Mitochondria and Long COVID: Why Cells Stop Making Energy After Infection

The morning after the fever finally broke, Sarah felt something was wrong in a way she could not quite name. She was no longer sick in the conventional sense: the cough had faded, the breathlessness had eased, and her test had come back negative. But the tiredness that had settled into her body during the acute phase of COVID-19 was not lifting the way ordinary tiredness lifts. A short walk to the kitchen left her needing to sit down. Reading a paragraph twice still left it opaque. Sleep brought no refreshment. Weeks became months, and the condition that doctors began calling long COVID became a permanent fixture of her life.

Sarah's experience is shared, in its broad outlines, by an estimated 65 million people worldwide as of 2026, according to figures compiled by the Long COVID Alliance and corroborated by successive waves of prevalence research. That number represents roughly one in ten people who contracted COVID-19 during the pandemic, and it includes individuals who had severe initial illness and those who experienced only mild acute symptoms. Long COVID cuts across age groups, fitness levels, and vaccination status in ways that confounded early expectations. What links so many of these cases, researchers are now beginning to understand, is not the virus still circulating in the body but the wreckage it left behind in one of the most fundamental structures in human biology: the mitochondria.

What Post-Viral Fatigue Actually Feels Like

Before examining the biology, it is worth dwelling on the phenomenology, because long COVID fatigue is often misunderstood and therefore under-appreciated by clinicians and the public alike. This is not the fatigue of a person who slept badly, nor the tiredness that follows a hard week of work and dissolves over a restful weekend. Long COVID fatigue is categorically different, and the distinction matters for understanding its underlying mechanism.

Patients consistently describe a total-body exhaustion that is disproportionate to any effort expended. Climbing a flight of stairs can produce a crash lasting hours or days. Concentrating on a document can trigger the same collapse. The fatigue is accompanied by what patients call brain fog: difficulty retrieving words, impaired short-term memory, slowed processing speed, and a pervasive sense of cognitive dimming that feels physical rather than psychological. Sleep, rather than restoring energy, often leaves patients feeling worse in the morning than the night before. And crucially, the harder a patient pushes, the harder they fall.

This last phenomenon has a clinical name: post-exertional malaise, or PEM. It is perhaps the most diagnostically distinctive feature of long COVID, and it is the feature that most directly points toward a fundamental problem with cellular energy production. In a healthy person, exercise produces a temporary energy deficit that the body rapidly compensates for by upregulating mitochondrial function, increasing oxygen delivery, and synthesising more ATP. In a long COVID patient with PEM, that compensatory response appears to be impaired or absent. The energy deficit triggered by exertion is not corrected; it compounds. The cellular machinery for making energy, it seems, is broken.

How SARS-CoV-2 Disrupts Mitochondrial Function

The Virus Targets the Cell's Power Plant Directly

To understand what SARS-CoV-2 does to mitochondria, it helps to recall what mitochondria do. These double-membraned organelles are responsible for the vast majority of cellular energy production through a process called oxidative phosphorylation. Inside the inner mitochondrial membrane, a series of protein complexes collectively called the electron transport chain passes electrons from nutrients through a cascade of reactions that ultimately drive the synthesis of adenosine triphosphate, ATP, the universal energy currency of biology. Complex I, the first and largest of these protein complexes, is the entry point for the entire chain and the site at which the most catastrophic damage from SARS-CoV-2 appears to occur.

SARS-CoV-2's non-structural proteins, particularly NSP12 (the viral RNA polymerase) and NSP7, have been shown to directly interfere with mitochondrial Complex I function. Research published in Nature Communications demonstrated that these viral proteins physically interact with Complex I subunits, impairing electron transfer and reducing the proton gradient across the inner membrane that drives ATP synthesis. The result is a cell that is consuming oxygen at a normal rate but converting far less of that oxygen into usable energy, generating instead a surplus of reactive oxygen species, the damaging byproducts of incomplete electron transfer.

ACE2 Receptors: The Mitochondria's Vulnerability

SARS-CoV-2 is well known to use the ACE2 receptor as its primary entry point into human cells, and most early discussion focused on ACE2 expression in lung epithelial cells. What became clear over subsequent years of research is that ACE2 is also expressed on the outer mitochondrial membrane in multiple cell types, including cardiomyocytes and skeletal muscle cells. This mitochondrial ACE2 expression provides the virus with a route not just into the cell but into the organelle itself.

