Introduction: Metabolic signatures of heart failure (HF) phenotypes generally portray static, single time point differences between disease and control groups. Dynamic changes in metabolite in response to cues such as glucose loading are rarely determined in human HF studies.
Methods: This was a multi-centre study that involved HF patients with reduced ejection fraction (EF) and controls. All patients underwent a protocol that involved fasting overnight, followed by an early morning oral glucose drink (75 gram). Blood sampling at (Time 1) baseline (prior to glucose drink), (ii) (Time 2) 2 hours after glucose drink, (iii) (Time 3) 3 hours after baseline were performed. Targeted high mass spectrometry measured 65 circulating acyl-carnitines in serum (nanomolar).
Results: This analysis consisted of 19 subjects (mean age 53±10 years; 8 females (42%) (HF n=9 (all EF <50%); controls n=10). All subjects were non-diabetic, as confirmed by fasting glucose and glycated haemoglobin levels. At Time 1, there were significant differences in short chain [(C2: 10050±2056 vs 8165±1572, p=0.024); (C3: 662±199 vs 478±143, p=0.02)], long chain (C16-OH: 7±1.3 vs 6±1.4, p=0.036) and medium-and long chain di-carboxylates [(C8-DC: 47±14 vs 36±9, p=0.039; (C18-OH/C16-DC: 6±1.5 vs 5±1.2, p=0.038)] between HF and controls, demonstrating higher accumulation of these metabolites among HF patients compared to controls. Following glucose loading, metabolite levels decreased at Time 2 and Time 3, in both HF patients and controls. However, metabolite levels decreased more robustly among controls compared to HF patients after glucose loading. The magnitude of change from Time 1 to Time 3 (ratio of time 3: time 1) for short chain (C2: 0.6±0.11 vs 0.7±0.08, p=0.022), long chain (C16:2: 0.5±0.1 vs 0.6±0.1, p=0.036) and medium- chain di-carboxylates (C8-DC: 0.6±0.1 vs 0.7±0.1, p=0.018) was reduced for HF compared to controls.
Conclusions: We observed novel, dynamic differences in rates of metabolite clearance in HF patients. These pilot findings, observed in the absence of diabetes, suggest that a state of reduced metabolic flexibility is present among HF patients. These hypothesis-generating data warrant further investigation into potential modifiable metabolic pathways in HF.