Tobacco Documents Online

](;j(n EriddNua[ Vul. 5'.. Nrx 9, pp. 627-1129. l.vy9 Cinrtlehc rV 1?99PLew~.Srlenc~lr,t All ri~A.n rcrervrei. VARIANCE AND DISSENT Dissent Rebuttal to the Paper by Enstrom ORP 5-115699K-}re fnx,r msnrr r'll SOtt95.4154(99)t5.P3514, Lewis H. Kul(er DSrnnTasnt!>r Ermratuux+., Unrvrnsrryur Prr•rara~xcH fiawrn,q-n; 5cnooros 1'Lwatc lisACrH, The anicle. "Snx+kinc Cessati,a, 6Mortaliry Trends Aunmg Two lfS lruptdatiuns," compares the changec in rcl- arivc risks nf all causes, lunscancer. and selecred other dir- c,vres ineluding cRrdiov,•tacular disease in a largc cohort of U.S. World War I vetetans aged 55-64 at entry, followed up from 1954 to 1979 and from the NHANES I Epideminl- ogy Foli.rw-up Study of U.S. adults aged SS-74 ac entty and folloa+etl up from 1971-1992. The study suggeats that smoking cessation dld not d.ange the relative risk of morml(ty comparing mnokers ver.cus nr,n..ruukm at the baseline axaminatiun- It is un- clear what the ttucputpcxsex are of this articlr. If eme pur- posc wnc_to determine the effects of smoking ce4sation on older individuals, it would tnake tma:h mo(e sense to study cohorts of oldes individuais who were smokas at ency and followed up over time to compato rates of diaeace among ccnokets, and the ex-smokers.Evest auchairelysis could be limited becaur•e of che diffetmces. in heatth chnraccerlstics between the smokers and thaex-smokere, e.pecinllyin the older age grcnti», It is obviovs diat many i+ider individuala who quit smoking dn so hecause. of ill health and nor be- cause of an attrmpt to itnprove their health starus. Previuus cpidcmiological stuciles have attcmp<cd to eval- uute the effects of a lawer cholcscerol level, deecease in lmdv weigltc. towering of ayscolich)ood ptcxsure, and a, an, and morraliry amnng okicr individuals. All of the¢ saulies have generated inmest but have been cerinvcly Aawed be- cnuce of the. failure to consider the ineEaact of rlxhealth of the oldu ittdividuals on the risk factua. In utlur words, muuc of these studies (including the present article) pnmarl ly measure the adverse effects of health and aging on risk fac- tors rather than The ¢(fccts of the risk factors *nd chur(ges in the risk f.+ctnrs on diaate. Thus, the assutnpdon in this ar- ticle is that smoking ctssudt>ft does not reduce the eaccae relarive risk amemg smakets compared with nonemokers without regard for the reasons for smoking cc.wtion for an elderly cohorr- Note that for the 1954-1979 cohert of Adde,.. co..egonyanee ee- t.ewe r(. K.dkr• 4t.tf.. (b.PH Univen(ry M Piush.nal,, ttraduare Sehmt of Publlc Heubh, ()qmw.en.t of Epfd.aninl- r,ev, A5f7 t:n6trse ftalL (30 Dz55nm Srreet Pirubundu PA 15261. Aacepted 4 March 1999. World War I vctt7uns aged 55-64 at entry In I some of thc veterans would have been at leac age, and by 1975, 84 yeara. Failure co note a elnnRr: in relative risk betw and noncmokers over time can be due to a si in mortality ramol~,+ a large number of ex-s amokcrs. Changes in other risk factors such as hypectensioa could r<suli in a decline in mtw smaketx and nunsmokers. in a similar type of analysis, but using a ptnpecZive de- sign, Rosenbeum u a)_ [1] evaluated changes 14 tnotculiry over time It. relation to cigarette Smnking in Mortality ,Fallowback Survey (NMFS) and Health Interview Survey (N1HS) fran,1970 iween 1966 and 1966, the ail-rsunc tnoatal'tty N•arional ;a National 1987- He- ecUaed for both nonsmokers and for tac-smokers, but nntlfaxcutrettc smakers. Lung canecr motralitq'.inucased fen tenc and ex-smokr.rs eomparing the 1970 years. 