A Longitudinal Study of Coronary Heart Disease
By OGLESBY PAUL, M.D.,
Downloaded from http://ahajournals.org by on March 17, 2024
MARK H. LEPPER, M.D., WILLIAm H. PHELAN, M.D.,
G. WESLEY DUPERTUIS, PH.D., ANNE MACMILLAN, B.Sc.,
HARLLEY MICKEAN, PH.D., AND HEEBOK PARK, M.S.
S INCE the Fall of 1957, a long-term study
of coronary heart disease has been in
progress at the Hawthornie Works of the
WTesterril Electric Company in Chicago under
the auspices of the IUniversity of Illinois Col-
lege of Medicine and Presbyterian-St. Luke 's
Hospital. The study was unidertaken in the
belief that coronary heart disease was a dis-
ease resulting from the interplay of multiple
factors and that there was ineed to delineate
these factors further. The data presented
herein represent the initial compilation of
data centering on this problemrl; a discrimi-
niate funietioln analysis is to be undertaken
iiext. The report which follows has been made
possible through the efforts of a large group
of physicians aiid other scientists who have
volunteered their time in the project to per-
init the accumulation of a large body of data,
a portion of whieh is presenited below.
Acknowledgment must also be made to the
W\estern Electric Company, which has beenl
nost cooperative and helpful throughout, to
its employees who have participated in the
study, and to the agencies and doniors par-
tially recognized below* who have finaneed
the survey since its inception.
Plan of Study
It was planned to study the eharacteristies
of a nlale population believed to be free from:i
clinieal coronary disease through annual inter-
views an-d examinations carried oni over a peri-
od of at least 3 years. The attributes of those
Fronm the Department of Medicine, University of
Illinois College of Medicilne, Chicago, Illinois.
*Gratitude is expressed to the Aniierican Heart
sociation, the Chicago Heart Associationl, the Illilnois
Heart Association, The United States Public Health
Service, The Otho S. Sprague Foundation, The Re-
search and Education Coiniiittee of The Presbyteriani-
St. Luke's Hospital, The Illini Founidation, and muany
private donors. Particular thainks are dlue to the gein-
erous help of Mrs. Tiffany ]Blake.
iien developing clinical coronary disease in
the course of the study were to be compared
to those men not so afflicted. The Hawthorne
Works of the Western Electric Conmpany was
selected both because of its reasonable prox-
imity to the Medical Center, and because of
the expressed interest of its Medical Depart-
rient in such a project. Over 20,000 persons
were einployed in the Hawthorne Works in
1937, imainly in clerical and light assembly
work, although a significant group was emn-
ployed at positionls requiring vigorous labor.
The group was comuposed chiefly of second-
anid third-generationi Americans of Polish and
Bohemian ancestry.
All mnen aged 40 to 55 who had been em-
ployed at the Works for 2 or muore years (N
3,397) were assigoned an idenitificationi nunmber
by a ranidomnizing process. Melubers of the
randomly selected group were then invited to
participate inl the study participation was
entirely voluntary-by beinig called from the
random list. In this process the medical rec-
ords of the Company were inspected and those
men for whom there was evidence of angina
pectoris by description or of myocardial in-
faretion confirmed by history anid electrocar-
diographic chaniges were excluded, as were
mnen in certain additioiial categories (deaths
subsequelnt to preparation of the list, trans-
fers, permanient disabilities, leaves of ab-
senee); this group numbered 127 men.
Invitations to participate were sent by letter
after preliminary discussion and explanation
at the top management and supervisor level,
As- and publicity through the plant bulletin and
company magazine. It was planned to invite
a sufficient number to provide a population
of approximately 2,000 men. Of the 3,102
men asked to participate, 2,080 or 67 per cent
of those iiivited agreed to participate and
wvere exanuilned the first year. Of this group,
20 Circulation, Volume XXVIII, July 1963
 CORONARY HEART DISEASE
44 were excluded from subsequent anialysis
because the first examiniation indicated the
presence of coronary heart disease. Forty-
seven others were not included in the analysis
because of special problems in follow-up un-
related to the eardiovascular system. The base
population therefore available at the end of
the first year for analysis amounted to 1,989
men. At the end of 4 years, 38 mnen had died
and 105 men had withdrawn from the project
or left the Company. Over 94 per cent of the
survivors of the original group had thus re-
mained in the study. The group labeled "non-
coronary " in subsequent references will be
the base population less men not followed,
the new coronary cases, and in certain in-
stances a small number of men for whom
adequate data were lacking. After the first
year the men were, in general, examined on
an anniversary-month basis, although in many
inistances, this was not possible due to conm-
mitments at work, travel, or illniess.
No therapeutic suggestions of any kinid
were given to the men participating in the
study, and the study itself was labeled a
Downloaded from http://ahajournals.org by on March 17, 2024
Health Survey to draw attention away fron
the cardiovascular system.
The diagnosis of angina pectoris was made
by two or more internists in each ease and
was based upon published criteria.1 The diag-
nosis of myocardial infaretioni was made with
use of the clinical history plus characteristic
electrocardiographic findings.1 Where differ-
ence of opinion occurred, decisioln as to diag-
nosis rested with the senior investigator, who
