Document wr0oMORoXOyJ828RNYvJoeoJ3

DownloadSearchable PDFRandom document
\' \\'?re ;enice. rest in *ported 2est in recomnimals ihr to lission nental ise its clemiPrac- =om, reitted mtal h in use- for lem. Dr. edi501rch ines1ar- ici- !SI'SY ce. ial M ah r. 5- f, h Smoking and Causes of Death Among US. Veterans: 16 Years of Observation EUGENE ROGOT. MA JAMES L. MURRAY, DVM. MS TO OESCFUBE T H E RELATIONSHIP BETWEEN TOBACCO USE and mortality experience, the Public Health Sen-ice in 1954 initiated a study of a cohort of U.S. veterans. I n the first reports on this study, covering the first 2 5 yean of the cohort's mortality experience, Dorn provided evidence that cigarette smokers are subject to increased risk of dying from lung, cancer as well as from cardiovascular disease, certain respirator): diseases, ulcers, and cirrhosis of the liver ( I , ? ) . Kahn, reporting on the mortality experience of this cohort after 8Vz years of observation, through December 1962 (31, had results which were similar to, and reinforced, Dorn's. T h e cohort has since been followed for a total of 16 years, through December 1969, and its general mortality has been reported over that period (4-6). We are now able to report on the specific causes of death of members of the cohort during this 16-year period. Materials and .Methods T h e study population, the study design, and other re- M r . Rogot is a statutician, .Epidemiology Branch. National Heart, Lung, and Blood Institute, Federal Bldg., Rm. 2C-08, Bethesda, M d . 20205, and Dr. Murray is an epidemiologist, Biometry Branch, National Cancer Institute. Tearsheet rcquerti to Mr. Rogot. lated details have been described in previous reports (1-31. T h e initial study population consisted of 293,958 U.S.veterans 31-84 years of age who held U.S. Gov- ernment life insurance policies in Deceniber 1953. Beginning in January 1954. questionnaires on sniokins habits were mailed to these policyholders, of whom about 200.040 (68 percent) responded. These respondents are called the "1954 cohort." Reginninc in January 1957, a second questionnaire, similar to the first, was mailed to those not responding in 1954. I t elicited about 5O.ooO additonal replies, raising the response rate to 85 percent. These respondents are called the "195i cohort." T h e remaining persons are called the 4' n*o reply group." All the smoking classifications used in our current report are taken from these questionnaires, and for the sake of convenience we considered them to have been as of Januar). 1 both for the 1954 and the 1957 cohort. T h e numben of persons in each cohort by age were as follow: Age group 1954 cohort 31-54 ......... 55-64 65-74 754+ ......... ......... ......... 34,473 135,820 25,002 -1,525 Tom1 ... 198,820 1957 cohort 8,441 26.579 13,583 523 49,226 S o reply group 8,417 3 1,468 5,603 424 45,912 May-Juno 1080. VoI. 91. No. 3 513 Almost d l the policyholders were white males, drawn mainly from the middle and upper socioeconomic classes. .\lost were veterans of \Vorld tVar 1. T h e Veterans .-\drninistratron provided the names of all po1ic)iiolders known to have died in the study period (Jan. I , I95C-Dec. 31, 1969). T h e VX s t a f f made specnl searches to ascertain deaths for about 75,000 penons !;.hose insurance policies had lapsed during the period. The overall mortality followup of the fact of death and the year of death is considered to be almost LOO percent complete. FuH details of the various followup procedures have appeared in the previous reports ( 1 3 ) . With respect to the cause of death, specid searches were made by the American Cancer Society for the 34,644 death certificates that were needed to complete the latter part of the 16-year followup. T h e society was successful in collecting dl but 2,541 death certifi- cates. Causes of death were coded according to the Seventh Revision of the International Classification of Diseases, 1955. The underlying cause of death was used. Overall, the iinderlyins cause was available for 37.6 percent of 311 deaths ( b b l e 1) . Table 1. Number of oersons, number of deaths. and number of death certificates found and not found. by cohort, U.S. veterans study. 1954-69 from 1951 eohorr 1957 NQ fODlV cohort group Total Number of persons . . Number of deaths . . Oeath certificates: Found . . . Not found . . . . . .. . . . . . . . . . Percent found . ... 198.820 49 226 69.858 16.877 68.446 16,521 1.412 356 98.0 97.9 45,912 20.828 20.055 773 96.3 293.958 107,563 105.022 2.541 97.6 Table 2. Observed deaths, expected deaths, and mortality ratios ( O s E ) for current cigarette smokers (category 2. .) and ex-cigarette smokers (category 1, ,), by cause of death, U.S. velerans study, 195469 Exzlgaretce m a k e n 1 Cause of deata [71h R o v l s i ~IC0 numben) All causes ....... . . . .. . ... .. . . . ... . . . .. . ........... 36,143 20.857 1.73 14,487 11.962 1.21 All cardiovascular diseases (330-334. 400468) . , .Coronary heart disease (420) . . . .. . ..... .. ... . .Rheumatic fever (4OWO2. 