Tobacco Documents Online

BKITISH AMERICAN TOBACCO Copy Request Form Request Number: 23 Organization: Physicians for a smoke free Canada Request Details Request Date: ............... .~ .................. • el • o* o~ m• •o ee •o•••••e•o•o olm File Number: ...~..~..~...~.. ........ Box Number: .....~...~...~ .......... Page Range: First Page Last Page ~" ~. C.~ C) ~ ~ ~-~ Requested By: (Print Name) ........ ~ ........................ Details below will be filled in by Depository Staff Only II Copy Details • o• ill o• • • oee • .eel• • •.m•Q ••• .ao D a t e:..Q..2./~.~'/..~..~. ............... Time:•../~..../.~. ~ ........... Copy Checked By:... .................................... .............. . . ... Date: Time: ............. o•o• o• • l• • e• i, • • • • • • •o • • • • •l,•• • e=• i,•o • Je•••• • •• •o•• • • •o III Delivery Details Checked By: ...... Date: • • eeoc• • •e • • •••• oo•••o e•eoeeo • le••e••e•*•D.0•••o•• Sent By: .......................................... Date: ..................... BATCo document for PFSFC 1 March 1999

p Integrated League Tables (Revised verslon of ¥JCR 246, q 874) Dr. ¥.J.C. Roe 6th January, lg78 FJCR 282 R 43 ~o Background 1.1 A multi-component league table was published in the German ¢onsumerist Journal "Test" during March, 1975. In this table deliveries of d~y particula~e matter (PM), nicotine (Nie), CO and oxides of nitrogen (NOx) were reported for some 37 cig~ette brands. No attempt was made to list brands according to overall merit. Subsequently, however' in Switzerland a Dr. Herzfeld (for a consumer organisation) devised a "merit indeX' based on deliveries of the same 4 components. The first Herzfeld Index p~blished in .February, 1976 was superceded by a second in April, 1977,~lthela~e~ deliveries of the 4 c==ponen~ ~ere first expressed as percentages of a fictitious cigar6tte which yielded: Dry PM =~ 30 mE Sic = 2 mg CO = 6~ V/Y Nox = 1500 ppm The four percentages were added together and the sum itself expressed as a percentage of Z00 (why of 300 an~ no~ of 400 is unclear). Cig~ettes were then put into one of 5 categories according to this latter percentage figure. 1.2 In March, 197T, the story was taken up by a ¢onsumerist Journal C in Relgium, except £hat the NOx figures were not taken into account. Instead, a figure ~as Riven for sides~ream CO. In this case the overall "merit index" resulted in cigarette brands finding ~heir way --- lnto one of $ categories. BATCo document for PFSFC 1 March 1999

-2- °. 1.3 Readers Digest (US) published a "triple-gas-rating" far 28 brands of cig~re~te aC the end of 1976. Nothing sophisticated was involved - the del~veries of CO, NOx and hydrogen cyanide were simply added together! 2m Health risks from smokin~ 2.1 Epidemiological evidence has suggested associations between cigar- "ette smoking and a wide variety of diseases and causes of death but only five of these associations merit really serious consider- ation in terms of disease incidence and strength of associa~ion. ~ney are:- (i) (ii) (~Ai) car) (v) lung cancer chronic bronchitis emphysema coronary heart disease low birth weight and complications of same. In all fi~s cases the associations are much stronger for cigarette smoking than for cigar or pipe-smoking. 2.2 Cigarette smoke, cigar smoke and pipe smoke each consist of thousands of known components together wi~h a large number of un- known ingredients. A few of the large number of smoke ingredients have been found to exhibit effects An laboratory tests tha~ might be relevant to the possible causation of the above five diseases. Thus: (i) trace amounts of known carcinogens have been isolated from smoke tar, (it) skin tumours have been produced by the repeated application of cigarette, pipe or cigar tar to the skin of mice or rahbits, (tii) some of the gases in smoke e.g. NOx, HCN have been found to increase mucus production by respiratory mucous membranes and to inhibit (temporarily) ciikal activity, (iv) nitrogen dioxide has been claimed ~o cause emphysema in rats (although the evidence is disputable), (v) changes which mighY conceivably be ~elevan~ to coronary ~rtery.dlsease, have been C~ C~ C~ BATCo document for PFSFC 1 March 1999