undertook an additional interview and phys-
ical examination. The assumption of death
from coronary disease was made on the basis
of the most reasonable conclusion to be drawn
from information obtained from the family,
physicians' and hospital records, death cer-
tificates, anid coroners' reports. Unfortu-
nately, postnmortem examinations were rarely
made.
A family history was provided by each par-
ticipant, who completed a form with the help
of his family. Annual additions to the history
were made by interview. The medical history
Circulation, Volume XXVIII, July 1963
21
and physical examination were modeled after
that employed in the Framinghanm Project.2
The dietary data were obtained by a 1-hour
interview with a trained dietitian.3 Supple-
mentary information was obtained from
forms completed by the wives, and by inter-
views with a sample of the wives. Plastic food
models were employed to assist in portion
estimates, as were data regarding food pur-
chases and preparationi obtainied from the
Company cafeteria.
Physical activity on the job was assessed
by reference to a Company eode, the validity
of which was confirmed for work in the shops
by oxygen studies on the job in a small sample
of 36 men. The code included five grades of
physical activity in the shops and four of en-
vironmental factors in the offices. Physical
activity off the job was estimated by use of a
questionnaire administered by a trained in-
terviewer fromn the personnel department of
the Company.
A routine 7-foot chest x-ray was taken, as
were an electrocardiogram with a direct-writ-
ing instrument with standard 12 leads plus
lead V4R, urinalysis, blood hemoglobin level
(with a colorimeter), and blood cholesterol
(Abell-Kendall method). Body fatness was
measured with a skin-fold caliper. Somato-
typing was performed by standard measure-
ments plus photographs in three views of each
man clothed in a brief supporter, with typing
by the Sheldon method.4 ljipoprotein lipase
measuremenlts were made the third year by
the technic of Pilgeram.5
Blood pressures were taken at the begin-
ning of each physical examination with the
participan-t seated; left and right arm meas-
urements were found to be similar on statis-
tical analysis and the left arm readings are
referred to below.
Psychologic studies were also made and will
be reported separately.
For the purpose of this report, the diagno-
sis of angina pectoris is given wheni a man
exhibited only this manifestation of coronary
disease in the absence of aortic stenosis or
other obvious underlying factors. The diag-
 22
nosis of myocardial infaretion is given when
myocardial infarction was adequately docu-
rnented whether or not angina had been pres-
ent on previous examinations (this occurred
in onie of 28 infarets). The diagnosis of death
from coronary disease is listed regardless of
previous documentation of angina pectoris or
myocardial infaretion, providing adequate
evidenee for such listingf was available. A
history of sudden death was not considered
adequate evidenee by itself.
In the data presented below, differences
between the groups are investigated with re-
spect to data obtained from the first exam-
inationl unless otherwise stated.
Results
New Coronary Cases
After 4 years anid 3 months of the study,
88 meii had beeii elassified as having devel-
oped clinieal coronary heart disease as fol-
lows: angina pectoris, 47 men; myocardial
infaretion, 28 men; deaths from coronary
disease, 13 men. Approximately one new case
per 100 men per year was thus diagnosed.
Downloaded from http://ahajournals.org by on March 17, 2024
These coronary men had a mean age 6 months
older than the men in the noneoronary group,
with a distribution curve skewed somewhat
to the right.
The group under study was probably not
more or less coronary prone than the men who
were invited but did not accept or respond.
This is shown by the fact that at the end of
4 years and 5 months, 21 of the 1,022 who
declined or did not respond had died from
all causes and 45 of the 2,080 who had ac-
cepted had died from all causes. The mor-
tality percentages are 2.1 and 2.2 per cent
Table 1
History of "Noncardiac" Chest Discomfort
Noncoronary cases Coronary cases
(1764) (87)
Yes 16 31a/c
Co
No 84%l 69%c
Data relating to history of "'noneardiac " chest
discomfort as obtained at the time of the initial
exasninationi anid its relation to subsequent recogni-
tiosi of clinical coronary heart disease.
PAUL ET AlI.
respectively. The total mortality for the 2,168
nmen in the group of 5,397 whose numnber
never came up for consideration was 3.3 per
cent, which compares with a total mortality
of 3.1 per cent for the 3,102 invitees plus the
127 men referred to previously who were con-
sidered but not called.
Historical Data
Family History
There was no significant difference betweeni
the coronary and noneoronary groups when
analyzed for the historical evidence of both
parents dying over 70 years of age, over 65
years of age, or dying over 60 years of age.
The ages of living parents were adjusted to
expected age at death aceordiing to the Ameri-
canl Experienee Mortality Table. There was
ino significant differenee between the average
age of death of the mother of these two
groups, but there was with the age of the
father (the age of the fathers in the noneor-
onary group exceeded that of the coronary
group by 3.4 years, on the average). Infor-
mation on the grandparents' age at death did
not appear to be a factor associated with the
incidence of coronary disease.
History of Chest Discomfort
It was of considerable interest to follow the
men who at the time of the first examinationl
gave a history of soime type of chest discom-
fort that was not considered adequate for a
