Introduction: Arteries and veins are lined by endothelial cells that play a critical role in regulating blood flow. Despite being exposed to oxygen-rich blood, endothelial cells are characterised as predominantly glycolytic, with estimates suggesting that 85% of their ATP derives from glycolysis rather than mitochondrial oxidative phosphorylation. This paradigm has shaped vascular research for decades, with mitochondrial contributions interpreted through non-energetic roles – such as calcium buffering and reactive oxygen species signalling – rather than ATP production. However, we previously demonstrated that mitochondrial ATP production is required for endothelial calcium signalling and vasodilation in intact mesenteric arteries (Wilson et al., Function, 2022). If mitochondria contribute only ~15% of endothelial ATP, why does mitochondrial dysfunction have such catastrophic consequences for endothelial function? We aimed to resolve this paradox by systematically evaluating the contribution of OXPHOS to endothelial ATP production and determining its relevance to cardiovascular disease.

Methods: We performed a systematic review and meta-analysis of studies using the Seahorse Real-Time ATP Rate Assay to simultaneously measure glycolytic and oxidative ATP production rates in endothelial cells. Quantitative meta-analysis used random-effects models with the DerSimonian-Laird method to pool estimates of the ratio of oxidative to glycolytic ATP production across 39 datasets from 31 studies spanning 17 endothelial cell models. Patient-level differential expression analysis of single-nucleus RNA-seq data (GSE183852; 37,895 cardiac endothelial cells from 13 failing and 25 non-failing hearts) was performed to determine whether endothelial metabolic pathways are disrupted in heart failure.

Results: Meta-analysis revealed that OXPHOS contributes approximately 59% of basal endothelial ATP versus 41% from glycolysis (pooled ln ratio = 0.36, 95% CI: 0.21–0.52, P<0.0001; ratio = 1.43, 95% CI: 1.23–1.68) – a four-fold increase over the prevailing ~15% estimate. This finding was robust across species, vascular beds, and endothelial cell models, confirmed by leave-one-out sensitivity analysis (OXPHOS contribution 57–60% across all iterations; I² = 98.9%). In human heart failure, patient-level differential expression analysis revealed comprehensive energetic remodelling in cardiac endothelial cells: 8 of 12 core metabolic pathways were significantly downregulated, including all ATP-generating pathways — glycolysis (−41%), mitochondrial transport (−43%), TCA cycle (−28%), OXPHOS (−26%), and fatty acid oxidation (−26%). Strikingly, glucose transport was unchanged despite profound suppression of all downstream glucose-dependent ATP production pathways. Analysis of electron transport chain gene expression (92 genes; per-donor log-normalised mean) revealed a progressive monotonic decline across normotensive controls (0.367 ± 0.09, n=12), hypertensive controls (0.33 ± 0.08, n=13; −13.5% vs normotensive), and hypertensive heart failure patients (0.25 ± 0.08, n=5; −37.6% vs normotensive, p=0.02, Welch’s t-test).

Conclusions: These findings overturn the prevailing model of endothelial glycolytic predominance and resolve a long-standing paradox in vascular biology. Oxidative phosphorylation, and not glycolysis, is the primary source of endothelial ATP and is functionally required for endothelial calcium signalling and vasodilation. In cardiovascular disease, coordinated collapse of all ATP-generating pathways – without glycolytic compensation – characterises endothelial energetic failure in heart failure. Strikingly, mitochondrial bioenergetic deterioration begins at the hypertensive stage, before overt heart failure, identifying endothelial mitochondrial metabolism as an early and progressive therapeutic target in cardiovascular disease.