Acute Kidney Injury and Multiorgan Crosstalk

Acute kidney injury changes more than filtration. It alters fluid distribution, electrolyte and acid–base balance, solute handling, drug exposure, and inflammatory communication. At the same time, cardiac, pulmonary, hepatic, and systemic disease can impair renal function through several pathways. The kidney is therefore both a participant in multiorgan dysfunction and a source of measurements that reflect the wider illness. Interpreting those measurements requires separating concentration, excretion, injury, and recovery.

77.1 Acute Kidney Injury Describes a Changing Functional State

Acute kidney injury, or AKI, encompasses different causes and combinations of functional impairment and tissue injury. Reduced filtration can accompany altered perfusion, venous congestion, inflammation, obstruction, toxic exposure, or intrinsic glomerular and tubular disease. Similar creatinine and urine-output patterns can arise through different mechanisms and require different explanations.

Functional changes and structural injury overlap without being identical. Glomerular filtration can fall through altered vascular or filtration pressures before extensive cell death occurs. Conversely, tubular stress and injury may develop before a large serum creatinine rise becomes apparent. Tubular transport, barrier function, and microvascular regulation can be disturbed unevenly across nephrons.

The starting kidney matters. Chronic nephron loss, vascular disease, prior injury, and reduced functional reserve change the response to a new stress. An apparently ordinary baseline creatinine can coexist with low muscle mass and limited filtration reserve. A later value must be interpreted against the person's preceding trajectory rather than an assumed universal starting function.

Kellum and colleagues analyzed creatinine and urine-output trajectories in a large intensive-care cohort. Concurrent abnormalities and their persistence were associated with worse outcomes than isolated or shorter abnormalities. These observations support using both domains and time; they do not prove that changing a measurement alone will reverse the associated risk. 1

AKI can also be a marker of severe systemic illness. Infection, shock, congestion, or toxic exposure may injure several organs simultaneously. Finding kidney and lung dysfunction together does not establish that one caused the other. Establishing a causal contribution requires timing, plausible pathways, and evidence that distinguishes shared causes from downstream effects.

The detailed filtration, oxygenation, transport, congestion, and replacement mechanisms were developed in Chapters 43–51. This chapter connects them to the wider physiological system. Its focus is how changing renal function modifies other organs and how those organs feed back on the kidney.

77.2 Creatinine Concentration Depends on Production, Distribution, and Removal

Serum creatinine is a concentration within a changing distribution space. Its time course depends on creatinine generation, movement between compartments, renal filtration and secretion, extrarenal handling, and any extracorporeal removal. A change in filtration does not immediately produce a new steady concentration, and a falling concentration does not necessarily identify recovering filtration.

Let M be the amount of creatinine in an assumed equilibrated distribution compartment in mg, V its volume in dL, and C the concentration in mg/dL. For a simple accounting snapshot:

In a constructed example, M is 900 mg and V is 40 L, equivalent to 400 dL. Concentration is 900 ÷ 400 = 2.25 mg/dL. If the distribution volume becomes 45 L, or 450 dL, while the amount is unchanged, concentration becomes 900 ÷ 450 = 2.00 mg/dL. The lower result reflects dilution without any removal of creatinine.

This is an illustrative compartment, not a fixed normal body-water fraction or a bedside correction formula. Real patients continue generating and excreting creatinine, and administered fluid does not equilibrate instantly. Distribution, sampling, and balance estimates introduce uncertainty.

Macedo and colleagues examined fluid accumulation and creatinine-based AKI recognition in critically ill patients. Their modeled fluid adjustment suggested that dilution could delay recognition of severity. The study supports considering distribution volume; it does not make a calculated adjustment a direct measurement of filtration or a universally validated treatment trigger. 2

Generation changes as well. Muscle loss and altered metabolism can lower production during prolonged illness. Doi and colleagues demonstrated reduced creatinine production in septic nephrectomized mice, where renal clearance could not explain the difference. This experimental isolation supports a mechanism, not a quantitative correction applicable to every septic patient. 3

Steady-state estimated-filtration equations assume relationships that changing creatinine can violate. During rapid decline or recovery, the present concentration contains information about the preceding interval. Serial values become more useful when fluid balance, production, native excretion, and treatment are placed on the same timeline.

