Tobacco Documents Online

BEFORE THE FEDERAL TRADE CO[,~MISSION CIGARETTES AND RELATED MATTERS ~THODS TO BE EMPLOYED IN DETErmINING CO, "TAR" AND NICOTINE CONTENT COMmeNTS IN RESPONSE TO THE FEDERAL REGISTER NOTICE OF JANUARy 18, 1979

CQM~]ENTS IN RESPONSE TO THE FEDERAL REGISTER NOTICE OF JANUARY IZI 1979 The fQl!owing comments in response to the Com~Lis- sion's Notice of January 18, 1979 (44 Federal Reqister 3777) are submitted on behalf of six companies: American Brands, Inc., Brown & willi~mson Tobacco Corporation, Liggett Group Inc., Lorillard, A Division of Loew's Theatres, Inc., Philip M~rris Incorporated, ~nd N. J. K~yncids Tobacco Co~any. The Federal Register announcement invited comments as to three specific questions. Before turning to those questions, there are two preliminary observations which must be made. First, the undersigned wish to emphasize that a number of eminent scientists are of the view that cigarette smoking has not been scientifically established as causing the diseases or conditions in human beings which have been asso- ciated with cigarette smoking. Further, there is no ingre- dient or ingredients as found in cigaret<e smoke which have been shown to produce such diseases or conditions. The test- ing for and publication of CO content by the Commission will, in the companies' view, imply to consumers without scientific proof that CO content has relevance or siqnifica~ice in terms ~l of health. Second, the undersigned wish to set forth in the record the fact that the commission has not honored or adhered

2 - to the procedures which it promised to follow in connection with the development of the methodology for the determination of CO content. Specifically, in a letter dated March 14, 1977, Russell Hatchl, a Co~nission official, wrote to counsel for the undersigned stating that: "We realize that the new machine will have to be tested for 'tar' and nicotine and the re- sults compared with the old machine, and ~hat CO in cigarette smoke will have to be tested with other laboratories which have had ex- perience with determining CO by other methods as well. After these tests are completed, and only then, can the methodology be fi- nalized and reduced to writing. At thi~ ti~e a collaborative study will be conducted to determine how well the method works in other laboratories with experienced and inexpe- rienced people running the test. After a successful collaborative test we will write the method which will be used by the Cori~nis- sio~'s laboratory~Lr The Commission simply has not followed the approach outlined by Mr. Hatchl. It is our understanding that the CoW, mission reduced to writing its methodology without running tests in conjunction with other laboratories which have had experience in determining CO by other methods. Nor is nhere any sug- gestion as to the conduc~ of the collaborative study prom- ised by Mr. Hatchl. The Co~ission's failure to honor its co~nit~ent in this regard is not ~e~ely a procedural matter. Unless procedures of this type have been complied with, the Co~ission's m~thodology cannot have an adequate scientific basis and should not be considered for or used in obtaining results which will be disseminazed no the publlc. ~D

L - 3 - With these general observations in m~nd, the under- signed now address the three specific inquiries posed bF the Federal Register annoonceme~t. A. Should The Co~mmission Laboratory, In Performing A CO Analysis, Use The "New Smoking Machine" Described In The Paper By Messrs. Merfeld ~d Pillsbury? On the basis of the information made available bF the Co~mmission the answer to this first question is clearly "no,lj i.e., the Co~Imission should not use the smoking machi~ described by Merfeld and Pillsbury to conduct CO determina- tions. Th~r~ a~e a nu~er of separate and wholly-indcpendent r~ons which lead to t~is concl~slo~. FIRST, there are no published data obtained from smoking cigarettes on this machine which establish that this machine yields accurate and r~roducible results. The Commiss±on has not ~ven published anF data on the machine's performance. How could any scientist possibly agree that this machine could be used without docualcntation that it performs satisfactorily? The absence of data is particularly slgniflcant here because this smoklng machine represents a radical departure in de~iga from smoking machines which ~re ~sed through~ut the world and which have proven themselves through loilg experience. What the commission has s~pp!ied thus far is merelF a d~scr~ption of the physi~l ~ppearanc~ of the machine and its mechanical operating principles. This provides little or

- 4 - no basis for concluding that izs operating capability, relia- bility, a~uracy or precision is satisfactory. These attributes can only be evaluated by close examination of significant q~antities of experimental data generated by use of the machine. This insistence upon actual data from the machine is not mere procedural haggling. There are numerous possible or probable sources of ~rror in CO measurement associated with this design and the proposed analytical method~ These include~ for example, the cumulative error due to the analysis of the individual puff, the effect of "dead volume" associated with the gas transport tubing, the response time and accuracy of the non-dispersive analyzer, and the precision associated with the data analysis equipment. In the "tar" and nicotin~ determina- tions, it remains to be seen whether or noL the new machine will produce data comparable to the machine used in the pre- sently accepted procedure, small differences in puffing profiles, in cigarette holding techniques, and in butt length determination, as well as the necesslty to draw clearing puffs of room air after each cigarette, could contrlbute to differ- ences in "tar" and/or nicotine values from those which would be derived from the standard machine. NO data upon whlch to judge the presence, absence or significal~ce of these differ- ences have as yet been made public. Moreover, what is required, before i~ could be agreed that the Co~ission~s m~chine is satisfactory, are not merely &1 O O O O O