410-416) ... . . . . .. .. ..Hypertensive heart disease (440-143) . . . .. .... Myocardial degeneration (422) . . . . . . . . ... . . , . . .Other heart disease (421, Cor pulmonale (434.4) .4.3.0.-.4.3.4.) ....... .... . . . . .. .. .. .. .. .. . . . . . . . .., ... .. . ... ... ... ... , . .. ._. ._........ . ... ...... . . . . ..._.... ..,_... .........,...... . 21.413 13.845 301 909 700 525 44 13.572 8.787 272 590 427 316 8 1.58 1.58 1.11 1.54 1.64 1.66 5.57 9.027 5.823 154 389 306 209 11 7.865 5,028 147 333 255 186 5 1.15 1.16 1.04 1.17 1.20 1.13 (') Stroke (330-334) .......................................... Peripheral vascular disease (453) . . . . . .. . Aortic aneurysm (451) Other aneurysm (452) Phlebitis and pulmonary . . ........_........... . ....._.. embolism (_46_3-_46..6..)........... .. .. . . Other circulatory disease (45d-456, 460-462, 467468) . . . . . .. .. .. ..... .. .. .. .. .. . . . Hypertension (444-447) .................................... 2.728 20 900 16 214 143 198 2.075 6 172 2 175 55 134 1.32 3.52 5.23 ('1 1.22 2.59 1.47 1.279 2 253 6 109 29 83 1,254 3 98 1 102 32 77 1.02 (7 2.58 (7 1.07 .91 1.08 Hypertensive disease (440-147) General arteriosclerosis (450) . . .Parkinson's disease (350) . . . ,, .Benign tumors (210-239) . . , .. .. .. .. .. ... ... .,, ... .... ., .. .. .. ,, .. ...... .. . . , . .. .. ............ ... ... .., . .. . . . . . . . . . . . . .. .. .. .. .. ,. . . . . .. .. . . .. 1,107 870 57 73 724 1.53 554 1.57 180 .32 33 2.23 472 374 74 27 410 1.15 343 1.09 106 .70 18 1.52 Ulcer of duodenum (541) . . . . . . . ..._........ . . . . . . ........ 216 60 3.61 55 34 1.64 Ulcer of stomach and gastrojeiunum (540. 542) . . . . . . . . . .. ... . . 149 32 4.60 54 19 2.90 Aneurysm of aorta and other cardiovascular syphilis (022, 023) . . 23 7 3.20 8 4 ('1 Ne6p0h0r-it6is0.3)ne.p.h.r.o.s.is. .a.n.d. . o. .th.e. .r .k.i.d.n.e.y. .d.i.s.e.a.s.e.s. .(.5.9.0.-5.9.4.,. . . . . . 349 261 1.34 155 149 1.04 Cirrhosis of fiver (581) ... .. .. Choletithiasis and cholecystitis .... .... . (584-586) . .. . . . . . , . , . . .... . . .. .. .. .. . . . . .. . . . . . . . . 404 72 150 2.69 60 1.20 114 33 81 1.41 38 -87 Diabetes (260) .......................... .Benign prostatic diseases (610-614) . . ... Pulmonary tuberculosis (001. 002) . . . . . . ............................................,....... ., . . , . 215 51 81 221 .97 65 .79 36 2.27 100 39 36 130 .77 40 .96 21 1.72 :Only ar-cigaretm $makers m e JloOOed smoking c i g a r e i i ~ s lor raasoni other than Qny$icmns.oraers. 2 O+E values are based an exoected numoerr 10 2 decimal places. Raiior are not m o w n for ooserved walues of less than 20 NOTE: For exolanalion of smoking calegories. sae 1. 215. I to the ation of 3th \\as .raiiable I d nurr cohor - -Tarat 293.958 107.563 105.022 2.541 -97.6 ) and - -1 -O+E 1 -1.21 1.15 1.16 1.04 1.17 1.20 1.13 ('1 1.02 ('1 2.58 ('1 1.07 .91 1.08 1.15 I.09 .70 .52 .64 .90 . ('1 .04 41 87 77 96 -72 now ) 2--r\eg~lar smoker now or current smoker !that is- as of 19j-I for the 1954 cohort or 1 3 7 for the 1957 cohort ! 6-Sever smoked or es-smoker, but iinknoun which ;-sever slnoked or current smoker. but unknown which 8-Current or es-mokcr, but unkno1s.n \\,hich 9-Uiiknowi tvi-iethcr ne\.er siltoked. es-sllloker: or current smokcr (Categclrie3 6-9 represent the differcnt kinds o i unknowns. i Current 3rnokint catcgories Sonsinoker (lifetime categories 0, 1: and 6) Regular smoker (lifetime category 2 ) Unknown if nonsmokcr or regular smoker (lifetime categories 7-9) For ease of relerence: the Iifetinie smokinq catesories for the three forms of tobacco studied are shown as 3-digit nutliben, Luith the first digit representing cigarettes: the second cigars, and the third pipes. Thus, for esample. 126 111eans those persons who a t the time they answered the smoking questionnaire were escigarette smokers but were currently smoking cigars Table 2. (Continued) Cause of death (7fh Revision ICD numbers) Cigarere smokers Observed dearns Expected deatns 0 - E3 Ex-cigarette smokers 1 Observed dmlhS Expecred deaths 0 - E2 Influenza and pneumonia (480-481.490-493) . . . . . . . . . . . . Pulmonary fibrosis, bronchiectasis, atelectasis (525-527.0) Emphysema (527.1) . .. . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . Bronchitis (500-502) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bronchitis and emphysema (50 Asthma (241) . .. . . . ... . . .. . ...........I...... . ...........I.... All respiratory diseases (241,480481,490-502, 520-527) Suicide (E97bE979) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Accidents and violence (E80bE965,E980-E999) . . . . . . . . . . . . . Suicide, accidents. end violence (E800499 All cancers (140-207) . . .. . .. .:. . . .Cancer Cancer of of pbhuacrcyanlxc(a1v4it5y-1(4184)0-.1.4.4.).... . . 460 144 1,201 163 1,364 90 2.107 376 780 1.156 7,608 110 92 Cancer Cancer Cancer of of of .esinstotoempshatiacngheus(s1(1(551125)-01).5..3..)............... . . . ........ .... .. .. . . . . . . . . . ..... . .... . ... .. ...... ..... ... . . ..... I . . Cancer Cancer of of rectum (154)... liver and biliary .......... passages ........ (155) . . . . .. .......... .... ..... . . ...... . . . Cancer Cancer fo of plaarnycnrxea(s16(11)57. ). ................. . . ................... . . . . ......... .... .... . . ... ... , .. Cancer of lung and bronchus (162.1. 162.8. 163) , . . . . . _. . . . . . . . . .Cancer of prostate (177). . . . . . .. .. . . . . . . . . . . . .. . . . ... . .Cancer of kidney (180) . ..... ........ .. . _ . . . . . . . . . _ . .. . . .Cancer of bladder and other urinary organs (181). . . . . . . . ... . . Malignant lymphomas (200-201. 203. 206) . .... . . . . .. ...... ... 156 390 662 239 176 459 94 2.609 660 175 326 370 -Leu.-k..*e_m*"i..a Cancer of ..-,&.,(204,,207) .** k.i*Ly-L.. brain (193) ... ... . . ...,..-_..... .... ... .. . . . . .. . . . . . . ...... . ... . . .. .. ..... ..... All other cancers ......................................... 111 defined conditions (780-795) . . .. .. . . . . . . , . .. . ... . All other diseases ........................................ Unknown (no death certificate) . . . . . ... . . . . .. . .. . . . . . . . . . ... 333 160 597 372 948 849 259 48 81 32 113 27 488 282 633 915 3.590 26 7 1.78 3.02 14.82 5.11 12.07 3.28 4.31 1.33 1.23 1.26 2.12 4.22 14.05 24 6.43 257 1.52 597 1.11 215 1.11 75 2.33 256 1.79 8 11.49 231 11.28 504 1.31 124 1.41 151 2.16 347 1.07 207 152 407 220 576 '390 1.61 1.05 1.47 1.69 1.65 2.17 163 59 313 45 358 29 628 154 329 483 2,816 24 11 35 158 431 131 78 170 22 517 360 82 126 188 175 85 223 159 396 285 165 27 50 19 69 15 302 136 326 461 2.025 14 4 15 143 334 116 44 145 5 130 304 68 90 789 119 79 227 130 322 21 7 .99 2.1 5 6.22 2.43 5.20 1.95 2.98 1.14 1.01 1.05 1.39 1.67 ('1 2.41 1.10 1.29 1.13 1.79 1.17 4.78 3.97 1.18 1.21 1.41 .99 1.47 1.08 .98 1.22 1.23 1.31 3 Only extigarette smokers who stopped smoking cigarellel for reasons other Ihan pnyficidns' orden. 2 0 - Evalues are bamed an expacted number, to 2 decimal places. RatlOs are no1 shown for observed vllUeS 01 le111 Ihan 20. NOTE. For explanation of smoking calagorirs, see 0 . 215. Yay-Juna 1980, Vol. 95. No. 3 215 and either had never smoked a pipe or were ex-pipe smokers. Cateqory 200 denotes persons who a t the time of the questionnaire were current cigarette smokers but had never jmokcd cipars or a pipe. Category 2.. denotes all current ciqarette smokers and includes catesones ;?00-209, 210. 11 1, and so forth. T h e method of analvsis used in this report is similar to that described in an earlier one (6).T h e probability of dying within 1 year by specified cause of death ( y ) and single year of aqe (x) is first estimated for the nonsmokers (000 catesory!. .A given year of age attained at any time in the 16-year period is considered. +h person who is x years old in 1954 is counted as x 1 year old in 1355, x f 2 years old in 1956, and so forth, provided he survives. X person who dies is counted as entering each year from 19% up through the year of death. The following table may sewe to illustrate. Age at:ainrd 1954 31 32 33 34 ............................................ 31 32 33 3+ 81 ........... 81 Age in- 1953 32 33 31 1956 33 3.1 1957 ... 1969 31 46 47 .(8 49 85 86 87 ... 99 T h e numbers of persons entering a e v e n age (x) are pooled (along a diiqonal) and constitute the total persons at risk for the given q e (.V.t-). The numbers of deaths from a specified cause ( y ) for a given age are collected in the same way. Dividing y by .Vx provides the I-year probability of death from that cause for that age (q r v ) . T h e values of qsv for nonsmokers were obtained for each year of age ( x ) from 31 to 99 for each cause of death ( y ) as described. By applying the probability qZv to the number of persons at risk for a given smoking group ( N x ), an "expected" number of deaths (Ery) was obtained. T h e expected values for the specified cause y were summed over all 3ga, separately by cohort, and then pooled. A mortality ratio of the causespecific mortality risks for the given smokinq . p u p as compared with nonsmokers was then formed by divid- ing the observed number of deaths by the expected number. Results T h e number of persons in each cohort, the number of deaths in each cohort, and the number of de`aths for which death certificates were found or not found are shown in table 1. T h e 3 percent of deaths for which certificates were not found \&`eretreated as unknown with respect to the cause of death 3nd discounted from the analyses of specific causes of death. Because of the smdI number of these deaths, their effect upon mortality ratios is considered ngligible. Table 2 provides a summary of the mortality experience for selected causes of death over the entire study period in terms of mortality ratios. T h a t is, the mortality ratios both for current (3t time of the questionnaire'! cigarette smokers and for ex-cigarette smoke n who had Stopped smokinq for reasons other than physicians' orden were compared with the mortality ratios for the nonsmokers. Searlv all the mortality ratios for the cigarette smokers kvere greater than unity. The highest ratios (those greater than 5) were observed for cor pulmonale (5.57), aortic aneuq-sm (5.23), emphysema and bronchitis ( 14.821, cancer of pharynx ( 14.05), cancer of esophagus (6.