77.3 Urine Volume, Solute Excretion, and Filtration Are Different Quantities

Urine volume represents the water remaining after filtration and tubular handling. It depends on filtered water, reabsorption, osmotic load, concentrating ability, hormones, and drug effects. A decrease can reflect conservation, reduced filtration, obstruction, or tubular dysfunction in different combinations. Substantial urine production does not exclude AKI.

A diuretic can increase water and sodium excretion by reducing tubular reabsorption without repairing the process limiting filtration. Conversely, a kidney recovering filtration may initially excrete dilute urine because concentrating function remains impaired. The amount of one solute excreted cannot be inferred from water volume alone.

Let UCr and PCr be urine and plasma creatinine concentrations in the same units, here mg/dL. Let Vdot be measured urine flow in mL/min. A timed creatinine clearance, CCr in mL/min, is:

For a constructed interval, UCr is 80 mg/dL, PCr is 2 mg/dL, and urine flow is 1 mL/min. Clearance is (80 ÷ 2) × 1 = 40 mL/min. In a second interval, urine flow doubles to 2 mL/min while UCr falls to 40 mg/dL and PCr remains 2. Clearance is (40 ÷ 2) × 2 = 40 mL/min again.

More urine has not increased the measured clearance in this pair. The calculation assumes an accurate collection and a representative plasma concentration. Tubular secretion means creatinine clearance is not identical to glomerular filtration, and rapid nonsteady changes make its interpretation less exact.

Collection obstruction, incomplete drainage, documentation intervals, and changes in body size or fluid balance also affect urine-output interpretation. The persistence of low output adds information, but does not establish one cause or mandate one fluid response.

Urine microscopy, urinary solutes, imaging, and injury or stress biomarkers can add different dimensions. Their specificity, timing, and extrarenal influences need consideration. They complement the physiological account rather than turning every rise in urine output or fall in a biomarker into proof that the kidney has healed.

77.4 Renal Perfusion Depends on Arterial, Venous, and Tissue Conditions

Renal blood flow depends on pressure differences and vascular behavior, while filtration additionally depends on glomerular and tubular pressures, oncotic forces, and filtration properties. Arterial pressure is one boundary of this system. Renal venous and interstitial conditions form other important boundaries.

For an illustrative organ-scale comparison, assume mean renal arterial pressure is 75 mmHg and renal venous pressure 5 mmHg, giving an arterial–venous difference of 70 mmHg. In another state, arterial pressure is 80 and venous pressure 20 mmHg, giving 60 mmHg. The higher arterial pressure accompanies a lower gross pressure difference.

These values share a reference and represent constructed renal pressures, not routine assumptions that MAP and CVP exactly equal pressures at both renal boundaries. The subtraction is not the glomerular filtration-pressure equation and cannot predict GFR without vascular, tubular, and tissue information.

Mullens and colleagues linked venous congestion with worsening renal function in advanced decompensated heart failure. Their hemodynamically characterized observational cohort supports attention to downstream conditions. It does not prove that output never matters or establish a universal CVP target. 4

Abdominal pressure, renal edema, and capsular constraint can alter drainage and tubular pressures. A distended venous system can coexist with low effective arterial filling, especially when cardiac function and vascular capacity are abnormal. Added volume may then increase congestion more than useful forward flow.

Supply also depends on oxygen content and regional distribution. In conscious septic sheep, Lankadeva and colleagues demonstrated medullary hypoperfusion and hypoxia despite increased total renal blood flow. The model supports regional vulnerability beyond a global flow measurement; it does not provide a human pressure target or universal drug response. 5

Filtered sodium delivery and tubular reabsorption create substantial oxygen demand, but reduced filtration does not make the injured kidney metabolically uniform or protected. Transport inefficiency, inflammation, diffusion conditions, and microvascular redistribution can disrupt the relationship. Pressure, flow, filtration, and tissue oxygen availability therefore require separate interpretation.

77.5 Lung Disease and Respiratory Support Can Modify Kidney Function

Lung failure can threaten renal oxygen availability through hypoxemia, altered cardiac output, inflammation, and increased respiratory work. The resulting renal response depends on the whole circulation rather than arterial oxygen tension alone. Raising saturation may improve content while another change reduces flow.

Positive-pressure ventilation changes the pressures surrounding the heart and great veins. Depending on transmission and reserve, it can reduce venous return, increase right ventricular load, or change left ventricular ejection conditions. A rise in measured CVP may reflect transmitted external pressure, true congestion, or both. Its renal implication cannot be derived from a fixed fraction of airway pressure.

