Urinary metabolomics signature of animal and plant protein intake and its association with 24-h blood pressure: the African-PREDICT study

Hypertension affects over 1 billion adults globally, and dietary interventions are central to its prevention and management. Global recommendations emphasize higher intake of plant-based foods and limiting saturated fat and sugars. Metabolomics offers tools to elucidate metabolic pathways linking diet to disease and to refine dietary guidance. In Sub-Saharan Africa, low consumption of fruits, vegetables, whole grains, and legumes, together with high sodium intake, are leading dietary risk factors for CVD. Prior work indicates plant-based diets are often associated with lower BP and CVD risk, whereas higher intake of animal proteins—especially red and processed meats—is linked to higher hypertension and CVD mortality risk. These contrasts may reflect differences in intrinsic amino acid profiles, fat content (including saturated fat), micronutrients, haem-iron, fibre, phytochemicals, and interactions with the gut microbiome. The study aimed to determine whether plant and animal protein intakes are differentially related to non-protein dietary components and to a urinary metabolomics profile associated with 24-h BP. The authors hypothesised that participants with low animal, high plant protein intake (low-risk dietary group) would differ in BP, body composition, non-protein dietary components, and metabolomic profiles compared with those with high animal, low plant protein intake (high-risk group).

Evidence from nutrient composition databases shows animal protein sources generally contain higher levels of essential amino acids, including branched-chain amino acids (BCAAs) and aromatic AAs, while plant protein sources are richer in nonessential AAs. Epidemiological and metabolomics studies have associated higher BCAA levels and their short-chain acylcarnitine byproducts with elevated BP and risk of hypertension, while lower levels of certain nonessential AAs (e.g., glycine, serine, histidine) have been linked with higher BP and hypertension risk. High-quality plant-based diets have often been linked to reduced BP and lower CVD mortality, whereas higher animal protein intake is associated with increased risks of hypertension and CVD mortality. In Sub-Saharan Africa specifically, low intake of plant-based foods and high sodium intake are prominent dietary risk factors for CVD. These data collectively suggest a plausible link between protein source, metabolomic signatures, and BP, motivating further validation across diverse populations and the identification of biomarkers to support dietary assessment.

Study design and participants: The African Prospective study on the Early Detection and Identification of CVD and Hypertension (African-PREDICT) recruited 1202 young adults (20–30 years) from communities around Potchefstroom, North West Province, South Africa. Inclusion criteria included clinic BP < 140/90 mmHg at screening, no self-reported chronic disease or treatment, HIV-uninfected, and not pregnant or lactating. For this analysis, N = 1008 participants with both metabolomics and protein intake data were included (49.8% women; 49.9% Black and 50.1% White). Ethical approval was obtained (NWU-00001-12-A1), and the study is registered (NCT03292094). All participants provided written informed consent. Data collection: Demographic and SES data were obtained via validated questionnaires. Dietary intakes (protein, fibre, fats) were assessed using three 24-h dietary recall interviews (including one weekend day), using a standardized kit and five-step multiple-pass method. Anthropometry included body weight, height, and waist circumference measured in triplicate; BMI calculated as kg/m². Physical activity was monitored with a triaxial accelerometer over seven days to estimate total energy expenditure (TEE), corrected for body weight (kCal/kg/day). Blood pressure: Ambulatory BP monitoring (ABPM) measured BP every 30 minutes during the day (06:00–22:00) and hourly at night (22:00–06:00) using a Card(X)plore device on the non-dominant arm. A total of N = 979 participants had ≥12 successful day and 24 successful night readings. Mean successful inflation rate was 88% ± 12.1%. Biological samples and assays: Following an overnight fast (from 22:00), early morning spot urine and blood were collected, processed, and stored at −80°C. Serum total cholesterol, HDL-C, LDL-C, GGT, CRP, and HbA1c were measured (Cobas Integra 400plus). Urinary amino acids and acylcarnitines were quantified in spot urine using LC-MS/MS (Agilent 6410 with 1200 series LC); a 50% CV filter was applied and metabolites with >50% zero values were excluded. Twenty-four-hour urine collections were obtained per WHO/PAHO guidelines to measure sodium and potassium (Cobas Integra 400plus). Estimated salt intake (g/day) was calculated as: UNa (mmol/L) × Uvolume (L/24h) × 23 / 390. Dietary protein grouping: Participants were stratified into low- and high-risk dietary protein groups using sex- and race-specific tertiles of plant and animal protein intakes. Low-risk: lowest tertile of animal protein and highest tertile of plant protein. High-risk: lowest tertile of plant protein and highest tertile of animal protein. Statistical analysis: Data distributions were inspected (skewness, kurtosis, histograms, Q-Q plots). Group comparisons used independent t-tests (normally distributed) and Mann–Whitney U tests (skewed). ANCOVA assessed BP differences between low- and high-risk groups adjusting for BMI, total cholesterol, LDL-C, salt intake, potassium excretion, Na+/K+ ratio, total energy intake, dietary fibre, or saturated fat intake. In the total cohort, associations of total, plant, and animal protein intakes with 24-h BP were analysed via Spearman correlations and multiple linear regression (pairwise deletion), adjusting for sex, race, SES, age, BMI, TEE, Na+/K+ ratio, LDL-C, dietary fibre, and saturated fat. Mediation analyses tested whether BMI, saturated fat, or other covariates mediated protein source–SBP relationships. Associations between urinary amino acids/acylcarnitines and dietary protein intakes were assessed by Spearman correlations (unadjusted and adjusted for sex, race, age, BMI, TEE, and total energy). Differences in metabolite levels between dietary groups were tested with Mann–Whitney U; ANCOVA further adjusted for BMI and total protein or BMI and total energy. Partial correlations were computed using matrix subcommands. Multiple testing was controlled using Benjamini–Hochberg q-values. Associations of 24-h BP with metabolites differing between groups were assessed by Spearman and multiple linear regression (adjusting for sex, race, age, BMI, TEE, and total energy). Model diagnostics included residual P-P plots and checks of homoscedasticity; 24-h BP was normally distributed.

