Glass Fiber Contamination of Cigarette Filters: an Additional Health Risk to the Smoker?'

Home , Eclipse (cigarette), Premier (cigarette)

Vol. 7. 96 7-979. November 1998 Cancer Epidemiology, Biomarkers & Prevention 967

Glass Fiber Contamination of Cigarette Filters: An Additional Health Risk to the Smoker?’

John L. Pauly,2 Hans J. Lee, Edward L. Hurley, 118/120, 98%; Menthol, n 120/120, 100%). Likewise, K. Michael Cummings, Joel D. Lesses, and 99% of NEW Eclipse had glass fibers on the redesigned Richard J. Streck filter (Regular, n = 119/120). In contrast, glass fibers Department of Molecular Immunology [J. L. P., H. J. L., J. D. L., R. J. S.], were never observed on the filters of conventional United Microscopy Ultrastructure Facility [E. L. H.], Cancer Control and States filter cigarettes that had been used as controls Epidemiology [K. M. C.], Roswell Park Cancer Institute (a unit of the New (n 0/120, 0%). In a study of Eclipse (n = 60), the York State Department of Health), Buffalo, New York 14263 number of glass fibers contaminating the filter surface ranged from S to 55. Glass fibers as well as fiber fracture items [aspect ratio, <3:1 (e.g., particles, fragments, bits, Abstract chips, flakes, specks, and dust)] were discovered in the We report here the results of studies documenting the pack residue. The average number of glass fibers in the contamination of a cigarette-appearing smoking article residue of a pack of Eclipse was 7,548 (SE ± 3443; range, labeled Eclipse with glass fibers, fragments, and particles. 1,164 to 26,725 glass fibers/pack; n 7 packs). The thin Eclipse, a product of the R. J. Reynolds Tobacco and fragile glass fibers of the insulation mats had most Company (RJR), was commercialized in June of 1996. likely been broken and fragmented in the high-speed Eclipse is unlike conventional cigarettes in that, like its multiple-step Eclipse manufacturing operation. predecessor Premier, it is designed to heat and not burn Invariably, puffing on Eclipse discharged glass fibers and tobacco. The purpose of Eclipse was to simplify the glass particles from the filter into the smoker’s mouth. chemical composition and reduce the biological activity of Subsequently, the bioresistant glass fibers and the mainstream and sidestream smoke and to achieve a microscopic glass dust are inhaled and/or ingested. significant reduction of environmental tobacco smoke. In Contamination of Eclipse filters with glass fibers and Eclipse, tobacco pyrolysis is reduced by a carbon fuel rod glass dust poses a potential and unnecessary health that serves as a heat source for generating an aerosol hazard to uninformed consumers. Eclipse is a paradigm having nicotine and tobacco flavor. The carbon rod, at of the health danger that may be imposed by technically the tip of the cigarette, is insulated and bound with two complex tobacco articles and nicotine delivery devices wrapping mats of glass fibers. Recently, Eclipse has been promoted by an unregulated industry to smokers modified to address consumer complaints of burdensome worldwide, many of whom are addicted to nicotine and draw and off-taste. The redesigned Eclipse, which we who seek a less hazardous cigarette. have termed the NEW Eclipse, has an unconventional filter-appearing mouthpiece that consists of a cellulose Introduction acetate cylindrical bundle with a central hollow tunnel. In our analysis of Eclipse, glass fibers (length:width aspect In 1987, RJR3 announced Premier, and a market test was ratio, 3:1) were: (a) observed protruding from the tip; conducted from October 1988 through February 1989 (1-4). (b) identified on the white cigarette wrapping paper; (c) Premier was one of several smoking articles defined by RJR as viewed on the surface of the cork-appearing tipping “NEW CIGARETTE Prototypes That Heat Instead Of Burn paper; (d) found in the pack residue; (e) discovered lying Tobacco” (1-1 1). The stated objectives of these cigarette-ap- freely on the cut surface of the filter by both light and pealing devices, as set forth in an RJR monograph (10), were electron microscopy; (J) harvested from the filter with to: (a) simplify the chemical composition of mainstream and adhesive tape; and (g) displaced when Eclipse was sidestream smoke emitted by the “NEW CIGARETTE” (10); smoked mechanically. In a study of Eclipse that had not (b) minimize the biological activity of the mainstream and been removed from carefully opened packs, we observed sidestream smoke emitted by the NEW CIGARETTE: and (c) that 95% of the filters were contaminated with glass achieve significant reduction of environmental tobacco smoke fibers (Eclipse: Regular, n 114/120, 95%; Muds, n from the NEW CIGARETTE (see also RJR reports, Refs. 6, 12, and 13). The performance of Premier in the market evaluation was

poor (2-5, 7, 8, 1 1 ). Smokers rejected Premier because of Received 6/26/97; revised 7/20/98; accepted 7/27/98. off-taste, difficulty in lighting, and other problems (2-5, 7, 8, The costs of publication of this article were defrayed in part by the payment of 1 1), Premier was reconfigured to address these shortcomings page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. I Supported in part by NiH Grants CA-67827 and Roswell Park Cancer Institute Cancer Center Core Grant P30 CA16056. 2 To whom requests for reprints should be addressed, at Department of Immu- 3 The abbreviations used are: RJR, R. J. Reynolds Tobacco Company: SEM. nology, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY scanning electron microscopy: SEM/EDX, SEM energy dispersive X-ray micro- 14263. Phone: (716) 845-8538: Fax: (716) 845-8906: E-mail: pauly@SC3102. analysis: NTP, National Toxicology Program: PAH. polycyclic aromatic hydro- med.buffalo.edu. carbon.

(2-5, 7, 8, 1 1). After 3 years of development, Eclipse was available information, we conclude that fibrous glass materials

introduced (2-5, 7, 8, 1 1). Features of Premier and Eclipse have are carcinogenic. .. “ (17). been described in more than 20 U. S. Patents (9, 10), and patent protection has been sought in more than 1 10 countries (10). Materials and Methods In June of 1996, Eclipse began selling in Chattanooga, TN Eclipse, NEW Eclipse, and Conventional Cigarettes. (1, 3). Soon thereafter, the market test was expanded. In 1998, Eclipse from RJR was the test article of this study. The Regular, Eclipse is being sold in other United States cities. Additionally, Milds and Menthol Eclipse articles were purchased from 1996 Eclipse is being sold in different countries under various trade through 1997 from retail vendors in Chattanooga, TN. The names (Ref. 1 ; see “Materials and Methods”). In September of NEW Eclipse articles were bought from retail stores in Lincoln, 1997, RJR introduced a modified version ofEclipse to Lincoln, NE and Chattanooga, TN; these were procured in 1998. We NE ($2.69 per pack, including tax; Ref. 14). We have designated the neengineered Eclipse as the NEW Eclipse. When analyzed also Eclipse that had been licensed by RJR for sale in Sweden (Inside, Regular, and Menthol; Svenska Tobaks, AB, compared with the original Eclipse, significant changes in the Stockholm, Sweden), and in Germany (HiQ, Regular; German NEW Eclipse included: (a) improved lighting; (b) filter mod- subsidiary of RJR).4 Premier smoking articles from RJR were ification; (c) more tobacco taste; and (d) 80% less secondhand provided by colleagues who had purchased them during the test smoke; the corresponding value cited by RJR for the original market period from retail dealers. Conventional United States Eclipse was 90% (14). filter cigarettes were used as controls. The conventional ciga- Eclipse has the outward appearance and geometry of a rettes were of different: (a) brands (e.g., premium and dis- conventional cigarette (Refs. 1, 3, 4, 9, and 10; Fig. 1) and is count); (b) manufacturers (e.g., RJR and other United States obtainable in four king-size styles: Regular, Milds, Menthol, tobacco companies); (c) styles (e.g., Regular, Mild, and Men- and Menthol Milds (1, 3). thol); and (‘0 packaging (e.g., hard and soft pack). All had been Opening Eclipse with a razor (Fig. I) revealed that it is a purchased from retail merchants in Erie County, NY. multiple component, technically sophisticated smoking device having many items not found in conventional cigarettes (1-5, 7, Inspection of Eclipse and Conventional Cigarettes. Packs of 9). Essential to the design of both Premier and Eclipse was the Eclipse and conventional cigarettes were opened carefully in goal that tobacco pyrolysis was to be reduced greatly (6, 10). the laboratory and by personnel familiar with the purpose and This was accomplished with the use of a carbon fuel nod (Refs. scope of our investigation. Safeguards were used to protect both 1, 3, 4, and 6-10; Figs. 1 and 2, a and b) that, when ignited, Eclipse and conventional cigarettes from airborne fiber and would heat reconstituted tobacco to generate a low-smoke particle contaminants. A partial listing of the precautions used nicotine containing aerosol that imparts a tobacco taste (1, 3, 4, are as follows: (a) after unwrapping and discarding the cello- 7, 9, 10). phane film, the flip-top lid was reflected and taped to maintain Unlike a conventional cigarette, Eclipse does not burn the pack in an open position. The front aluminum foil was lifted down to the filter when smoked; the spent Eclipse retains its off carefully and discarded. The back aluminum foil was re- length (Fig. 1; Refs. 4, 9, and 10). When smoked, Eclipse moved by cutting with clean scissors. All manipulations were displays an ash-appearing tip that consists mostly of an insu- performed without contacting or moving the Eclipse. Particular