Lung volume and pulmonary vascular behavior matter. Recruitment may improve gas exchange and some vascular conditions, while overdistension can increase right ventricular loading. Right ventricular dysfunction can reduce forward delivery and raise systemic venous pressure simultaneously, exposing the kidney to both supply and drainage problems.

Respiratory gases add further pathways. Hypercapnia and acidemia modify vascular tone, neurohumoral responses, and the acid load handled by the kidney. Renal bicarbonate adaptation depends on remaining function and time. A kidney with limited acid excretion cannot supply the same compensation as an otherwise healthy kidney exposed to sustained carbon dioxide elevation.

There are also potential inflammatory links. Imai and colleagues compared ventilatory strategies in rabbits with aspiration-related lung injury and studied effects of their plasma on renal epithelial cells. The injurious strategy increased renal epithelial apoptosis and dysfunction; a small associated human analysis was correlational. This supports a possible systemic mediator pathway without proving its magnitude in every ventilated patient. 6

The response to a ventilator change should consequently follow mechanics, gases, cardiac performance, congestion, and renal trends together. A temporal association helps identify a contribution but does not exclude sepsis, obstruction, drug exposure, or another simultaneous cause. Chapters 39–40 and 75–76 develop the detailed respiratory mechanisms.

77.6 Injured Kidneys Can Influence Distant Barriers and Inflammatory Signaling

Kidney-to-lung interaction is not limited to retained water. Injury can change circulating mediator production, metabolism, and clearance, while distant endothelial and epithelial cells respond to the altered environment. These pathways can coexist with hydrostatic edema, infection, and direct lung injury.

An increased circulating cytokine concentration does not identify its source. It can reflect greater release from kidney or other tissues, reduced clearance, altered distribution, or combinations. Urinary and plasma measurements also sample different compartments. Demonstrating accumulation alone does not establish that the mediator causes organ dysfunction.

Klein and colleagues studied ischemic AKI and bilateral nephrectomy in mice. Interleukin-6 deficiency reduced lung inflammatory and permeability responses; antibody inhibition reduced selected inflammatory measures. These manipulations support a causal contribution within the tested models, while nephrectomy and ischemic injury are distinct preparations. The findings do not establish clinical benefit from cytokine removal or blockade in heterogeneous human AKI. 7

Distant effects can also involve the brain. Liu and colleagues observed inflammatory changes, altered barrier permeability, and reduced locomotor activity after severe ischemic AKI in mice. The experimental exposure and behavioral measure do not directly define human delirium, a brain-injury threshold, or a dialysis-responsive syndrome. They demonstrate a possible biological link requiring careful translation. 8

In critically ill people, common causes are difficult to separate. Sepsis can simultaneously injure kidney, lung, brain, and endothelium. Sedatives, acid–base changes, hypoxemia, and retained metabolites can contribute to altered consciousness during the same interval. Assigning all of that dysfunction to a single renal inflammatory pathway exceeds the evidence.

The same caution applies to recovery. Lower circulating concentrations after extracorporeal treatment may demonstrate extraction without proving suppression of tissue production or reversal of downstream injury. Mechanistic experiments help identify possible connections; clinical strategies need independent evaluation of the relevant patient outcomes and adverse effects.

77.7 Retained Fluid and Solutes Feed Back on Heart, Lung, and Muscle

When renal excretion cannot match input and endogenous release, sodium, water, potassium, acid-associated solutes, and other substances accumulate according to different balances. Their consequences depend on distribution and rate as well as amount. One serum concentration cannot summarize the entire retained burden.

Sodium and water retention can increase venous and pulmonary hydrostatic pressures. Interstitial lung water may reduce compliance and impair exchange, increasing respiratory work and changing the required support. Rising abdominal and systemic venous pressures can then worsen renal drainage, creating a feedback between retention and impaired excretion.

The cardiac effect depends on reserve. Additional filling can improve stroke volume in an underfilled responsive circulation, while a constrained or failing ventricle may mainly generate higher upstream pressure. A patient can be edematous and still have inadequate effective filling or forward flow. Removing fluid can relieve congestion but can also reduce return if removal outpaces vascular accommodation.

Potassium retention can alter resting membrane potential, conduction, and excitability. The extracellular concentration also reflects cellular release and uptake, insulin, adrenergic effects, acid–base conditions, and tissue injury. Total-body potassium and the measured plasma fraction are related but distinct quantities; a concentration change need not represent the same change in total amount.

