Metabolic Failure, Mitochondrial Dysfunction, and Multiorgan Injury

Organ function requires oxygen and substrates to reach cells, metabolic pathways to transform them, and energy production to remain matched to cellular work. Critical illness can disturb each step in different combinations. Mitochondrial dysfunction is one important mechanism, but neither an elevated lactate nor an abnormal peripheral-cell assay establishes uniform failure of energy production throughout the body.

70.1 Metabolic Failure Can Arise at Several Levels

A cell can lack usable energy because delivery is inadequate, substrates cannot enter the relevant pathways, mitochondrial conversion is impaired, or demand exceeds available capacity. These mechanisms can coexist. Increasing supply may correct one while leaving another unchanged.

The distinction extends from circulation to intracellular organization. Adequate cardiac output does not guarantee uniform capillary delivery, while adequate local oxygen tension does not guarantee effective oxidative phosphorylation. Conversely, low oxygen consumption can reflect reduced demand rather than an inability to use oxygen.

Metabolic failure is therefore a mismatch between the energy and biochemical functions required and those that can be sustained. It is not defined by one oxygen value, one metabolite, or an isolated mitochondrial abnormality. Ion transport, contraction, synthesis, secretion, immune activity, and repair have different energetic requirements and reserves.

Brealey and colleagues studied muscle biopsies from 28 critically ill septic patients and nine surgical controls. Energetic and respiratory-chain abnormalities were associated with illness severity and outcome. The study supports a bioenergetic contribution in the sampled tissue, but its observational design does not establish that one defect caused every organ failure or that all tissues behaved identically. 1

Functional impairment also need not imply widespread irreversible cell death. Cells can reduce specialized work, alter substrate use, or reorganize metabolism while remaining viable. Such changes may conserve resources in some circumstances, but calling all organ dysfunction protective “hibernation” would overstate an incompletely established hypothesis.

At the other extreme, continuing ATP insufficiency can undermine membrane gradients, calcium control, structural integrity, and survival. The same organ can contain relatively preserved regions, metabolically altered cells, and areas of irreversible injury. Their proportions change with time and with the persistence of the initiating insult.

The physiological task is to locate the limitation and assess its consequences. Whole-body delivery, regional perfusion, cellular conversion, demand, and recovery capacity must be considered together rather than arranged as competing explanations.

70.2 ATP Content and ATP Production Are Different Quantities

Adenosine triphosphate, or ATP, transfers energy to many cellular processes. The measured ATP pool is an amount or concentration at a moment; production and utilization are rates. A preserved pool can coexist with very high turnover, and a reduced turnover can preserve the pool when demand falls similarly.

A simplified balance makes this distinction explicit:

The left side is the rate of change of the ATP amount in the defined sample. The terms on the right represent oxidative ATP production, glycolytic ATP production, and ATP utilization. All require matching units and normalization, such as nmol ATP/min/mg tissue protein.

For a constructed example, assume oxidative production of 80, glycolytic production of 20, and utilization of 100 nmol/min/mg protein. The pool is stable because 80 + 20 − 100 = 0. If utilization rises to 140 with production unchanged, the initial balance becomes −40 nmol/min/mg protein. Holding those artificial rates constant for two minutes gives a loss of 80 nmol/mg protein.

This is not a patient value or a safe duration before injury. Real cells change production, utilization, phosphocreatine buffering, and adenine-nucleotide handling. Continuing a linear calculation until ATP becomes zero would ignore those responses and the earlier consequences of energetic stress.

ATP hydrolysis also depends on the surrounding chemical state. ATP, adenosine diphosphate (ADP), inorganic phosphate, pH, and magnesium influence the usable free-energy change. Equal ATP concentrations need not provide identical capacity to drive work if those conditions differ.

The location of production matters as well. ATP-dependent processes are organized within cellular microdomains, and transfer systems connect production to sites of demand. A whole-tissue average can conceal local deficits or contributions from different cell types.

Thus, ATP concentration, respiratory rate, ATP-production capacity, and actual cellular work should remain distinct. Their relationships provide the explanation; none is a complete substitute for the others.

70.3 Oxidative Phosphorylation Requires an Intact Coupling System

Oxidative metabolism transfers electrons from reduced substrates through the respiratory chain to oxygen. Reduced nicotinamide adenine dinucleotide (NADH) supplies electrons through complex I, while succinate oxidation supplies electrons through complex II. Electron transfer proceeds through shared carriers and complexes III and IV, with oxygen reduced to water at the terminal step.