lation mat of glass fibers (Fig. 1 ; Refs. 1 , 4, 9, and 10). Eclipse care was exercised in not touching the filter, the major test site burns very little paper, which, in conventional cigarettes, is of this study; (b) Eclipse articles were examined immediately impregnated with burn accelerators and smoke modifiers (Fig. after opening the package so as to reduce the exposure and risk 1; Refs. 10 and 15). The burning (-900#{176}C; Refs. 4 and 10) to contaminants; (c) the filters were inspected for glass fibers carbon nod is insulated with two glass fiber mats (Fig. 1 ; Refs. without removing the Eclipse or conventional cigarettes from 4, 9, and 10). The primary function of the glass mats is to: (a) the pack. After inspection of the filters for glass fiber contam- secure the position of the carbon fuel rod; (b) insulate the inants, the top was closed, and the pack was stored in a burning rod; and (c) direct the heat from the rod to the adjacent stationary upright position in a closed plastic container; (d) tobacco column (10). Other components of Eclipse have been conventional cigarettes, which had been handled in a similar described in detail elsewhere (9, 10). manner, were used as assay controls; (e) for some tests, the The Eclipse filter resembles that of a conventional ciga- Eclipse and conventional cigarettes were taken out of the pack. rette; however, the filter nod has been truncated and is approx- In these assays, examinations for glass fibers and particles were imately one-third the length (Fig. 1; Refs. 3, 9, and 15). The conducted immediately after removing the Eclipse and conven- filter rod of conventional cigarettes consists of thousands of tional cigarettes; and (f) samples for SEM or SEM/EDX were Y-shaped cellulose acetate fibers (Ref. 16; see also Ref. 15). transported and stored in closed boxes and examined immedi- Y-shapcd cellulose acetate fibers that are similar in appearance ately after removing. are used in the filters of Eclipse and NEW Eclipse. Examination of the Tip of Eclipse. The tip of Eclipse was Observations obtained in our initial inspection of Eclipse inspected to assess the integrity of the two glass fiber insulation raised concern that the thin and fragile glass fibers of the mats. For this evaluation, a pack was inverted, and the paper- insulation mats may break during manufacturing and contam- board was cut carefully from the bottom of a pack with scissors mate the surface. To address this hypothesis, a study was (Fig. 2) without contacting or disturbing the Eclipse articles. conducted to: (a) determine whether the glass fibers and par- The tip was scrutinized with a stereo-zoom microscope. After tides are present in the pack residue and the surface of Eclipse; completing the microscopic examination of the glass mats and and (b) if present, is there evidence to assume with reasonable carbon rod, the Eclipse articles were removed carefully and certainty that the glass fibers would be inhaled and/or ingested individually with forceps from the pack and positioned at when Eclipse is smoked as intended. The rationale for under- taking these studies is derived in part from animal and human studies that have for many years addressed the health risks of inhaled glass fibers; in evaluating these findings, it was con- 4 Eclipse is also marketed in Japan (AIRS); however, AIRS was not available for cluded in 1994 that: “From our comprehensive review of the our study.

Fig. 1. Components of Eclipse.

Upper panel, nonsmoked Eclipse; ‘ w 54 middle panel, Eclipse after being completely smoked; lower panel, nonsmoked Eclipse that has been cut open to illustrate: a, carbon fuel rod; b, insulation mats of glass fibers; c, tobacco paper; d, column wrapping of white paper with an interior lam- 0 mate aluminum foil heat barrier; e, distal column of reconstituted tobac- . co; f proximal column of reconstituted tobacco; g, tipping paper with a cork-like appearance; h, laser- generated holes for side-stream ven- tilation and smoke dilution (we have I enlarged the size of the holes to fa- cilitate viewing their position and number); and i, filter of cellulose acetate fibers. Eclipse length, 83 o;a mm. 0 0 0

different angles so as to permit critical microscopic viewing of processed in the same manner but which contained either: (a) the tip at different perspectives. no pack residue (i.e., microscope slide and micropore mem- Analysis of the Pack Residue for Glass Fibers and Particles. brane control); or (b) residue from packs of control cigarettes. The debris at the bottom of the pack of Eclipse and conven- For nonquantitative morphological and chemical analysis, the tional cigarettes was examined for glass fibers, particles, and residue was collected directly (e.g. , no water, no membrane, no fragments. In this investigation, the top of the pack was opened, dry ing, and no mounting) onto a glass microscope slide or SEM and all of the Eclipse articles were removed. Then, the pack was stub (see below). inverted, and the residue was tapped from the pack directly into Fibers, defined by the National Toxicology Program a cartridge that is marketed to analytical laboratories for the (United States Department of Health and Human Services), quantitative analysis of airborne fibers and particles (Zefon, were recognized as structures having a length:diameter aspect Inc., St. Petersburg, FL). We selected a standard cartridge that ratio of 3: 1 or greater ( 18). Glass structures having an aspect had been preloaded with a 25-mm diameter micropore (0.45- ratio of <3: 1 were defined as particles; the term particle is used m) membrane. The plastic disposable cartridge was clamped herein as a generic term that includes other descriptive expres- in an upright position to a vertical pole of a laboratory stand that sions including fragments, flakes, bits, chips, specks, and dust. had been placed in a laminar flow hood. Afterward, the pro- In identifying glass fibers and particles by polarized micros- tective seal was removed, and the cartridge was opened. Then, copy, we used protocols and schema used routinely for fiber the aluminum foil liner of the cigarette package was extracted identification by forensic scientists ( 19). Unlike all other fibers and flushed thoroughly with a stream of fiber-free water that and particles, glass items do not polarize light; this is defined as was directed into the cartridge. The pack residue sample that isotropic (19). This unique optical property, transparency, size had been transferred to the assay cartridge was then deposited [i.e., diameter(s)], cross-sectional shape, surface morphology, gently onto the membrane by low-vacuum filtration. During color, terminus (i.e. , end) appearance, and other optical and this step, the cartridge walls were rinsed thoroughly with a morphological characteristics were used to distinguish Eclipse- stream of hospital grade water. Thereafter, the micropore mem- derived glass fibers and glass particles from other items found brane was removed and dried at 56#{176}Cina covered plastic Petri in the pack residue. dish. Filter Tipping Paper Tested for Glass Fibers. The surface of After placing the membrane onto a clean glass slide, the the cork-appearing tipping paper of Eclipse and conventional disc was cleared, and a coverglass was mounted using Cytoseal cigarettes was tested for glass fibers. For this purpose, the filter 60 (Stevens Scientific, Riverdale, NJ). The number of glass was rolled one complete rotation on a strip of clear adhesive fibers was determined by inspecting the membrane with a tape. The tape was then adhered to a clean microscope slide, microscope that had been configured for viewing with white, and the number of glass fibers collected was enumerated with polarizing, and phase-contrast lighting as well as epifluorescent lighting with different filter configurations. The number of a microscope configured for viewing with white and polarizing light. glass fibers in each of five microscopic fields (X 10 objective) of each membrane was counted and recorded. The mean num- Filter Surface Examined with a Stereo-Zoom Microscope. ber per field and the viewing field area were then used to Packs of Eclipse and conventional cigarettes were opened, as calculate the total number of glass fibers in the residue of a described above, and the faces of the filters were inspected for pack. glass fibers with a stereo-zoom microscope (Stereomaster; Controls consisted of micropore membranes that had been Fisher Scientific). Each filter face, and portions thereof, was

: ‘r

A A-A -AA.