- 3 - reported in the lininE of blood vessels of. animals exposed acutely to CO (bu~ attempts to reproduce these findings have failed) p (¢I) both raised COHh levels and exposure to nicotine have shown to reduce exercise tolerance in patients with anEina pe@~oris~ and (vii) exposure of preEnan¢ animals continuously hiEh levels of CO has, in some experiments, been associated wlth low birth weiEh¢. 2.3 The kind of information summarized in paragraph 2.2 falls a long way short of proof that any of the agents mentioned cause any of the dlseases listed In paraKTaph 2.1. ~oreover if'is clear that none of the agents listed are the sole causes of any of the diseases @oncerned. The most that can be said with any" confidence is: (1) w that there are probably components in clEarette tar which, if inhaled, predispose ~o chronic lunE disease including cancer. (ll) that in smokers who inhale smoke in the same way and to the same extent, less harm is likely to ste~ from smoke low in irritant compon~:~ts than from smoke hiEh in such compon- ents. (Ill) HiEh doses of CO and/or nicotine are probably harmful to patients who already have coronary artery disease but it is by no means certain that either component alone or in combination with each other or with other smoke components increases the incidence of atherosclerosls of coronary vessels. 3e The nature of the contribution of particular smoke components to the risk of partlcular dlsesses 3.1 The best guess that anyone can make about the basis of the assoc- C~ C~ C~ BATCo document for PFSFC 1 March 1999

iation between smoking and lung cancer is that it has something to do with the condensible matter. Not all Investigators are convinced ~hat the vapour phase (VP) constituents do not include relevant carcinogens and some think that the VP may act Go-¢arclnogenlcally. Virtually nobody believes that any one constituent oZ smoke is solely responsible for Its ca~clnogenlc effects on the lung: multiple InEredlents acting addltlvely or synerglstlcally seem .to be involved and, almost certainly, some constltuents' of ts~T exhibit antagonistic effects. Without better knowledge of mechanisms it is a reasonable hypothesls that the rlsk of lung oancer will be reduced In smokers who inhale less tar, and despite the argume=t that some smokers adjust their inhaling habits to get a fixed dose of nicotine, since tar and nicotine deliveries tend to me~e in same direction, ther~ is likely to be a~ overall benefit frc~ across the board reductions in tar delivery. 3.3 3.3 • ° ~he contribution of smoking to the causation of chronic-bronchitis is not so elea~ cu~. Occupational Sactors and air pollution are obviously of considerable importance and there is no clear indication of which constituents of smoke are especially implicated. In the short term~cough =nd mucus production are reduced by reductions in tar delivery and possibly also by the use of charcoal fll~ers which reduce ~he delivery of cert%in irritant gases. However, it is uncertain whether these short term benefits reflect long-term benefits In ~erms of risk of :chronic bronchitis. The nature of the association between smoking and emphysema is klso obscure. The laboratory demonstratlon that nltro~en dloxlde may cause emphysem& is somewhat unconvincing. Presently the vogue is to believe that the phagocy~osls of smoke particles by lung macro- phages which then release pro~eolytie enzymes Is to blame and that humans who are genetically deficient in certain antlproteolytlc enzyme factors (e.g.~1--antltrypsln) are especially at rlsk if they smoke. But this vogue may not last. If it does then a tar league table would be the best index. A merit index which took into account NOx and other Ea@es would have no subs~antlal sclentlfi¢ bamis and migh~ be misleading. co BATCo document for PFSFC 1 March 1999

3.4 In the case of coronar7 heart disease, ~wo constituents of smoke, nicotine and carbon monoxide have attracted Che most attention, aS poeEible a~ternative co-fa¢¢ors with dietary factors being the prime cause. If Russell is right and the best index of relative Bagel7 would be the ra~io oZ n~co~ine ~o tar and/or the ratio O~ nicotine to CO, and a "merit index°' which in some way adds___ dd the ta~, nicotine and CO deliveries together would be wholly misleading. 3.5 At present nobody knows which ingredients of smoke ~ffec¢ the birth weight of babies. Mos¢ attention has been paid to CO but nicotine and particulate phase materials are not necessarlly without e£~ec:, Tables for Tar and Nicotine BATCo document for PFSFC 1 March 1999

have some value. stay. In any case one suspects, that they are here to 4.2 The reason why Governments publish t~r and nicotine, tables is so that members of the public ca~ choose the lower tar and nicotine .delivery cigarettes which they a~e advised to choose on health grounds. 4.3 • For the most part nicotine and tar deliveries move up and down together. This is because the majority of the nicotine is in the "particulate phase of smoke so that factors (e.g. filtra~ion, ventilation) which affect the one also affect the other. 4,4 The scientific case for nicotine tables is less clear than that for tLr tables. There is considerable evidence that a proportion of .smokers adjust their cigarette consumption and puffing characteristics in order to get the dose of nicotine to which they have grown accust- omed. The possibility thus arises that such a smoker will extract similar amouL~s of tar and nicotine from cigarettes irrespect~ve of their position in the league tables, or if they canlt completely compensate in this waytthat they will consume more cigarettes per day. In fact the more likely situation is that such smokers will not be satisfied by low nicotine delivery cigarettes and will choose stronger brands. 4.5 All in all there is no strong case on health grounds for publishing nicotine tables but on the other hand, ~he inclusion of figures for nicotine in tar league tables does not introduce any serious source of confusion for smokers. Nor is it likely to induce smokers to indulge in the smoking habit in a more harmful way ~han they o~her- wise would. C~ BATCo document for PFSFC 1 March 1999

