Critical Care Nephrology · Two-Week Intensive · Lecture 1 of 9

AKI Essentials

Definitions, staging, mechanism & epidemiology of acute kidney injury in the ICU

40 minutes Nephrology Fellows Week 1

Koyner, Handbook of Critical Care Nephrology (2021) · Ronco, Critical Care Nephrology 3e · NTUH Yunlin Branch

Learning objectives

By the end of this session you will be able to…

Stage AKI with KDIGO 2012 and name the assay, weight and baseline errors that systematically mis-stage ICU patients.
Separate transient from persistent AKI and place the patient on the AKI → AKD → CKD continuum.
Explain septic AKI at the level of microcirculation, PAMP signalling and cell-cycle arrest — not "ischaemia".
Quote the operating characteristics (and failure modes) of FeNa, FeUrea, urine microscopy and the furosemide stress test.
Use the epidemiology — adjusted mortality by stage, long-term CKD/ESKD hazards — to argue for urgency at the bedside.

Koyner Ch 5–7 · Ronco Ch 11, 13, 16, 55 · KDIGO 2012 AKI Guideline · ADQI 16 (Chawla, Nat Rev Nephrol. 2017)

The burden — and why the number matters

Not "the creatinine is a bit up". A graded, dose-dependent, independently-adjudicated predictor of death and of the next decade of kidney disease.

Koyner Ch 6 · Ronco Ch 13 · AKI-EPI 2015 · Kellum JASN 2015 · See KI 2019

Incidence and the dose–response with death

57.3%of ICU patients, week 1 (AKI-EPI)

21.6%pooled, hospitalised adults

5–11%receive KRT

51.1%mortality, stage 3 on both axes

KDIGO stage	Adjusted OR, hospital death	Read it as
Stage 1	1.68, p = 0.11	Not significant after adjustment
Stage 2	2.95, p = 0.005	The inflection point
Stage 3	6.88, p < 0.001	Independent of APACHE

The honest caveat

Cross-sectional, first ICU week — prevalence, not incidence. Rest "mild AKI kills" on duration, not on stage 1.

Hoste EAJ et al. Intensive Care Med. 2015;41:1411 (AKI-EPI) · Susantitaphong P et al. Clin J Am Soc Nephrol. 2013;8:1482 · Kellum JA et al. J Am Soc Nephrol. 2015;26:2231 · Koyner Ch 6

The bill arrives after discharge

Outcome after AKI	Pooled HR (See 2019)
New or progressive CKD	2.67
ESKD	4.81
Death	1.80

Graded by stage — and the risk persists when Cr "normalises"
Highest after dialysis-requiring AKI; repeat episodes compound it
+86% CV mortality, +58% heart failure, +40% acute MI
Only ~5% of survivors ever see a nephrologist

What your discharge summary must carry

Peak stage, cause, KRT, and the baseline used
Nephrotoxins stopped — when to restart RASi/SGLT2i
Cr + urine ACR at 3 months
Named follow-up: stage 3, KRT-treated, unrecovered

See EJ et al. Kidney Int. 2019;95:160 · Odutayo A et al. J Am Soc Nephrol. 2017;28:377 · Koyner Ch 6, 51 · Ronco Ch 23, 29

Definitions, staging & phenotype

Twenty years of consensus criteria built on two indirect, confounded signals — and the 2026 attempt to fix them.

Koyner Ch 5, 7 · Ronco Ch 11, 22 · KDIGO 2012 · KDIGO 2026 draft

Two decades of consensus — and what changes next

Year	System	Key innovation	Legacy problem it created
2004	RIFLE (ADQI, Bellomo)	First consensus; ≥25% GFR fall; UO axis	Needs a baseline ICU patients lack
2007	AKIN (Mehta)	48-h rise ≥0.3 mg/dL; 3 stages	Assay noise became "AKI"
2012	KDIGO	Merged: 48-h absolute or 7-day relative	Excretory function only — no structure
2017–20	ADQI 16 / KDIGO nomenclature	AKD; transient vs persistent; recovery	Almost never coded
2026	KDIGO AKI/AKD update	One AKI+AKD frame; damage biomarkers; e-alerts	Draft — not policy yet

