Coronary Heart Disease Epidemiology - Circulationcirc.ahajournals.org/content/circulationaha/48/1/185.full.pdf · Coronary Heart Disease Epidemiology Revisited Clinical and Community

Coronary Heart Disease Epidemiology Revisited

Clinical and Community Aspects

By FREDERICK H. EPSTEIN, M.D.

SUMMARYEpidemiological research relating coronary heart disease risk to risk factors has been reviewed

and condensed into simple, diagrammatic form. The predictive power of risk factors for detect-ing susceptibles in the population is presented. While much is now known about long-term predic-tion, there is an urgent need for tests to signal impending heart attacks. Against this background,the potential for preventive action is discussed in the light of current knowledge. The "clinical-individual" and "community" approaches to prevention are contrasted. To what extent is preven-tion on the community level warranted, pending the results of controlled prophylactic trials? Theview is presented that advice on preventive measures to the population at large is justified at thistime, based on education and planned change but short of an all-out campaign which must awaitconclusive data from intervention studies.

Additional Indexing Words:Coronary heart disease Risk factors and coronary heart diseaseHypertension Smoking and coronary artery disease

Epidemiology has been well described as thebasic science of community medicine or, as statedvery aptly by the late Dr. Thomas Francis, Jr.,"clinical investigation on the community level." Nofield of medicine has advanced more rapidly over soshort a space of time, perhaps 15 years, thancoronary heart disease epidemiology. As a result,there is now probably no other disease which canbe predicted, in terms of predisposing factors, withsimilar accuracy and power. With the advent of thisknowledge, the terra incognita between clinicalcardiology and epidemiology is now becominginhabited by preventive cardiology,' communitycardiology and precoronary care.2 The emphasis,therefore, is not only on the acutely ill patient whowill, if within reach, receive care under any cir-cumstances, or the patient with symptoms whoshould seek care, but that much larger group in thepresymptomatic stage, the susceptibles in the pop-ulation who need long-term preventive care.

From the Center for Research in Diseases of the Heart,University of Michigan, Ann Arbor, Michigan.

This paper represents part of the Fourth Annual GeorgeC. Griffith Scientific Lecture presented in Los Angeles,under the auspices of the Los Angeles County Heart Asso-ciation in Los Angeles, California on October 13, 1971.

Address for reprints: Dr. Frederick H. Epstein, Director,Center for Research in Diseases of the Heart, University ofMichigan, Ann Arbor, Michigan 48108.

Circulation, Volume XLVIII, July 1973

Serum cholesterol

The Burden of Coronary Heart DiseaseThe burden of coronary heart disease can be

measured in terms of morbidity and mortalitystatistics. It is best, however, to start with the living,particularly the "middle-aged man" in his forties orfifties since more is known about the epidemiologyof coronary disease among such men than thosewho are younger or older, or women of any age. Byand large, the frequency of the disease amongwomen is a third or a quarter of the male rate;however, considering the high male rate, thefrequency in women is by no means low.1 3 Sincethe majority of available data apply to white men inthe more favored social groups, there is great needto learn more about the epidemiology of coronaryheart disease among the other ethnic and socialgroups in the community.

Figure 1 shows a diagram for coronary diseaseamong middle-aged men, and is based on the sumtotal of available epidemiological data." 3' 4 Itrepresents general magnitudes rather than precisefigures but can be viewed with a fair deal ofreliance. Among 1,000 men, about ten percent, or100 men, will experience a "heart attack" within thenext ten years. At least a quarter of these manifestas sudden deaths which occur before the patientreaches the hospital.5' 6 A minimum of 25 deaths,more likely closer to 30, is a conservative figure fordeaths within an hour. The relative frequency ofsudden deaths is somewhat lower among women

185

by guest on May 8, 2018

http://circ.ahajournals.org/D

ownloaded from

http://circ.ahajournals.org/

EPSTEIN

CORONARY HEART DISEASE AMONG MIDDLE-AGED MEN*

l00 Men

10 Years

Sudden Myocardial An inaDeaths Inforction Pectoris25 75 50 850

I 1othDeaths Survivors20 55 No Coronary Care Unit10 65 CCU

*Numbers represent approximations

Figure 1

Coronary heart disease among middle-aged men.

than men but the difference is not marked.5 Thenational problem of coronary heart disease hingesin large part on these persons who cannot possiblybenefit from any advance in acute coronary care.

