We have reanalyzed the data according to the convincing arguments forwarded by trained statisticians and epidemiologists stating that analysis involving other issues than asthma ought to be analysed in the random sample alone and not as forwarded including the symptomatic sample. However, the reanalysis has not changed our conclusions or interpretation of the results.
Methods
The present study was designed to investigate risk factors for asthma. Our purpose was therefore to investigate a case‐enriched study‐sample. We recruited 1,191 subjects for clinical investigation, and in 295 subjects we analysed for enzyme activity of which 179 subjects recruited as random sample.
Results
Gender difference in GPX1‐activity was first reported among the 295 subjects to be 3.7 (95% CI: 2.0–5.5) U/g protein. In subjects recruited as random sample the difference in activity was 3.6 (95% CI: 1.2–6.0) U/g protein. Association between GPX1 activity and selenium in males was still significant when restricted to subjects recruited as random sample The difference in GPX1 activity per Δ 10 μg/L serum selenium was 2.5 U/g protein (95% CI: 1.1–3.9). Adjusted for smoking the corresponding difference was 2.3 U/g protein (95% CI: 0.8–3.7). presents the combined effects of sex, genotype, and smoking status (current smokers vs. former or never smokers) on GPX1 activity in subjects recruited as random sample as it should have been presented.
Discussion
Restricting the number of subjects eligible for analysis of associations between GPX1‐activity and gender, genotypes, smoking and selenium to subjects recruited as random sample had only minor impact on results. In males selenium and smoking are associated to GPX‐activity, whereas in females GPX1‐genotype is associated to GPX‐activity.
Reference
- Malling TH, Sigsgaard T, Andersen HR, et al. Sex determines the influence of smoking and gene polymorphism on glutathione peroxidase activity in erythrocytes. Scand J Clin Lab Invest 2009; 69: 295–302