Bellomo R et al. Crit Care. 2004;8:R204 · Mehta RL et al. Crit Care. 2007;11:R31 · KDIGO AKI Work Group. Kidney Int Suppl. 2012;2:1–138 · Chawla LS et al. Nat Rev Nephrol. 2017;13:241 · KDIGO 2026 AKI/AKD public-review draft, March 2026 · Koyner Ch 5

KDIGO staging — the language of the unit

Stage	Serum creatinine	Urine output
1	1.5–1.9× baseline, or ≥0.3 mg/dL in 48 h	<0.5 mL/kg/h for 6–12 h
2	2.0–2.9× baseline	<0.5 mL/kg/h for ≥12 h
3	≥3.0× baseline, or Cr ≥4.0 mg/dL, or any KRT	<0.3 mL/kg/h ≥24 h, or anuria ≥12 h

Key point

Stage on the higher axis — but prediction is best when both are used.

Two staging traps

The ≥4.0 mg/dL route still needs an acute rise — stable CKD 5 is not stage 3. Starting KRT is stage 3 by fiat.

KDIGO AKI Work Group. Kidney Int Suppl. 2012;2:19–36 · Kellum JA et al. J Am Soc Nephrol. 2015;26:2231

Staging arithmetic at the bedside

Bedside numbers

Actual wt 118 kgIBW 72 kgBaseline Cr 0.9Day 2 Cr 2.1UO 480 mL / 12 h

2.1 ÷ 0.9 = 2.3× baseline → creatinine stage 2

480 ÷ 118 ÷ 12 = 0.34 mL/kg/h (actual) → stage 2. IBW: 0.56 → no UO criterion.

→

Stage 2 either way here — but in obesity the weight you pick decides whether AKI exists.

Three habits that prevent mis-staging

Declare the weight convention in the note — the commonest silent inconsistency
Read UO in blocks (6/12/24 h), never a single hour
State the baseline: "0.9, clinic, 4 months ago"

KDIGO 2012 · Koyner Ch 5 (limitations of urine-output criteria) · Fliser D et al. Nephrol Dial Transplant. 2012;27:4263 (ERBP position)

Urine output: earlier than creatinine, easier to corrupt

Why it earns its place

Falls hours to days before creatinine
Oliguria with a normal creatinine still predicts death and KRT
Continuous, free, charted — the only real-time renal signal

Why it fails

Diuretics uncouple flow from filtration — a low-GFR kidney can still chart 60 mL/h
Appropriate oliguria: hypovolaemia, ADH — low UO can be the kidney working
Needs a catheter, an accurate chart, an agreed weight
Non-oliguric ATN is common — normal UO never excludes AKI

Common pitfall — and why it happens

Teams anchor on the printed creatinine while UO sits on a flowsheet nobody totals. Read the 6-, 12- and 24-hour blocks before speaking.

KDIGO 2012 · Kellum JA et al. J Am Soc Nephrol. 2015;26:2231 · Macedo E et al. Kidney Int. 2011;80:760 · Thurau K, Boylan JW. Am J Med. 1976;61:308 · Koyner Ch 5

The baseline creatinine problem

KDIGO is a change definition — and in the ICU the denominator you pick decides the stage.

Best

Outpatient Cr 7–365 days before admission

Acceptable

Lowest inpatient Cr this admission — under-calls AKI already present

Last resort

Back-estimate from eGFR 75 — over-calls AKI in CKD, under-calls it in the young

Never

Admission Cr after transfer — the peak, not the floor

What the mis-call costs

True baseline 1.8 back-calculated to 0.9: instant "stage 2" — RASi stopped, surgery cancelled, a needless KRT conversation.

KDIGO 2012 · Siew ED et al. Clin J Am Soc Nephrol. 2012;7:712 · Thomas ME et al. Kidney Int. 2015;87:62 · Koyner Ch 5

Creatinine: lagging, diluted, muscle-dependent, assay-dependent

Lag

Needs 24–72 h to reach a new steady state — early injury is invisible.

Dilution

+10% total body water lowers Cr ~10% with no change in GFR.

