Model 3 was further adjusted for use of vitamin/mineral supplements.
Subsequent analyses were further adjusted for use of ACEIs or ARBs, blood pressure, lipid levels, A1C, and smoking.
Case and control patients with a recorded diagnosis of type 2 diabetes were matched on age, calendar time, and general practice, and the multivariate conditional logistic regression analyses were further adjusted for use of oral antidiabetes drugs, insulin, estrogens, smoking BMI, diabetes duration, and HbA1c (A1C).
Similar results held if we further adjusted for uses of β2 agonists and anti-cholinergic bronchodilators (model 4 in Additional file 1).> -wrap-foot> aVariables were dichotomized as 1 = yes, 0 = no (baseline) unless stated otherwise.
We performed sensitivity analyses, further adjusting for use of hormone replacement therapy (HRT) in regression models in women only and also exploring for birth cohort effects.
Analyses of current daily exclusive snus use or smoking versus no current daily tobacco use were further adjusted for past use of the other type of tobacco.
In our primary analyses of the potential interaction of blood Pb concentration with dietary nutrients, we fit models stratified according to intake from food and then further adjusted for supplement use (yes, no), thus treating supplement use as a surrogate for confounding by socioeconomic status (SES), especially in the stratum of high intake.
We conditioned on matched sets (matched on sex, region, and age within 5 years) and further adjusted for snuff use among male smokers, smoking among male snuff users, and education, using this last variable as an indicator of social class.
As BMI violated the proportional hazard assumption, Cox models stratified on quartiles of BMI were used to adjust for the confounding effect of BMI; Model 3, further adjusted for drug use at enrollment, including lipid lowering drugs, renin-angiotensin system inhibitors, oral anti-diabetes drugs and insulin.
27 We further adjusted for hormone use (oral contraceptives and HT) and reproductive factors.
Models were further adjusted for ever use of other antidiabetic agents (metformin, sulfonylureas, insulins, and other oral hypoglycaemic agents), defined as at least one prescription between cohort entry and the year before index date.
