Clinical Cases in Applied Critical Care Physiology

All twelve vignettes and patient values in this chapter are invented for teaching. Their paired observations isolate useful relationships while retaining uncertainty about causes and responses. The cited studies support physiological mechanisms or particular clinical comparisons; they do not supply the case data. Each case connects what was measured, where and when it was measured, the assumptions behind a calculation, and the additional information needed to interpret the patient. None specifies a universal treatment target or protocol.

80.1 Arterial Pressure Improves While Oxygen Delivery Falls

A patient with infection-associated circulatory failure initially has warm extremities, mean arterial pressure of 55 mmHg, and measured forward cardiac output of 8 L/min. After several simultaneous changes in support, pressure rises to 65 mmHg while output falls to 6 L/min. Arterial oxygen content remains 12 mL O₂/dL. The pressure transducer remains correctly referenced, and the flow measurements are assumed comparable.

Let DO₂ be global oxygen delivery in mL O₂/min, CO cardiac output in L/min, and CaO₂ arterial content in mL O₂/dL. The factor 10 converts L to dL.

Initial delivery is 8 × 12 × 10 = 960 mL/min. Later delivery is 6 × 12 × 10 = 720 mL/min. The higher pressure accompanies 25% less global oxygen delivery. This is compatible with increased vascular resistance, but it does not identify which intervention changed tone, filling, contractility, or pulmonary loading.

If oxygen consumption were maintained at an assumed 300 mL/min, required global extraction would rise from 300 ÷ 960 = 31.25% to 300 ÷ 720 ≈ 41.7%. The calculation describes a requirement, not proof that every tissue can increase extraction sufficiently. Local flow distribution and diffusion remain hidden.

The low content also deserves explanation through hemoglobin, saturation, and sampling. A satisfactory oxygen tension cannot establish adequate hemoglobin-bound transport. New right ventricular dilation, reduced filling, arrhythmia, or impaired contraction could coexist with the distributive process.

ARISE FLUIDS compared early fluid/vasopressor strategies in emergency-department septic shock. The earlier-vasopressor, restricted-fluid strategy did not improve the primary days-alive-and-out-of-hospital outcome, although pulmonary edema was less frequent. This strategy comparison cannot identify a preferred action from the constructed pressure change alone. 1

Serial interpretation follows forward flow, content, congestion, regional perfusion, and demand alongside pressure. An improved pressure is meaningful, but its physiological cost and contribution to organ support remain separate questions.

80.2 An Early Hematocrit Conceals a Substantial Red-Cell Loss

A patient has an assumed pre-event blood volume of 6 L and hematocrit of 0.40, or 40%. An acute internal hemorrhage removes 1.5 L of well-mixed whole blood. No fluid redistribution or replacement is initially included in this deliberately simplified snapshot.

Red-cell volume is blood volume multiplied by hematocrit. It initially equals 6 × 0.40 = 2.4 L. The shed blood contains 1.5 × 0.40 = 0.6 L of red cells. Remaining blood volume is 4.5 L and remaining red-cell volume 1.8 L. Hematocrit is 1.8 ÷ 4.5 = 0.40, still 40%, despite loss of one quarter of the red-cell amount.

Now stipulate that 0.75 L of cell-free fluid enters and remains in the circulation without further bleeding or cell addition. Blood volume becomes 5.25 L, and hematocrit becomes 1.8 ÷ 5.25 ≈ 0.343, or 34.3%. The lower later concentration reveals dilution of an already reduced cellular amount. It does not date the loss to the later blood sample.

Real hemorrhage includes continuing exchange, treatment, sampling uncertainty, and possibly ongoing bleeding. The values are invented, and a peripheral hematocrit is not an exact measurement of the average fraction in every vascular compartment. Neither sample provides the missing volume independently.

Suppose pressure transiently rises after support while abdominal findings worsen. Improved pressure has not established control of the source. Volume support, oxygen-carrying cells, and hemostatic components address different deficits, while mechanical source control addresses continued escape of blood.

TOWAR's 2026 comparison of prehospital whole blood with blood components did not demonstrate lower 30-day mortality with whole blood. The active-comparator result does not establish product equivalence or explain whether this patient's bleeding has stopped. 2

The useful reconstruction combines source, timing, estimated loss, fluid movement, concentration trends, forward flow, and tissue consequences. A concentration can remain familiar while circulating mass and reserve have changed greatly.

