(P29-031-23) Dietary and Lifestyle Inflammation Scores: The Boston Puerto Rican Health Study

Objectives: To test the utility of the Dietary Inflammation Score (DIS) and Lifestyle Inflammation Score (LIS) in Puerto Rican older adults.

Methods: The DIS (previously validated in non-Hispanic White and Black adults) was created by published methods (19 foods, beverages, and micronutrient supplements) using mean dietary intakes from baseline and visit 2 (v2, at ~2.2 y; n=1514). The LIS (smoking, physical activity, alcohol use, and BMI) was created from v2 data (n=1203). Correlation and multiple linear regression were used to examine associations between the DIS and LIS with markers of inflammation (CRP, IL-6, TNFα) and a biomarker score (BMS; the sum of inflammatory markers following standardization). One-way ANOVA was used to test mean CRP across tertiles of DIS and LIS. Scores were recreated after adaptation of dietary intake across 27 culturally tailored food groups, and modification of lifestyle factors, for future validation in a Hispanic cohort.

Results: Mean age was 59 y ± 7.7, 72% female, and 51% had serum CRP >3mg/L. The DIS (range: -11.7 to 9.7) was not correlated with CRP, TNFα, IL6, or BMS (r=0.08, 0.05, 0.12, 0.07, respectively). The LIS (range -0.8 – 2.4) was moderately correlated with CRP r=0.30, IL6 r=0.35, and BMS r=0.24, but not with TNFα r=0.09. There were no significant differences in CRP across DIS tertiles (P range 0.78-0.98); mean CRP was 6.0 ± 12.3 in tertile (t) 1, 5.9 ± 6.6 in t2, and 7.2 ± 12.7 in t3. Mean CRP was 4.1 ± 8.2 in LIS t1, 6.52 ± 13.5 in t2, and 7.8 ± 14.4 in t3; significant differences were observed between mean LIS t1 vs t2 (P < 0.001), t1 vs t3 (P < 0.001), and t3 vs t2 (P < 0.001). Higher IL-6 was related to higher DIS (β 0.03±0.01, P=0.004) and LIS (β 0.30±0.036 P < 0.001), after adjustment. Higher BMS was related to higher DIS (β 0.06±0.02, P=0.02) and LIS (β 0.81±0.08, P < 0.001) after adjustment. Higher CRP was related to higher LIS (β 0.48±0.05, P < 0.001) after adjustment; TNFα was not related to scores (P range 0.17 - 0.50). Adapted DIS and LIS ranges were -0.84 to 0.93 and -0.34 to 0.84, respectively.

Conclusions: DIS was not related to CRP or TNFα in a cohort of older Puerto Rican adults but was related to IL-6 and BMS. LIS was strongly associated with CRP, IL-6, and BMS, but not with TNFα. Lifestyle factors are better predictors of inflammation than diet in this population. Alternatively, the use of a culturally adapted DIS may be relevant in future research.

Funding Sources: NIH