Once inside, or closely associated with mitochondrial membranes, the virus triggers a cascade of events that compound the direct Complex I interference. Mitochondrial membrane potential, the electrical gradient across the inner membrane that drives ATP synthesis, becomes destabilised. Calcium handling, which mitochondria regulate in concert with the endoplasmic reticulum, becomes dysregulated in ways that impair muscle contraction and neuronal signalling. And mitochondrial dynamics, the constant cycle of fusion and fission by which mitochondria maintain their network and eliminate damaged components, shifts toward fragmentation: the healthy, interconnected mitochondrial networks seen under fluorescence microscopy in healthy cells are replaced by smaller, isolated, dysfunctional fragments.

The NAD+ Crisis and Viral Replication

Perhaps the most elegant mechanism through which SARS-CoV-2 sabotages mitochondria involves a molecule called nicotinamide adenine dinucleotide, or NAD+. NAD+ is the essential cofactor that feeds electrons into Complex I of the electron transport chain; without adequate NAD+, oxidative phosphorylation cannot proceed at full capacity. Viral replication, it turns out, is enormously NAD+-hungry: SARS-CoV-2 commandeers the cell's NAD+ pool to fuel the synthesis of its own genetic material, depleting the supply available to mitochondria. Viruses have also been shown to activate poly(ADP-ribose) polymerase (PARP) enzymes as part of the cellular stress and DNA-damage response triggered by infection, and PARP activation consumes NAD+ at a prodigious rate.

This NAD+ depletion creates a cellular energy crisis that persists long after the virus itself has been cleared. The enzymes responsible for NAD+ biosynthesis, particularly NAMPT (nicotinamide phosphoribosyltransferase), appear to remain suppressed in some long COVID patients, preventing the cell from restoring its NAD+ pool to pre-infection levels. The mitochondria are left operating in a state of substrate starvation: the machinery for making ATP is structurally impaired by Complex I damage and also starved of the cofactor it needs to run. This dual mechanism helps explain why long COVID fatigue can be so severe and so persistent. You can read more about the broader role of mitochondrial dysfunction in chronic disease to understand how this energy deficit plays out across multiple organ systems beyond COVID alone.

The ROS Storm: When Damaged Mitochondria Become Toxic

Reactive oxygen species, or ROS, are a normal byproduct of mitochondrial metabolism. Under healthy conditions, a small fraction of the electrons flowing through the electron transport chain escape to react with molecular oxygen, forming superoxide and other oxidising molecules. These are rapidly neutralised by antioxidant enzymes including superoxide dismutase and catalase, and in modest amounts, ROS even serve useful signalling functions within the cell. The problem in long COVID, and in mitochondrial dysfunction more broadly, is when this balance tips.

When the electron transport chain is disrupted, as it is by SARS-CoV-2's attack on Complex I, electrons that should move efficiently through the chain instead leak and react with oxygen at elevated rates. The resulting ROS overproduction overwhelms the cell's antioxidant defences, causing a state of oxidative stress. ROS at toxic concentrations damage mitochondrial DNA (which, unlike nuclear DNA, has limited repair mechanisms), oxidise the lipid membranes that mitochondria depend on for structural integrity, and inactivate the very electron transport chain proteins that are already compromised. The damage is self-amplifying: impaired mitochondria produce more ROS, which impair the mitochondria further.

The consequences extend far beyond the mitochondria themselves. Oxidative stress signals from damaged mitochondria activate inflammatory pathways, contributing to the low-grade chronic inflammation that characterises long COVID. ROS also damage the endothelium lining blood vessels, impairing vasodilation and reducing blood flow to tissues that already struggle to extract enough oxygen to feed their depleted mitochondria. In the brain, neuronal mitochondria subjected to oxidative stress produce the cognitive symptoms of brain fog through mechanisms that include synaptic dysfunction and impaired neurotransmitter synthesis. This intersection of mitochondrial damage, oxidative stress, and chronic inflammation is explored in greater depth in our article on reactive oxygen species, DNA damage, and biological ageing, which places these mechanisms in the context of the body's long-term response to cellular stress.

What the Research Shows: Mitochondrial Biomarkers in Long COVID Patients

Muscle Biopsies and Electron Microscopy Evidence

The claim that SARS-CoV-2 damages mitochondria is not speculative: it is supported by a converging body of direct observational evidence from patient samples. Skeletal muscle biopsies from long COVID patients with prominent fatigue and post-exertional malaise have revealed, under electron microscopy, mitochondria with abnormal morphology: swollen, with disrupted cristae (the inner membrane folds where ATP synthesis occurs) and evidence of membrane degradation. These structural abnormalities are accompanied by functional deficits measurable in vitro: mitochondria extracted from long COVID patient muscle cells produce significantly less ATP per unit of oxygen consumed than those from healthy controls, a finding consistent with uncoupling of the electron transport chain.

