AKI workup

Staging follows the KDIGO 2012 AKI guideline, which is still the published version. KDIGO released a 2026 AKI/AKD guideline as a public-review draft in March 2026; comment closed on 11 May 2026 and the final text had not been published as of October 2026. Where the draft differs, the tool says so rather than applying it:

• It adds cystatin C (a rise of 1.5× baseline or more within 7 days) and validated kidney-damage biomarkers as criteria, and stages by separate creatinine (C1–C3), urine-output (U1–U3) and biomarker (B0/B1) components.
• Urine output is a mean rate on ideal body weight, with stage 2 at more than 12 h and stage 3 at more than 24 h (or anuria for more than 12 h). KDIGO 2012 wrote "12 hours or more" and "24 hours or more".
• It defines transient AKI (48 h or less) versus persistent AKI (more than 48 h, up to 7 days).

AKI is present (KDIGO 2012, 2.1.1) with any of: creatinine rise of 0.3 mg/dL (26.5 µmol/L) or more within 48 h; creatinine 1.5 times baseline or more, known or presumed to have occurred within 7 days; urine under 0.5 mL/kg/h for 6 h.

Stage (Table 2): 1 — 1.5–1.9 × baseline or a 0.3 mg/dL rise; urine under 0.5 mL/kg/h for 6–12 h. 2 — 2.0–2.9 ×; under 0.5 mL/kg/h for 12 h or more. 3 — 3.0 × or more, a rise to 4.0 mg/dL (353.6 µmol/L) or more, or KRT started; under 0.3 mL/kg/h for 24 h or more, or anuria for 12 h or more. The patient takes the worse of the two stages.

A creatinine of 4.0 mg/dL counts as stage 3 only after the definition is met by the 48-hour or 7-day rule; KDIGO 2012 dropped the older "acute rise of 0.5" requirement in favour of that. The ratio is rounded to the 2 decimals shown before it is compared. The rise and the 4.0 mg/dL (353.6 µmol/L) cutoff are judged on the values exactly as typed, in the unit typed; if baseline and current were typed in different units, both are converted to µmol/L (× 88.4) and compared there. Converted mg/dL values are shown truncated, so a value under a cutoff never displays at it.

Rise over more than 7 days is not AKI by these rules. KDIGO 2012 calls a creatinine rise of more than 50% lasting under 3 months acute kidney disease (AKD); the 2026 draft keeps that criterion.

Each pair must be in the same unit; the tool converts urine creatinine to the serum unit. Conventional reading: FENa under 1% suggests prerenal physiology and over 3% an intrinsic cause; FEUrea 35% or less suggests prerenal and over 50% intrinsic (Gotfried 2012; Carvounis 2002).

• Best case. In 8 studies of oliguric patients without CKD or diuretics, a 1% FENa cutoff had pooled sensitivity 95% and specificity 91% (Abdelhafez 2022).
• Diuretics. They raise FENa in prerenal states; in 5 studies of patients on diuretics, specificity fell to 54%. FEUrea was proposed for this, but a 2024 meta-analysis found it no better overall (pooled sensitivity 66%, specificity 75% at 35%).
• CKD. FENa can exceed 1% in prerenal states with chronic kidney disease (Steiner 1984). With CKD or diuretics, pooled specificity was 66%.
• FENa under 1% does not mean prerenal. It is also reported in contrast nephropathy, pigment nephropathy (rhabdomyolysis, hemolysis), acute glomerulonephritis, acute obstruction, sepsis, heart and liver failure, burns and transplant rejection (Steiner 1984; Gotfried 2012).
• Cirrhosis. Very low values (often under 0.1–0.2%) fit hepatorenal syndrome better than tubular injury, but FENa should not be used alone (KDIGO 2026 draft).
• The KDIGO 2026 draft advises keeping these indices for selected prerenal-versus-ATI questions because of their limited overall usefulness.

Patterns summarised from the KDIGO 2026 draft (section 2.5). The draft suggests manual microscopy by trained staff when the cause is unclear (2D), because automated urinalysis is not reliable enough to identify ATI, GN or AIN. In one biopsy study, granular casts had 100% specificity and 78% sensitivity for acute tubular necrosis. Urine eosinophils are a weak test for AIN: at a 1% cutoff, sensitivity 31% and specificity 68%; at 5%, sensitivity 20% and specificity 91%. A bland sediment supports a hemodynamic cause but does not exclude early tubular injury.

"AEIOU" is a memory aid. It carries no thresholds and leaves out the word that matters most: refractory. The criteria used here come from the KDIGO 2026 draft (Table 41) and the STARRT-AKI standard arm, which allowed KRT for K⁺ 6.0 or higher, pH 7.20 or lower or bicarbonate 12 or lower, PaO₂/FiO₂ 200 or lower with volume overload, or AKI persisting 72 h or more. Even then, starting was not mandatory.

• Acidemia: the 2026 draft suggests IV bicarbonate over KRT for severe metabolic acidosis with pH under 7.20 unless another urgent indication exists (2C).
• Uremia: KDIGO 2012 advises against starting on a single urea or creatinine value, and to weigh trends and the wider clinical context instead (5.1.2).
• Timing without an urgent indication: the 2026 draft recommends a deferred strategy over early KRT (1C). In STARRT-AKI (2,927 patients), 90-day mortality was 43.9% with accelerated and 43.7% with standard initiation; dialysis dependence among survivors was 10.4% vs 6.0%, and adverse events 23.0% vs 16.5%. In AKIKI 2 (as summarised in the KDIGO 2026 draft), waiting longer still (until urea exceeded 50 mmol/L or an urgent indication appeared) was linked to higher 60-day mortality (55% vs 44%), so deferral should not become indefinite delay.
• Relative indications in the draft: persistent stage 3 AKI, severe non-kidney organ dysfunction made worse by AKI, worsening illness, an expected solute load (tumor lysis, rhabdomyolysis, hemolysis), and limited physiological reserve.