Author + information
- Mark Hamer, PhD* (, )
- Emmanuel Stamatakis, PhD,
- Mika Kivimaki, PhD,
- Gordon D. Lowe, DSc and
- G. David Batty, PhD
- ↵*Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London, WC1E 6BT, United Kingdom
We thank Dr. Onat for his comments on our recent paper (1). Regarding the first issue, it would indeed be inappropriate to model cotinine as a continuous parameter because it has a large variation and highly skewed distribution. That is why we transformed cotinine into a categoric variable, which is considered a standard epidemiologic approach to deal with this issue. When creating a categoric variable, it is inevitable that arbitrary cut points will be used, and this might be one reason for the lack of linearity observed. Dr. Onat alludes to the medium secondhand smoke (SHS) category having a borderline significantly lower risk; however, in fact, the 95% confidence intervals for cardiovascular disease (CVD) death in the fully adjusted model (0.62 to 1.19) provide no robust evidence for an association. Furthermore, the results in never-smokers do suggest a clear linear association between SHS and CVD death (hazard ratios for medium and high SHS categories are 1.28 and 2.22, respectively). The inclusion of ex-smokers in the main analyses might have considerably masked the true effects of SHS because ex-smokers already have an excess risk of mortality (2), and thus the presence of ex-smokers in the low SHS-exposed group may dilute the effects.
In relation to the issue regarding C-reactive protein (CRP), we agree that these data should be viewed cautiously because it is unclear if CRP does play a causal role in CVD etiology or is merely a biomarker (3). In the present studied cohort, CRP is independently associated with CVD events, although it does not add prognostic significance to established models such as the Framingham risk score (4).
Last, we are certainly not in agreement with Dr. Onat's opinion that our data call for reassessing the acceptance of an adverse relation between active and passive smoking and CVD risk. Our data clearly show a strong association between active smoking and CVD, which is consistent with a plethora of previous epidemiologic evidence over the past few decades. We believe our data also support an association between SHS and CVD, especially in never-smokers. Future large-scale studies, such as the recent data from the European Prospective Investigation into Cancer and Nutrition (5), are required to further explore this potentially important public health issue.
- American College of Cardiology Foundation