Acute Kidney Injury
By the end of this post you will be able to:
- Recognize and stage acute kidney injury using the KDIGO criteria, against an established baseline creatinine.
- Distinguish pre-renal, intrinsic and post-renal injury using the bedside assessment and the urine.
- Initiate first-line ward management — fluids, nephrotoxin review, relief of obstruction — and recognize the triggers for calling nephrology.
TL;DR
What you need to know before seeing your first patient on your IM rotation with this condition
Estimated read time: 3 minutes
- What it is: An abrupt fall in kidney function over hours to days, from mild loss to complete kidney failure. Any one of three criteria diagnoses it: a creatinine rise of >26 µmol/L within 48 hours, a rise greater than 1.5× the usual baseline within the past 7 days, or urine output under 0.5 mL/kg/h for 6 hours.
- How it presents: As a number on the morning bloodwork, not a symptom. At the bedside: volume status, a palpable bladder, the medication list. The JAMA Rational Clinical Examination found no single physical examination finding reliably identifies hypovolemia; weigh several together.
- Initial tests to order: Labs — creatinine, CBC, electrolytes, urea, BNP, and where the picture suggests it, CK, SPEP/SFLC. Imaging — ultrasound KUB. Urine — urinalysis with microscopy. Unexplained AKI with a suspected renal cause leads to a biopsy.
- If unstable — indications for renal replacement, AEIOU: Acidosis · Electrolytes · Intoxication · Overload · Uremia — each meaning refractory to medical therapy. No single number triggers dialysis. Refractory hyperkalemia is the commonest on a ward — the hyperkalemia episode is eight minutes on managing it.
- If stable — stabilize first, then treat by category: Pre-renal — fluid, transfusion, pressure support; a resuscitation bolus is 500 mL of crystalloid over less than 15 minutes. Renal — treat the identified cause. Post-renal — relieve the obstruction. Across all three: hold any nephrotoxic medication — NSAIDs, ACE inhibitors and ARBs, aminoglycosides, vancomycin — and metformin, which is renally cleared rather than nephrotoxic. Dose-check the rest.
Listen
Read Around the Case
Based on KDIGO 2012 Clinical Practice Guideline for Acute Kidney Injury, Canadian Association of Radiologists Guidance on Contrast-Associated AKI (2022) and NICE NG148 (acute kidney injury, updated 2024). No Canadian guideline is dedicated to acute kidney injury; the Canadian Society of Nephrology has published a commentary on the KDIGO recommendations.
Estimated read time: 21 minutes
The AKI Sequence
Confirm it, and stage it against a baseline
Volume status · bladder scan · the medication list
The cheapest high-yield test, and the one most often skipped
Scan first. Ultrasound if no cause is found, or obstruction is likely
Strict intake/output, daily weight, repeat creatinine
1. Is It Acute, and Where Is the Injury?
Estimated read time: 6 minutes
Staging is arithmetic, and it needs a denominator. Find a previous creatinine — a prior admission, family physician bloodwork, a pre-operative panel. Where none exists, the lowest value this admission is the usual stand-in; say so in the chart.
KDIGO has been the source of these criteria since 2012. Its first major update went to public review in March 2026 and will widen the framework to acute kidney disease; until then these definitions stand.
Two things the table does not say out loud:
- Creatinine lags filtration. A normal value on the day of the insult does not mean normal kidneys; a plateau is not recovery.
- The urine criterion is unusable unless somebody measures. A patient can meet stage 1 on output alone and be recorded normal because nobody charted it
Acute, chronic, or acute-on-chronic? The trajectory usually answers it; where it does not, look for the marks chronic disease leaves:
- Small kidneys with poor corticomedullary differentiation
- A normocytic anemia
- Raised parathyroid hormone, disordered calcium and phosphate
- None is decisive alone, and chronic disease plus a new acute injury is the commonest version
A rising creatinine during decongestion is not automatically AKI
The heart failure patient whose creatinine climbs while being diuresed is the commonest confusion here, and Week 0 Day 3 met that patient two days ago. In DOSE, worsening renal function during decongestion was common and did not track with worse outcomes.