80.3 Decongestion Improves Circulation While Creatinine Rises

A patient with biventricular dysfunction has edema, impaired breathing, mean arterial pressure of 80 mmHg, and central venous pressure of 18 mmHg. Both pressures use a common reference. During a later observed state, pressure is 76 mmHg, CVP 10 mmHg, and forward output has increased from 4 to 4.4 L/min. Breathing improves, but creatinine rises from 1.6 to 1.9 mg/dL.

The gross arterial-to-central-venous difference changes from 80 − 18 = 62 mmHg to 76 − 10 = 66 mmHg. Thus, a lower arterial pressure can coexist with a larger measured difference and better forward flow. This subtraction is not the glomerular filtration-pressure equation: renal venous pressure, tissue pressure, vascular resistance, and filtration properties remain unmeasured.

Relief of right ventricular distension may improve septal interaction and left ventricular filling. Lower pulmonary venous pressure can reduce hydrostatic lung burden. Conversely, further removal could eventually reduce useful filling or coronary support. The response describes the current operating state rather than unlimited tolerance of continued decongestion.

The creatinine rise could reflect altered filtration, distribution, or structural injury. Its direction alone cannot decide whether the whole response was beneficial, and improvement in edema does not establish that the rise is harmless.

In a CARRESS-HF substudy involving preexisting worsening renal function, further creatinine worsening during intensive removal was associated with tubular-marker increases. Increased markers also accompanied greater decongestion and better 60-day creatinine recovery. These observational associations within a selected trial cohort do not prove harmlessness or justify continued removal universally. 3

Follow the timing of creatinine, native excretion, pressure, flow, residual congestion, and other renal insults. New hypotension, electrolyte disturbance, or deteriorating perfusion would change the interpretation even if visible edema remained. Neither a biomarker nor a filling pressure contains the entire balance of benefit and injury.

80.4 Better Lung Mechanics and Oxygen Content Hide a Circulatory Cost

A passively ventilated patient with ARDS has tidal volume of 450 mL, valid plateau pressure of 28 cm H₂O, and total end-expiratory pressure of 12 cm H₂O. Later, tidal volume remains 450 mL while plateau pressure is 31 and total PEEP 16 cm H₂O. Holds are assumed interpretable, without effort or important leakage.

Let CRS be respiratory-system compliance in mL/cm H₂O and VT tidal volume in mL. The communicating respiratory-system driving pressure is the difference between plateau and total end-expiratory pressure.

Initially, driving pressure is 16 cm H₂O and compliance 450 ÷ 16 ≈ 28.1 mL/cm H₂O. Later, driving pressure is 15 and compliance 30 mL/cm H₂O. These global changes are compatible with improved participation or operating mechanics, but do not identify recruitment separately from expansion of already open units.

Arterial oxygen content improves from 17 to 18 mL O₂/dL while measured forward output falls from 5 to 4.2 L/min. Delivery changes from 5 × 17 × 10 = 850 to 4.2 × 18 × 10 = 756 mL/min. Better content and compliance coexist with less global oxygen delivery.

Venous-return limitation, pulmonary vascular loading, right ventricular reserve, or another simultaneous change could explain the flow reduction. The measurements do not isolate one cause. A chest-wall contribution and regional overdistension can remain hidden within the global mechanics.

ART demonstrated increased mortality with its tested recruitment-maneuver and compliance-titrated PEEP bundle compared with a low-PEEP strategy. It did not isolate every PEEP change or prove that this constructed patient has suffered the trial's mechanism of harm. It does show why a plausible mechanical improvement cannot establish outcome benefit. 4

The integrated interpretation follows participating lung, effort, gases, forward flow, right-heart geometry, and congestion. The achieved airway pressure is one boundary of the system, rather than a complete measure of its benefit.

80.5 A Modest Airway Pressure Accompanies Strong Patient-Generated Distension

An assisted patient with heterogeneous lung injury appears to have an acceptable airway-pressure display but makes forceful inspiratory efforts. For two constructed end-inspiratory observations, assume communicating alveolar pressure is adequately represented by a zero-flow airway measurement, and esophageal pressure supplies a representative global pleural estimate. Both use the same atmospheric reference.

In the first state, alveolar pressure is 22 cm H₂O and the pleural estimate −8 cm H₂O. In the second, they are 26 and −2 cm H₂O. The estimated end-inspiratory transpulmonary differences are 22 − (−8) = 30 and 26 − (−2) = 28 cm H₂O.

