Tobacco Documents Online

...................... C yH American oronar eart ~eart Disease Associatio~ Monograph in +u~o+ a+ Seven Countries Edited ln3 Ancel Keys, Ph.D. Director, Laboratory of Physiological Hygiene, University of Minnesota, School of Public Health, Minneapolis, Minnesota. THE AMERICAN HEART ASSOCIATION. INC., NEW YORK 1970 TIEX 0007117.001

. . • i. T~e Study Pre~r~m a~d Objectives ...................... Iol il. Methods ........................................ : ...... I-9 ~ Conterlt$ " III. Classificatlon--Codes and Cohorts ...................... T-14 IV. Five-Year Follow*up of Employees of Selected U. S. Railroad Companies ............................................. I-~O V. Five-Year Follow-up in Dalmatia and Slavonia .......... T-40 VI. Five-Year Experience in Finland ......... : .............. VII. Five-Year Experience in Rural Italy .................... T-63. VIII. Zutphen, A Town in the Netherlands .................... IX. The Greek Islands of Crete and Corfu .................... T-~ X. Rural Southern Japan .................................. T-101. XI. Five Years of Follow-up of Railroad Men in Italy ........ Ll13. XII. Three Cohorts of Men Followed Five Years in Serbia ...... XIII. Multiple Variables ..................................... XIV. Prevalence ..................................... i ....... XV. Prognosis of Coronary Heart Disease Found at Entry ...... T-1~8 XVi. The Electrocardiogram in Prediction of Five-Year Coronary Heart Disease Incidence Among Men Aged Forty through Fifty-nine ............................................. T-15~ XV|I. The Diet ............................................... T-lb'2. XVIII. Some Problems ......................................... Summary .............................................. T-I~ References ............................................ : I-198 - Appendix .............................................. 1-199 TIEX 0007117.002

CORO: Sum,nary Amo: old my of five IN AN international cooperative study on the thereafter, blood sample, and qualitative times epidemiology of coronary heart disease urinalysis. Frequent exchange of professional CHD-f (CHD), international teams examined 12,770 personnel between national teams and diag- same c men aged 40 through 59 years in Finland, nostic guidelines were provided to assure toris at Greece, Italy, Japan, the Netherlands, the maximum comparability of examinations and and E United States, and Yugoslavia. Strictly stan- diagnoses. 4.7 am dardized methods and criteria were used, and Average diets of all cohorts except the men w all items that could be measured and analyzed railroad men ~vere estimated from surveys on accour~ centrally were handled at the University of random samples of the cohorts. The dietary with a Minnesota-all tabulations and statistical surveys involved weighing all items consumed 87 m~ work, classification of electrocardiograms, di- during seven days of the survey and ~vere myoca agnoses and causes of death, chemical analy- repeated in different seasons. Nutrients were alive f ses of a serum cholesterol, dietary items, and estimated from chemical analyses of compos- for e~ menu composites, ites of replicate meals and menus as ~vell as exan~ The 12,770 men included 11 cohorts of 500 to from tables of food composition developed to men 1,000 men each in rural Yugoslavia (three), cover local foodstuffs. The diets of the U.S. ECG Finland (two), Italy (twol, Greece (t~vo), and railroad tnen were estimated by calculation Japan (two), the men examined comprising from dietary interview and recall records, pector an average of more than 95~ of all men aged supplemented by visits, to the homes of a years 40-59 living in geographically defined areas, small subsample, with In Yugosla~ia there were two additional After entry examination the men were cohorts of men aged 40-59, one comprising followed, with checks by an internist on men in a small .agricultural center ~vith some mortality and major morbidity several times a food-processing industry (Zrenjanin), the year, and then were re-examined, as at entry, other comprising members of the faculty of after five years. Complete five-year re-exam- the University of Belgrade. In the Netherlands inations covered 94.2~ of all survivors, and at the cohort was a statistically drawn sample of least some information about health status was four out of nine men aged 40-59 in Zutphen, obtained for ahnost all of the other men; a small commercial to~vn with light industry in fewer than 0.5 of 1~ were lost to follow-up. central Holland. In the United States the Prevalence of Coronary Heart Disease cohort of 2,571 men ~vas made up of employees of railroad companies in the Great differences between cohorts in age- northwestern sector of the country, the standardized prevalence rate of CHD were eligible men being drawn by sampling the observed at entry, ECG evidence of old companies, locations, and occupations. For myocardial infarction being many times high- er in the U.S. and in Finland than in comparison, a sample of 768 railroad men Yugoslavia, Greece, Italy, and Japan, with aged 40-59 was enrolled in Italy. Zutphen being intermediate. Angina pectoris The examination procedure included stan- and CHD diagnosed on "softer" clinical and dardized questionnaires on family status, ECG criteria tended to show similar popula- work, personal habits and medical history, tion differences. The prevalence rates of CHD anthropometry, including subcutaneous fat of the cohorts tended to be directly related to measurements, physical examination by in- characteristics of the cohorts in regard to ternists using a standardized protocol and blood pressure and serum cholesterol but not record forms, 12-lead ECG, three-minute in regard to relative body weight or body ' exercise test w/th ECG repeated immediately fatness or smoking habit~ of the cohort~. 1-186 S~t~st 1 to Ci~cM~ios, Fob. XLI ~ul XLII, ,4p¢il 1970 TIEX0007117.003

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-187 Prevalence at Entry--Prognosis ship to relative body ~veight, blood pressure, Among 129 men with a diagnosis of definite or serum cholesterol. old myocardial infarction at entry, at the end Deaths of five years 27 were dead from CHD, or 13.6 ,'e times the CHD death rate of men iudgcd During five years there were 588 deaths in al CHD-free at entry of the same a~e in the the entire study population, 158 from CHD. g- same cohorts. Prevalence cases of angina pec- Figure $1 summarizes the age-standardized "e toris and of CHD diagnosed on softer clinical death rates, cohorts being combined by d and ECG criteria experienced, respectively, countries. Among the U.S. railroad men 62 of 4.7 and 6.8 times the CHD death rate of the 125 deaths were due to CHD; in Finland 38 ~e men without e~Sdence of CtlD at entry. CHD of 111 deaths were due to CHD; in the Netherlands 16 out of 50 deaths ~vere thus n accounted for 725 of all deaths among men Y with any diagnosis of CHD at entry. Among accounted for. For all other col~orts combined, d 87 men with a diagnosis of definite old only one out of eight deaths ~vas due to CHD. That low proportion did not reflect a high e myocardial infarction at entry and who were death rate ascribed to other, non-CHD causes. e alive five years later, 29 did not meet criteria In most of the cohorts, low CHD death ;" for e~'idence of CHD at the five-year re- rates were associated with low all-causes s examination. Of 100 five-year survivors among deaths, the standard basis of comparison for ;. men who at entry had less definite clinical and CHD and all-causes deaths being the deaths ~ ECG signs of CHD or who had classic angina expected to match the five-year experience of • pectoris, 41 met no criteria for CHD five equal numbers of white men in the United ~ years later. The five-year prognosis of men States with the same age distribution, the with CHD at entry showed no clear relation- expected numbers being calculated from life TIEX 0007117.004

1-158 KEYS C( tables and death rates in U.S. vital statistics gave O/E=0.545, the ratio of observed to for 1962 (approximately in the middle of the expected CHD deaths was 24/139.2, or five-year follow-up). O/E =0.172. In these ten cohorts there were The least favorable all-causes death rates, 157 fe~ver all-causes deaths than expected; the an expressed as ratios of observed deaths to ~ relative rarity of CHD in those cohorts c~ deaths expected to match U.S. experience, accounted for 115 of the 157 expected total in( were east Finland (O/E = 1.34); Slavonia, deaths that did not occur. In contrast, in the 10. Yugoslavia (O/E = 1.22); ~vest Finland other six cohorts combined, all-causes deaths th, (O/E = 0.96) ; Crevaleore, Italy (O/E = 1.00); were 98.4% of those expected to match the faz Zutphen, Netherlands (O/E = 0.95); and U.S. general population of U.S. white men. 10 railroad men (O/E = 0.82). The unfavorable Among U.S. railroad men age-standardized mortality in Slavonia was accounted for by death rates from all causes were higher among death causes scarcely seen in the other co- switchmen (physically active) than men in i th, horts-tuberculosis, suicide, cirrhosis of the sedentary occupations while the reverse ~vas 1 Ft liver, and acute alcoholism. The U.S. railroad true of CHD deaths, but in neither case was CO men, being fitlly employed at entry were the difference statistically significant. The tl~ expected to have a lower death rate than U.S. same ~vas true when comparison was made of white men in general. The other ten cohorts men free of CHD at entry, fo: were expected to provide 344.7 deaths; the co observed number was only 188 deaths, Incidence o~ Coronary Heart Disease be O/E = 0.545. Among 12,529 men iudged to be free from Fi Great differences in CHD death rate CHD at the entry examination, the five-year in accounted for most but not all of the CHD experience, in a hierarchy ot? mutually differences in all-causes deaths. In the same exclusive diagnostic categories, was as fol- ra! ten cohorts for ~vhich all-causes death rate lo~vs: 116 deaths from CHD, 113 nonfatal Figure $2 Age-standardized average yearly CHD incidence rates per 10.000 of 1~..5~9 men aged 40-59. judged to be free of CHD at the outset, followed for five years. Non-fatal CItD incidence in Japan is not precisely indicated because the relevant 5-year clinical and ECG records were not independently viewed at the University of Minnesota center. MEN 40-59, C HD-FREE AT ENTRY CHD INCIDENCE! 10,0001YEAR ~32 GREECE ~53 YUGOSLAVIA ,o Vols. XLI ~ XLII, Al~l 1970 TIEX 0007117.005

