Tobacco Documents Online

I I I I I I I I 1 I I I I I I I I I I Eur J Respir Dis (1986) 69 (Suppl. 146), 129-137 SMOKING HISTORY, CIGARETTE YIELD AND SMOKING BEHAVIOR AS DETERIMINANTS OF SMOKE EXPOSURE. R. B. BRIDGES1, J. W. HUMBLE2, J. A. TURBEK2 AND S. R. REHM3 1: Oral Biology, Dentistry, University of Kentucky, Lexington, KY. 2: Tobacco and Health Res. Inst., Univ. of Kentucky, Lexington, KY. 3: Medicine, VA Medical Center, Lexington, KY. ABSTRACT This study examines how smoking history, cigarette yield and smoking behavior relate to smoke exposure as determined by smoke constituents and their metabolic products in peripheral blood. Recruited without regard to the nicotine yield of their cigarette, male smokers smoked their own cigarettes j" libitum, including one cigarette five minutes prior to venipuncture. Smokers had signifi- cant (p=0.0001) elevations of serum thiocyanate, blood carboxyhemo- globin, plasma nicotine, and cotinine concentrations each of which was significantly associated with past 24-hour cigarette consump- tion. The nicotine yield of the cigarette significantly correlated with plasma cotinine concentrations and with the smoking behavior variables. Most notably, smokers consuming lower nicotine yielding cigarettes exhibited an increased total puff volume per cigarette, suggesting that smokers of low nicotine yielding cigarettes compen- sate for these low yields by their smoking behavior. However, the fact that lower plasma cotinine concentrations are present in smokers of low-nicotine cigarettes suggests that this compensation is incomplete. That smoking behavior variables relate to smoke exposure was demonstrated by a significant linear correlation be- tween plasma nicotine and mean puff interval in the total smoking population and between plasma nicotine and total puff volume per cigarette in a subpopulation smoking a single brand of cigarette. These data suggest that smoking history, nicotine yield of the cigarette and smoking behavior are all determinants of smoke expo- sure. Further, although smokers of low-yield cigarettes appear tc compensate by puffing larger volumes per cigarette, this compensa- tion appears to be inadequate to attain an equivalent smoke expo- sure. Thus, if tobacco related obstructive pulmonary diseases are associated with increased smoke exposure, a lesser degree of lung injury may be expected in smokers using low-yield cigarettes. INTRODUCTION Aside from possible genetic or other individual differences in susceptibility, the development of obstructive lung disease is likely related to a dose-dependent exposure to cigarette smoke (13). The exposure to the components of cigarette smoke is com- plex, being dependent upon a number of variables, some of which are interrelated. These variables include: the yield of the cigarette smoked; the present and cumulative smoking history; the smoking behavior (e.g., puff volume, duration and interpuff interval); and the absorption of smoke constituents (7). The absorption of smoke constituents is in itself dependent upon smoke chemistry, inhala- tion behavior, lung morphology and other physiologic parameters. The purpose of this cross-sectional study was to ascertain the major determinants of cigarette smoke exposure in a group of rela- tively young smokers with the ultimate intent of attempting to relate this smoke exposure to the development of obstructive lung disease. Cigarette smoke exposure was ascertained by measuring blood concentrations of smoke constituents/metabolites including thiocyanate, carboxyhemoglobin, nicotine and cotinine. We then 129

