Diet, Iq'mik Smokeless Tobacco Use and Cardiometabolic Risk among Yup'ik Alaska Native People living in Southwest Alaska
Yup'ik Alaska Native people live in rural communities in Southwest Alaska. Many Yup'ik people retain aspects of their traditional lifestyle, including a subsistence diet and use of iq'mik smokeless tobacco. We aimed to characterize associations between these aspects of Yup'ik lifestyle and biomarkers of cardiometabolic (CM) status (e.g., lipids, blood pressure, glucose, adiposity). We conducted three analyses using Center for Alaska Native Health Research adult study participants. First, we tested the reproducibility and reliability of previously identified dietary patterns: "processed foods"; "fruits and vegetables"; and "subsistence foods". We used confirmatory factor analysis to measure reproducibility and composite reliability (n=267) and intraclass correlation coefficients for test-retest reliability (n=113). Next, we characterized associations between dietary pattern quartiles and biomarkers of CM status using regression adjusted for confounders (n=637). Finally, we characterized the association between current iq'mik use and biomarkers of CM status using regression models adjusting for measures of Yup'ik lifestyle or adjusting for a propensity score (n=874). The results confirmed the dietary patterns based on acceptably correlated factor loadings for 17 of the 18 foods used in the analysis and model fit criteria were all above the 0.90 threshold. Composite and test-retest reliability were respectively 0.73 and 0.66 for "processed foods", 0.72 and 0.54 for "fruits and vegetables", and 0.56 and 0.34 for "subsistence foods". Comparing participants in the fourth to first quartile for each dietary pattern, we identified significant associations between "processed foods" and log triglycerides (β=0.11); "fruits and vegetables" and diastolic blood pressure (DBP) (β=2.87) and HbA1c (β=-0.08); and "subsistence foods" and log triglycerides (β=-0.10) and DBP (β=-3.99). Current iq'mik use was significantly associated with log high-density lipoprotein cholesterol (β=0.05), log triglycerides (β=-0.07), HbA1c (β=-0.05), log fasting blood glucose (β=-0.02), log waist circumference (β=-0.04), and log body mass index (β=-0.04), using either adjustment method. Therefore, we confirmed the reproducibility and reliability of the dietary patterns, and found evidence that both diet and iq'mik use were associated with CM risk. The dietary associations align with results from previous studies, and can be used for future research and development of health interventions. However, the iq'mik findings are novel requiring further study.