (p < .001). Approximately 25% of pa-tients in both groups who smoked cigarettes when theyentered the trial reported at follow-up contacts thatthey had quit. There was no difference between the twogroups with respect to the proportion who smokedduring the follow-up period. The severity of angina,use of medication, rates of recurrent nonfatal myocar-dial infarction and hospitalization, as well as otherindices of quality of life are topics of separate re-ports.24' 25

Mortality. There were no significant differencesamong survival rates for the two treatment groups (fig-ure 3) or for any of the subgroups (figures 4 through 7).A total of 65 deaths occurred during the entire follow-up interval through April 15, 1983. Thirty-six patientsin the medical group and 29 in the surgical group died.

Patients assigned to medical therapy experienced anestimated annual mortality of 1.6%. For patients withsingle-vessel disease treated medically, annual mortal-ity was 1.4%; in patients with double-vessel diseasethe rate was 1.2% and in patients with triple-vesseldisease, 2.1%. In patients assigned to medical treat-ment who had moderately impaired ventricular func-tion (ejection fraction less than 0.50) the annual mor-tality rate was 3.3%.

The estimated annual mortality rate in patients as-signed to surgical therapy was 1.1%. In patients withsingle-, double-,and triple-vessel disease, annual mor-tality rates were 0.7%, 1.0%, and 1.5%, respectively.In the surgically assigned patients with ejection frac-

943


Dow

nloaded from


CASS INVESTIGATORS

tions less than 0.50 there was an annual mortality rateof 1.7%.

There were no statistically significant differencesamong survival curves for medical and surgical groupswhen patients were grouped according to presence ofsingle-, double-, or triple-vessel disease (figure 4), byclinical subgroup (figure 5), ejection fraction, or by acombination of number of diseased vessels and ejec-tion fraction (figures 6 and 7). There was no statistical-ly significant difference between patients with leftanterior descending coronary artery lesions at baselinebetween medical and surgical groups.

The mortality in patients in the medical group wasincreased in the presence of impaired ventricular func-tion (ejection fraction of less than 0.50) when com-pared with that in those with ejection fractions of atleast 0.50 (p < .0001); the survival curve for patientsin the surgical group who had impaired ventricularfunction and that for those with normal ventricularfunction (ejection fraction of at least 0.50) were notsignificantly different (p - .30).A number of additional subgroups were formed

based on patient characteristics at entry into the studyand differences in mortality among the groups wereexamined. Variables used in the formation of groupsincluded presence or absence of myocardial infarction,degree of myocardial jeopardy,26 history of hyperten-sion, presence or absence of ST depression on a restingelectrocardiogram, and history of peripheral vasculardisease. No statistically significant differences be-tween the patients in the medical and surgical groupswere observed in these subgroups.

DiscussionPatients with coronary artery disease vary widely

with respect to severity of symptoms, extent and distri-bution of coronary artery lesions, ventricular function,and prognosis. For example, recent CASS registry es-timates of annual mortality in patients managed non-operatively range from 1.3% in those with single-ves-sel disease and well-preserved ventricular function to12.5% in those with triple-vessel disease and impairedventricular function.27 This CASS registry study in-cluded patients with coronary disease of a wide rangeof severity. Noninvasive descriptors appear to addindependent prognostic information. 14 Given the largereported range of annual mortality rates in subgroupsof patients treated medically, a treatment strategy ofimmediate surgery may be advantageous in some pa-tients but not in others.CASS, which incorporates both a registry and a

randomized trial, is designed to permit observational944

:D

LXIC)

o-J

LX

Q

a-

YEAR

P = 0.3

N ALL RANDOMIZED PATIENTS100 3 381. 37 *. 288.... 17a 100

N 4 382 7

80 .... ............ ... .................. 80

60 ............ ................ , . 60

40 .................. ............ 40LEGEND

20 °- MEDICALLY ASSIGNED 20SURGICALLY ASSIGNED 20

C 00 1 2 3 4

%SURVfVAL100 98 97 96 94100 98 98 97 96

54

5

9295

6

9092

FIGURE 3. Six year cumulative survival rates for all study patients.Numbers of patients in medical and surgical groups followed to eachyear appear on the curve. At time zero the number assigned to medicaltherapy is topmost; thereafter, the number refers to nearest data point. Inthese and all subsequent survival curves, the error bars represent one-half of 97.5% confidence intervals for the cumulative percent survival ateach year with the use of the Greenwood standard error.

data analysis across the entire spectrum of patientsundergoing diagnostic coronary arteriography (CASSregistry) as well as a more rigorous comparison ofmortality in well-defined subgroups randomly as-signed to medical or surgical therapy (CASS random-ized trial). This methodologic approach delineates thepopulation from which trial participants are drawn,which in turn permits more confidence in applying trialfindings to the general population of patients undergo-ing coronary arteriography.

In 71% of the 16,626 registry patients undergoingcoronary arteriography at CASS randomizing sitesduring the recruitment period, assignment to medicalor surgical therapy would not have been a point ofcontention. As illustrated in figure 1, 28% of thesepatients had either normal coronary arteriograms or nolesion causing 70% or greater luminal diameter reduc-tion and no left main lesion causing 50% or greaterluminal diameter reduction. An additional 4.9% hadno operable vessels. Thus, 33% of CASS registry pa-tients were not candidates for surgery and were there-fore excluded from participation in the trial.