71rin was cspeciallytrue for women. Th crease in morcabry tor coronary heatt' discdse never smokers. current smakces, aad for ex-g The increased relative risk in current never-senokers for Iungcaneer vrm consistentw studies. The increaae in relacivc risk for smri clon of both the inaeasinp duration of smaki (cohort eflect) and also probably tlie intensity smnkiryr. The relative risk of lung etmeer about 50% for ex-smokers but tema(nedat fivefold compared with never-anolcers, whereas risk for coronary heatc discasc among ex :amoke with nonsmokns reached uniry. The atticle agrspecific analysix. The folloer-ug from the Amesican C'ance.r specrive srvdice in 1959 and 1982 have results thece two studies 121. Acain, drere was an incre rive risk of lung cancer in smokcrs versus whieh. again, can be au.r.l.wt<.1 to the inereast of „nwkup,. (t.c., the cohort effecrs), especially age groups and, Probably, the Interwlry of ci The current article in this journal suggests been a failure to note a decl(ne7n lutgr cancer the recent decreases in c(garette smoking. This false. The lung cancer death rares arc declining h thc cur- rt in laier was a de- r tltoe for ers. ets versus ptdvitun it a 11tnc- over time eigarette reoxed by rwohuldny he relative compared ided no ety Pco mimic in rela- nsmaketa, duration the older trbkvfg. thete has In apice of oomeru is n younger

age /,aoulu and the rare of ittercace in older age grarn5, e.vtx- cielly for men, has slowed [3]. The reascuu for chese differ- ences, as noted in the Disery..don of rhe article by Rosen- baum, are the cnhurr effeces of cigarctte smoking over time, especially in chese older rrKc groups. The eutrent aruclee clearly has vcry aerious limitations drat, csseadally, tnakc any Interpretatioos of very quescion- abk value. The authors presume that these older individual stopped smoldng as n`good health behavior." This is falla- cious. Qlder indivldtrals are more likely to stoP smoking be- cause of change in real or perceived iliness and disability. It Is highly unlikely that, at age. 75, individuuls begin to quit smoking because they ltec thc metsagc about good health. The associarud diseases and disability ccatld, clearly, heas- s<~ciated with increased risk af dcarh [4L Thus, the pre- sumptirm that all uf the individuals stopped smoking by fhe cime ahc lasr individual [cmch.es 105 has norclevance to the quesrion about whether smoking cessation reduces the risk of morbidity and raortality among dderindividuala who ere healthy. It is Iliogical to cotnpare individuals who stapped srrmoking at 30 arrd 40 years of age with individuals who stupped smoking at 60, 70, and 80 years of age. Individualc stopped snwking because of ill health cowani the end of their Ii&. It is likely that most of t1'te cigarette smoken will have stopped smoking, as the authots note. This has no rel- cvance to the question of che value of stopping smoking among "healthy individuals," ' For example, in rhe Multiple Risk Factor Imervcntion Trial 15.61 which included sme of the mnsr extcnsivic effnrts to increase smoking cessat'ion among "healthy" middle- aged high-risk men aged 35-57 years in 1973-1975, cnly abcmt I Iwro of r}re usual care participenr.i etvppe.1a smoking over 7 ycars cornpared with appruximaeeiy 20% of the spc- cial carc pan.icipants. -Fhm, the high percentage of quitting emong older itidividuals is not related to aggressive health education nte'Jbages end intetventinh programs hut is cer- talnly related m health problems. ln the MRFIT. there was no dliference in lung ranccr ruur- elirybecwRtn vwual and special care afrcr t6 yexrs offoliow•up [6], The lack of benefit was not due to the failure to roducc =he risk o( lung cancer following hmg-term amo4itg cesra- rion but rarher to the difficultien ( i.c., the low pcrcenmge of quitters) of healthy middle-aged mep who consumed 40+ cigarettes per day and were at high risk of lung cancer. Only 284 of 1639 special intervention and 154 of 1652 usual ratec mon who amokedd grcaret than 40 cigarettes per d ay at brrse- line had quit smoking, 77tese itti7ividuals who cAncinncd to smoke had a very high lung cattcer mcxtahry, and there wax actually, in thc observariona(evaluation of this scudy, a re- duction in risk of lung caneer