I I I I I I I I I I I I I I I I I I I ti sought to determine how parameters of cigarette consumption, yield of the cigarette smoked, and smoking behavior related to smoke exposure. • T_HODS METHODS For this study, 170 male smokers were recruited without re to the yield of the cigarette smoked or smoking history and pared with 170 age-matched male non-smokers . Subjects for study completed an extensive questionnaire concerning medical pulmonary histories, tobacco and drug usage, alcohol and cc consumption, and demographics. Nine smokers and two non-smc indicating uses of tobacco products other than cigarettes excluded from the data analysis. Smokers provided data on ci ette brand and number of cigarettes smoked per day for each f year interval to •the present time. The nicotine yield of cigarette smoked was determined from lists available from •Federal' Trade Commission (8). Smokers were requested to smokE libitum and to smoke one of their own cigarettes at 8 am, minutes prior to venipuncture. This latter cigarette was sa using a cigarette holder-pneumotachograph to measure number puffs, intervals between puffs, duration and volume of each and total volume puffed per cigarette (1). This pneumotachoE was calibrated by drawing puffs of standard volume inserting subject's unlit cigarette in the cigarette holder. The t concentrations of smoke constituents/metabolites were determin< previously described (6). Statistical analysis of the data utilized Student's t test unpaired data and Pearson's correlates for regresaion ana: using appropriate SAS programs (10) and an IBM 3083 computer. data are expressed as means tS.E.M. RESULTS The characteristics of the smoking and nonsmoking popula, are given in Table i. The smoking and nonsmoking population: not differ significantly in age, weight or tea consumption. ever, the smokers consumed significantly more coffee and alcol beverages than did their nonsmoking counterparts. Table 1. Characteristics of Smoking and Nonsmoking Population: Nonsmokers Smokers P Age (years) 37.2:0.8 37.8t0.8 0.59 Weight (1Cg) 79.6t1.0 77.6t0.9 0.13 Tea Consuaptiont 1.02=0.12 0.94t0.15 0.72 Coffee Consumptiont 1.70=0.18 3.8g=0.30 0.0001 Alcohol Consu.ption$ 2.55t0.29 8.73t1.06 0.0001 •Results for 161 smokers and 168 nonsmokers. tExpressed as cups per day. $Expressed as ounces of pure alcohol consumed per week. There was considerable variability among smokers in term the smoking history parameters (Table 2). Consistent with relatively young age, the smokers had a relatively brief sm history (20.3 years smoking, 24.6 pack-years) and had smoked a of 25 cigarettes in the past 24 hours. The smokers consum variety of cigarette products including both non-filter and f cigarettes, with a relatively high mean nicotine yield (1.0 mg arette) for the overall population. As expected, smokers had significantly (pS0.0001) higher thiocyanate, blood carboxyhemoglobin, and plasma nicotine and nine concentrations than did the nonsmokers (Table 3). F

I I I I I I I I I I I Table 2. Smoking Historyf MEAN t S.E.M. RANGE Nicotine yield 1.00 t 0.02 0.28-1.56 Cigarettes past 24 hours 24.8 t 0.9 5-70 Years smoking 20.3 t 0.9 3.7-53.7 Pack-years 24.6 t 1.0 3.8-60.0 •Data for 161 smokers. nicotine and cotinine concentrations, being specific markers for tobacco smoke exposure, were the better indicators to differentiate between smokers and nonsmokers (data not shown). Significant linear correlations existed between the log of the cigarettes consumed in the past 24 hours and the blood c~~centrations of each of the smoke constituents/metabolites. In each case the log of the cigarettes in the past 24 hours was a better predictor of Hood concentrations of smoke constituents/metabolites than were the linear plots. However, the log of the cigarettes consumed in the past 242hours was the best predlgtor of plasma cotinine concentra- tion (R =0.132, pS0.0001) (Fig. 1), followed2in order by carboxy- hemoglobin (R 20.101, ps0.0001), nicotine (R =0.058, pS0.002) and thiocyanate (R =0.033, pS0.02). Table 3. Blood Concentrations of Smoke Constituents/Metabolitese. Nonsmokers Smokers P Serum Thiocyanate ( u M) 98.3t2.6 163.43.7 0.0001 Carboxyhemoglobin () 2.2.t0.1 7.at0.2 0.0001 Plasma Nicotine (ng/ml) 2.30.3 31.1t1.3 0.0001 Plasma Cotinine (ng/ml) 2.9t0.4 384.0t12.5 0.0001 •Results for 161 smokers and 168 nonsmokers. r.0.373 080.0001 .n E w c v W 2_ ~ U . - ~ 2 300 J ' 00 • _ ~ . 40 . ~ ~ < 20o • f- • - • 2 ' J d . . . s . Ioo o . - C - . . ZZJ 5 10 iS 20 30 40 50 60 70 6" CIGARETTES PAST 24 HOURS al~ Fig. 1. Relationship of p lasm a cotinine conc en tration with ~ cigarette consumpti on i n the past 24 hou rs . ~ ~ ~ The smoking population was arb itrarily divided into six groups I-~ according to relatively narro w ra nges of nicotine yield of their ~ cigarettes (Table 4). Smokers of high-nicotine cigarettes in groups 5 and 6 had significantl y di fferent mean age, years smoked I - - 131