Complexes I, III, and IV support proton movement across the inner mitochondrial membrane. Complex II transfers electrons without directly pumping protons. The resulting electrochemical gradient combines electrical and concentration components and provides the driving force for ATP synthase, or complex V.

ATP production also requires ADP and inorganic phosphate, appropriate transport across mitochondrial membranes, and a sufficiently intact inner membrane. Electron transport alone is therefore not the entire process. Impaired substrate oxidation, respiratory complexes, membrane coupling, ATP synthase, or nucleotide transport can each limit the overall result.

Protons can return through pathways that do not generate ATP. Some leak is physiological, while increased leak or uncoupling can increase the oxygen required for a given ATP yield. Greater oxygen consumption can consequently accompany unchanged or reduced useful phosphorylation, depending on demand and coupling.

Conversely, low respiration does not necessarily mean the respiratory machinery is damaged. Limited substrate delivery, low ADP availability, or reduced cellular work can suppress oxygen consumption in an otherwise capable system. Distinguishing capacity from current activity requires controlled measurements and knowledge of the conditions.

Glycolysis supplies ATP independently of mitochondrial oxygen reduction, although its net yield per glucose is smaller than that of complete oxidation under suitable conditions. Its contribution differs greatly among tissues and states. Mature red cells lack mitochondria and depend on glycolysis, illustrating why a mitochondrial model cannot be imposed on every cell.

There is no single fixed ATP yield to infer from a measured oxygen flux across all patients. Substrate entry, proton leak, transport costs, and the assay conditions affect the relationship. Respiration is an important observation, but converting it into actual ATP supply requires additional assumptions or measurement.

70.4 Oxygen Delivery and Extraction Do Not Localize the Cellular Defect

The circulation's oxygen-delivery relationship and the Fick principle describe different quantities:

Q is blood flow, CaO₂ arterial oxygen content, and CvO₂ mixed venous oxygen content. With Q in L/min and contents in mL oxygen/dL, each expression requires multiplication by 10 dL/L to give mL oxygen/min. DO₂ is delivered oxygen, while VO₂ is net oxygen consumption under the assumptions of the measurement.

Assume an illustrative flow of 6 L/min, arterial content of 16 mL/dL, and mixed venous content of 12 mL/dL. Delivery is 6 × 16 × 10 = 960 mL/min. Consumption is 6 × (16 − 12) × 10 = 240 mL/min, giving an extraction fraction of 240 ÷ 960 = 25%.

Those global values do not reveal whether every organ receives an adequate share or whether extraction is uniform. Parallel regions with limited flow, rapid transit, or altered exchange can contribute to the final venous mixture. A relatively high venous oxygen content can coexist with regional hypoxia.

The same finding can also accompany reduced whole-body demand, for example after muscle work or thermogenic activity diminishes. Impaired cellular utilization is another possibility, but it cannot be uniquely diagnosed from the venous measurement. Sampling a central venous location adds further regional weighting rather than supplying the same quantity as mixed venous blood.

The calculation assumes sufficiently representative contents, flow measurement, and temporal conditions. Oxygen consumption estimated from a content difference is sensitive to measurement error, especially when the difference is small. It is not a direct assay of mitochondrial ATP production.

Increasing oxygen delivery may help when supply limits function, but cannot automatically correct defective substrate handling, coupling, or another cellular limitation. The relevant question is whether the intervention changes the limiting step and improves function, not simply whether a global delivery number rises.

70.5 Substrate Handling and Lactate Reflect More Than Oxygen Availability

Glucose, fatty acids, amino acids, ketones, and lactate can contribute to metabolism in different tissues. Their availability, cellular uptake, enzymatic entry, and hormonal regulation determine whether they support ATP production or another function. A high circulating substrate concentration does not prove effective intracellular use.

Pyruvate links glycolysis to several pathways. Pyruvate dehydrogenase supports conversion to acetyl-coenzyme A for oxidation, while lactate dehydrogenase interconverts pyruvate and lactate in association with cytosolic redox conditions. Lactate formation regenerates the oxidized form, NAD⁺, required to sustain glycolytic reactions; it is not restricted to cells without oxygen.

