Acute Kidney Injury

In depth

The one-page summary above is what the PDF prints. These sections go further.

Pathophysiology and creatinine kinetics

The kidneys take about a fifth of cardiac output, yet the outer medulla runs near hypoxia because sodium transport in the thick ascending limb consumes its oxygen. ICU injury is rarely single: hypotension, sepsis, congestion, abdominal pressure, nephrotoxins and pigment add up.

Sepsis-associated AKI, AKI within 7 days of sepsis onset (ADQI 28, 2023), is not simply ischemia: 60 of 159 animal studies found renal blood flow unchanged or increased, tracking cardiac output (Langenberg 2005). Microcirculatory shunting, inflammation and a metabolic switch in tubular cells cause injury despite flow.

Perfusion depends on the arterial-to-venous gradient inside a capsule. In decompensated heart failure, worsening renal function tracked admission CVP (18 vs 12 mm Hg) at every level of cardiac index (Mullens 2009), and intra-abdominal hypertension compresses renal veins the same way.

Creatinine is late and indirect: it rises 24–48 h after an insult, once nearly half of function is lost (KDIGO 2026 draft). Fluid dilutes it, liver disease and low muscle mass lower its production, and trimethoprim or cimetidine raise it by blocking tubular secretion without changing GFR. Cystatin C ignores muscle mass; the draft suggests it when creatinine is unreliable (2B).

Definitions: KDIGO 2012, the 2026 draft and outcomes

KDIGO 2012 merged RIFLE (2004) and AKIN (2007). Severity predicts death: in AKI-EPI (Hoste 2015; 1,802 ICU patients), 57.3% developed AKI in the first week, adjusted odds of hospital death were 2.9 at stage 2 and 6.9 at stage 3, and 47.7% of those with AKI left hospital with an eGFR below 60, against 14.8% without.

The baseline drives the label: the 2026 draft prefers an outpatient creatinine from 8–365 days before admission, since an admission value misses AKI that began at home.

Work-up: urine chemistry, sediment, biomarkers and imaging

FENa measures how much filtered sodium escapes; underperfused but intact tubules reabsorb nearly all of it. In a 2022 meta-analysis (Abdelhafez; 19 studies, 1,287 patients), a 1% cutoff had 90% sensitivity and 82% specificity overall, 95% and 91% in oliguric patients without diuretics or CKD, but 83% and 66% with diuretics or CKD.

Loop diuretics raise sodium but not urea excretion, so FEUrea under 35% supports a prerenal state on diuretics (89% of such patients; ATN averaged 58.6%; Carvounis 2002). FENa can fall under 1% in ATN from contrast, pigment or early sepsis and in GN or obstruction (Kidney360 2022), and values under 0.1–0.2% favor hepatorenal syndrome. The 2026 draft reserves urine chemistry for selected cases and suggests microscopy (2D).

Urine TIMP-2·IGFBP7, cell-cycle arrest markers (NephroCheck, FDA-authorized 2014), predicted stage 2–3 AKI within 12 h with an AUC of 0.80 in 728 critically ill adults (Sapphire, Kashani 2013); risk rose sharply above 0.3 (ng/mL)²/1000, where sensitivity was 92% (Topaz 2014). An NGAL test is FDA-cleared (2023) only for ICU patients aged 3 months to 21 years.

Ultrasound finds hydronephrosis or a distended bladder; get it for anuria, poor catheter output, stones, pelvic tumors or prostate disease (2026 draft). A post-void residual over 100 mL means incomplete emptying (Merck 2026).

Perfusion, fluids and abdominal pressure

Pressure: KDIGO 2012 pairs vasopressors with fluids in vasomotor shock (1C); the 2026 draft suggests a MAP above 65 mm Hg (2C), perhaps higher with chronic hypertension. In SEPSISPAM (Asfar 2014), chronically hypertensive patients treated to 80–85 mm Hg needed less RRT without a survival difference, but the higher target raised atrial fibrillation overall; SSC 2026 accepts 60–65 at age 65 or older (Sepsis page).