- A creatinine rising while the jugular venous pressure falls and breathing eases is behaving differently from one rising while the patient stays congested
- The question is not “has the creatinine gone up” but “is this patient still volume overloaded?” If so, the diuretic usually continues
Where is the injury?
Localizing the injury is a clinical judgement, not a single test. The history, volume assessment, bladder scan and medication list do most of the work. The urine is one input that helps point you to a bucket — cheap, quick, often not sent. NICE recommends a dipstick in everyone as soon as AKI is suspected; microscopy adds what a dipstick cannot. Read it alongside the bedside assessment, not in place of it.
Two caveats belong next to the finding, not in a footnote:
- Granular casts are not exclusive to acute tubular injury
- Scattered fine granular casts alone carry little diagnostic weight
- Whether they are seen depends on who looked
2. Pre-Renal AKI
Estimated read time: 4 minutes
Fractional excretion of sodium and urea
What it is, and how to read it. The proportion of filtered sodium the kidney fails to reabsorb. It answers one question: are the tubules still holding on to salt? Send a spot urine sodium and creatinine with paired serum values, drawn at the same time.
- FeNa below 1% — the tubules are conserving sodium, fitting a pre-renal picture. It also fits early obstruction and some glomerulonephritis, so it narrows rather than decides.
- FeNa above 2% — sodium is being wasted, which fits tubular injury. Between 1% and 2% is genuinely indeterminate.
- On a diuretic, use the fractional excretion of urea instead — below 35% suggests pre-renal, with the caveat below.
How well it performs:
- Best in the clean case. A 2022 meta-analysis found pooled sensitivity 90% and specificity 82% at the 1% threshold, rising to 95% and 91% in oliguric patients with no chronic kidney disease and no diuretics.
- Diuretics damage the specificity, not the sensitivity — 80% and 54% (95% CI 31–75) in that subgroup. Because specificity falls, it is a low FeNa — the pre-renal answer — that stops meaning much.
- The workaround is weaker than taught. In a 2007 study the fractional excretion of urea had a specificity of 33% on diuretics — it “cannot be used as an alternative tool because it lacks specificity.” A 2026 meta-analysis is kinder at 87%, with heterogeneity over 90%.
Read it in the direction above, as one input among several. The next three sections take the buckets in order.
- Which fluid — two separate questions. KDIGO recommends an isotonic crystalloid (0.9% saline, Ringer’s lactate or Plasma-Lyte) over a colloid such as albumin or starch. Whether a balanced solution beats saline is a separate question.
- Prefer a balanced solution, on a modest signal. SMART is the only trial of the three with a kidney primary outcome and it favoured balanced — major adverse kidney events 14.3% vs 15.4% (OR 0.91, 95% CI 0.84–0.99) — though no individual component reached significance, and BaSICS and PLUS, powered on mortality, found none. Large volumes of saline also cause a hyperchloremic acidosis. Saline is not an error.
- How much, and when to stop. NICE CG174 sets a resuscitation bolus at 500 mL of crystalloid over less than 15 minutes, reassessing after each. Once perfusion is restored, more fluid is not a kidney treatment.
- Why the usual suspects matter here. NSAIDs and ACE inhibitors or ARBs blunt the autoregulation that protects filtration when perfusion falls — which is how a combination tolerated for years turns dangerous the week the patient gets dehydrated. Hold any nephrotoxic medication; the drug table in section 5 has each class and its restart plan.
- Two things that are not treatments. KDIGO and NICE both advise against loop diuretics to treat AKI, except for volume overload. Low-dose dopamine carries a 1A recommendation against it — “renal-dose dopamine” has been tested and does not work.