diagnosis of angina pectoris. Table 1 shows
a significantly higher development of clinical
coronary disease in those men who gave such
a history, as contrasted with those who did
not give such a complaint (p < 0.001).
History of Cough and Shortness of Breath
A chronic cough was stated to be present
by 25 per cent of the men in the coronary
group (88 cases) as contrasted with 12 per
cent of the noneoronary men (1,778 cases).
Likewise, 18 per cent of the men in the coro-
nary group indicated shortness of breath on
effort as compared with 11 per cent of the
men in the noneoronary category. Both of
these differences are significant (p < 0.001
for the former and < 0.025 for the latter).
Circulation, Volume XXVIII, July 196.q
 CORONARY HEART DISEASE
Table 2
Somatotype Distribution in Noncoronary and Coronary Groups
Category
Eiido. dominanit
Meso. dominant (endo. > ecto.)
Meso. dominant (enido. = ecto.)
Meso. dominant (endo. < ecto.)
Eiido. = Meso.
Meso. = Ecto.
Ecto. dominiant
Somatotype distribution in the noneoronary and coronary groups, listing observed
and expected numbers. In the mesomorphic dominant category are listed the secondary
dominant or equal components.
History of Peptic Ulcer and Gallbladder Disease
Ten per cent of 1,794 noneoronary men
gave a history the first year of peptic ulcer
as compared with 18 per cent of the 88 coro-
nary cases; this difference is significant (p <
0.025). There was no significant relationship
with a history of gallbladder disease.
Weight Gain
The pattern of weight gain and loss from
Downloaded from http://ahajournals.org by on March 17, 2024
age 20 to age 40 was not appreciably differenlt
in the two groups.
Physical Characteristics and Observations
Height and Weight
The mean height and weight for the nlon-
coronary population (1,776 men) were 68.7
inches and 171.1 pounds, and for the coronary
group (88 men) 68.4 inches anld 172.0 pounds.
These differences are not statistically signifi-
cant. The correlation coefficients for height
versus weight in the two groups also were not
significantly different.
Body Fatness
A significant differenee in skin-fold thick-
ness for the same two groups was demon-
strated: the mean triceps skin-fold thickness
measured 1.39 em. for the 1,776 noneoronary
men and 1.52 cm. for the coronary group, and
the mean scapular thicknesses were 1.81 cm.
and 1.98 cm., respectively. The difference be-
tween the two groups is signifieant at the
0.025 level for the triceps mneasurement and
at the 0.01 level for the scapular thickness.
Circulation, Volume XXVIII, July 1963
2M3
Noncoronary cases Coronary cases
(1790) (87)
Obs./Exp. Obs./Exp.
253/259 19/13
742/743 37/36
156/153 4/7
124/120 2/6
217/221 15/11
101/99 3/5
197/195 7/10
Somatotype Data
Somatotype of individuals was scored by
the Sheldon technic. The subjects fell lnatu-
rally into three groups: (1) endomorphic
dominance, i.e., the endomorphic component
exceeded the other two components, (2) ecto-
morphic dominance, and (3) all those whose
mesomorphic component was at least as high
as either of the other components. The last
category is by far the largest. It was decided
that a closer scrutiny of the mesomorphic
dominant group was desirable. Accordingly,
the mesomorphic group was further parti-
tioned into groups defined by the comparative
sizes of the other components. The groups
actually considered are clearly defined inl
table 2.
It has been reported that mesomorphic dom-
inance in body build is positively related to
the presence of clinical coronary disease."
Our material does not support such a conclu-
sion, and instead suggests that endomorphic
dominance may be important, in that there
is an excess of coronary cases in the group
characterized by endomorphic dominance (p
< 0.1). A larger series will be needed to con-
firm this latter point.
Pulse Rate
The meau pulse rate of the coronary group
(80.7) was not significantly higher than that
of the conitrol group (78.5). The distribution
of pulse rate in each group was miiarkedly dif-
ferent, however, with a p value of 0.025
 24 PAUL ET AL.
Table 3 a survey limited to a single season may give
Pulse Rates in Noncoronary and Coronary Group different results from one in which the data
Noncoronary cases Coronary cases
were obtained at another time of vear (fig. 1).
Rate (1766) (88)
Ophthalmologic Observations
< 69 166 c 18WZ No relation between the presencee of areus
70-79 30'S
80-89 31%s
3 51,%c
senilis and corollarv heart disease was noted.
90-99 166 C
There were not enough examples of xaiithe-
II
C/ lasma to permit statistical analysis. It was
> 100 %<
1 1%
somewhat surprising to find that a definite
statistical significance was present between
Table 4
the observation of arteriovenous nicking as
Mean Blood Pressure in Noncoronary and Coro-
nairy Groups (Left Arm Readings) noted the first year and the subsequent devel-
opment of coronary disease. In 1,793 noneoro-
Noncoronary cases
(1788)
Coronary cases
(87)
nary meii and 86 coronary cases in whonm data
135 146
were available, the pereentage having this
87 92 finding weas 4 per cent and 13 per cent re-
spectively, a differencee which is significant at
the 0.005 level.
(table 3). In particular, the coroniary group Presence of " Tension"
appeared to have a higher percentage of ex- In view of the preoccupation of some ihives-
treme cases. tigators with impressions of "tension" or
Blood Pressure tenseniess in coronary or supposed coronary-
Table 4 lists the data relating to blood pres- prone individuals, it is worth while to record
sure recordings in the coronary aiid the non- that the internists making these examinations
coronary categories. As reported by others, a
Downloaded from http://ahajournals.org by on March 17, 2024