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-189 definite myocardial infarctions, 219 cases of per 10,000 with 95g confidence limits of +6 classic angina pectoris, and 113 men given the and +278. diagnosis of CHD on the basis of less rigid In general, the several categories of CItD and specific clinical and ECG criteria. The diagnosis tended to show much the same crude average annual rate for all CHD picture for differences in rates between incidence in these men initially CHD-free ~vas cohorts, but the rates for angina pectoris were 102.3 per 10,000. For CHD incidence meeting not in close conformity with the rates for the the hardest criteria-deaths and definite in- obiective and standardized diagnoses of CHD farcts-the corresponding rate was 36.5 per death and nonfatal infarction. In the Nether- 10,000 men. • lands angina pectoris apparently contributed a The age-standardized CHD incidence rate lower proportion of total CHD incidence than for men CHD-free at entry differed greatly in in the other areas; there is no way of deciding the several cohorts, the extremes being in whether this apparent peculiarity reflects a Finland and the U.S. on the high side xvith the difference in the manifestation of the disease cohorts in Japan, Greece, and Yugoslax-ia on or diagnostic conservatism of the responsible the low side. Figure $2 summarizes the data cardiologist. for the cohorts combined by countries. Within Differences Between Cohorts ,Risk Factors countries there were no significant differences Examination of the representation in the between cohorts in CHD rates except for rural several cohorts of the so-called risk factors Finland where the total CHD incidence rate shows that most of those factors, whatever in the east (272 ± 59) xvas definitely higher may be their influence within cohorts, cannot than in the west ( 130 ± 39). The difference in explain the observed differences in the inci- rate, east minus west Finland, was 142 cases dence of CHD. Figure $3 shows that cigarette Figure $3 Percentage'o! men regularly smokin;~ at least lO cigarettes every day. The lengll|s of the narrow solid bars are proportional to the age-sta.dardized CHD incidence rates amo.g men CHD-free at entry as given in Figure S~. MEN 40-59, % SMOKING > 10 CIGARETTES/DAY 59%..TAPAN 49 %. C~EECE ----. 49 %, YUGOSLAVIA __43%, ITALY 44 "/., NETHERLANDS ~ '~.. u.s 50%, FINLAND NARROW,5OI.IO BARS SHOW CHD INCIDENCE RATE $11~m~t I :o Ci~c~ion, Yol:. XLI ~md XI.11, tlf~i~ 1970 TIEX 0007117.006

1-190 KEYS MEN ~,0-59, "/o SEDENTARY ', ~, JAPAN ~18'/., GREECE 30%, YUGOSLAVIA 14 %, ITALY --F 24 %, NETHERLANDS 60 %, US. 10%, FINLAND NARROW,SOLID BARS SHOW CHD INCIDENCE RATE Figure Percentage o| men sedentary or engaged only in very light physical ~etivlty. Figure 55 Percentage of men with relative body weight of 110 or more. MEN 40-59.% WITH RELATIVE WEIGHT >110'/, 2'/., JAPAN ~'22222] ~11'I., GREECE 19%, YUGOSLAVIA 33'/,, ITALY 13%, NETHERLANDS :32%, U.S. R~. 15%, FINLAND NARROW.SOLID BARS SHOW Clio INCIDENCE RATE $~1~ I to Ck'c~io~, Vols. X.LI ,~g XI.II, Al~'il 1970 TIEX 0007117.007

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-191 MEN 40-59,'/. WITH ~" SKINFOLDS • 28 rnrn ~ ~%, JAPAN ~ 11%, GREECE ~__~_=__ ._ ___~ 29"/.. YUGOSLAVIA -- -- 28 °/,. ITALY 32%, NETHERLANDS 63"h RY. FINLAND NARROW.SOLID BARS SHOW CHD INCIDENCE RATE Figure SO Percentage of men with values of -~8 or more mm for the sum of the skinfolds over the triceps muscle and over the tip ot' the scapula. smoking cannot be involved as an explanation; they were much less prone to CHD than cigarette smoking habits do not differ much either the Finns or the Americans. The con- between the various cohorts or cohorts corn- clusion is that the group trends in these vari- bined by country as in figure $3. ables have nothing to do with the group Differences between the cohorts in the pro- trends in CHD incidence. portion of the men who are sedentary or Blood pressure and the prevalence of physically inactive do not explain the differ- hypertension are more interesting, as shown ences betxveen the cohorts in the incidence by figures $7 and $8. There is some tendency of CHD. The data are summarized by coun- for the incidence of CHD to be related to the tries in figure $4. prevalence of hypertension in the cohorts; at There were large differences between co- least hypertension was less common in the horts in body fatness and relative body cohorts with the lowest incidence of CHD. It ""weight, but as illustrated by figures $5 and $6, is not possible, of course, to insist that the consideration of neither obesity nor relative blood pressures recorded for the several weight helps to explain the population differ- cohorts are strictly comparable, especially for enees in CHD incidence. The U.S. railroad diastolic pressure. Common instructions were men tend to be relatively much heavier and issued about recording blood pressure, but fatter than the men in any of the other groups, uncertainties remain in the absence of identity but the thin and relatively lightweight Finns of the environment ond of a measuring device match the Americans in CHD incidence. The without human intervention. The role of blood Italians were the most often overweight and, pressure in CHD incidence is more reliably although less fat than the Americans, were examined within cohorts. certainly far more obese than the Finns, yet Figure $9 indicates that the incidence rate Sml~l~m~t I to Cicc~4~io~, Vols. XLI ,~nd XLII, Atoll 1970 "FLEX 0007117.008

1-192 KEYS Figure $7 Age-standardized percentage of inert ~ith resting systolic blood pressure of 160 mm or more. MEN 40-59,'/, MEN WITH SYSTOLIC B.P.->160 1:~8 'I, ,]APAN ~~--~ 12.5 '1., OREECE __'114%, YUGOSLAVIA -~ ~:_ ____~_ ~ 18.3"/., ITALY 21.2 "/,, NETH. 16.4%, U.S. RY ---- 18.7, FINLAND NARROW.SOLID BARS SHOW CHD INCIDENCE RATE MEN 40-59, % HYPERTENSIVE (DIAST. _> 9~ ram) 9.3%, JAPAN I ~-~--.-~.~ 11.9°I,,OREECE 15.'/'I,, YUGOSLAVIA ~-~:.~ 22.1"I.. ITALY 'I.. NETH. . .... 2Z3 "I,, U~. RY 20 2'I,, FINLAND NARB.~ID BARS S~ CHD INCIDENCE RATE Fibre S8 " A~-s~ndardi~ ~ntage of men with resting diastolic bl~ premm of ~ mm or morn. 5~ I ~o Ci¢¢~, Voi$. ~ ~ ~1, ~ 1970 TIEX 0007117.009

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-193 Figure $9 Percentage of men with serum cholesterol values over ~50 mg per deciliter. MEN 40"59, % CHOLESTEROL >250mg/dl ~7%, JAPAN 14 %, GREECE 7%, YUGOSLAVIA - --:-~ --> 13 %, ITALY 32 %, NET HERLANDS 39%, U.S. RY NARROW.SOLID BARS SHOW CHD INCIDENCE RATE MEN 40-59. % DIET CALORIES PROVIDED BY SATURATED .IEA. 7%, GREECE 10 °/o, YUGOSLAVIA 19%o NETK 18%, U.S. RY 20%. F[NLANO NARROW, SOLID BARS SHOW CHD INCIDENCE RATE F|~zre Average percentage of total dietary calories provided by saturated fatty acids.. 5"~1~¢n~ I to Ci~czlatiom, Vols. XLI ,rod XLII, A~il 1970 TIEXO007117.010 ~ •

