Tobacco Documents Online

. ." report ~flong-tcrm Follow-up re.sults from the FtarninghamSmdy, "." ,. Foun~i~e.~evide.n~-~ ttaat .~he. filler,dgz~ett~: ,of,the I~0~ and ,ear|y ... - |9?0~eonFerred any protection from coronary heart discasc(CHD) and ~_,~ested :that more suspicion should be ralse~ about the -smoked) concluded that smokers of filter cigarettes have a lower mortalky tha.~ido smokers of plain cigarettes, for those diseases mbst0 . TIIELANCET, SILFI"F.MBER 19,198i " : -. w~c able to examine such n~or po~cnt~I confounders ~d .rule .thcm Out ~s cxplainers of a particular result. . : - " " " Lee ~onlends that our fit{dlngs art too limited and makes the fatal error that all you have tO do is. m~.e N hrg~ and ~¢sults that arc • .~mall become signlficant. In our smdy--,and th~s ~ primarily why we published the data at such an early stage--the ~udc mtcs of CHD were, ifanyrNng, ~ighcr in the filter dgarctte smoker, s,.hnd fxtm*~cr numbers :m~'.g~t p~obalfl~ onty ~scr~e *to m~c that :mote people that at an early stage the returns on filter dgarettes do not look good and, as wc mcqtioncd, this may bc due tf~ the fact that ~a~I:~R, foonoxlde~_~: ver~ powe.~la]. ~ " iS • "fiherlplain rcla~dye risk ~fO- 94t,still 'had,w~de 95.% confidence.limits (0"70Li "25) an=d' again even'flle'redu~ed mortality in non-smokers. • compared with smokers (18%) w~ not significant. Sample size • were if any~hlng worse off. .... : .... .. - " • But there is an even more fundamental 6om~demtion here•Lee " "'" : places great stock in his statistical tests, but it is well known (in ~ "statistical drclea at least) that in the large Amcr[.can Ca.nccr Sg,ciety ",..- .. studies virtually any number .difference in the num~ato~ is ".. significantly different from any mher number. The ultimate "" .consMeration does not then become statistical significance but • biological and human consideration. For example, take Lee's beat • tesuh, a 25% fall in lung cancer rates in filter s~okers. It means in .. '" "'ptaellcal terh~ that a pack a day smoker ofcigatettes goes from 20 • dme~ the risk oflung ean~cr withplaln clgarette~ to 15 times the risk - ofI~g cani:er with a fillet cigarette. I do not personally get much sa~t~a~t io.n enco~r~g}ng someone to pu r~uc a habit, which incrcas '" t,her~skoflunfIcan'oct IS'times. " "'., " " Thus~ I feel that we should l~ concerned about this unsuccessful ca'To r t by tobacco companies to produce a ",~d'e cigarette"; Consider the number of people (addicts) who have rationalised their health • away l~causc they want to believe the propagan~la that a safe problems were even more marked for the: analy*es by age group..., cigarette is ~ the way. How long sfiould we play this game~ Is itnot F~ more emphasis should have ~n phced on ~c limkadons of " ~tter to put the emphasis on quhtlng bemuse swltch~ng ~ such a the Framiagham Study data. ~cfindings do not providcrcliablc . hollow promise at pr~cnt~_ ' " evidence on which ~o doubt the ~nefits of the switch from plain to " " filter ~garctt~.'. r • . . . _ _ "N~fio~l H~tt Imlitute, . .. . .. 2~ Cedar R~ '. " ! .... H~ Disease Epidcmiol~y Study, . . ~,uo.S~2 ~ P~TE~N.~ Ftamlnl~am, M~chu~¢tt~ 01701, U.S.& WILLIA~P~ ~ %'This letter has been shown to Dr Castdli, whose reply follows.--E~. L. SIRe--It is indeed true that the sample siz~ and number of coronary heart disease (CHD) death* may not be large enough in Framlngham to demonstrate a benefit in the range of'10-20% in the filter smokers. It ~ less likely, however~ that we would ndss such a benefit in the total CHD experiene% and it is important to look at the total experience not just the deaths. ' ". . The large studies of Lee and Crar finkel~ do have the advantage of being able to estimate small effects with great confidence• However, they have to rdy on information collected under less stringent control than in the Framlngham Study where close surveillance and very carefully obtained information are traded off against large sample size. Mr Lee leaves the i~prcssion that it is only sample size that explains the reason why the studies he rites differ from the results of the Framlngham Study. There are oth~ very important com, Merations that he has left out. In practical cliM~al dad s~,:ntific terms h means tllat not enough' is known about the individ,~Is in Lee and his colle~es' studie~ to ru!e o,ut that a fu~d~mr nta! Nas ~ -:h,lpi~g theh rt~'~tt. For eaxample~ are the l~ple who swhch in l'd~ studies to a filter dgarette gso h~Ith eemeious in at:tee ~: a} s (,t~e', exerdsc, and so on) to explain why they did better in CHD e.u ",,¢ rues? In the Framingham data we NALOXONE REVERSAL OF ISCHAEMIC NEUROLOGICAL DEFICIT~ Sm~--We have been discussing the use of "naloxone in ccrebrov~scular accidents since June, 1980~ when we observed an instance of reversal of coma and hemiparesis after naloxonc. The paper by Dr Baskin and Dr Hosobuchi (Aug. 8, p. 272) prompts us to relrort this ease. An 85-year-old woman was admit ted on June 25~ 1980, because of severe heart failnre with pulmonary oedema. She ~as alert and Free ofneurologlcal defidts. She was given standard therapy, including morphine 0hree b~travcnous doses of 5 mg during the first I0 h, the last dose being given at 0930 hours on the second day). The pu!monary oedcma cleared. However, in the late afternoon of the • second day, the patient became unconscious and parezis ofthe right • arm~ bilateral pyramidal signs, respiratory depressipn~ and bilateral mlosis developed. At 1910 hours we decided to give na]oxone 0.4 mg i.v. Within 2 rain the respirator,/" rote doubled, the pat{eat wotze up~ moved the ~r~.Meusly parctk right arm, and was able to speak. The miosis disappo~red. The effect on consciousness lasted about 5 mln hut the tesl~irs.tory rate w~ normal for several honrs. In view ~of this observalion~ ~,¢ were rductan( to accept the hypothesis of a morphine induced deterioration more than 9 hour* after the last mnr£h,.'ne dtx~e and we felt:areal the naloxotae dose of 0-4 mg i.v. I H~tt.~'ae VM, r ) Js. $:.~k~s at~ hcahk: ~ a~llen bctwettl T1052800~