Fig. 2. Glass fibers in Eclipse. a, two packs of Eclipse. as viewed after inverting and removing the bottom wrapping paper. b, close-up view of . Eclipse, from the previous panel. Illustrated are the two mats of glass fibers that sandwich the ./ I tobacco paper (black arrow). Also shown is the carbon fuel rod that is insulated by the glass mats. Open arrow, a bundle of glass fibers displaced from the insulation mats. c, detailed view of the tips of two Eclipse articles that illustrates a multitude of glass fibers that form the insulation mats and carbon particles (arrow) displaced from the fuel rod. d, glass fibers protruding from the tips of two Eclipse that had been removed carefully from a freshly opened pack. Arrow, one of several glass fibers that were seen on the surface of the cigarette. e, view with a light microscope of the residue of a pack of Eclipse. Shown are numerous glass fibers of variable length. Also illustrated are glass fiber-derived particles, three of which have been denoted with an arrow (X50). f view with a SEM of the broken end of a glass fiber that was found in the pack residue of Eclipse. Binder is present as randomly dispersed amorphous dabs as well as a film-like substance (arrow) dripping from the glass fiber surface (bar, 1.0 sm).

scrutinized without removing the articles from the pack. Ini- precise focusing; (b) both the glass and cellulose fibers are tially, the field of observation was adjusted to permit imaging translucent. Furthermore, light was reflected from the smooth of the whole filter (15-fold magnification; X 1 .5 objective and surface of both fiber types. When compared with vertical light- x I0 ocular). For a more detailed viewing of the filter, we used ing, oblique shadow-casting illumination provided better di- the variable magnification afforded by the stereo-zoom optics. mension and enhanced the contrast of the glass and cellulose The magnification was usually within an X8 to X40 range. The acetate fibers; and (c) glass fibers and particles were seen both resolution at these magnifications was about 73 and I 35 lines! atop and within the column of large cellulose acetate fibers. mm, respectively.5 Consequently, resolving the glass at different depths within the Glass fibers upon and within the Eclipse filter were seen filter was facilitated by adjusting repetitively and simulta- using a conventional stereo-zoom microscope and standard neously both the illumination and focus. illumination. We learned, however, that better viewing of the We established the following technique whereby the filter glass fibers was achieved with duel gooseneck fiber-optic of each Eclipse in a pack could be analyzed efficiently and lamps from a 4.8-W halogen light (Dolan Jenner Industries, expediently for glass fibers with a stereo-zoom microscope. In Inc., Lawrence, MA). The lamps were placed in an opposing this method, one hand was positioned on one of the gooseneck left- and right-hand position. Each lamp was positioned at an halogen lamps to direct oblique lighting onto the cut surface of oblique angle, usually 5#{176}to15#{176}tothe horizontal. the cigarette filter, while the other hand was used simulta- The advantages afforded by this illumination scheme were neously to adjust the focus of the microscope. as follows: (a) the Eclipse and conventional cigarette filter are Filters Assayed for Glass Contaminants with a Fiber Trans. composed of numerous cellulose acetate fibers, and these fibers fer Test. We have developed a test for harvesting loose fibers are oriented longitudinally to form the cylinder-shaped filter and particles on the surface of a cigarette filter for analysis (16). plug. Microscopically, the on-end view of the filter fibers The transfer assay, as used for detecting glass fiber contami- revealed a carpet-like appearance. Inspection of glass fibers on nants of the Eclipse filter, was conducted as follows. We this morphologically complex surface at high-power required affixed to a board a 3 X 2-inch piece of clear tape (Scotch SuperClear; 3M Company, St. Paul, MN). An Eclipse or control cigarette was orientated perpendicular to the tape and posi- 5 H. Kaneko (Carton Optical Industries, Tokyo. Japan), personal communication. tioned directly above the tape. Then, the end of the filter was

touched to the adhesive side of the tape. Immediately thereafter, a-c). In many instances, glass fibers was seen protruding from the same filter was touched a second time to an adjacent spot on the insulation rolls (Fig. 2, a and b). Glass strands projecting the tape. The tape was then pressed, sticky-side down, onto to outwardly from the mat were detectable with the naked eye and a clean, fiber-free, 3 X 2-inch glass microscope slide. The two were visible readily with a hand-held magnifying glass. Greater contact spots were then examined with a polarizing microscope, definition of the glass rods extending randomly from the cut and the glass fibers were counted and the number recorded. surface of the Eclipse tip was obtained with a stereo-zoom Filters Inspected for Glass with an Electron Microscope. microscope (Fig. 2d). The flat face of the test filter was examined for glass fibers and An assay was performed to define the relative frequency of particles with a SEM. In this analysis, a precleaned and new glass strands extending from the Eclipse tip. In this test, we razor blade was used to slice carefully a 3-4-mm transverse evaluated Eclipse (n = 60) that had been obtained from three section from the mouth end of the Eclipse filter. Particular care different Eclipse packs (A, Regular; B, Milds; and C, Menthol). was exercised not to touch the face of the filter with fingers or Each Eclipse was placed on the stage of a stereo-zoom micro- instruments so as to avoid transferring any extraneous glass scope and examined individually. The following scoring plan fibers. Using standard SEM methodologies, the test filter sec- was used to categorize the number of glass fibers that extended tion was glued in an upright position upon an aluminum SEM 1 mm or more from the tip: group I (<5 fibers/Eclipse); group stub. The filter was coated with gold (Edwards Model 360 II (5-15); and group III (> 15). Values recorded were as fol-

Evaporator; Grand Island, NY). Then the filter surface was lows: group I: pack A = 1/20 Eclipse, B = 3!20, and C viewed with an Autoscan SEM (Etec, Inc., Hayward, CA) or as 12/20; group II: A = 9/20, B = 1 l!20, and C = 5/20; and group described below. III: A 9/20, B = 3/20, and C = 1/20. Thus, 63% of the SEM/EDX Analysis of Glass Particles. Confirmation of par- Eclipse (n = 38/60) displayed five or more glass fibers that tides as glass, undoubtedly arising from Eclipse glass fibers, protruded 1 mm or more from the tip. was established using two different methods: (a) polarized Glass Fibers Observed on the Paper Wrappers of Eclipse microscopy (Ref. 19; see above); and (b) analytical scanning Articles. Glass strands were observed on the white paper electron microscopy energy dispersive X-ray microanalysis wrapper of the Eclipse articles, including the shaft and tip (Fig. (20). In the EDX analysis, we used an Hitachi 5-800 field- 2i’). No tests were performed to define the frequency with emission SEM (Hitachi Scientific Instruments, Mountain View, which glass fibers occurred on the white paper wrapping be- CA) with an accelerating voltage of 25 kV. X-ray spectra were cause we thought that it would be more informative to analyze collected with a PGT X-ray microanalyzer probe (Princeton the mouthpiece for glass fiber contamination. Gamma-Tech, Princeton, NJ). The X-ray spectra recorded for Glass Fibers Seen on Cork-like Tipping Paper. An exami- the suspect glass particles was compared with Eclipse-derived glass fibers. We have recently used this methodology for de- nation was undertaken to learn whether glass fibers were pres- fining the elemental (chemical) composition of inhaled path- ent on the cork-like mouthpiece. Of the 16 Eclipse articles that des in human lungs (21). we examined, all tipping papers had been contaminated with glass strands. Included in this analysis were both Eclipse (n Mechanical Smoking of Eclipse. The release of glass fibers 12) and NEW Eclipse (n 4). Each Eclipse was from a from the filter of Eclipse articles was tested with a smoking different pack and had been collected randomly. The number of machine. The smoking machine was constructed at the Roswell glass fibers on the tipping paper ranged from 5 to 44 (mean ± Park Cancer Institute (Electrical Bioengineering Instrumenta- SE, 23 ± 2.4). tion Center and Occupational and Environmental Safety). The Charcoal particles, undoubtedly derived from the fuel rod, machine was engineered to permit cigarettes to be smoked with were also observed frequently on the tipping paper of Eclipse preselected smoking conditions. The machine smoking param- eters could be varied to approximate the smoking behavior of and NEW Eclipse. Also discovered on the tipping paper of smokers (6). In these studies, the machine settings were as Eclipse, NEW Eclipse, and Camel cigarettes were cellulose follows: (a) puff volume, 35 cc; (b) puff duration, 2.0 s; (c) puff acetate particles and broken fibers. vacuum, 10 mm H20; (d) puff frequency, one every 60 s; and In contrast, no glass fibers or charcoal particles were (e) total puffs per Eclipse, -10. observed on the tipping papers of conventional filter cigarettes The glass fiber release assay was performed in the follow- (Camel, RJR; n = 4). Likewise, no glass fibers were observed ing manner. The filter surface of an Eclipse was examined with on the negative control slides (n = 4). a stereo-zoom microscope. X- and Y-coordinates were estab- Glass Fibers Found in Eclipse Pack Residue. Having ob- lished by positioning four small marks with a black felt tip pen served glass fibers protruding from the tip of Eclipse articles on the filter tipping paper. Then, for each filter surface, a map and present on both the white wrapping paper and cork-appear- was drawn of the location and orientation of glass fibers that ing tipping paper, we undertook a study to learn whether glass were observed with a stereo-zoom microscope. Thereafter, the fibers were present in pack residue of Eclipse. For all Eclipse Eclipse was positioned onto the smoking machine, ignited with packs examined, numerous glass fibers were present in the pack a butane cigarette lighter, and smoked mechanically. After residue (Fig. 2e). The glass fibers were differentiated effort- smoking, the filter surface was reexamined and compared with lessly from the few other fiber types that were found in the the original map. The displacement of one or more glass fibers residue and which included paper strands and cellulose acetate was recorded as a positive test. filter fibers. Glass fibers in the pack residue, however, were indistinguishable from the glass fibers of the mats at the tip of Results the Eclipse. Displaced and Protruding Glass Fibers Observed on Eclipse Glass material in the pack was morphologically heteroge- Article Tip. The two insulation mats of glass fibers on the tip neous. Fibers were always observed and were of variable of Eclipse are shown in Figs. 1 and 2. An on-end inspection of length. Remarkably long glass fibers (>200 m) as well as the tip revealed marked variation in the integrity of the glass short fibers were observed in all packs. Glass that was recog- mats; frequently, clusters of glass fibers were displaced (Fig. 2, nized as components of the shattered glass fibers were present