Class III or IV angina was present in 36.5% ofregistry patients; patients with left main coronary le-sions causing at least 70% luminal diameter reductionmade up an additional 1.5%. Coronary artery bypasssurgery results in either complete relief or partial im-provement in symptoms of chest pain in a majority ofpatients with severe angina7' 8; patients with severe an-gina were therefore excluded from the randomized tri-al. Bypass surgery improves longevity in patients withsevere left main artery lesions,'6 21 so randomization ofthese patients would not have been ethical. Thus, 38%

CIRCULATION


Dow

nloaded from



N 1 DISEASED VESSEL107 106 106 105 101 731O,, - a 100N 105 105 105 1

80 7 80

60 ........... .............. .. ...... , 60

40 ..................... 40LEGEND

20 ' e- MEDICALLY ASSIGNED 20SURGICALLY ASSIGNED

00 1 2 3 4 5

%SURVVAL100 99 99 98 94 93100 98 98 97 96 96

42

N 2 DISEASED VESSELS100 146 157 157 155 121

YEAR 0%SURVIVAL

_ 100100

P = 0.52

100

80

60

40

20

YEAR -0%SURVIVAL

_ 100_ 100

P 0.18

9998

29798

39798

4 5

95 9498 95

N 3 DISEASED VESSELS135 122 121 119 16 10

, *~~~~~~~~... T..... 1.. 1 0 0..9-123

IN 13213 12 13 B

... ... 10..

60

LEGEND

-... 1*- MEDICALLY ASSIGNEDSURGICALLY ASSIGNED

9899

2

9698

3

9597

60

40

20

.U4 5

91 8994 93

due to geographic, language, or psychologic barriers.The findings of the CASS randomized trial are mostclearly applicable to the remaining 2099 patients,12.7% of all patients undergoing diagnostic coronaryarteriography during the trial recruitment interval and44% of patients for whom the best therapeutic choice isa matter of the physician's judgment and patient pref-erence rather than generally accepted practice.

Execution of the CASS protocol was successful.Patients randomly assigned to medical or surgical ther-apy had similar clinical and anatomic features (tables 1

Ut

zLU

a-

YEAR

P = 0.7

N GROUP A100 254 252 252 251 242 194 100N040 255.254.254.247 167

260~~~~~~~~~~6

60 ....................... ..............; 60

40

20

...............................

LEGEND- MEDICALLY ASSIGNED .

SURGICALLY ASSIGNED

40

20

IO0 1 2 3 4 5

%SURVrVAL99 99 99 95

100 98 98 97 98 96

N GROUP B54 52 52 5 0 4.

-N 5i3

45 3

LEGENDMEDICALLY ASSIGNED .........................SURGICALLY ASSIGNED

100

80

60

40

20

,9

100'

§ Bo> 6

:Du) 60

zLU 400rLUtL 20

40

20

0 . OYEAR 0 1 2 3 4 5

%SURVtVAL100 98 94 93 85 85

- 100 100 100 98 96 96

P 0.057

FIGURE 4. Five year cumulative survival in groups with single-,double-, and triple-vessel disease, with disease being defined as at least70% luminal diameter reduction.

of registry patients were excluded from this study be-cause surgical therapy was known to be effective forrelieving symptoms or prolonging life.The remaining 29% or 4780 patients in the CASS

registry (unshaded portion of figure 1) had clinical andanatomic features that make the choice of treatmentdifficult: symptoms were not severe, coronary anato-my was surgically approachable, and the effect surgery

would have on longevity was not known. Of this groupof 4780 patients, 56% were excluded by the CASSprotocol for a variety of scientific and practical reasons

including prior coronary bypass surgery, severe con-

gestive heart failure, age over 65, or inaccessibility


LI

zLJ

L...LU

100'

80

60

40

20

YEAR%SURVIVAL

_ 100100

P 0.79

N GROUP C82 77 77 76 75

.T10076

_

.7 80

LEGEND*o_ MEDICALLY ASSIGNED

SURGICALLY ASSIGNED

9699

40

20

,2 3 4 5

94 91 90 8999 97 96 69

FIGURE 5. Five-year cumulative survival in clinical subgroups. A,Patients with mild-to-moderate angina and with ejection fractions of at

least 0.50. B, Patients with mild-to-moderate angina and ejection frac-

tions of less than 0.50. C, Angina-free survivors of well-documented

acute myocardial infarction.

945

-J

>

u)

zLJ

YEAR

P =0.4

80AL

D

U) 60-

z

40-0

ay-LU

tL 20-

,,0 -11.- -

-N 1S7 145...

............. .....................

............... ...