Production

Sarcopenia, liver failure, sepsis cut generation. "Normal" 0.6 in cachexia can be GFR ~40.

Interference

Trimethoprim, cimetidine, cobicistat block secretion — Cr rises, GFR unchanged.

Kinetic eGFR

Uses the rate of change plus volume of distribution — the only creatinine tool honest on day 1.

Cystatin C

Muscle-independent; creatinine–cystatin C CKD-EPI 2021 is the most accurate. Confounded by steroids and inflammation — no free lunch in sepsis.

Chen S. J Am Soc Nephrol. 2013;24:877 (kinetic eGFR) · Inker LA et al. N Engl J Med. 2021;385:1737 · Koyner Ch 5, 16 · Ronco Ch 25 (functional biomarkers)

Poll 1 · staging under uncertainty

62 F, ICU day 3 after emergency laparotomy for perforated diverticulitis. No pre-admission bloods anywhere in the system. Admission Cr 2.6 mg/dL → today 1.9 mg/dL. She has received 9 L of crystalloid and is +7.5 kg. UO 0.45 mL/kg/h (IBW) for the last 14 h. Albumin 2.1 g/dL, CK 180 U/L.

A. No AKI — creatinine is falling · B. Stage 1 by urine output only · C. Stage 2 by urine output, and the falling creatinine is dilutional · D. Back-calculate a baseline from eGFR 75 and stage on that

Vote, then defend your answer in one sentence

KDIGO 2012 §2 (definition & staging) · Kellum JA et al. J Am Soc Nephrol. 2015;26:2231 · Koyner Ch 5

Transient, persistent, AKD, CKD — one continuum

0–48 H

Transient AKIReverses ≤48 h; usually haemodynamic

→

>48 H

Persistent AKIPersists after haemodynamics corrected; biomarker-positive

→

7–90 D

AKDAKI criteria, or GFR fall >35%, or Cr rise >50%

→

>90 D

CKDGFR <60 or damage beyond 3 months

AKD is commoner than AKI — it catches the "creeping creatinine" patient
Recovery is undefined; the 3-month creatinine is the fair one
Renal angina = risk × injury — who deserves a biomarker

Why the phenotype beats the stage

"Will this reverse?" drives every real decision. "What stage?" drives none.

Chawla LS et al. Nat Rev Nephrol. 2017;13:241 (ADQI 16) · Levey AS et al. Kidney Int. 2020;97:1117 · James MT et al. JAMA Netw Open. 2019;2:e191795 · Goldstein SL, Chawla LS. Clin J Am Soc Nephrol. 2010;5:943 · Koyner Ch 5, 51 · Ronco Ch 22

Predicting who progresses: scores, stress tests, biomarkers

Furosemide stress test — a functional read-out

Furosemide 1.0 mg/kg IV (1.5 if loop-exposed); replace losses mL-for-mL. 2-h UO <200 mL predicts stage 3. Euvolaemic patients only.

What the models can and cannot do

Renal angina index AUROC 0.74–0.81
AKI mortality models AUROC <0.7 externally
Machine-learning ~0.88 — discrimination ≠ calibration

CHAWLA 2013 · furosemide stress test

n = 77. 2-h UO <200 mL predicted stage 3: AUC 0.87, sens 87%, spec 84% — ahead of NGAL.

BigpAK-2 · Lancet 2025

n = 1176 post-surgery. KDIGO bundle cut stage 2–3 AKI at 72 h: 22.3% → 14.4%, NNT 12.

Chawla LS et al. Crit Care. 2013;17:R207 · Rewa OG et al. J Crit Care. 2019;52:109 · Zarbock A et al. Lancet. 2025;406:2782–2791 (BigpAK-2) · Meersch M et al. Intensive Care Med. 2017;43:1551 (PrevAKI) · Koyner Ch 7, 16

Mechanism & etiology

The three-compartment framework still works — provided you remember that the ICU patient has three compartments failing at once.

Koyner Ch 5, 8, 36, 40 · Ronco Ch 11, 18, 39, 45

Three compartments, one patient

Prerenal ~30%

Hypoperfusion without structural injury. Diagnosable only in retrospect.

