AKI Staging Bridge — KDIGO 2012 Single-Axis vs KDIGO 2026 Three-Axis (C/U/B)
“KDIGO 2026” is ambiguous. Name the guideline. The KDIGO 2026 Anemia in CKD guideline is published — Kidney Int. 2026 Jan;109(1S):S1-S99, executive summary at 109(1):44-56 [16]. The KDIGO 2026 AKI/AKD guideline is not. It exists as a public review draft dated March 2026; the comment window was extended to 11 May 2026 and has closed [14]. As of 2 August 2026 a PubMed search returns no published version, and the guideline has not appeared in Kidney International. Everything in this note attributed to “2026” is drawn from that draft and can change before final publication.
Why this note exists
A learner reading the UDPA AKI lecture sees a staging table with axes labelled C, U, and B. The same learner opening the mastery AKI evaluation page or chapter 6 of the textbook sees three numbered stages and no axes at all. Nothing on either surface tells them these are the same disease described by two different guidelines a decade and a half apart. That gap is the problem this note closes.
The editorial decision is settled. Andy’s call, 1 August 2026:
“Use the 2026 guidelines. I’ll go back and edit if they change but its important to use something that is 14 years newer and based in evidence.”
So 2026 is the operative framework for teaching on this site. The 2012 system is not retired here — it is taught as the system the published world still runs on, which is a different job and still a necessary one.
Part 1 — KDIGO 2012: one axis, three stages
The 2012 guideline merged the two competing systems that preceded it. RIFLE came out of the Acute Dialysis Quality Initiative in 2004 with five tiers keyed to creatinine, GFR, and urine output [2]. The Acute Kidney Injury Network trimmed it in 2007, added the absolute 0.3 mg/dL creatinine rise, and fixed a 48-hour window [3]. KDIGO 2012 kept the AKIN 0.3 mg/dL threshold, kept the RIFLE 1.5-times-baseline ratio, extended the ratio window to seven days, and collapsed everything into three stages [1].
AKI definition (2012)
Any one of the following:
- Serum creatinine rise of at least 0.3 mg/dL (26.5 µmol/L) within 48 hours
- Serum creatinine rise to at least 1.5 times baseline, known or presumed within the prior 7 days
- Urine output below 0.5 mL/kg/h for 6 hours
Staging (2012)
| Stage | Serum creatinine | Urine output |
|---|---|---|
| 1 | 1.5-1.9 × baseline, or ≥0.3 mg/dL absolute rise | <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 rise to ≥4.0 mg/dL, or start of RRT, or (age <18) eGFR fall below 35 mL/min/1.73 m² | <0.3 mL/kg/h for ≥24 h, or anuria ≥12 h |
Provenance note: the 2012 table above is reproduced from the vault summary of the 2012 guideline and matches the tables carried on this site’s textbook and mastery pages. The pediatric eGFR clause in stage 3 was not re-verified against the primary Kidney Int Suppl document, which has no PubMed record and is not held locally. The 2026 tables in Part 2 were read directly from the primary draft PDF.
One rule governs the whole table: stage by the worse criterion. Creatinine at stage 2 with urine output at stage 3 is a stage 3 patient. This is the single most tested feature of the 2012 system and the one students most often invert.
Stage 1 is the mildest stage and it is not benign. Chertow studied 19,982 admitted adults, 9,210 of whom had two or more creatinine values. After adjustment for age, sex, admission diagnosis, severity of illness, and CKD, a creatinine rise of ≥0.5 mg/dL carried a 6.5-fold increase in the odds of death (95% CI 5.0-8.5), a 3.5-day longer stay, and nearly $7,500 in excess cost [4]. Large rises were rare; modest ones were common and did the damage. Do not round a small delta away.
Where the 2012 system still governs
This is the part that keeps 2012 alive, and it is not sentiment.
- The published literature is indexed to 2012. Essentially every AKI trial, cohort, registry, and meta-analysis of the last decade defines its endpoint as KDIGO 2012 stage 2 or stage 3. When you read a KRT-timing trial, the entry criterion is a 2012 stage. Translating those results forward is your job as the reader, not the trialist’s.
- Point-of-care evidence tools have not moved. A query run against OpenEvidence on 2 August 2026 for the three-axis system returned this: “A ‘KDIGO 2026’ three-axis AKI staging system is not identifiable in the retrieved literature. The operative KDIGO document remains the 2012 Clinical Practice Guideline for AKI.” That is the expected answer for a guideline that has not been published, and it is what a student will see if they check your teaching against a database.