Stress signaling can increase glycolytic flux and cellular work while oxygen-dependent pathways remain active. In oxygenated rat skeletal muscle, James and colleagues linked epinephrine-stimulated lactate production to pump activity without significant changes in ATP content or oxygen consumption. This experimental mechanism does not quantify its contribution in human sepsis or exclude concurrent hypoperfusion and impaired disposal. 2

Lactate can also be oxidized elsewhere or used for glucose synthesis. Its blood concentration reflects production, use, distribution, and clearance. Hepatic and renal dysfunction, changes in perfusion, adrenergic exposure, and altered metabolic demand can all affect the result.

Consequently, a falling concentration can reflect reduced production, increased disposal, dilution, or combinations. A calculated percentage decline is not necessarily a direct measurement of clearance. Likewise, an elevated lactate-to-pyruvate ratio requires attention to sampling and the compartment studied; it is not a universal mitochondrial-failure test.

Critical illness also mobilizes endogenous substrates while insulin resistance and altered synthesis change their handling. Supplying more nutrient does not guarantee more useful energy or restored cell function if uptake, oxidation, or demand matching remains abnormal. Detailed endocrine and gastrointestinal regulation belongs to Chapters 65 and 69.

The useful interpretation connects substrate availability with flux and function. Blood metabolite concentrations provide clues, but their causal meaning depends on the pathways and organs responsible for generating and removing them.

70.6 Redox Signaling and Injury Depend on Location, Amount, and Timing

Mitochondrial electron transfer can generate reactive oxygen species, or ROS, alongside its main energy-converting function. Controlled redox signals participate in adaptation and host defense. Injury develops when production, localization, and removal create damaging exposure rather than simply because any ROS is present.

Oxidative and nitrosative modifications can affect membranes, proteins, and mitochondrial DNA. Some effects are reversible, while others damage components that require repair or removal. The measured abundance of an oxidized product does not identify the exact source or prove that all effects are harmful.

Brown and Cooper demonstrated reversible inhibition of respiration by nitric oxide in brain synaptosomes and isolated cytochrome oxidase. The interaction depended on oxygen conditions. This supports a biochemical mechanism for altered respiration without establishing a fixed nitric-oxide threshold or the dominant cause of organ failure in every patient with sepsis. 3

Ischemia and reperfusion provide a different temporal sequence. Metabolites can accumulate while oxygen supply is restricted, then participate in abrupt redox changes when flow returns. Restoring circulation remains essential even though the transition can also generate injury.

Chouchani and colleagues demonstrated a contribution from succinate accumulation and oxidation to mitochondrial ROS during reperfusion in experimental models. Rapid succinate oxidation can support reverse electron transport at complex I and ROS generation under those conditions. The findings establish a metabolically organized mechanism, rather than nonspecific “oxidative stress,” but do not prove that the same pathway dominates all sepsis or that experimental inhibition improves general ICU outcomes. 4

Calcium handling, membrane potential, substrate supply, and inflammatory signals interact with these processes. Loss of membrane integrity or abnormal permeability can disrupt coupling and contribute to cell death. A transient change in respiration, however, need not mean that such irreversible injury has occurred.

The therapeutic inference must therefore retain specificity. A mediator can be protective in one location or interval and harmful in another. Reducing a broad oxidative marker is not equivalent to restoring mitochondrial function, preserving host defense, or demonstrating clinical benefit.

70.7 Recovery Requires Quality Control and Rebuilding

Mitochondria continually change through protein turnover, remodeling, division, fusion, and selective removal. These processes help separate or repair damaged components and maintain a functional population. Their regulation determines recovery capacity as well as immediate performance.

Mitophagy refers to selective autophagic handling of mitochondria. Appropriate removal can prevent damaged organelles from persisting, while excessive loss without replacement can reduce energetic capacity. Impaired removal can also be harmful. Neither more nor less autophagy is automatically favorable under every condition.

Biogenesis involves coordinated production of mitochondrial components encoded by nuclear and mitochondrial genomes, together with membrane and organelle assembly. Increased expression of a biogenesis-related transcript is evidence of signaling, not proof that a complete new population of functional mitochondria has already appeared.

Carré and colleagues examined early muscle biopsies in 16 critically ill patients and surgical controls. Survivors showed early differences in biogenesis-related and antioxidant responses, alongside energetic and structural measurements. The observations support a relationship between recovery programs and outcome, but do not establish that forcing one transcriptional pathway would reproduce survival. 5

The need for rebuilding also explains why recovery can lag behind correction of circulation. A cell may receive adequate oxygen while respiratory proteins, membrane organization, or metabolic coordination remain abnormal. Conversely, apparent normalization of one enzyme measurement does not establish restored organ work or reserve.