Intrinsic ~60%

Tubule, interstitium, glomerulus or vessel — overwhelmingly acute tubular injury.

Postrenal <10%

Small in numbers, large in consequence — and the only fully reversible compartment.

The mechanistic list that actually helps

Drop the three buckets for parallel mechanisms: haemodynamic · microcirculatory · endothelial · thrombotic · inflammatory · toxic · congestive · obstructive. Name every one that is running.

The commonest consult error

Naming one cause and stopping. Sepsis plus hypovolaemia plus vancomycin plus IAP 20 is the modal ICU patient — each addend has a different fix.

Koyner Ch 5 (Table 5.3, pathophysiologic mechanisms) · Uchino S et al. JAMA. 2005;294:813 (BEST Kidney: sepsis in ~50% of severe ICU AKI) · Ronco Ch 11

Prerenal physiology — autoregulation and where it breaks

Kidney takes ~25% of cardiac output; GFR autoregulates over a textbook MAP range ~80–180, shifted right by HTN, age, CKD
Afferent tone: myogenic reflex, TGF, prostaglandins — NSAIDs abolish
Efferent tone: angiotensin II preserves filtration fraction — ACEi/ARB abolish
Perfusion pressure = MAP − CVP (or − IAP). Congestion is prerenal

Spectrum, not dichotomy

Prerenal and ATN: one process — no marker finds the transition. Declare the fluid-challenge endpoint in advance.

SEPSISPAM · NEJM 2014

n = 776 septic shock; MAP 80–85 vs 65–70; mortality NS. Chronic-HTN subgroup: less KRT (31.7% vs 42.2%), more AF.

Asfar P et al. N Engl J Med. 2014;370:1583 · Koyner Ch 5 · Ronco Ch 18 (renal blood flow and perfusion pressure)

Sepsis-associated AKI: the model that broke the model

Up to 50% of ICU AKI — and it is not ischaemia: renal blood flow is often high.

Microcirculatory failure — shunting despite adequate flow
Innate immune signalling — PAMPs engage tubular TLR2/TLR4
Cell-cycle arrest — G1 is the cell's defence; TIMP-2/IGFBP7
Metabolic downregulation — oliguria is defence, not necrosis
Endothelial injury — glycocalyx shedding, capillary leak

What follows at the bedside

More MAP and more fluid do not buy filtration. Past adequate perfusion, fluid is congestive. Source control and time are the therapies.

SSC 2026: MAP ≥65 (60–65 if ≥65 y), balanced crystalloid, no starches.

Takasu O et al. Am J Respir Crit Care Med. 2013;187:509 · Langenberg C et al. Kidney Int. 2006;69:1996 · Bellomo R et al. Intensive Care Med. 2017;43:816 · Poston JT, Koyner JL. BMJ. 2019;364:k4891 · Surviving Sepsis Campaign 2026, Crit Care Med. 2026;54:725–812 · Koyner Ch 36 · Ronco Ch 21, 27

Acute tubular injury — ischaemic and toxic, with numbers

Ischaemic ATN

Watershed is the outer medulla — S3 and thick ascending limb work at PO₂ 10–20 mmHg
Shock, arrest, cross-clamp, bypass, ECMO
Muddy-brown granular casts, RTEC; course 1–3 weeks

Nephrotoxic ATN — thresholds worth remembering

Vancomycin: AUC₂₄/MIC 400–600; trough-only dosing is obsolete
Aminoglycosides: non-oliguric AKI day 5–10, K⁺/Mg²⁺ wasting
Amphotericin deoxycholate ≫ liposomal; distal RTA first
Crystal (acyclovir, methotrexate): volume + alkalinisation
Pigment: risk above CK ~5000; target UO 200–300 mL/h

Key point

Of septic + ischaemic + nephrotoxic, you can modify only the last two. ~20% of ICU drugs are nephrotoxic.