Fluid: balanced crystalloid cut major adverse kidney events from 15.4% to 14.3% in SMART (15,802 ICU adults); the 2026 draft recommends it over saline except in traumatic brain injury (1B), and crystalloid over albumin, gelatin or starch (1B). Excess fluid harms: in PICARD (Bouchard 2009; 618 patients), overload above 10% of baseline weight at dialysis start doubled the adjusted odds of death (OR 2.07), and overloaded patients not dialyzed recovered less often.

Abdominal pressure (WSACS 2013): intra-abdominal hypertension is a sustained IAP of 12 mm Hg or more (grade I 12–15, II 16–20, III 21–25, IV over 25); abdominal compartment syndrome is a sustained IAP over 20 with new organ dysfunction; abdominal perfusion pressure is MAP minus IAP. Measure through the bladder with no more than 25 mL of sterile saline, supine, at end-expiration, zeroed at the midaxillary line, whenever a risk factor such as massive resuscitation is present (1C).

• Analgesia and sedation; a brief trial of neuromuscular blockade
• Nasogastric or rectal decompression; avoid a positive fluid balance
• Percutaneous drainage of intraperitoneal fluid
• Decompressive laparotomy for overt abdominal compartment syndrome

Cirrhosis: hepatorenal AKI needs albumin plus a vasoconstrictor, terlipressin or norepinephrine (2026 draft 1C; GI and Liver page).

Nephrotoxins: antibiotics, contrast and pigment

Vancomycin toxicity tracks exposure: the 2020 ASHP/IDSA/PIDS/SIDP consensus targets an AUC of 400–600 mg·h/L because AUC-guided dosing lowered nephrotoxicity. In the 2026 draft review, AUC dosing carried an AKI relative risk of 0.68, continuous infusion an odds ratio of 0.47, and pharmacist-led monitoring a relative risk of 0.59.

Adding piperacillin-tazobactam raised AKI in a meta-analysis of 24,799 patients (22.2% vs 12.9%; OR 2.68 vs vancomycin with cefepime or a carbapenem), but not significantly against those comparators in 968 critically ill patients (OR 1.43; Luther 2018). Part is pseudo-AKI from blocked creatinine secretion: in 739 ICU patients, creatinine-defined AKI rose (RR 1.34) without changes in cystatin C, dialysis or death (Miano 2022). ACORN (2023; 2,511 adults, 77% on vancomycin) found no difference in stage 3 AKI or death with cefepime vs piperacillin-tazobactam (7.0% vs 7.5% and 7.6% vs 6.0%); cefepime caused more delirium and coma.

Aminoglycosides enter proximal tubular cells through megalin; AKI follows 5–25% of courses, classically nonoliguric after 5–7 days (Chou 2022). KDIGO 2012 suggests avoiding them when an alternative exists (2A), once-daily dosing (2B) and levels for courses beyond 48 h (2C), with the trough undetectable by 18–24 h. Other culprits: NSAIDs with an ACE inhibitor or ARB plus a diuretic (the triple whammy), conventional amphotericin, acyclovir crystals, colistin, calcineurin inhibitors, and interstitial nephritis, about 70% drug-induced (2026 draft).

Contrast: many post-contrast creatinine rises have other causes (2026 draft). ACR–NKF (2020) calls the risk overstated and gives prophylactic IV saline for AKI or eGFR under 30 not on dialysis, considering it at 30–44 in high-risk cases. In PRESERVE (5,177 high-risk angiography patients), death, dialysis or a persistent creatinine rise at 90 days was 4.4% with bicarbonate vs 4.7% with saline, and 4.6% with acetylcysteine vs 4.5% with placebo. The 2026 draft recommends periprocedural 0.9% saline for high-risk adults without overload (1B), the lowest dose, no deferral of needed studies, and stopping metformin for AKI or eGFR 30 or less until function is stable at 48 h.

Pigment: myoglobin and hemoglobin precipitate in tubules, release free iron and constrict renal vessels. In 2,083 trauma ICU patients, renal failure followed CK above 5,000 U/L in 19% vs 8% below it, and bicarbonate with mannitol changed neither renal failure, dialysis nor death (Brown 2004); the 2026 draft supports IV crystalloid and advises against forcing urine with mannitol or loop diuretics. AAST 2022 targets urine 1–3 mL/kg/h, up to 300 mL/h (Trauma page: 200–300); burn protocols 75–100 mL/h (Burns page). If urine stays low, more fluid only overloads; RRT may be needed. Expect early high potassium and phosphate with low calcium, later hypercalcemia, and compartment syndrome.

