Tobacco Documents Online

Accession number 1657 Lori//ard Researcii Center Greensboro A CRITICAL REVIEW OF "TOWARD A LESS HAZARDOUS CIGARETTE," RESEARCH FROM NCI'S SMOKING AND HEALTH PROGRAM Submitted by: R. W. Slaven Report number: Date:May ag , 1981 Summary or Abstract: Results of toxicity and tumorigenic activity tests conducted on experimental cigarettes are summarized. Several of the statis- tical analyses were found to be unreliable. The implications of findings on nicotine, hydrogen cyanide and acrolein are discussed. /fr XC: Dr. A. W. Spears Dr. Mr. Dr. Ms. Dr. F. C. R. M. H. J. L. B. S. J. Schultz Tucker Patterson Ireland Minnemeyer Library

INTRODUCTION During the 1970's the National Cancer Institute, under the auspices of its Smoking and Health Program, conducted a series of experiments designed to evaluate the relative effects on labora- tory animals of smoke and smoke condensates derived from especial- ly prepared experimental cigarettes. These experiments,generally entitled "Towards a Less Hazardous Cigarette," were carried out in four distinct steps from their inception in 1970 to the issuance of the last report in September, 1980. The data and conclusions in the public reports (1-5) are the main subjects of this review. Reference is also made to a separate study (6) which is concerned with results obtained using cigarettes equipped with experimental filters. EXPERIMENTAL OUTLINE Experimental cigarettes were prepared by a different major American cigarette manufacturer for each set of tests. All con- structions were assigned a code number for subsequent identifica- tion. The variations incorporated in these cigarettes are outlined below. The cigarettes were analyzed at several different laboratories for major chemical constituents and physical parameters of the con- struction, blend, smoke and smoke condensate. The major tests con- ducted to determine the effects of smoke were in vivo and in vitro ciliary transport inhibition and inhibitions of phagocytosis by alve- olar macrophages. The effects of the condensate were tested by skin- painting bioassays on mice and cytotoxicity studies (for descriptions of these tests see Appendix A). Statistical analyses of the data generated were carried out independently of the experimental labora- tories. Summary of Series #1 Cigarettes Results The standards in this and subsequent studies were the 1R1 cigarette developed by the University of Kentucky and a standard experiment blend, SEB, prepared to represent an average unfiltered American cigarette marketed in 1970. The composition of SEB is given in Table 1. A total of 21 modifications were prepared for this first study. Table 1 Composition of SEB Glycerine 2.80% Invert Sugar 5.30 Flue-Cured 32.54 Burley 20.04 Maryland 1.06 Turkish 11.09 Reconstituted Sheet 27.17

- 2 - The construction was altered by changes in paper porosity and the tobacco cut. The blend itself was modified by variations in the stem and leaf content, incorporation of blend components into the reconstituted sheet (RTS) and the addition of nitrate. The differences in test results generated by these variations were not great. In fact, only the mouse skin paintings with smoke condensate gave statistically different activities. of the seven variations which gave significantly lower tumorgeni- city than SEB-l, six involved cigarettes made completely from different types of RTS. The seventh was a cigarette made en- tirely of stems (Ref. 1, pp. 11, Table 4). Some improvement was found with use of high porosity paper. Low porosity paper and lamina only blend scored very poorly being twenty-first and nineteenth respectively, in rank by number of surviving mice. Doubling the nitrate content had no significant effect. Correlations were drawn between the activities of the con- densate and the characteristics of the blend such as TVB's, total nitrogen and nicotine content, and physical characteristics of the cigarettes such as average weight, burn rate, pressure drop and others. The results were, in general, as expected with higher activity related to increased tar yields (Ref. 1, pp. 132, Table 15; pp. 133, Table 18). The vapor phases and smoke condensates were analyzed for chemical constituents and the correlations of these constituents derived. In the condensates,nicotine, isoprene and palmitic acid showed the highest correlations with increased ad- verse activity (Ref. 1, pp. 134, Table 22). Interestingly, poly- aromatics and phenols did not correlate as well. From the gas phase only HCN was shown to have a positive relationship to in- creased activity (Ref. 1, pp. 138, Table 24). The question of the appropriateness of these gas phase analyses will be discussed in another section of this report. Summary of Series #2 Results The variables in these experiments included weight of the cigarettes, aeration of the burning zone, use of high and low nicotine tobaccos, the concentration of fertilizer used in culti- vation, and fatty alcohol treatmen for sucker ontrol. Addition- ally, the non-tobacco fillers NSM ~ and Cytrel were tested. These materials are referred to as ASM-A and ASM-B, respectively, in the report (Ref. 2). In this series the toxicity tests gave a considerably larger range of.results than was seen in Series #2. This is especially apparent'with ASM-A and ASM-B. ASM-B had the lowest activity in the inhibition of ciliary transport, inhibition of phagocytosis and cytotoxicity tests. ASM-A was third best in the first two of these and second in the last (Ref. 2, pp. 92, Table 1). It was in the skin-painting bioassays where a large difference between these materials was seen. With ASM-A the number of mice to exhibit tumors was only 13 to 18% of the standard, whereas, ASM-B recorded 36% to 40% more. ASM-A was the best material tested and ASM-B the worst. Con- densates from both cigarettes were high in polycyclic aromatics. 89771784

