Modulation of circulating levels of advanced glycation end products and its impact on intima-media thickness of both common carotid arteries: CORDIOPREV randomised controlled trial

Modulation of AGE Levels and Carotid IMT: CORDIOPREV RCT (Gutierrez-Mariscal 2024)

TL;DR

This pre-specified sub-study of the 5-year CORDIOPREV RCT (n=809; coronary heart disease patients; Mediterranean diet vs. low-fat diet) shows that Mediterranean diet (MedDiet) prevented MG elevation while the low-fat diet significantly increased circulating methylglyoxal (MG) over 5 years. In patients whose carotid IMT did not worsen (“non-increased IMT” subgroup), MedDiet produced upregulation of GloxI (GLO1) and AGER1 mRNA expression, increased sRAGE, and was associated with greater IMT regression. This is the longest-duration (5-year) RCT demonstrating dietary modulation of the glyoxalase system in humans, providing high-quality human evidence that MedDiet-associated polyphenols and dietary pattern differences can maintain GLO1 transcriptional activity and MG homeostasis in CHD patients.

Background and rationale

The CORDIOPREV trial (CORonary Diet Intervention with Olive oil and cardiovascular PREVention) was a parallel-arm RCT enrolling 1002 patients with established CHD at Hospital Reina Sofia (Córdoba, Spain). The main trial compared a Mediterranean diet (20% fat from olive oil; ≥3 servings/week fish; ≥2 servings/day legumes) versus a low-fat diet (<30% fat; <10% saturated) for cardiovascular outcomes. This sub-study examined whether dietary-induced differences in circulating AGE metabolism — specifically MG levels and expression of the MG-detoxifying enzyme GLO1 (GloxI) — mediated the atherosclerosis benefit of MedDiet as measured by carotid IMT.

Rationale: MG is the primary AGE precursor detoxified by the glyoxalase system (glo1 + GLO2 + GSH). GLO1 is Nrf2-regulated; dietary polyphenols (olive oil polyphenols, quercetin, resveratrol in red wine) activate Nrf2 and may upregulate GLO1 transcription — a mechanism by which MedDiet could reduce MG-AGE-driven vascular damage. AGER1 (Advanced Glycation End products receptor 1; also called DDOST or OST48) is a protective receptor for AGEs that promotes AGE clearance; sRAGE is the soluble form of RAGE, which acts as a decoy receptor, reducing pro-inflammatory RAGE signaling.

Study design

• Design: Pre-specified parallel-arm RCT sub-study within CORDIOPREV
• Primary trial NCT: CORDIOPREV (NCT00924937; Córdoba, Spain)
• N for this sub-study: 809 patients (432 MedDiet; 377 low-fat) who completed both baseline and 5-year carotid ultrasound assessments; 193/1002 excluded in two stages: 63 did not complete baseline carotid study, then 130 of the remaining 939 did not complete the 5-year follow-up (reasons: disapproval of technique, refusal, death, withdrawal)
• Duration: 5 years
• Participants: Established CHD patients; mean age and demographics per the CORDIOPREV main cohort (not re-specified in this sub-study)
• Dietary interventions: Mediterranean diet (olive oil enriched; ≥3 fish servings/week; ≥2 legumes/day; ≥2 fruits/day; ≥3 vegetables/day) vs. low-fat control (<30% fat; <10% saturated fat)
• Outcome measures:
  1. Serum methylglyoxal (MG) levels — measured by competitive ELISA (OxiSelect Methylglyoxal Competitive ELISA Kit, Cell Biolabs) at baseline and 5 years; not LC-MS/MS (ELISA measures protein-adducted MG-H1 imidazolone as a stable protein/peptide-associated form, inter-assay CV 5.8%, intra-assay CV 5.1%)
  2. GloxI (GLO1) mRNA expression — qRT-PCR with validated primer (qHsaCID0011227)
  3. AGER1 mRNA expression — qRT-PCR
  4. Plasma sRAGE levels — ELISA
  5. Reduced glutathione (GSH) — measured as glyoxalase system cofactor surrogate
  6. Intima-media thickness of common carotid arteries (IMT-CC) — bilateral carotid ultrasound

Results

MG levels over 5 years

The primary MG finding was a significant between-arm divergence in MG over the 5-year period ¹:

Arm	Baseline MG (µg/mL)	Change at 5 yr	p (within-arm)
Low-fat diet	3.05 ± 0.08	Δ+0.44 ± 0.04	p = 0.029 (increase)
Mediterranean diet	3.18 ± 0.07	Δ−0.03 ± 0.05	ns (stable)
Between-arm difference	—	—	p < 0.001

Note: no baseline difference between arms (p=0.380; Table 2 in paper). MG measured by competitive ELISA (OxiSelect Methylglyoxal Competitive ELISA Kit, Cell Biolabs), not LC-MS/MS. The ELISA detects stable protein/peptide-associated MG-H1 (imidazolone adducts) rather than free MG; values reflect protein-adducted MG burden.