Hepatorenal syndrome — when more fluid is the wrong answer
- In advanced cirrhosis with ascites, a rising creatinine may be a functional, pre-renal-looking injury with a bland sediment and very low urinary sodium
- The pre-renal reflex misleads here. Terlipressin with albumin improves renal function in type 1 disease, at the cost of ischemic and respiratory adverse events
- For a clerk the job is recognition, not management — call for help
3. Renal (Intrinsic) AKI
Estimated read time: 4 minutes
Renal causes divide four ways — tubular, interstitial, glomerular, vascular. These are the same four answers the urine table above points to, read from the other direction: there a finding suggests a bucket, here the bucket decides management.
Nephrotic syndrome sits awkwardly: the lesion is glomerular, but the AKI is often pre-renal, through low oncotic pressure and depleted effective arterial volume. The Intern at Work episode covers it; section 2 treats the volume problem.
Contrast-associated AKI. Practice has changed here, and an older guideline will mislead you. The Canadian position, from the Canadian Association of Radiologists:
- Never delay an emergency scan for a creatinine
- Prophylaxis only below an eGFR of 30, where intravenous hydration is preferred; above 30, none
- Do not routinely stop drugs to reduce risk, metformin excepted
- N-acetylcysteine is not recommended, nor statins started for this purpose, theophylline, ascorbic acid or allopurinol. PRESERVE found no benefit from bicarbonate or acetylcysteine, and AMACING found withholding prophylactic hydration non-inferior in high-risk patients
- No routine follow-up creatinine
- One boundary matters. This is about preventing injury before contrast. It says nothing about which drugs to hold in a patient who already has AKI — that is section 2
Acute interstitial nephritis deserves its own note, because the steroid question comes up on every ward round.
- Suspect it when the sediment shows white cells without infection and a new drug started in the preceding weeks — antibiotics, proton pump inhibitors, NSAIDs, increasingly immune checkpoint inhibitors. Rash, fever and eosinophilia are frequently absent
- Stopping the drug is the least controversial intervention
- The steroid evidence is retrospective. A 2025 systematic review included 23 studies, only three randomized, and concluded high-quality trials are still needed
- What it agrees on is timing, not dose or duration. Delay of about a month was associated with failure to return to baseline and fibrosis on repeat biopsy; in 182 cases each day of delay carried an odds ratio of 1.02 for a worse creatinine at six months, while high-dose treatment beyond three weeks added nothing
- For a clerk: stop the drug, and if steroids are given the argument is for giving them early rather than long
4. Post-Renal AKI
Estimated read time: 3 minutes
- Scan the bladder first. Minutes, no radiology, and the one investigation a student can do alone at 2 a.m.
- A high residual means outlet obstruction, and a catheter is both test and treatment — the fastest intervention available
- Then image, selectively. NICE timings: ultrasound within 24 hours when no cause is identified, within 6 hours for suspected pyonephrosis
- Do not routinely scan when the cause is already identified
- Upper tract obstruction goes to urology — immediately for pyonephrosis, an obstructed solitary kidney or bilateral obstruction, with nephrostomy or stent within 12 hours
- The three grades. Mild — pelvis and calyces dilated, pelvicalyceal pattern retained, cortex unaffected. Moderate — pyramids flatten and calyces balloon outwards (a cauliflower appearance), cortical thickness preserved. Severe — pelvis and calyces ballooned, corticomedullary differentiation lost, cortex thin
- Grading is qualitative, and nobody has agreed on it. NephroPOCUS is explicit that no system is universally accepted — “mild to moderate” is acceptable for point-of-care purposes
- Then watch for post-obstructive diuresis. Output can run to litres. Most is physiological; the pathological form causes dehydration, electrolyte derangement and, untreated, death. Incidence spans 0.5% to 52%, reflecting competing definitions
- This is where strict intake/output stops being paperwork — replacement is decided from the recorded output
Do not clamp the catheter. Draining a distended bladder in stages is taught almost everywhere, on the reasoning that rapid decompression causes bleeding and hypotension. A 1997 review found hematuria in 2–16% after quick complete relief, rarely significant, blood pressure that usually normalized, and no studies supporting clamping at all. It was then tested directly: a randomized trial of 294 men found hematuria in 11.3% with gradual drainage versus 10.5% with rapid. Drain the bladder.