Glass Fibers Seen On Eclipse Filter

REGULAR liii MILDS MENTHOL IllhlIllIllIllllllhlNEW

20 2 Fig. 3. Glass fibers seen on an Eclipse filter. IE The filters of Eclipse (Regular, Milds, and Menthol) and NEW Eclipse were examined 16 with a stereo-zoom microscope for glass fibers (n 24 packs). The number of Eclipse and NEW Eclipse with glass fibers on the filter for 12 each pack is illustrated (n = 20 cigarettes/ pack: ordinate). Of the 480 Eclipse articles examined, 471 (98%) were contaminated with glass fibers. In contrast, glass fibers were never 8 observed on the filters of conventional cigarettes, including premium or discount cigarettes of RJR and other cigarette companies (not illustrated). 4

0 I 23456123456123456123456

Eclipse Cigarettes (Pack Number)

in large numbers. These included glass particles, fragments, different Eclipse styles were: Regular, 95.0% (n = 1 l4!l20);

specks, and dust (Fig. 2e). Milds, 98.3% (n 1 18!l20); and Menthol, 100% (n = 120! We counted the glass fibers (3: 1 aspect ratio) present in 120). the residue of each of seven randomly selected packs of Eclipse. Glass fibers were discovered also on the filters of Eclipse The average number of glass fibers present in an Eclipse pack that had been obtained from other countries (Hi-Q and Inside). was 7,548 (SE ± 3,443; the range was from 1,164 to 26,725 Similarly, the filters of Premier were also contaminated with glass fibers!pack; n 7 packs). In contrast to our findings of glass fibers. the Eclipse pack residue, glass fibers were not observed in the More recently, RJR introduced Eclipse with a new filter

residue of packs (n = 7) of conventional cigarettes. design. We included these NEW Eclipse into our studies to A coating, probably a glass fiber binding agent (10), was determine whether the defect of glass fiber contamination of the observed on the surface of some of the glass fibers. Glass fibers filter had been corrected. Almost all of the NEW Eclipse from both the Eclipse tip and pack residue contained the coat- articles examined had filters with glass fiber contaminants. In

ing. The glass fiber bonding agent was seen with conventional our assay of 120 NEW Eclipse, all but 1 (n = 1 l9!120; 99.2%) white light and polarizing microscopes, and its presence on were contaminated with glass fibers (Fig. 3). Eclipse glass fibers could be examined critically on the high- Conventional cigarettes were also analyzed for glass con- resolution afforded by a SEM (see below). taminants. Cigarettes assayed included: Winston and Doral Glass Fibers Discovered on Eclipse Filter. The discovery of (RJR); Marlboro and Basic (Philip Morris, Inc.); Kool and GPC binder-coated glass fibers in the pack residue suggested to us (Brown and Williamson); L&M and Lark (Liggett Group); and that the glass fibers, particles, and dust may have originated Newport and Kent (Lorrilard). Of the 10 packs examined, glass from the insulation mats at the tip of the Eclipse. This obser- fibers were never observed on the filters of these cigarettes (n = vation prompted us to inquire as to whether the glass fibers and 0!200 cigarettes). Thus, glass fiber contamination of the filter is particles from the insulation mats had moved onto the face of unique to Eclipse and could not be attributed to a common the Eclipse filter. To this end, examinations were performed to manufacturing flaw that would corrupt the filters of conven- ascertain whether glass fibers, particles, and fragments were tional cigarettes. present at the mouth end of the Eclipse cellulose acetate filter. The contamination of Eclipse, HiQ, Inside, and NEW Our inquiry revealed that glass fibers were observed easily and Eclipse with glass fibers was commonplace, extensive, and routinely on the Eclipse filter. Moreover, as will be described in global. Furthermore, test results of Premier illustrate that glass greater detail below, the glass fibers observed on the filter were fiber contamination of the filter had persisted for years. indistinguishable by SEM and SEMIEDX from glass fibers in Glass Fibers on Eclipse Filters Identified with a Fiber the pack residue. Transfer Test. Glass fiber contamination of the Eclipse filter Our findings were documented further in a comparative was corroborated further in the fiber transfer test (see “Mate- assay of Eclipse (Regular, Milds, and Menthol; six 20-count rials and Methods”). In this test, we analyzed the filters of eight packs of each of the three styles). Conventional premium and Eclipse from each of three randomly selected packs. Of the 24

discount cigarettes (n = 10 packs) were used as controls. Most Eclipse filters analyzed, all were contaminated with glass fi-

all (98%) of the Eclipse filters (n = 352/360) were comipted bers; two-thirds (n = 16!24) had 20 or more fibers (Fig. 4). The with glass fibers (Fig. 3). Values recorded for each of the number of glass fibers harvested ranged from 5 to 55 per

Glass Fibers On Eclipse Filter (Fiber Transfer Test)

Fig. 4. Glass fibers on an Eclipse 70 filter defined in the fiber transfer test. G) 18888881 Pack 1 Results of an experiment in which 60 the transfer test (see “Materials and ‘j- Methods”) was used to define glass o) Pack 2 fibers on filters. In this analysis. (I) eight randomly selected Eclipse cig- 0. 50 7A Pack 3 arettes were removed from three dif- C.) ferent Eclipse packs. The number of w glass fibers present on each of the 24 40 Eclipse filters is illustrated (glass fi- a) bers per filter; range. 5 to 55: 0 mean ± SE, 26.2 ± 2.6). In contrast. Cl) 30 I- using the same assay procedure, a) glass fibers were never observed on -o 20 the filters of 24 conventional ciga- Ii- rettes (not illustrated). U)

Ct’) 10 0 0 123456781234567812345678

Eclipse filter (mean ± SE = 26.2 ± 2.6; Fig. 4). In the same observe the tissue or item at low power and then, without assay, glass fibers were never (n = 0!24) observed on the filters changing the field of view, advance to a higher power. Accord- of conventional cigarettes (not illustrated). ingly, we would screen an Eclipse filter at low power for glass Glass Fibers on Eclipse Filter Revealed by SEM. High- fibers. Having identified a glass fiber of interest, the micro- resolution viewing of the glass contaminants of Eclipse and scope was adjusted to provide viewing with a higher magnifi- NEW Eclipse filters was achieved with a scanning electron cation. Operationally, a portion of the Eclipse filter was ob- microscope (Figs. 5 and 6). Detailed viewing of glass fibers and tamed randomly and viewed at a relatively low magnification of glass particles on Eclipse and NEW Eclipse filters was obtained -x 12,500. Having identified a glass fiber, a photograph was easily and frequently. Two morphological features that were collected for documentation. Thereafter, the same fibers were particularly useful in distinguishing microscopically the glass viewed at a high power. However, “charging” occurred often, fibers and cellulose acetate fibers of the Eclipse filter are and this release of electrons caused the glass fiber to jump from illustrated in Fig. 5, a and b: (a) the glass fiber had a rod-shape, the field of view. In some instances, the glass fiber was ob- whereas the cellulose acetate fiber had a distinct Y-shape; and served on another position on the filter; in other cases, it could (b) the size of the cellulose acetate fiber was invariably larger not be found. The displacement and release of the glass fiber than that of the glass fiber. was due to an excess charge injected by the electron beam that The position and orientation of the glass fibers on the filter could not flow to ground. The build up of charge in the surface varied markedly. Most of the glass fibers were observed cellulose acetate fibers and the glass fibers repelled the two lying upon the filter surface (Fig. 5, a and d). Some glass rods, fiber types, resulting in the ejection of the glass fibers from the most of which were relatively short, were observed positioned electron-enriched area. Recall, however, in all electron micros- behind a cellulose acetate fiber (Fig. 5, a and b). Some glass copy applications, the specimen (e.g., Eclipse filter) is viewed fibers were hidden deep within the cellulose acetate bundle in a sealed air-tight vacuum column positioned on a low- (Fig. 5a). Notable was the frequent appearance of glass fibers vibration air-table. Thus, the displacement of the glass fiber was of different lengths that were observed standing upright (Fig. caused by an electrical, and not mechanical, action. Sc) and lying down (Fig. Sd) on a wing of a Y-shaped cellulose Length and Diameter of Glass Fibers Defined by SEM. The acetate fiber on the surface of the filter. morphology and dimensions of the glass fibers and glass par- In each of the cases cited above as well as for all other tides were analyzed by SEM (Fig. 6). Long (>200 p.m) and critical examinations, we failed to detect any evidence that the short glass fibers were observed lying upon the surface of glass fibers on the filter surface were affixed, connected, cellulose acetate filter fibers (Fig. 5, c and d). Glass fiber bonded, or glued. Specifically, we were unable to identify diameters were relatively uniform (-5 pm; range, 4-6 jim). binding by the glass fiber bonding agent or the cellulose acetate plasticizer (e.g., triacetin). Thus, in many instances, glass fibers For some glass fibers, the ends displayed sharp transverse rested freely upon the filter surface and in a position whereby breaks, whereas other glass fibers displayed uneven, frag- it would be reasonable to predict that they would be released mented ends (Figs. 2fand 5, a-c). The surface of the glass fiber unabated and readily during smoking. was relatively smooth. Fiber binder on the glass fibers was Evidence in support of our prediction that the glass fibers prominent (Fig. 2f and 5a) and amorphous (Fig. 5, a and b). could be released readily was obtained in a serendipitous ob- Glass Fragments Identified by SEMIEDX. Glass rod frag- servation made in studies in which the Eclipse and NEW ments and debris were observed by light microscopy in the pack Eclipse filters were examined with a SEM. For all light and residue of Eclipse (Fig. 2e). SEM examination of the pack electron microscopy observations, it is a common practice to residue showed that the observed particles, thought to be from

