Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Environmental tobacco smoke and mortality
- Hisato Takagi, Toshiyuki Tanabashi, Norikazu Kawai, Takayoshi Kato (19 August 2006)
|
|
|||
|
Hisato Takagi, consultant cardiovascular surgeon Shizuoka Medical Centre, Shizuoka 411-8611, Japan, Toshiyuki Tanabashi, Norikazu Kawai, Takayoshi Kato
Send response to journal:
|
Wen and colleagues(1) analysed data of over 72 000 women who had never smoked and concluded the following in the abstract of their article: “exposure to environmental tobacco smoke is related to moderately increased risk of all cause mortality and mortality due to lung cancer and cardiovascular disease”. However, significant results in abstracts are common but should generally be disbelieved.(2) As mentioned in Table 4, neither ever exposure to tobacco smoke from husbands (hazard ratio 1.11, 95% confidence interval 0.99 to 1.25), ever exposure at work (1.08, 0.91 to 1.28), ever exposure in early life from family members (0.98, 0.85 to 1.14), nor ever exposure to all three sources (1.15, 0.95 to 1.41) was associated with significantly (i.e. statistically) increased all cause mortality. According to a subgroup analysis, only current exposure from husbands (1.15, 1.01 to 1.31) was associated with significantly increased all cause mortality. Therefore, the statement in the abstract should be not “exposure to environmental tobacco smoke is related to moderately increased risk of all cause mortality” but “exposure to environmental tobacco smoke is NOT related to increased risk of all cause mortality” or “ONLY CURRENT EXPOSURE to tobacco smoke FROM HUSBANDS is related to increased risk of all cause mortality”. Furthermore, neither ever exposure from husbands (1.09, 0.74 to 1.61), ever exposure in early life (0.88, 0.55 to 1.43), nor ever exposure to all three sources (1.03, 0.57 to 1.87) was associated with significantly increased mortality due to lung cancer (table 4). Only ever exposure at work (1.79, 1.09 to 2.93) was associated with significantly increased mortality due to lung cancer. Therefore, the statement in the abstract should be “ONLY EXPOSURE to tobacco smoke AT WORK is related to increased risk of mortality due to lung cancer”. Finally, neither ever exposure from husbands (1.18, 0.92 to 1.51), ever exposure at work (0.92, 0.64 to 1.32), ever exposure in early life (1.26, 0.94 to 1.69), nor ever exposure to all three sources (1.45, 0.95 to 2.22) was associated with significantly increased mortality due to cardiovascular disease (table 4). According to a subgroup analysis, only current exposure from husbands (1.37, 1.06 to 1.78) was associated with significantly increased mortality due to cardiovascular disease. Therefore, the statement in the abstract should be “exposure to environmental tobacco smoke is NOT related to increased risk of mortality due to cardiovascular disease” or “ONLY CURRENT EXPOSURE to tobacco smoke FROM HUSBANDS is related to increased risk of mortality due to cardiovascular disease”. Abstracts of research articles must reflect studies fairly and present the results without bias,(2) because abstracts are often the only part that is read, and only about half of all results initially presented in abstracts are ever published in full.(3) Journal editors should scrutinise abstracts more closely and demand that research protocols and raw data be submitted with the manuscript.(2) 1 Wen W, Shu XO, Gao YT, Yang G, Li Q, Li H, et al. Environmental tobacco smoke and mortality in Chinese women who have never smoked: prospective cohort study. BMJ 2006; 333:376. 2 Gotzsche PC. Believability of relative risks and odds ratios in abstracts: cross sectional study. MBJ 2006;333:231-4. 3 Scherer RW, Langenberg P, von Elm E. Full publication of results initially presented in abstracts. Cochrane Database Methodol Rev 2005;2:MR000005. Competing interests: None declared |
|||