wrere asked to cheek at the end of the exam-

positive relationship is apparent between ele-
vations of the systolic level and the develop-
ment of coronary disease (p < 0.001) and
also with the diastolic level (p < 0.03). The
mean systolic blood pressure of the coronary
cases was 146 as compared with 135 for the
noneoronary population, and comparable fig-
iniationi (whiclh occupied approximatelv 45
minutes) wlhether they conisidered the mian
examiined to be "relaxed'" or "teiise.i' No
significant differeniee was seen but there was
a slight excess of meni classified as "relaxed"
8 7 -

ures for the diastolic level were 92 and 87.

8 6- ,_ I~ ~ ~ ~ ~iST YR.
Isolated systolic blood pressure elevation bv 8
6

-~~~~~~~~4h R

itself was uncommon. The meani diastolic level 8 5

\
_ <
/P_~~~~~4TH
.', AV. IN SEASON
YR.

for the angina group was 92, for the infaret N 2

2 ND YR.
group 90, and for the men who died from
coronary disease 94. Sixteen per cenit of the
83- '. , .e3 RD YR.
noneoronary group had a diastolic level of N
'. -
100 or over; this rose to 28 per cent for the I
8 2-
angina group, 29 per cent for the infaret
group, and 33 per cent for those who died 81 -

from coronary disease. A discussion of the

WINTER SUMMER
blood pressure data in relation to the family
SPRING AUTUMN
history has been presented elsewhere.7 It is Figure 1
well to remember that in any such population- Pattern- of seasonal blood pressure change (dias-
study, fluctuations of the blood pressure lev- tolic pressure only) as observed over 4 yeacrs. The
els with the season of the year occur and that systolic pattern is similar.
Circulation, Volume XXVIII, July 1963
 CORONARY HEART DISEASE
who subsequently developed clinical coronary
disease.
Laboratory Observations
Hemoglobin Levels
No significant difference was shown among
the noneoronary, angina, and infaret groups
in mean hemoglobin levels although the trend
was in the direction noted by others. The
mean hemoglobin level for the noneoronary
men was 14.8 Gm. per cent, for the angina
cases 15.0 Gm. per cent, and for the infaret
cases 15.2 Gm. per cenlt.
Cholesterol Levels
As anticipated, a definite relationship was
demonstrated between higher blood choles-
terol levels as found the first year, and sub-
sequent incidenee of clinical coronary heart
disease (table 5). This relationship is highly
significant (p < 0.01). The mean cholesterol
level for the noneoroniary population was
quite high, 247 meg. per cent, and for the eor-
onary groups was 272 mg. per cent. Despite
this mean difference, almost 30 per cent of
the coronary cases had blood cholesterol levels
Downloaded from http://ahajournals.org by on March 17, 2024
below 240 mg. per cent.
A second and perhaps more important ob-
servation was the demnonstration of a signifi-
can-t difference in behavior of the coronary
subgroups. A comparison of cholesterol levels
in the noneoronary cases with the levels of
47 angina cases, the 28 infaret cases and 13
menl who died of coronary disease, is indicated
in table 6. The stepwise progression of the
mean cholesterol levels is clearly seen, fron
the lowest levels in the men as vet not iden-
tified as having coronary disease to the high-
Table 5
Cholesterol Levels in Noncoronary and Coronary
Groups
Noncoronary cases Coronary
Mg. % (1794) (88)
100-204 20C 7%
205-239 2 9 thf
240-256 15%G 15cc
257-274 11%o 17%
275-309 14%l 22%
> 310 12Cc 18%Co
Circulation, Volume XXVIII, July 1963
25
260-
250-
--,0958
240- j'\ \ , .
1 - 9
' AV/ERAG E
2 30-
:, I#
//
"
220-
WINTER SUMMER
SPRING AUTUMN
Figure 2
Seasonal variation it serumn cholesterol lev,els a(s
observed over^ 4 years. It will be noted that oul,
one season is inconzsistent wcith the trend (Summer
1959). The reason for this is unknown.
est level in those who have died from coronarv
disease.
A third observationi in the cholesterol de-
terminations supports the work of Thomas et
al.8 These authors founld a seasonal variation
in cholesterol levels in a small group of pris-
oilers who were studied throughout the year.
As will be seen in figure 2, our data show a
consistent seasona[ variationi, the highest levels
beinig encountered in the winter with the
lowest in the springfo or sum-mer. In this
study, cholesterol determinations were not
made on each man for each season; rather,
the means for each seasoni were determined
with the data procured from annual examina-
tions. The men were often but not invariabl-
examined in the same seasoni each year. That
the variatioll is not identical with the seasoni
variation in blood pressure recordings can be
cases seen by comparisoni with figure 1. Our dieti-
tians have also analyzed seasonal dietary
changes and find that in this population, they
are apparently insignificant although more
beer is consumed in the months with higher
temperatures and somewhat more fruit is in-
cluded in the diet. There is no change in the
 26
Cholesterol Levels in Noncoron-ary asnd Coronary
Noncoronary cases
(1794)
Mean 247.3
K 256 mIg. %C 63Xc-
> 256 mng. Cc 37'%o
hours of work or type of work related to the
seasons although we have determined that
men are more physically active in the sumnler
than the winter. The magnitude of the varia-
tionis present is considerable and inmportant in
any study in which attempts to manipulate
the blood cholesterol levels are made.
Lipoprotein Lipase Measurements
During the third year anld during the first
half of the fourth year, lipoprotein lipase de-
terminations were made on the whole group.
ILipoprotein lipase cofactor levels were also
m-iade. These studies were carried out through
the courtesy of Dr. John Olwin of the Coagu-
lation laboratory of Presbyterian-St. Luke's
Hospital. There was no evidenee of a rela-
Downloaded from http://ahajournals.org by on March 17, 2024
tionship between the presence of clinical cor-
oniary disease and levels of either lipoprotein
lipase or lipoprotein lipase cofactor.
Blood Sugar Determinations
During the third year of the project, each
man was asked to drink a solution containing
100 Gm. of glucose, and a blood sample for
glucose determination was drawn approxi-
mately 2 hours later. The mean blood-glucose
levels showed no significant differences be-
tween the noneoronary and coronary groups
(mean noneoronary blood-glucose level 106.7
mg. per cent, mnean coronary level 106.8 lug.
per cent). Of interest was a significant het-
erogeneous variation within each of the four
coronary groups. The angina group was the
most variable, followed by the noncoronarv
anid infaret groups, respectively. The deceased
group appeared to be the least variable, but
the estimated standard deviation is based
only five observations. The differenee in vari-
ability between the angina and infarct group
was particularly striking.
PAULI ET AL.
Table 6
Subygroups
Angina Infarct Deceased
(47) (28) (13)
263.6 7V
'- _284.5
48%'c 43( c 277%
3
5 2 Vc 57cc 7 7 -'I
Electrocardiograph Findings
A limited survey of the electrocardiographi(
observations has shown as yet iio significant
association between the presence of -ventrieu-
lar prenmature beats or the finding of a coini-
plete left or right bundle-branch block with
the development of eliiiical coronary disease.
The numbers in these categories are as yet
small. However, there was an association
found between all types of ST-segment or T-
wave abnormalities (in the absence of a coni-
plete bundle-branch block or QRS abnormali-
ties consistent with myocardial infaretion).
Thus 10 (11 per cent) of the 88 electrocar-
diograms of the nlen subsequently identified
as coronary cases showed these changes on the
first electrocardiographic tracinig, whereas
only 49 (2.7 per cent) of the 1,797 electrocar-
diograms in the as yet noneoronary group
revealed this type of abnormality. This is
highly significant (p < 0.005).
Environmental Factors
Place in, the Company
There was no indication that current type
of work in the Company was a factor predis-
posiiig to the development of coronary disease
(table 7). A study of the prior work history
likewise failed to demonstrate any significant
differences.
Physical Activity off the Job
Physical activity off the job was estimated
from the answers to series of questions posed
to each mnan by a trained interviewer. These
questions included reference to blocks walked
to and from work, flights of stairs climbed,
on hours devoted to household chores and yard
work, and past and present participation in
sports. No striking differences were encoun-
tered between the coronary and noneoronarv
Circulation, Volume XXVIII, July 1963
 CORONARY HEART DISEASE 27