LEGENDMEDICALLY ASSIGNEDSURGICALLY ASSIGNED

(I1

zLUJC-),-LUJa-

u l I

o

ip-143 i'!-a 11.., ...................... 49tnffu

r

I .......................-

I

_j n


Dow

nloaded from


CASS INVESTIGATORS

N A. EF >= 0.50M4 280 280 279 272

N IN W 16" 214-100291

100286 265 A5N 277 207'

80 .. ... ... ... 80

60 ............ ...... .......... 60

z40 .. .. 40

LEGENDa 20 MEDICALLY ASSIGNED 20SURGICALLY ASSIGNED

0 0

10(

>'8(

6

z

02

NB. 1 DISEASED VESSEL & EF >= 0.5085 94 94 94 89 67

Sir.. I...100

N 9 3 9

10-. .. 80

0o .-.. ..60

10 40- LEGEND

0 - .. MEDICALLY ASSIGNED -20SURGICALLY ASSIGNED

0 ~~~~~~~~~~~~~0YE-AR 0 1 2 3 4 5 YEAR 1 2 3 4 5

%SURV1VAL %SURVIVAL100 99 99 96 95

.-100 99 99 99 96 95

98 98 96 95 100 98 98 97 96 96

P 0.55 P= 0.74

C. 2 DISEASED VESSELS & EF >= 0.50so. 98 Oa go 77

1004 . .-.,a - 0:.: ".1 i. 11004-

N i23 123 122 94

80 80

60 .. .... .. 60

40 ... .. 40LEGEN

MEDICkLLY ASSIGNED ................... 2020 SURGICALLY ASSIGNED

0 0

%SURVIAL100 99 99 99 97

::: 100 98 98 98 97 96

P 0.13

0. 3 DISEASED VESSELS & EF >=- 0.50go 70 89 69 67 54

100 70 - j--0 - - ----3-..l - -T---- 100. N 88 a8 67 8 -8

Bo80-

v() 60 -

zL' 40 40of - LEGEND

CL 20 - SURGEDIALLY ASSIGNED ....20.SMRGICALLY ASSIGNED 2

0 0YEAR 0 2 3 4 5

%SURV1VAL¶00 98 98 97 96 94100O 100 99 99 96 94

P = 0.91

FIGURE 6. Five-year cumulative survival rates for patients with ejection fractions of at least 0.50 (A) and patients with ejectionfractions of at least 0.50 and single- (B), double- (C), and triple- (D) vessel disease. Luminal diameter reduction of at least 70%was required for diagnosis of vessel disease.

and 2). Vital status was known on all patients but onelost to follow-up after 4½2years. Compliance to thera-peutic assignment was influenced by the developmentof worsening angina in those assigned medical therapyand by reconsideration in a small proportion of thoseassigned to surgery. Crossover rates were similar to

.80 -

vs 60-

zLi 40-

10L

P - 0.085

those observed in both the European and Veterans Ad-ministration trials. 17. 23

For all analyses reported in this article patients wereconsidered to belong to the group to which they were

initially assigned (intention to treat) regardless of sub-sequent therapy they received. In clinical trials as in

so

vs 601

z

401Li0- 20

0 2 3 4 5IRVIVAL)o 99 95 91 87 84

o 99 99 97 96 Q92

,B. 1 DISEASED VESSEL & EF < 0.50-11 11 11 8-a 8 5

'80

60

40

20

YEAR 0 1 2 3 4 5%SURVIVAL

too 100 100 91 81 81100o 100 100 100 100 100

P =0.95

,C. 2 DISEASED VESSELS & EF < 0.50 2235 35 28 28 28

10029 2

80 32 3

6

z

so

4

50

100o

80

60

40

YEAR 0 1 2 3 4 5

XSURV1VAL0- 100 100 94 91 91 88

100 100 100 100 92

P - 0.85

ND. 3 DISEASED VESS17LS & EF < 0.5036

41 41 40 39 32ioor -~~~~~~. 100

35~~~~~~~~~3:D 60... 25 60

40 . 40- LEGEND

C 0 . MEDICALLY ASSIGNED ....... .......2020.. SURGICAL-LY ASSIGNED

YEAR 0 1 2 3 4 5100VVA 97 94 92 83 801,00 98 98 95 93 90

P - 0.063

FIGURE 7. Five-year cumulative survival rates for patients with ejection fractions of less than 0.50 (A) and patients with

ejection fractions of less than 0.50 and single- (B), double- (C), and triple- (D) vessel disease. Vessel disease was diagnosed as

described in figure 6.

:D

a--

LEGEND

MEDICALLY ASSIGNEDSURGICALLY ASSIGNED

LEGEN0o---00 MEDICALLY ASSIGNED :..... 121SURGICALLY ASSIGNED ......

IM . tloo

0

946 CIRCULATION

CK


Dow

nloaded from



the practice of medicine some patients do not complywith therapeutic assignment, raising the question ofwhether a more appropriate analysis might be compari-son of outcome in groups created by treatment receivedrather than treatment assigned. However, analysis bytreatment received violates the process of randomiza-tion; groups created in this post hoc fashion are quitelikely to be unbalanced for both known and, moreimportantly, unrecognized prognostically importantcharacteristics. An unbiased comparison of outcome isdifficult, if not impossible, when analysis is by treat-ment received. Analyses of CASS mortality data bytreatment received and by means of censoring at thetime of crossover do not change the results. Theseanalyses are not reported since they are of dubiousscientific merit in the context of a randomized trial.

Analysis by assigned treatment influences the hy-potheses that can be tested because of the slow, steadyrate of crossover to surgical therapy in the medicalgroup. Given the observed degree of noncompliance totreatment assignment, CASS cannot definitively testthe true contrast in outcome between medical and sur-gical therapy, since it is impossible from a practicalstandpoint to restrict therapy during follow-up. Pa-tients assigned to the medical group who developedworsening angina during the course of the study werenot denied the opportunity to seek improvement bybypass surgery and surgically treated patients withworsening angina were treated medically as well.However, the CASS tests the practical hypothesis thata strategy of prompt elective surgery results in a differ-ent mortality rate than a strategy of deferred surgerycontingent upon worsening symptoms.