While provisions of speedy mobile ambulance andcare units for patients who do not die instantaneous-ly is essential, it is still uncertain how many livescan be saved by such units and how they should beorganized most effectively.7Another 75 men will experience a myocardial

infarction over the ten-year period and reach thehospital. Despite recent reports from Britain thathome care might be safer for some of these patients,this study was designed in such a way that theconclusions must remain unsettled.8 Within a

month, from 25-30 percent (approximately 20) ofthe 75 patients will die, mostly within the first fewdays; this estimate tends to be on the low side ofreported hospital mortality before coronary care

units existed. In a coronary care unit, ten of the 20patients will probably be saved from these "electri-cal deaths." Most of the remaining ten persons diebecause of profound myocardial damage and wouldbe candidates for new hearts, artificial or trans-planted, to be discussed shortly. The prerequisitefor saving them would be a good temporary assistdevice-a top priority need.The 65 survivors of myocardial infarction and the

50 men who develop angina pectoris are about fivetimes more likely to die within the next five or tenyears than the 850 men without a coronary event.9The need for primary prevention is implicit in thisdiscussion. Nor is this merely a matter of life or

death because many survivors do not return to a

fully productive life.10 A systematic communityview of these people will provide a new under-

standing which cannot come from the ad hoc studyof individual patients.Angina pectoris, involving another 50 men, has

epidemiological features which distinguish it fromsurvivors of myocardial infarction despite thesimilarity in prognosis.'1 The accumulated experi-ence with angina patients from the several longitu-dinal epidemiological studies has only been partial-ly exploited. Recent work in Edinburgh on what theWHO group working on ischemic heart diseaseregisters has termed "unstable angina" is ofinterest.12 This refers to chest pain of recent onset,half-way between angina and myocardial infarc-tion, including "coronary insufficiency." All physi-cians in a defined area in Edinburgh refer suchpatients to a study center. Preliminary experiencesuggests that few die suddenly but a substantialnumber progress to frank myocardial infarction.

This observation raises a question of profoundimportance. How many of the 100 patients withheart attacks (fig. 1) have premonitory symptoms?Even more important: are any of these specificallypredictive of sudden death? Work is now proceed-ing on several fronts.'3 At least a quarter to a thirdof the patients who experience a heart attack haveseen a physician within the previous week, eitherbecause of cardiac symptoms or seemingly unrelat-ed complaints. This is clearly a matter for intensestudy, with a view toward professional and publiceducation.

Before leaving this deceptively simple chart, itmay be mentioned that a screening survey ofmiddle-aged men in the United States would yieldabout five men in a hundred with probable,manifest coronary heart disease; two of the fivewould have angina, the remaining three beingsurvivors from myocardial infarction.3 In about 20percent of the survivors, the attack will have been"silent" or "unrecognized," as evidenced by theappearance of significant Q waves not previouslypresent. The prognosis in such men is similar tomyocardial infarction survivors.'4 Depending on thecriteria of an abnormal exercise electrocardiogram,one other man could be added to the five as havingdemonstrable coronary heart disease by this test.15Are exercise electrocardiograms worthwhile inasymptomatic persons, as opposed to a diagnosticaid in patients with symptoms? Medicine used to bea matter of doing what seemed best. Now, thequestion of cost-benefit pervades much of theplanning. The matter of large-scale exercise electro-cardiogram testing is but a minor illustration of thisvery large problem.


186



ownloaded from


CORONARY HEART DISEASE EPIDEMIOLOGY

CIRCUMSTANCES of DEATH from CORONARY HEART DISEASE (CHD)

Combined Framingham and Tecumseh Experience

Total CHD Deaths (Men &Women)183

Died with other severe Died without ofher severeirreversible disease irreversible disease

-48 135

Died in Hospial Died outside Hospital50

With priorseverej Without priorheart disease severe heart disease

5* 45

Unexpected orless than I hour Otheafter admission

25 20*

With prior severe Without priorheart disease severe heart disease

51,11 80

rs *Cardiac ReplacementF Candidates

Condened from the Report by the Task Force on Cordiac Replacement, Na+ionol Heort Institute. October, 1969

Figure 2Circumstances of death from coronary heart disease.

Considerations of cost-benefit were a major factorin appointing a task force to assess cardiacreplacement, not only from the medical and ethicalbut also the economic point of view.16 Somefindings relate to some previous remarks (fig. 2).By 1968, there were 183 cardiac deaths under age65 in the combined experience of the Framinghamand Tecumseh studies. It is of interest that aquarter (48) died with concomitant, severe non-cardiac, life-limiting illnesses which alone wouldhave made them unsuitable as candidates forcardiac replacement. Of the remainder, 50 died inhospital; three had prior heart disease of suchseverity that they would not have been candidatesfor replacement while as many as 20 of the 50 couldonly have been saved by a temporary assist deviceprior to replacement. Five of the 85 persons whodied outside the hospital might have lived if theyhad been given a new heart earlier. This and otherinformation led to the estimate that there would bemaximally 200,000 candidates for new hearts peryear in the Un'ted States-given a temporary assistdevice, fully efficient nation-wide emergency trans-portation facilities and other prerequisites. Cost, assuch, did not turn out to be a decisive limitingfactor.