Impaired net acid excretion alters bicarbonate and other buffers, while the lungs determine carbon dioxide elimination. Added respiratory compensation can increase breathing demand, and limited ventilatory reserve can make the combined disturbance more consequential. Neither potassium nor pH has one universally tolerable value independent of rate, context, and coexisting abnormalities.

Phosphate, magnesium, calcium, and nutritional balance also influence myocardial and skeletal-muscle function. During recovery or extracorporeal treatment, retention can give way to depletion. The relevant physiology is therefore the changing input, release, redistribution, and removal of each species, connected to the organs that depend on it. Chapters 45–47 and 52–57 develop those specific mechanisms.

77.8 Hepatic Disease Changes Renal Circulation and the Meaning of Solute Markers

Advanced hepatic disease can combine splanchnic vasodilation, altered effective arterial filling, neurohumoral sodium retention, cardiac limitations, inflammation, and portal or abdominal congestion. Substantial ascites and edema can coexist with a circulation that provides inadequate renal perfusion support. Total fluid amount does not establish its effective distribution.

Hepatorenal dysfunction is consequently not explained by arterial pressure alone. Changes in vascular capacity and renal vasomotor regulation interact with cardiac reserve and venous conditions. Infection, hemorrhage, nephrotoxic exposure, and intrinsic kidney injury may coexist. The presence of cirrhosis does not establish that every episode of AKI is one purely functional syndrome without tubular injury.

Liver and kidney also share solute processing. Urea concentration depends on hepatic production, protein intake and catabolism, distribution, and renal or extracorporeal removal. Reduced production can blunt a rise despite impaired excretion, while tissue breakdown or gastrointestinal blood digestion can increase production. Urea and creatinine need not move together.

The kidneys contribute to ammonia generation and its partition between urinary excretion and return to the circulation. Liver handling, acid–base conditions, potassium balance, and substrate availability modify that system. A blood ammonia value therefore reflects interorgan production and handling rather than a direct measure of renal filtration or one isolated organ's injury.

CONFIRM tested terlipressin versus placebo with albumin strongly recommended in both groups among adults with trial-defined hepatorenal syndrome. Verified reversal occurred more often with terlipressin, but respiratory adverse events were more frequent and a survival advantage was not established. This population-specific result demonstrates why improved renal-function endpoints and whole-patient benefit require separate assessment. 9

It does not imply that a creatinine decrease proves recovery from every hepatic–renal mechanism. Changes in distribution, production, hemodynamics, and intrinsic injury can contribute simultaneously. Following liver function, circulation, respiratory state, and native kidney handling together provides a more coherent account than assigning all retained solutes to filtration failure alone.

77.9 Drug Exposure Can Change Through Renal and Nonrenal Pathways

AKI can reduce elimination of a parent drug, an active metabolite, or both. Changes in protein binding, pH, distribution volume, tissue perfusion, transporters, and hepatic metabolism can also alter exposure. The direction is drug-specific and evolves during illness, recovery, and extracorporeal support.

For a highly protein-bound substance, the unbound fraction available for distribution, elimination, and receptor interaction can change when albumin falls or competing solutes accumulate. A lower total concentration need not mean a proportionally lower active concentration. Conversely, a large distribution volume can reduce plasma concentration while substantial drug remains outside plasma.

Kirwan and colleagues associated AKI severity and duration with reduced elimination of a midazolam probe in critically ill patients, proposing altered CYP3A activity as a contributor. The study did not establish one universal hepatic-clearance correction or isolate every illness-related influence. 10

A 2026 observational study by Shudofsky and colleagues found higher dose-normalized exposure to the glucuronidated midazolam metabolite in AKI and chronic kidney disease, without a significant association for the parent drug or its preceding hydroxylated metabolite. The small cohort illustrates why parent and metabolite kinetics should be distinguished; it does not validate a universal monitoring or dosing strategy. 11

Neurological or respiratory findings can consequently reflect several interacting processes: retained endogenous solutes, drug or metabolite accumulation, altered brain sensitivity, inflammation, sleep disruption, or primary cerebral disease. Reduced consciousness during AKI cannot automatically be labeled either uremia or sedation from one concentration measurement.