Patients undergoing coronary artery bypass graftingas part of the CASS in the period from 1975 to 1979enjoyed excellent results. The operative mortality of1.4% is similar to the currently reported surgical risk.8The number of grafts fashioned and bypass patencyrates are similar to those reported in other recent series.The incidence of perioperative myocardial infarction,determined according to a vigorously applied algo-rithm, was 6.4%.24 This rate is difficult to comparewith the reports of others because of wide variability inthe criteria used to diagnose this complication, but it iswithin the range of rates reported in studies relying onelectrocardiographic Q wave criteria.8Recommended medical therapy is outlined in the

CASS protocol. '1 Patients were referred to the partici-pating CASS sites from a large number of practicingphysicians. After entry into the randomized trial, pa-tients were returned to the care of their personal physi-cians. The recommended outline of medical treatment

was forwarded to these physicians. No attempt wasmade to further standardize medical therapy; the re-sponsible physician was free to choose optimal therapyaccording to his or her best judgment. At each follow-up visit, symptoms and therapy were carefully re-viewed and recorded.

The survival of CASS patients assigned to medicaltherapy is excellent. When the CASS was initiallydesigned, review of available published data suggestedthat the annual medical mortality rate in patients withproven coronary artery disease would be approximate-ly 4%. Calculations have been published estimatingthe statistical power of CASS to detect clinically im-portant differences in survival for an assumed annualmedical mortality of 2% and 4%. 10The annual mortality in patients assigned to medical

therapy is 1.6%, a figure considerably lower than thatestimated during trial design and lower than the annualmedical mortality of 3.3% noted in the European Col-laborative Study23 and than the 4.3% reported in theVeterans Administration Study.'7 The excellent long-term results in CASS patients assigned to medical ther-apy reduces the power of the experiment to detect areduction in mortality afforded by prompt electivecoronary artery bypass. Thus, if medical mortality forthese patients is in truth 1.6% a year, a 50% reductionin mortality would be discerned with the present sam-ple size with a probability of .65. A 30% reductionwould be detected with a probability of .26.28 Indeed, asample size of over 4600 would have been required todetect a 30% reduction in mortality with 90% suretygiven the observed CASS medical mortality rate.28 Theclinical significance of the observed nonsignificant re-duction in estimated mortality for the individual pa-tient is quite small (1.1% vs 1.6% annual risk). Thedifference in the annual mortality rates for patientswith good ventricular function at entry is even smaller(0.8% vs 1.0%).

Only 160 patients with ejection fractions of less than0.50 participated in the present study. As shown infigure 7, the survival curves reveal a statistically non-significant advantage for the patients assigned to surgi-cal therapy, especially in those with three-vessel dis-ease. Because of the small sample size, a largeimprovement in survival due to immediate surgicalintervention cannot be excluded as a possibility. Forexample, if medical mortality in this subset of patientsis 3.3% a year, a sample size of 650 would be requiredto discern a surgically induced 50% reduction in 5 yearmortality in this subgroup with 90% surety. A 30%reduction would require a sample size of over 2000.28

Analysis of data from the 1319 patients qualifying

947Vol. 68, No. 5, November 1983


Dow

nloaded from


CASS INVESTIGATORS

for randomization but declining to participate (the ran-domizable group for whom therapy was chosen in con-sultation with the personal physician rather than byrandom assignment) reveals an outcome in those treat-ed medically similar to that reported in the randomlyassigned patients who received medical therapy. Pa-tients who agreed to be randomly assigned to treatmentdid not differ in long-term survival when comparedwith those who declined randomization. 15

In contrast to the European Collaborative Study, theCASS does not show a significant decrease in survivalrates for patients with three-vessel disease and ejectionfractions greater than 0.50, or in patients with two- orthree-vessel disease involving the left anterior de-scending coronary artery. The subgroup in this study inwhich a trend in favor of surgery (those with three-vessel disease with ejection fractions less than 0.50)was revealed was not included in the European Study.The reasons for the apparent differences between re-sults of the two studies may lie in variations in referralpatterns, the medical therapy used in the two studies,the exclusion of patients with single-vessel disease inthe European study and, perhaps most importantly, theinclusion of patients with more severe angina in theEuropean trial. The excellent survival rate observed inCASS medical patients is the major difference betweenthe results of the two studies.

Neither the Veterans Administration Study nor theCASS shows a significant overall difference in out-come when medical and surgical therapy are com-pared. Since the Veterans Administration Study wasinitiated, changes have been made in the process ofselection of patients for coronary arteriography and insurgical technique (with a lowering of operative mor-tality) and there have been improvements with respectto graft patency and perioperative infarction rate, aswell as more effective application of medical therapy,including administration of nitrates and /3-blockingdrugs. These secular changes make close comparisondifficult. In contrast to the Veterans AdministrationStudy, CASS included patients with less impairmentof ventricular performance at entry. The small CASSsubgroup with impaired ventricular performance moreclosely resembles patients included in the VeteransAdministration trial. An analysis of CASS subgroupsof patients with prior myocardial infarction, ST seg-ment depression apparent on the resting electrocardio-gram, and history of hypertension does not reveal sta-tistically significant survival differences betweenpatients assigned to receive medical and those assignedto receive surgical therapy.CASS trial results are most applicable to patients