The higher airway pressure accompanies the smaller estimated distending difference because patient-generated negative pressure has decreased. These are end-inspiratory pressure-difference estimates, not tidal driving pressures or direct measurements of every region's stress. During flow, airway-opening pressure would include resistive losses and could not simply replace alveolar pressure.

Esophageal position, balloon conditions, mediastinal influence, regional pleural gradients, and timing limit the estimate. A globally representative signal can still conceal stronger local deformation. Tidal volume at the airway also does not reveal all gas movement between regions.

Yoshida and colleagues combined a clinical observation with injured-animal experiments to demonstrate effort-related pendelluft and concentrated dependent inflation. This supports a potential mechanism of hidden regional loading; it does not prove that every spontaneous breath is harmful or that one suppression strategy improves survival. 5

Effort also increases respiratory oxygen demand and changes venous return and left ventricular transmural load. Reduced visible effort could represent useful unloading, altered drive, fatigue, medication effects, or neurological deterioration. Its meaning requires ventilation, gas exchange, muscle activity, and circulation to be followed together.

A satisfactory saturation or machine pressure therefore cannot establish a tolerable patient–ventilator workload. The case calls for explaining how pressure is generated and distributed, while preserving uncertainty about regional exposure and the net consequences of changing assistance.

80.6 A Lower Peak Pressure Leaves Less Time for Expiration

A constructed patient with severe airflow obstruction receives passive constant-flow ventilation. Tidal volume is 0.50 L, frequency 20 breaths/min, and inspiratory flow 0.50 L/s. There is no inspiratory pause or additional triggered breath. These assumptions make delivered timing calculable.

The cycle lasts 60 ÷ 20 = 3 seconds. Inflation takes 0.50 ÷ 0.50 = 1 second, leaving 2 seconds for expiration. Now reduce inspiratory flow to 0.25 L/s while holding tidal volume and frequency unchanged. Inflation takes 2 seconds, leaving only 1 second for expiration.

Peak airway pressure may fall because flow-related resistive pressure decreases. That display can accompany less complete emptying and increasing end-expiratory volume. Minute ventilation remains 10 L/min in both states, but neither minute ventilation nor peak pressure establishes the trapped-volume burden.

In the observed vignette, expiratory flow continues into the next breath, measured venous pressure rises, and forward output falls. Hyperinflation could contribute through adverse thoracic pressure and right ventricular loading. A simultaneous pneumothorax, airway-device problem, arrhythmia, or drug effect remains possible; the temporal pattern does not establish one exclusive diagnosis.

Tuxen and Lane varied ventilatory patterns in nine sedated, paralyzed patients with severe obstruction. Hyperinflation and circulatory effects changed with volume and expiratory opportunity, while peak pressure did not reliably represent lung volume. Their experimental settings are not prescriptions for another patient. 6

A global hold-derived end-expiratory pressure also requires passive conditions and communicating pathways. Closure can conceal pressurized regions, while active expiration can increase pressure without the same volume change. The timing calculation identifies one plausible contributor without measuring all regional emptying.

The useful comparison follows actual delivery, flow, volume, valid pressure measurements, gases, and circulation. A later carbon dioxide improvement obtained through greater ventilation could still carry an adverse mechanical cost. Correcting the displayed number and improving the integrated system are related but distinct achievements.

80.7 Creatinine Falls Without Demonstrated Native Kidney Recovery

A patient with systemic inflammation and fluid accumulation has creatinine of 2.4 mg/dL. For an invented equilibrated compartment, assume creatinine amount is 840 mg and distribution volume 35 L, or 350 dL. The concentration is 840 ÷ 350 = 2.4 mg/dL.

Suppose volume subsequently becomes 42 L, or 420 dL, while amount remains unchanged in this isolated snapshot. Concentration becomes 840 ÷ 420 = 2.0 mg/dL. The apparent improvement has required no creatinine removal. Real patients continue producing, excreting, and redistributing creatinine, so this is not a clinical correction formula or a fixed body-water assumption.

Macedo and colleagues showed how modeled fluid adjustment could alter recognition of AKI severity in a critically ill observational cohort. The work supports considering dilution, while a modeled concentration remains distinct from measured filtration. 7

Now assume a complete collection has urine creatinine 100 mg/dL, plasma creatinine 2 mg/dL, and urine flow 0.6 mL/min. Creatinine clearance is (100 ÷ 2) × 0.6 = 30 mL/min. In another collection, urine flow is 1.5 mL/min and urine creatinine 40 mg/dL, with plasma unchanged: clearance remains 30 mL/min. More urine has not established more clearance. Secretion and nonsteady conditions further separate this estimate from exact GFR.