Extracorporeal clearance adds another pathway whose effect depends on unbound fraction, distribution, molecular properties, circuit operation, and residual native clearance. A nominal effluent flow does not establish removal of every drug, and treatment interruption can change exposure before serum creatinine provides a useful clue. Restored kidney function can reverse the balance again. The physiological task is to identify the relevant compound and elimination pathways rather than apply one renal-function number to all medications.

77.10 Extracorporeal Support Changes Several Organ Interfaces at Once

Renal replacement therapy can remove selected solutes and fluid, supply buffer, and modify electrolyte balance. It substitutes for parts of kidney function without reproducing all tubular, endocrine, metabolic, and immune roles. A corrected plasma concentration does not establish restoration of the native organ's full contribution.

Fluid removal can reduce pulmonary hydrostatic burden and venous congestion. At the same time, extraction from the vascular compartment can reduce effective filling when interstitial refill and cardiovascular accommodation cannot match it. The same net removal can therefore have different consequences at different stages, even in the same person.

Solute exchange changes gradients between plasma and tissues. Rapid plasma changes can precede equilibration in muscle or brain, producing redistribution and later rebound. The extracorporeal circuit samples a circulating compartment; it does not instantly expose or empty every tissue store. Water shifts follow the effective gradients and the permeability of the relevant barriers.

Acid–base effects also connect organs. Buffer transfer and removal of acid-associated solutes alter extracellular chemistry, while carbon dioxide generated or redistributed during buffering still requires appropriate pulmonary elimination. Changes in pH and ionized calcium can affect vascular and cardiac behavior during the same interval.

Electrolytes, nutrients, and medications can leave with treatment, while replacement fluids can supply others. Thermal exchange changes heat balance and sometimes vascular tone. Circuit clotting, blood–surface contact, vascular access, and anticoagulation add exposures and potential harms unrelated to the native filtration defect.

These mechanisms mean that more extracorporeal exchange is not synonymous with more complete organ recovery. Lowering a circulating inflammatory mediator does not prove that tissue signaling has stopped, and controlling urea does not ensure adequate nutrition or neurological recovery. Native excretion and metabolism continue to evolve alongside the circuit.

Chapters 50 and 51 provide the transport accounting and detailed cardiovascular effects. Here, the central connection is that support perturbs the lung, heart, brain, metabolism, and drug handling together. Its benefits and costs must be assessed across those interfaces rather than inferred from one achieved laboratory value.

77.11 Clinical Outcomes and Native Recovery Need Their Own Evidence

Mechanistic plausibility identifies what an intervention might change. Demonstrating solute removal, increased urine flow, a lower venous pressure, or improved oxygenation establishes a physiological response under the measured conditions. It does not establish the net effect on survival, neurological function, sustained renal recovery, or treatment complications.

The 2026 ICRAKI trial compared initial intermittent with continuous renal replacement in critically ill adults already requiring support. It randomized 554 patients and analyzed 535. Intermittent treatment did not establish noninferiority for the day-90 composite of death, treatment dependence, or persistent creatinine elevation. The trial stopped early for a mortality signal: death occurred in 63.1% versus 53.5%, respectively, with hazard ratio 1.30 (95% confidence interval 1.04–1.63). This important result concerns the tested initial strategies; early stopping limits effect precision, and the mechanism is not established. 12

Failure to establish noninferiority is not by itself proof of superiority for the composite. Observed complication frequencies alone cannot identify which pathway mediated an outcome difference. Strategy effects include the way care is delivered and its interaction with patient conditions, rather than one isolated transport variable.

Native recovery must also be separated from supported performance. During extracorporeal treatment, a stable creatinine can reflect sufficient circuit removal despite little kidney clearance. After interruption, a rise can reflect continued production and redistribution, not necessarily a new injury. Urine volume may improve while solute excretion remains limited.

Muscle loss can make a later creatinine appear reassuring even when filtration reserve remains reduced. Conversely, a transient concentration rise during changing fluid balance does not alone identify irreversible tissue damage. The trajectory requires input, distribution, native handling, treatment delivery, and organ function.

Longer-term vulnerability can persist after the acute episode appears resolved. Follow-up interpretation therefore considers the previous baseline, residual function, recurrent exposures, and the wider illness. A return of one laboratory measurement toward its earlier value is one observation within recovery, rather than its complete definition.