° . O' So League Tables for CO 5.1 It is doubtful whether raised COHb levels produce adverse effects in healthy individuals although they probably reduce exercise tolerance in subjects who already have coronary Lrter7 occlusion. For the latter reason, if CO were the only potentially noxious component of smoke, it would not be unreasonable to T.hink in terms of publishing CO league tables. This is not, however, the position and few doubt that, collectively, the thousands of substances which go to made up the 'tar' are likely to contribute most to the harm- fulness of smoke." C0 is under suspicion mainly in relation to . cardiovascular disease and effects on the fetus, but nicotine is by no means entirely clear of suspicion in relation to either of these possible health effects. 5.2 Unlike tar and nicotine which tend to rise and fall together, CO delivery can to some extent be made to vary independently of tar and nicotine delivery. Thus it would, within limits, "be possible "to have medium tar and nicotine with low CO delivery cigarettes or low tar and nicotine with higher CO delivery cigarettes. ~f a Government published tables which contained such 'choices' what advice would they then E'ive to consumers? 5.3 5.4 There is no formula for equating ~he risks from tar, nicotine and CO and simply to advise the consumers to choose a brand low in all 3 constituents is to ignore the facts referred to in paragraph 4.4 above• A vocal body of medical opinion is on record as stating that the formula for a safer cigarette is one that delivers smoke with a high nicotine:low CO ratio rather than a low everything ¢igLTette which will be (a) rejected by those who smoke for the nicotine or (b) 'over-smoked' by comparison with the smoking param- eters used for determination of the various deliveries. One must conclude, therefore, tha~ ~he addition of CO delivery to existing tar and nicotine tables would at best serve to.confuse smokers into lwi~ching to brands that they might smoke in such a way that they are at greater risk than before they switched. c~ c~ BATCo document for PFSFC 1 March 1999

- 8 - q 6o Leasue Tables for other components 6.1 Smoke consists of thousands of constituents. The composition of smoke changes rapidly during the few seconds after its generation and is subject ot numerous chemical interactions during this time. Hany of the constituents other than carbon dioxide and water are present in very small or trace amounts. The detection of the presence of some of them represent triumphs of analytical chemistry. The accuracy of the methods of measuremen~ of many of the constituents in the amounts in which they are present is poor and frequently varies depending on the concentrations of other constituents. There is little precise knowledge about the potentiality of m~ny of the constituents for biological effec~ and every reason to expect mult- iple biological interactions. Thus it would not be su~prisin; if the potentiality for effect of one component depended to the extent of 10-fold or I00-fold on the concentration of another component. Finally, it is clear that the way a cigarette is smoked has consid- erable effects on the composition of the smoke produced. 6.2 Even if (t) it were ~ossSble to measure accurately the concentrations of a number of minor components of smoke, (ii) there were no biological interactions and (fit) it were reasonable to ignore the fact that smokers smoke differently, it would still be impossible to produce any meaningful InteErated league table tha~ took the deliveries of such ¢onstltuents into account alonE with those of tar as & whole, nicotine, and carbon monoxide. In the first place it is not clear whether in the concentrations in which they are present any of the components has any measurable harmful effect on health. In the second place even if one could define adverse effects on health of individual components, there is no formula for equating one kind of health risk with another. 6.3 Individuals differ in the spectra of diseases to which they ere especially susceptible or resistant. One par$1cular component may be harmful in one way to one individual but not to another whilst another componen~ may be harmful In the same or another way to the second individual but not to the first c~ c~ c~ BATCo document for PFSFC 1 March 1999

a I -° 6.4 If, therefore, the Government or other body. decided to publish deliveries of say the 16 individual components of cigarettes listed in the Hun~er Commi~tee Guidelines, brand by brand, they would in effect produce no~hing more than a complex and uninter- pre~able table. Many of the numbers would be qualified by large confidence limits and there would be no back-up of scion=trio informa~ion ~o answer any member of ~he public who enquired whether he should choose a high X low Y or a high Y low X brand. Most members of the public would be confused and some of them misled. Turthermore, much of the impact of the present, simple, tar and nicotine ~ables would be lost. 0 "Merl~ Indices" in the interpretation of multicomponen~ league tables 7.1 If the Government or another body took the matter further and attempted to i~egrate the figures for multiple components into a single index of relative merit, they could only do so by making "assumptions for wh±ch there is little or no medical scientific or toxicologicaI basis. The wide limits of acoura¢~ for the measure- ments of some components might, depending on the method of calcul- ation of the merit index, fluff the significance of differences in delivery of components which can he measured with narrower confidence limits. The more components that were integrated, the lees effect differences in any one component would have on the overall index. Thus there would be a tendancy for indices for different brands to approximate towards the mean for all brands. FJCR 7.2 Theoretically the number of different ways of 'integrating' the deliveries of widely different chemicals with the potential for widely different biological effects is legion. Since there is no scien~ific basis for choosing any one particular method from the many, it behoves anyone who makes the choice go list the articles of faith on which he relies. peculiarly his own. He may ~hen find tha~ his religion is c~ c~ c~ ~J BATCo document for PFSFC 1 March 1999