Prevention bundles and the furosemide stress test

KDIGO 2012 manages risk by stage: stop nephrotoxins, secure volume and perfusion pressure with functional monitoring, follow creatinine and urine output, avoid hyperglycemia and consider alternatives to contrast. In cardiac surgery patients with TIMP-2·IGFBP7 above 0.3, this bundle cut AKI from 71.7% to 55.1% (PrevAKI, Meersch 2017); the 2026 draft recommends it for high-risk surgical patients (1B) and drug stewardship for anyone at risk (1C).

What fails: low-dose dopamine raises urine output without preventing AKI or death (KDIGO 2012, 1A against), and fenoldopam and natriuretic peptide are not advised. Diuretics neither prevent (1B) nor treat AKI (2C) nor shorten RRT (2B), but remain first-line for overload, as boluses first and before extracorporeal removal (2026 draft, 1C). In cardiac surgery, IV amino acids (2 g/kg ideal weight daily for up to 3 days) cut AKI from 31.7% to 26.9% and stage 3 AKI from 3.0% to 1.6% (PROTECTION 2024; draft 2A).

The furosemide stress test probes tubular function, since furosemide must be secreted by proximal tubules to act. Protocol (Chawla 2013): AKIN stage 1–2, not volume depleted, baseline eGFR 30 or more; furosemide 1.0 mg/kg IV, or 1.5 mg/kg after a loop diuretic in the past 7 days; hourly urine for 6 h, replaced mL for mL with lactated Ringer’s or saline. Under 200 mL in 2 h predicted stage 3 with 87.1% sensitivity and 84.1% specificity (AUC 0.87; 77 patients), beating urinary biomarkers in the same cohort (Koyner 2015).

Later studies were less consistent: the 2026 draft rates its accuracy low to moderate across 12 studies, yet suggests it for euvolemic or hypervolemic adults with stage 1–2 AKI (2C). In the FST trial (Lumlertgul 2018), 6 of 44 responders (13.6%) needed RRT, and early RRT for non-responders did not lower 28-day mortality (62.1% vs 58.3%). The result is prognostic, not a treatment.

Medical management and drug dosing

Hyperkalemia threatens first: with ECG changes give IV calcium, shift with insulin and dextrose, and remove potassium with a binder, a loop diuretic if there is urine, or dialysis (Electrolytes page). Remove potassium from fluids and feeds, and review ACE inhibitors, ARBs, spironolactone, trimethoprim and heparin.

Acidosis: in BICAR-ICU (Jaber 2018; 389 patients, pH 7.20 or less), 4.2% bicarbonate to pH 7.30 did not significantly change overall 28-day survival but raised it from 37% to 54% in the prespecified AKIN 2–3 stratum. BICARICU-2 (Jung 2025; 627 patients with pH 7.20 or less and stage 2–3 AKI) found no difference in 90-day mortality (62.1% vs 61.7%) but less kidney replacement therapy (35% vs 50%). The 2026 draft suggests bicarbonate over RRT for pH under 7.20 without another urgent indication (2C).

Nutrition (KDIGO 2012): 20–30 kcal/kg/day; protein 0.8–1.0 g/kg/day without dialysis, 1.0–1.5 on intermittent RRT, up to 1.7 on CRRT or when hypercatabolic; enteral first; never restrict protein to delay dialysis. The 2026 draft warns against excess (2.2 g/kg/day or more raised mortality with AKI in one ICU trial) and accepts glucose of 140–180 mg/dL (KDIGO 2012: 110–149).

Steady-state equations (Cockcroft-Gault, CKD-EPI) overestimate GFR while creatinine climbs and underestimate it in recovery; the 2026 draft says to dose from creatinine and urine trends and to consider kinetic eGFR (2B). Loading doses stay full because resuscitation expands distribution, and admission AKI resolved within 48 h in 57.2% of infected patients (Crass 2019), so many pharmacists delay cutting wide-margin antibiotics.

