Assessing Validity of Self-Reported Dietary Intake within a Mediterranean Diet Cluster Randomized Controlled Trial among US Firefighters

Collecting dietary intake data is associated with challenges due to the subjective nature of self-administered instruments. Biomarkers may objectively estimate the consumption of specific dietary items or help assess compliance in dietary intervention studies. Our aim was to use a panel of plasma and urine biomarkers to assess the validity of self-reported dietary intake using a modified Mediterranean Diet Scale (mMDS) among firefighters participating in Feeding America's Bravest (FAB), an MD cluster-randomized controlled trial. In our nested biomarker pilot study, participants were randomly selected from both the MD intervention group ( = 24) and the control group ( = 24) after 12-months of dietary intervention. At baseline data collection for the pilot study (t = 12-months of FAB), participants in the control group crossed-over to receive the MD intervention (active intervention) for 6-months. Participants in the intervention group continued in a self-sustained continuation phase (SSP) of the intervention. Food frequency questionnaires (FFQ), 13-item-mMDS questionnaires, 40 plasma fatty acids, inflammatory biomarkers and urinary hydroxytyrosol and tyrosol were analyzed at both time points. Spearman's correlation, -tests and linear regression coefficients were calculated using SAS software. Overall, the mMDS derived from the FFQ was highly correlated with the specific 13-domain-mMDS (r = 0.74). The concordance between the two questionnaires for low and high adherence to MD was high for all the participants in the parent trial (κ = 0.76). After 6 months of intervention in the pilot study, plasma saturated fatty acid decreased in both groups (active intervention: -1.3 ± 1.7; = 0.002; SSP: -1.12 ± 1.90; = 0.014) and oleic acid improved in the SSP ( = 0.013). Intake of olive oil was positively associated with plasma omega-3 ( = 0.004) and negatively with TNF-α ( < 0.001) at baseline. Choosing olive oil as a type of fat was also associated with higher levels of plasma omega-3 ( = 0.019) at baseline and lower TNF-α ( = 0.023) at follow up. Intake of red and processed meats were associated with lower serum omega-3 ( = 0.04) and fish consumption was associated with lower IL-6 at baseline ( = 0.022). The overall mMDS was associated with an increase in plasma omega-3 ( = 0.021). Good correlation was found between nutrient intake from the FFQ and the corresponding plasma biomarkers (omega-3, EPA and DHA). In this MD randomized controlled trial, some key plasma bioma