There are at least some 160,000 deaths, half ofthem sudden, from coronary heart disease in theUnited States per year under age 65.17 The vastCirculation, Volume XLVIII, July 1973

majority must be considered premature. Eighty-thousand sudden deaths per year among persons inthat age range represent a staggering figure-asd1fficult to fathom and comprehend in its totalimpact as some of the individual tragedies hiddenbehind it.

Predicting RiskIt was shown how many men in the general

population develop coronary heart disease in itsvarious manifestations. How can such men beidentified prior to the clinical event so thatpreventive measures might be instituted?Most is known about serum cholesterol, blood

pressure and smoking. When these three risk factorsare viewed in combination (table 1), two or allthree are in the upper range, as defined, among 38percent of the men-a rather sorry commentary onthe risk factor status of so-called healthy Americanmen! These 38 percent of the men generate 59percent, close to two-thirds, of all subsequentevents of heart attacks. The actual risk, in terms ofincidence rates, is almost 9 times higher when allthree risk factors are elevated, as compared withmen in whom all three are in the lower range. Thedata come from the Report of the Inter-SocietyCommission for Heart Disease Resources on thePrimary Prevention of the Atherosclerotic Diseases,4based on the National Cooperative Pooling Project,

187



ownloaded from


EPSTEIN

which provides the largest and longest availableexperience for prediction estimates and representsthe combined experience of the Framingham,Albany and Los Angeles studies, the two Chicagostudies and the original Minneapolis study.When it is the aim to assess risk in a single

individual, the most useful measure is incidencerate, which permits the calculation of risk, i.e., theprobability, of developing clinical disease associatedwith one set of characteristics as compared withanother. An approach which is more telling forassessing risk in the population is to ask to what ex-

tent a minority of the population generates a ma-

jority of the new events of disease. Such data whichdefine the proportion of the population exposed toexcessive risk and the proportion of new eventsoccurring in that high risk group, combined withincidence rates, provide the scientific basis forscreening programs and permit tentative estimatesof the impact of preventive measures on the diseasein the population. It is therefore worthwhile topursue further the question of the distribution ofrisk factors and their predictive power in thepopuilation.

Using the criteria and definitions shown in table1, it appears that slightly more than a third of thepopulation at risk, in terms of two or three factorsin the upper range, generates about two-thirds ofthe subsequent cases! The choice of other defini-tions for what constitutes an elevated level will, ofcourse, yield different results. For instance, if serumcholesterol and blood pressure are called "elevated"when they are in the upper third (tertile) of age-

specific frequency distribution, and smoking isagain defined as any amount of cigarettes, almost 25percent of the men pooled from seven studies willsmoke and will have both serum cholesterol andblood pressure in the top tertile; they will generatealmost 50 percent of the new events (unpublisheddata based on the experience of the five studiesquoted in table 1, combined with the RailroadWorkers study, also coordinated in Minneapolis,and the Tecumseh Study). In the Tecumseh Study,as another example, blood glucose one hour after a

standard load has been measured along with thethree risk factors already considered. Elevatedvalues are defined by the upper tertile of serum

cholesterol, blood pressure and blood glucose, or

smoking a pack or more a day. In this totalpopulation, 15 percent of the men aged 35-64 havetwo, three or all four of these risk factors in theupper range and they generate about one-third ofthe new events of myocardial infarction and suddendeath over a seven-year period (unpublisheddata).Could this prediction be enhanced by triglyceride

measurements? The Framingham group reportedthat triglyceride-rich lipoproteins do not add to thepredictive value of serum cholesterol among men'8

but such an independent effect is seen in the datafrom a prospective study in Sweden.19 TheFramingham and Swedish groups do not report on

a possible, concomitant influence of hyperglycemiawhile our group in Tecumseh, so far, has no

prospective data on serum triglycerides. However,there are such data from the Tecumseh Study based

Number of Incidence RateCombination Men Events Within 10 Years

None elevated 1249 28 20

Any one elevated 3320 171 48

Any two elevated 2178 198 90

All three elevated 595 38 % 82 59 % 171

Total 7342 479

Data from National Cooperative Pooling Project,American Heart Association

Table 1Prediction of Myocardial Infarction and Sudden Death According toThree Risk Factors:* Hypercholesterolemia > 250 mg/100 ml, Hyper-tension > 90 mm Hg diast., Cigarette Smoking in Any Amount