Fig. 5. Glass fibers on Eclipse filter imaged with a SEM. Glass fibers on the Eclipse filter were seen frequently. readily, and in great detail with a SEM. All panels present views look- ing down onto the cut surface of an Eclipse filter face. These panels illustrate that the glass fibers were not bound to the plastic-like cellulose acetate filter fibers. Furthermore, the perspectives illustrate the ease with which the glass fibers would be released into the smoker’s mouth. a, the surface of an Eclipse filter exhib- iting three glass fibers (arrows). The glass fibers were distinguished promptly from the large Y-shaped cellulose acetate filter fiber (*). Binder or binder-like material (sharp arrow) is shown streaming from the surface of one of the glass fibers. Bar, 10.0 sm. b, glass fiber (length, -135-140 sm), balanced on a curved, Y-shaped cellulose acetate filter fiber. This view illustrates further the ease with which glass fibers and cellulose acetate fibers were distinguished morphologically. Bar, 10 sm. c, single short glass fiber (arrow), standing freely on its broken end, on the surface of a cellulose acetate filter fiber. Bar, 1.0 jzm. d, broken short glass fiber lying upon the surface of one of the three arms of a Y-shaped cellulose acetate filter fiber. Bar, 1.0 m.

glass fiber breakage, were morphologically heterogeneous and tests were performed in which Eclipse articles were smoked included glass chips, specks, and dust. To substantiate unequiv- mechanically to determine whether glass fibers were released ocally that these items were glass and not of other material, the during smoking. Positive results were recorded for 80% of the

chemical element composition of the particles was defined by Eclipse articles studied (ti = l2!15). SEM/EDX. Noteworthy was that the comparison of the maps drawn Fig. 6d shows a view of a single glass fiber and three for the glass fibers on the Eclipse filters before and after adjacent particles that are morphologically dissimilar. The his- smoking often revealed the appearance of previously tograms ofthe elemental analysis ofthe glass fiber (Fig. 7) were unidentified glass fibers. The newly discovered glass fibers compared with those of the suspect particles. The histogram of were thought to have been drawn up from an undetected the fiber revealed a major peak of silicon and smaller peak of position, most likely deep within the colonnade of cellulose calcium (Fig. 7). The SEM/EDX profile of the glass chip (Fig. acetate fibers, to the filter surface during mechanical 7, inset) was indistinguishable from that of the fiber. These and smoking. other SEM/EDX profiles of fragments, particles, chips, and items 1 .0 p.m were authenticated as being of glass. We concluded, therefore, that these glass particles and fragments Discussion originated from fracture and breakage of the Eclipse glass In studies reported herein, we have: (a) observed glass fibers fibers. and other glass contaminants lying freely on the cut surface of Glass Fibers Released from Eclipse Filter During Smoking. filter by light and electron microscopy; (b) failed to detect any From our extensive examinations of Eclipse filters with a evidence to suggest that these glass fibers were bound or glued stereo-zoom microscope and SEM, we concluded that the num- to the filter surface with binding agents, plasticizers, or other ber, size, position, orientation of the unbound glass fibers, materials; (c) discovered that the glass items were liberated fragments, and particles on the cellulose acetate filter surface from the filter by touching to tape. Inevitably the glass items would suggest that some of the glass items would be released would move easily from the cut surface of the filter to the into the smoker’s mouth during smoking. Notwithstanding, smokers lips and tongue; (d) shown that the glass fibers lying

Fig. 6. Glass fiber contamination of NEW Eclipse filter and SEM confirmation of glass particles. This figure illustrates glass fiber contamination of the customized filter of a NEW Eclipse and documents the presence of glass particles as originating from Eclipse glass fibers. a, a view with a stereo-zoom microscope of two glass fibers (arrows) that are observed ea.sily as protruding toward the opaque, black-appearing. hollow central cavity of the cellulose acetate filter of a NEW Eclipse. The displaced glass fibers of the NEW Eclipse can be distinguished a quickly. even in this low-power view, from the single. thick. curved cellulose acetate filter fiber that has been dislocated and is prominent between the two arrows: b, SEM view illustrating the heterogeneity in the size of the Eclipse- derived glass contaminants that range from don- gated fibers to dust-like specks (bar, 60 pm); c, high-power view of a single glass particle. Ar- row, a sharp fracture plane where this particle broke from a glass fiber. Binder or binder-like material is shown on all surfaces of the particle (bar, I .0 sm); d. SEM view of a glass fiber (right side) and adjacent morphologically heterogeneous particles: confirmation that the three particles were glass was achieved by SEMIEDX (see Fig. 7: bar, 2 .tm).

on the cut surface of the filter were moved by charging that the glass particles were chemically indistinguishable from the occurred during SEM observations; and (e) glass fibers on the glass fibers as defined by SEMIEDX. filter were displaced during mechanical smoking of Eclipse. The observed glass fibers and particles on the Eclipse Comiption of the filter with glass fibers was documented filter were undoubtedly from the glass fiber insulation mats. in tests of Premier, Eclipse, NEW Eclipse, HiQ, and Inside. We envision that the brittle and thin glass fibers of the mats Thus, glass pollution of the filter of Eclipse was widespread, were fractured and shattered during high-speed Eclipse man- collective, and continuous. Experiments similar to those per- ufacturing, particularly during the process in which the glass formed using Eclipse were also carried out on conventional mats and carbon fuel rod are bisected transversely. In sup- cigarettes, and in no instance was even one glass fiber discov- port of this perception is the observation that fragments of ered. plastic fibers, such as cellulose acetate fibers, forming the The glass fibers observed on the filter, in the pack residue, mouthpieces of filter cigarettes tend to become separated and on the white wrapper paper and tipping paper ofthe Eclipse from the filter mouthpieces during cutting and contaminate articles were optically and morphologically indistinguishable the area around the filter (Ref. 22; see also Refs. 16 and 38). from the glass fibers of the insulation mats at the tip. Moreover, Thus, the glass particles would be produced by cutting

4.0k

3’ 3.6k Glass fiber :: Glass chip 3.2k

2.4k 4; a $ s 7 #{149} #{149} I#{149} - 1.2k

80#{128}

404 HH

0 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00

Energy (keY)

Fig. 7. Glass fibers and glass particles defined by SEMIEDX. SEM/EDX profiles defining the elemental (chemical) composition of an Eclipse-derived glass fiber and glass particle (Fig. 6). The large histogram is the SEMIEDX profile of the glass fiber illustrated in Fig. 68. This SEM/EDX profile was indistinguishable from SEMIEDX profile (small histogram. inset) recorded for the glass chip shown in Fig. 6d.

