Reduced ATP-to-phosphocreatine ratios in neuropsychiatric post-COVID condition: Evidence from 31P magnetic resonance spectroscopy, 2025, Schilling+

[Nightsong]

Abstract:​

Background​

Post-COVID condition (PCCo) affects 5-10% of individuals following SARS-CoV-2 infection, with cognitive disturbances being a major feature. Central hypotheses regarding its pathophysiology include disturbed cell energy metabolism and oxidative stress pointing to mitochondrial dysfunction. However, brain energy metabolism remains unexplored.

Methods​

We investigated cerebral high-energy phosphate metabolism in 27 PCCo patients and 23 fully recovered controls using whole-brain 31P-MRSI at 3T. ATP/PCr ratios were quantified throughout the brain and analyzed with voxel-based and regional statistics including correlations with neuropsychological performance (Montreal Cognitive Assessment and Trail Making Test Part B). Statistical analysis employed voxel-wise comparisons with age as covariate, followed by region-of-interest analysis of cingulate cortex subdivisions.

Results​

PCCo patients showed a significant cluster of reduced ATP/PCr ratios centered on the cingulate cortex. Regional analysis revealed consistent reductions across anterior (ACC), mid- (MCC), and posterior (PCC) cingulate cortices. Lower ATP/PCr ratios in the ACC specifically correlated with poorer cognitive performance. Exploratory analyses revealed a trend toward higher intracellular pH in the MCC with significant negative correlation between pH and ATP/PCr observed only in patients, suggesting disease-specific alterations in pH regulation and bioenergetic homeostasis. Subgroup analysis showed similar metabolic alterations in PCCo patients meeting ME/CFS criteria.

Conclusions​

Our study provides first in vivo evidence of impaired brain energy metabolism in PCCo, with anterior cingulate dysfunction directly linked to cognitive impairment. The observed pH–ATP/PCr relationship suggests broader disruption of cellular bioenergetic regulation. These findings support mitochondrial dysfunction as a key pathophysiological mechanism and may inform targeted therapeutic strategies.

Link | PDF

 

[Eleanor]

Slow phosphocreatine recovery was mentioned in the WASF3 study https://www.s4me.info/threads/wasf3...me-2023-hwang-et-al.34776/page-13#post-583206

Wonder if anyone's looking at this in larger/better identified numbers of patients.

 

[V.R.T.]

Slow phosphocreatine recovery was mentioned in the WASF3 study https://www.s4me.info/threads/wasf3...me-2023-hwang-et-al.34776/page-13#post-583206

Wonder if anyone's looking at this in larger/better identified numbers of patients.

I hadn't made that connection- that's really interesting!

 

[SNT Gatchaman]

Slow phosphocreatine recovery was mentioned in the WASF3 study […] Wonder if anyone's looking at this in larger/better identified numbers of patients.

An apparatus for performing submaximal foot exercise to deplete phosphocreatine (dorsiflexing the foot against 30% of the maximum weight lifted by each subject) and to measure its recovery kinetics in the tibialis anterior muscle by 31P-MRS was used.

Measuring the regeneration of phosphocreatine (PCr) after its utilization by exercise in skeletal muscle using 31 P-MRS can noninvasively assess mitochondrial ATP synthesis capacity. In this test, the subjects perform a brief exercise using their feet while the 31 P-PCr spectra of their leg muscle are obtained. The PCr recovery time constant (Tc) is calculated by fitting the postexercise PCr amplitudes to a monoexponential recovery curve. The experiment is repeated after a 25-min rest following the first scan. Using this technique, the PCr Tc in the tibialis anterior muscle of the patient (S1) was found to be markedly prolonged at 80 s compared with reported values of 37 to 40 s in healthy volunteers.

Demonstrated in similarly exercising calf muscles in 1H and 31P MR Spectroscopy to Assess Muscle Mitochondrial Dysfunction in Long COVID (2024)

MR spectroscopy revealed a higher phosphocreatine recovery rate time constant (92.5 seconds ± 35.3 vs 51.9 seconds ± 31.9 [P < .001]; mean difference, 40.6 seconds [P ≤ .001]) and lower mitochondrial capacity (mean difference, 0.16 mmol/L per second; P = .008) in participants with PCC.

 

[V.R.T.]

