Tobacco Documents Online

-3- Tf at any subsequent exaalnaciou a cumour is judsed Co have disappeared a reKression is recorded. For the purposes of cabulaCin8 the ~nc£dence of tmnours, the time of appe~arance of a Cumour is taken as the tlme of appearance of the papilloma, recorded as described in the previous paragraph. The condition o£ maliKnancy is noC diagnosed unCll autopsy; the time of onset o£ maliSnancy ~s taken as the t~me o£ death of the animal carryLng the cumour. Tumours are resarded as malignant if they produce mmtastases, or Lf they /~filtrate the dermis and penetrate the tunics muscularls, or i£ they infiltrate the dermis and penatrsta the basal mmabrana of the epithelium. Occasionally tumou=s sho~td 4u~iltracion oE the da~mis, but ~[t could noc be cl~rly escabllshed thaC penmcration of the basal membrane had occ~ed. Such turnouts were recorded as sho3r~a4 the f£~st siKns of real £8~ancy. The results o~ the expe=Lmmnt are shown in Tables I to 18. These ~bles show Che ~c~denca of tumour-bea:iuqL a~imsLs ~n the I8 emper/a~ntal Stoups of animals. The c~sour-b4~r~ ma/~aSs /~ mn axp~r/~mncal poup msy be allocated to the £oLlowiJ~ seven clasaas:- 1. kn~nals hey/Jag no lesion. 2. An~nals haviz~ lesions all of vhlch pex3naneuatly reSrassad. 3. An/~aZs having lesions all o£ which are uncharmcterlsed. A:~Lma~s hevi~s lesions, sows o£ which penmmeuCZy :egressed the :ama£nd~c bataK unchazacCe=£sed. 5. An~nals havlz~ at least one coutirmed amour. 6. An/s~Zs hev£nZ at least one turnout show~ng the £irsC siSas o~ mli~ancy, but no malignant cumomr. 7. An/~aZs heviJzS at least one malignant t~ur. BAT Co LTD - MINNESOTA TOBACCO LITIGATION OO r',o

INDEX HAS INDICATED GAP IN BATES RANGE HERE

V An animal must fall into one and only one of the classes 1 to 5. Classes 6 and 7 are sub--divisions of class 5. The lesions borne by an animal in an experimental Stoup may be allocated to one of the follovinS five classes:- 1. Lesions which permanently resress. 2. Lesions which are tmc~eracterised. 3. Les~ons which are beniBn. A. Lesions which show the first si~na of mal£Snancy. 5. Lesions which are maliKnant. & lesion must fail into one and only one of these five eI&sses. In order co analyse the incidence of t~nour-b~aring anlmals, the t~ne of appearance of ttmour-b@ar~ animals in the ex?er/ment is tabulated as in Tables I to 18. The firs= column shows the duration of the experimene divided into 4-reek periods. The times recorded in the date for each animal e.K- time of appea~anne of turnout, time of death, etc., are recorded in weeks. The second column is defined as the number of animals alive end tumourless st the beEinnln8 of the time period. The third column is defined as the number of aIL'|.msls becoming tumour-boaring durinS the 4-week period. Col,~ A is defined as the number of animals dylng w~thout becoming turnout-be&tins du=inS the #-week period. The fifth eolutn is defined as the mmber of animmls allve vithout a maliE~ant turnout at the belinnLng of the &-weak period. The sixth column is defined as the ntmbe: of animals which have no mal i~Lnt t~mour at the bqi=ins of the time--period and which produce at toast one maliKnanC turnout durinE the 4-~sek period. Since the onset of mal~ancy is dafinld is the date of death o£ an animal b~ar~ng a malignant turnout, columns six may be equivalently defined as the number of an/male dying with a malisnant BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-5- tcuuour during the 4-week t~ne period. (SLmilerly column 5 may be equivalently defined as the total number of animals alive st the beginning of the 4-week period. ) The seventh column is d~ined as the number of ani:als dying vlthout a malignant turnout during the 4~eek period. Thus if the rows of the table ere numbered from i to i then for columns 2, 3, 4 and 5, 6, 7 Ni+I " Ni - DIi - D2i where Ni is the number of animals ac risk aC the beginning of rime period £, D1i is the number of Qls becoming ctanour-b~ar/ng during rime period i, and D2i is the number of animals dying during eke period i without becomins Cumour-bear£ng. N1 is the number of animals encer£ng the experiment at the beginnins of week 1. The animals ~n an axperin~ntml group enter ohm tables according co their occupancy of the seven classes described above. Thus aninuals from classes 3, 4 and $ will ante: column 3 of the Cable; az~mals from classes 1 end 2 will enter column 4 of the table; animals in class 7 will enter column 6 of the table. Columns 3, 4, 6 and 7 are totalled. These tables form the basis of the statistical analysis. 5ubsidlary tables of the numbers of --4m-ls in the seven classes and of the numbers of lesions in the five class., are shown in Tables 19 to 38. Table 28 is • stamnary of Tables 19 Co 27, end Table 38 is 8 sugary of Tables 29 to 37. Zt v~ll be noted Chat Tables i to 18 frequently show that the numbers of animals in the experimental groups were greater Chart Chose stated at the beginning of section 2. These extra animals are called "I~PLACZIEHT AH~S" and were taken into the experiment Co replace animals dying during the first 4 weeks of ceat~enc. The number of replacement animals used in r_he 18 Kroups were as follows'-. BAT Co LTD - MINNESOTA TOBACCO LITIGATION O