188



ownloaded from



on prevalence.'t Only 15 percent of the men withmanifest coronary disease did not have one or moreof five risk factors in the upper tertile range, thefactors being cholesterol, triglyceride, glucose onehour after challenge, systolic blood pressure andsmoking a pack a day or more. These data bear on afrequent statement that physicians often see pa-tients with coronary heart disease who are "risk-factorless." This, of course, depends on one'sdefinition of the troublesome term "normal."According to the definitions used in the aboveexample from a general population, patients withcoronary disease devoid of risk factor elevations arequite uncommon. To be sure, the same statement istrue for men without manifest coronary disease-only 20 percent of Tecumseh being in the lowerrange for all five variables, confirming what wasalready said about risk factor status in Americanmen. The situation among American women is notmuch different.20

It may seem surprising that the foregoingdiscussion has not included obesity as a risk factor.Most epidemiological data indicate that obesitymakes only a small, independent contribution to therisk of myocardial infarction and sudden death ifother risk factors, notably blood pressure, areincluded in the prediction. Blood pressure andobesity are strongly associated and it would appearthat the ill-effects of obesity are largely mediated byits effect on blood pressure and other correlatedrisk factors-in particular, triglycerides, hypergly-cemia and, to some extent, serum cholesterol. Thus,obesity is an important contributor to risk only inassociation with other risk factors. Presumably, interms of mechanisms, what is involved is not onlycaloric excess but its source; Keys pointed out longago that, in some countries like Italy, obesity maybe common but coronary heart disease fairlyuncommon because the relatively low proportion offat in the diet maintains serum cholesterol in alower range. Moreover, the relation betweenphysical activity and coronary heart disease hasbeen deliberately omitted from discussion becauselimitations of space preclude doing justice to thisimportant field.There are other predictors of risk. A very simple

questionnaire on chest symptoms and the restingelectrocardiogram read according to the MinnesotaCode gives remarkable predictive power: symptomsor ECG findings were recorded among 16.4 percentof the men aged 35-59 and these generated over athird (37 percent) of the new coronary events.21Quite likely, indices of psycho-social stress, behaviorCirculation, Volume XLVIII, July 1973

patterns of personality, scores for physical activityand ratings for family history would add some fur-ther predictive power, taking into account that pre-dictiveness is not simply additive because of thecomplex biological and statistical interactions be-tween the many risk variables. Unfortunately, addi-tioa of variables not only puts more persons withevents into the numerator but also increases thedenominator. The need is for an approach whichwill concentrate a majority of the new eventsamong a minority of the population at risk. Multi-variate analysis methods attempt to achieve thisaim.Such a biostatistical technique, the multiple

logistic function, was developed by Cornfield andapplied to the Framingham data in 1967.22 Thismathematical function ranks persons at risk in termsof any number of variables and assigns to eachindividual a score which measures the probabilityof developing a heart attack within a specifiedperiod of time. Using the median value of adiscriminant score distribution for seven variablesas the cutting point, 67 of the 88 new events, orthree-quarters, occurred among half the populationabove the median. Just over half of the new eventsoccurred among the 30 percent of the men with thehighest scores. Keys and his colleagues havereported that the multiple logistic function may putas many as 49 percent of the new major coronaryevents among men aged 40-59 into 20 percent ofthe population at risk.23 As might be expected,multivariate risk functions refer only to groups fromwhich they were derived and cannot necessarily beapplied to different populations in the same or othercountries. Mathematical risk functions of this sortalso have much potential application in coronaryrisk factor detection programs and multiphasichealth testing systems.

In the foreseeable future, risk functions can mostlikely be derived placing half of the new events into20 per cent of the population at risk. This may beillustrated in a diagram (fig. 3) where 50 (15 + 35)of the 100 "new events" (see fig. 1) are seen tooccur among the 200 men at the highest risk, out ofa total of a thousand men. Ideally, one would wishto put as many as perhaps three-quarters of the newevents into the high-risk, upper 20 per cent of thepopulation, but this is hardly realistic unless itbecomes possible to measure processes which areconnected more closely with the pathologicalprocesses in the coronary arteries and the myocardi-um than are currently recognized risk predictors.

189



ownloaded from


EPSTEIN

CORONARY HEART DISEASE and RISK FACTORS:

HYPOTHETiCAL MODEL! for MIDDLE AGED MEN

14000 Men

SuddenDeafhs

15

HGHRISK200

10 Years

-iMnardiona Neither35 150

NOT HGHRISK80Q

10 Years

Sudden M ocardia NeherDe10hs 4070acfion10 40 750

'See textFigure 3

Coronary heart disease and risk factors: hypothetical modelfor middle-aged men.

