Dietary patterns identified by latent class analysis in relation to the risk of cardiovascular disease: Tehran lipid and glucose study.

Background: Several data-driven reduction techniques have been developed to derive dietary patterns, each with different underlying assumptions and approaches to data handling. Here we aimed to determine the major dietary patterns of Tehranian adults using Latent Class Analysis (LCA) method, and assess the association between extracted patterns and cardiovascular disease (CVD) risk.

Methods: 1849 adult men and women, aged ≥ 30 years, from the participants of the third phase (2006-2008) of the Tehran Lipid and Glucose Study (TLGS), without a CVD history were included. Baseline dietary intakes were estimated using a validated 168-items semi-quantitative food frequency questionnaire. Dietary patterns were obtained by LCA method. Adjusted Hazard Ratios (HRs) and 95% confidence intervals (CIs) of CVD were calculated for the association of incident CVD and extracted dietary patterns.

Results: LCA classified the participants into four exclusive classes; named as "mixed pattern", "healthy pattern", "processed foods pattern", "alternative class". After adjustment for confounding variables, there was no significant association between LCA-derived classes and CVD incidence.

Conclusion: In this cohort of Tehranian adults, dietary patterns identified using the LCA method were not significantly associated with CVD risk over 10 years of follow-up. These findings suggest that LCA-derived dietary classifications may have limited predictive utility for CVD in this context. Future studies should consider combining LCA with other dietary assessment methods, incorporating repeated dietary measurements, and evaluating population-specific dietary behaviors to better understand diet-disease relationships.