+3),cancer of 1aVn.x ( 1 1-49), and cancer of lung and bronchus ( 11.281. Mortality ratios for ex-cigarette smokers for these causes were much lower than those for current cigarette smokers, dthough in all instances the mortality ratios were greater than 1.0. In terms of obsemed deaths minus expected ( 0 - E ) for 311 causes, an excess of 15,286 deaths was associated with current cigarette smoking. By cause, the hrqest 0--E was for all cardiovascular diseases, which accounted for nearly 8,000 excess deaths. About 5,000 of these excess deaths were from coronary heart diseze, 728 nere from aortic aneuqxn, 633 were from stroke, 383 were from hypertensive disease. and 316 were from general ~rteriosclerosis. About 4,000 excess deaths associated with current cigarette smoking were noted for a11 cancers. More than half of this excess (2,378 deaths) was from lung cancer. T h e number of excess -deaths from all respiratory diseases was 1,619, with bronchitis and emphysema accounting for 1,251. For ex-smoken, the mortality ratios for nearly all cardiovascular causes of death shown were close to unity (1.20 or less). T h e only notahle exception ivas aortic aneurysm with a mortality ratio of 2.58. Few conditions in table 2 exhibit mortality ratios of less than I.0. T h e only exception of note is Parkinson's disease with a n 0--E o10.32 for current cigarette smokers and 0.70 for ex-cigarette smokers. bfonality ratios for selected causes of death are shown in table 3 for "pure" cigarette smokers (cateqory 200), "pure" cigar smokers (020), and "pure" pipe smokers (002). T h e pure cigarette smokers had much hiqher values than the pure cifsar or pure pipe smokers for each cause shown. i\lso, for each cause, mortality ratios were somewhat higher for the 200 qroiip than for the 2.. group (see table 1); for ail causes, ratios fl t. 3. 2; c 5: \\ 3. b \I fc i n CC fi sr fc ci Oi 0 b. sr dt 31 cc ar e: TE C; CZ cc Str ~ Inf Ao Re 8rr LUf 1 .` of the rtaiity v esmrire F. the quesrndthan tality tality than \cere n'sm 'r of ncer zhus for rent Jity -E\ ,ted Test ac1 of ise. !he. 0111 mt re "S -ana 111 to as IS 1- e e Y 0 1 i [or ti:c 20(! croup \\?re 1.86 conipared \\it11 I .;j for lilt 2 . . &TOLl1J. For J I I ~ ` 3 i i . co~f ~dt*zili. p r c cirar sniokers showed an esccrs iiiortdity oi 1.5 percent romparrd \vitli non. C I I I O ~ C : >1 Ol111 c a t c z o n : pure pipe sniokers showcd an pscccs rnortaliry oi 8 pcrrent rompared with non.tiioLtlj. For cirrar sitiokr.rs. the highest mortality ratios .wcre fdr aor:ic a t i r ~ i ~ s i:n2.041. lung canrer f 1.66! and dl raiirers 1 1 . 3 2 , . .\Iortality ratios were near or Ibcio\\. unity for stroke l . O T ! , influenza and pneumonia (!.82 . and IeSi'iKltGQ'diseases !0.84!. For pipe smokers. the highest mortality ratios were for bronchitis anti emphysema (2.53\ lung cancer ~ (2.14 aortic aneuFsni (2.07" respirator). diseases ~ ( 1.44#.and all cancers ( 1.39 1. llortality ratios were near or below I .O lor ail cardivvascuiar dkeases ( 1.04 . corona? heart disease ( 1.011. stroke (0.99: and in- fluenza and pneumonia (0.97;. hlortality ratios according to the amount of tobacco smoked are shown in figures 1 and 2 for selected causes for both current cigarette smokers and those escigarette smokers who stopped smoking for reasons other than physicians` orders. T h e obsemed numbers of deaths are given in tables 4 and 5. A d e a r gradient by the amount of tobacco smoked is evident for current sl:iokers and. also, for ex-smokers. For each cause of death, the mortality ratio varies directly with the aniount snioked. \$`hen the mortality ratios for the ex-smokers are compared with those for the current smokers (figs. 1 and 2!, the cs-smokers have a lower mortality ratio at each level of smoking for each cause shown. In figure 3: mortality ratios for es-cigarette smoken wlio had stopped smokinx for reasons other than physicians' orders are shown for celected causes of death, by the nuinher of yean of smoking cessation. (The obs e n e d numbers of deaths for this %roup u e given in table 6.) Initially. we dilided these es-cigarette stiiokers into 5-year groups according to the length of time that they had stopped smoking cigarettes as of the beginning of folloivup. IVith each year of follo\vup, the length of time that smoking had been discontinued was increased by 1 year, and the excigarette smokers were classified into one of the five categories shown in figure 3. T h e approsimate mid- point of the 5-year intenal was used to measure the length of time the person had stopped smoking. These midpoints were -4 = 2, B = 7 , C = 12. and D = 17. For esample, an ex-smoker in the 19% cohort who had stopped smoking less than 5 yean as of the start of the study would be counted in the A p u p for the first 3 years of follo\wp. in the B group for the next 5 years of followup, in the C group for the next 5 years, and in the D group for the last 3 yeus-provided