This model which, however, is realistic assumes thata major coronary event occurs within ten yearsamong 25 percent (50 out of 200) of the men athigh risk, while the risk among the remaining 800men is four times less. Sudden deaths are alsoshown to be more common among high risk men, assuggested by unpublished data from the TecumsehStudy.A good many experts advocate limiting preven-

tive action-at the present state of knowledge-tohigh risk individuals, because they consider scientif-ic evidence still insufficient to warrant a preventiveeffort directed at the entire population. If half ofthe events in the top risk group could be preventedby such limited action-and there is no good reasonto believe that the effect would be greater, short ofstarting prevention in childhood-there will now be75 instead of a hundred major events. Such a 25percent reduction, while not negligible, would stillnot substantially reduce the national burden ofatherosclerosis and its major complication.

Potential for PreventionThe burning issue of coronary heart disease

prevention is intimately tied to the question ofcausality. Is the relation between serum cholesterol,blood pressure, smoking, carbohydrate intolerance,and perhaps others a matter of cause-and-effect?24 25 At present, the evidence is onlycircumstantial but, at least for the first three risk

factors mentioned, very persuasive and strong. Theprovision of direct, rather than circumstantial, evi-dence requires controlled, preventive trials whichFredrickson has called "The Indispensable Or-deal."26 Yet, ever since the conclusion of the Diet-Heart Feasibility Study, reported in 1968,27 therehas been protracted and at times heated discussionover its design and even whether the indispensableordeal is, in fact, indispensable. Meanwhile severalEuropean groups are decisively forging ahead withpreventive trials.28 All the argument and hesitationin this country which has been the pioneer in thefield is not entIrely beyond comprehension. Trials ofthis sort present formidable logistic and technicalproblems, quite apart from their cost. Fortunately,a new phase has now been entered. The Task Forceon Atherosclerosis29 has strongly backed the criticaland crucial need for preventive trials, building onearlier recommendations,24 27 and contracts forcollaborative intervention studies are now beingawarded.30There are two distinct, though related issues. The

first (fig. 4) concerns the preventive approachwhich the individual physician should take towardhis individual patient or, rather, pre-patient beforehe becomes a patient. The idealized curve in thefigure depicts the hypothetical frequency distribu-tion of risk factor levels, defined as shown on theabscissa, in the population at risk. In reality, thesefrequency curves, as in the case of serum cholesterol


190



ownloaded from



CORONARY HEART DISEASE PREVENTION: Clinical-individual Approach

FREQUENCY OFRISK FACTOR LEVEL PREVENTIVE ACTION:

or Others

Figure 4

Coronary heart disease prevention:proach.

clinical-individual ap-

or blood pressure, are more or less skewed to theright. There will be levels where there would belittle argument that preventive action is essential,such as a serum cholesterol level of 300 mg/100 ml.However, what is the lowest level above whichprophylactic treatment becomes essential? Thebroad, vertical bar, in lieu of a line, is intended toindicate that there is no sharp line of demarcation.To some extent, the decision is clinical. For theremainder, it depends on what one considers highrisk and to what extent one desires to reduce it,based on the epidemiological information presentedearlier. The important point is to realize that thereis a risk factor level above which failure to institutepreventive action might be considered more seriousthan the risk of taking an action for which there isas yet no scientific basis in terms of direct evidencefrom a controlled, preventive trial. In the middle-range, a decision on the desirability of actiondepends on one's views on the strength of thecircumstantial evidence in favor of causality and, toput the matter perhaps too bluntly, on how longone thinks, in terms of probabilities, a person

should be permitted to live without experiencing a

heart attack.In the low range of risk, the decision on

preventive action is considered "conditional." Themain issue here is not the adult but his children.There is no serious doubt that true primaryprevention of atherosclerosis must start in child-hood. Before telling a man who has maintained a

serum cholesterol level of, say, 195 mg/100 ml atage 35 in the face of eating one or two eggs a dayand much saturated fat, that he has lucky genes andcan go on eating without bothering, one shouldCirculation, Volume XLVIII, July 1973

know about his children's levels. Data from thestudies in Tecumseh, reported in 1965 and 196631 32and recently confirmed (unpublished data), show arather striking relationship between parents' andchildren's risk factor levels; this leaves open thequestion to what degree the correlation is due toenvironmental and genetic factors. While therelationships between children and their parents aremarked, it cannot be assumed-returning to the"conditional" range in the figure-that a parentwith a low level necessarily has a child whose levelis or will remain low.The second issue relating to preventive action