groups, although, in general, men in the for- Table 8

mer category seemed to spend more time in Cigarettes Smoked "Most of Adult Life"
household and yard work and fewer hours a Noncoronary Coronary
week in sports than did men in the latter cat- Number of
cigarettes per day
cases
(1786)
cases
(87)
egory. Thus 38.6 per cent of the coronary
Nonsmoker 33% 23%
men denied participation in bowling, golf, 1/4 Pack ( 1- 7) 7C% 2%
baseball, hunting, or fishing as contrasted 1/2 Pack ( 8-12) 11% 9 %G
with 23.4 per cent of noncoronary men, and 3/4 Pack (13-17) 12% 6%o
when the coronary men did engage in these 1 Pack (18-22) 30% 47%
activities, they tended to do so for shorter 1-1/4 Pack (23-27) 27% 3%cl
periods. 1-1/2 Pack (28 +) 6% 9%

Tobacco
No difference was found between the groups
relative to the habit of cigar smoking, or to
the habit of pipe smoking. A highly signifi-
cant relation to the use of cigarettes was ap-
parent. The data secured in answer to the
question asked of the participants to indicate
the number of cigarettes smoked most of their
adult life are seen in table 8. The association
of cigarette smoking most of adult life with
the subsequent development of clinical coro-
nary disease is significant, with a p value of
0.005. It is also significant in regard to
current smoking habits, also with a p value
Downloaded from http://ahajournals.org by on March 17, 2024
was a relation between the number of cig,ar-
ettes smoked and the manifestation of coro-
nary disease identified (table 9), which rela-
tion was significant at the 0.005 level. The
stepwise trend analogous to that observed
with cholesterol levels is to be noted.
Diet
The assessment of diet3 was done by inter-
view as indicated above. It is recognized that
such a method yields at best only an approxi-
mation of food intake. In a population of this
size, metabolic balance studies were clearly

