Tobacco Documents Online

PHiILIP M'ORIRIS U'. S. A. I N T E R - 0 F F I' C E C' 0'~ R R E S P 0 N f, E N I C E i, S J' R1'CHIMiOND, VIRGINCA To: . Mr. W. McDowelll'. p,rom: . F. Ryan Subject: . PANDORA (Finger Occlusion) C 0 N F I D E N T`"I A L The final report fram NeuroCommunitations Research Lab!(NCRL)~ on finger occlusion„ titled "Investigation of Hand Behaviors DuringiCi'lgarette Smokingg which may affect Tar and Nicotine Delivery", is attached. It is based on unobtrusive observations of 113 people engaged'in watch.ingiand evaluatiing TV co,:nmerciais (38 full--flavored, 319 flavor-llow, and 36 ultra-low deiiviery smokers) who:took a total of over 1100 puffs while smoking their own brand of cigarettes. Their behavior was recorded'on TV tapes and later scoredl by three independent judges, who were iln g_~nera:l agreement about the inciidence& and duration of finger occlusions, usiing very conservative ("worst case"), criteria. RESULTS: ----_ -'" _Forty percent'of__alll_ recorded_puffs_ were reported to; be subject to at least some irocidental finger occlusion of dilution holes. The incidence of coverage does not seem related to the ,del ivery of the cigarettes. The duration of coverage during puffs averagedl24% for the study as a whole (e.g. 24% of the smoke generatedi was generafea--WFh-ilie fingers were on holes, 76% while fingers were off holes) and was NOT greater amongi ul'tra-lowAel1very smokers than among smokers of other delivery levels. The proportion of coverage was in fact lowest among the ultra low delivery smokers. These observations aren in direct confliet with the hypothesis that the ultra-low, tar smokers willl learn to smoke differently thian:smokers of higher delivery products. That they in fact smoked in a similar manner is confirmed by the observed number andlduratiiorn of puffs taken, which were generally similar across groups, as wel l as by the f inger coverage data. To evaluate the effect of fingers covering.holies an internal study was conducted by W. Houck andl J. Ne,pomuceno in wh ich 31' smokers were asked to intentional ly cover dillution holies with paint smeared fingers. The number of holes or relative area of ESP band, in the case of Marlboro, coveredlby paint was measured for a number of P.MI. brands, and the effects of such blockage on dilution was calcullated. Cigarettes were then machine smokedlwith and without such dillution blockage. Such smoking is analagous to that reported by Kozlowski as cited in the recent Consumer Reports article,, but differs in that a) the holes were not completely occluded but occludedd only to the exte,nt that the paint smear d'ata suggested was appropriate, andlb), the smoking conditions used FTC standardK of puff count, p~uffduiratio~n, interpuffintervalis, etc., rath~erthan t~h~eclaimed bigger and longer puffs at shorter intervals used by Kozlowski. There were, of course, some increases in FTC tar observed when all puffs were taken wlith some of the holes bloeked, as shown in the Tablie I below:

-2- Table I Hole Blockage Effects on FTC Tar (Houck &. Nepomucino) Brand Extent of Hole Blockage FTIC Tar (mg/cig) Holes Blocked (as per paint smear) FTC Tar (mg,/cig), Holes Unblocked (normal C1 data) Marlboro 85 65% 16.6 16.3 M'arlboro Liohts 85 58% 10'.7 10.2 Merit 85 47% 8.8 7.9 Mer i~t Ultra Lights, Cambr iidge 85 44% 50% 7.0 2.5 4.5 1. 0' We then calcullated the effects of the NeuroCommunications hole coverage data on theestimate~d delliveryof FTC tar to,thesmokers. We assumed that any observed incidence of hole coverage actuallly produced the maximum coverage observed in the l aboratory -an overest imate of the true s ituat ion wh ichiis b ilased towards i'nfl'ating tar del ivery numbers to the h ighest vallue 1 ilkely. Thuis a f iinger placement whichimight--had we been able to measure it with compIete accuracy-- have covered' 10q of a Marllboro's holes was treated as if it hadi in fact covered 65% of its holes, thus rai'sing~ Marlboro's norma~l 16.3' mg del ivery rate to a. 16.6 mg rate during the coverage time. To get the caliculated delivery effect of observed hole coverage on ai Marlboro wee thenitook the proport ioniof puff ing t ime that holes were blocked and~ unblocked for smokers of full-flavored cigarettes from the data of the NeuroCommuniications report (26.6% ofth~e puff ing, t imehad at lea~stsome holes occluded byfull l fl'avored smokers, leaving 73I.4/ of the time with hioles not occluded)„ multipl ied 0 these proportions times the appropriiate delivery levells, and then summed the two jV component products to get the r•el at i've tar for Mar lboro: (. 266 x 16. 6)+ (.734 O x 16. 3) = 116, 4 mg,. ~ These cal'cul'at ions were made for eachi of our pr inc ipal 85mm products of ~ W int'erest, us ing the observed proport ions ofcover•ed andunicovered t imes and the, data frorn Houck and Nepomucenio. ~

-3- Table 2 Calcullated Fffects, of Finger Coverage of 'ibles, Based on Laboratory Datai app.lied to Fiield Observations Brandl % Blocked Time % Unbllocked Tiime Tar Blockedl (mg(cig) Tar Unb.locked (mg,/'ciig) Total Calculated (mg/cig) Difference Mar lltroro~ 85 26. 6 73. 4 16.6 1', 6. 3' 16.4 +.1 Marll. Lt 85 241.4 75.6 10.7 10.2' 10.3 +.1 Merit 85 24.4' 75.6 8.8 7.9 8.1 +.2' Merit Ultra 21.0 79.0 7.0 4.5 41.9 +.4 Cambridge 85 21.0 79.0 2.5 1.0 1.3' +.3 The Table 2'data show, the calculated deliveries taking finger blockage into account ini the next to: the last column, with the last column representing the increase from a co~mpl'etely-unblocked-on-alil-oecas ions cond it ion to a part ial ly- blocked-on-some-occas ions cond it iion. As can be seen, the effects of the coverage are trivial, and would in all likelihood suggest no change in the reported FTC tar numbers, which are giveni in whole numbers rather than decimals., In sum, th~ere is 1 ittle that is disappointing and much that is encouraging in this study. If the "worst case" observations have so liittle effect, then a lless than worst case interpretation may be wasted! effort. We wiilll„ however, follow up on the duration probilem, and will check all their observations here using more sophisticated measurement procedures. Furtherinfo~rmati~on on noni-P"~-1 brands will be available once we have scrutinized the NCRL raw data.. FJR':yll Attachment