1-194 KEYS of CHD tends to be directly related to the significant relationships between habitual distributions of serum cholesterol values. Since physical activity and any measure of CHD the average serum cholesterol values of the iucidence rate were found. If there were a cohorts tended to be directly related to the true excess of 15 to 20~ CHD among average proportion of calories pro~'ided by sedentary men, the present material would be saturated fats in the diet, it is not surprising to too small to prove it. find the picture shown in figure SI0. The Analysis was made of the relationship, in CHD incidence rates of the cohorts are just as men judged to be CHD-free at entry, between closely related to the dietary saturated fatty CHD incidence and certain ECG abnormali- acids as to the serum cholesterol level. Serum ties at entry. Ischemic type of ST depression cholesterol averages and CHD incidence rates after exercise was associated with a CHD were not found to be related to the percentage incidence rate more than double the rate in of calories provided by protein or polyunsatu- men without that abnormality. The bad rated fatty acids in the diet and were only outlook for men with that ECG abnormality slightly related to total fat calories, persisted when matching was done on blood Average relative body weight, as well as pressure, relative weight, serum cholesterol, average body fatness, tended to be inversely physical activity, smoking habits, and age. related to the average dietary calories per unit Other post-exercise ECG abnormalities inves- of body mass. It was indicated that, on the tigated inclnded junctional ST depression, average, relative obesity and overweight is negative T waves, and various arrhythmias, more a reflection of underexpenditure of none of which proved to be associated with calories rather than of overconsumption, significantly more CHD incidence than could Risk Factors Within Cohorts easily occur by chance. On the other hand, among men not judged as CItD-free, later The analysis of the relationships within CHD deaths were significantly increased when cohorts consistently indicated the importance the entry ECG showed large Q waves, nega- of blood pressure and serum cholesterol, tire T waves, or atrial fibrillation. When "hard" criteria of CHD death and infarction were used for diagnosis, CHDMultivariate Differences Between Cohorts incidence was not significantly related to The coeff;cients for the multiple logistic either relative body weight or to body fatness, equation for CHD risk, obtained by Truett et When all CHD diagnoses were used in al. (1967) from Framingham data, were computing incidence rate, there was a weak tested with the present material. Comparison tendency for the rate to be related to relative of absolute numbers of CHD cases with those body weight as well as to body fatness. That "predicted" is improper because of lack of tendency was not statistically significant when identity in the diagnostic methods and the the confounding influence of blood pressure question of comparing five-year observed rates was removed, with predictions based on 12 years of follow- CHD incidence was significantly related to up. However, analysis in terms of ratios of smoking habits in the U.S. railroad men but rates would seem to be reasonable to consider. not in the European cohorts. All-causes death When both observed and predicted CHD rate was also related to smoking habits in the incidence rates for the various cohorts are U.S. railroad men but not in the other expressed as ratios of the observed and cohorts, predicted rates of the U.S. railroad men, the In the U.S. railroad cohort the CHD correspondence between observed and pre- incidence rate of the men in sedentary dieted ratios was unexpectedly good; the occupations was about 16~ higher than the coefficient of correlation was r=0.83. A rate of the physically more active switchmen, detailed multivariate analysis with new solu- but the difference was not statistically signifi- tions to the multiple logistic is in progress. cant. In the other cohorts no statistically Besides the approach of Truett et al. (1967), $~1~1~'~.~ I ~a Ci~io,~, Yoh. ~ ~ XLI1, ,'l;~-il 1970 TIEX 0007117.011

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-195 which assumes multivariate normality and In spite of the bad experience in Slavonia, equality of variances and covariances, that of for all five Yugoslav cohorts combined, all- Walker and Duncan (1967) without those causes deaths totaled 123 compared with 175.5 assumptions is being used. expected for age-matched white men in the United States, O/E=0.70. The all-causes Total Mortality and Coronary Heart Disease mortality of the three Italian cohorts was also There was no indication that the incidence favorable with O/E = 0.80 (114 deaths ob- of CHD was inversely related to the incidence served, 142.3 expected to match U.S. white of any other disease or that, in effect, rarity of men), as was that of the Japanese men with CItD in a cohort was compensated by an excess of other affliction. In general, the total O/E =0.78 (47 deaths, 60.3 expected). The Greeks had by far the best experience ~vith all-causes death rate reflected the death rate from CHD with the result that the all-causes only 21 deaths instead of 70.5 expected, death rate was remarkably low in several of O/E =0.30. CHD was relatively much less the cohorts with the lowest incidence of CHD. common in all of these cohorts than in the In Slavonia, an apparent exception, unusual Finns, the men of Zutphen, and the U.S. causes of death were involved, railroad men. Sapplam~,~t 1 to C'h, cad~tlo~, Volt. Xl.l ,,mR XI.II, dp~l 1970 TIEX 0007117.012

J References ADELSON S, KEYS A: Diet and sonie health charae- Dalmatia and Slavonia. Aeta Med Seand (suppl) teristics of 123 business and professional men. U.S. 460: 147, 1967 Department of Agriculture Publ ARS 6211, 1962 CttaPXfXN JM, COEI~K.E LS, DIxon ~V, ET AL: Clinical AxnEnsoN JT, GV,~r,'DE F, .'~,IATsI~.. iOTO Y, ET ^t: Glu- status of a popnlation group in Los Angeles under cose, sucrose and lactose in the diet and blood observation for two to three years. Amer J Public iipids iu man. J of Nutrition 79: 349, 1963 Health 47 (pt 2): 33, 1957 ANTONIS A, BERSOHN P: Influence of diet on serum Cr~zAxo BN, PERLMAN LV, OST~.X~DEa LD Jn, L~r lipids in South African white and Bantu prisoners. ^L: Relationship of premature systoles to coronary. AmerJClinNutr 10: 484, 1962 heart disease and sudden death in Tecumseh Art~vAxts C, DO.',TAS AS, LEgOS D, ET AL: Rural epidemiologic study. Ann Intern Med 70: 1159, populations in Crete and Corfu, Greece. Acta Med 1969 Stand (suppl)460: 209, 1967 Committee on the Coronary Circulation: Statistical Association Medico-actuarial Mortality Investigations, review of eases with clinically diagnosed m.vocardial vol. 1. New York, Assoc Life Insurance Med DiE infarction and with anginal pains iu several districts and Actuarial Soc Amer, 1912 in Japan. Jap Circ J 21: 1, 1957 BLACKI~U~tN H, KEYS A, Si.xio.,,-sox E, L~r ,~L: The DAVENPORT CB: Body build and its inheritance. electrocardiogram in population studies: A classi- Carnegie Institute, Washln~ton. D. C., puhl no 329. fication system. Circulation 21: 1160, 1960. 1923 B~cK~vax H, Pa~tax RW, .~×~ KE~'S A: The inter- D.~w~E~ TR, Moon~ FE, MA.xx GV: Coronary heart relations of electrocardiographic findings and phys- disease in the Framingham Study. Amer J Public ical characteristics of middle-aged men. Acta Med Itealth 47 (pt 2): 4, 1957 Stand Suppl 460: 316, 1967 Djor~njEvlc B, JoslPOVlC V, NEDELJKOXqC SI, ET ~L: BLOOXr WL, Emzx MF: Inactivity as a major factor in Men in Velika Krsna, a Serbian village. Acta Med adult obesity..Metabolism 16: 679, 1967 Scand (suppl) 460: 267, 1967 BRO.N-rE-STEWART B, Krxs A, Bnoc): JF: Serum- DJORDJEVlC B, SL%[IC B, Si.XtlC A) ETAIL, %VITI-I TIlE cholesterol, diet and coronary heart disease. Lancet ~rLt" or ~e SrAT|S~ZC|~,,¢ Tonotaov~c, P: Dietary. 2: 1103, 1955 studies in connection with epklemiology of BnO.'~xE-STEw.~nT B, AxTox~s A, G^LES L, ET .*L: heart diseases: ResalEs in Serbia. Voedin~ 26: Effect of feeding different fats ou serum-cholesterol 117, 1965 level. Lancet 1: 521, 1956 DOYL~ JT, HEstaxs S, HmLrnOE ttE, rx aL: o Baoz~ J, .~.,,'o ALEX~'~Or'n H: A note on estimation Prospective study of degenerative cardiovascular of the components of variation in a two-way table, disease in Albany. Report of three years' ex'peri- Amer J Psych 60: 629, 1947 ence. 1. Ischemic heart disease. Amer J Pnblic BaozE~: J, Bt'z~.,,'.~ R, M~rac F: Population studies on ltealth 47 (pt 2): 25, 1957 serum cholesterol and dietary fat in Yugoslavia. DriEd'ross 17: Incklence of myocardial infarction in Amer J Clin Nutr 5: 279, 1957 various communities in Israel. Amer Heart J 45: vx~B~zc~xt FSP, DALD~a~.a" L: Town of Zutphen, 749, 1953 the Netherlands. Acta Med Scand (snppl) 460: Dmg~rvss F, Toon M, AG.MON J, ET AL: Observations 191, 1967 on myocardial infarction in Israel. Cardlo!ogia 30: Brat.LL~ BA, REr~ RB, M.,wn J: Physical activity of 387, 19o-'7 obese and non-obese adolescent girls appraised by FmAxz~ F, FmA~z~ AI..BI~TI A, FEI~o-LL'zzh G, ~-r motion picture sampling. Amer J Clin Nutr 14: ~: Dietary surveys in connection with the epidemi- 211, 1964 ology of heart disease: Results in Italy. Voeding 25: Bvzh'~,~ R, F~aBgn E, KE~S A, ~ .~L: Diets of rural 502, 1964 families and heads of famihes in two regions of FmAszx F, ~,'~D F~DA~Z~ AL~En~ A: Dietary SUE- Yugoslavia. Voeding 25: 629, 1964 veys in connection with the epidemiology of heart Bt~'~x R, KgYs A, BnODXnEC A, ET XL: Dietary disease: Reliability, sources of variation and other surveys in rural Yugoslavia: It. Chemical analyses data from nine surveys in Italy. Voeding 28: 244, of diets of Dalmatia and Slavonia. Voeding 27: 31, 1967 1966 FIDANZA F, PUDDU V, DEL VECCltlO A, AND KEYS A: Btrza~A R, KEYS A, BaODanEC A, ET AL: Dietary Men in rural Italy. Acta MOd Stand Suppi 460: survey in rural Yogoslavia: III. Comparison of three 116, 1967 methods. Voeding '~7: 99, 1966 GnrzzL~ JE: Continuity correction in the Xe-test for Btr-z~o, R, KEYs A, MOaaCEK I, ET at: Rural men in 2 x 2 tables. American Statistician 21: 28, 1967 1-196 S,~l~n~t I to Circ~l~tio~, l~oh. XH ~nd X[,II, Alwil 1970 TIEX 0007117.013

." CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-197 DENHARTOG C, VAN SCHAIK TH FSM, DALDERUP LM, and professional men followed fifteen years. ~'r ~,L: Diet of volunteers participating in a long Circulation 28: 381, 1963 term epidemiological field survey on coronary heart KEYs A, A.'~D Gn^NDE F: Body weight, body composi- disease at Zutphen, Netherlands. Voeding 26: 184, tion and calorie status. Pp 13-43 in Modern Nu- 1965 trition in Health and Disease (MG Wohl, and RS DE.'~HAarO¢ C, Bvzlr~A R, FIDANZA F, ~-r XL EDs: Goodhart, eds.) Lea and Febiger, Philadelphia, Dietary Studies and Epidemiology of Heart Disease. 1964 The Hague, Sticht. Wetensch. Voorlichting Voed- KEYs A: Dietary survey methods in studies on ingsgebied, 1968 cardiovascular epidemiology. Voeding 26: 464, HUENEMANN RL, HA.MPTO.~ ~IC, SHAPIRO LR, Ev 1965 A~,: Adolescent food practices associated with KEYs A, ANDEaSO.~ JT, GaXXOE F: Serum cholesterol obesity. Fed Proc 25 (pt I): 4, 1966 response to changes in the diet. Metabolism 14: KANNEL WB, DAWBErt TR, KAGAN A, E'I" AL: Factors 747, 1965 of risk in the development of coronary heart KEYs A, AaAVXX~s C, SDa~N H: Diets of middle-aged disease-six )'ear follow-up experience, the Fram- men in two rural areas of Greece. Voeding 27: 575, ingham Stud)'. Ann Intern Med 55: 33, 1961 1966 KAnvoxE.~" MJ, BLO.MQVlST G, KALLIO V, ET AL: KEYS A, Pan~l,~ RW: Serum cholesterol response to Men in rural east and west Finland. Acta Med changes in dietary lipids. Amer J Clin Nutr. 19: Stand (suppl) 460: 169, 1967 175, 1966 K~-vs A, Baozm~ J: Body fat in adult man. Physiol KEYs A: Blood lipids in man-a brief review. J Amer Reviews 33: 245, 1953 Diet Ass 51: 508, 1967 K~vs A: Cholesterol problem. Voeding 13: 539, 1952 Kr'.Ys A: Current status of research on the epidemiol- (a) ogy of coronary heart disease. Jap Circ J 32 (No. KEYS A, ~r ^~.: Epidemiological studies related to 12): 1669, 1968 coronary heart disease: characteristics of men aged KEYs A: Sernm cholesterol and the question of "nor- 40-59 in seven countries. Acta Med Scand Suppl mal." Pp 147-170 in Multiple Laboratory Screen- 460, 392 pp, 1967 ing (PS Strandjord and E Benson, eds.) Academic KEYs A: ttuman atherosclerosis and the diet. Press, New York, 1969 Circulation 5: 115, 1952 (b) KE~s A, VwAxco F, Ronamv~z-.Mtxo.~ JL, K~:YS A, KL~tCaA N: Diets of middle-aged farmers in Studies on the diet, body fatness and serum choles- Japan. Amer J Clin Nutr, 1970. In press teroi in Madrid, Spain. Metabolism 3: 195, 1954 KEYs A: Atherosclerosis: A problem in newer public Klx~t~aA N: A farming, and a fishing village in health. J Mount Sinai Hosp NY 20: 118, 1953 Japan-Tanushimaru and Ushibuka. Acta Med (a) Scand (suppl) 460: 231, 1967 KeYs A: Prediction and possible prevention of Klxn;~ N: Analysis of 10,000 post-mortem examina- coronary disease. Amer J Public Health 43: 1399, t'ons in Japan. In World Trends in Cardiology: I. 1953 (b) Cardiovascular Epidemiology, edited by A Keys KEYs A: Obesity and degenerative heart disease. Am and PD White, Hoeber-Harper, New York, 1956, J Public Health 44: 864, 1954 p 22-33 KEYS A, FID.~,NZA F, SCAm~ V, ET AL: Studies on L~.~REa~ HA, BEmCSO.~" DM, Srxx~t.~a J, ~r AL: serum cholesterol and other characteristics on Totally as)anptomatic myocardial infarction: Esti- clinically healthy men in Naples. Arch Intern Med mate of it in the living population. Arch Intern (Chicago) 93: 328, 1954 Med (Chicago) 106: 628, 1960 KEYs A: Weight changes and the health of men. LYt~ AM: Coronary disease as an underwriting Chapter 8, pp 108-118, in Weight Control (ES problem. Trans Soc Actuaries 15: 324, 1963 Eppright, P Swanson, and CA Iverson, eds.) Iowa MAt~,~taos H: Relation of nutrition to health-a State College Press, Ames, Iowa, 1955 statistical study of the effect of war-time on K~vs A: Diet and the epidemiology of coronary heart arteriosclerosis, cardiosclerosis, tuberculosis and disease. JAMA 164: 1912, 1957 diabetes. Acta Med Scand (suppi) 246: 137, Km's A, K~.~tr~ N, Kusvr~WA A, ~ AL: Serum 1950 cholesterol in Japanese coal miners: A dietary MA~"r~L N, HAEXSZ~L W: Statistical aspects of the experiment. AmerJClinNutr. 5: 245, 1957 analysis of data from retrospective studies of l~s A, K~.~tva~ N, Kustrr~WA A, ~r AL: Lessons disease. J Nat Cancer Inst 22: 719, 1959 from serum cholesterol studies in Japan, Hawaii and MA~aT~ N: Chi-square tests with one degree of Los Angeles. Ann Intern Med 48: 83, 1958 freedom: Extensions of the Mantel-Haenszel pro- l~vs A, Kxavo.~Ex MJ, Fmx.~zx F: Serum-cholesterol eedure. J Am Star Assoc 58: 690, 1963 studies in Finland. Lancet 9~: 175, 1958 M~,~a'~L N, G~r:~nousE SW: What is the continuity Km's A, T~,YLon HL, Bt~c~va.'~ HW, ET ~: correction? American Statistician ~2: 27, 1968 Coronary heart disease among Minnesota business M~,s-r~ AM, ROsENr~t~ I: Exercise as an estinaation $~i~t~mc~t I to Circulation, Volt. Xld and XLll, ~pril 1970 _ TIEX 0007117.014