If extracorporeal therapy then lowers serum creatinine, external removal adds another explanation. Native recovery cannot be inferred from a circuit-supported concentration or water output alone.

ICRAKI's 2026 comparison did not establish noninferiority of initial intermittent therapy to continuous therapy and found higher mortality with the intermittent strategy. Its early stopping and selected population matter; the result does not identify the mechanism or turn a falling creatinine into proof of recovery. 8

The timeline should connect production, fluid distribution, urine solute handling, circuit delivery, and wider organ function. Several improving measurements can coexist with continuing kidney dysfunction or congestion.

80.8 Lactate Improves While pH Falls

A patient initially has arterial PCO₂ of 20 mmHg and gas-calculated bicarbonate of 12 mmol/L. A paired chemistry sample reports total CO₂ of 13 mmol/L, sodium 140 mmol/L, chloride 106 mmol/L, albumin 2 g/dL, and lactate 8 mmol/L. Later, arterial PCO₂ is 28 mmHg, gas-calculated bicarbonate 14 mmol/L, chemistry total CO₂ 15 mmol/L, chloride 112 mmol/L, and lactate 3 mmol/L; sodium and albumin are unchanged.

For this illustration at 37°C, use pK′ = 6.10 and dissolved-CO₂ coefficient 0.030 mmol/L/mmHg. With bicarbonate in mmol/L:

The first pH is 6.10 + log₁₀(12 ÷ 0.60) ≈ 7.40. Later it is 6.10 + log₁₀(14 ÷ 0.84) ≈ 7.32. The bicarbonate increase is outweighed by the relative PCO₂ increase. A falling lactate does not guarantee a rising pH.

Using sodium minus chloride minus chemistry total CO₂, the uncorrected anion gap falls from 21 to 13 mmol/L. Low albumin lowers the expected gap. With an explicitly assumed albumin reference of 4 g/dL and the empirical adjustment of 2.5 mmol/L per g/dL difference, adjusted values are approximately 26 and 18 mmol/L. Figge and colleagues' observational work supports this albumin effect; the adjustment does not identify every unmeasured ion or supply a universal laboratory reference interval. 9

The chloride increase relative to unchanged sodium suggests an additional acidifying contribution, while changing ventilation, production, renal handling, and other ions require assessment. Gas-calculated bicarbonate and measured chemistry total CO₂ are different quantities and need appropriately timed samples.

The coherent account follows respiratory and metabolic contributions together. Near-reference initial pH did not exclude substantial opposing disturbances, and the later lower pH does not prove that lactate production worsened. No single compensation approximation can establish that only one process is present.

80.9 A Cerebral Pressure Difference Depends on Where Pressure Is Referenced

A head-injured patient has mean arterial pressure of 85 mmHg referenced at heart level. The relevant cerebral arterial level is assumed 20 cm higher, while intracranial pressure is 20 mmHg at an appropriate cranial reference. Subtracting 20 directly from 85 would combine different hydrostatic levels.

For this constructed blood-column estimate, density is 1060 kg/m³, gravity 9.81 m/s², and height 0.20 m. Pressure difference equals density × gravity × height: approximately 2080 Pa. Using 133.3 Pa/mmHg, this is approximately 15.6 mmHg.

Estimated arterial pressure at the cerebral level is therefore 85 − 15.6 = 69.4 mmHg. The conventional CPP difference is 69.4 − 20 = 49.4 mmHg, rather than the unadjusted 65 mmHg. The numbers demonstrate reference dependence and are not treatment boundaries.

If heart-level MAP later becomes 90 and ICP 25 mmHg at the same assumed height, the difference remains 90 − 15.6 − 25 = 49.4 mmHg. Higher systemic pressure has not improved the calculated cerebral difference because intracranial pressure rose equally.

Actual cerebral flow also depends on vascular tone, autoregulation, carbon dioxide, oxygen content, venous outflow, and regional injury. ICP is not necessarily the sole effective downstream constraint under every condition. A single pressure subtraction cannot quantify regional flow or oxygen availability.

COGiTATE tested the feasibility of autoregulation-guided CPP management after traumatic brain injury. It supported delivery of the individualized strategy within that trial, while being designed for feasibility and safety rather than powered proof of improved functional outcomes. Its result does not validate the invented pressure values as individualized targets. 10

The same discipline applies to trends after positioning or respiratory changes: reference geometry and real physiology can change together. Interpret pressure, gases, circulation, neurological observations, and available regional monitoring on the same timeline rather than treating a calculated CPP as a complete brain assessment.