Eclipse. In addition, glass fibers would be shattered during not a cigarette, and that Eclipse should be classified as a subsequent operations, including rapid packaging and ship- nicotine delivery device (5, 1 1). Others have perceived Eclipse ping. The glass debris attaches nonspecifically to all surfaces as a ploy to challenge clear-air regulations banning cigarette of the Eclipse articles. Most importantly, the glass fibers, smoking in designated areas (4, 7). Some analysts who have particles, and fragments translocate to the cut surface of the studied Eclipse believe that a nicotine-addicted smoker will mouthpiece, where they were found atop and entangled accept these drawbacks and use a less favorable, smokeless, loosely within the carpet-like filter surface. nicotine-dispensing smoking article in a restricted clear-air Eclipse is packaged standing on its tip. Thus, the glass environment until a full-flavor conventional cigarette can be fibers protruding from this end would be susceptible to break- smoked in an unrestricted area (4, 7, 1 1). age resulting from jostling that would occur during daily trans- Accordingly, the goals for Eclipse declared by RJR, the port by the smoker, particularly if the Eclipse package was only perception by others of RJR motives, and the anticipated use of partially filled. Furthermore, the likelihood of glass debris in Eclipse have varied markedly. Likewise, the health benefits to the pack residue contacting and adhering to the cut filter face be realized by the Eclipse and NEW Eclipse user have been the would be greatly increased. Studies reported herein were con- subject of debate (4, 7, 9, 1 1, 14). Regardless of the intent, RJR ducted using Eclipse from freshly opened packs. We would has spent a considerable amount of time and money on Eclipse anticipate, therefore, finding a larger number of glass fibers, during a time when their share of a highly competitive market particles, and fragments on the filters of Eclipse articles of has declined (14). in-use packs. A leaflet that accompanies each pack of Eclipse instructs Eclipse has been viewed favorably by some who perceive the smoker to “take an extra puff and a longer draw while it as a smokeless cigarette that yields relatively low amounts of holding a flame to the carbon tip.” The increased vigor of tar, nicotine, and ash (4, 5, 7, 13). It has been argued that puffing would increase further the probability that glass fibers, Eclipse provides an alternative smoking article for health-con- particles, and fragments would be discharged into the smoker’s scious smokers who are: (a) unwilling to abstain from smoking; mouth during smoking. In an apparent attempt to correct the

(b) nicotine-addicted and are unable to quit; and/or (c) con- hard draw, the mouthpiece of the NEW Eclipse was redesigned. cerned with the health risks of passive smoke and who wish to It consists of a filter-appearing rod (diameter, -7 mm) having reduce the exposure of smoke to children, family members, and a central longitudinal passageway. This tunnel (diameter, -4 others (1-3, 4, 6, 7, 9-14). mm) spans the truncated filter rod (length, -9 mm). Thus, the In contrast, some reviewers have asserted that Eclipse is short hollow mouthpiece of the NEW Eclipse, shown in Fig. 6a,

is thought to achieve little or no filtration of the smoke, includ- fibers that had been supplied previously had been shown to fuse ing both the tar and vapor-phase components. When viewed in the smoking article during testing (27). Schuller claimed that: on-end, glass strands were seen readily protruding from the “. . . all purchases of ‘C’ glass fiber by RJR were understood by white cellulose acetate bundle into the central, black-appearing both parties to be used only for developmental purposes and not tunnel (Fig. 6a). commercial production of a new cigarette product” (27). RJR has not revealed to the Eclipse smoker the use of glass Arguments were also presented in the lawsuit to provide fibers. Known Eclipse smokers included thousands of individ- indemnification of Schuller by RJR (27-29). These arguments uals who were paid participants of test panels that were assem- related to a draft of a “Specialty Supplier Agreement” for the bled in different cities. Today, Eclipse articles are available on sale of glass fibers (27). The draft contained a provision pro- the open market and are being smoked by people in different viding that Schuller would: “. . . defend, indemnify and hold countries (United States, Germany, Sweden, and Japan; Refs. 1 harmless [RJR]. .. against all claims, damages, losses and ex- adn 14). RJR did not disclose to the consumer the presence of penses ... attributable to ... (a) bodily injury, or death ... glass fibers in Eclipse in: (a) a package warning label; (b) an (27). instructional brochure that was included in each pack; (c) promotional video tapes; or (6) newspaper advertisements. The lawsuit further declares that: “Schuller and RJR have The health risks and hazards associated with glass fibers in never resolved either Schuller’s policy concerns or the scope of RJR’s Premier and/or Eclipse smoking articles have been ad- indemnification or other contractual protections that RJR would dressed specifically in two documents. The first document is a provide Schuller in the event that the two companies could U.S. patent, filed in 1992 and awarded in 1994, to inventors establish a continuing business relationship” (27). Schuller Serrano et al. (23) and assigned to Philip Morris, Inc. The Philip withdrew the case on May 5, 1995 after reaching a settlement Morris patent describes the construction of a smoking article with RJR (30, 31). The terms of the accord were not disclosed delivering a nicotine-containing flavored aerosol rather than other than Schuller would no longer supply RJR with glass conventional tobacco smoke.6 This invention relates to a: “... fibers (31). smoking article in which the sensations associated with the The association of fibrous glass and cancer has been smoking of tobacco are achieved without the burning of the addressed in a document authored by investigators from the tobacco.” The aerosol is generated by a convective and radia- Environmental Carcinogenesis Program of the National Insti- tive heat transfer from a burning pure carbon rod. A claim is tutes of Environmental Health Sciences, Duke University Med- made that the article generates substantially no sidestream ical Center, and the Occupational Safety and Health Adminis- smoke. The intent, as well as some design features, are similar tration of the United States Department of Labor (17). These to RJR’s prototypes (7), Premier and Eclipse (2, 3, 5, 6). examiners declared “that fibrous glass materials are carcino- In presenting the background of their invention (23), genie ...“ (17). This assessment included an analysis of data Philip Moms makes reference to a U.S. patent assigned seven from both laboratory experiments (e.g., animal studies) and years previously to RJR for an Eclipse-like smoking article epidemiological studies (human data; Ref. 17). Furthermore, (26). Philip Morris argues that the RJR device: “. . . suffers they concluded: “. . . that on a fiber-per-fiber basis, glass fibers from a number of drawbacks. First, the resilient glass fiber may be as potent or even more potent than asbestos” (17). insulating jacket is difficult to handle on modern mass production machinery without special equipment. Second, the glass Letters presenting opposing view points have been presented fibers may become dislodged during shipping and migrate (32-34). These letters were followed by a concluding response through the pack to rest on the mouth end of the article, giving and rebuttal (35). RJR has discussed their perception of the rise to the potential for the inhalation of glass fibers into the potential toxicity associated with the use of the glass mat smoker’s mouth” (23). In an additional warning, Philip Morris insulator in the Premier- and Eclipse-like “NEW CIGA- claims: “It is a further objective of this invention to avoid the RETTE” prototypes (10). No results or discussions, however, potential for inhalation of glass fibers by a smoker of such an were presented for RJR’s proposed analysis of “glass fibers” in article” (23). “mainstream smoke” (Table 3.4.3-1 in Ref. 10) or subsequent The second document was a lawsuit filed against RJR for RJR publications. using glass fibers in their new cigarettes (27). The litigation Health risks of fibrous glass (i.e. glasswool; respirable against RJR was filed in March 1995 by Schuller International, size) have been reviewed also by the NTP, of the United States Inc.8 In this action, Schuller sought to bar its glass fibers from Department of Health and Human Services. For many years being used in: “the manufacture of a new cigarette product for they have and continue to assign glasswool in a category for a sale to the public” (27). In the summer of 1993, RJR asked “substance or group of substances and medical treatments Schuller to supply between 20,000 and 50,000 pounds of “C” which may reasonably be anticipated to be a carcinogen” (Ref. glass fiber to RJR by December 1993 (27). “C” glass fiber had 36; see also Ref. 17). Other agents listed in this grouping a higher heat tolerance than ‘T’ glass fibers. The ‘T’ glass include: (a) ceramic fibers (respirable size); (b) PAHs [n 15; e.g., benzo(a)pyrene]; (c) dioxane; (d) polychlorinated biophe- nyls; and (e) 1,l-bis(p-cholorophenyl)-2,2,2-trichloroethane 6 Philip Morris, Inc. has announced the development of a battery-powered, (DDT; Ref. 36). The NTP notes further that “there is sufficient low-smoke, ultra-low tar, cigarette-appearing smoking system called Accord (24, 25). This smoking device, produced to address health concems for environmental evidence for the carcinogenicity of glasswool in experimental smoke and potential fire hazards, is presently being evaluated in test panels of animals” (17). smokers for consumer acceptability (24). Inhaled glass fibers (37) and glass dust (38) have been 7 Eclipse was announced to the public on Nov. 27, 1994 by the The New York Times (4). Eclipse was evaluated for two years in 20 different states and by test isolated from the lungs of affected patients with lung disease, panels having collectively more than I 2,000 smokers but without the knowledge and the glass material was identified unequivocally by SEMI of the Food and Drug Administration (4). EDX (37). 8 Schuller was known formerly, and is known currently. as Johns Manville. Johns Manville was the target in the 1980s of numerous claims related to health hazards In January 1998, the NTP upgraded crystalline silica (res- associated with inhalation of asbestos fibers (27, 28). pirable size; e.g., quartz; Ref. 39) to that of a “known human