the person sun-ived. Sitnilarly, a person in the B group at the s t a r t would be counted in the B group for the first 3 yean of followup, in the C group for the next 5 years, and so forth. Persons in the E group at the start of the study were counted in the E y o u p through- out. T h e assumption was made that a person who was an es-smoker at :he beginning of the study would ;emain an es-smoker throughout the follo\\*upperiod. For all causes of death, a consistent inverse relation- ship esisted between the years that the persons had Table 3. Observed deaths, expected deaths. and mortality ratios (0-E) for pure cigarette smokers. pure cigar smokers, and pure pipe smokers for selected causes of death. U.S. veterans study, 1954-69 Cause 01 death (7th Revision ICD numbers) Pura c8qaretrC smokers (cafegory 2001 Pure cigar smokers (caregory 020) Pure pipe smokers (caregory OO2l O b s m e d Expectea Observed EXWcted Observed ExDecfed dettlhs dearns O A E 1 dearns oaalhs 0 - E1 deaths a w n s 0 - E1 All causes ....................... 15.091 8,112 1.86 2,653 2.302 1.15 1.545 1,432 1.08 Cardiovascular diseases (330-334. Cancers. all 400-468) . . . . . sites (140-207) ............................ Coronary heart disease (420) ............ Stroke (330-334) ....................... Influenza and pneumonia (480481. 490-4931 Aortic aneurysm (451) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. Respiratory 490-502. diseases 520-527) (.2.4..1.,.4..8.0.4..8.1.,...... Bronchitis and emphysema (500-502. 527.1). Lung cancer (162.1. 162.8. 163) . . . . . . . . . . 8,920 3,138 5,740 1,172 200 359 879 568 1.095 5.257 1,401 3,414 796 96 68 185 43 91 1.70 2.24 1.68 1.47 2.08 5.28 4.75 13.13 12.06 1.681 510 1,077 267 28 38 51 10 41 1,522 386 965 249 34 19 61 14 25 1.10 1.32 1.12 1.07 .82 2.04 .84 (*) 1.66 984 307 606 157 22 24 57 22 - 32 948 1.04 237 1.29 596 1.02 159 .99 23 .97 12 2.07 39 1.44, 9 2.53 15 2.14 1 eased on e x p e c t e d number to 2 decimal DIaces. * Ratio not shown lor obsened Values 01 leas than 20. NOTE: For eaDlanalion 01 srnokmg Calegorter. fee p. 215. MapJune 19110. Vol. 95, No. 3 217 .! I stopped sniokinq 2nd the rnonaliry ratio. Ex-smokers pattern for coronary heart disease was similar, w i t h who had stopped smoking for 20 or more yean had the E group having a mortality ratio of 1.05. For 2 mortality ratio iairlv close to unity (1.091. stroke. the pattern differed in that all ex-smoker groups T h e pattern oi mortaiirv for dl cardiovascular except B had mortality ratios close to I-ranging diseases was similar to that for a11 causes: the mortality from 0.96 for the C group to 1.03 for the D group. m i 0 for the E group was close to unity (1.04). T h e .For aortic aneurysm, the pattern appeared to be inverse Figure I. Morrality rarios for selected causes of death with mortality ratios of less than 5. for current cigarette smokers (category 2. ) and ex-cigarerte smokers (category 1. .). bynurnber of Cigarettes Smoked daily. U S . veterans study, 195469. figure 2. Mortality ratios for selected causes of death with mortality ratios of less than 25. for current cigarette smokers (category 2. .) and ex-cigarette smokers (category 1. .). by number of cigarenes smoked daily, U.S. veterans study, 1954169. Ex-smokers' 0Current smokers 129 1 d2.03 ..- 1 582.29 All causes Ail cardiovascular diseases (330-3344,00-468) 1.55 1 Respuarory diseases (241,480-481, 490-502. 520-5271 4.06 7.24 1.95 Cancers, all sites (140-207) . Coronary heart disease (420) Bronchitis and emphysema (500.502, 527.1) n23.70 Stroke (330-334) ld125 I Influenza and pneumonia (480-481.490-4931 Less than10 per day 10-20 21-39 perday perday Cigarettes smoked 40or more per day 2 Oncv si.c'qarerrermorers wna itaopaa smoung lor -cpion$ otner man onysaclans oraers. Lung cancer (162.1,162.8,163) Less ChanlO per day 10-20 perday 2149 40 or perday mare per day Cigarettes smoked ' Onlv + i - q a r c t ? e smokem m a stoooea r m o ~ m 9tor reaJons other than onysn. cuns araao. 210 Public Health Report. - .. ~ . -e-.'- * .. I . with 5 For groups inpg group. nvene h with nokers L by 154-69. 3s well, but the mortality ratio for the E yroup ICY 1.60. For all cancers. the pattern was also inversc, but it dicered from that for the cardiovascular diseases in that the A group eshibited a higher rnonality ratio than the current cigarette snioliers. T h e mortality ratio for the E y o u p H'Y 1.27. For lung cancer. the A group had a higher m0Kdity ratio than current cigarette smokers (18.83 compared with 11.28). However, strong reductions in the mor- tality ratio were obscned for the other c r o u p as the years of smokin? cessation increased. The final value was 2.10 for the group that had stopped r~nokingfor 20 or more yean. It-ith the esceprion of :he mortality ratio for the A group! respirator?. diseases as a whoie showed a consistent downward gradient with increases in the number of yean of smoking cessation. T h e E p u p had a mortality ratio of 1.34. For the B group, how- ever, the mortality ratio for these diseases was greater 9 5 ii' -'0 Figure 3. Mortality ratios for selected causes of death for current cigarette smokers (category 2. .) and ex-cigarette smokers (category 1. .), by number of years of smoking cessation. US. veterans study, 1954-69. 