I I I I I I I I I I I I I I y and pack-years smoking history than did those in the first groups. Therefore, in order to avoid the possible complicatin effects of age and smoking history characteristics, further statis tical analyses were limited to groups 1 through 4. Table 4. Characteristics of Smokers According to Nicotine Yield 0 Their Cigarette. Range of Nicotine Group Yield• N Age (years) Cigarettes Past 24 Hours Years Smoked Pack-Year 1 0.28-0.43 5 36.1t2.5 19.4t5.4 16.8t3.0 23.8t7.2 2 0.50-0.70 16 35.2f2.2 23.1t2.6 15.9±1.7 20.2±2.6 3 0.71-0.90 22 40.0t2.8 24.7t6.6 20.5f2.9 23.5t6.7 4 1.05-1.10 65 36.1t1.2 25.6±1.6 19.1±1.2 24.1t1.5 5 1.11-1.20 17 30.6.t1.4t 24.2t2.6 13.5t1.5t 16.4±2.2 6 1.40-1.60 14 48.4t2.it 24.1±3.0 33.9t2.6 t 37.9t3.2 fTwenty-two subjects smoked generic cigarettes for which th nicotine yield was unavailable. tSignificantly (pS0.05) different from groups 1 through 4. The blood concentrations of smoke constituents/metabolites ii the first 4 groups are given in Table 5. Although smokers of low• yield cigarettes had lower plasma nicotine and cotinine concentra• tions than did smokers of higher yield cigarettes, these decreasec concentrations were not proportionate to the relative decrease ir yield of the cigarette smoked. For example, comparing smokers ir groups 4 and 1, there was a 68% decrease in the mean nicotine yielc of the cigarettes with only a 44.6% and 37.4% decrease in mear plasma nicotine and cotinine concentrations, respectively. Although serum thiocyanate and blood carboxyhemoglobin were alsc decreased (19.2% and 25%, respectively), these decreases were smaller in magnitude, suggesting that exposure to the gas phase components of cigarette smoke (i.e., cyanide and carbon monoxide) remained relatively high despite the reduction in the principal pharmacologic agent of the particulate phase, nicetine. Table 5. Blood Concentrations of Smoke Constituer.ts/Metabolites in Groups of Smokers According to Nicotine Yield of their Cigarettef. Mean Nicotine Carboxy- Plasma Plasma Tield Thiocyanate hemoglobin Nicotine Cotinine Grcup (mg/cigarette) ( pmoles/1) (S) (ng/ml) (ng/ml) 1 0.34 132.2t17.5 5.7f1.0 18.0t1.4 256t92 2 0.56 160.7t 9.1 7.4.0.6 29.7t5.3 330±37 3 0.76 175.9±11.3 7.6±0.5 28.4s3.1 351±28 4 1.06 163.6t 6.0 7.6t0.3 32.5±1.7 409t19 g Change* -68% -19.2% -25% -44.6% -37.4% •The percentage change in the means comparing groups 4 and 1 That the yield of the cigarette affects exposure to smoke was verified by the linear relationships between yield and plasma nicotine or cotinine concentrations. The nicotine yield of the cigarette correlated marginally with plasma nicotine (r=0.168, p:0.08) and significantly with plasma cotinine (Fig. 2). However, the yield of the cigarette did not correlate significantly with either serum thiocyanate (r=0.082, pa0.4) or blood carboxyhemo- Llobin (r=0.105. p=0,29) concentrations. 132 I \ 00

I I I I I I I 1 I I I I I I I 1 I I ~ E aa ~ W 2 2 O U Q 900 eoo 700 600 500 400 2 to 300 Q a 200 loo y= 1 74.0:+222.9 r=0.257 os0.00s . ' . . .: _ .: ~ . . . .~ . . . . . . . . . . ~. ~ . : . . . . . . . ... . . . . . . . ~ . : . . . .. . . . . . • . 041. ~ . x ~0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1 .I 1.2 NlCOTNE YlELD (eg/ciyar.tt.) Fig. 2. Relationship of plasma cotinine with the the cigarette smoked. nicotine yield of The smoking behavior variables for the first 4 groups of smokers according to the nicotine yield of their cigarettes are given in Table 6. Although not significant, smokers consuming the lowest nicotine yielding cigarettes (group 1) tended to have a higher number of puffs per cigarette. However, these smokers had a significantly higher mean puff volume (pS0.001) as compared to the other 3 groups. These changes in smoking behavior probably account for the highly significant (pS0.0001) increase in total puff volume observed in the saiokers of low-yield products in group 1. Further, a highly signif:.cant linear decrease in total puff volume per cigarette was observed with increasing nicotine yield in these four groups of smokers (Fig. 3). Table 6. Smoking Behavior in Groups of Smokers According to the Nicotine Yield of their Cigarettef. Group 1 Number of Puffs Per Cigarette 13.2t2.8 Mean Puff Volume (milliliters) 85.4t10.60 Total Volume Per Cigarette (milliliters) 1141.6.t281.4t 2 11.4t1.4 63.7t 8.0 686.4t 89.7 3 11.1.t0.9 54.1t 2.6 577.9t 40.8 4 10.6t0.5 52.2t 1.8 528.