.. 1-198 KEYS of cardiac function. J Amer Coll Chest Physicians RotxE P, PEI.;.r;.ARIXEN M, KARVOXE>," MJ: Dietary 51: 347, 1967 studies in connection with epidemiology of heart . MATHEWSO.'~ FAL, BRmaETOX DC: A-V block: U. of diseases: Results in Finland. Voeding 25: 384, ;~ Manitoba follow-up study reports-series 1963. 1964 Trans Ass Life Insur Med Dir Amer 48: 210, Rose GA, Bt.ACKB~'laX H: Cardiovascnlar survey 1964 methods. WHO Monograph Series no. 56, 1968 Mx~Ea J: Overweight-causes, cost and control. Preu- SC~OaXACEL ttE: Connection between nutrition and tice-Hall, Englewood Cliffs, New Jersey, 213 pp, mortality from coronary sclerosis during and after 1968 World War II. Documenta de Medicina Geogra- ~ ~IETROPOLITAN: Metropolitan Life Insurance Co phica et Tropica (Amsterdam) 5: 173, 1953 Statist Bnll. No 23, 1942 SELTZER CC: Some re-evaluations of the Build and MoRms JN, Kxcxx A, P.~TT~SOX DC, ~ ^L: Blood Pressure Study, 1959, as related to ponderal Incidence and prediction of ischaemic heart disease index, somatot39e and mortality. New England J in London busmen. Lancet 2: 553, 1966 Med 284: 254, 1966 Morm~s JN, .MAnn JW, HEADY JA, ET AL: Diet and SOCIETY OF AcTvxmm: Build and Blood Pressure plasma cholesterol in 99 bank men. British Med J Study. Society of Actuaries, Chicago, Illinois, 1959 1: 571, 1963 STA.XILEP. J, LINDBERG HA, BER~SOX DM, ET AL: N~,TIOXAL HEALTH S~'av~Y: Weight by height and Prevalence and incidence of coronary heart disease age of adults, United States, 1960-62. U.S. Dept. in strata of the labor force of a Chicago industrial H.E.W., Nat Center Health Stat Ser I1, No 14, corporation. J Chronic Dis 11: 405, 1960 1966 STAXtt.ER J: Cardiovascular diseases in the United Nl~V YOP, K HEART ASSOCIATION': Diseases of the States. AmerJ Cardiol 10: 319, 1962 heart and blood vessels. Nomenclature and criteria SrEvA.~i~ PA, HE.~LD FP JR, M:,VER J: Caloric intake for diagnosis. Sixth edition. Little, Brown and Co., in relation to energy" output of obese and non-obese Boston, 463 pp, 1964 adolescent boys. Amer J Clin Nutr 7: 55, 1959 O~AXCOV.~ K, .~XD 4tEJoX S: Dietary studies in con- Sr~0xt A, JEXSEX AR: Mortalit.v from circulatory nection with epidemiology of heart diseases: Re- d'seases in Norway. Lancet 1: 126, 1951 suits of surveys in Czechoslovakia. Voeding 26: TAYLOa HL, Kt-~PETAR E, Km's A. ET ^L: Death rates 71, 1965 among physically active and sedenta~' emplo.vees of PAUL O, L~-:vvk-P, hill, P~EL.*.~ WH, ET AL: A the railroad industry. Amer J Public Health 52: longitudinal study of coronary heart disease. 1697, 1962 Circulation 28: 20, 1963 TAYLOR I-IL, PantRy RW, BLACKBURN H. ET PEARSOX ES: Choice of statistical test illustrated on Problems in the analysis of the relationship of the interpretation of data classed in a 2 x 2 table, coronary heart disease to physical activity or its Biometrika 34: 139, 1947 lack, with special reference to sample size and PE~mxEx M: Chemical analysis in connection with occupational withdrawal. In Physical Activity in dietary surveys in Finland. Vocding 28: 609, Health and Disease, edited b.v K Evanz, K Lange 1967 Andersen. Oslo, Universitetsforlaget, 1966, p. 242- PEE~n~X~X M, KIvxoj.,, S, JOaT~.~ L: Comparison 261 of the food intake of rnral families estimated by TAYLOR HL, .MOXT~ M. Pv~D~; V, ~r ,~L: Railroad one-day recall and precise weighing methods, employees in Rome. Acta Mud Scand (suppl) 460.' Voeding 28: 470, 1967 250, 1967 PLACKETT RL: Continuity corrections in 2 x 2 tables. TAYLOR HL, BL¢C|CUVRX H, BROZEr: J, Ex XL: Railroad Biometrika 51: 427, 1964 employees in the United States. Acta Mud Scand RE,~nXGTOX RD, SC~ORX MA: Determination of (suppl) 460: 55, 1967 number of subjects needed for experimental TOOR M, IC~TCH~LSKY A, AcMO~ J, ~-r At.: Sernm epidemiologie studies of the effect of increased lipids and atherosclerosis among Yemenite immi- physical activity on incidence of coronary heart grants in Israel. Lancet 1: 1270, 1957 disease--preliminary considerations. In Physical Tatr~r-r J, CORNFIELD J, KANNEL W: Multivariate Activity and the Heart, edited by MJ Karvonen, AJ analysis of the risk of coronary heart disease. J Barry, Springfield, I11, Charles C Thomas, 1967, p Chronic Dis 20: 511, 1967 311-319 V^ST~XF~,~ I, KANERVA K: Arteriosclerosis and war- Rxo.~ JW: The development of height weight tables time. Ann Mud Intern Fenn 36: 748, 1947 from life insurance data. Manuscript in the Division W~,LXEa SH, Do,~cx~ DR: Estimation of the of Chronic Disease, U.S. Public Health Service, probability of an event as a function of several Washington, D. C., 1952 independent variables. Biometrika 54: 167, 1967 Rot,','~ P, PEKKARINEN ~l, KARVONEN MJ, ET AL: Diet YANO L, UED~ S: Coronary heart disease in and cardiovascular disease in Finland. Lancet 2: Hiroshima, Japan. Yale J Biol Med 35: 504, 173, 1958 1963 $~pplcmcm 1 to Cifc~o~, Voh. XLI ~ XLII, ,4p~il 1970 TIEX 0007117.015 ~ ~..~,.~ ~.~:~-~-~.~ ~-~ ~,,,,,,- ~.~ .z- ~ ~.~.....~.- ~.'?~ .........._-.:7~.-~~ ~-:~-.~--~-.-~',.-,~-~er'~'-"~-:, .~ ,~....~.~, :.~ ~ ........ . "-~" ,.... =_. ~.~_~.y:.~..r::...~=.~m~..~.:~,~:~:~.~,.~,:~ ...:--'~z ~-,_~

Appendix Relative Body Weight W HEN the present studies were started, through 1953 (Society 1959). The data suffer the only tables widely used for the from the same defects as those summarized by calculation of relative body weight were the the actuaries more than half a century ago, tables of average weight at given height and and the presentation and analysis are scarcely age of Davenport (1923) ~vho put into 'more satisfactory (Seltzer 1966). Heights and convenient form the data from life insurance weights still refer to persons as "ordinarily examinations made over the years 1885 to dressed," but what that means is not clear; pre- 1900 and assembled in "Medico-Actuarial sumably, more men removed their jackets for Mortality Investigation" (Association 1912). the measurements than was the case in the Heights and weights were recorded as "cus- 1890's. However, even a cursory inspection of tomarily dressed," including shoes, and the the figures in this 1959 "Build and Blood Pres- fact that there was very marked terminal digit sure Study" shows that in many cases, perhaps preference for 0 and 5 indicated that many the majority, no measurements were actually weights were not measured but were merely made. For example, the distributions of height estimated by the examiner or stated by the show strong concentrations on even numbersof applicant (Rion 1952). It has been suggested inches. Among men aged 40--49, 18.95 were frequently that the height added by wearing 68 inches tall, 9.15 were 69 inches tall, and shoes might compensate for the weight added 16.6g ~vere reported as being 70 inches tall. by being "customarily dressed" (as of many Since all scales used to measure stature are years ago) and therefore allow the use of the graduated to fractions of an inch, it is obvious Davenport tables ~vith data recorded for that in a large part of the material, the heights height and weight in. stocking feet and light reported were not based on any measurements underclothing, but no actual trials testing that at all. It is safe to suggest that the same was idea have been reported. The serious limita- true of the reported body weights, but the tions of the Davenport tables were recognized extraordinary way in which the actuaries (Keys and Brozek 1953), but it was also grouped and averaged weights in the "Build realized that even grossly faulty tables may and Blood Pressure Study" makes a critical serve as a useful basis for many comparative analysis impossible. purposes. Accordingly, in default of other For men aged 40 through 59 years, as at bases for comparison, in the present study entry in the present studies, the average relative body weight was computed as 100 weights at equal height and age reported in " times the observed body weight divided by 1959 are little different from those published the ~vei~ht given in the Davenport tables for in 1912. In the later report men at all heights men of the same height and age. The less than six feet are listed as being a little Davenport tables, smoothed and converted heavier-l,, to 5 pounds-than some 50 years into the metric system, were published in the earlier. However, such a comparison makes no previous general report on the present studies (Keysetal. 1967). allowance for the fact that clothing in Shortly after the present studies were in 1935-1953 was not the same as in 188,5-1900. operation, the (U.S.) Society of Actuaries From the data of the insurance companies, the published a massive compilation of data best guess is that at equal age and height the representing almost five million insurance average weight of middle-aged American men policies in the United States and Canada and increased about 5~ in 50 years (Keys and covering examinations during the years 1935 (~rande 1964). $~pp~t 1 tO Circulation, Volt. XL~ a~d ~LI[, dp~ 1970 1-199 TIEX 0007117.016

1-200 KEYS APPENDIX TABLE 1 ALL COHORTS, ENTRY MEASUREMENTS. Cutting points below which are to be found 10, 50 and 90 percent of the men. For more details see Keys et al., 1957, pp. 355-382-. For Zrenjanin and Belgrade faculty see Section XII. Age 40-4~ Age 45-49 Age 50-54 Age 55-59 COHORT VARIABLE 10 50 90 10 50 90 10 50 90 10 50 90 U.S. RR Height 166 175 183 169 175 182 168 173 182 168 173 182 ReL Wt. 88 105 122 88 103 120 88 104 120 86 102 120 ~- Skinfolds 17 32 49 17 32 46 19 34 $2 20 32 49 Sys. B.P. 114 130 157 116 133 163 119 138 172 120 139 175 Dias. I~ P. 71 82 99 72 84 100 74 87 102 73 86 104 Serum Chol. 184 234 292 182 234 297 192 235 297 191 243 295 Dalmatia Height 166 175 182 166 173 182 165 173 182 165 172 180 Rel. Wt. 87 94 108 80 93 111 78 90 107 77 88 111 ~ Skinfolds 10 15 34 10 15 31 ? 14 28 9 13 32 Sys. B.P. 115 136 160 120 135 164 120 137 165 120 135 165 Dias. B.P. 70 85 95 70 80 96 70 82 98 70 82 99 Serum Chol 146 182 251 141 185 237 136 186 241 142 188 246 Slavonia Height 161 170 178 162 168 175 159 166 175 161 168 176 Rel. Wt. 81 95 112 80 94 120 77 88 1!4 77 91 111 ~- Skinfolds 10 15 28 9 15 33 9 13 31 9 14 27 Sys. B.P. 111 130 165 116 130 163 115 131 161 115 140 171 Dias. B.P. 68 79 97 69 80 100 68 80 96 68 84 100 Serum Chol 149 196 249 147 197 255 152 ZOO 260 146 194 256 E. Finland Height 161 168 177 161 168 177 !60 168 174 159 167 175 Rel. Wt. 83 94 110 g2 94 110 79 93 111 78 90 110 ~- Skinfolds 10 13 28 10 15 31 9 14 30 8 14 32 Sys. B.P. 124 141 164 125 140 173 130 149 179 130 153 184 Dias. B.P. 75 87 100 78 88 102 77 90 104 78 90 107 Serum Chol 193 265 328 208 272 335 208 262 340 190 259 317 W. Finland Height 162 173 180 164 171 178 163 172 178 162 170 178 Rel Wt. 86 78 115 82 96 115 82 ~)? 115 80 95 114 ~- Skinfolds 11 16 32 11 16 32 11 16 35 10 16 31 Sys. B.P. llZ 133 159 119 135 156 118 139 165 121 143 177 Dias. B.P. 67 80 92 70 80 92 70 82 100 72 82 97 Serum Chol 201 248 314 201 255 319 197 257 323 195 251 305 Crevalcore Height 161 169 176 160 168 175 160 168 176 158 167 175 Rel. Wt. 88 105 128 86 103 124 87 102 122 86 101 125 r. Skinfolds 12 23 40 11 21 38 12 22 37 12 22 41 Sys. B.P. 120 136 161 120 142 169 1.27 147 180 130 157 185 Dias. B.P. 73 84 98 75 87 100 77 88 104 78 90 104 Serum Chol 156 194 256 146 194 257 150 198 246 152 204 257 The late Dr. Louis Dublin of the Metropoli- =recommended" weight. Insurance experience tan Life Insurance Company championed two was that persons at the upper end of the concepts concerning weight that have had relative weight distribution tended to have an much influence but are a constant source of unfavorable mortality experience. Further, eorffusion. The "Met"published tables of there is no obvious good reason why people ~ideal" weight, later more modestly called should continue to gain weight after growth in $~l~nn~s ! to Ci~d, alo~, Volt. XIJ ~d YJdlo d~ii 1970 TII~X 0007117.017