Demonstrated in similarly exercising calf muscles in 1H and 31P MR Spectroscopy to Assess Muscle Mitochondrial Dysfunction in Long COVID (2024)

So that's three papers showing similar findings? And in different parts of the body?

 

[SNT Gatchaman]

Excessive Intracellular Acidosis Of Skeletal Muscle On Exercise In A Patient With A Post-Viral Exhaustion/Fatigue Syndrome: A 31P Nuclear Magnetic Resonance Study (1984)

The time course of recovery of [phosphocreatine] after exercise is shown in fig 2. Although this was somewhat slower than that observed in healthy controls after the same exercise protocol, recovery of calculated free [ADP] after exercise was not abnormally slow.

The slow phosphocreatine recovery suggests an aerobic defect, but intracellular pH fell to very low values on two of the four exercise tests and this in itself has been shown to delay phosphocreatine recovery.

Impaired oxygen delivery to muscle in chronic fatigue syndrome (1999)

we found that CFS patients had a moderate reduction in oxidative capacity, as measured by the rate of recovery of phosphocreatine after submaximal exercise

Negative results in this Fukuda criteria study though.

Muscle metabolism with blood flow restriction in chronic fatigue syndrome (2004)

Muscle metabolism was measured in the medial gastrocnemius muscle with 31 P-magnetic resonance spectroscopy. Muscle oxygen saturation and blood volume were measured using near-infrared spectroscopy. CFS and controls were not different in hyperemic blood flow or phosphocreatine recovery rate.

 

[SNT Gatchaman]

Selected quotes from introduction —

³¹Phosphorus magnetic resonance spectroscopic imaging (³¹P MRSI) enables in vivo investigation of cellular energy metabolism in the brain by quantification of the adenosine triphosphate (ATP) to phosphocreatine (PCr) ratio. In addition to high-energy phosphate compounds, ³¹P MRSI provides access to intracellular pH estimates, which are tightly coupled to mitochondrial proton gradients and ATP production.

The primary aim of this study was the in vivo investigation of changes in cerebral energy metabolism in patients with PCCo compared to completely recovered control subjects using whole-brain ³¹P MRSI. Further, we investigated the association of cerebral energy metabolism with measures of cognitive performance and disease severity, and explored potential alterations in intracellular pH regulation as an additional indicator of disrupted mitochondrial function.

We hypothesized that, compared to fully recovered control subjects, patients with PCCo would show reduced cerebral ATP levels, as indicated by lower ATP/PCr ratios, altered intracellular pH, and a relationship between regional ATP/PCr ratios in cognitively relevant brain areas and performance on neuropsychological assessments.

Noting in limitations discussion —

Another limitation might be that we could not include any 'COVID-naïve' subjects. This is due to the fact that at the time of the study, there were insufficient numbers of non-infected individuals. However, since we were interested in the specific changes related to PCCo and not in the correlates of the infection itself, our control group of infected recoverees is the more relevant one for this research question.

 

[SNT Gatchaman]

Selected quotes from methods —

The remaining group consisted of 27 patients with PCCo (mean age 41.19 years, 16 female, 11 male) and 23 completely recovered control subjects (mean age 41.70 years, 15 female, 8 male). All 27 patients met the WHO criteria for PCCo.

16 patients additionally met the Canadian Consensus Criteria for myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) as part of the disease spectrum of PCCo (mean age 43.81 years, 11 female, 5 male).

Age, sex, and body mass index (BMI) were included as covariates in all statistical models to account for known demographic and metabolic influences on cerebral energy metabolism.

 

[SNT Gatchaman]

Selected quotes from results —

Whole brain statistical analysis revealed a large significant cluster of reduced ATP/PCr concentration in PCCo patients […] centered on the cingulate cortex. The cluster was significant at cluster level with a cutoff of p<0.001

Given the functional heterogeneity of the cingulate cortex, with distinct roles in cognitive control (anterior cingulate cortex, ACC), motor function and conflict monitoring (midcingulate cortex, MCC), and memory and visuospatial processing (posterior cingulate cortex, PCC), we performed additional region-of-interest analyses

we observed a significant negative association between ATP/PCr and age, as well as higher ATP/PCr ratios in females […] consistent with previously reported age- and sex-related differences in brain energy metabolism. No significant associations between ATP/PCr and body mass index were observed. […] analyses did not reveal a significant interaction between sex and pathology, indicating that the main group effects reported in the manuscript are not driven by sex differences.