% 20.3 fSignificantly (pS0.001) greater than the other groups (2-4). tSignificantly (p5 0.0001) greater than the other groups (2-4). The importance of studying smoking behavior variables is dependent upon the ability to demonstrate relationships between these variables and the blood concentrations of smoke constituents /metabolites. Examination of the linear correlates between these smoking behavior variables and blood concentrations of smoke con- stituents/metabolites of smokers in groups 1 thru 4 (Table 7) revealed relatively few significant, linear correlations. Signifi- cant negative correlations were observed between serum thiocyanate concentrations and the mean puff volume and the total volume per cigarette. The rationale for these significant correlations is not 133

1 ~ 1 ~ ~ ~ ~ ~ 1 ~ ~ 1 ~ ~ 1 V = r--48S.0. + 1023 ~ ~ 2000 P.-0.407 1•0.0001 S 1800 1600 W 1400 1 V 1200 1000 ~ . ~ 800 . . W . :.S W 600 . . . . ~~ . .S 0 . . . ...~- . 400 . : .~ ~ ~. 4 .. . -ILI : . - « p 200 F ~ 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 x 1.2 NICOTTE YELD */ciparntt.) Fig. 3. Relationship of total puff volume per cigarette with tt nicotine yield of the cigarette smoked. Table 7. Relationships of Smoking Behavior with Blood Concentrations of Smoke Constituents/Metabolitesf. Serum Carboxy- Plasma Plasma Thiocyanate # Puffs/cigarette -0.088 hemoglobin 0.046 Nicotine 0.125 Cotinine 0.016 Mean puff interval 0.048 -0.093 -0.295t -0.114 Mean puff volume -0.195t -0.070 -0.039 -0.139 Total volume/cigarette -0.225t -0.054 0.058 -0.117 *Pearson's correlates for 108 smokers (Groups 1 tarough 4) tp50.05 tpS0.01 40 35 30 25 20 IS 10 5 0 240 360 480 6o0' 720 840 940 1080 TOTAL PtFF VOLtJ1/E /CiGARfTTE (.it.n) Fig. 4. Relationship between plasma nicotine concentration an total volume puffed per cigarette in a populatio smoking a single brand of cigarette (nicotine yield 1.0 me/ciQarette). 134 ~

I I I I I I I 1 I I I I I I I I I apparent. A significant, negative correlation was also observed between plasma nicotine and the mean puff interval. Although a significant correlate was expected between total puff volume and plasma nicotine, none was detected in the total smoking population. Since nicotine yield of the cigarette had already been shown to affect smoking behavior variables, it seemed likely that the variation in nicotine yield might confound any significant linear relationships between smoking behavior variables and plasma nico- tine. Therefore, the relationships between smoking behavior vari- ables in smokers with an equivalent nicotine yielding cigarette were sought in the largest, available sub-population of smokers consuming the same brand of cigarette. Thirty smokers in this study consumed a single brand of cigarette with a nicotine yield of 1.05 mg/cigarette. In addition to the significant negative corre- lation between mean puff interval and plasma nicotine concentration (r=0.438, p=0.02), the total volume per cigarette was also signifi- eantly correlated with plasma nicotine (Fig. 4). On the other hand, number of puffs was only marginally related to plasma nico- tine (r=0.306, p=0.10). DISCUSSION 0o concentrations of thiocyanate, carboxyhemoglobin, nico- tine and cotinine are at best only estimates of smoke exposure. The concentrations of these constituents, while related to smoke exposure, are also dependent upon pharmacokinetic parameters such as volume of distribution, clearance time and half-life (7). In this study, we sought to determine how cigarette consump- tion, nicotine yield of the cigarette and smoking behavior relate to cigarette smoke exposure as measured by blood concentrations of these smoke constituents/metabolites. This cross-sectional study differs from other previous studies in that these parameters were determined in a relatively