American Journal of Epidemiology Advance Access originally published online on February 22, 2008
American Journal of Epidemiology 2008 167(7):882; doi:10.1093/aje/kwn023
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
LETTERS TO THE EDITOR |
RE: "CIGARETTE SMOKING AND ERECTILE DYSFUNCTION AMONG CHINESE MEN WITHOUT CLINICAL VASCULAR DISEASE"
1 Department of Health Sciences Research, Mayo Clinic, Rochester, MN 55905
2 Southern California Permanente Medical Group, Pasadena, CA 91101
(e-mail: gades.naomi{at}mayo.edu)
We read with great interest the recent article in the Journal by He et al. (1). This research adds to the growing body of evidence that smoking may have yet another untoward consequence—one that hits men below the belt. The authors' thoughtful discussion helped put their findings in perspective. As we read the article, however, we felt that several points deserved further comment.
First, the authors' assessment of the history of clinical vascular disease appears to have been based solely on self-report, since the clinical examination, with the exception of a blood sample for assessment of serum cholesterol and triglyceride levels, was not diagnostic for clinical vascular disease. The reliance on self-reporting may have resulted in inaccurate information on clinical vascular disease in this cohort. This information could be confirmed by medical record review for a sample of participants to determine whether respondent bias might have been a problem in this study.
Of greater concern, though, is the presentation of results in the authors' tables 2 and 3 (1, pp. 806–807). In these tables, He et al. presented risk estimates and etiologic fractions for alcohol drinking, physical activity, body mass index, hypertension, serum cholesterol level, and diabetes, adjusted for age and educational attainment, without presenting the bivariate associations. This may account for their statement in the Discussion: "Our study did not find a significant association of diabetes, overweight, and dyslipidemia with erectile dysfunction" (1, p. 808). As the authors pointed out, this is contrary to the findings of many previously published studies. While age adjustment is often a knee-jerk approach in epidemiology, the adjustment for age may have negated a real association between erectile dysfunction and diabetes, a generally accepted pathologic (etiologic) factor (2). Because age is associated with diabetes, adjustment for age will account for age differences in the comparison. However, this may have the untoward effect of removing the portion of the true association between diabetes and erectile dysfunction that is coincidentally associated with age. Granted, the judgment as to whether this adjustment removes a confounding effect or overadjusts the results is somewhat subjective. Nonetheless, the reader is hampered in making that judgment if the bivariate associations are not presented.
Finally, He et al. suggest that the best evidence for this association would come from a prospective cohort study with a large, population-based sample. We agree wholeheartedly and would like to point them to our 2005 paper, published in the Journal (3). There we found that in a cohort of Caucasian men, current smokers (vs. former and never smokers) in their forties had the greatest relative odds of erectile dysfunction (relative odds (RO) = 2.74, 95 percent confidence interval (CI): 0.44, 16.89) as compared with men in their fifties (RO = 1.38, 95 percent CI: 0.51, 3.74), sixties (RO = 1.70, 95 percent CI: 0.82, 3.51), and seventies (RO = 0.77, 95 percent CI: 0.27, 2.21). Compared with men who had never smoked, men who had smoked at some time had a greater likelihood of erectile dysfunction (age-adjusted odds ratio = 1.42, 95 percent CI: 1.00, 2.02), and there was a dose response (3).
The results and conclusions presented by He et al. (1) are noteworthy and add to our knowledge about the association between smoking and erectile dysfunction. These findings may help to drive home an antismoking message to recidivist men. They also highlight the fact that presentation of both bivariate and multivariate results could substantially aid readers in understanding the effects of confounding versus overadjustment.
 |
ACKNOWLEDGMENTS |
|---|
Conflict of interest: none declared.
|
References |
|---|
 TOP References |
|---|
- He J, Reynolds K, Chen J, et al. Cigarette smoking and erectile dysfunction among Chinese men without clinical vascular disease. Am J Epidemiol (2007) 166:803–9.
[Abstract/Free Full Text] - Burke JP, Jacobson DJ, McGree ME, et al. Diabetes and sexual dysfunction: results from the Olmsted County study of urinary symptoms and health status among men. J Urol (2007) 177:1438–42.[CrossRef][Web of Science][Medline]
- Gades NM, Nehra A, Jacobson DJ, et al. Association between smoking and erectile dysfunction: a population-based study. Am J Epidemiol (2005) 161:346–51.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. He, K. Reynolds, and D. Gu THREE OF THE AUTHORS REPLY Am. J. Epidemiol., April 1, 2008; 167(7): 882 - 883. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||