Koyner Ch 5 (Table 5.4), Ch 8–9, Ch 49 · Rybak MJ et al. Am J Health Syst Pharm. 2020;77:835 (vancomycin AUC) · Bosch X et al. N Engl J Med. 2009;361:62 · Perazella MA. Clin J Am Soc Nephrol. 2019;14:1101

Intrinsic AKI beyond the tubule — the diagnoses you cannot afford to defer

Site	Disease	Discriminating clue	Time-critical action
Interstitium	Acute interstitial nephritis	New PPI, β-lactam, NSAID; sterile pyuria, WBC casts	Stop the drug; no recovery by day 5–7 → biopsy
Glomerulus	RPGN / vasculitis	Dysmorphic RBCs, RBC casts; low C3/C4	ANCA, anti-GBM, C3/C4 today; anuric anti-GBM rarely recovers
Small vessels	TMA	Schistocytes, ↓platelets, ↑LDH; normal PT/aPTT	ADAMTS13 before plasma exchange; PLASMIC score; eculizumab
Large vessels	Atheroemboli, embolism	Sudden anuria after instrumentation; livedo, blue toes	Recognise it and stop instrumenting

Koyner Ch 5 · Ronco Ch 47, 50 · Perazella MA, Rosner MH. Clin J Am Soc Nephrol. 2022;17:1220 · Moledina DG, Perazella MA. Clin J Am Soc Nephrol. 2017;12:2046 (AIN)

Nephrotoxins — the compartment you can actually change

Mechanism	Agents	Signature
Afferent constriction	NSAIDs, CNI, vasopressors, contrast	FeNa <1%, reversible
Efferent dilation	ACEi, ARB	Cr rise <30% is expected
Tubular toxicity	Aminoglycosides, vancomycin, amphotericin	Non-oliguric, wasting, granular casts
Interstitial nephritis	PPI, β-lactams, quinolones	Sterile pyuria, WBC casts
Intratubular crystals	Methotrexate, acyclovir, sulfonamides	Crystalluria hours after a bolus
Osmotic nephropathy	Starch, dextran, mannitol	Vacuolated proximal tubules
Thrombotic microangiopathy	CNI/mTOR, gemcitabine, VEGF	Schistocytes, normal coagulation

ACORN · JAMA 2023

n = 2511; cefepime vs pip-tazo. AKI or death day 14: OR 0.95. The pip-tazo signal is secretion interference.

Koyner Ch 5 (Table 5.4), Ch 8 · Qian ET et al. JAMA. 2023;330:1557 (ACORN) · Ronco Ch 39

Organ crosstalk — when the kidney is the messenger

Cardiorenal syndrome (Ronco types 1–5)

Congestion, not low output, drives type-1 AKI
CVP tracks worsening renal function better than cardiac index
Decongestion is the therapy — a Cr rise during effective diuresis is not injury

Hepatorenal syndrome — AKI (HRS-AKI)

ICA 2019: diagnose by KDIGO criteria — the fixed Cr ≥2.5 threshold is gone
No response to albumin 1 g/kg/day × 2 d + diuretic withdrawal
Not purely functional — inflammation causes structural injury

CONFIRM · NEJM 2021

n = 300 HRS-1; terlipressin + albumin. Reversal 32% vs 17% — but more respiratory failure and no survival benefit.

Ronco C et al. J Am Coll Cardiol. 2008;52:1527 · Mullens W et al. J Am Coll Cardiol. 2009;53:589 · Angeli P et al. J Hepatol. 2019;71:811 (ICA) · Wong F et al. N Engl J Med. 2021;384:818 · Koyner Ch 40, 42 · Ronco Ch 44, 45

The first 24 hours of the consult

A reproducible sequence — exposure history, the bedside, the sediment, the chemistries, the probe — and an explicit decision about biopsy.