associatad with smoking ces- sacion. Given the difGculties of unoEing ceseatlon amutrg heavy cigarette amokerx uttd theit risk uf lung cancer, it is hard to bolievc, in a strictly obscrvational study with no in- tmendon, that a 6rge number of healthy individtrals are stoppirrlx6rnclcing in onkr no reduce their risk of lung cancer. In a reccnr sludy of a popul:,nott sarrµle Spain, 477 men aged older than 65 wete e vival from 1986 to 1994- Rciativc risk of d among cigarette smokers und 1.5J in ex pared to nonsmukers. The relative ri.k- was among men who had quit nrttokking after age to wntimmi smokerr 171. Furtaer .+mokers ( baseline) reponed a much higher prcvaleuc hezlch in boch heart snd respiratory disC lic consistent with the view ihar many olde duced their smokin>R because of ill health. tum i3atcelona, aluated for sur- d, wad 2.1-f ,ld ckers as corn- educed to 0.77 rw coompared Ider than 65 at nf FaiT or poor , which would individuals r4~ Therc is also an important period berwe amokutg ccs- s-ation and rnduetion of risk of lurtg cana:er. 1 dividuai.,vho smp smoking do nor reducc their risk of l cancer r„ that of the ncver-sutukers (7). Rmcatg oldor i ividu.ds wh., have smoked for many years, the reductio of chc risk nf lung cancer will probably he modext. even after smoking cessation. The high risks of lung cancer for individu- als would persist for meny years. This will he pecfally, true for thac older individuals who have smnked many yr,as and have decreased lung ftmcNnn [8,91. The 4ecr<ased lung funcuon of cigarette smokers and ex-smokets Iss been asso- ciated with an inc[ea,aed rislr of lung cancer a d is probably a rtueasvn nf a dose and biologic effect of ciga ire smoking. Catdlovasc~siar diseuse i.a chr other rrtatim use of death among cigarette smokets, and the catdi r risk assnci- wce¢l with smoking will decrcase with advatrcl age. Therc ace seveml reasons for this. First, extensive a[l lerosut ir, the major dctccminant of the risk of caniitw ular disease-, huiividuals who develop cxteneive adt krn.cis and smoke cigarertes ate likely to die, before the reach these older ages that have been included in the rrenr studies used in the atticle by Enstmm- Thus, smnk surviving to ulder ages are less iikely to have severe a lemsis, poa- sibly owing to their other risk factorx such as -density ii- poprotein choloterol, high-density tipopr ein ehnics- rerol, blood ptcssure, and dialxtes, or perha to generic susceptibility- Countries in which the b tstul preva- lencc of a4tcrtrcletusis is relatively low, such the Asian Padfic-cuuntries or Southern Europe. report Ia very high prevalence of cigarette smoking. especially jmonK men. There ate, however, rclatively low races of h attack. Second, in the older age groups the causes o espe- cially, cardiovascular diseases, arc nor well d ned as they arc in younger age groups. Mirelassifica[lon of causes of death wouki decra,ae the risk associated wi eigare.etc smoking estd coronary heact disease mortality i the relative risk of cigarette smoking ftm the misclassified d- ase is less than that for coronary hcart disease. Third, other discasex btxocne intportant de ' ana of cardiovaseular discase with aging and inerea ire prcva- lence with older age. This is espe:cially true of d bet~ mel- litus atul systolic hypenensinn-Thus, with incr - g age in thc cohort, the effeccrs of other diaxeases auch as abeees and elcvared sysmlic hypercension become major d reemlL,ans I