- 3 - ASM-B condensate was higher in phenols than ASM-A but not out of the range of ordinary tobacco (Ref. 2, pp. 46, Table 8). Thus, no good correlation for the high adtivity of ASM-B in the mouse dermal bioassay was found. The other variations in blend and construction were of little significance with the exception of the use of high and low nicotine tobaccos. As in the first report, increases in adverse physiological responses, including tumorigenicity, were found to correlate with , increased nicotine content of the smoke and smoke condensate. Summary of Series #3 Results Changes in paper porosity, filter design and the effect of several additives, including sugar, humectants and cocoa, were tested in this series. Additionally, ASM-A and ASM-B were retested. The use of additives gave the most interesting results from this series. The addition of magnesium nitrate at a 5.7% w/w level gave a reduction of tumor bearing mice in the skin-paintings of 24% to 48% depending on dose level (Ref. 3, pp. 137-8, Tables 2 and 3). Un- fortunately, this compound had deleterious effects in the cilia toxi- city and cytotoxic potency tests (Ref. 3, pp. 168, Table 1). Cocoa powder exhibited a positive correlation with adverse effects in all three tests (pp. 33, Table 10 and pp. 168, Table 1). The deletion of sugar or glycerin had no statistically significant results. The use of high porosity paper and an air dilution filter gave very low activity in the skin painting bioassay (Ref. 3, pp. 36, Table 12). However, very little beneficial effect from this con- struction was noted in either the ciliatoxicity or cytoxicity tests. The retesting of the artificial smoking materials gave very similar results to those found in Series #2. Summary of Results of Series #4 The majority of the cigarettes prepared in this series consisted of refinements on the variations tested previously, such as nicotine content, paper porosity, bright and burley content and repeated tests on ASM-A and ASM-B. No great differences from the previous results were found. New variations included different processes for RTS production, changes in the puffed tobacco and the use of insecticide-free tobacco. None of these changes gave results significantly different from the standard. A cigarette made entirely from expanded stems displayed the lowest correlation with tumorigenic activity in this series. This result was not surprising in view of the effect shown by an all-stem cigarette in Series #1. Summary of Experimental Filter Results Tests on a variety of filter constructions were carried on during the mid 1970's. This work dealt only with the toxicity tests and no dermal bioassays were done. The results were reported 89771785

4 separately (6) with no referencing to the work already discussed. The variations in filter design included the use of cellulose, cellulose acetate and charcoal; addition of acid, base or oxidi- zing agent; and air dilution. It was found that addition of po- tassium bicarbonate, polyol or polyolimine to a cellulose filter and air dilution all contributed to an overall lowering of smoke toxicity. Analysis of Experimental and Statistical Methods While in general, this study was done in a highly competent manner there are certain aspects which bear close scrutiny. One of the overall problems with the program was its size and duration. The data was basically collected on variations of a 1970 vintage unfiltered cigarette. Of the 118 different cigarettes prepared in the main set of work exactly 5 had filters (only 2 being air diluted) and 3 had high porosity paper. Thus, much of the data generated does not bear directly on the concerns of the 1980's marketplace. The use of additives to mitigate deleterious effects of ciga- rette smoke was one facet of this study. Yet the work done with nitrate is less than complete. Previous to this project several reports had appeared in the literature which indicated that addi- tion of nitrate to the tobacco blend suppressed tumorigenic activi- ty in dermal bioassays. Cigarettes prepared in Series #1 with 0.71% nitrate, approximately double the natural level, did not ex- hibit this effect. Yet cigarettes with 5.7% nitrate made in Series #3 were quite effective in this regard. No comment is made in the report summary about these results nor was any attempt made to find an effective threshold. The addition of nitrate has the unfortunate side eff ect of increasing the potency of the smoke in the toxicity tests (Ref. 3, pp. 173, Table 4). This problem was overcome by addition of 6.9% zinc oxide. No comment was made on this result and the matter was not investigated further. In Series #1 the Code 13 and 14 cigarettes contained only stems and scored very well in the tumorigenicity test. The impression is left that this was due to the low tar levels of these cigarettes. The bearing that the high nitrate level of the stems had on this result was not analyzed. In summary, the whole topic of nitrate addition and the use of zinc oxide to suppress toxic side effects was investigated only sketchily. The:handling and reporting of the nitrate data is symptomatic of a problem in the construction of the reports themselves. The overall report summaries are extremely short and present only the barest highlights of the work. This is probably to insure the ob- jectivity of the report but, in fact, makes deciphering'the work more difficult. This is especially true in regard to relating work from different series. Such relations must be minimized due to vari- ations in blends and construction. However, work on areas such as nitrate, paper porosities and artificial tobacco substitutes was, carried on in several series and necessary appropriate commentary is not given. The summaries also give no rationale for changes made during the course of the program. For example, in the first series of experiments the ciliatoxicity tests are done both in vitro and in vivo.i The in vivo testing was then dropped. No rationale 89771786