out of the question. Attempts at validationi

of 0.005. Further, it was noted that there of the interview technic were done as men-
tioned above; these showed an excellent cor-
Table 7 relation when a sample of the men (20) were
Job Classification interviewed by two dietitians on different
Noncoronary Coronary days. A comparison of the interview method
cases cases with records kept by the men (50 the first
(1797) (88)
Shop occupations year, 364 the third year) in a specially pre-
physical denmand code 1 .3 % pared 7-day food-intake diary revealed that
2 38.0 %s 29.5% the interview approach (which embodied a
3 15.6%Xo 17.0% 1-week meal pattern plus a 28-day recall)
4 4.0%C/ 5.7% gave higher levels for 21 nutrients than did
5 .2%o the diary kept by the men. It appears that
Office occupatioiis the interview tends to exaggerate to some ex-
work condition code 1 32.2% 39.8%o tent the quantities of food taken in, while the
9 8.2%o 6.8% diary underestimated quantities. The men
3 1.4% 1.1%o often conceded that they were embarrassed
4 .2% to list in some cases all that they really had
Job classification of noncoronary anid coronary consumed. Because of the extensive experi-
cases. Code 1 in the shop represents the least physical ence of our dietitians with these men, the in-
effort; code 5, the most strenuous and sustainied effort. formation obtained from the wives and other
Code 1 in the office represents the most ideal environ-
mental condlitions in. terms of noise, temperature sources to supplement the data, and the simi-
clhanige, etc.; code 4, the Imlost distuirbing type of larity of the information obtained on succes-
enivironmenit. sive years (being described in greater detail
Circulation, Volume XXVIII, July 1963
 28 PAUL ET AL.
Table 9
Cigarette Data in Noncoronary and Coronary Subgroup
Cigarettes adult life Cigarettes now
Noncoronary Noncoronary
cases Angina Infarct Deceased cases Angina Infarct Deceased
(1786) (47) (27) (13) (1789) (47) (28) (13)
Nonsmokers and smokers of
1/2 pack per day or less 51%o 40 cl 266% 31%o 58 % 51%l 36% 31%
Smokers of more than
1/2 pack per day 49%Xl 60% 74% 69% 426ro 4 9% 64% 69%

elsewhere), we believe the dietary data to be
a reasonable approximatioii of the truth.
Table 10 lists the dietary information for
the coronary and noncoronary groups. No
significant differences are apparent for these
nutrients.
In an attempt to isolate the high- and low-
fat groups from the point of view of risk, the
296 men eomprising the top 15 per cent of
the whole population in terms of calories from
fat were contrasted with 296 men represent-
ing the bottom 15 per cent in terms of fat
intake. It will be seen from table 10 that only
20 per cent of the fat calories came front veg-
Downloaded from http://ahajournals.org by on March 17, 2024
etable fat, so that this is in essenee a coni-
parison of high and low aniimal fat groups.
Table 11 gives the important findings in these
two populations. Worthv of comnment is the
fact that of the 88 coronary eases, 14 have
appeared in the high-fat intake group and
16 in the low-fat group.*
There was no demonstrable relation between
responses to questions regarding salt intake
in the coronary and the noncoronary groups.
The second year, the mnen were, in addition,
asked to taste three samples of soup contain-
ing no salt, a moderate amount of salt, and a
heavy amount of salt to indicate their prefer-

ence as to flavor. Here, too, no positive rela-

Table 10 tionship could be found.
Daily Average fo, Nutrients A positive relation was encounitered, lhow-
ever, between coffee intake and coronarv dis-
Noncoronary Coronary
cases cases ease. Because of subsequent discovery of
(1797) (88)
variations in portionl size for coffee intake
Total calories 3174 3082 during some of the first interviews o0llv the
Animal protein (Gin.) 90 90
26
data for a total of 1,108 of the men seen coil-
Vegetable protein (Gin.) 28

Animal fat (Gm.) 118 116 secutively during the last 9 months of the first
Vegetable fat (Gm.) 34 32 year are listed in table 12. For this group, a
Carbohydrate (Gm.) 304 299 significant correlation between the use of cof-
Calcium (Gin.) 1.08 1.10 fee and the later developmenit of coronary
Phosphorus (Gm.) 2.33 2.16
Iron (Mg.) 17.'2 16.6 disease is seen (p < 0.025).
Cholesterol (Mg.) 757 721
Total sat. fatty acids (Gnm.) 59 59
*We have attempted to assess the constanecy of diet
in these two groups. In the third year of the project,
Total unsat. fatty acids (Gm.) 83 80
a complete diet interview was performed iii a sample
Linoleic acid (Gm.) 12.28 11.57
.69 .66
conmposed of 60 men from each category. At this
Linolenic acid (Gm.) time, the samnple group initially characterized as
Arachidonic acid (Gin.) .98 .97
having a high-fat intake obtaiined 45 per cent of its
Vitanmin "Al" (I.U.) 12,470 11,750 calories fronm fat, ancd the sample group from the
Thiamine (Mg.) 1.66 1.55
original low-fat categor-y obtained 39 per cent of
Riboflavin (Mg.) 2.52 2.51 its calories fronm fat. The difference iil the fat
Niacin (Mg.) 22.8 22.2 intake of the tw o groups has thus lessened, a result
Vitamin "'C" (I.U.) 105 101 not unexpected in viewi- of the pheniomenon of regres-
Vitamin "'D"' (LU.) 198 196 sion ton-ard the mean upon repeated testing.
Circulation, Volume XXVIII, July 1963
 CORONARY HEART DISEASE 29

Table 11
Comparison of Groups wvith High and Low Fat Intake
High fat Low fat
(296 cases) (296 cases)
Median C% calories from fat 49.3 36.1
Average cholesterol (Mg. c) 253.8 246.8
Average height (inches) 69.0 68.4
Average body weight (lb.) 168.9 175.0
Average triceps skin fold (cm.) 1.4 1.4
Average scapular skin fold (cm.) 1.8 1.9
Average blood pressure (left arm) 133/86 139/90
Shop work 171 Cases-median 176 Cases-nmedian
grade 2.3 grade 2.4
Office work 121 Cases-median 120 Cases-median
grade 1.2 grade 1.2
New coronlary cases 14 16