large population of healthy, male smo- kers who smoked ad libitum their own brand of cigarette. In addition, the sub7ect~or this study were not selected according to indices of smoking history or type of cigarette smoked, accoun- ting for the observed large variation in these parameters. By selecting the subject population in this manner, relationships between parameters were not optimized due to the confounding effects of the other variables. However, the study population is probably more representative of the general smoking population and no changes in the smokers' habits have been introduced by the experimental design. The best indicator of cigarette consumption in the past 24 hours was p1asms cotinine probably as a result of both it: rela- tively long half-life (15.8 hours) (4) and its specificity as an indicator of cigarette smoke exposure. However, 24-hour cigarette consumption still accounted for only 13.9% of the variation in plasma cotinine concentration. Although blood concentrations of carboxyhemoglobin are affected by exposure to environmental carbon monoxide, physical activity and a relatively short half-life of 2-4 hours (15), it was still a good indicator of cigarette consumption in the past 24 hours. On the other hand, nicotine [as a result of its short half-life of 1.9 hours (11)] and thiocyanate [being affected by diet (5) and having a half-life of 14 days (9)] were relatively poor indicators of cigarette consumption in the past 24 hours. Controversy exists in the literature as to the relation be- tween blood levels of smoke constituents and the nicotine yield of the cigarette (2,11). In the present study, we have demonstrated that the nicotine yield of the cigarette affects the blood concen- trations of smoke constituents/metabolites, especially cotinine. 135

I I I I I I I I I I I I I I I I y Dividing the smoking population in groups according to the nicotir, yield of the cigarette smoked, a trend toward decreased concentra tions of all smoke constituents/metabolites was observed with th lower nicotine yield of the cigarettes. Decreased plasma nicotin and cotinine concentrations were linearly associated with decrease nicotine yield of the cigarette smoked, without taking into accoun individual differences in daily cigarette consumption. However these decreases in plasma nicotine or ootinine were not equivalen to the decrease in nicotine yield of the cigarettes, suggesting a incomplete compensation by smokers of low-yield cigarettes. Previous studies have demonstrated compensation by smokers o low-yield cigarettes by means of either increased cigarett~ consumption (12) or changes in smoking behavior (8). Compensatioi in the present study was not due to increases in cigarette consump• tion since all the groups had an equivalent mean daily cigarett( consumption. Therefore, we sought to determine if compensation b: smokers of low-yield cigarettes was due to differences in smokinE behavior. •A highly significant increase in total puff volume per cigarette was observed in smokers of low-yield cigarettes. Thi: increased total puff volume per cigarette appeared to be due t( increases in both the number of puffs per cigarette and the mear puff volume. Further, a significant linear increase in tota: volume per cigarette was observed with decreasing nicotine yielc for this total population of smokers. Collectively, these datz suggest that smokers of low-yield cigarettes compensate primaril} by puffing a larger, total volume per cigarette. Relationships between the smoking behavior variables and thE blood concentrations of smoke constituents/metabolites were alsc sought in order to demonstrate that these variables are related tc smoke exposure. Due to its relatively short half-life and speci- ficity, plasma nicotine seemed more likely to correlate with these smoking behavior variables. In the total smoking population, plasma nicotine concentration was observed to increase with decreasing mean puff intervals. This observation is consistent with the fact that shorter intervals between puffs raises the pyrolysis temperature of the burning cigarette rod and thus, in- creases the nicotine delivery. No significant