with mild-to-moderate angina and to angina-free survi-vors of myocardial infarction with surgically ap-proachable coronary artery disease. CASS registrydata reveal that 12.7% of all patients undergoing diag-nostic coronary arteriography at the 11 CASS sitesfrom August 1975 to May 1979 meet these criteria.More importantly, this subset includes 44% of thosepatients (unshaded area of figure 1) for whom thetherapeutic choice between surgery and medical man-agement is a matter of judgment rather than widelyaccepted practice. In these patients a treatment strategyof prompt elective coronary artery bypass surgery doesnot result in improved survival when compared with astrategy of medical management unless worseningsymptoms make surgery necessary. In patients withunimpaired ventricular performance on entry into thestudy (ejection fraction of at least 0.50), the survivalcurves are indistinguishable. In those with impairedventricular performance, a statistically nonsignificanttrend in favor of surgery is evident, and may representa true survival advantage for patients assigned to surgi-cal therapy. Insufficient numbers of patients in thissubgroup prevent rigorous testing of this hypothesis,but further follow-up of CASS randomly assigned pa-tients is planned and may clarify this trend.

In patients similar to those enrolled in the CASStrial, a strategy of medical management until worsen-ing symptoms require surgical palliation does not carrya mortality penalty. Furthermore, deferring surgerydoes decrease immediate health care costs and pre-serve both veins and mediastinum if surgery is requiredlater. Thus, the decision to proceed with bypass sur-gery in these patients can be reduced to a question ofwhether the current level of symptoms is acceptable tothe patient. These findings have important implica-tions for both cardiologists and cardiac surgeons intheir diagnostic and therapeutic efforts.CASS

Clinical centersUniversity of Alabama in Birmingham, Birmingham, Prin-

cipal Investigators: William J. Rogers, M.D., Richard 0.Russell, Jr., M.D. (former); Associate Investigators: AlbertOberman, M.D., Nicholas T. Kouchoukos, M.D.; Data Tech-nicians: Rosa Kimerling, Mary Lynn Mandy, Glenda Barnes,R.N. (former), Marlene Means (former), Phyllis Newsom (for-mer); Others: Benigno Soto, M.D., P. Nath, M.D., Sanford P.Bishop, Ph.D., Sammy Scalisi, Helen Boerner, Martha Arm-strong, Victoria Talley, Jeanne Arnold (former), Matt Beavers(former), Jim Kitts, P.A. (former), Toni Price (former), MarianOberman (former), Suzanne Bearman (former), Myrna Gold-stein (former), Sharon Stein (former), Morty Roach (former),Steve Williams (former), William Brodie, R.N. (former).

Albany Medical College of Union University, Albany, Prin-cipal Investigators: Eric D. Foster, M.D., Julio A. Sosa, M.D.;Associate Investigators: Joseph T. Doyle, M.D., Martin F.

CIRCULATION948


Dow

nloaded from



McKneally, M.D., Joseph B. Mcllduff, M.D., Harry Odaba-shian, M.D. (former), Thomas M. Older, M.D.; Study Coordi-nator: Joyce A. Carpenter; Data Technicians: Sally Abele, Ste-phen Cameron, William Olewnick, Frank Strickler.

Boston University Medical Center and Evans Memorial De-partment of Clinical Research, Boston, Principal Investigator:Thomas J. Ryan, M.D.; Associate Investigators: Robert L.Berger, M.D., David P. Faxon, M.D., David C. Heuter, M.D.(former), Mirle A. Kellett, M.D., John R. McCormick, M.D.,Edgar C. Schick, M.D., Donald A. Weiner, M.D., Laura F.Wexler, M.D.; Project Managers: Carolyn H. McCabe, B.S.,Sandra Pfund, R.N. (former); Data Technicians: Sally S. Cut-ler, B.S., Patricia DeAngelis, R.N. (former), Patricia Hollum(former), Mary-Grace Lett; Financial Manager: WandaMcGovern, B.A.Loma Linda University School of Medicine, Loma Linda,

Principal Investigators: C. Joan Coggin, M.D., Melvin P. Jud-kins, M.D., Ellsworth E. Wareham, M.D.; Associate Investi-gators: Winston A. Mitchell, M.D. (former), Patrick M. Mo-loney, M.D.; Data Technicians: Isla Andersen, Carol Mason,Pat Meyer; Project Coordinator: Eileen Judkins, R.N.

Marshfield Medical Foundation, Inc., Marshfield Clinic,Marshfield, WI, Principal Investigators: William 0. Myers,M.D., Richard D. Sautter, M.D.; Associate Investigators: Rob-ert D. Carlson, M.D., John N. Browell, M.D., Edgar L. Koch,M.D. (former), Jefferson F. Ray, III, M.D., George S. Strauss,M.D. (former), Dieter M. Voss, M.D., Dean A. Emanuel,M.D., Richard H. Ulmer, M.D.; Data Technicians: AnitaArndt, R.N., Research Associate (former), Barbara Ault, R.N.,Research Associate (former), Carrie Kraemer, R.N., ResearchAssociate; Other: Rosie Weber.

