Berryman · Clinical nutrition (Edinburgh, Scotland) 2021 · cross-sectional study · n=10,112

Greater protein intake at breakfast or as snacks and less at dinner is associated with cardiometabolic health in adults.

Cited 21 times in the scientific literature.

Level 4 - case-series / case-control

Cross-sectional analysis of NHANES survey data

PubMed 33583662 · doi:10.1016/j.clnu.2021.01.018 · record verified 2026-08-29

What was done

Researchers evaluated cross-sectional data from 10,112 US adults (aged ≥19 years) in the NHANES 2013–2016 dataset to examine associations between protein intake at specific eating occasions and cardiometabolic health markers. Deciles of individual usual intake (IUI) for protein were calculated for breakfast, lunch, dinner, and combined snacks. Regression models adjusted for age, age², gender, ethnicity, physical activity, poverty income ratio, protein intake at other meals, and total energy intake at the evaluated occasion, using a significance threshold of P < 0.0042.

What was found

Protein intake (10th to 90th percentiles) ranged from 5.9 ± 0.1 to 22.6 ± 0.3 g/d at breakfast, 14.0 ± 0.1 to 34.6 ± 0.4 g/d at lunch, 24.3 ± 0.3 to 46.8 ± 0.2 g/d at dinner, and 4.9 ± 0.1 to 16.5 ± 0.2 g/d for snacks. Breakfast protein was inversely associated with diastolic blood pressure (-0.39 ± 0.10 mmHg per decile, P = 0.0003) and systolic blood pressure (-0.40 ± 0.13 mmHg per decile, P = 0.0038), and positively associated with HDL-cholesterol (0.75 ± 0.16 mg/dL per decile, P = 0.0001). Dinner protein was positively associated with insulin concentrations (0.77 ± 0.23 uU/mL per decile, P = 0.0025) and HOMA-IR (0.32 ± 0.09 per decile, P = 0.0017). Snack protein was inversely associated with diastolic blood pressure (-0.41 ± 0.09 mmHg per decile, P < 0.0001) and CVD risk score (-0.0018 ± 0.0004 per decile, P = 0.0001). Protein intake across occasions was not associated with BMI, waist circumference, glucose, total cholesterol, LDL-cholesterol, or triglycerides.

Why it matters

These findings suggest that the timing and meal-specific distribution of daily protein intake, particularly shifting protein toward breakfast rather than dinner, may be relevant for cardiometabolic risk profiles. However, longitudinal and interventional trials are required to determine whether changing protein distribution directly improves health outcomes.

Limits

The cross-sectional design cannot determine temporality or establish causality. Dietary intake was self-reported and susceptible to measurement error. The abstract does not distinguish between protein sources (such as animal versus plant sources) or account for concurrent macronutrient substitutions.

Cited by