concerns the community approach (fig. 5). The aimis to shift the whole risk factor distribution curve tothe left so that the total population be exposed to alower level of risk. This would require the adoptionof preventive measures by the entire population-and at an early age. Here the matter of modifyingnational eating habits to lower serum lipids is, ofcourse, the main area for controversy. It is askedwhether such far-reaching changes are justifiedwithout supportive evidence from controlled, pre-ventive trials33' 34 or whether the circumstantialevidence and the results from trials alreadyconducted,35 despite all their shortcomings,24' 25warrant preventive action on the community levelnow. 4, 36What is the solution? For the practicing physi-

cian, the matter is relatively simple. He will beguided by the location of the pre-patient within thecontinuum "prevention essential, desirable or condi-tional." The practical details of prophylactic treat-ment are not being considered here; the focus ofthis discussion is on "who" and "why" rather than"how." Iatrogenic disease is a problem in curativemedicine. Iatrogenic anxiety is an even greaterpotential problem in preventive medicine. A person

CORONARY HEART DISEASE PREVENTION: Community ApproachSHIFTING DISTRIBUTION CURVEFREQUENCY OF - IN THE POPULATION

RISKFACTORLEVEL ACTR LVE

RISK FACTOR LEVEL

Figure 5Coronary heart disease prevention: community approach.

191

~AIVIX,



ownloaded from


EPSTEIN

in the "prevention essential" category must bemotivated most strongly but, at any level or risk,instilling fear is probably not only a poor, but an

ineffective motivating force. Those in need ofprevention must be told with sympathetic care that,though not a certainty, their chances of developingheart trouble are greater than if their risk factorlevels were lower, that it would be sensible,therefore, to do something about it. There is at leastas much need for the blessed physician inpreventive as in curative medicine!The foregoing remarks assume that a decision has

already been made whether, in a given person,

prophylactic measures are indicated. Some guide-lines toward this decision were provided in thediscussion related to figure 4. It would obviously behelpful to devise sharply defined criteria for whoshould, might and should not be treated. This isdifficult for two reasons. First, as already men-

tioned, there is no level for any of the single riskfactors under consideration where risk changesabruptly from "low" to "high"; in general, the lowerthe level, the less the risk. All the same, to give an

example, the tendency would be, dependingsomewhat on age, to let a diastolic blood pressure

of 90 mm Hg or below go untreated, to treatpressures of 100 or above, and to worry about thosehovering between 90 and 100, pending the resultsof preventive trials for "borderline pressures." Thedecision in this twilight zone brings up the seconddifficulty: if a man with a diastolic pressure of 96also had a serum cholesterol level of 280 mg/100 mlwhich, even after treatment, could still remainrather high, one would lean more toward antihyper-tensive therapy than at a concomitant serum

cholesterol of, say, 200 mg/100 ml. It is here thatthe total risk score from the multivariant function,which takes into account cumulative effects, as-

sumes practical value for the treatment of individ-uals; it is also a population screening device andresearch tool. Knowledge of this total risk facilitatesa decision on "treatment versus no treatment" withregard to the separate "ingredients," such as bloodpressure or serum lipids, from which the score isderived.On the community level of prevention, interrela-

tionships between the health professions, officialand voluntary health agencies, government, agricul-ture and industry complicate the matter. If one

disagrees with those who maintain that no action on

this level can be taken without firm, supportingscientific evidence from definitive preventive trials,one must also disagree with the opposite position,

that the evidence is strong enough to launch an all-out national campaign to change eating habits andother factors, taking such actions which, althoughconsidered preventive to the development ofcoronary artery disease, have not been so proved.Such a frontal attack could not fully succeed, in anycase, without supporting evidence from a decisivepreventive trial! The need is to work withagriculture and the food industry, to improvelabelling and advertising, to spread knowledge onpreventive cardiology among the public and theprofession and, last but not least, to establish pilotpreventive programs. These aims will be accom-plished more effectively in a less heated andpolarized atmosphere and this is already beginning.The above remarks particularly apply to advice andactions involving living habits like eating, smokingand exercise. There has been less controversyregarding drug prophylaxis where controlled trialsare also underway both here and abroad.There is an apparent paradox in advocating

preventive advice addressed to the population atlarge to change certain habits of living and, at thesame time, calling for scientific proof from preven-tive trials that such advice will be based onscientific fact. There is a justification for thisstrategy. As shown before, around half of the newevents of heart attacks occur among persons whoare not at top risk in terms of the three "major" riskfactors. Without buttressing evidence from preven-tive trials, there is currently no way of sufficientlymotivating people in this critical "middle-risk"category to change their habits or to takeprophylactic medication. Without the full weight ofevidence from preventive trials, there can never bean all-out national effort to prevent atherosclerosisand its consequences, starting in youth.