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-201 APPENDIX TABLE 1 page 2 Age 40-44 Age 45-49 Age 50-54 Age 55-59 COHORT VARIABLE 10 50 90 10 50 90 10 50 90 10 50 90 Montegior- Height 159 165 174 157 165 172 156 163 171 156 162. 171 gio Rel. Wt. 82 99 117 84 98 120 81 94 114 79 94 121 Y- Skinfolds 9 15 28 10 16 32 9 14 28 9 14 26 Sys. B.P. 112 128 145 119 134 158 119 137 161 120 142_ 173 Dias. B.P. 70 78 91 70 80 93 70 81 98 72 83 97 Serum Chol. 157 192 249 155 2.00 257 157 199 248 162 198 2.63 Zutphen Height 165 175 183 166 175 183 166 174 182 163 172 182 Rel. Wt. 85 99 111 85 97 111 82 97 113 82 97 110 ~ Skinfolds 14 24 38 13 23 37 13 24 37 14 22 38 Sys. B.P. 12.5 140 160 120 140 165 120 140 175 12.2 145 176 Dias. B.P. 75 90 100 73 90 104 75 90 108 75 88 106 Serum Chol 177 233 2.85 186 235 298 187 22.7 289 177 226 292 Crete Height 159 166 174 158 166 174 159 166 174 158 165 173 Rel. Wt. 82 94 112 79 91 111 75 92 112 77 88 103 ~- Skinfolds 10 14 30 10 14 28 9 15 27 10 14 ?-3 Sys. B.P. 112 131 155 116 137- 155 115 135 164 119 138 174 Dias. B.P. .68 80 96 70 80 97 70 81 95 70 83 94 Serum Chol 156 198 260 154 199 251 158 210 ?-70 163 208 257 Corfu Height 159 167 176 158 166 177 159 166 172. 158 164 17;~ Rel. Wt. 80 94 114 79 93 115 77 92 110 77 90 111 ~Skinfolds 10 16 30 10 15 31 10 14 31 10 14 30 Sys. B.P. 109 130 154 110 130 160 I11 134 164 111 135 167 Dias. B.P. 70 81 92 70 80 98 71 81 97 70 81 95 Serum Chol 146 193 262 147 203 259 162 202 258 154 194 251 Tanuschi- Height 155 167- 167 1.54 161 168 153 159 167 152 160 165 maru Rel. Wt. 80 89 98 77 86 100 75 84 100 75 84 99 ~ Skinfolds II 15 23 I0 15 2a. I0 14 26 II 15 26 Sys. B.P. 102 120 145 I0"8 128 152 ii0 137- 160 I12 138 183 Dias. B.P. 51 68 83 58 70 90 60 72 90 60 78 I00 Serum Chol 109 167 277 III 165 259 114 178 266 116 168 ~57 Ushibuka Height 153 160 167 153 160 167 151 158 165 150 159 168 Rel. Wt. 77 91 102 78 89 llJl 77 87 97 72 84 96 Y- Skinfolds Sys. B.P. 110 126 152 107 17-8 165 112 135 176 114 140 181 Dias. B.P. 60 75 90 64 76 97 65 80 96 67 80 96 Serum Chal 106 147- 181 109 143 182 103 137 179 107 144 184 stature is completed. So why not propose that weight;of ostensibly "healthy" persons of given the ideal body weight should be the average sex, ag~, and height, and it seemed reasonable weight at around 25 years, of age? "Ideal" to believe that, in part, this refle&s differences weight tables were published that were simply in skeletal or "f~ame" type. With no actual " based on the average weights of insurance basis in measurement of _,'ame," or even a policy applicants of given height at about age suggestion as to how "frame~. type should be 25 (Metropolitan 1942). These tables also objectively evaluated, the aetuaries produced attempted to allow for the fact that people a table of heights and weights corresponding vary in skeletal type. The raw insurance to that notion about frame. In effect, the company data showed great variation in body distribution of weights at .given.-,beight was $~pld~a~ I ~o ~.i~'c~I,aio~, Vol.~. XLi ,rod X£21, Al~il 1970 TIEX0007117.018

1-909. KEYS APPENDIX TABLE I page 3 Age 40-44 Age 45-49 Age 50-54 Age 55-59 COHORT VARIABLE 10 50 90 10 50 90 10 50 90 10 50 90 Rome RR Height 160 166 175 159 166 172 158 165 172 157 164 173 Rel. Wt. 90 108 12.7 90 108 129 87 106 125 82 108 126 ~ Skinfolds 14 2.6 44 15 2.7 45 13 2.5 40 13 2.6 39 Sys. B.P. 119 135 158 118 138 160 118 138 162' 122 142' 174 Dias. B.P. 70 86 102 72. 89" 106 72 89 105 78 90 103 SerumChol. 154 207 2.5? 159 2.06 2.60 159 2.09 267 158 2.04 261 Velika Height 163 171 178 163 170 178 160 "168 176 161 168 177 Krsna Rel Wt. 81 89 109 78 88 109 78 88 103 76 86 101 Z Skinfolds 10 13 23 9 13 25 9 13 22 9 12 24 Sys. B.P. 109 124 141 110 12.8 148 110 130 157 115 130 160 Dias. B.P. 69 78 90 70 80 93 70 80 96 70 80 98 Serum Chol. 121 154 191 120 157 2.01 116 159 204 126 155 207 divided into thirds ~vhich were then labeled as employees in the present study, the averages "small," "medium," and "large," or as "light," for men aged 45-54 reported from the "medium," and "heavy." That procedure, National Health Survey (1966), and the applied to life insurance applicants in their centers of the ranges of the weights for given twenties, resulted in the familiar three-column height recommended by the Metropolitan Life tables of "ideal" or "recommended" body Insurance Company (1942). In the National weight. Health Survey the number of men in the Appendix table 2 gives, for men aged 40 sample is small-only 547 men aged 45-- through 59, weights at given height as M-and confidence in the data is reduced by reported by the U.S. insurance industry, the the digit preference for even numbers of corresponding averages from the U.S. railroad inches in height. APPENDIX TABLE 2_ Average body weight at given height and age as reported by U.$. life insurance actuaries in 1912 (Association 1912; Davenport 1923) and 1959 (Society 1959); as found (smoothed) for U.S. railroad employees in the present study; as reported from the National Health Survey (1966); and weights recommended by the Metropolitan Life Insurance Company. The Iatter weights are the centers of the ranges recommended for the stated heights. HEIGHT 40-49 Years 50- 59 Years 45-54 Years 20-60 Years inches ~.. 1912 1959 U.S.Ry. 1912 1959 U.S. Ry. Nat. H.S. Metropolitan 64 143 148 153 144 149 149 157 130 65 147 152 156 149 153 158 161 134 66 150 156 159 153 157 159 162 138 67 155 161 163 157 162. . 164 166 142 68 160 165 168 162- 166 170 172 146 69 165 169 173 167 170 176 170 150 70 170 174 17o 172 175 175 181 155 71 176 178 180 178 179 180 189 159 7Z 182 183 184 184 185 184 181 164 73 188 187 189 190 189 193 185 168 "/4 195 197'- 197 197 194 195 201 173 TIEX0007117.019