77.12 Integrate the Kidney with the Whole Patient's Timeline

Consider a constructed patient with lung injury, impaired right ventricular reserve, and increasing fluid accumulation. Mean renal arterial and venous pressures are assumed to change from 75 and 5 mmHg to 80 and 20 mmHg. The gross pressure difference falls from 70 to 60 mmHg despite a higher arterial pressure. Actual filtration still depends on intrarenal vascular, oncotic, and tubular conditions.

Suppose creatinine falls from 2.25 to 2.00 mg/dL during the same interval. In the earlier accounting model, expansion from 40 to 45 L with unchanged amount explains exactly that change. This does not prove dilution is the only mechanism in a real patient; it shows why the concentration alone cannot establish renal improvement.

Now apply the earlier collection example with plasma creatinine fixed at 2 mg/dL. Urine flow doubles while urine creatinine concentration halves, and the constructed clearance remains 40 mL/min. Water handling has changed without a corresponding increase in this clearance estimate. Improved urine volume and lower serum concentration can therefore coexist with persistent limitations and worsening congestion.

The lung adds competing requirements. Removing retained fluid might improve hydrostatic conditions, whereas reducing effective filling could impair forward flow. A respiratory adjustment might improve oxygen content while increasing right ventricular load or transmitted venous pressure. The net renal consequence cannot be selected from one improved number.

If extracorporeal support subsequently lowers urea, potassium, and creatinine, it demonstrates useful external handling of those solutes. Native recovery remains a separate question. Changes in consciousness or ventilation also require drug and metabolite exposure, acid–base state, and the primary illness to be considered.

The integrated account follows measured and assumed quantities separately, aligns them in time, and asks which mechanisms each observation supports. Kidney dysfunction can be an effect, a contributor, and a marker of the same critical illness. Understanding those roles makes multiorgan physiology more coherent without converting uncertainty into a universal fluid, pressure, urine-output, or replacement protocol.

References

1. Kellum JA, Sileanu FE, Murugan R, et al. Classifying AKI by urine output versus serum creatinine level. J Am Soc Nephrol. 2015;26:2231–2238. doi:10.1681/ASN.2014070724.

2. 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.

3. Doi K, Yuen PST, Eisner C, et al. Reduced production of creatinine limits its use as marker of kidney injury in sepsis. J Am Soc Nephrol. 2009;20:1217–1221. doi:10.1681/ASN.2008060617.

4. Mullens W, Abrahams Z, Francis GS, et al. Importance of venous congestion for worsening of renal function in advanced decompensated heart failure. J Am Coll Cardiol. 2009;53:589–596. doi:10.1016/j.jacc.2008.05.068.

5. Lankadeva YR, Kosaka J, Evans RG, et al. Intrarenal and urinary oxygenation during norepinephrine resuscitation in ovine septic acute kidney injury. Kidney Int. 2016;90:100–108. doi:10.1016/j.kint.2016.02.017.

6. Imai Y, Parodo J, Kajikawa O, et al. Injurious mechanical ventilation and end-organ epithelial cell apoptosis and organ dysfunction in an experimental model of acute respiratory distress syndrome. JAMA. 2003;289:2104–2112. doi:10.1001/jama.289.16.2104.

7. Klein CL, Hoke TS, Fang WF, et al. Interleukin-6 mediates lung injury following ischemic acute kidney injury or bilateral nephrectomy. Kidney Int. 2008;74:901–909. doi:10.1038/ki.2008.314.

8. Liu M, Liang Y, Chigurupati S, et al. Acute kidney injury leads to inflammation and functional changes in the brain. J Am Soc Nephrol. 2008;19:1360–1370. doi:10.1681/ASN.2007080901.

9. Wong F, Pappas SC, Curry MP, et al. Terlipressin plus albumin for the treatment of type 1 hepatorenal syndrome. N Engl J Med. 2021;384:818–828. doi:10.1056/NEJMoa2008290.

10. Kirwan CJ, MacPhee IAM, Lee T, Holt DW, Philips BJ. Acute kidney injury reduces the hepatic metabolism of midazolam in critically ill patients. Intensive Care Med. 2012;38:76–84. doi:10.1007/s00134-011-2360-8.

11. Shudofsky K, Abma J, Driessen J, et al. Association of acute kidney injury and chronic kidney disease with concentration-to-dose ratios of midazolam and its main metabolites in critically ill adults. Fundam Clin Pharmacol. 2026;40:e70108. doi:10.1111/fcp.70108.

12. 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.