Studies measuring mitochondrial biogenesis, the process by which cells produce new mitochondria to replace damaged ones, have found it to be reduced in long COVID patients relative to both healthy controls and COVID-recovered individuals without ongoing symptoms. This matters because mitochondrial biogenesis is the cell's primary mechanism for recovering from mitochondrial damage. It is driven by a molecular switch called PGC-1 alpha (peroxisome proliferator-activated receptor gamma coactivator 1 alpha), whose activity appears to be suppressed in a subset of long COVID patients. Cells that cannot efficiently generate new, healthy mitochondria are cells that accumulate damage over time rather than clearing it.

Metabolomic Signatures

Metabolomics, the large-scale study of the small molecules produced and consumed by cellular metabolism, has provided some of the most compelling evidence for mitochondrial dysfunction in long COVID. Several research groups, including teams at Harvard Medical School and the University of Cambridge, have identified distinctive metabolic signatures in the blood of long COVID patients that reflect impaired mitochondrial function: elevated levels of lactate (consistent with a shift toward anaerobic metabolism), reduced levels of metabolites in the tricarboxylic acid cycle that mitochondria use to process nutrients, and depleted levels of NAD+ and its precursors.

A 2023 study published in Cell by researchers at Yale School of Medicine identified cortisol dysregulation, reactivation of latent Epstein-Barr virus, and serotonin depletion alongside mitochondrial metabolic abnormalities in long COVID patients, suggesting that the mitochondrial dysfunction is part of a broader systemic dysregulation rather than an isolated event. This systems-level perspective is important: targeting mitochondria alone may be insufficient if the inflammatory, hormonal, and viral reservoir dimensions of the condition are not also addressed.

Post-Exertional Malaise: Why Pushing Through Makes Things Worse

The instinct to push through fatigue, to "exercise your way back to health," is so deeply embedded in Western medical culture that it has caused serious harm to long COVID patients. Graded exercise therapy, a treatment that has some evidence base in certain fatigue conditions, was initially recommended by some clinicians for long COVID and ME/CFS alike. The results were, for many patients, disastrous. This is not a failure of willpower or motivation on the part of patients; it is a predictable consequence of the underlying biology.

Exercise in a healthy individual triggers a cascade of adaptive responses: mitochondria upregulate their biogenesis programme, muscle fibres increase their oxidative capacity, and the cardiovascular system becomes more efficient at delivering oxygen to working tissues. These are the processes by which fitness improves. But these adaptive responses depend on mitochondria that are capable of responding to the exercise signal, and in long COVID patients, they are not. When a patient with impaired mitochondrial function and depleted NAD+ attempts exercise, the cell cannot ramp up ATP production to meet the increased demand. Instead, it falls back on anaerobic glycolysis, which is far less efficient and produces lactate as a byproduct. Lactate accumulation contributes to muscle pain and a cascade of inflammatory signalling that exacerbates the mitochondrial dysfunction. This is the mechanism of post-exertional malaise: exercise that should stimulate repair instead amplifies damage.

The clinical implication is that pacing, not pushing, is the appropriate management strategy for long COVID patients with PEM. Pacing involves carefully matching activity level to the patient's available energy envelope, avoiding the exertion threshold at which PEM is triggered, and building in recovery periods that allow whatever mitochondrial repair capacity the patient retains to operate. It is a frustrating and difficult discipline for people accustomed to activity, but the evidence increasingly supports it as the approach least likely to cause harm while disease-modifying treatments are developed.

Emerging Treatments Targeting Mitochondria

Antioxidant Strategies: NAC and CoQ10

Given the centrality of oxidative stress to mitochondrial damage in long COVID, antioxidant interventions have attracted considerable research attention. N-acetylcysteine, or NAC, is a precursor to glutathione, the most abundant and important endogenous antioxidant in human cells. Glutathione is substantially depleted in long COVID patients, and restoring it via NAC supplementation theoretically addresses one of the key amplifying mechanisms of mitochondrial damage: the failure of antioxidant defences to contain the ROS storm generated by impaired Complex I.

Small clinical trials of NAC in long COVID have produced encouraging preliminary signals, with some patients reporting reduced fatigue and improved cognitive function over 8 to 12 weeks of treatment. Larger randomised controlled trials are underway to determine whether these effects are robust and reproducible at scale. Coenzyme Q10, a molecule that serves as an electron shuttle within the electron transport chain and also has antioxidant properties, is being studied in several trials for long COVID fatigue. Patients with genetic disorders causing CoQ10 deficiency present with symptoms closely resembling those of severe long COVID, providing a biological rationale for trialling CoQ10 in the acquired mitochondrial dysfunction of long COVID. Early results from European trials have been mixed, with some showing improvements in fatigue scores and others showing no significant effect over placebo, likely reflecting the heterogeneity of long COVID as a condition.