Koyner Ch 5, 18 · Ronco Ch 31, 33, 55

A consult checklist that never fails

History & exposure

Volume losses — quantify, don't characterise
Hypotension: read the flowsheet. Minutes below MAP 55 predict AKI dose-dependently
Contrast timing, surgery, bypass and cross-clamp times, ECMO
Full reconciliation: NSAIDs, RASi, SGLT2i, herbal
Baseline function — ask, then verify a real lab value

Exam & bedside data

Perfusion pressure, not MAP: MAP − CVP, or − IAP
JVP, oedema, capillary refill, mottling
IAP in every distended patient — ≥12 is IAH, ≥20 with organ failure is ACS
Bladder scan and catheter flush — the only reversible cause
Skin, eyes, joints — the systemic-disease survey

Koyner Ch 5, 43 · Walsh M et al. Anesthesiology. 2013;119:507 · Kirkpatrick AW et al. Intensive Care Med. 2013;39:1190 (WSACS) · KDIGO 2012 §2

Urine microscopy — the cheapest biopsy, with real numbers

Muddy-brown granular casts + RTEC → ATN. Cast score ≥2: OR 74 — one of the largest bedside likelihood shifts
Their absence in suspected prerenal: NPV 91% — reassuring only at low pretest probability
Dysmorphic RBCs → glomerular; WBC casts → interstitial; eosinophils add nothing
Crystals: uric acid, calcium oxalate, acyclovir, sulfadiazine
Dipstick blood without RBCs → myoglobin or haemoglobin

Why this test fails in practice

Casts lyse within hours; automated "microscopy" doesn't count them. Spin a fresh specimen yourself, within the hour.

Perazella MA et al. Clin J Am Soc Nephrol. 2008;3:1615 · Chawla LS et al. Am J Kidney Dis. 2008;51:402 (cast scoring index) · Cavanaugh C, Perazella MA. Am J Kidney Dis. 2019;73:258 · Koyner Ch 5

Urine chemistries: what they promise and what they deliver

Index	"Prerenal"	"ATN"	The caveat
FeNa	<1%	>2%	Uninterpretable after diuretics; also low in contrast and HRS
FeUrea	<35%	>50–65%	Survives loop diuretics; needs an intact proximal tubule
U_Osm	>500	<350–400	Meaningless in CKD, DI, osmotic diuresis
BUN/Cr	>20:1	10–15:1	Raised by GI bleeding and steroids

DARMON 2011 · the negative study

FeUrea 41% vs 32%, AUC 0.59 — no better than chance.

Key point

Supporting evidence, never a verdict. A fluid challenge with a declared endpoint answers faster.

Darmon M et al. Crit Care. 2011;15:R178 · Vanmassenhove J et al. Crit Care. 2013;17:R234 · Ronco Ch 55 (Box 55.1) · Koyner Ch 5

Obstruction and the probe — the two minutes that change everything

Postrenal AKI: where it comes from, how it hides

Prostate, pelvic malignancy, stones, clot, neurogenic bladder
Intratubular counts too: crystals, myeloma casts
Anuria alternating with polyuria is classic — but partial obstruction can have normal UO
Post-obstructive diuresis: replace half to two-thirds, hypotonic

What the ultrasound can and cannot answer

Hydronephrosis → obstruction, but false negatives early or encased
Size & echogenicity — <9 cm with a thin cortex means chronic
Bladder — the two most useful minutes
Doppler RI >0.75 tracks persistent AKI, but is heavily confounded

Common pitfall — and why it happens

A Foley reads as "obstruction excluded", so nobody flushes it. Blocked catheters are common, silent, free to fix.

Koyner Ch 5, 18 · Ronco Ch 33 · Darmon M et al. Intensive Care Med. 2011;37:68 (resistive index)

Putting it together — and when to biopsy

Feature	Prerenal/transient	ATI
Volume response	Falls in 24–72 h	No response; may worsen
Sediment	Bland, hyaline; 0	Granular casts + RTEC; ≥2
Damage biomarkers	Negative	Positive
Course	Hours to days	1–3 weeks; some KRT

Biopsy changes management when…

Active sediment — RPGN, vasculitis, anti-GBM
No plausible haemodynamic or toxic explanation
No recovery beyond 3–4 weeks
AIN; culprit drug can't be stopped

Weigh against

Platelets <50, uncorrectable coagulopathy, uncontrolled HTN
Bleeding ~1–2%; transjugular if risk high

Ask this before booking

Will I start immunosuppression tonight if the biopsy shows what I suspect? If no, wait.

Koyner Ch 5, 16 · Ronco Ch 31 · Corapi KM et al. Am J Kidney Dis. 2012;60:62

Cases, special hosts & pitfalls

Same criteria, completely different differential — and the two cognitive errors that account for most missed diagnoses on the consult service.