carcinogen.”9 Glass fibers and microparticulates, as is known opinion was expressed by Fan and Revere (42) in 1958 in for crystalline silica, may induce inflammation, fibrosis, hyper- describing their observation of the discharge of particles from plasia, or other nonmalignant pulmonary diseases (17, 36-39). cigarette filters: “In view of the accumulated medical evidence We (16, 40, 41) and others (42, 43) have documented the concerning the greater irritation to the respiratory tract from a discharge of different types of foreign bodies from cigarette solid particle with a coating of hydrocarbon than from either the filters, including cellulose acetate filter fibers (16, 40, 41), particle or hydrocarbon alone, the need to eliminate solid par- asbestos fibers (43), and carbon from cigarettes with charcoal ticulate material from the smoke streams would seem to be filters (40, 42), cigarette paper fibers (42), crystalline materials indicated.” (“cellulose acetate additives”; Ref. 42), and unidentified parti- Eclipse-derived glass released into the smoker’s mouth des (42). In studies in which conventional cigarettes were may also become entrapped in the saliva and, subsequently, smoked mechanically, crystalline material and filter material swallowed. Thus, some of the glass fibers and glass micropar- additives were released (42); the size of these particles were ticulates are likely to be ingested. Accordingly, the mouth, within the submicron range (1/20 to 1/40 tim; Ref. 42). The respiratory tract, and digestive system are at risk to toxin-coated release of various fibers and particles from cigarette filters was bioresistant glass discharged from Eclipse. In this context, glass documented with a transmission electron microscope in studies funded by the Brown and Williamson Tobacco Company (42). in foods is recognized by the Food and Drug Administrations as These comprehensive scientific findings were presented in a extraneous matter for which there is zero tolerance. 109-page monograph that was published -40 years ago (42). Our studies have documented glass fiber contamination of Philip Morris, Inc. has known and tested routinely, beginning as the filter and substantiates with reasonable certainty that glass early as the l970s, cigarette filters for the “fallout” of cellulose fibers would be inhaled andlor ingested when Eclipse is smoked acetate fibers (44) and charcoal (45). as intended. These findings and known risks associated with Today’s smoker is unaware that cigarette filter elements of glass fibers delineate and designate Eclipse as a defective different origin and size are being released into their mouths product. during smoking. These consumers are uninformed that tar- and The filter of conventional cigarettes is perceived as both toxin-coated biopersistent filter components will be inhaled efficient and safe. However, the American Thoracic Society, and/or ingested during normal smoking behavior. Battery by the Medical Section of the American Lung Association, these filter elements is not part of the smoking experience. adopted in 1995 this official statement: “Epidemiological stud- Today, both conventional and nonconventional smoking ies have shown that brands of cigarettes that contain less tar and articles are being marketed as having a greater consumer ben- nicotine only marginally reduce the risk of lung cancer mor- efit, including reduced health risks and decreased environmen- tality. Similarly, little difference in mortality has been found for tal smoke. Paradoxically, the filter was designed to absorb and lifelong filter versus nonfilter smokers and for persistent smok- trap tar in cigarette smoke. Tobacco tar binds rapidly and ers who switch from nonfilter to filter cigarettes” (50). tenaciously to most all surfaces (e.g. glass, plastic, cotton, and Many U.S. patents are being awarded yearly to tobacco paper). The binding of tobacco tar, to glass fibers used in companies for technically complex multiple component ciga- prototype cigarette filters has been known for many years. rettes consisting of tobacco-like material with assorted chemi- Likewise, the use of glass fiber filters (e.g., Cambridge filters, Ref. 15) in smoking machines, for the quantitative assessment cal additives (1, 4). Some will be promoted as safer cigarettes of tar, as well as glass flasks, used for the condensation of with lower health risks (23). Other low-smoke, nicotine-deliv- tobacco smoke for extracting different tar components, is ery devices, also having the appearance of cigarettes, will be known widely (see also Ref. 46). Furthermore, it is widely introduced to avoid clean-air ordinances (4). accepted that the adsorption of tobacco tar onto glass fibers is Cancer prevention, particularly smoking cessation, will nonspecific, efficient, and rapid (47-49). In this context, we continue to be a primary goal of health organizations through- have observed that tobacco tar binds to the Eclipse glass fibers. out the world. There exists, however, no regulatory body that The enhanced health hazard associated with cigarette provides an independent analysis of cigarettes. Unlike the prod- smoking and fiber/dust inhalation has been discussed by many ucts of the food, drug, and cosmetic industries, smoking articles scientists. The synergistic effects of tobacco smoking and in- and chewing tobacco of the tobacco industry remain unregu- haled asbestos have been well documented. We hold the opin- lated. Thus, the question posed to the medical, scientific, and ion that an increased health risk may arise from the inhalation public health communities remains unanswered as to who will of tar-coated glass and/or cellulose acetate fibers and particles. define the health risks imposed on the nicotine-addicted, unin- Support for this expectation is derived from studies similar to formed smokers who participate in worldwide market tests of that reported herein. For example, the adsorption of PAHs, experimental tobacco articles. from tobacco smoke, onto asbestos fibers has been postulated We believe that smokers have the right to be educated for the known syncarcinogenic effect of cigarette smoke and fully of any and all health hazards and risks associated with asbestos (48, 49). This “PAH carrier hypothesis” proposes that particular brands of cigarettes that are offered for sale. RJR has asbestos fibers carry PAH into the cells and influence their promoted Eclipse as a “great-tasting cigarette with less second- metabolism. An alternative hypothesis predicts that the fiber hand smoke, no ashes, no lingering odor, and practically no acts as a complete carcinogen when it is chronically retained in staining on walls, windows, and curtains” (RJR’s promotional the bronchial wall (49). In this “fiber implantation hypothesis,” the irritants and toxins contained in tobacco smoke serve as video). It is what RJR has failed to tell consumers about Eclipse efficient cocarcinogens by damaging the bronchial mucosa and that concerns us. Contamination of Eclipse filters with glass impairing the bronchial clearance mechanisms (49). A similar fibers and glass dust may pose an additional health risk to the smoker. Eclipse is a paradigm of the health danger that may be imposed by technically complex tobacco articles and nicotine delivery devices promoted by an unregulated industry preying 9 NTP. 9th Report on Carcinogens (Attachment 3). January 1998. Also see IARC Classification Group I: “carcinogenic to humans.” upon addicted smokers who seek a less hazardous cigarette.

Acknowledgments Houck, W. G. Philip Morris, Inc. Smoking article. U. S. Patent 5.345.951, Sept. 13, 1994. We acknowledge the assistance of Stephen Bamhart, Cathleen Nasca, Lynn Hrusa, Charles Stuffier of Medical Photography, Elena Greco of Medical Illus- 24. Collins, G. Analysts mixed in reviews of Phillip Morris’ new smoking trations, and Gayle Ablove of the Medical and Scientific Library. We also thank system, p. 1. New York Times, Oct. 24, 1997. Dr. Richard Cheney for critical review of the manuscript. 25. Counts, M. E., Dccvi, S. C., Fleischhauer, G. S., Hajalogol. M. R., Hayes, P. H., Higgins, C. T., Houck, W. G., Jr., Keen, B. J., Jr., Laroy, B. C., Lipowicz. P. J., Miser, D. E., Nichols, C. H., Stevens, W. H., Subbiah, M., Watkins, M. L., Note Added in Proof and Wrenn, S. E. Electrical smoking system for delivering flavors and method for Pauly and co-workers have reported earlier this year the discovery of cellulosic making same. U. S. Patent 5,666,978, Sept. 16, 1997. and plastic fibers in surgically excised human lung tissue and malignant tumors 26. Banerjee, C. K., Fattier, E. G., Reynolds, J. H., IV, Ridings, H. T., (Pauly, J. L., Stegmeier, S. J.. Allaart, H. A., Cheney, R. T., Zhang, P. J.. Mayer, Sensabaugh, A. J., Jr., Shannon, M. D., and Shelar, G. R. R. J. Reynolds Tobacco A. G., and Streck, R. i. Inhaled cellulosic and plastic fibers found in human lung Co. Smoking article. U. S. Patent 4,714,082, Dec. 22, 1987. tissue. Cancer Epidemiol. Biomark. Prey. 7: 419-428, 1998). 27. Schuller International, Inc., a Delaware Corporation (plaintiff) versus R. J. Reynolds Tobacco Company, a New Jersey Corporation (defendant). United States References District Court for the District of Colorado, case no. 95-5-535. March 6, 1995.