Renal replacement therapy: indications and timing

KDIGO 2012 starts RRT emergently for life-threatening fluid, electrolyte and acid–base changes and otherwise weighs trends over single BUN or creatinine values. The trials gave AEIOU numbers: AKIKI started for BUN over 112 mg/dL, potassium over 6 mmol/L (5.5 despite treatment), pH under 7.15, pulmonary edema needing over 5 L/min of oxygen or FiO₂ over 50%, or oliguria beyond 72 h; STARRT-AKI used potassium 6.0 or more, pH 7.20 or less or bicarbonate 12 or less, PaO₂/FiO₂ 200 or less with overload, or AKI persisting 72 h.

Early starts have costs: catheter bloodstream infections 10% vs 5% (AKIKI), and adverse events 23.0% vs 16.5% with more dialysis dependence (STARRT-AKI). Practice now plans early, starts for an indication or persistent severe AKI, and does not drift past AKIKI triggers. The 2026 draft recommends deferred over pre-emptive initiation (1C) and suggests CRRT or acute peritoneal dialysis over intermittent HD for instability (2C); see the CRRT page.

Nursing care, recovery and follow-up

• Renal: hourly urine output against weight; keep a catheter only while hourly measurement guides care (CDC 2009 lists it for critical illness); bladder scan before catheterizing for retention.
• Fluid: total every drip, flush and feed; daily weight on the same scale; report the cumulative balance at handoff.
• Cardiac and respiratory: continuous ECG while potassium is high or rising; crackles, JVD, an S3 and rising oxygen need signal overload.
• Neuro: uremic or drug-related encephalopathy (confusion, asterixis, myoclonus, seizures), especially with cefepime, opioids or gabapentinoids.
• Hematologic and skin: uremic bleeding; dialysis catheter care; edematous skin breaks down.
• Medications: reconcile daily; no NSAIDs; renal doses and levels; question contrast and nephrotoxin orders.

Relieving an obstruction can unleash post-obstructive diuresis: over 200 mL/h for 2 consecutive hours or over 3 L in 24 h. Record output every 2 h, check potassium, magnesium, phosphate and creatinine every 12–24 h, and replace part of the output as ordered, for example 75% of the previous hour with 0.45% saline (Halbgewachs 2015). Recovering ATN can behave the same way, with urine output rising before creatinine falls.

AKI raises later risks of CKD, cardiovascular events and death (ADQI 28). KDIGO 2012 asks for evaluation 3 months after AKI; the 2026 draft recommends coordinated follow-up (1B), restarting RAS inhibitors when indicated (1B), discharge summaries that name the AKI episode, and sick-day teaching: pause diuretics, RAS inhibitors, NSAIDs and metformin during vomiting, diarrhea or fever, with a clear restart plan (2D).

Exam pearls

• KDIGO AKI: creatinine up 0.3 mg/dL in 48 h or to 1.5 × baseline in 7 days, or urine under 0.5 mL/kg/h for 6 h; stage by the worse criterion.
• Prerenal: FENa under 1%, urine Na under 20, specific gravity over 1.020, BUN:creatinine over 20:1, bland sediment. ATN: FENa over 2%, muddy brown casts, specific gravity fixed near 1.010.
• On diuretics, use FEUrea: under 35% prerenal, over 50% ATN.
• RBC casts mean GN; WBC casts mean interstitial nephritis or pyelonephritis; a heme-positive dipstick without RBCs means myoglobin.
• Sudden anuria: think obstruction or a blocked catheter first; bladder scan, flush, ultrasound.
• Contrast prevention is isotonic IV fluid for high-risk patients; acetylcysteine and bicarbonate add nothing; hold metformin with AKI or eGFR 30 or less.
• Rhabdomyolysis: CK over 5,000, tea-colored urine; early isotonic fluid to a high urine output; potassium is the early killer.
• No renal-dose dopamine; diuretics treat overload, not AKI; a furosemide stress test under 200 mL in 2 h predicts progression.
• Dialysis follows AEIOU; starting early on a stage alone does not lower mortality (AKIKI, IDEAL-ICU, STARRT-AKI).
• Intra-abdominal hypertension is IAP 12 or more; compartment syndrome is IAP over 20 with new organ dysfunction; measure via the bladder with 25 mL or less.