0Current cigarette smokers 0Exagarette smokers A: stopped less man 5 years B: stopped 5-9 years C: StopW 10-14 ye&% D: stopped 15-19 years E: stopped 20 or more years AU causes 1.o 0 2.0 r ABCDE ~ o r t l caneurysm (4511 Respvatory diseases (241. 480-481. 490-502. 520-527) 05 [ 5 0 15 10 Bronchitis and emphysema (500-502.527.1) 5 0 20 r A'BCDE A'B C 0 E All cardiovascular diseases (330-334. 400-4681 -t ABCDE Cancers, aH sltes (140-207) Coronary heart disease (4201 ABCDE 2.0 r .i 0 I I Lung cancer (162.1. 162.8. 163) I,- Stroke (330-334) 1 i-TtiB C D E A Influenza and pneumonm (480-481. 490-493) A'BCDE A- B* C D E Way-Juna 1980. Vol. 95. No. 3 219 Table 2. Observed deaths from selected causes among current cigarette smokers (category 2. .), by number of cigarettes smoked daily, U.S. veterans study, 195449 Cause 01 death [71h Revisron ICD numbers) Cigarecles smored daily <IO 1040 11-59 40s. All causes 4,792 17.457 11,514 2,380 Cardiovascular diseases (330-334.4 0 M 6 8 ) . . 2.995 Cancers, all sites (140-207) . . . . . . . . . . 891 Coronary heart disease (420) . . . . . . . . . . . . . . 1,932 Stroke (330-334) . . . . . 435 lnfluenza ana pneumonia (480481,490-4931 . . Aortic aneurysm (451) . 66 69 Respiratory diseases (241,480-481. 490- 493, 500-502, 520-527) . . . . . . . . . . 211 Bronchitis and emphysema (500- 502, 527.1) . . . . . . . . 105 Lu1n6g2.c8a.nc1e6r3()16.2..1.,. . . . . 156 10.594 3.502 6.859 1,321 230 474 1,007 645 1.118 6.533 1.291 2,632 583 4,216 838 816 155 137 27 297 60 731 158 501 1,063 113 272 NOTE: For explanation 01 smoking categories. see p. 215. Table 5. Observed deaths from selected causes among ex-cigarette smokers (category 1 . .), by maximum number of cigarettes they had smoked daily, U.S. veterans study, 1954-69 Cause Of d..th (7th Revwon I C 0 numaen) . -Harlmum number of cioarenes smoked dally <rO i b m 2149 40s. All causes . . . . . . . 3.068 6,651 3.436 1,332 Cardiovascular diseases (330-334,400-468) .... 1,994 Cancers, all sites (14W207) ........... 561 Coronary heart disease (420) . . . . . . . . . . . . . . . . 1.256 Stroke (330-334) ....... 303 Influenza and pneumonia (480481. 490493) . . . Aortic aneurysm (451) . . . Respiratory diseases (241, 37 38 480481,490-493, 500-502. 520-527) .... Bronchitis and emphysema (500-502 527.1) ...... 82 29 Lung cancer (162.1. 162.8. 163) ................ 39 4,753 1.283 2,684 581 78 123 296 172 228 2,092 707 1,393 272 33 64 176 107 173 788 265 490 123 15 28 74 50 n NOTE: Include only deaths of ex-cigaretts smokers 'who aroppsd smoking cigarsttea for reasons other lhdn physicians' orden. For explanation of smoktng categories. see page 215. Table 6. Observed deaths from selec!ed causes among ex-cigarette smokers (category 1. .) by years of smoking cessation, US. veterans study, 1954-69 Cause ot dealh (71h Revision I C 0 numbers) <5 All causes ............................................ 384 cardiovascular diseases (330-334,400468) ..................... 224 Cancers, all sites (143-207) .................................. 99 Coronary heart disease (420) .................................. 150 Stroke (330-334) ............................................ 16 Influenza and pneumonia (480481.490-493) .................... 0 Aortic aneurysm (451) ....................................... 2 Respiratory disease 241,480481,490493. 500-502, 520-527) ..... io Bronchitis and emphysema (500-502. 527.1) .................... 8 Lung cancer (162.1,162.8. 163) ................................ 47 Years at Smoking cessation E9 10-14 15-19 20+ Unknown 1,441 892 305 599 98 8 34 82 58 86 2,445 1.488 488 997 167 25 66 140 101 100 2,767 1.707 5t5 1,101 243 36 53 140 75 115 6,049 3,832 1,160 2,418 604 83 BO 209 90 123 1,401 a84 249 558 151 11 1s 47 26 46 NOTE: Includes only deaths of excigarette smokers who s~oppsdsmoking cigarettes lor reasons other thin physiclanr' orden. For explanation of smoking calegorter. see page 2 1 5 . than for the current cigarette smoken (5.04 versus .4.3 1 ) For bronchitis and emphysema, the mortality ratios were somewhat similar to those for lung cancer, in that those 'Or the *49 B7 and .Foups were 'lose to or higher than the d o for current cigarette smokers. For the E Group, the mortality ratio was '2.64. For influenza and pneumonia, the mortality ratios for the ex-smokers were all much less than the 1.77 value for the current cigarette smokers. T h e C agvup had the highest value, 135; the D p u p , with 1.11, and the E group, with 0.91, were near unity. ~i~~~ In p ~ e m l ,the results for the longer followup period (16 yean) were in close aqeement with those of Kahn ( 3 ) for the First 8% yean of the study, as the following mortality ratios for current cigarette smokers for jelected causes show: 220 Publlc Heailh Rapofla ~ ~ , -1.332 7aa 265 490 123 15 20 74 50 -77 mea xpla- ssa- - -i o w n -01 84 49 58 51 I1 18 17 !6 -16 Of "P 1, id in `3 :d .