Koyner Ch 38–39 · Ronco Ch 40, 42

Case · 68 M, septic shock from a urinary source, ICU day 2

Norepinephrine 0.4 µg/kg/min plus vasopressin 0.03 U/min. Receiving piperacillin-tazobactam and vancomycin (trough-dosed, no AUC). Abdomen distended and tense; 6.8 L positive since admission. Long-standing hypertension on an ARB, held on admission.

Baseline Cr 1.1Cr 2.4 mg/dL / 24 hUO 0.4 mL/kg/h × 8 hMAP 62CVP 16Bladder pressure 19 mmHgLactate 3.1pH 7.28 / pCO₂ 34 / HCO₃ 16K 5.1Plt 148

Stage it. Then name the four mechanisms operating simultaneously — and the one number on this slide that changes your haemodynamic target.

Think 60 seconds · answer on the next slide

Koyner Ch 5, 36, 43 · Asfar P et al. N Engl J Med. 2014;370:1583 · Kirkpatrick AW et al. Intensive Care Med. 2013;39:1190 (WSACS) · Ronco Ch 49

Four mechanisms, one creatinine

Stage 2

2.4 ÷ 1.1 = 2.2× baseline; UO at 8 h not yet qualifying — not stage 3.

Sepsis

Microcirculatory injury, TLR-mediated tubular stress, G1 arrest. Source control is the therapy.

Nephrotoxins

Vancomycin to AUC-guided dosing today. ACORN clears pip-tazo; the ARB remains.

Congestion

IAP 19, CVP 16 → perfusion pressure ≈ 43 mmHg. Decompress and decongest — do not fill.

The decision fork

6.8 L positive: more fluid raises IAP and lowers perfusion pressure. Target MAP 80–85, decompress, run net-negative.

What would change the plan

Falling platelets + schistocytes → TMA. Rash + eosinophilia → AIN. Anuria → flush the catheter.

Koyner Ch 5, 8, 36, 43 · Asfar P et al. N Engl J Med. 2014;370:1583 · Kirkpatrick AW et al. Intensive Care Med. 2013;39:1190 · Ronco Ch 49, 49

Two hosts where the differential inverts

Cardiac surgery-associated AKI

Up to 45% by KDIGO; 1–2% need KRT
Bypass-specific: haemolysis, non-pulsatile flow, atheroemboli, recent contrast
Bypass duration and re-exploration are the strongest predictors
Works: delay elective surgery after contrast, avoid intraoperative hypotension, KDIGO bundle. Doesn't: dopamine, mannitol, fenoldopam

Obstetric AKI

Pregnancy creatinine is low (0.4–0.8) — a "normal" 1.0 is already substantial AKI
Preeclampsia/HELLP commonest, then haemorrhage and sepsis
TTP antepartum, complement HUS postpartum — HELLP is the great mimic
Delivery is often definitive. Magnesium is renally cleared — reduce and follow levels

Koyner Ch 38, 39 · Thakar CV et al. J Am Soc Nephrol. 2005;16:162 · Meersch M et al. Intensive Care Med. 2017;43:1551 · Fakhouri F et al. Clin J Am Soc Nephrol. 2012;7:2100–2106 · Ronco Ch 40, 42

Two errors that cost kidneys

Pitfall 1 — premature closure on "ATN"

"ATN" is the default because it is usually right — exactly what makes it dangerous. GN, TMA, interstitial nephritis and obstruction all have time-critical therapy. Before the word goes in the note: sediment yourself, platelets/LDH/haptoglobin/smear, every drug in 14 days, bladder imaged.

Pitfall 2 — trusting derived ratios in the wrong host

Every ratio has a numerator and a denominator; critical illness moves both. Liver disease and malnutrition lower BUN (false "intrinsic"); GI bleeding, steroids, catabolism raise it (false "prerenal"). Sarcopenia and sepsis flatten Cr generation; transport blockers raise it. Interpret indices in the direction the host is already biased.