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-203 APPENDIX TABLE 3 SMOKING CLASSIFICATION Laboratory of Physiological Hygiene University of Minnesota Punch Cigarettes Punch Pipe 0 Never 0 Never 1 Stopped <1 year 1 Stopped<l year 2 Stopped 1-9 years ~- Stopped 1-9 years 3 Stopped 10 or more years 3 Stopped I0 or more years 4 Now < 5/day 4 Now < 3/day 5 Now 5-9/day 5 Now 3-4/day 6 Now 10-19/day 6 Now 5-9/day 7 Now 20-29/day 7 Now 10-19/day 8 Now 30 or more/day. 8 Now 20 or more/day 9 Now only occasional 9 Now only occasional Punch Cigar Punch Formerly 0 Never 0 < 5 cigs. /day I Stopped < I year I 5-9 cigs. /day 2 Stopped I-9 years 2 I0-19 cigs. /day 3 Stopped 10 or more years 3 20-29 cigs. /day 4 Now 1/day 4 30 or more cigs. /day 5 Now 2-4/day 5 Light pipe, no cigs. 6 Now 4-8/day 6 Heavy pipe, no cigs. 7 Now 8 or more/day 7 Light cigar, no cigs. 8 8 Heavy cigar, no cigs. 9 Now, only occasional 9 Pipe + cigar, no cigs. SJ~.~ 1 ~o C~.¢a,l~ti~, V~.~. XII ,rod XLII, A~t~il 1970 TIEX 0007117.020

I--°04 KEYS C£ APPENDIX TABLE 4 U. S. RAILROAD MEN. Comparison of the distributions of blood pressure and serum cholesterol concentration of men aged 40-49 and 50-59 with the distributions of values for men of the same age in studies at Framingham, 1. Mass., Albany, N.Y. and Chicago (Paul's and Stamler's studies}. Men with coronary heart disease were excluded from all of the distributions. We are 4. grateful to Doctors William Harmel, Joseph To Doyle, Oglesby Paul and Jeremiah Stamler for making their data available and to Dr. Frederick H. Epstein for the compilation of those data. 6. Variable Item U.S. Ry. Framingharn Albany Chicago_P Chicago S 7. 40-49 50-59 40-49 50-59140-49 50-59 40-49 50-59 40-49 50-59 Systolic BP N~ all men 1227 1220 771 621 1267 42S 1252 758 580 691 " " <120 ram. 0% 17 I0 16 12 14 13 18 13 23 15 9. '° " ~<140 ram.,% 67 52 59 50 63 56 64 55 72 55 " " <160 mm.,% 91 76 88 78 89 80 88 84 94 83 I0. ~ " " <180 ram. ,% 98 93 96 92 97 92 97 96 99 94 Diastolic BP N, all men 1223 1217 771 621 1267 425 1252 758 580 691 12. ( " " < 80 ram.,% 32 24 22 22 18 20 19 17 34 27 " " < 90 ram.,% 70 59 57 56 63 57 57 54 76' 68 14. ( " " <100 mm., % 89 83 85 79 84 78 85 81 93 89 " " <II0 mm.,% 97 94 94 92 94 91 94 94 98 95 16. ~ Cholesterol N, all men 1206 1208 753 608 1223 416 1252 758 580 691 ~' <2.00 mg/d]~% 19 14 27 27 26 25 19 15 20 17 " <230 mg/d],% 44 41 56 56 53 53 40 40 45 42 18. R " <~260 mg/d],% 71 67 82 77 77 77 70 67 72 71 " <290 mg/dl~% 87 87 92 91 91 92 80 80 89 88 19. ~ 20. P 21. H - 23. D c 24. S 25. B, 27. S~ 31. ~ NOYE~ $~ I ~ Ci¢¢~d, aioa, VoiL XLI a~l XLII, ~1~il 1970 TIEX 0007117.021

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-205 INTERNATIONAL INITIAL DATA FORM (Revised May 1968) 1. Serial No. 2. Examination: Place 3. Date: Day.__ Mo. Yr. 4. Name 5. Father's Name. 6. Birth Date: Day. Month. Year 7. Occupation 8. Occupation Code (See Occupation Code Sheets) 9. Physical Activity Code Code: O = Bedridden; I = Sedentary-Light; 2 ----- Moderate; 3 = Heavy-Very Heavy. 10. Height (cm,). 11. Weight (kg.) 12. Cigarette Code 13. Pipe Code 14. Cigar Code 15. Stopped Code (See Smoking Code Sheet) 16. Marital Status ]7. No. of Children ]8. Marriage Age. ]9. Alive, or Cause of Death: Father (code) Mother (code) Code: Alive = O. Dead: Violence = 1; Infectious Disease = 2; Other = 3. 20. Present Age or Age at Death: Father (years). Mother (years). 21. Hematocrit 22. Urinalysis Code: (Protein). (Sugar) Code: None = 0; Slight = 1; Definite = 2 23. Diet Code: Code: 0 = No Special; 1 = Reducing; 2 = Diabetic; 3 = Ulcer;, 4 = Other Special 24. Sitting Height (cm.) 25. Bi-acromial Diameter (cm.). 26. Bi-cristal Diameter (cm.). 27. Skinfold (ram.): Upper Arm ?8. Subscapular 29. Upper Arm Circumference (cm.) 30. Total Vital Capacity (c.c.). 31. aA Second Expiration (c.c.). 32. 1 Second Expiration (c.c.) NOTES: $appl~t I ~ Ci~csl~ti~, Volt. XIJ ,rod XIJI, AI~I 1970 TIEX 0007117.022

1-206 KEYS "Col. No. Identifying Data Date ': I-7 MEDICAL HISTORY FORM , (Laboratory of Physiological Hygiene, University of Minnesota) , (Instructions: The examiner may feel free to use his accustomed method of history taking, but answers to all items are requested. An item left blank will be tabulated as information unknown or not determined. Therefore, if the irfformatiou is ' obtained either yes or n._9_o must be circled. Make comments at end of history, identified with item numbers. ) Name Serial No. Age Col. Col. No, PAST HISTORY No. SYSTEM REVIEW /.f no irfformation operator punches O. CARDIOVASCULAR Circle either yes or no. Is there a history of: Circle 31 Y N Postural dizziness Yes No 32 Y N Headache 33 Y N Attacks of blurred vision 1 Z Is there a history of: 34 Y N Tinnitus 8 Y N Rheumatic fever or chorea 35 Y N Syncope 9 Y N Scarlet fever 36 Y N Any shortness of breath 10 Y N Dipthe ria at all II Y N Pneumonia or Pleurisy Y N At rest 1~- Y N Chronic Bronchitis Y N Light effort 13 Y N Bronchial Asthma Y N Moderate effort 14 Y N Pulmonary Tuberculosis Y N Severe effort 15 Y N Peptic Ulcer Y N Other 16 Y N Gall bladder disease- stones 37 Y N Orthopnea 17 Y N Kidney disease- stones 18 Y N Diabetes Mellitus- glycosuria 3_8 Y N Nocturnal Dyspnea 19 Y N Thyroid disease-Goiter 39 Y N Palpitations 40 Y N Any chest discom/ort at all Z0 Y N Liver disease-Jaundice • Y N On effort 21 Y N Arthritis -Gout Y N On excitement 22 Y N Stroke Syndrome Y N After meals 23 Y N Venereal illness, specify Y N Other 24 Y N Other illness~ specify ?-5 .... 41 Y N Peripheral edema 4g Y N Urinary frequency-nocturia Has suhiect been told he had: 43 Y N Varicose veins or ~6 Y N Heart attack Phlebitis 27 Y N Heart trouble 44 Y N Claudication 28-. Y N Heart murmur 45 Y N Other 29 Y N H~gh blood pressure 46 30 IS HEART DISEASE BY HISTORY IF CARDIOVASCULAR DISEASE OR 47 Y N PRESENT or SUSPECT HYPERTENSION PRESENT OR SUSPECT 48 Y N SUSPECT ONLY PLEASE EMPLOY SUSPECT CVD HISTORY FORM AFTER COMPLETING THIS FORM. COMMENT: S~t I to Circ.~d~ai~, Fob. ~ ~ XLil, A$~il 1970

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-207 MEDICAL HISTORY FORM page 2 Col. Col. No. No. 49 60 RESPIRATORY OTHER SYSTEMS Is there a history of: Is there a history of 50 Y N Frequent involvement of significant: the chest with colds 61 Y N EENT complaint 51 Y N One or more chest illnesses 62 Y N Gastrointestinalcomplaint the past 2 years 63 Y N Genitourinary complaint 52 Y N Apart from chest illness 64 Y N Neuromuscular complaint is there daily or frequent 65 Y N Other significant conaplaint cough specify in cornrnent 53 Y N Was there formerly a FAMILY HISTORY (IN ANY daily cough, explain PARENT=GRANDPARENT, 54 Y N Is phlegm produced with SIBLING--SPECIFY WHICH) the cough 66 Y N Heart attack Y N Through the year 67 Y N Other heart trouble Y N In winter only 68 Y N High blood pressu.re Y N With chest illness only 69 Y N Strokes 55 Y N Has phlegm ever contained 70 Y N Diabetes blood 71 Y N Other~ specify 56 Y N Is there wheezing or 72-73 AGE ONSET SYMPTOMS tightness in the chest PRESPYOPIA Y N Through the year HISTORY RELIABILITY Y N Seasonal RATING Y N With chest illness only Y N Good 57 Y N Frequent hoarseness Y N Fair IS CHRONIC BRONCHO- Y N Poor PULMONARY DISEASE BY ITEMS IN RESERVE HISTORY Y N 58 Y N PRESENT OR SUSPECT Y N 59 Y N SUSPECT ONLY Y N Y N (DETAILED CHEST DIAGNOSIS AT END OF PHYSICAL EXAMINATION FORM) COMMENT: (Give year or age of illnes or surgery if pertinent. Identify remarks with Item Number. ) Examiner's Initials $e~em ! to CiecM~tio~, Volt. XLI ~i XLII, Apeil 1970 TIEX 0007117.024