No positive association was foulnd between
the development of coronary disease and the
ilntake of alcoholic beverages.
Discussion
The data presented clearly cainnot have too
broad an applieation. The population under
survey was limited to men in an age group
known to be susceptible to coroniary disease.
Further, the popula-tion is an industrial one
Downloaded from http://ahajournals.org by on March 17, 2024
ticular importance is the link between a his-
tory of any type of chest discomfort and the
later recognition of coronary disease. This
finding confirms what many elinicians have
observed. It should be commented that it is
probably more difficult to elicit a history of
classical symptoms of disease in men who
come for this type of screening examination
as compared with those who present them-
selves at the office of a physician proffering a

and in background relatively homogeneous

and stable. Also, as yet, only a moderate complaint. This also further points to the
niumber of individuals have been identified as vagaries of angina pectoris in its manifesta-
having clinieal eoronary disease. As we have tions. Agreement with the criteria employed
watehed the data unfold, however, there has in the diagnosis of angina pectoris in this
been a remarkable consistency to it, and a study was evidenced by a review of the first
trend once revealed has shown greater anid 30 of the angina cases by a team of three
greater statistical significance as new cases physicians (one from Chicago, one from Ala-
have been added. We have waited to this bama, an-d one from England).9 The impor-
point before presenting the material available tance of the association with a prior history
in order to try to have a body of data reason- of peptic ulcer is uncertain, but may reflect
ablv suitable for analysis. As one fragments certain of the problems to be discussed in the
the coronary group into subgroups, the num- companion paper on the psyeholog-ic studies
bers of course become smaller and conclusions completed on the group.
more tenuous.
The picture that has unfolded is clearly Table 12
that of a multifactorial disease. Positive re- Inttake of Coffee
lationships have been established between
Noncoronary cases Coronary cases
coronary disease and the presence of elevation Cups/month (1108) (54)
of systolic and diastolic blood pressure, eleva- 0- 49 17% 16%o
tions of the blood cholesterol levels, and heavy 50- 99 36% 28%o
use of cigarettes as reported by others. The 100-149 246% 15%C/o
association with abnormalities in the electro- 150-199 15% 2 2%
cardiogram was also not unexpected. Of par- > 200 7% 19%