- Examinations follow published guidelines. A guideline still in draft has no published version for a question writer to key an answer to. Treat 2012 as the exam answer until the final AKI/AKD guideline appears in Kidney International and question banks catch up — a lag measured in years, not months. (This item is editorial judgment about exam behavior, not a sourced claim; no examination body has been asked.)
- Institutional protocols, e-alerts, and quality measures are hard-coded to it. Best-practice advisories, sepsis bundles, and AKI e-alerts fire on 2012 creatinine logic. Measure denominators do not change because a draft appeared.
Part 2 — What 2026 changes, and why
Three structural changes. Each one exists because the 2012 system had a known blind spot.
Change 1 — Structure joins function in the definition
KDIGO 2012 defined AKI entirely by function: creatinine and urine output, both filtration markers. The 2026 draft adds a structural criterion — an elevated, clinically validated, regulatory-qualified biomarker of kidney damage within the prior 7 days (Practice Point 1.1.1, Table 1). It also adds cystatin C at 1.5 times baseline as a functional criterion, with Recommendation 1.1.1 (2B) suggesting cystatin C where creatinine is less accurate.
The problem being solved: creatinine is a late, insensitive, biased marker. It rises only after a substantial fraction of filtration is already lost, it lags injury by 24-48 hours, and it is a function of muscle mass, so it under-reads injury in the cachectic, the elderly, and the cirrhotic. Cystatin C does not depend on muscle mass. The CKD Prognosis Consortium meta-analysis pooled 11 general-population studies (90,750 participants) and 5 CKD cohorts (2,960 participants) with standardized creatinine and cystatin C. Cystatin C-based eGFR found a prevalence of eGFR below 60 of 13.7% against 9.7% by creatinine, and — the point that matters — wherever cystatin C reclassified a person to a higher eGFR their risk was lower, and wherever it reclassified them lower their risk was higher. Net reclassification improvement was 0.23 (95% CI 0.18-0.28) for death and 0.10 (95% CI 0.00-0.21) for end-stage renal disease [11]. Note the ESRD interval touches zero.
The damage-biomarker case rests on two lines of evidence. First, patients who are biomarker-positive but creatinine-negative are not well. Pooling 2,322 critically ill patients across 10 prospective studies, the NGAL-positive/creatinine-negative group (445 patients, 19.2% of the sample) went on to RRT in 2.5% of cases against 0.0015% in the doubly negative group (OR 16.4, 95% CI 3.6-76.9), with hospital mortality of 12.4% against 4.8% and longer ICU and hospital stays [7]. These are observational pooled data, not a trial. Second, cell-cycle-arrest biomarkers predict what creatinine has not yet shown. Urinary TIMP-2 × IGFBP7 was derived and then validated in the Sapphire cohort — 744 enrolled, 728 analyzed — for moderate-to-severe AKI (KDIGO stage 2-3) within 12 hours, reaching AUC 0.80 where no previously described marker exceeded 0.72 [8]. A 420-patient study then confirmed it against AKI adjudicated by three nephrologists masked to the test result [9]. The ADQI consensus conference formalized the functional-versus-damage framework that the 2026 draft now encodes [10].
A damage biomarker tells you an injury signal is present. It does not tell you the cause, it does not tell you the treatment, and no trial has yet shown that acting on a biomarker-only signal changes patient outcomes. B1 buys you attention and time. It does not buy you an intervention.
Change 2 — Three independent axes replace one composite stage
Practice Point 1.2.1 stages severity by individual components or combinations of creatinine, urine output, and damage biomarker (Table 6):
| Axis | Level | Criteria |
|---|---|---|
| C — serum creatinine | C0 | Does not meet creatinine criteria |
| C1 | ≥0.3 mg/dL (26.5 µmol/L) rise, or 1.5-1.9 × baseline | |
| C2 | 2-2.9 × baseline | |
| C3 | ≥3.0 × baseline, or rise to ≥4.0 mg/dL (353.6 µmol/L), or initiation of RRT | |
| U — urine output | U0 | Does not meet urine-output criteria |
| U1 | <0.5 mL/kg/h for 6-12 h | |
| U2 | <0.5 mL/kg/h for >12 h | |
| U3 | <0.3 mL/kg/h for >24 h, or anuria for >12 h | |
| B — damage biomarker | B0 | Negative |
| B1 | Positive |
The design intent is stated in the table footnote: the relevant components can be used to classify severity according to the measurements actually collected for that individual. A ward patient with daily chemistries and no hourly urine records is a C-something / U-unknown patient, and the notation says so instead of silently defaulting to the creatinine stage.
