This observation supports previous findings that Vif directly interacts with CBFb

terin and IgE, and ANA and IgMRF were calculated using logistic regression models, analyses were adjusted for the same variables, as listed previously. The associations are expressed as odds ratio with CIs. We evaluated effects of cumulative duration and daily dose, DDDs, number of prescriptions, adherence and potency of statins on the immune status. To test P for trend of the associations across increasing quartiles of cumulative duration and daily dose, DDDs, number of prescriptions, adherence or potency of statins, the median values of these different exposure aspects of statin use were assigned to each quartile and used as a continuous variable in the linear regression model. Because of a potential modifying effect due to the presence of cardiovascular diseases, the analysis was stratified according to history of cardiovascular diseases. Changes in the immune system with aging and sex differences have been reported. Therefore, age- and sex-stratified analyses for the evaluation of effect modification were carried out. To study the relationship between statin use and the levels of CRP and IgE antibodies during follow-up, we used linear mixed-effects models for the analysis of repeated measurements with adjustment for the matching and baseline co-variables, as described above. The model deals with the correlation between repeated measurements in a subject, and allows subjects to have unequal gaps and numbers of observations. Only subjects with at least two serological measurements were included in the model. The random effects of the model include a random intercept and/or slope of time. In the model where we only included a random intercept, specification of a random slope did not change the results in a relevant way. Data are 2883-98-9 presented as ‘s with CIs 23103164 and denotes the adjusted percentage change in the levels of CRP and IgE antibodies compared between and within statin users and nonusers. P<0.05 was considered statistically significant using a 2tailed test. Data were missing on several variables as listed in table 1. The missing values were imputed by the multiple imputation method by using the Markov Chain Monte Carlo Method. All original outcome and co-variables presented in table 1 were included in the imputation model. Twenty imputation sets were created, analysed and pooled by the MIANALYZE procedure. Baseline measurements of complete and imputed cases were compared based on the means and frequencies. Disease history before date of examination Fibrates Antihypertensive drugs Anti-diabetic drugs 18201139 Aspirin HRT d Antibiotics Antidepressants PPIs e a BMI = Body mass index b HDL = High density lipoprotein c NSAIDs = Nonsteroidal anti-inflammatory drugs d HRT = Hormone replacement therapy e PPIs = Proton pump inhibitors doi: 10.1371/journal.pone.0077587.t001 analyses are presented. All analyses were performed using SAS version 9.2. Inflammation Measurements of immune parameters are presented in Results Baseline characteristics Baseline characteristics of the study population are presented in table 1. The study comprised 332 statin users of whom 196 were men with a mean age of 59.5 years and 331 non-users of whom 194 were men with a mean age of 59.3 years. In statin users, total cholesterol levels and BMI were higher and the level of HDL cholesterol was lower than in non-users. Hypertension and cardiovascular disease were more frequently reported among statin users than non-users. As a result of these differences, antihypertensive drugs,