\lortal:!y rafior ,for current rrnolicrr .Vorfdify ratio3 /or rx-Jmokers Cdurt of dcafh .-\I1 CPUICI . . . . . . . . . . . . . Cardiovascular diseases . . Cancer,. all rites . . . . . . . . Coronary heart di.casc . . . Stroke . . . . . . . . . . . . . . . . (nflticnza' and pneumonia .-\ortic ancun.rrn ........ Parkinson`s disease . . . . . . .Bronchitis and 1.ung cancer .e.m.p.h.y.s.e.m.a. a!.: yrars 1.71 1.62 2.08 1.61 I .+o 1.53 5.15 0.93 8.65 10.88 , 16 years 1.73 1.58 2.12 1.58 1.32 1.78 5.23 0.31 12.07 11.28 Causr 01 denfh 8!,i pars .411 caiizes ................. CCCStoaarrrnoodcknireorarsra..y.sac.1hu.1e.laa.b.rri.tte.dr.di.i~.s.e.e.a.a..as..ee..s..................... ABIPLnarouforrnlkntguiiccnehcsniaaotuninnsce`aseuanrrdndyids.mep.ean.mse..epu..hm...y...os...ne...mi.a..a...................... 1.77 1.21 1.49 1.21 1.07 .93 2.75 36 7.64 4.71 1 6 ?cars 1.21 1.15 1.39 1.16 1.02 .99 2.58 .70 5.20 3.97 The agreement between the data for the two periods is reiiiarkably good? considering that there were differences in the methods used to derive mortality ratios as well as minor differences in the classification of some categories. T h e one difference worth noting is that for bronchitis and emphysema, for which the mortality ratio was 8.65 for the fint 8fi-year period compared with 12.07 for the entire period. This difference suggests that mortaiity ratios for regular cigarette smokers have been increasing over time. A further comparison of the bronchitis and emphysema mortality ratios for the two periods by the number of cigarettes smoked yielded the following results: Mortality ratios for bronchitis and emphysema Cigarettes smoked daily 8% years L241e100s--so32r90thma..no..r..eIO............................................................. 4.14 8.73 11.13 15.02 I 6 years 4.84 11.23 17.45 2 1.98 T h e preceding comparisons suggest that mortality ratios have been increasing over time at all levels of smoking. with the most pronounced increase being evident for persons smoking 21 or more cigarettes a day. This relationship, however, was not found for lung cancer, nor for aortic aneunsm, two diseases that are also closely associated with cigarette smoking: hfortality ratios for lung cancer Cigarettes smoked daily 8% years L241100e-s-2so30r9thma..on..r..e1..0....................... 4.76 9.05 16.93 23.63 16 years 3.89 9.63 16.70 23.70 Mortnlity ratior lor aortic aneurysm 8% years 2.16 5.58 6.55 7.21 16 years 2.29 5.46 6.36 7.18 In general, close agreement was obtained over the two periods of followup in the mortality ratios for excigarette smokers who had stopped smoking for reasons other than physicians' orders. .4lthough agreemcnt was closc. mortality ratios tended to be lower for the 16-year period than for the 8)'~year period. This \\.as true for a11 causes shown except those for which the mortality ratios were less than 1.0. This result is consistent with the invene relation between the number of years of smoking cessation and the mortaiity ratio (fig. 3 ) , Our study showed that in the case of some diseases, the mortality risk for ex-cigarette smokers who stopped smoking for reasons other than physicians' orders returned to normal [that is, to nonsmoker levels) alniost immediately after cessation of smoking, whereas for other diseases the return to normal was more gradual. T h e fint group includes stroke and the combined category of influenza and pneumonia: the second group includes cardiovascular diseases as a whole and coronary heart disease. For still other diseases, althoush the mortality ratios declined with the length of time smoking was discontinued, substantial excess risks remained even after 20 years of cessation. In this third group are aortic aneurysm. bronchitis and emphysema, and lung canrer-diseases with very high mortality ratios for the current cigarette smokers. As stated earlier, all the smoking classifications in this report were taken from the original questionnaires and for this analysis are considered to be unchanged over the whole followup period. Short of a resurvey, there is of c o u n e no way of knowing just how riiany people actually changed their smoking habits and in what ways. I n view of the age distributions of the starting cohorts (for esample, 87 percent of all the nonsmoken were 55 years or older at the outset of the study), it seems reasonable to assume that relatively few nonsmokers began smoking during the course of the study. It is probable that a number of "current" smokers stopped smoking during the course of the study and also that a number of "es-smokers" took up smokins again. Based on these assumptions, the mortality ratios obsewed for current cigarette smokers would tend to be too low, and those for es-cigarette smokers would tend to be too high. Thus, the ratio of 1.73 for current cigarette smokers is an understatement, whereas for Uay-Juna 19110. Vol. 95.. No. 7 221 , ,. I .d-- ex-ciqarette vnokers, the mortality mtio of 1.21 is an over~tatenient. In other words, the mortality risk associated with smokinq is greater than indicated by our