1-208 KEYS PHYSICAL EXAMINATION FORM p. 1 The examiner may feel free to pursue his accustomed routine of physical examination, checking abnormalities on the right and encircling the appropriate punch number on the left. Check only if present. Comment on findings or diagnosis at end of form~ identifying the item by number. Col. No. 1-5 Study Population No. 17 0 Normal neck 6-9 Examination No. 18 1 Unequal carotid pulse Z Distended or pulsatile vein.~, Hair : upright I0 I Full growth 3 Enlarged Thyroid Z Receding forehead Diffuse Nodular~ 3 Receding forehead-bald spot Single Nodule 4 Bald dome 4 Other Eyes: Thorax and Lungs: II I Exophthalmos Measure circumference at nipple 12 l Arcus senilis level in max. inspiration. Re- 13 l Xanthelasma L R measure after max. rapid forced Pupil abnormality expiration with mouth open. 14 0 Normal fundi 19-21 Inspiratory circum, in. Fundi not well seen 22-24 Expiratory circum, in. 15 1 Arteriosclerotic Fundi 16 Hypertensive Fundi, class: 25 0 Normal chest and lung exam. 1 General narrowing only 26 1 Thorax abnormal 2- A-V nicking and/or focal Z7 1 Lung abnormal narrowing 3 Hemorrhages and/or exudates Check Abnormalities Present: 4 Papilledema Expiratory wheezes heard Check Abnormalities Present: without stethoscope __ Widened Light reflex Expiratory lag evident ~ General arterial narrowing ~ Tachypnea at rest ~ Tortuousity of arteries __ PA diameter prominent __ Copper or silver wiring ~ Anterior chest deformity A-V nicking Posterior chest or spine Focal arterial narrowing ~ deformity __ Venous engorgement ~ Diaphragm excursion decreased ._ Hemorrhage ~ Dullness to percussion L__ " Exudate- Hyperresonance L__ R__ Papilledema Abnormal breath sounds L Aneurysms Absent absolute cardiac dullness Diabetic retinitis ~ Inspiratory wheezes Lens opacity Expiratory wheezes ~ Abnormal pulsations ~ Inspiratory tales or rhonchi, Other "spe cify Expiratory rales or rhonchi, Normal ear, nose and throat ~ specify ~ Pharyngeal injection__ Abnormality, specify ~ 1970 TIEX 0007117.025

• CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-209 I~HYSICAL EXAMINATION FORM p. 2 Heart: Abdomen: 2-8 0 Normal heart 35 0 Normal abdomen Apex outside MCL__cm from 36 1 Hepatomegaly MSL--circle which interspace: Z Splenormegaly 3 4 5 6 7 3 Other, comment Abnormal apex impulse c omme nt: Pulse s : 2.? 1 Heart considered enlarged 37 0 Normal pulses Thrill palpable 38 1 Abnormal pulses found S Systolic Diastolic Radials scler.~tic Location: Possible abdominal aneurysm__ Basal Apical__ Femoral pulse diminished L. sternal border Left__ Right__ Gallop rhythm present D. pedis diminished or absent comment: Left__ Right__ 30 1 Arrhythmia present, specify P. tibial diminished or absent Tones abnormal Left__ Right__ Check tones: Distant Water hammer pulses present__ A2. Accent.__ MI accent__ 39 1 Dupuytren's contracture or A2. Dimin..__ Opening snap__ palmar thickening present Slight__ Moderate Left Murmurs: Advancedu Right__ Grade murmurs: 1Barely Audible Other: 2. Faint, 3 Moderate, 4 Loud, 5 Very Loud, 6 Loudest Possible Lower Extremities: 40 0 Normal 31 1 Systolic murmur present; 1 Elevation pallor L R Check location:Grade 1-6 1 Dependent rubor L R 1 Decreased temp. L__ R__ Precordial .1 Dependent edema L__ R__ L. sternal border 1 Varicose veins r. R Apical Pitch-quality 1 Leg ulcer L R Pulmonic 1 Other Aortic Transmission 41 1 Vascular abnormalities found Other 2 More than 1 systolic murmur__ Upper Extremities: 42 0 Normal 32 1 Diastolic murmur present; check 1 Clubbing location: Grade 1-6 1 Other, comment Precordial L. sternal border Skina Lymphatics: ~' Apical(mitral)m Pitch-quality 43 0 Normal 1 Petechaie Pulmonic__ 1 Cyanosis Other, comment Aorticu Transmission 1 Adenopathy Other 2. More than one diastolic murmur Neurmomuscular: Change on position or exercise__ 44 1 Abnormality found, comment 33 1 Murmur(s) considered organic General Appearance: 34 1 Murmur(s) considered non- 45 1 Excellent health significant 46 1 Fair health ~7 1 Poor health, comment Sttpplcmgat i to Ciraldatio~, Fol~. 2(J.d and XLll, tlpril 1970 TIEX 0007117.026

1-210 KEYS PHYSICAL EXAMINATION FORM p. 3 Blood Pressure Hypertension: Supine, right arm, at end of examina- Give clinical inapression concerning tion- 2 successive readings allowing Hg to presence of heart and/or vascular return to zero between, read to nearest Z disease on hypertensive basis. Fixed ram. mark, record 4th & 5th phase diast, criteria of BP combined with findings w/ll be applied later for comparability S D(c. hange) D(absence) with other studies. 48-56 Elevated BP without cardiovascular invol~enaent ( > 140/90) 57-65 75 I Hypertensive ~eart disease 76 1 Hypertensive vascular disease Diagnostic Impression: Checked on basis 77 I Hypertensive and coronary of history and physical heart disease 66 0 No heart disease 79 Other vascular disease 67 I Heart disease possible only Present or Suspect 68 I Heart disease diagnosed present I Peripheral arteriosclerotic dis. 69 I Coronary insufficiency Z Peripheral venous disease (angina pectoris by history) 3 Thromboangiitis 70 I History of myocardial infarct 4 Cerebral arteriosclerotic dis. 71 1 Chronic heart disease of prob- Specify: able coronary origin (1+2=5; 1+4=6; other comb. =7) 72 1 Rheumatic heart disease, specify lesion: 80 1 Chronic bronchopulmonary MI AS disease--present or suspect. MS PI Bronchial asthma AI TI Pulmonary emphysema ... 73 I Hea;t disease, unknown etiology Chronic bronchitis Bronchiectasis 74 1 Check if other disease present: Pulmonary fibrosis Pulmonary heart disease Pneumoconiosis Congenital heart disease pulmonary tuberculosis Syphilitic heart disease Other, specify: Thyrotoxic heart disease Myxedema heart disease 81 l Other significant disease-- Neurocirculatory asthenia present or suspect. Check: Other heart disease Diabetes mellitus Specify: Gall bladder disease__ Hepatic disease~ Renal disease -.. Thyroid disease COMMENT: (Identify by Item Number) Peptic ulcer Operated Arthritis Other, specify in cornrnents~ 82 1 Obesity 83 1 No significant disease present 84-85-86 Physician's Initials $~*l,~l~,n~t 1 to Circ~lasio~, Vols. XLI s~d Xl, ll, ~lo~ii 1970 TIEX 0007117.027

CORONARY HEART DISEASE IN SEVEN COUNTRIES 1-211 Minn. Form 20 International Collaborative Studies Cause of Death Coding Subject Name Area Serial No. Date of Birth: Mo Day Yr__ Date of Death: Mo Day.__Yr__ DOCUMENTATION: Autopsy Death in hospital Death not in hospital: Medical witness Lay witness Unwitnessed, found dead TYPE OF DEATH: Sudden and unexpected death: Time from onset of symptoms: less than 15 rain. 15 rain. to 2 hrs. With accompanying anginal pain Without accompanying anginal pain but characteristics of cardiac mechanism . (Other non-cardiac mechanism coded below by cause) ' Coronary event with death: Documented M.I. (code 410. 0,410.9,41Z) Possible M.I, (code 410.0,410. 9,41Z) Angina pectoris, coronary insufficiency (code 41 I) ~ Other cardiac event with death: Congestive heart failure, unspecified (code 427.0) Cardiac arrhythmia, not sudden, unspecified (code 4Z9. 9)~_~ Cardiac death, unspecified (code 4Z9. 9) . Non-cardiac death: Cause known (specify) ~ (codel ,, Cause unknown including found dead or dying and other witnessed death (codes 795,796. Z, 796.3,796.9) (specify) (code) CAUSES OF DEATH: (D.C. = Death code, 8th revision, ICDA) Principal Cause Major Contributory Causes - ~.. First Second Third , D. ~. Revise D.C. Revise D.C. Revise D.C. Revise Comments: Ss~ple~l I to Circshtlo~, Vol~. XLI ~ XLII, April 1970 TIEX 0007117.028