Prevention ProgramsPrevention programs primarily provide a service,

as opposed to being a vehicle for research. There isprobably agreement that such programs shouldpresently be on a pilot basis in selected populationsin order to learn the best ways to integrate theminto comprehensive health care systems. Earlierenthusiasm for coronary risk factor screening, inisolation or as part of multiphasic health testing, hasnow given way to a more sober assessment. Itwould be a gross disservice to tell people that theyare at high risk without teaching them preventivemeasures. Screening is but the first step in anintegrated preventive program. On a national scale,these are enormous undertakings in terms of


192



ownloaded from



organization, trained personnel and cost. At least 20percent of American men are at top risk towardcoronary disease and another 30 or 40 percent are atless risk although still excessive! This in itself arguesin favor of community prevention programs,supplemented as needed by the clinical-individualprevention approach. The total problem cannot besolved until new, comprehensive preventive andcurative health delivery systems are created. In themeantime, there should be carefully designed andcontrolled pilot programs, with emphasis not onlyon modifying risk factors, but also on testingdifferent approaches to motivating health behaviorand achieving optimal cost-effectiveness.

At present, a man age 35 stands a 30 percentchance of developing a myocardial infarction byage 65 if he is in the highest ten percent risk rangein terms of the multiple logistic function score; inthe lower 30 percent risk range, his chances are lessthan ten percent, still an appreciable figure!37 Forsudden death, the difference is similar but evensharper.3 To get the vast majority of the populationinto the low risk range presently enjoyed by only aminority is the ultimate target.

References1. STAMLER J: Lectures on Preventive Cardiology. New

York and London, Grune and Stratton, 19672. LOWN B, KLEIN MD, HERSHBERG PI: Coronary and

precoronary care. Amer J Med 46: 705, 19693. EPSTEIN FH: The epidemiology of coronary heart

disease; a review. J Chronic Dis 18: 735, 19654. THE PRIMARY PREVENTION OF THE ATHEROSCLEROTIC

DISEASES: Position Paper for the IntersocietyCommission for Heart Disease Resources Prepared bythe Atherosclerosis Study Group (STAMLER J,Chairman) and the Epidemiology Study Group(LILIENFELD AM, Chairman). Circulation 52: A55,1970

5. GORDON T, KANNEL WB: Premature mortality fromcoronary heart disease-the Framingham Study.JAMA 215: 1617, 1971

6. KULLER L, COOPER M, PERPER J: Epidemiology ofsudden death. Arch Int Med 129: 714, 1972

7. LoWN B, VASSAUX C, HOOD WB JR, FAKIRO AM,KAPLINSKY E, ROBERGE G: Unresolved problems incoronary care. Amer J Cardiology 20: 494, 1967

8. MATHER HG, PEARSON NG, READ KLQ, SHAw DB,STEED GR, THORNE MG, JONES S, GUERRIER CJ,ERAUT CD, MCHUGH PM, CHOWDHURY NR, JAFARYMH, WALLACE TJ: Acute myocardial infarction:Home and hospital treatment. Brit Med J 3: 334,1971

9. THE FRAMINCHAM STUDY: An Epidemiological Study ofCardiovascular Disease (KANNEL WB, GORDON Teds.). Section 25: Survival Following Certain Cardio-vascular Events. Washington DC, USGPO, 1970


10. NAGLE R, GANCOLA R, PIcToN-ROBINSON I: Factorsinfluencing return to work after myocardial infarc-tion. Lancet 2: 454, 1971

11. KANNEL WB, FEINLEIB M: Natural history of anginapectoris in the Framingham Study. Amer J Cardiolo-gy 29: 154, 1972

12. FULTON M, DUNCAN B, LUTZ W, MORRISON SL,DONALD KW, KERR F, KIRBY BJ, JULIAN D, OLIVERMF: Natural history of unstable angina. Lancet 1:860, 1972

13. THE PRODROMAL SYMPTOMS OF MYOCARDIAL INFARC-TION AND SUDDEN DEATH. Report on a WorkingGroup Regional Office for Europe. Copenhagen,World Health Organization, March 1-3, 1971 (EURO8204-3)

14. KANNEL WB, McNATMARA PM, FEINLEIB M, DAWBERTR: The unrecognized myocardial infarction. Four-teen-year follow-up study in the Framingham Study.Geriatrics 25: 75, 1970

15. DOYLE JT, KINCH SH: The prognosis of an abnormalelectrocardiographic stress test. Circulation 41: 545,1970

16. CARDIAC REPLACEMENT: Medical, Ethical, Psychologi-cal and Economic Implications. A Report by an adhoc Task Force on Cardiac Replacement. NationalHeart Institute, National Institutes of Health, U. S.Public Health Service, October 1969

17. MORIYAMA IM, KRUEGER DE, STAMLER J: Cardiovascu-lar Diseases in the United States. Cambridge,Harvard University Press, 1971