Group comparisons revealed consistently reduced ATP/PCr ratios in PCCo patients across all three cingulate subregions compared to controls (ACC: p < 0.040; MCC: p < 0.001; PCC: p < 0.004). Further, in the patient group correlations with neuropsychological performance were specifically observed in the ACC. Lower ATP/PCr ratios in the ACC correlated significantly with poorer performance in the Montreal Cognitive Assessment (MoCA) (p < 0.013) and longer completion times in the Trail Making Test part B (TMT-B) (p < 0.037), indicating an association between energy metabolism and cognitive dysfunction in PCCo patients in the ACC – a key component within the significant whole-brain cluster and a region crucial for cognitive processing.

When using Spearman rank correlation, TMT-B association was no longer significant (p = 0.199). Exclusion of the two patients with the longest TMT-B completion times likewise resulted in a loss of statistical significance.

While group differences in intracellular pH did neither reach significance for whole brain nor regional analysis, the patient group exhibited a trend towards higher pH values in the medial cingulate cortex (p = 0.07). Notably, a significant negative correlation between pH and ATP/PCr was observed in patients (p = 0.02), but not in controls (p = 0.5).

In subgroup analyses comparing PCCo patients with and without ME/CFS, no significant group differences in ATP/PCr ratios were found, suggesting similar metabolic alterations across these clinical subgroups.

 

[SNT Gatchaman]

Selected quotes from discussion —

Our 31P MRSI findings provide robust in vivo evidence for impaired cellular energy metabolism in the brains of patients with PCCo. By demonstrating significantly reduced ATP/PCr ratios in a large cluster centered around the cingulate gyrus, we present the first direct evidence of deficient high-energy phosphate metabolism in the central nervous system of PCCo patients.

The anatomical specificity of our findings is particularly notable. Our whole-brain analysis identified a large cluster of metabolic dysfunction encompassing the cingulate gyrus. […] targeted analysis revealed that lower ATP/PCr ratios in the ACC correlated with poorer performance on neurocognitive tests, suggesting a direct link between metabolic dysfunction in this key brain region and cognitive impairment

The observed reduction in ATP/PCr ratio can, in principle, arise from either a decrease in ATP availability or an increase in PCr levels. […] we normalized both metabolites to phosphorus mono- and diesters (PEstr = PC + PE + GPC + GPE). […] we observed a spatially extended reduction of ATP/PEstr in PCCo patients that showed substantial overlap with the significant ATP/PCr cluster. In contrast, increases in PCr/PEstr were limited to small, lateralized regions and did not show consistent or widespread effects, suggesting that the reduced ATP/PCr ratios in our data are caused by a reduction in ATP rather than by elevated PCr levels.

Under physiological conditions, cellular ATP concentrations are tightly buffered by the PCr–creatine kinase (CK) system, such that increases in ATP consumption are initially compensated by rapid regeneration of ATP from PCr via the CK reaction, resulting in a preferential decline of PCr while ATP levels are maintained. Consequently, a reduction in ATP in the presence of relatively preserved PCr is difficult to reconcile with an isolated impairment of oxidative phosphorylation and instead points to a limitation of CK-mediated energy buffering or intracellular energy transfer. Such a scenario may arise from impaired CK activity or disrupted functional coupling between mitochondrial ATP production and cytosolic ATP utilization (46), potentially in combination with a mild to moderate reduction in mitochondrial oxidative phosphorylation as the primary production site of ATP (47).

The ACC, which showed the most significant metabolic-behavioral correlations, plays a pivotal role in allocating cognitive control resources and integrating cognitive, sensory, and emotional information processing

The implications of our findings extend beyond PCCo to other conditions characterized by fatigue and cognitive dysfunction, such as myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Interestingly, our subgroup analysis comparing PCCo patients with and without ME/CFS criteria did not reveal significant differences, suggesting that the metabolic alterations we observed may be common to both conditions, potentially reflecting shared pathophysiological mechanisms. This observation aligns with the clinical overlap between these conditions

 

[SNT Gatchaman]

With regard to ME/CFS, PEM with sensory hypersensitive, fatiguability associated with physical, cognitive and emotional activity, this sentence stands out —

The ACC, which showed the most significant metabolic-behavioral correlations, plays a pivotal role in allocating cognitive control resources and integrating cognitive, sensory, and emotional information processing