Massachusetts General Hospital, Boston, Principal Investi-gators: J. Warren Harthome, M.D., W. Gerald Austen, M.D.,Eldred Mundtn, M.D. (former); Associate Investigators: RobertE. Dinsmore, M.D., Stephen Miller, M.D., Frederick Levine,M.D. (former), Dimitrios Linos, M.D. (former), John McDer-mott, M.D. (former), Frederick Poulin, M.D. (former), SayedQuadri, M.D. (former), Vincent DiCola, M.D.; Data Techni-cians: Emily Gmitter, Lucy Matson, May Beth Connor (for-mer), Jane Porter (former), Mary Ferraro (former), StephenGottlieb (former), Margaret Fleury (former), Elizabeth Chan-trell (former), Margaret Brungraber (former), Patricia A. Gal-vin (former), Susan Rossi (former), Andrea Freeman (former),Carol McGlinchey (former), David Bindman (former).Mayo Clinic, Rochester, Principal Investigator: Robert L.

Frye, M.D.; Associate Investigators: Michael B. Mock, M.D.,Ronald E. Vlietstra, M.D., Bernard J. Gersh, M.D., HartzellV. Schaff, M.D., Hugh C. Smith, M.D., David R. Holmes,M.D.; Data Technicians: LaVon Hammes, Susan Brevig, JuneDunfee, Mary Curran, Kathy Kremer; Secretaries: Daphne Lin-coln, Karen Fabian.Miami Heart Institute, Miami Beach, Principal Investigator:

Arthur J. Gosselin, M.D.; Associate Investigators: Paul S.Swaye, M.D., Ernest Traad, M.D., Parry B. Larsen, M.D.(former); Data Technicians: Hazel J. Yon, Rose P. Callen,Marilyn Baram (former), Evelyn Ananias (former), Maria Walz(former), Debrah Fiore-Bradley (former), Donna Dobbs (for-mer); Secretaries: Alexis Benjamin (former), Beth Hendel (for-mer), Rosita Mila (former), Claudette Broomfield (former),Maria Canet (former).

Institut de Cardiologie de Montreal, Montreal, Principal In-vestigator: Martial G. Bourassa, M.D.; Associate Investigators:Claude Goulet, M.D., Jacques Lesperance, M.D., Bernard R.Chaitman, M.D. (former), Claude M. Grondin, M.D., B.Charles Solymoss, M.D., Yves Castonguay, M.D., LucienCampeau, M.D.; Research Associate: Margot Methe, R.N.;Data Technicians: Micheline Labbe, Claudette Faille, Celine

Guilbeault, Manon Bellemare (former), Diane Tourville (for-mer).New York University Medical Center, New York, Principal

Investigators: Ephraim Glassman, M.D., Michael Schloss,M.D.; Associate Investigator: Wayne Isom, M.D.; Data Tech-nicians: Deborah Walzer, Carolyn Mungo, Ruth Gonzalez.

St. Louis University School of Medicine, St. Louis, PrincipalInvestigators: George C. Kaiser, M.D. (Director), J. GerardMudd, M.D.; Associate Investigators: Naseer Ahmad, M.D.,Henrick B. Bamer, M.D., John E. Codd, M.D., David M.Harshman, M.D., John P. Merjavy, M.D., D. Glenn Penning-ton, M.D., Dennis H. Tyras, M.D., Robert D. Wiens, M.D.,Vallee L. Willman, M.D., Hillel Laks, M.D. (former), Alex-ander Mark, M.D. (former); Data Technicians: Polly PorterFogarty, Project Administrator, Jean LoPiccolo, Assistant Proj-ect Administrator, Karen Kelley, Marcea Freeman, EstherMuller, Rex Tannahill, Judi Willenbrink, Kristine Allen (for-mer), James Bertram (former), Douglas Clermont (former),Claudia Day (former), Barbara Fineberg (former), Kelly Fitz-gerald (former). Dorothy Hoffman (former), Mary McCarthy(former), Susan Neill (former), Pat Otto (former), GeorgeSchroeder (former), Diane Waack (former), Marc Zelenka (for-mer).

St. Luke's-Roosevelt Hospital Center, New York, PrincipalInvestigator: Harvey G. Kemp, M.D.; Associate Investigators:Airlie Cameron, M.D., George E. Green, M.D., John E.Hutchinson III, M.D., Haroutune A. Mekhjian, M.D.; DataTechnicians: Lillian Agosto, Carlos Coronado, Patricia Gries-hammer, Petronella Marshall, Joanne Oh, Amy Patton, RosalieSinkler, Synthia Vega-Piontek, Monica Wachter, Susan Walsh,Paula Zakos, Patricia Zorena.

Stanford University Medical Center, Stanford, Principal In-vestigator: Edwin L. Alderman, M.D.; Coinvestigators: JamesSilverman, M.D., Edward B. Stinson, M.D.; Associate Investi-gators: Francis Koch, M.D., Paul Cipriano, M.D. (former);Data Technicians: Adriana Krauss, Geeta Charjee, CandaceRabiroff, Anne Schwarzkopf; Sandra Avery (former); Staff As-sistant: Susan Groechel.

The Medical College of Wisconsin, Veterans AdministrationMedical Center, Milwaukee, Principal Investigators: Harold R.Brooks, M.D., Felix E. Tristani, M.D., Lawrence Bonchek,M.D. (former), Michael Keelan, Jr., M.D. (former); AssociateInvestigators: Virinderjit S. Bamrah, M.D., Robert J. Flemma,M.D., C. Vincent Hughes, Jr., M.D., Donald Tresch, M.D.(former); Data Technicians: Ronald Mack, Patricia Ryan, Mar-lene Bednarz (former).