18. KANNEL WB, CASTELLI WP, GORDON T, McNAMARAPM: Serum cholesterol, lipoproteins, and the risk ofcoronary heart disease-the Framingham Study. AnnInt Med 74: 1, 1971

19. KANNEL WB, McNAMARA PM: Serum lipid fractionsand risk of coronary heart disease- the FraminghamStudy. Minnesota Med 52: 1225, 1969

20. EPSTEIN FH: Epidemiology of Cardiovascular Disease:Respective Role of Neural and Non-Neural Factors.In Symposium on Neural and Psychological Mecha-nisms in Cardiovascular Disease (ZANCHETTI A, ed.).Milan, International Society of Cardiology and I1Ponte, 1972 (In press)

21. RoSE G: Predicting coronary heart disease from minorsymptoms and electrocardiographic findings. Brit JPrevent Soc Med 25: 94, 1971

22. TRUETT J, CORNFIELD J, KANNEL W: A multivariateanalysis of the risk of coronary heart disease inFramingham. j Chron Dis 20: 511, 1967

23. KEYS A, ARAVANIS C, BLACKBURN H, VANBUCHEM FSP,BUZINA R, DJORDJEVIC BS, FIDANZA F, KARVONENMJ, MENOTTI A, PUDDU V, TAYLOR HL: Probabilityof middle-aged men developing coronary heartdisease in five years. Circulation 45: 815, 1972

24. MASs FIELD TRIALS OF THE DIET-HEART QUESTION:Their Significance, Timeliness, Feasibility and Ap-plicability. American Heart Association MonographNo. 28. New York, American Heart Association, Inc,1969

25. CORNFIELD J, MITCHELL S: Possible effects on coronaryheart disease of intervention on selected risk factors.AMA Arch Environmental Health 19: 382, 1969

193



ownloaded from


EPSTEIN

26. FREDRICKSON DS: The field trial: Some thoughts on theindispensable ordeal. Bull N Y Acad Med 44: 985,1968

27. THE NATIONAL DIET-HEART STUDY: Final Report.Circulation 37 (suppl I), March 1968

28. REPORT OF A WHO WORKING GROUP. Methodology ofMultifactor Preventive Trials in Ischemic HeartDisease. Rome, November 1970. EURO 5011 (3)Regional Office for Europe. WHO, Copenhagen,1971

29. ARTERIOScLERosIS: Report by National Heart andLung Institute Task Force on Arteriosclerosis.National Institutes of Health, June 1971 (DHEWPublication No. NIH 72-137).

30. COOPER T: Arteriosclerosis, policy, polity, and parity.Circulation 45: 433, 1972

31. JOHNSON BC, EPSTEIN FH, KJELSBERG MO: Distribu-tions and familial studies of blood pressure and serumcholesterol levels in a total community-Tecumseh,Michigan. J Chron Dis 18: 147, 1965

32. DEUTSCHER S, KJELSBERG MO, EPSTEIN FH: Relation-ships between toxemia of pregnancy and essentialhypertension in a total community. Am J Epid 85:72, 1967

33. FREDRICKSON DS: Mutants, hyperlipoproteinemia andcoronary artery disease. Brit Med J 2: 187, 1971

34. MEADE TW, CHAKRABARTI R: Arterial-disease research:Observation or intervention? Lancet 2: 913, 1972

35. STAMLER J: Acute myocardial infarction-progress inprimary prevention. Brit Heart J 33: 145, 1971

36. STAMLER J, EPSTEIN FH: Coronary heart disease: Riskfactors as guides to preventive action. PreventiveMed 1: 27, 1972

37. DOYLE JT, KANNEL WB: Coronary Risk Factors: 10Year Findings in 7446 Americans. Pooling Project,Council on Epidemiology, American Heart Associa-tion. Presented at the VI World Congress ofCardiology, London, England, September 6-12,1970

Circuklion, Volume XLVIII, July 1973

194



ownloaded from


FREDERICK H. EPSTEINCoronary Heart Disease Epidemiology Revisited: Clinical and Community Aspects

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1973 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.48.1.185

1973;48:185-194Circulation.

http://circ.ahajournals.org/content/48/1/185Wide Web at:

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

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 in the

Permissions in the middle column of the Web page under Services. Further information about this process is availableOnce the online version of the published article for which permission is being requested is located, click Request

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Circulation Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:



ownloaded from

http://circ.ahajournals.org/content/48/1/185

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

http://www.lww.com/reprints

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


Coronary Heart Disease Epidemiology - Circulationcirc.ahajournals.org/content/circulationaha/48/1/185.full.pdf · Coronary Heart Disease Epidemiology Revisited Clinical and Community

Documents