Evaluation of Preconception Dietary Patterns in Women Enrolled in a Multisite Study.

Background: Diet indices are widely used in nutritional research across communities but do not "capture" the full extent of diet variability across multiple countries. Empirically derived dietary patterns can provide additional information because they reflect combinations of foods potentially associated with health outcomes. Limited studies have evaluated preconception dietary patterns in heterogeneous populations.

Objectives: In the multisite Nutritional Intervention Preconception and During Pregnancy to Maintain Healthy Glucose Metabolism and Offspring Health (NiPPeR) study, the secondary aims included: ) derive pooled and site-specific preconception dietary patterns, and 2) evaluate these patterns using anthropometric measures and metabolic biomarkers.

Methods: Women planning pregnancy ( = 1720) in the United Kingdom, Singapore, and New Zealand completed interviewer-administered harmonized FFQs and lifestyle questionnaires at recruitment. Across-cohort ("pooled") and site-specific dietary patterns were derived, and associations between dietary pattern scores and BMI, waist-to-hip ratio, plasma lipids, and glycemia assessed using multivariable linear regression, expressing results as SD change in outcome per SD change in dietary pattern score.

Results: The pooled analysis identified 3 dietary patterns: "Vegetables/Fruits/Nuts" ("Healthy"), "Fried potatoes/Processed meat/Sweetened beverages" ("Less Healthy"), and "Fish/Poultry/Noodles/Rice" ("Mixed"). The "Healthy" and "Less Healthy" pooled pattern scores were highly correlated with their corresponding site-specific dietary pattern scores ("Healthy": ρ = 0.87-0.93; "Less Healthy": ρ = 0.65-0.88). Women with higher scores for the "Healthy" pooled pattern had a lower waist-to-hip ratio (standardized β: -0.10; 95% CI: -0.18, -0.01); those with higher scores for the "Less Healthy" pooled pattern had a higher BMI (standardized β: 0.17; 95% CI: 0.09, 0.24), higher LDL cholesterol (standardized β: 0.10; 95% CI: 0.01, 0.19), and less optimal glucose profiles. However, we noted higher adherence to the "Healthy" pooled pattern with higher BMI.

Conclusions: The "Healthy" and "Less Healthy" pooled patterns were comparable to the corresponding site-specific patterns. Although the associations between these patterns and objective anthropometric/metabolic measures were largely in the expected directions, future studies are required to confirm these findings.This trial is registered at clinicaltrials.gov (NCT02509988).