- 5 - for this major change is given in any part of the text. Also, one of the primary reasons to repeat the artificial smoking material work was a problem with the solubility of the conden- sates. This is discussed in the experimental section on prepara- tion of the condensates but never in the summaries. Another pro- blem is that the summaries deal almost exclusively with results from the dermal bioassays. The toxicity tests are barely mentioned. There are several compounds such as carbon monoxide, nitrogen oxides and formaldehyde which increase potency in the tumorigenicity ' test but decrease it in the various toxicity tests. The tumori- genicity results are the only ones mentioned in the summaries. There are certain compounds where the whole concept of correlating their concentrations to certain test results is highly dubious. For ex- ample, carbon monoxide concentration is measured in the tobacco smoke. Yet the tumorigenicity test is run with smoke condensate. Clearly, any mathematical relationship drawn is essentially meaningless. Additionally, the analytical data, of several important chemical con- stitutents is of low quality. Methods available for measurement of such smoke components as formaldehyde and nitrogen oxides are not as reliable as some others. This results in widely varying correlations between concentrations and potency results. In the case of oxides of nitrogen the correlation of concentration to increased inhibition of cilia transport is 0.47 in one series and -0.05 in another. The former number indicates that inhibition is moderately dependent on oxides of nitrogen concentration while the latter means there is essentially no relationship. Which of these is to be believed? Similarly, the formaldehyde concentrations in the cytotoxicity tests range from 0.85 in the experimental filter work to -0.04 in Series #4. These anomalies demonstrate that the data in these reports must be understood in terms of their experimental and chemical reliability. There are many statistics generated that have little or no validity. This point can best be emphasized by noting that one of the most positive correlations to tumorgenic activity was drawn to the moisture content bf the smoke. The nomeclature used in the statistical sections can also be confusing. In the Series #1 dermal bioassay results, the various blends and their components are related to two numbers PF and PFH. PF is the percentage of mice who survive the test tumorfree by staff observation. PFH is the the percentage tumorfree by histophathologidal testing. PFH is a 28% higher than PF in Series #1. In Series #2 this diff erence is 19% but both numbers are called PF and only the data tables are labelled as to whether PF is from visual or verified data. The visual data is dropped altogether from Series #3 and #4 and all PF's are based on verified data. This makes it very easy to compiie two completely separate sets of correlations. The discre- pancy can be serious. For example, in Series #1 the Code 16 cigar- ettes, which represent a simple variation in RTS density ranked second best by PFH and tenth by PF. In addition to simple correlations of potency levels to indivi- dual physical and chemical parameters, correlations are also drawn to groups of such parameters. The.groupings were generally the chemical composition of the blend, smoke and smoke condensate and the physical characteristics of the construction. In the first two series the re- lationships between members of different groups were analyzed. Thus, 89773787