80.10 A Higher Venous Oxygen Content Does Not Diagnose Mitochondrial Failure

A septic patient initially has measured forward output of 4.5 L/min, arterial oxygen content of 18 mL O₂/dL, and genuinely mixed venous content of 12 mL/dL. Later, after changes in temperature, breathing effort, sedation, and circulation, output is 5 L/min, arterial content remains 18, and mixed venous content is 14 mL/dL. Assume comparable steady sampling without extracorporeal gas exchange or an important unaccounted shunt.

Let VO₂ be whole-body oxygen consumption in mL/min and CvO₂ mixed venous content in mL/dL. Under the stated Fick conditions:

Initially, consumption is 4.5 × (18 − 12) × 10 = 270 mL/min. Later it is 5 × (18 − 14) × 10 = 200 mL/min. Delivery has increased from 810 to 900 mL/min while global extraction has fallen.

The observations could reflect reduced metabolic demand, impaired extraction, altered distribution, or combinations. Less respiratory work and lower temperature can reduce demand; shivering could oppose that effect. There is no universal temperature coefficient that converts the readings into an expected consumption change.

A high venous oxygen value cannot establish mitochondrial failure, and preserved global consumption cannot exclude local supply limitation. A central venous sample would introduce a different mixing assumption and should not silently replace the mixed venous content used here.

Suppose an assay also shows low platelet respiratory-chain activity. Protti and colleagues found that platelet and skeletal-muscle mitochondrial enzyme findings did not track one another in their septic-shock study. This tissue-specific result cautions against using one accessible cell population to diagnose the state of every organ. 11

Mitochondrial interpretation additionally requires the assay substrate, normalization, oxygen conditions, viable-cell composition, and biological timing. A pressure, lactate, venous saturation, or ex-vivo enzyme result can contribute evidence without independently measuring organ ATP sufficiency or identifying a beneficial metabolic treatment.

80.11 Ejection Fraction Falls While Useful Forward Flow Improves

A patient with substantial mitral regurgitation has constructed left ventricular end-diastolic volume of 180 mL and end-systolic volume of 72 mL. Total chamber stroke volume is 108 mL and ejection fraction 108 ÷ 180 = 60%. Assume 68 mL per beat returns to the atrium, with no other important regurgitation or shunt.

Forward aortic stroke volume is therefore 108 − 68 = 40 mL. At 110 beats/min, forward output is 40 × 110 ÷ 1000 = 4.4 L/min. The vigorous-looking chamber empties much more blood than it delivers usefully to the systemic circulation.

In a later measured state, end-diastolic volume is 160 mL, end-systolic volume 80 mL, and regurgitant volume 25 mL. Total stroke volume is 80 mL and ejection fraction 50%. Forward stroke volume is 55 mL. At 95 beats/min, forward output is 5.225 L/min.

With arterial content held at 18 mL O₂/dL, delivery rises from 4.4 × 18 × 10 = 792 to 5.225 × 18 × 10 = 940.5 mL/min. A smaller ejection fraction accompanies greater forward transport. These are paired observations rather than a prediction of one procedure or drug response.

Loading, contractility, geometry, rhythm, and the backward opening can all change. Reducing regurgitation removes one ejection pathway and can expose limited ventricular reserve. Hashimoto and colleagues associated estimated loading changes with acute ejection-fraction reduction after successful mitral edge-to-edge repair in a selected retrospective cohort. Their model-dependent findings do not establish that every fraction decline is benign or exclude myocardial injury. 12

Volume measurements must concern compatible beats and loading states. Subtracting independently uncertain volumes can amplify error, and mixed lesions invalidate the simple conservation boundary. Pulmonary venous pressure, symptoms, forward flow, and organ perfusion add information about whether the later state is favorable. Neither a total chamber fraction nor a lower regurgitant estimate supplies that judgment alone.

80.12 Successful Membrane Gas Exchange Does Not Establish Complete Support or Recovery

A patient on venovenous extracorporeal support has circuit flow of 4 L/min, premembrane oxygen content of 12 mL O₂/dL, and postmembrane content of 20 mL/dL. Assume paired, representative samples under stable circuit conditions. Net membrane oxygen transfer is 4 × (20 − 12) × 10 = 320 mL/min.