Low-fat diet significantly increased circulating MG over 5 years; MedDiet maintained stable MG levels. The divergence between arms was statistically significant at p < 0.001.

Mechanistic interpretation: The low-fat diet likely increases glycolytic flux (higher carbohydrate proportion of energy intake) without compensating Nrf2/GLO1 upregulation, resulting in net MG accumulation. MedDiet polyphenols (olive oil oleuropein/hydroxytyrosol, resveratrol from wine, quercetin from vegetables) activate Nrf2 → maintain GLO1 transcription → sustaining MG detoxification capacity.

GloxI (GLO1) and AGER1 expression — subgroup analysis

Gene expression data were analyzed in the subgroup of patients with non-increased IMT-CC (i.e., patients whose atherosclerosis burden did not progress over 5 years; n=408: 166 low-fat + 242 Mediterranean diet):

Gene	MedDiet Δ	Low-fat Δ	p (each arm)
GloxI (GLO1 mRNA)	Δ+0.08 ± 0.07	Δ−0.10 ± 0.07	p < 0.05 (each arm)
AGER1 mRNA	Δ+0.22 ± 0.04	Δ−0.19 ± 0.09	p < 0.05 (each arm)

In non-worsening patients, MedDiet upregulated both GLO1 and AGER1 transcription while low-fat diet downregulated both. This bidirectional divergence supports the interpretation that MedDiet actively maintains the cellular AGE-defense apparatus — not merely slowing a decline shared by both arms.

Caution: These are mRNA levels from PBMCs specifically (peripheral blood mononuclear cells isolated from blood, RNA extracted with RNeasy kit, quantified by real-time qPCR on the Bio-Rad CFX Maestro platform; primer qHsaCID0011227 for GloxI; housekeeping genes GAPDH and ACTB). Whether GLO1 protein or enzymatic activity changes in parallel, and whether tissue-level GLO1 in the vascular wall matches PBMC expression, is not established by this study. The PBMC source is important context: PBMC gene expression may not reflect vascular smooth muscle or endothelial cell responses.

sRAGE — non-increased IMT subgroup

In the non-increased IMT-CC subgroup:

• MedDiet: sRAGE increased by 187.3 ± 22 pg/mL (p = 0.004)
• Low-fat: sRAGE decreased by −5.44 ± 12 pg/mL (ns)

Rising sRAGE acts as a decoy receptor competing with cell-surface RAGE for AGE ligands — a protective mechanism reducing RAGE-driven NF-κB activation and chronic-inflammation.

IMT-CC outcomes

In the non-increased IMT-CC subgroup:

• MedDiet: mean IMT-CC change = −0.101 ± 0.011 mm (regression; p = 0.009 vs. low-fat)
• Low-fat: mean IMT-CC change = −0.056 ± 0.002 mm

In the increased IMT-CC subgroup, both diets showed IMT increases with no significant difference between arms — suggesting AGE-metabolism modulation is selectively beneficial in those who respond metabolically.

Mechanistic model

MedDiet polyphenols (olive oil; resveratrol; quercetin; luteolin)
    ↓ Nrf2 activation
    ↓ GLO1 (GloxI) transcriptional upregulation
    ↓ Sustained MG detoxification capacity
    ↓ Stable circulating MG → reduced MG-H1/CEL/MOLD formation
    ↓ Upregulation of AGER1 (protective AGE clearance receptor)
    ↓ ↑ sRAGE (decoy receptor; reduces pro-inflammatory RAGE signaling)
    ↓ Reduced chronic vascular inflammation
    ↓ Reduced carotid IMT progression (−0.101 mm vs −0.056 mm over 5 yr)

This model is consistent with the directionality of all measured endpoints but remains associative — the sub-study design cannot fully establish causality between GLO1 induction and IMT benefit, only co-occurrence within MedDiet arm.

Context within GLO1 pharmacology landscape

Prior human evidence for GLO1 induction was limited to an 8-week crossover RCT of isolated compounds (trans-resveratrol + hesperetin; Rabbani 2021; see glo1 § GLO1 inducers). CORDIOPREV 2024 extends this in two key dimensions:

1. Duration: 5 years vs. 8 weeks
2. Intervention complexity: Whole dietary pattern vs. isolated polyphenol pair

CORDIOPREV establishes that the GLO1 induction signal is durable over 5 years under a realistic dietary intervention — the most clinically relevant translation context. However, unlike Rabbani 2021, CORDIOPREV did not directly measure GLO1 enzymatic activity (only mRNA), and the whole-diet design precludes attribution to specific polyphenol components.

