Does the Inclusion of Free Sugars as Opposed to Total Sugars in Nutrient Profiling Models Improve Their Performance?...

Nutrient profiling (NP) models characterize the healthfulness of foods. Few NP models have been validated, and nutrients included in their algorithm do not always reflect the most recent scientific evidence.

This study aimed 1) to evaluate the validity of NP models against a diet quality measure and cardiometabolic risk factors in French-Canadians and 2) to compare the validity of each model when replacing total sugars by free sugars in their algorithm.

The PRÉDicteurs Individuels, Sociaux et Environnementaux cross-sectional study was used to test original and modified versions (i.e., including total or free sugars, respectively) of 3 NP models: Health Star Rating (HSR) system, Nutri-Score, and Nutrient-Rich Food index 6.3. Data from web-based self-administered 24-h recalls completed by 1019 adults were used to calculate energy-weighted NP-derived individual scores for both versions of each model. Associations between individual scores and the Healthy Eating Food Index 2019, as well as 14 biomarkers covering anthropometry, blood pressure, blood lipids, glucose homeostasis, and inflammatory biomarkers, were assessed using multivariable linear models.

Higher quality of foods consumed, as assessed by all 3 original models, was associated with higher Healthy Eating Food Index 2019 (adjusted R²: 0.43-0.55), and with lower body mass index (BMI) (β: from -0.16 to +0.48 kg/m²; P ≤ 0.0001), diastolic blood pressure (β: -0.08 to +0.30 mm Hg; P ≤ 0.04) and triglycerides (β: -0.01 to +0.02 mmol/L; P ≤ 0.002). Original HSR and Nutri-Score were also associated with lower waist circumference and HOMA-IR, lower insulin (HSR only), and higher HDL cholesterol (Nutri-Score only). Replacing total sugars by free sugars in each model only slightly increased the number of associations observed with biomarkers.

All 3 models are associated with diet quality and some biomarkers of health status in French-Canadians, although no model outranks the others. Replacing total sugars by free sugars has little to no effect on NP models' performance, therefore not supporting this approach for now.