Reduced specialized function can lower ATP demand during this interval, but the tradeoff may include weakness, reduced transport, impaired barrier function, or delayed repair. Whether a change is an adaptive reduction in work or a manifestation of injury must be established from its consequences rather than its description.

Persistent inflammation, substrate imbalance, immobility, and ongoing organ stress can alter both damage and replacement. The relevant trajectory is therefore the balance between injury, removal, repair, and the capacity required for function. A single early measurement cannot describe that evolving process.

70.8 Tissue and Time Determine the Meaning of Mitochondrial Findings

The heart, kidney, liver, brain, skeletal muscle, platelets, and immune cells have different work patterns, substrate preferences, and mitochondrial organization. They also encounter different perfusion and inflammatory environments. One circulating-cell result cannot be assumed to represent their shared state.

Protti and colleagues measured respiratory-chain biochemistry in platelets and skeletal muscle from 30 patients with septic shock and surgical controls. The platelet changes did not track the muscle measurements. This directly challenges substitution of an accessible blood-cell assay for the energetic state of another tissue. 6

Timing can change the direction of a comparison. Early cellular activation, later impairment, recovery, and selection of surviving cells can produce different observations. A cohort sampled only when stable enough for repeated measurement may not represent patients who deteriorate before sampling is completed.

Normalization is equally important. A respiratory rate per viable cell differs from a rate per tissue mass, total protein, mitochondrial protein, or an enzyme used as a marker of mitochondrial content. These denominators answer different questions. A lower rate per tissue mass could reflect fewer mitochondria, lower capacity per organelle, a different cell mixture, or combinations.

Citrate synthase activity is often used as a mitochondrial-content marker, but it remains a biological measurement rather than an immutable count of organelles. Changes in its own expression or activity can affect a ratio. Reporting a normalized enzyme activity without its denominator can conceal the physiological interpretation.

Morphology also requires qualification. Swelling, altered shape, or changed network organization can accompany injury or remodeling but do not quantify ATP production by themselves. Similarly, mitochondrial DNA abundance and respiratory transcript levels describe related components rather than actual organ-level energy flux.

An informative result therefore states tissue, cell identity, sampling time, preparation, measured function, and normalization. Apparent contradictions can arise because studies asked different questions, and agreement across one marker does not guarantee agreement across all levels of the system.

70.9 Respiratory Assays Measure Responses Under Defined Conditions

High-resolution respirometry measures oxygen consumption in a prepared sample. Intact cells retain many transport and regulatory processes, whereas permeabilized preparations permit more controlled substrate access. Standardized medium and patient plasma expose the cells to different environments, so their results need not reproduce the same in vivo state.

A common assay sequence measures basal respiration, respiration remaining after ATP-synthase inhibition, a maximum elicited by an uncoupler, and residual consumption after respiratory-chain inhibition. These are operational measurements requiring suitable correction and interpretation; the uncoupled maximum is not the patient's actual rate of useful ATP production.

The respiration reduced by ATP-synthase inhibition is often termed ATP-linked. The remaining corrected mitochondrial respiration is commonly interpreted as leak-associated. Their ratio can estimate coupling under the assay conditions, while the difference between maximum and basal rates describes an operational spare respiratory capacity.

For an illustrative corrected assay, choose basal respiration of 40, leak-associated respiration of 10, and uncoupled maximum of 70 pmol oxygen/s per million viable cells. ATP-linked respiration is 40 − 10 = 30, coupling is 30 ÷ 40 = 75%, and spare capacity is 70 − 40 = 30 in the same flux units.

In a second constructed state, basal respiration is 60, leak-associated respiration 30, and maximum 80. ATP-linked respiration remains 30, coupling falls to 50%, and spare capacity is 20. Higher basal oxygen consumption has not increased the ATP-linked component. Neither example directly measures ATP molecules produced or establishes a clinical target.

Moonen and colleagues' 2026 study explicitly used live-cell-normalized respiration in intact peripheral blood mononuclear cells, or PBMCs, under standardized conditions. Its design illustrates why medium, viability, cell composition, and the distinction between oxygen flux and ATP production belong to the interpretation. 7

Sample handling, drug carryover, oxygen conditions, substrate concentrations, and analytical choices add further constraints. A laboratory perturbation can reveal capacity or sensitivity without identifying the rate-limiting process inside the patient's organ. Assay detail improves understanding only when its physiological boundary is preserved.