5. Practical Considerations on the Ward
Estimated read time: 4 minutes
Every drug on the chart gets read again. Two questions: is it harming the kidney, and does the dose fit the kidney the patient now has?
Calling about dialysis. You will not be the one starting it, so the task is recognizing the need and making a call nephrology or critical care can act on.
- Have these ready before you dial: potassium and whether the ECG has changed; pH and bicarbonate; the creatinine and urea trend, not one value; urine output over 24 hours; oxygen requirement and fluid status; and whether dialysis fits their goals of care.
- Say which AEIOU indication is met, and what you have already tried. “Refractory” means treated and not responding — so name the treatment.
- Do not wait for a number. No creatinine or urea threshold triggers dialysis, and waiting for one is how the call gets made late.
- Nor start early to get ahead of it. STARRT-AKI, with AKIKI and AKIKI-2, found no benefit from an accelerated strategy and more dialysis dependence at 90 days.
Making the fluid record usable.
- A daily weight, same time and same scale, is often more reliable than a charted balance. Intake gets missed and incontinent losses are never measured, so balances drift
- The catheter is a decision with a downside. It relieves outlet obstruction and gives accurate hourly output; it also causes infection. Name the removal plan when you insert it — “for accurate output” is not an indefinite indication
Documenting it, and what happens after.
- Write the AKI into the discharge summary as a problem in its own right — peak creatinine, presumed cause, drugs held, which need restarting and when, and when function should next be checked
- A held ACE inhibitor with no restart plan is how a temporary hold becomes permanent. Week 0 Day 2 covers how to write it
- KDIGO advises review three months after an AKI for resolution, new chronic kidney disease, or worsening of existing disease — and that review is not automatic
- A Canadian randomized trial of a risk-guided handover raised the proportion of high-risk survivors getting recommended follow-up from 3% to 28% — the gap is organizational, not clinical
Related Podcasts
Kidney topics this post points at rather than covers.
Test Yourself
Click on your answer to reveal the explanation.
A 72-year-old man is on day 3 of an admission for cellulitis. His creatinine six months ago was 80 µmol/L; this morning it is 160 µmol/L. He feels well. His blood pressure is 128/76 mmHg and has been stable, heart rate 84, and he is afebrile. He is clinically euvolemic. His medications include ramipril, naproxen for gout, and intravenous cefazolin.
What is the most appropriate next step?
A 78-year-old man with a long history of benign prostatic hyperplasia is admitted with delirium. His creatinine is 340 µmol/L, up from a baseline of 95. A bladder scan shows 900 mL. A urethral catheter is inserted and drains 1.2 L immediately. Over the next 8 hours he passes 3.2 L of urine. His blood pressure has fallen from 132/78 mmHg on admission to 104/60 mmHg, and his heart rate is 96.
What is the most appropriate next step?
Reflect
Your patient’s creatinine has risen from 78 to 145 µmol/L over two days. You stop the NSAID, give 1 L of saline, and plan to reassess in the morning. The next day the creatinine is 190 µmol/L and there is 200 mL of urine in the bag. What did yesterday’s plan assume, and what do you do now?
Please do not include patient identifying details — no names, initials, dates, medical record numbers, or any detail that could identify a specific patient or encounter. Write about your reasoning, not about a real person.
Further Reading
- GuidelineKDIGO Clinical Practice Guideline for Acute Kidney Injury (2012)
- GuidelineCanadian Association of Radiologists Guidance on Contrast-Associated Acute Kidney Injury (2022)
- GuidelineNICE NG148 — Acute kidney injury: assessment, prevention and management
- CommentaryCanadian Society of Nephrology commentary on the KDIGO AKI guideline (AJKD, 2013)
- Infographic and podcastAcute Kidney Injury — MedicinePods infographic and Intern at Work episode 190
Week 0