Koyner Ch 5 · Perazella MA. Clin J Am Soc Nephrol. 2019;14:1101 · Ronco Ch 55

Poll 2 · commit to one mechanism

74 M, day 6 of ceftriaxone for pneumonia and day 3 of a newly started PPI. Cr 1.0 → 3.1 mg/dL. Urine output preserved at 1.1 mL/kg/h, bladder scan 30 mL, BP 145/85, afebrile since day 2. Maculopapular rash on both shins. WBC 12 ×10⁹/L with 6% eosinophils. Urine: sterile pyuria, WBC casts, protein 0.8 g/g, no dysmorphic RBCs. FeNa 0.7%. Renal ultrasound normal, RI 0.79.

A. Prerenal — the FeNa is 0.7% · B. Acute interstitial nephritis · C. Ischaemic ATN — the RI is 0.79 · D. Post-infectious glomerulonephritis

Hands up — then name the single finding that excludes each wrong answer

Koyner Ch 5 · Perazella MA, Rosner MH. Clin J Am Soc Nephrol. 2022;17:1220 · Moledina DG, Perazella MA. Clin J Am Soc Nephrol. 2017;12:2046 (AIN) · Ronco Ch 47, 50

Key takeaways

Stage with both axes; declare the baseline and the weight convention.
Ask "will this reverse?" — transient vs persistent beats peak stage.
Septic AKI is microcirculatory and inflammatory, not simple ischaemia. Fluid past adequate perfusion is congestion.
Know the numbers: cast score ≥2 → OR 74; FeUrea → AUC 0.59.
Bundles work in selected patients: BigpAK-2 cut stage 2–3 AKI 22.3% → 14.4%, with no mortality effect.
AKI is a doorway to CKD (HR 2.67) and ESKD (HR 4.81) — name the 3-month follow-up.

References & further reading

Where to go deeper

KDIGO AKI guideline. Kidney Int Suppl. 2012;2:1–138.
Hoste EAJ, et al. AKI-EPI. Intensive Care Med. 2015;41:1411.
Kellum JA, et al. J Am Soc Nephrol. 2015;26:2231.
See EJ, et al. Kidney Int. 2019;95:160.
Chawla LS, et al. ADQI 16. Nat Rev Nephrol. 2017;13:241.
Chawla LS, et al. Furosemide stress test. Crit Care. 2013;17:R207.
Zarbock A, et al. BigpAK-2. Lancet. 2025;406:2782–2791.
Qian ET, et al. ACORN. JAMA. 2023;330:1557.

Full list: (1) KDIGO 2012, Kidney Int Suppl 2012;2:1–138, with the 2026 AKI/AKD public-review draft; (2) Hoste EAJ et al., Intensive Care Med 2015;41:1411–1423; (3) Kellum JA et al., J Am Soc Nephrol 2015;26:2231–2238; (4) See EJ et al., Kidney Int 2019;95:160–172; (5) Chawla LS et al., Nat Rev Nephrol 2017;13:241–257 (ADQI 16); (6) Chawla LS et al., Crit Care 2013;17:R207 (FST); (7) Zarbock A et al., Lancet 2025;406:2782–2791 (BigpAK-2); (8) Qian ET et al., JAMA 2023;330:1557–1567 (ACORN); (9) Darmon M et al., Crit Care 2011;15:R178; (10) Perazella MA et al., Clin J Am Soc Nephrol 2008;3:1615–1619; (11) Asfar P et al., N Engl J Med 2014;370:1583–1593 (SEPSISPAM); (12) Takasu O et al., Am J Respir Crit Care Med 2013;187:509–517; (13) Koyner JL, Topf JM, Lerma EV, eds., Handbook of Critical Care Nephrology, Wolters Kluwer 2021 (Ch 5–9, 36, 38–40, 43, 51); (14) Ronco C et al., eds., Critical Care Nephrology 3e, Elsevier 2018 (Ch 11, 13, 16, 18, 22, 31, 33, 55).

Critical Care Nephrology · Two-Week Intensive

Thank you

Questions & discussion — bring a consult from this week and we will stage it, phenotype it, and argue about the fluid.

Next: Lecture 02 — Renal Hemodynamics, Shock & ICU Monitoring