Circulation, Volume XXVIII, July 1963

 30
It was particularly initerestinig to us to en-
counter the stepwise association of the blood
cholesterol level with the coronary sub-
groups.> To those who decry the significane
of blood cholesterol findings in populatioii
groups, this should be food for thought. This
evidenee helps to buttress the concept that
the patient with angina pectoris is not iden-
tical with the patient who develops a myocar-
dial infaret, or even who dies from the effects
of the underlying disorder. The limited ex-
tent to which this is meaningful in the indi-
vidual patient is however obvious.
We regret our failure to find any relation
between the clearing factor, lipoprotein lipase,
and this type of vascular disease. Certainly.
this is a field that needs more extensive plow-
ing and cultivation.
Finally, our dietarv data with all the limi-
tations of the methods employed to obtain
them appear worthy of presentation and val-
uable as a basis for comparison with other
groups. It is true that our population does
not include a low-fat group comparable to
populations found in certain other parts of
Downloaded from http://ahajournals.org by on March 17, 2024
the world. Perhaps, imore important, the fig-
ures presented do show the dietarv profile of
a population group living under free condi-
tions and exposed to a combination of envi-
ronmental influences not unlike those preva-
lent in most other parts of the United States.
The findings noted contrast with the step-wise
association observed with certain other fac-
tors and would seem to inject a healthv note
of caution at least toward attempts to alter
the American diet within the modest range
of fat intake we have described.
Summary
Certaini of the data accumiulated over the
course of more than 4 years in a prospective
study of coronary heart disease being carried
on in an industrial population have been pre-
sented.
In a base population of 1,989 men, 88 eases
of coronary heart disease have developed:
*Increasing severity of degree of atherosclerosis
with inereasing elevation of the blood cholesterol
levels ba s been noted experiimentally.10
PAUL ET AL.
angina pectoris, 47 men; mnyocardial infare-
tion, 28 men; death from coroniary, disease, 13
men. This approximates one case per 100 nmen
per year.
The development of clinical coronary heart
disease has shown an association with early
age of death of father, history of "noncar-
diac" chest discomfort, history of chronic
cough, history of shortness of breath, history
of peptic ulcer, presence of increased skin-
fold thickness, elevated blood pressure, AV
nicking in the fundi, elevated blood choles-
terol, ST and T abnormalities in the electro-
cardiogram, and use of cigarettes and coffee.
No relation was encountered between body
weight, mean blood sugar levels, lipoprotein
lipase levels, or diet (other than coffee), and
the development of coronary heart disease.
Similarly, there was no association with job
tvpe and no certain relation to physical ae-
tivity off the job.
Acknowledgment
The following physicians have been active in this
project for twvo or more years and their invaltiable
assistance is gratefully acknowledged: Maurice Albala,
Harry Bliss, Herschel Browvns, MIarvin Colbert, Henry
De Young, Peter Economou, Sanford Franzblau,
Robert Felix, John Graettinger, Buford Hall, Wallace
Kirkland, Joseph Muenster, Hyman Mackler, Adrian
Ostfeld, Robert Parsons, Charles Perlia, Norman
Roberg, Marvin Rosenberg, George Saxton, Armin
Schick, James Schoenberger, John Sharp, Jay Silver-
man, Donald Tarun, and Walter Wood.
Appendix
In rielation to the soinatotype material, r eferencee
no. 4 quoted, page 5, gives the following description
of the three components used. "Endomorphy means
relative predominance of soft roundness throughout
the various regions of the body." "Mesomorphy
means relative predomiinance of miuscle, bone, and
connective tissue." "Ectomorphy means relative pre-
dominance of linearity and fragility." The indi-
viduals were rated on a scale of one to seven for
each of these components.
In relationi to job classification, it may be helpful
to indicate some of the definitions employed in the
Company physical deniand code for w-ork in the shop.
For the purpose of job evaluationi, "'continuous"
meanis 50 per cenit or imiore of the timi-e, ''frequent'"
10 to 50 per cent of the time, "'occasional'" 2 to
10 per cent of the time. Thus, code 2 involves
(a) work that requires the continuous lifting of
Circulation, Volume XXVIII, July 1.963
 CORONARY HEART DISEASE
material weighing over 1 pound and up to 5 pounds
inclusive, or frequeint lifting of material weighing
over 5 pounids anid up to 25 pounds inclusive, or
occasional liftinig of mi-aterial weighing over 25
pounids and up to 60 pounds inclusive or equivalent
exertion pushing or pulling; or (b) work that requires
the employee to stand on one work position or walk
continuously or work involvinig conltinluous use of
arms raised in an unsupported position, lifting or
holding material weighing up to 1 pound inclusive.
Code 4 involves (a) work that requires the continu-
ous liftinig of m-iaterial weighinig over 25 I)ouii(ds
and up to 60 pounds or equivalent exertion pulling
or pushing; or (b) work that involves continuous
strain due to streniuous work positions.
The work-conidition code in the office occupations
refers to requiremients or conditions incidental to the
occupation or to surroundinigs that result in the tirinlg
or discomfort of the incdividual. These iielud(le abnor-
mal visual demanid, physical effort enduraitee, (lis-
or
tinetly objectionial surrounidiings insvolviing lheat, col(1,
dirt, noise, fumes, or hazards. Thus, the first- degiee
code in the office is essentially normial office sutirrotnd(1-
ings with minimmn tiring effect, such as ouild be
niecessarily present in any job. The seconid-degree co(le
includes some one individual objectionable surround-
ing or physical or visual demand existinig to al degree
distinctly beyond that elncountered in nornial salaried
occupations. The third code is a combination of two
or mlore objectionable conditions, each of whielh (Iiuali-
Downloaded from http://ahajournals.org by on March 17, 2024
31
height and weight data, body fatniess measurements,
meani pulse rates, mean bloocl pressure levels, anid
meani cholesterol levels. Analysis of variance was
employed for the cholesterol levels in the coroniary
subgroups, and Welch 's approximation was used for
the mean blood glucose levels.
References
1. Epidemiology of Cardiovascular Diseases: Mle-
thodology. Report of Princetoni Confereinee.
Am. J. Pub. Health, Suppl. 10, 50: 1, 1960.
2. DAVW-BER, T. R., MEADORS, G. F., AND MIOORE,
F. E.: Epidemiological approaches to heart
disease: The Framingham Study. Am. J. Pub.
Health 41: 279, 1951.
3. PAUL, O., AND MACMILLAN, A.: Nutritional habits
of mnidwestern workinigmen. Illinois MA. J. 120:
273, 1961.
4. SHELDON, W. H., STEVENS, S. S., AND TUCKER,
W. B.: The Varieties of Human Physique. New
York and London, Harper, 1940.
5. PILGERAM, L. O.: Deficiencies in the lipoprotein
lipase system in atherosclerosis. J. Gerontol.
13: 32, 1958.
6. GERTLER, M. M., AND WHrxM, P. D.: Coronary
Heart Disease in Young Adults: A Multidis-
ciplinary Study. Cambridge, Harvard Univer-
sity Press, 1954.
7. OSTFELD, A., AND PAUL, O.: Some observationis
on the inheritance of hypertension. Lane-et. In

fies ani occupation for no. 2 or some one inidividual, press.

very objectionable, surrounding condition.
For those statistically minded, the following conm-
ments may be made. The chi-square test of sig-
nificance was employed in determining the p values
for history of chest discomfort, history of cough
anid shortness of breath, history of peptic ulcer,
somatotype data, distribution of pulse rate, finding
of arteriovenous nicking, relationship of cholesterol
levels to coronary disease, the electrocardiographic
findings, and for the cigarette and coffee data. On
the other hand, t tests were appropriate for the
8. THO-MAS, C. B., HOLLJES, H. W. D., AND EISEN-
BERG, F. F.: Observations on seasonal varia-
tions in total serum cholesterol level among
healthy young prisoners. Ann. Int. Mcd. 54:
413, 1961.
9. RoSE, G. A.: The diagniosis of ischemic heart
pain and intermittent claudication in field sur-
veys. Bull. World Health Organ. In press.
10. HAss, G. M.: Experimiental iniquiry inlto patho-
genesis of human atheroarterioselerosis. Geri-
atrics 16: 581, 1961.

William Heberden 1710-1801

It is a strange play of fortune that Heberden was remembered by many generations of
medical students by virtue of a bizarre formula which the undergraduates called
"Heberden's ink." Originated by Heberden about 1760, it became official in the Dublin
Pharmaeopeia in 1826 under the title Mistura Ferqii Aromnatica.-Prefatory Essay, by
LEROY CRUMMER. WILLIAM HEBERDEN. An Intr-oductio)i to the Stucdy of Physic. New
York, Paul B. Hoeber, Inc., 1929, p. 21.

Circulation, Volume XXVIII, July 1963?