Yale University School of Medicine, New Haven, PrincipalInvestigator: Lawrence S. Cohen, M.D.; Associate Investiga-tors: Alexander S. Geha, M.D., Graeme L. Hammond, M.D.,Rene A. Langou, M.D., Richard Shaw, M.D. (former), StevenWolfson, M.D. (former); Data Technicians: Caroline Jacobs,R.N. (former), Abby Miller, B.A., Barbara Nemier.

Central electrocardiographic laboratoryUniversity of Alabama in Birmingham, Birmingham, Princi-

pal Investigator: L. Thomas Sheffield, M.D.; Associate Direc-tor: David Roitman, M.D.; Coinvestigator: Santosh Kansal,M.D.; Data Management Supervisor: Carol Troxell, B.S.; Ad-ministrative Assistant: Myrnie Driskill, B.S.; Data ProcessingCoordinator: Monika Whitcomb; Clerical Supervisor: EvonneLawrence; Secretary: Juanita Brasher; Clerk/Stenographer:Sherry Tew; Coding Clerks: Shirley Decker, Jo Ann Kelley,Paula Stanton, Jeanne Arnold.

Coordinating center

University of Washington, Seattle, Principal Investigators:Lloyd D. Fisher, Ph.D., Richard A. Kronmal, Ph.D. (former);

949Vol. 68, No. 5, November'1983


Dow

nloaded from


CASS INVESTIGATORS

Associate Investigators: Kathryn B. Davis, Ph.D., Mary JoGillespie, M.S. (Deputy Director), J. Ward Kennedy, M.D.,Richard A. Kronmal, Ph.D.; System Programming and DataProcessing: Lois Danks, Brian Leverson, Erik Lundberg, RuthMcBride (Supervisor), Rita Mines, Thomas Ormond, ShimonBenami (former), Mary Ann Betts (former), Harvey Friedman(former), Susan Harada (former), Kelly Landreth (former),Robert Mauritsen (former), Thomas Rawson (former); Analysisand Scientific Programming: Marion Athearn, Richard Ells,James Ferguson, David Fray, Cheryl Look, Stephen Luk,Charles Maynard (Supervisor), Mary Pettinger, John Wilson,Elisabeth Cardis (former), Leland DePriest (former), WilliamHalteman (former), Robert Jones (former), Paul Litwin (for-mer), Paula Roberson (former), Bruce Schaalje (former), An-and Seth (former), Dalice Sim (former), Vicki Stover-Hertz-berg (former), Ronald Thomas (former), Terrence Tong(former), William Vollmer (former), Homer Warner (former),Barbara Williams (former), Kathryn Long; Program Staff: JoanPitell (Supervisor), Jean Savage, Denise Borer (former), Cyn-thia Bush (former), Margaret Ellwood (former), Janice Halte-man (former), Andrea Jarvela (former), Donna King (former),Rhonda Moon (former), Carol Murray (former), Patricia Thorn-hill (former).

National Heart,' Lung and Blood InstituteNational Institutes of Health. Bethesda, Current Staff of the

Program Office: Eugene R. Passamani, M.D., Peter L. From-mer, M.D., Suzanne Mullin, R.N., M.P.H., Kent Bailey,Ph.D., Barbara Packard, M.D., Ph.D., Betty Nordan; PreviousStaff: Per T. Berje, M.D., David L. Demets, Ph.D., John B.Hanks, M.D., David M. McCarthy, M.D., Michael B. Mock,M.D., AlanL. Pinkerson, M.D., Robert A. Poirier, M.D., IvarRingqvist, M.D., Patrick M. Sullivan, Hans Wedel, Ph.D.;Central Pathology Laboratory: William C. Roberts, M.D.,Bruce F. Waller, M.D.; Policy Advisory and Data MonitoringBoard: Lloyd D. Fisher, Ph.D, Ralph Frankowski, Ph.D., Wil-liam T. Friedewald, M.D., L. Julian Haywood, M.D., ThomasKillip, M.D., Francis Klocke, M.D. (Chairman), Curtis L.Meinert, Ph.D., Richard S. Ross, M.D., David Sabiston,M.D., Harmon Smith, Ph.D., Jeremiah Stamler, M.D., JohnVasko, M.D., James H. Ware, Ph.D.; Chairman, SteeringCommittee: Thomas Killip, M.D., Henry Ford Hospital, De-troit.

References1. U. S. Department of Health, Education, and Welfare: Proceedings

of the Conference on the Decline in Coronary Heart Disease Mor-tality. Public Health Service, National Institutes of Health, NIHPublication No. 79-1610, May 1979

2. National Center for Health Statistics: Monthly vital statistics re-port. Vol. 23, No. 13, September 17, 1981

3. National Center for Health Statistics, Vital and Health Statistics:Utilization of short-stay hospitals: annual summary for the UnitedStates, 1980. DHHS Publication No. (PHS) 82-1725, March 1982

4. The Norwegian Multicenter Study Group: Timolol-induced reduc-tion in mortality and reinfarction in patients surviving acute myo-cardial infarction. N Engl J Med 304: 801, 1981

5. Hjalmarson A, Elmfeldt D, Herlitz J, Holmberg S, Malek I, Ny-berg G, Ryden L, Swedberg K, Vedin A, Waagstein F, Walden-strom A, Waldenstrom J, Wedel H, Wilhelmsen L, Wilhelmsen G:Effect on mortality of metoprolol in acute myocardial infarction.Lancet 2: 823, 1981

