776 



CoulUs, SUdley, and Sameb FTC Tar and Nicotine Yields and Markers of Exposure to Tobacco Smoke 



439 



TABLE 3 



MULTIPLE UNEAR REGRESSION MODELS' FOR PREOICTINO 



SAUVARY COTININE LEVEL (N(VML) 



' Fot t^M• motjau. compWI* data with standanl fm (S£) in pa/amhaaas. troifi 163 of 299 

 currant smokara wara includad 

 1 Daflnad aa an indcaior variawa - flO mm or graatar. 1 • l«aa than 60 mm 

 * p < O.OS 



colleagues (12) compared questionnaire responses on current 

 brand of cigarette smoked witli an empty cigarette pack obtained 

 within 2 wk of completing the questionnaire. Of 220 smokers who 

 returned cigarette packs, the overall accuracy for reports on the 

 current brand of cigarettes smoked was 74%: most discrepancies 

 were minor and dkj not result in misclassification of the usual brand 

 smoked. The statistical power of our study for detecting signifi- 

 cant effects of FTC yields on levels of biomarkers may have also 

 been limited by the small number of subjects using low-yield cig- 

 arettes (< 5 mg lar/cigarette). 



Although information is available on the prevalence of smok- 

 ing and on smoking habits in this study population (5) and other 

 Hispanic populations (13-16), little is known about the types of 

 cigarettes smoked (16). In New Mexico, the prevalence of smok- 

 ing is similar for Hispanics and non-Hispanic whites, but on aver- 

 age Hispanics smoke fewer cigarettes per day (5). However, the 

 prevalence of smoking may vary widely among different groups 

 of Hispanics within the United States (14, 16). 



The types of cigarettes smoked have been determined in sur- 

 veys of smokers in the United Kingdom (17) and the United States 

 (3, 18-21). Available data on the use of various cigarette types 

 in Hispanic populations of the United States are limited to brand 

 without specific tar and nicotine information (16, 22). In 1986, a 

 survey of U.S. smokers 17 yr of age and older found that 41% 

 reported smoking cigarettes with tar yields of 15 mg or less (3). 

 This proportion varied by racial group: 44% for whites, 22% for 



blacks, and 32% for "Wher' racial groups. The distributrans of types 

 of cigarettes smoked based on tar yield in our study (figure 1) are 

 most similar to those of blacks and other racial groups in the 1986 

 nationwide survey. 



The relationships among current smoking habits and various 

 biologic markers of smoking have been examined in smokers at- 

 tempting to quit (23-25) and In population-based surveys of 

 smokers (21, 26, 27). Among 353 Australian cigarette smokers in 

 a population survey. Pierce and coworkers (26) reported a corre- 

 lation coefficient of 0.33 between reported number of cigarettes 

 smoked per day and salivary cotinine. Using linear regression to 

 predict salivary cotinine, with cigarettes smoked per day, time since 

 last cigarette, and cigarette strength (filler or nonfilter) as inde- 

 pendent variables, the model explained only 13.6% of the vari- 

 ability of the salivary cotinine level. Maron and Fbrtmann (21) ex- 

 amined data from 713 California smokers and found that sex and 

 number of cigarettes smoked per day explained 28 and 22% of 

 the variability in carbon monoxide and salivary thiocyanate, 

 respectively. FTC nicotine yields explained only an additional 1 

 and 2% of the variability, respectively. 



Failure to find a relationship between FTC yields and biologic 

 markers of exposure to tobacco combustion products may be ex- 

 plained by several factors, including limitations of the method for 

 determining FTC yields, variation in patterns of smoking among 

 smokers, and individual differences in uptake and metabolism of 

 tobacco combustion products FTC yields are determined by 

 machines that smoke with a constant pattern: the machine pat- 

 tern of smoking may not accurately reflect actual smoking pat- 

 terns or the variability of smoking patterns among humans (4, 23, 

 28). Among smokers smoking the same type of cigarette, differ- 

 ences in smoking patterns (29). such as puff volume and inhala- 

 tion time and the practice of blocking the cigarette filter (30), may 

 affect the doses of toxic components of smoke delivered to the 

 lung. Furthermore, as smokers reduce the numbers of cigarettes 

 smoked or change to lower yield cigarettes, levels of nicotine and 

 of carbon monoxide decline, but less than predicted based on 

 the decrease in the number of cigarettes smoked (31) or the ciga- 

 rette tar yield (32). To maintain a constant nicotine level, changes 

 in smoking patterns may follow reductions in the numbers of cig- 

 arettes smoked or switching to cigarettes with low nicotine yields 

 (33). These findings illustrate the difficulties in measuring exposure 

 to tobacco combustion products and give evidence of the inac- 

 curacy of measures obtained from FTC figures of cigarette yields. 

 Other factors that may contribute to variability in measures of ex- 



' Comptate data from I9fl tubftcts wt% inchidad m th« lhr«* modilt r«(ated to FTC tar and ntcaUn* (Coiumn* 1 (o 3) and data Irom 1 

 wef« inctudsd in tha two modela rvtalad to FTC CO (Columna 4 and S) 

 ^ Bflcauaa at moalng dau and varying umpta •m, ffw rvgrMaton coafficwtta wy among Ihaaa modati 



* p < 06 



* Vim aino* laat ctganna daAnad aa an Indlcatar vartaUa: O - ao inin or graatar. I - laaa ttun 60 mm. 