Methylene Blue and Electron Transport Support

Methylene blue is an old compound with a new rationale in long COVID. First synthesised in 1876 and used historically as a stain, an antimalarial, and a treatment for methaemoglobinaemia, methylene blue can act as an alternative electron carrier in the mitochondrial electron transport chain, bypassing damaged components of Complex I and allowing ATP synthesis to proceed through an alternative route. In animal models of viral infection, methylene blue has been shown to preserve mitochondrial membrane potential and reduce ROS production. Small open-label trials in long COVID patients with prominent neurological symptoms, including brain fog and cognitive slowing, have reported subjective improvement in cognitive function, though the absence of placebo controls limits interpretation. Phase II randomised trials are now recruiting.

Low-Dose Naltrexone and Neuroinflammation

Low-dose naltrexone (LDN), typically prescribed at 1.5 to 4.5 milligrams per day rather than the 50-milligram doses used in addiction medicine, has attracted a significant patient community reporting benefit for long COVID symptoms. Its proposed mechanism is distinct from direct mitochondrial support: at low doses, naltrexone is thought to modulate microglial cells in the brain, the resident immune cells whose chronic activation contributes to neuroinflammation and the cognitive symptoms of long COVID. By reducing the inflammatory environment in which neurons operate, LDN may indirectly ease the burden on neuronal mitochondria that are struggling to maintain function under concurrent oxidative and inflammatory stress. Observational data from long COVID patient registries consistently places LDN among the interventions most frequently reported as helpful, and a small randomised trial published in 2025 found statistically significant improvements in fatigue and brain fog scores. Larger trials are now funded and underway.

NAD+ Precursors and Metabolic Restoration

Given the documented NAD+ depletion in long COVID, supplementation with NAD+ precursors, particularly nicotinamide riboside (NR) and nicotinamide mononucleotide (NMN), has attracted both patient and researcher interest. These compounds enter the NAD+ biosynthesis pathway at different points and have been shown in healthy adults and certain disease models to substantially raise intracellular NAD+ levels. The hypothesis in long COVID is that restoring NAD+ availability will remove one of the rate-limiting constraints on mitochondrial function, allowing the electron transport chain to operate at higher capacity even in the presence of residual structural damage. Clinical trials are underway at several centres, though results are preliminary. The broader science of NAD+ and mitochondrial ageing, which provides context for these interventions, is part of the emerging understanding of how AI-assisted monitoring can personalise metabolic interventions for complex chronic conditions where one-size-fits-all dosing protocols are likely to be inadequate.

The Promise of AI-Assisted Monitoring in Long COVID

One of the most significant challenges in long COVID management is the condition's heterogeneity. Patients who present with the same headline diagnosis of long COVID can have profoundly different underlying biology: some have primarily mitochondrial dysfunction, others have viral reservoir persistence, others are dominated by autoimmune mechanisms, and many have combinations of all three. This heterogeneity makes population-level treatment trials difficult to interpret and generic treatment protocols unlikely to help all or even most patients. What is needed is the kind of precise, individual-level biological characterisation that, until recently, was confined to academic research settings.

Artificial intelligence is beginning to make that characterisation accessible in clinical settings. Machine learning models applied to metabolomic data, wearable sensor outputs, and standard clinical laboratory results can identify subgroups of long COVID patients likely to have predominant mitochondrial dysfunction versus other pathological mechanisms, guiding targeted treatment selection. Heart rate variability monitoring through consumer-grade wearables, analysed by AI, can detect the autonomic nervous system abnormalities that accompany mitochondrial dysfunction and provide a real-time readout of a patient's energy state, helping them pace more effectively and avoid the PEM threshold. Some research groups are developing multiparametric AI platforms that combine wearable biosensor data, patient-reported symptom logs, and laboratory biomarkers into a continuously updated model of each individual's physiological status, flagging deterioration trends before they become symptomatic crashes.

This vision of AI as a precision monitoring layer for long COVID patients mirrors the broader trajectory of AI in medicine: not replacing clinical judgment but extending its reach, making continuous surveillance and personalised adjustment possible for a condition that demands exactly that level of attention. For a condition affecting 65 million people, with a healthcare system that cannot come close to providing that attention through conventional means, the AI layer may prove as important as any pharmacological intervention.