6. Beta-Blocker Heart Attack Trial Research Group: A randomizedtrialiof propranolol in patients with acute myocardial infarction. I.Mortality results. JAMA 247: 1707, 1982

7. McIntosh HD, Garcia JA: The first decade of aortocoronary bypassgrafting, 1967-1977, a review. Circulation 57: 405, 1978

8. Rahimtoola SH: Coronary bypass surgery for chronic angina1981: a perspective. Circulation 65: 225, 1982

9. National Institutes of Health consensus-development conferencestatement: Coronary artery bypass surgery: scientific and clinicalaspects. N Engl J Med 304: 680, 1981

10. Principal Investigators of CASS and their associates (Killip T,editor, Fisher LD, Mock MB, associate editors): National Heart,Lung, and Blood Institute Coronary Artery Surgery Study (CASS).Circulation 63 (suppl I): I-1, 1981

11. Campeau L: Grading of angina pectoris. Circulation 54: 522, 197612. Snedecor GW, Cochran WG: Statistical methods, ed 6. Ames,

Iowa, 1968, The Iowa State University Press13. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV,

Mantel N, McPherson K, Peto J, Smith PG: Design and analysis ofrandomized clinical trials requiring prolonged observation of eachpatient. IL. Analysis and examples. Br J Cancer 35: 1, 1977

14. Greenwood M: The natural duration of cancer. Reports on PublicHealth and Medical Subjects, 33, London: Her Majesty's Station-ery Office, 1-26, 1926

15. CASS Principal Investigators and their associates: Coronary ArterySurgery Study (CASS). A randomized trial of coronary arterybypass surgery. Comparability of entry characteristics and survivalin randomized patients and non-randomized patients meeting ran-domization criteria. (submitted for publication)

16. Takaro T, Hultgren H, Lipton M, Detre K, and participants in theVeterans Administration Cooperative Study Group: VA Coopera-tive Randomized Study for Coronary Arterial Occlusive Disease.II. Left main disease. Circulation 54 (suppl 3): 111-107, 1976

17. Murphy M, Hultgren H, Detre K, Thomsen J, Takaro T. andparticipants of the Veterans Administration Cooperative Study:Treatment of chronic stable angina. A preliminary report of surviv-al data of the Randomized Veterans Administration CooperativeStudy. N Engl J Med 297: 621, 1977

18. Detre K, Peduzzi P, Murphy M, Hultgren H, Thomsen J, ObermanA, Takaro T, and the Veterans Administration Cooperative StudyGroup for Surgery for Coronary Arterial Occlusive Disease: Effectof bypass surgery on survival in patients in low- and high-risksubgroups delineated by use of simple clinical variables. Circula-tion 63: 1329, 1981

19. Braunwald E: Coronary artery surgery at the crossroads. N Engl JMed 297: 661, 1977

20. Hiatt H: Lessons of the coronary bypass debate. N Engl J Med 297:1462, 1977

21. European Coronary Surgery Study Group: Coronary bypass sur-gery in stable angina pectoris: survival at two years. Lancet 1: 889,1979

22. European Coronary Surgery Study Group: Prospective randomizedstudy of coronary artery bypass surgery in stable angina pectoris.Second interim report. Lancet 2: 491, 1980

23. European Coronary Surgery Study Group: Long-term results ofprospective randomized study of coronary artery bypass surgery instable angina pectoris. Lancet 2: 1173,1982

24. CASS Principal Investigators and their associates: Coronary ArterySurgery Study (CASS), a randomized trial of coronary artery by-pass surgery. Myocardial infarction results. (submitted for publica-tion)

25. CASS Principal Investigators and their associates: Coronary ArterySurgery Study (CASS): a randomized trial of coronary artery by-pass surgery. Quality of life in patients randomly assigned to treat-ment groups. Circulation 68: 951, 1983

26. Alderman EL, Fisher L, Maynard C, Mock MB, Ringqvist 1,Bourassa MD, Kaiser GC, Gillespie MJ: Deterrninants of coronarysurgery in a consecutive patient series from geographically dis-persed medical centers, The Coronary Artery Surgery Study. Cir-culation 66 (suppl I): 1-6, 1982

27. Mock M, Ringqvist I, Fisher L, Davis KB, Chaitman BR, Kou-choukos, NT, Kaiser GC, Alderman E, Ryan TJ, Russell RO Jr,Mullin S, Fray D, Killip T, and the participants in CASS: Survivalof medically treated patients in the Coronary Artery Surgery Study(CASS) registry. Circulation 66: 562, 1982

28. Eisenhart C, Hastay MW, Wallis WA: Selected techniques ofstatistical analysis. New York, 1947, McGraw-Hill Book Com-pany, Inc.. p 259

CIRCULATION950


Dow

nloaded from


surgery. Survival data.Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1983 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.68.5.939

1983;68:939-950Circulation.

http://circ.ahajournals.org/content/68/5/939the World Wide Web at:

The online version of this article, along with updated information and services, is located on

http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer information about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. FurtherEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculationpublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:


Dow

nloaded from

http://circ.ahajournals.org/content/68/5/939

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circ.ahajournals.org//subscriptions/


Coronary Artery Surgery Study (CASS): randomized trial...

Documents

Transcript of Coronary Artery Surgery Study (CASS): randomized trial...