70.10 Immune Metabolism and Injured Organelles Can Amplify Systemic Dysfunction

Immune activation changes the energy and biosynthetic work required from cells. Different populations can increase or redistribute metabolic activity while other tissues lose function. This helps explain why increased blood-cell respiration can coexist with multiorgan injury.

In Moonen and colleagues' single-center 2026 cohort, 25 septic patients had higher basal, ATP-linked, maximal, and leak-associated PBMC respiration than matched controls during the first ICU week. These findings challenge universal suppression in that sampled population. Small numbers, cell mixtures, medication effects, and exploratory outcome associations limit stronger conclusions. 7

A separate 2026 study of isolated lymphocytes found lower complex-linked respiration associated with later persistent inflammatory, immunosuppressive, and catabolic features in a small subgroup. Different populations and assay conditions prevent treating the findings as a simple contradiction or a validated prognostic test. 8

Mitochondrial injury can also generate signals. DNA and formylated peptides released from damaged mitochondria can interact with innate immune pathways. Zhang and colleagues combined human injury observations with cellular and experimental models to demonstrate inflammatory consequences of mitochondrial material. This supports a link between tissue damage and systemic signaling without proving that every circulating mitochondrial marker measures viable-cell energetic failure. 9

Organ interactions then amplify the disturbance through several routes. Cardiac dysfunction alters delivery, pulmonary dysfunction changes oxygenation, and renal or hepatic dysfunction changes acid–base conditions, metabolites, drug exposure, and clearance. Endothelial injury and microvascular thrombosis further modify tissue access to substrates.

These loops need not progress uniformly or in a fixed order. One organ's functional decline can alter another's workload or environment even without a shared initial mitochondrial lesion. Conversely, restoring one function can reduce stress elsewhere before every cellular marker normalizes.

Multiorgan injury is consequently a network phenomenon involving circulation, exchange, metabolism, signaling, and organ work. Mitochondria participate as energy-converting structures and sources or responders to signals, while remaining embedded in the broader physiological system.

70.11 A Biochemical Effect Requires Separate Outcome Validation

A plausible metabolic intervention may improve one measured pathway without correcting the dominant cause of illness. Increasing substrate oxidation, changing redox signaling, or raising a circulating nutrient can affect several systems simultaneously. The net result depends on timing, tissue, deficiency state, and competing physiological costs.

Stacpoole and colleagues randomized 252 adults with severe lactic acidosis to dichloroacetate or placebo. The intervention changed lactate and pH but did not improve hemodynamics or survival. This historical trial directly illustrates why a biochemical response cannot be assumed to represent recovery of the patient's metabolic system. 10

An antioxidant rationale has similar limits. LOVIT randomized 872 adults with sepsis receiving vasopressors to intravenous vitamin C or placebo. The intervention increased the primary composite risk of death or persistent organ dysfunction at 28 days. The mortality component alone did not establish the same level of statistical certainty. This was a clinical drug strategy, not a direct trial of selectively repairing mitochondrial ROS generation. 11

Neither result means that every metabolic target is ineffective or that a demonstrated nutrient deficiency is irrelevant. They show that mechanism, biochemical modification, and patient benefit require separate evidence. Trials in a particular disease or selected population cannot supply a universal mitochondrial prescription.

Physiological support likewise needs an identified limitation. Improving oxygen content can matter when it constrains delivery, while reducing unnecessary work can change demand. Those effects do not prove correction of intracellular conversion, and their tolerability depends on the surrounding circulation and organ state.

The appropriate endpoint is not simply normalization of lactate, a redox marker, or a peripheral respiratory rate. Functional recovery, organ support, adverse effects, and patient-centered outcomes test whether the intervention helped the integrated system.

Mechanistic studies remain essential because they identify plausible targets and explain heterogeneity. Their strongest clinical use is to sharpen the question being tested, while retaining the distinction between a pathway that can be changed and a strategy that has been shown to improve outcomes.

70.12 Integrating Supply, Demand, Conversion, and Recovery

Consider the illustrative circulation from Section 70.4: flow of 6 L/min, arterial oxygen content of 16 mL/dL, and oxygen consumption of 240 mL/min. Delivery is 960 mL/min and extraction 25%. Suppose arterial content later reaches 20 mL/dL at the same flow, while measured consumption falls to 200 mL/min.