For prior human evidence context, see glo1 § GLO1 inducers and pharmacology.

Limitations

• Sub-study design. The MG/GLO1 endpoints were pre-specified in the sub-study protocol but the CORDIOPREV main trial was powered for cardiovascular events, not GLO1 expression. Statistical power for the gene expression analyses may be limited.
• mRNA only — no protein or activity data. Blood GloxI mRNA is a transcriptional surrogate; whether GLO1 enzymatic activity or protein in vascular tissue changed in parallel is not known. Blood mRNA may not reflect tissue-level GLO1 in the vascular wall.
• The non-increased IMT subgroup is retrospective. Patients were stratified post-hoc into “increased” vs “non-increased” IMT groups. This introduces survivor/responder bias — the beneficial GLO1 + AGER1 gene expression changes may characterize metabolic responders to MedDiet regardless of mechanism.
• No dietary compliance biomarker. Olive oil and fish adherence were self-reported; objective polyphenol exposure (urinary polyphenol metabolites) was not a reported endpoint in this sub-study.
• CHD population. Participants had established coronary heart disease (not a healthy aging cohort). Generalizability to normoglycemic aging populations is uncertain.
• MG measured in serum, not tissue. Circulating MG may not reflect intracellular or tissue-compartment MG in arteries, kidneys, or brain.

needs-replication — No independent 5-year dietary RCT with GLO1 gene expression as an endpoint has been published. This is a single-trial sub-study.

Supersession check (2026-05-20): A 2025 systematic review (Sánchez-González et al., Nutrition Reviews 2025; doi:10.1093/nutrit/nuae045; PMID 38719207; n=2935 from 6 trials) found that Mediterranean diet adherence reduces serum MG, CML, and improves GLO1 activity across multiple trial designs — broadly consistent with CORDIOPREV 2024. The systematic review does not include CORDIOPREV (published October 2024, likely after the review’s search cutoff) and does not supersede it. No meta-analysis or large RCT reaching a different conclusion was identified as of 2026-05-20. long-term-unknown — Whether the GLO1 upregulation and MG stabilization over 5 years translates to hard cardiovascular outcomes (MI, stroke) or aging endpoints (biological age acceleration) is not addressed.

Extrapolation

Dimension	Status
Pathway conserved in humans?	yes — glyoxalase system and Nrf2/GLO1 transcriptional regulation fully conserved in humans
Phenotype conserved in humans?	yes — this IS a human study (CHD patients)
Replicated in humans?	partial — first 5-year RCT showing dietary GLO1 modulation; no independent replication; consistent with shorter-term RCT data (Rabbani 2021)

See also

• glo1 — the principal MG-detoxifying enzyme; GLO1 inducers section
• methylglyoxal — substrate; circulating MG as a modifiable biomarker
• mediterranean-diet — the dietary intervention; broader cardiovascular evidence base
• advanced-glycation-end-products — upstream AGE biology
• chronic-inflammation — downstream hallmark; RAGE-NF-κB axis

Footnotes

Footnotes

1. cordioprev-2024-meddiet-glo1 · n=809 (432 MedDiet; 377 low-fat diet; from original n=1002 CORDIOPREV cohort) · rct · 5-year parallel-arm sub-study within CORDIOPREV trial · model: established CHD patients (Córdoba, Spain) · MG measured by competitive ELISA (OxiSelect Kit; not LC-MS/MS); baseline MG: low-fat 3.05±0.08, MedDiet 3.18±0.07 µg/mL (ns between arms) · ΔMG: MedDiet −0.03±0.05 µg/mL (ns) vs low-fat +0.44±0.04 µg/mL (p=0.029); p<0.001 between arms · GloxI mRNA (PBMCs): Δ+0.08±0.07 AU (MedDiet) vs Δ−0.10±0.03 AU (low-fat) in non-worsening IMT subgroup (n=408; p<0.05 each arm) · AGER1 mRNA: Δ+0.22±0.04 (MedDiet) vs Δ−0.19±0.09 (low-fat; p<0.05 each) · sRAGE: +187.3±22 pg/mL MedDiet vs −5.44±12 pg/mL low-fat (p=0.004 MedDiet) · IMT-CC: −0.101±0.011 mm (MedDiet) vs −0.056±0.002 mm (low-fat; p=0.009 between arms) in non-worsening subgroup · full PDF verified 2026-05-20 (doi:10.1186/s12933-024-02451-4) ↩