linear relationship was observed between total puff volume per cigarette and plasma nicotine in the total smoking population. However, removing the confounding effects of nicotine yield by studying a sub-group of smokers consuming the same brand of cigarette, linear increases in plasma nicotine concentration were significantly associated with both decreased intervals between puffs and increased total puff volume per cigarette. These data demonstrate that smoking behavior variables, especially puff interval and total puff volume per cigarette, are determinants of smoke exposure. Thus, in a cross-sectional study of a well-delineated group of relatively young smokers, we have observed that cigarette smoke exposure (espeoially as measured by plasma nicotine and cotinine concentrations) is dependent upon: i.) daily cigarette consump- tion; ii.) the yield of the cigarette smoked; and iii.) smoking behavior. Smokers of low-yield cigarettes appear to compensate for these lower yields by taking significantly larger puffs and a trend toward a larger number of puffs, but not by an increased cigarette consumption. Despite these compensatory mechanisms in smokers of low-yield cigarettes, these smokers did not achieve the same smoke dose as smokers of high-yield cigarettes. Thus, if tobacco related pulmonary diseases are associated with increased smoke exposure, a lesser degree of lung injury may be expected in smokers using low- yield cigarettes. 136

I I I I I I I I I I I I I I I I I ACKNOWLEDGEMENTS We thank Ms. Judy Combs for expert technical assistance; Ms. Donna Reed for data entry; and Dr. Nancy Haley of the American Health Foundation, Valhalla, N.Y. for the nicotine and cotinine analyses. This study was supported by grants from the Tobacco and Health Research Institute, University of Kentucky, Lexington, KY. REFERENCES 1. Anderson W.H., Bright M., Snider H.L.: Relation of smoking to cardiopulmonary disease. IN: Proceedings of the 5th Workshop Conference of the Tobacco and Health Research Institute of the University of Kentucky, November 1-2, 1979. 2. Benowitz N.L., Hall S.M., Herning R.I., Jacob P. (III), Jones R.T., Osman A.-L.: Smokers of low-yield cigarettes do not consume less nicotine. N. Engl. J. Med. 309: 139-142, 1983. 3. Benowitz N.L., Jacob P.(III), Jones R.T., Rosenberg J.: Interindividual variability in the metabolism and cardiovascular effects of nicotine in man. J. Pharm. Exper. Ther. 221: 368-372, 1982. 4. Benowitz N.L. Kuyt F., Jacob P.(III), Jones R.T., Osman A.-L.: Cotinine disposition and effects. Clin. Pharm. Ther. 34: 604- 611, 1983 5. Borgers D., Junge B. Thiocyanate as an indicator of tobacco smoking. Prev. Med. 8: 351-357, 1979. 6. Bridges R.B., Rehm S.R., Wyatt R.J.: Increased plasma concentrations of C9, Cl-inhibitor and a1-protease inhibitor associated with cigarette smoking. Life Sci. 36: 955-963, 1985. 7. Darby T.D., McNamee J.E., van Rossum J.M.: Cigarette smoking pharmacoki"tics and its relationship to smoking behavior. Clin. Pharmacokinetics 9: 435-449, 1984. 8. FEDERAL TRADE COMMISSION: Report of "Tar", Nicotine and Carbon Monoxide of the Smoke of 207 Varieties of,Domestic Cigarettes. February 1984. 9. Prue D.M., Martin J.E., Humbe A.S.: A critical evaluation of thiocyanate as a biochemical index of smoking exposure. Behav. Ther. 11: 368-379, 1980. 10. Ray, A.A. Ed. SAS User's Guide: Statistics. Cary, North Carolina: SAS Institute, 1982. 11. Russell M.A.H., Sutton S.R.,' Iyer R., Feyerabend C., Vesey C.J.: Long-term switching to low-tar low-nicotine cigarettes. Br. J. Addict. 77: 145-158, 1982. 12. Stepney R.: Consumption of cigarettes of reduced tar and nicotine delivery. Br. J. Addict. 75: 81-88, 1980. 13. Surgeon Ceneral: Smoking and Health: A report of the Surgeon General U.S. Department of Health, Education and Welfare Publication No. (PHS) 79-50066 (1979). 14. Tobin M.J., Sackner M.A.: Monitoring smoking patterns of low and high tar cigarettes with inductive plethysmography. Am. Rev. Respir. Dis. 126: 258-264, 1982. 15. Wald N., Howard S., Smith P.C., Bailey A.: Use of carboxyhaemoglobin levels to predict the development of diseases associated with cigarette smoking. Thorax 30: 133- 140, 1975. 07