- Sample: 1008 young adults (20–30 years), balanced by sex and race; 979 with valid 24-h ABPM. - Group differences (low-risk: low animal/high plant protein, N=102; high-risk: low plant/high animal protein, N=90): • Anthropometry: High-risk group had greater waist circumference (+6.73 cm; p < 0.001) and BMI (+2.90 kg/m²; p < 0.001). • 24-h BP: SBP higher in high-risk vs low-risk by 3.32 mmHg (p = 0.011); DBP not significantly different (p = 0.14). • Urinary electrolytes: Potassium excretion higher in high-risk (+11.5 mmol/day; p = 0.039); Na+/K+ ratio lower in high-risk (−0.67; p = 0.017); estimated salt intake not different. • Biochemical: Total cholesterol higher (+0.54 mmol/L; p = 0.002); LDL-C higher (+0.39 mmol/L; p = 0.003); GGT higher (p = 0.007); CRP higher (p = 0.007) in high-risk group; HDL-C and HbA1c not significantly different. • Dietary intake: High-risk consumed more total protein (+34.9 g; p < 0.001), more animal protein (+57.4 g; p < 0.001), less plant protein (−20.4 g; p < 0.001), lower fibre (−11.3 g; p < 0.001), higher total fat (+14.0 g; p = 0.007), higher saturated fat (+8.37 g; p < 0.001), higher MUFA (+6.56 g; p = 0.001); PUFA and energy intake not significantly different. Potassium intake was higher in high-risk (p = 0.031). - Adjusted BP comparisons: The SBP difference between dietary groups lost significance after adjustment for BMI (p = 0.19), LDL-C (p = 0.053), potassium excretion (p = 0.26), Na+/K+ ratio (p = 0.09), or saturated fat intake (p = 0.15). - Protein intake–BP associations (total cohort): Total, plant, and animal protein intakes correlated positively with SBP, but associations were not significant in multiple regression after adjustment for confounders. Male sex, BMI (positive; p < 0.001), and saturated fat intake (positive; p = 0.008) were independently associated with 24-h SBP. - Mediation: Relationships of plant and animal protein intakes with 24-h SBP were partially mediated by BMI and saturated fat intake. - Metabolomics and BP within low-risk group: Higher urinary levels of methionine (Std. β = −0.217; p = 0.034), glutamic acid (Std. β = −0.220; p = 0.031), glycine (Std. β = −0.234; p = 0.025), and proline (Std. β = −0.266; p = 0.010) were inversely associated with SBP; beta-alanine inversely with DBP (Std. β = −0.277; p = 0.020).

The study investigated whether differences in plant versus animal protein intake align with distinct urinary metabolomic profiles that relate to 24-h blood pressure in young South African adults. Participants consuming high animal and low plant protein exhibited higher SBP compared with those with high plant and low animal protein, alongside higher BMI, waist circumference, LDL-C, total cholesterol, CRP, GGT, and saturated fat intake, and lower fibre intake. However, when accounting for key confounders—particularly BMI and saturated fat—the between-group SBP difference and the crude correlations of protein intakes with SBP were attenuated and became non-significant. This indicates that the apparent association between protein source and SBP is at least partly explained by adiposity and saturated fat intake rather than protein per se. Mediation analyses supported BMI and saturated fat as intermediaries in the protein source–SBP pathway. Metabolomics findings showed that, within the low-risk (plant-rich) group, several urinary amino acids were inversely related to SBP or DBP, suggesting potential amino acid-related metabolic pathways that may be more favorable to BP regulation in the context of plant-forward diets. Overall, the results align with established dietary guidance promoting plant-based foods and lower saturated fat intake for BP control, and they highlight urinary metabolomic signatures that may serve as supportive biomarkers of dietary patterns and BP health.

In a large cohort of young adults, higher SBP observed in individuals with high animal and low plant protein intake was largely explained by higher BMI and saturated fat intake, indicating these factors—rather than protein source alone—drive BP differences. Plant-forward diets were associated with urinary amino acid profiles inversely related to BP, suggesting metabolomic pathways through which such diets may support BP regulation. The study contributes evidence that urinary metabolomic markers can complement dietary recall data in assessing diet–BP relationships. Future research should validate these urinary biomarkers across diverse populations, employ longitudinal and interventional designs to establish causality, and further disentangle the roles of specific dietary components (e.g., fibre, saturated fat) and metabolomic pathways in BP regulation.