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-6- 25 m8 75 30 c.p.i. i I0 g6 60 c.p. £. e,, l 4 3 120 c.p.i. 1 4 6 30 ¢.p.i. 4 2 3 B7 60 c.p.i. 120 c.p.i. 4 3 3 A torsi of 70 replacement animals were used, maki~ an overall total of 1582 animals uJed in the experiment. 4. ANALYSIS OF RESULTS Using the Weibull distribution, common values of k and w, and separate values of b were fitUed to the da~a from the eighteen groups of animals. A fit was obtained for all turnouts, and for malignant tumours. The est/-eates of ~he parameters are shown in Tables 39 and 40. The associated values of S and V ere shown in Tables 41 and 42 IS and V are the Weibull statistics and arm defined at the £ooc of the tables], t goodness of fit test for these parameter values was carried out, the results beinE shown in Tables 43 and 44. Inspection of Tables 43 and 44 shove that the values of the parameters given in Table 39 do not fit the data too well (X2 " 27.63 with 6 degrees of freedom, significant at 99.98Z) whereas the parameter values given in Table 40 do fit the data re~eouably well (X2 m 11.43 with 6 degrees of freedom, not siEnificanC). A closer inspect£on of Table 43 shows that the lack of fit is occurrlnE dur£~ the experimental weeks 65 to 80, Coo few ct~our-baarinll animals being expected duxin8 weeks 65--72, and coo -,any being expected in weeks 73 Co 80. The total expecLation for this time period, 191.4, is however fairly comparable wi=h chose observed, 214. r~ ~D BAT Co LTD - MINNESOTA TOBACCO LITIGATION

INDEX HAS INDICATED GAP IN BATES RANGE HERE

-7- It is difficult to identify reasons for the lack of fit of the parameter values of Table 39 co the a11-tumour data. Inspection of the data (Tables I to 38) shows that a feature of this experiment is the high toxicity of the col~Ita~,sllUeS used, as Lnd£cated by the death rate in the first 8 weeks of the experiment. The deaths in the first 8 weeks of the expeximent are shown in Table 45 and these deaths expressed as a percentage of the number of animals at the beginning of the experiment (including replacements) are shown in Table &6. Analyses of variance of the dace of Tables A5 and 46 were carried out, the results being shown in Tables A7 and A8 where it can be seen chat ali the main effects (tobaccos, cuts per inch, dose levels) are highly significant. Also, as would be expected, comparison of Table 47 with 48 shows thaC expressing the numbers of deaths in the first 8 weeks of the experiment as a percentage of the s~arting number in each group increalmS the significance levels (the F-ratios) of the main effects. The partitioning of the m,,,in effects indicates than the dose effect is m-tlraly log linear, and that the c.p.i, effect is h~hly Iog linear but with a significant degree of non- Einearicy. This analysis of the early deaths shows that (a) ~he condensates used were unexpectedly toxic, giving rise to very high death races (up to 64.5Z; see Table 46) in the first 8 weeks of the experiment. (b) the condensate from cigarettes contalnln6 tobacco B6 was sisw/ficmntly more toxic than that from cigarettes containing tobacco BT. (c) decreas/~g the cuts per inch siEuificantly increased the toxicity of the condensates produced by the cigarettes. BAT Co LTD - MINNESOTA TOBACCO LITIGATION r',,o ",.O

INDEX HAS INDICATED GAP IN BATES RANGE HERE