6 the dependencies of such things as nicotine or carbon monoxide yield on the burn rate or paper porosity were examined. These statistics are interesting in and of themselves but are unfortu- nately not included in the reports of the last two series. CONCLUSION It should first be noted that the observations in this report are in no way inclusive of all the data and results developed in the Smoking and Health Program. Only some of the more important results and obvious omissions and/or inconstituencies have been noted. Several additional points should be made. First, in all these tests hydrogen cyanide exhibited the most consistently high correlations with toxicity. Thus, steps to eliminate this com- pound are of high importance. The pair of hydrogen cyanide and acrolein,also showed very high correlations even though acrolein itself did not. If this is evidence of a synergistic effect then efforts to remove acrolein could be very beneficial in cutting the overall toxicity of smoke. The removal of hydrogen cyanide was one of the aims in the experimental filter work. It was shown that this could be partially accomplished by incorporating potas- sium bicarbonate into the filter. Such filters are also of interest for another reason. Examinations of the data reveals that on a per puff basis the potassium bicarbonate filter increased the nicotine yield 55%. This is one of the aims of the current nicotine project. Preparation of experimental cigarettes with such filters will be commenced in the near future. There is one physical characteristic of the cigarettes in this study which correlates very well with tumorigenic activity; this is the static burn rate. In fact, this correlates better than the tar yields do. Thus, it is possible that the burn rate alters the composition of the tar and its subsequent physiological effect. This could, for example, reflect a lowering of the relative amounts of high molecular weight polyaromatics. Research into this area could be very valuable. Comments are currently being made in the press about the higher nicotine/tar ratios in today's cigarettes and the concept of nicotine being a "co-carcinogen." The inference is then made that little benefit is derived by switching to a low tar cigarette. If it could be demonstrated that the high burn rate of today's cigarettes alters the tar composition in a positive manner an effective answer to the previous argument would be in hand. Finally, it should be noted that the data from the Smoking and Health Program was seminal in the evolution of the "co-carcinogen" concept for nicotine. Any discussion on this ill-understood area must include this work. Since work is still continuing on "co-car- cinogenic" activity, for example, Reference 7, this research must be closely monitored by the industry and its impact on public attitudes assessed. oD ~ N ~ GO Go

REFERENCES 1. Report No. 1. "Toward Less Hazardous Cigarettes," Gio B. Gori, 'ed., DHEW Publication No. (NIH) 76-905 (1976). 2. Report No. 2, "Toward Less Hazardous Cigarettes," Gio B. Gori, ed., DHEW Publication No. (NIH) 76-1111 (1976). 3. Report No. 3, "Toward Less Hazardous Cigarettes," Gio B. Gori, ed., DHEW Publication No. (NIH) 77-1280 (1977). 4. Report No. 4, "Toward Less Hazardous Cigarettes," Gio B. Gori, ed., DHEW Publication No. (NIH) - (1980). 5. Report No. 5, "Toward Less Hazardous Cigarettes: Summary," DHEW Publication No. - (1980). 6. "Chemistry and In Vitro Bioassay of Smoke from Experimental Filter Cigarettes," G. B. Gori, ed., DHEW Publication No. (NIH) 76-1076 (1976). 7. Danbury Report "A Safe Cigarette?," G. B. Gori, F. G. Boch, Cold Spring Harbor Laboratory (1980.

Appendix A Description of Testing Procedures Mouse Dermal Bioassays - ICR Swiss•female mice were the test animals for this study. The mice were caged in groups of ten after random selection. Several control groups were established. The first received no treatment at all, the second merely had their hair clipped. Hair clipping and skin painting with acetone comprised the next control and the last was painted with benzo (a) pyrene. The test animals were painted with two different levels of condensate, usually 25 or 50 mg, on a daily basis. Data were kept on survival rate, latent period of tumors, tumor volume, number and total dose of smoke condensate, acetone, or benzo (a) pyrene, group average body weights and experimental day of death. During the study, animals which died or were sacrificed were re- cropsied and all tissues fixed in formalin. The duration of these bioassays was 18 months. Treatment was discontinued at this time and surviving animals sacrificed and recropsied. Ciliatoxicity Tests In Vitro - The trachea was removed ~rom chickens immediately after sacrifice and mounted on a Lucite R trachea holder. The trachea was rinsed with Tyrode's solution to remove mucus and placed in an exposure chamber at an angle of 90 from the horizontal. A solution of 12% eggwhite in Tyrode's solution was pumped on to the lower end of the trachea to promote a constant and reproducible level of ciliary activity. Small numbers of tracer particles (lyco- podium spores or soot particles) were applied to the trachea and the intrinsic ciliary activity measured by time necessary to move the parti'cles 5 mm. The trachea was then exposed to the mainstream smoke of one experimental cigarette and the 5 mm movement time re- measured., Comparison of the movement times gave the percent in- hibition of ciliary activity. In Vivo - White Leghorn hens were anesthetized and strapped into an especially designed holder. The inner surface of the upper part of the trachea was exposed by manipulation of the larynx and movement of tracer particles observed as in the in vitro study. Inhibition of Macrophages - Macrophages harvested from rabbit lungs were exposed to a laboratory strain of Staphylococcus albus. The removal of the bacteria from the test suspension was measured after 48 hours. The test colonies were exposed to one 35 ml puff of smoke du'ring this period. Controls included colonies which contained either no macrophages or no smoke. The differences in bacterial counts were taken as a measure of the inhibition of the phagocytic activity of the macrophages. Cytotoxicity - Inhibition of mammalian cell growth in culture was assessed in this test. The KB human tumor cell line was exposed to aliquots of aqueous solutions of smoke condensate. After 72 hours of growth at 37°C, the cell sheets were washed, digested, and pro- tein levels measured. Controls included cultures with no smoke con- densate or no cells. -