Am J Epidemiol 2003; 158:1021-1022.
Copyright © 2003 by Johns
Hopkins Bloomberg School of Public Health
LETTERS TO THE EDITOR |
THE AUTHORS REPLY
School of Public Health, University of Alabama at Birmingham, Birmingham, AL 35294-0022
We thank Drs. Ebbert, Yang, and Tomar for their observations (1). Although we understand and agree with some of their concerns about our study (2), we believe that they are consequent to limitations of the database used in this research, limitations that were beyond our control, and concerns that exist about many epidemiologic studies. We stand by our results and interpretations.
Their first concern is the comparison of exclusive smokeless tobacco users with "non-tobacco users," a group that potentially includes pipe and cigar smokers. The paper states that other tobacco habits, such as pipe and cigar smoking, were not included in the "the formation of tobacco exposure categories" (2, p. 731). All four tobacco categories included pipe and cigar smokers. However, only about 5 percent of our non-tobacco users reported smoking cigars or pipes, nowhere near the 24 percent quoted in the letter by Ebbert et al. (1). Additionally, pipe and cigar smoking was more common in our data set among the smokeless tobacco users than among the non-smokeless tobacco users. We compared a group of individuals who used smokeless tobacco and smoked pipes or cigars with a group of individuals who did not use smokeless tobacco and smoked pipes or cigars to a lesser extent, and still we found no increased mortality risk. Limitations in the data did not allow us to effectively control for pipe and cigar smoking.
Their second concern is related to the lack of distinction between chewing tobacco and snuff and the effect this may have on our results. We agree that the increased hazard ratio observed among women may be because they used a different, more hazardous smokeless tobacco product. Nowhere in the paper do we claim that the results of our study are applicable to the smokeless tobacco products used today, which the literature suggests are safer than the smokeless tobacco products from 20 years ago because they contain substantially lower levels of tobacco-specific nitrosamines (3).
Their third concern relates to our choice of "ever" use of smokeless tobacco as our exposure category. This limitation was mentioned in the Discussion (2, p. 736). In addition to the 26 percent of subjects that we suggest used smokeless tobacco daily, the paper also states that 60 percent used it in the past year, additional evidence for regularity of use by our "smokeless tobacco users." If information on the duration or amount used had been available, we would have conducted the analyses accordingly. We clearly state in the concluding sentence that further studies are needed to determine if the duration and intensity of use are related to increased mortality (2, p. 736).
The last concern of Ebbert et al. is related to the overall interpretation of our results. We hope that we have made all attempts to present the data in a clear and impartial manner that would avoid misinterpretation. We do not state that use of smokeless tobacco is harmless; rather, we state that, in this population of smokeless tobacco users in the 1970s, smokeless tobacco use did not lead to increases in mortality for the diseases studied.
REFERENCES
- Ebbert JO, Yang P, Tomar SL. Re: "Chronic disease mortality in a cohort of smokeless tobacco users." (Letter). Am J Epidemiol 2003;158:000–00.
- Accortt NA, Waterbor JW, Beall C, et al. Chronic disease mortality in a cohort of smokeless tobacco users. Am J Epidemiol 2002;156:730–7.
[Abstract/Free Full Text] - Djordjevic MV, Brunnemann KD, Hoffmann D. The need for regulation of carcinogenic N-nitrosamines in oral snuff. Food Chem Toxicol 1993;31:497–501.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||