statistics, m d the benefits of stopping must also he greater than those shown. The one disease reported by Kahn for which the mortality ratio for ciqarette smokers \\asless chan 1.O was Parkinson's disease (paralysis agitansj . The dat3 for the longer followup period clearly confirmed this result. Like Kahn, we examined the possibility that this low ratio could be an artifact resulting from the inability of persons with this disease to continue y smoken. This possibility, however, was contradicted by the low risk observed among ex-cigarette smokers (mortality ratio 0.70). An artificially low ratio might also result if Parkinson's disease were mentioned on the death certificate in combination with other diseases 'that are strongly associated with a smoking history and that tend to be selected as the underlying cause of death. T h e data did not support this possibility either. Other prospective mortality studies (7,8),as well as two case-control morbidity studies (9,10), support our results. The negative association of Parkinson's disease with cigarette smoking needs serious study. References 1. Dorn, H. F.: The mortility of smokers m d nonsmokers. I n Roceedings of Social Statistics Section, hmerican Statistical Association, 1358, pp. 34-71, Xmericrn Stacirtical Association, Washington, D.C. 2. Dorn, H. F.: Tobacco conrumption and mondity from cancer and ocher diseases. Public Health Rep 7+: 581-593 I, 1959). 3. Kahn, H. A.: T h e Dorn study of smokinq and mortality amonq U.S. veterans; reporr on 8% years of observation. In Epidemiological approaches to the study of cancer and other diseases, edited by W. Haenuel. Nacional Cancer institute Monograph 19. Bethesda. Md., 1966, pp. 1-125. 4. Rogot E.: Smoking and general mortality among U.S. veterans, 1954-1969. DHEW Publication No. ( N l H ) i5544. Xacional Institutes of Health, Bechesda. Md., 1974. 5. Roqot, E.: Smokinq and mortality Yhonq U.S.veteranr. j Chronic Dis 17: 189-203 (197+). 6. Roqot, E.: Smoking and life expectancy among U.S. veterans. .%nJ Public Health 68: 1022-1095 (1978). 7. Hammond. E, C.: Smoking in relation to the death races of I million men and women. In Epidemiological approaches to the study of cancer and other diseases, edited by W. Haenszel. Sational Cancer Institute !donograph 19. Bethesda, !dd., 1966, pp. L27-204. 8. Doll, R., and Peto, R.: Mortality in relation to smoking: 20 yeus' observations on male British doctors. Brit >fed .J 2: 1525-1536 (1976). 9. Yefzqet, bl. D.,Quaufasel, F. .\., and Karl, V. C.: h retrospective study of smoking in Parkinson's disease. .bn j Epidcmiol 88: 149-158 (1968). IO. Kessler, I. I., and Diamond, E. L.: Epidemioloqic studies of Parkinson's disease. Am J Epidemiol9.C: 16-25 (1971). r ROGOT, EUGENE (National Heart, Lung, and Blood Institute). and MURRAY, JAMES L: Smoking and causes of death among U.S. veterans: 16 years of observation. Public Health Reports. Vol. 95, May-June 1980. pp. 213-222. In a 18-year mortality followup of some 293.000 insured U.S. veterans, specific causes of death were studied in relation to smoking status. The main results confirmed earlier find- ings. Mortality ratios for cigarette smok- ers as compared with nonsmokers were 1.73 for all causes of death, 1.58 for all cardiovascular diseases. 2.12 for all cancers, and 4.31 for all respiratory diseases. The highest ratios (those greater than 5.0) were observed for cor pulmonale. aortic aneurysm, emphysema and bronchitis, cancer of the pharynx, cancer of the esophagus. cancer of the larynx, and cancer of the lung and bronchus. The greatest excess in deaths in terms of observed numbers minus expected was found for the cardiovascular diseases. in particular for coronary heart disease. Mortality ratios for excigarette smokers who had stopped smoking for reasons other than physicians' orders were much lower compared with nonsmokers than the mortality ratios for current cigarette smokers: 1.21 for all causes, 1.15 for all cardiovascular diseases, 1.39 for all cancers. and 2.08 for all respiratory diseases. For most causes of death. the mortality ratios for excigarette smokers who had stopped smoking for reasons other than physicians' orders varied inversely with t h e number of years of cessation. For some diseases, the monality risk for the exsigarette smoker returned to normal almost immediately after the cessation of smoking, whereas for others, the return to normal was more gradual. The first group included stroke and the combined category of influenza and pneumonia; the second group included cardiovascular diseases as a whole and coronary heart disease. For still other diseases, although the mortality ratio declined with the length of time smoking was discontinued. substantial excess risks remained even after 20 years of cessation. In this third group were aortic aneurysm, bronchitis and emphysema. and lung cancer-diseases with very high mor- tality ratios for current cigarette smokers. Parkinson's disease remained :he one disease that clearly exhibited a negative association with cigarette smoking. - T !. r: s :c rl Il