The simple strategy and the possibility of considering the effect of adherence to the recommendations are lost by using quartiles as cut points.
Combined fish and shellfish intake, total fish intake, lean fish intake, fatty fish intake, and shellfish intake were categorized by using quartiles (based on the distributions in the subcohort).
Differences between quartiles (qrtl) of morning cortisol were examined by GLM (general linear model), using quartile 1 as reference.
Differences between quartiles (qrtl) of evening cortisol were examined by GLM (general linear model), using quartile 1 as reference.
Differences between quartiles (qrtl) of Δ-cortisol (logarithmic morning salivary cortisol – logarithmic evening salivary cortisol) were examined by GLM (general linear model), using quartile 1 as reference.
In order to identify a subgroup of patients with a particularly poor prognosis, we also split patients into PSA/Ki67 groups using quartile 4 (≥5.4%) as cut-off value for Ki67 high; PSA high/Ki67 quartile 4-low (121/237, 51%, of men managed by watchful waiting), PSA high/Ki67 quartile 4-high (11/237, 4.6%), PSA low/Ki67 quartile 4-low (78/237, 33%), and PSA low/Ki67 quartile 4-high (27/237, 11%).
We estimated dose response models by linear regression for individual PBDE congeners using quartiles, with quartile 1 as the referent group.
In the models using quartiles, p-trend was calculated by modeling the quartiles, coded as 1, 2, 3, and 4, as a continuous variable.
When using quartiles, one quartile increase of AURKA-CN reduced the hazard by 0.63 (95% CI: 0.40 0.97, P=0.04).