17~nt nf morraiiry. This sinration may lx. 4ifferent in wuteren than in meti. Women develop their cardiava.culer discase ac a later age, and the e(ftccs of elgFrcttc smoking on older vrnrrcrrr and nnokioy cessacion may be auhsrrrntially greater than chcy arc amnng very o1d menn in the Manitoba Follow-up Study [1Q), 1948-.1953, the incidenca of curditrvaccu/ar coremary heart disenscin- creased dramaricallv witlh aKC. The prevalence of 2iP,arette smoking taortality dKreases wirh iner.arsing age from 64% arnxwnd apr: 40 to 22% at age 75. Aoth.diabetes and syswF'ie blood pressure increased wlth aga in this cohort. The retx- cive risk of iSchemic hqve discasc for xmnkers verws rtutr smokets wns 1.6 at age 40 and 0.87 at age 70, but it drops m 0-71 at age 75, consistent with ptcwitus nhservacions. 7hus, thers Is probal+ly a very different association of ciKsrette smoking or ex-smnking andcoronary hcatt disease rLck ;jttrtmg alder men compared with mirJdle-aged and younger men-a r.nhorr survival effect. In the Cardiovasatlar Fkalth Study. of chase older than 65, the relari..e risk of . myocardial inforction among smoker.c vernrts nonsmokcrs wzs 1.10 (957o confidence interval = 0.78-155). T1lbrela- tfve risk esrimares. howevcr,wcre higher fer wnmen than mrnh11. ln summary, the data from En,attomn pruvide little usefui informasion. Thc population in the scudy. is hea~rilyV weighted by old<r.tnen. The reasonswhy indi>•iduals stop smoking are nor explained, not are any data pccaenred on the smoking cc5sad.m ratcs and hcaltli ehatacceristics of sho ex-esnokm, curtcnc stcurkecs. or nonsmokers. Tlutr~ is soLd cvidencc in the IiteratLee oCthe t.flet:ts ed ctgareae smoking nn lung cancer. corunaty leert disc-use, and nther dlstases and on total m<+nalitY. There is equally potverful cvidence of the effects of imnkinq te,c<ation ori changes in rates of coronacy heart diseaseattd&mg•ecc+n cAanges in luirg cancer nnd, also, in all-cavsetaurrality (12j. The snudy of cigarette sntohing among oldcYindividuals, especially men; requires a cateful avaluation of the health ;tanm of the ittdividuals prior to smoking cessacion. Srxnc epidemiologiscs eppear to have a Fnacination with changes 829 in risk fnctnne thac occur ~idc aging and are, cleafly, ;scnci- aced with the aging proceCS anrl disability, hu[ a,~T• ttur nec• es.arily risk Fa~rs for disca~e II. References t. EnsrroalJE.Sm<dc;ngcesmGatandnuvmlittrtenA u.glwo Uni~d Statcs pepularions. ) Clin F.Mdvm(ol 1999; • 813-825, 2_ Thun M, DuyLally C, r:allt L flasrdcrr W. Hcm C, Fxccss /Loztality among Cigarette emukces: Chu.gc+ in a 0.year in- rarvnL Am 3 Pubfic Hmlth 1995; 85: 1223 1230. 3. Nat:.m:d Ccnter fnr Hralrh Srarisrica. Fiedda Un ed Statu. i)96-87 and injury Charthorrk. F{yattsville, N(:H.l't 1997. DHHS publiatcion no. (PHS) 97-1232, 9/9 4. FotsFn L. Bfartveit K, $fptnd,d A, Edna T-H, eycr HE, 9chei B. FX-smokers and risk uf hip fr,+cmtF J Public Health 1998; 88: 14ti1-1483. i.'Shatm $J. Kuller 11-1, N.nv, Jn, 4tt the Re.earch Gtaup. Aswtcfat.on tcnYeetr tiaseline rlak fac+t , cigarecte amnlUng, and CHD mrrrtahry sitcr 10.5 ycort. Pteev Wed 1991 ; 20: 6551G9. 6, SharenB), KaBer 1.H. Kjetshvg MO. Sramlet ), ua JK' Curlet )A, cr alA f,u thc MRF1T Research G:vup.. £7udc- a,iat 1997; 7- 125-136. 7- Sunytr J. Lamarca R, Alonso J. Smaltit4; afrer CS ye.,rc - sed mortality in Barcehrna; Spnin.Am J Epide ioi 1998; 148i 575-580. . 8- Rulg3ntet A. A4mcer ). Ann; JM. RclatiPnslclp of ith be- havlors tt+five-year motcal'uy in sui eldrrly cnhurt Age Age- ing 1995;:24: 113-119. 9. PiJbt t:r.Fcvkens F,l. ICrnrohout (7. Self-rarerl hexl ; mortal- iry. aod clunn;c dirw~ in cldorly nr<n: The Zu[p Swdy, 1985-I990- Am J Epidemid 1993; 13& 840-848. to. TaccRB,ManfredaJ.CuddyT6.Thcdfeccofege . riskfm:- rnrs for ischentic hea:, "rlmease; The Manirnhn- •Il.,w-up St.rdy, 194&-1993. An#. Bpiderniol 1998; 8: 415,-4- 1. 1 l. Psaty BM. Fudxri; CD; Kvller I.TI.'8iW DE. Rai 'u PM, Pnlak JF, n al. Trudirional risk faetots and suhel' '-diu-ase measures aa Aicd;cr.x. rd firse myucrrdial in.i In older adnlts- The Cardio-a.cular Health SpidY. An:h I tetn Med 1999. In Prerd. 12. Nxri,urai Inscitures ofHealth. Changes in ~ e-rclated- dlenase risksatuf rheir implie:`rcinn farprevcnritrn eantroL Mrrnctgreph & Nethcsda. MU: National Institutes Healrh, Naponal Cancer inn.rnrc; Fd.ru:,ry 1997. NIH blication no. 97-4213.