- 5 - data obtained from smGking on that machine. Rather the per- formance of this machine should be checked against machines and methods the reliability and accuracy of which are know~. One cannot establish scientifically sound analytical proce- dures by governmental edict. SECOND, the Co*rJnission's smoking machine should not be used because its conceptual basis, i.e., separately anal~z- ing the smoke from each puff, is not acceptable. It is very difficult to colmnent even on the hardware described in the Merfeld-Pillsbury paper because an adequate description of this complex machine has not been made available. Specif- ically, there is no schematic diagram of th~ hydraulic system or of the syrlnge-manifold system with information as to approximate vol~mes. There is no timing diagram showing the time relationships in the complete (dual) system. A computer is shown in a diagram in the presentation, but is not described i~ the text. Some functional description should have been included. Notwithstanding these deficiencies, it can skill be confidently stated that the new machine proceeds upon the wrong basis by separately analyzing cO from each puff. This is because determining the CO delivery of cigarettes on an individual per puff basis in~roduces systematic and random errors substantially in excess of :hose encountered in the Co~issio~'s current "tar~' and nicotine testing. O ¢ O O O

- 6 - Under the approach proposed in the Merfeld-Pillsbury paper, per cigarette values will be the summation of individually measured and calculated per puff values. The errors in final cigarette values will be greatly magnified using this approach. Following the Commission's proposal and assuming that a cigarette takes I0 puffs, each one of the 20 smoking positions will yield more than 50 pieces of CO data (i0 values per cigarette, 5 cigarettes per port position, plus clearing puffs). Each of these values will have an error associated with it. These data must then be combined to obtain a final per port result comparable to the "tar" and nicotine measurements. Moreover, the mere handling and manipulation of massive quantities of data i~herently leads to error. In addition to the "tar" and nicotine analyses, the procedure proposed by Merfeld and Pillsbury, if applied to the analysis of 175 brands, would require approximately 175~000 more deter- minations. Added to this, the number of smokings i~volved will be increased by 25~ when the monitor cigarettes are included, and the necessary clearing puffs must also be counted. This brings to well over 200,000 the number of pieces of data which must be processed for the CO analysis. Thus, the determi- nation of "tar" and nicotine which the Commission,s laberatory now performs on a market sample of about 175 cigarette varieties -- originally promised to be released semi-annually, but now approach- ing a once-a-year basis -- will be further weighted down with an extra burden of nearly a quarter-million additional measurements. 500000; 15

- 7 - This should emphasize the impracticality of making the analy- sis for carbon monoxide on a puff-by-puff basis. Also, use of the per puff approach has one other inherent defect. Accurately determining CO content on low delivery cigarettes on a per puff basis is practically impos- sible. This is because of the extremely low concentrations of CO to be measured, especially in the initial puffs from a cigarette ~nd in the clearing puffs. The error in each per puff determination may be as great as the CO value itself. On a related matter, the undersigned companies have serious reservations about the reliability of the built-in calibration check. In the Merfeld-Pillsbury paper we found no indication of the calibration range to be used or the accuracy and precision obtained during measurements for carbon monoxide. This could be especially important at very high and very low concentrations of carbon monoxide. For example, the choice of a 0-I0 scale range which would be suitable for the majority of cigarettes would seriously reduce the accuracy of measurements on highly ventilated low-"tar" cigarettes where very low CO concentrations may be. encountered. Also, measurements of the low concentrations in fractional puffs and in clearing puffs will have reduced accuracy. On %he other hand, the selection of a lower range scale may be too low for carbon monoxide concentrations in many cigarettes, particularly in the latur puffs. &l ==

- 8 - The equipment used, such as small diameter tubing and miniature manifolds, is inherently unreliable. Speci- fically, the small diameter tubing which is incorporated to minimize dead volume may create pressure problems in trans- ferring samples to the IR cell for analysis, This would translate to variability in analyzer response. Also, puff- by-puff analysis will require that a tremendous number of samples pass through the small tubing and %he IR ce~l. This increases the chances of fouling and clogging. Frequent cleaning will be required along wi~n careful attention to cell alignment to protect the accuracy of the end determination. THIRD, even from the limited information presented, it is clear that there are numerous deslgn deficiencies in the machinery described. For example, it has been reported in the literature that CO delivery of cigarettes increases with increasing puff number and increases during the puff. From these observa- tions, it is unlikely that there is such a thing as a reliably representative portion of a puff applicable to all cigarette brands and tees. The FTC machlne allows one second for "the gas phase to thoroughly mix" prior to exhausting the gas to the [R analyzer. This is an extremely ~hcrt time if one expects ~q the smoke of all tlr~es of cigarettes to become thoroughly mixed O prior to analysis. It is probable that complete mixing of the O gas from a given puff may not occur within the one second al- lowed for this purpose. This may be even more critical for

- 9 - fractional puffs. Obviously, incomplete clearing of the system will permit mixing of the gas from two different cigarettes, The new machine is mechanically complicated and may be more prone to breakdown or malfunction than that Phipps and Bird machine currently used by the ETC and the Tobacco Insti- tute testing laboratories. Specifically, it can be expected that the miniature components, such as solenoids, t~bing and fittings to minimize dead volume, will not funotlon properly and will Introduce inaccuracies and pose a maintenance problem if this device were run on a daily basis. Also, the Merfeld-Pillsbury machine uses plastic syringes rather than glass syringes which hav~ worked over a long period of time. There is no basis for concluding that th~ plastic syringes will perform as well, and they may pro- duce leaks. Both a thel~isto~ and a string cutoff means are mentioned but ne preference was indicated in the paper. The undersigned have previously explained why the thermistor cutoff should not be used. Since a gas flame from a hand-held lighter is pro- posed, it is probable that variable amounts of carbon monoxide from the flame will be drawn into the first puff. This could have a significant effect on the analysis of cigarettes yield- ing very low per puff ccncennrations of carbon monoxide. Cl