The outlet contains 800 mL O₂/min, but 480 mL/min entered the membrane already present in blood. Counting all outlet oxygen as newly added would overstate transfer. Recirculation affects inlet conditions; an additional arbitrary correction to the measured content-difference flux could count that influence twice.

Suppose an independent assessment valid for this configuration gives systemic output of 6 L/min and arterial content of 17 mL/dL. Delivery is 1020 mL/min. VV circuit flow is not added to cardiac output to claim 10 L/min of systemic flow: returned blood must still traverse the native heart and lungs.

Later, arterial content rises to 19 mL/dL while systemic output falls to 4.5 L/min. Delivery is now 855 mL/min. Better arterial oxygenation accompanies less global transport. The pair does not assume unchanged membrane transfer, native exchange, venous content, or demand.

Schmidt and colleagues' small VV physiological study demonstrated different influences of circuit flow and sweep on oxygenation and carbon dioxide elimination. It supports separating these functions, without supplying universal settings or proving clinical benefit from a corrected gas alone. 13

Changing support also changes respiratory drive, ventilation, cardiac loading, and the conditions under which native organs are assessed. With VA return, additional questions concern regional mixing and ventricular afterload; the VV accounting cannot simply be transferred to that different pathway.

Across these cases, the durable discipline is to identify the measured compartment, preserve units and references, connect pressure with forward flow and content, and follow demand and organ consequences over time. Supported performance, causal recovery, and outcome benefit require distinct evidence even when several monitor values improve together.

References

1. ARISE FLUIDS Investigators, ANZICS Clinical Trials Group, ACEM Clinical Trials Network. Vasopressors or fluids in early septic shock. N Engl J Med. 2026;395:1387–1398. doi:10.1056/NEJMoa2516225.

2. Sperry JL, Guyette FX, Cotton BA, et al. Prehospital resuscitation with type O whole blood for trauma and hemorrhage. N Engl J Med. 2026;394:2317–2328. doi:10.1056/NEJMoa2602167.

3. Rao VS, Ahmad T, Brisco-Bacik MA, et al. Renal effects of intensive volume removal in heart failure patients with preexisting worsening renal function. Circ Heart Fail. 2019;12:e005552. doi:10.1161/CIRCHEARTFAILURE.118.005552.

4. Writing Group for the ART Investigators, Cavalcanti AB, Suzumura ÉA, et al. Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs low PEEP on mortality in patients with acute respiratory distress syndrome: a randomized clinical trial. JAMA. 2017;318:1335–1345. doi:10.1001/jama.2017.14171.

5. Yoshida T, Torsani V, Gomes S, et al. Spontaneous effort causes occult pendelluft during mechanical ventilation. Am J Respir Crit Care Med. 2013;188:1420–1427. doi:10.1164/rccm.201303-0539OC.

6. Tuxen DV, Lane S. The effects of ventilatory pattern on hyperinflation, airway pressures, and circulation in mechanical ventilation of patients with severe air-flow obstruction. Am Rev Respir Dis. 1987;136:872–879. doi:10.1164/ajrccm/136.4.872.

7. Macedo E, Bouchard J, Soroko SH, et al. Fluid accumulation, recognition and staging of acute kidney injury in critically-ill patients. Crit Care. 2010;14:R82. doi:10.1186/cc9004.

8. Gaudry S, Boubaya M, Louis G, et al. Intermittent or continuous renal-replacement therapy in acute kidney injury. N Engl J Med. 2026;395:1677–1690. doi:10.1056/NEJMoa2608397.

9. Figge J, Jabor A, Kazda A, Fencl V. Anion gap and hypoalbuminemia. Crit Care Med. 1998;26:1807–1810. doi:10.1097/00003246-199811000-00019.

10. Tas J, Beqiri E, van Kaam RC, et al. Targeting autoregulation-guided cerebral perfusion pressure after traumatic brain injury (COGiTATE): a feasibility randomized controlled clinical trial. J Neurotrauma. 2021;38:2790–2800. doi:10.1089/neu.2021.0197.

11. 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. doi:10.1371/journal.pone.0096205.

12. Hashimoto K, Nagai T, Kamioka N, et al. Increased afterload in patients with acute reduction in left ventricular ejection fraction following mitral valve transcatheter edge-to-edge repair. Echocardiography. 2025;42:e70095. doi:10.1111/echo.70095.

13. Schmidt M, Tachon G, Devilliers C, et al. Blood oxygenation and decarboxylation determinants during venovenous ECMO for respiratory failure in adults. Intensive Care Med. 2013;39:838–846. doi:10.1007/s00134-012-2785-8.