1 . Anonymous. RJR tests new technology intemationally. Tobacco Intemational, 28. Gerlin, A. Manville unit sues an RJR unit to bar fiberglass from cigarettes. p. 62, April, 1997. p. B8. The Wall Street Joumal, March 8, 1995. 2. Glantz, S. A., Slade, J., Bero, L. A., Hanauer, P., and Barnes, D. E. The 29. Mallory, M. For Reynolds, where there’s smokeless. . . a Manville lawsuit Cigarette Papers, pp. 1-539. Berkeley, CA: University of Califomia Press, 1996. could stub out the Eclipse cigarette. p. 39. Business Week, March 27. 1995. 3. Glass, C. Testing Eclipse. Tobacco Reporter, June, pp. 22-24, 1996. 30. Order of Dismissal (see above, Ref. 30), by Judge Daniel B. Span. May 16. I995. 4. Hilts, P. J. Little smoke, little tar, but still a lot of nicotine. New York Times, pp. 1 and 38, Nov. 27, 1994. 3 1 . Anonymous. Manville unit settles suit naming R. J. Reynolds Tobacco. p. 1. The Wall Street Joumal, May 17, 1995. 5. Benowitz, N. L., Jacob, P., III, Slade, J., and Yu, L. Nicotine content of the Eclipse nicotine delivery device. Am. J. Public Health, 87: 1865-1866, 1997. 32. Hesterberg, T. W., and Chase, G. R. Letter to the editor. Commentary on “Fibrous glass and lung cancer.” Am. J. Indust. Med., 30: 1 1 1-1 12, 1996. 6. DeBethizy. J. D., Borgerding, M. F., Doolittle, 0. J., Robinson, J. H., McManus, K. T., Rahn, C. A., Davis, R. A., Burger, G. T.. Hayes, J. R., Reynolds, 33. McConnell, E. E. Letter to the editor. Commentary on “Fibrous glass and J. R., IV, and Hayes, A. W. Chemical and biological studies of a cigarette that lung cancer.” Am. J. Indust. Med., 30: 109-I 10, 1996. heats rather than bums tobacco. J. Clin. Pharmacol., 30: 755-763, 1990. 34. Weiss, W. Letter to the editor. Epidemiology of fibrous glass and lung cancer. 7. Feeder, B. J. A safer smoke orjust another smokescreen. New York Times, p. Am. J. Indust. Med., 30: 105-108, 1996. 1., April 12, 1996. 35. Infante, P. F., Schuman, L. D., and Huff, J. Fibrous glass insulation and 8. Kluger, R. Ashes to Ashes, pp. 1-807. New York: Alfred A. Knopf, 1996. cancer: response and rebuttal. Am. J. Indust. Med.. 30: 1 13-120. 1996. 9. Pauly, J. L., Streck, R. J., and Cummings, K. M. U.S. patents shed light on 36. U. S. Department of Health and Human Services. Sixth Annual Report on Eclipse and future cigarettes. Tobacco Control, 4: 261-265, 1995. Carcinogens, pp. 1-473. Research Triangle Park, NC: National Toxicology Program, National Institutes of Environmental Health Sciences, 1991. 10. R. J. Reynolds Tobacco Company. Chemical and Biological Studies of New Cigarette Prototypes that Heat Instead of Bum Tobacco, pp. 1-744. Winston- 37. Takahashi, T., Munakata, M., Takekawa, H., Homma. Y., and Kawakami, Y. Salem, NC: R. J. Reynolds Tobacco Company, 1988. Pulmonary fibrosis in a carpenter with long-lasting exposure to fiberglass. Am. J. Indust. Med., 30: 596-600, 1996. I 1. Warner, K. E., Slade, J., and Sweanor, D. T. The emerging market for long-term nicotine maintenance. J. Am. Med. Assoc., 278: 1087-1092, 1997. 38. Radenbach, D., Morgenroth, K., Staud, R. D., and Magnussen. H. Pulmonary risks caused by exposure to glass fiber dust. Pneumologie 44 (Suppl. I ): 250-251, 12. Smith, C. J., McKarns, S. D., Davis, R. A., Livingston, S. D., Bombick, B. R., 1990. Avalos, J. T., Morgan, W. T., and Doolittle, D. J. Human urine mutagenicity study comparing cigarettes which bum or primarily heat tobacco. Mutat. Res., 36!: 1-9, 39. Intemational Agency for Research on Cancer (WHO). In: IARC Monographs 1996. on the Evaluation of Carcinogenic Risks to Humans. Vol. 68. Silica, some silicates, coal dust and para-armid fibrils, pp. 1-506. Lyon, France: IARC. 1997. 13. Steele, R. H., Payne, V. M., FuIp, C. W., Rees, D. C., Lee, C. K., and Doolittle, D. J. A comparison of the mutagenicity of mainstream cigarette smoke 40. Pauly, J. L., Stegmeier, S. J.. Mayer, A. G., Lesses. J. D.. and Streck, R. J. condensates from a representative sample of the U. S. cigarette market with a Release of carbon granules from cigarettes with charcoal filters. Tobacco Control, Kentucky reference cigarette (1R4F). Mutat. Res., 342: 179-190, 1995. 6: 33-40, 1997. 14. Anonymous. RJR takes smokeless test. p. 1. USA Today, Oct. 27, 1997. 41. Borowicz, J. S., Streck, R. J., and Pauly, J. L. Contamination of cigarette filters with cellulose acetate fibers and particles documented in studies of adult 15. Browne, C. L. The Design of Cigarettes, Ed. 3, pp. 1-1 19. Charlotte, NC: smokers. Am. Assoc. Cancer Res., 39: 336. 1998. Filter Products Division, Hoechst Celanese Corp., 1990. 42. Farr, W. K., and Revere, A. Examination of Whole Cigarette Smoke by Light 16. Pauly, J. L., Allaart, H. A., Rodriguez, M. I., and Streck, R. J. Fibers released and Electron Microscopy, pp. 1-109. New York: The Life Extension Foundation, from cigarette filters: an additional health risk to the smoker? Cancer Res., 55: 1958. 253-258, 1995. 43. Longo, W. E., Rigler. M. W.. and Slade, J. Crocidolite asbestos fibers in 17. Infante. P. F., Schuman, L. D., Dement, J., and Huff, J. Fibrous glass and smoke from original Kent cigarettes. Cancer Res., 55: 2232-2235, 1995. cancer. Am. J. Indust. Med., 26: 559-584, 1994. I 8. U. S. Department of Health and Human Services. Seventh Annual Report on 44. Johnson, V. C. Fallout of cigarette filter material. Inter-office correspond- Carcinogens Summary-1994, prepared for the National Institutes of Environmen- ence. Philip Morris, Inc., Jan. 7, 1970. tal Health Sciences, pp. 1-473. Washington, DC: U. S. Govemment Printing 45. Johnson, V. C. Particle fall-out. Inter-office correspondence. Philip Morris, Office. 1994. Inc., Apr. 14, 1970. 19. Petraco, N. A guide to the rapid screening, identification, and comparison of 46. Bednarcyk, N. E. Tobacco smoke filters, pp. 1-263. Park Ridge, NJ: Noyes synthetic fibers in dust samples. J. Forensic Sci., 32: 768-777, 1987. Data Corporation, 1972. 20. Goldstein, J. I., Newberry, D. E., Echlin, P., Joy, D. C., Romig, A. D., Jr., 47. Pinter, K. G., and Calhoun, E. J. Glass microfibers. U. S. Patent 3,412,737, Lyman, C. E., Fion, C., and Lifshin, E. Scanning Electron Microscopy and X-ray Nov. 26, 1968. Microanalysis: A Textbook for Biologists. Materials Scientists, and Geologists. 48. Pots, F. A hypothesis for explaining the syncarcinogenic effect of cigarette Ed. 2, pp. 1-820. New York: Plenum Publishing Corp., 1992. smoke and asbestos. In: A. P. Wemer and D-L. Felton (eds.), Biological Inter- 21. Pauly, J. L., Bush, P. J., Allaart, H. A., Swift, M. R., Green, C. E., and Streck, action of Inhaled Mineral Fibers and Cigarette Smoke. Proceedings of an Inter- R. J. Size, morphology, elemental composition and airway distribution of diverse national Symposium/Workshop held at Battelle Seattle Conference Center, April types of inhaled particles in human lung specimens defined by scanning electron 10-14, 1988, Seattle, Washington, pp. 51-62, Columbus, OH: Battelle Press. microscopy and energy-dispersive X-ray microanalysis (SEMIEDX). Proc. Am. 1989. Assoc. Cancer Res., 39: 336, 1998. 49. Mossman, B. T., Cameron, G. S.. and Yotti, L. P. Cocarcinogenic and tumor 22. Chehab, F., Fietkau, S., Amold, P-F., Juschus, T., and Schmick, C. Method promoting properties of asbestos and other minerals in tracheobronchial epithe- of and apparatus for decontaminating the exposed surfaces of filter mouthpieces hum. Carcinog. Compr. Surv., 8: 217-238, 1985. in smokers’ products. U. S. Patent 5,645,087, July 8, 1997. 50. Hanrahan, J. P., Sherman, C. B., Bresnitz, E. A., Emmons, K. M., and 23. Serrano M. A., Houghton, K. S., Lanzillotti, H. V., Sanders, E. B., Lilly, Mannino, D. M. Cigarette smoking and health. Am. J. Respir. Cnt. Care Med., A. C.. Jr., Hayward. C. R., Hearn, J. R., Losee, D. B., Jr., Fleischhauer, G. S., and 153: 861-865, 1996.

J L Pauly, H J Lee, E L Hurley, et al.

Cancer Epidemiol Biomarkers Prev 1998;7:967-979.

Updated version Access the most recent version of this article at: http://cebp.aacrjournals.org/content/7/11/967

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/7/11/967. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.