Delivery is now 6 × 20 × 10 = 1200 mL/min. Extraction is 200 ÷ 1200 ≈ 16.7%. Under a steady-state Fick calculation, mixed venous content would be 20 − 200 ÷ (6 × 10) ≈ 16.7 mL/dL. These are constructed measurements, not recommended delivery or extraction targets.

The higher venous content and lower extraction have several possible explanations. Reduced work, changes in regional distribution, altered exchange, and impaired utilization could contribute. A lactate elevation would add evidence of altered production–disposal balance but would not select one of those explanations by itself.

If an accompanying PBMC assay showed increased respiration, it would describe the sampled cells under laboratory conditions. It would neither exclude impaired muscle function nor establish that increased immune-cell metabolism caused the organ dysfunction. A low muscle enzyme measurement would require its own tissue, normalization, and timing context.

The integrated assessment consequently asks where the limiting process resides and whether it is changing. Are oxygen and substrates reaching the relevant regions? Has demand changed? Is conversion or coupling impaired? Are injured components being cleared and rebuilt? The answers must be connected to organ function and trajectory rather than inferred from one global marker.

Several explanations can remain active simultaneously. Improved circulation may coexist with delayed cellular recovery, while inflammation maintains high immune-cell demand and reduced organ clearance alters metabolite concentrations. Physiological recovery is the evolution of those relationships, not necessarily simultaneous normalization of every measurement.

Metabolic failure brings the book's central relationships together: pressure and flow, exchange and content, transport and regulation, cellular work and energy conversion. Keeping their distinct quantities visible makes the patient's response more intelligible and prevents a compelling single mechanism from replacing the integrated physiology.

References

1. Brealey D, Brand M, Hargreaves I, et al. Association between mitochondrial dysfunction and severity and outcome of septic shock. Lancet. 2002;360:219–223. Original muscle-biopsy study.

2. James JH, Wagner KR, King JK, et al. Stimulation of both aerobic glycolysis and Na+-K+-ATPase activity in skeletal muscle by epinephrine or amylin. Am J Physiol. 1999;277:E176–E186. Original experimental study.

3. Brown GC, Cooper CE. Nanomolar concentrations of nitric oxide reversibly inhibit synaptosomal respiration by competing with oxygen at cytochrome oxidase. FEBS Lett. 1994;356:295–298. Original biochemical study.

4. Chouchani ET, Pell VR, Gaude E, et al. Ischaemic accumulation of succinate controls reperfusion injury through mitochondrial ROS. Nature. 2014;515:431–435. Original experimental study.

5. Carré JE, Orban JC, Re L, et al. Survival in critical illness is associated with early activation of mitochondrial biogenesis. Am J Respir Crit Care Med. 2010;182:745–751. Original muscle-biopsy study.

6. Protti A, Fortunato F, Caspani ML, et al. Mitochondrial changes in platelets are not related to those in skeletal muscle during human septic shock. PLoS One. 2014;9:e96205. Original tissue-comparison study.

7. Moonen HPFX, Slingerland-Boot R, de Jong JCBC, et al. Progression of peripheral blood mononuclear cell mitochondrial function during the early phase of sepsis in intensive care unit patients. Sci Rep. 2026;16:8828. Original repeated-measurement study.

8. Nedel W, Kopczynski de Carvalho A, Rodolphi MS, et al. Persistent inflammation, immunosuppression, and catabolism are associated with impaired lymphocytic mitochondrial metabolism during the early phase of sepsis: a single-center, prospective cohort study. Anaesthesiol Intensive Ther. 2026;58:164–171. Original exploratory cohort study.

9. Zhang Q, Raoof M, Chen Y, et al. Circulating mitochondrial DAMPs cause inflammatory responses to injury. Nature. 2010;464:104–107. Original human and experimental study.

10. Stacpoole PW, Wright EC, Baumgartner TG, et al. A controlled clinical trial of dichloroacetate for treatment of lactic acidosis in adults. N Engl J Med. 1992;327:1564–1569. Original randomized trial.

11. Lamontagne F, Masse MH, Menard J, et al. Intravenous vitamin C in adults with sepsis in the intensive care unit. N Engl J Med. 2022;386:2387–2398. Original randomized trial.