Kellum’s analysis of 32,045 critically ill patients, of whom 23,866 (74.5%) developed AKI, is the evidence that the axes carry different information. In-hospital mortality and RRT rates ran from 4.3% and 0% with no AKI to 51.1% and 55.3% when creatinine and urine output both reached stage 3. Outcomes were worse at every stage when both criteria were met rather than one, and duration of AKI predicted long-term outcome independently of severity [5]. Collapsing that into one number, as 2012 does, discards a signal that was already measured.
The 2026 draft also tightens the urine-output wording: Table 1 specifies a mean urine volume of less than 0.5 mL/kg/h based on ideal body weight for at least 6 hours. Neither qualifier appears in the 2012 table as reproduced across this site.
Subclinical AKI — normal creatinine, normal urine output, positive damage biomarker — is a stageable condition in 2026 and does not exist at all in 2012. This is the one cell of the crosswalk with no counterpart in either direction. When a page teaches subclinical AKI, it has already left the 2012 framework whether it says so or not.
Change 3 — The AKI-to-AKD continuum
The 2012 guideline stopped at 7 days and had little to say afterward. It recommended evaluating kidney function at 3 months and managing persistent impairment as CKD, and that was the extent of the framework. The gap mattered: Coca’s meta-analysis of 13 cohort studies found AKI associated with a pooled adjusted hazard ratio of 8.8 (95% CI 3.1-25.5) for incident CKD, 3.1 (95% CI 1.9-5.0) for ESRD, and 2.0 (95% CI 1.3-3.1) for death, graded by AKI severity [13]. Read the CKD interval before you quote the point estimate — it spans an order of magnitude. Patients were falling into a window the guideline did not name.
The 2026 draft names it, following the ADQI 16 consensus report [12]:
| Term | Definition (2026 draft) |
|---|---|
| Transient AKI | Rise in creatinine or cystatin C, or reduced urine output, lasting ≤48 h |
| Persistent AKI | The same, lasting >48 h and up to 7 days |
| AKD | Any of: AKI by functional or structural criteria; GFR <60 mL/min/1.73 m²; GFR fall ≥35 mL/min/1.73 m² from baseline; creatinine rise >50%; or a marker of kidney damage (albuminuria, hematuria, leukocyturia) — duration ≤3 months |
| AKI complete resolution | Creatinine or cystatin C <1.2 × baseline within 7 days |
| AKI partial resolution | Creatinine or cystatin C ≥1.2 to <1.5 × baseline within 7 days |
| AKD complete resolution | Creatinine or cystatin C <1.2 × baseline, or eGFR >80% of baseline, within 3 months |
| AKD partial resolution | Creatinine or cystatin C ≥1.2 to <1.5 × baseline, or eGFR <80% to >66% of baseline, within 3 months |
| Recurrent AKI | A new AKI episode after partial or complete resolution of a previous one |
Practice Point 1.3.2 draws a distinction worth teaching on its own: AKD without preceding AKI is a different problem from AKD following AKI, with different causes and different management. A patient discovered at eGFR 40 with no prior creatinine is not the same patient as one who fell to 40 after a documented septic AKI.
The transient/persistent split is not cosmetic. Uchino studied 20,126 hospitalized patients; 3,641 (18.1%) met RIFLE criteria for AKI, and of the 1,600 who recovered in hospital, 1,172 (73.3%) did so within 3 days. That transient group — the creatinine bump that came back down — still carried an adjusted odds ratio for hospital mortality of 2.26 (95% CI 1.85-2.76) against patients with no AKI at all [6]. “It resolved” is a description, not reassurance. AKD as a defined entity has also proven identifiable in prospective multi-country data: screening 4,311 at-risk adults presenting to facilities in Bolivia, Brazil, South Africa, and Nepal confirmed AKD in 2,922 (67.8%), and the authors’ closing point was the rate of persistent kidney disease and death after discharge — the AKD window doing exactly what the category was created to make visible [15].
Part 3 — The crosswalk
Translating a 2026 code to a 2012 stage
Rule: take the higher of C and U. Discard B. That number is the 2012 stage. If both C and U are 0, the patient has no 2012 stage — 2012 does not classify them as having AKI.
| 2026 code | 2012 stage | What is lost in translation |
|---|---|---|
| C0 / U0 / B0 | No AKI | Nothing |
| C0 / U0 / B1 | No AKI — no 2012 address exists | Everything. The entire finding disappears. |
| C1 / U0 / B0 or B1 | Stage 1 | Which axis drove it; the damage signal |
| C0 / U1 / B0 or B1 | Stage 1 | Which axis drove it; the damage signal |
| C1 / U1 / B0 or B1 | Stage 1 | That both axes were positive — the worse-prognosis group [5] |
| C2 / U0 or U1 / B0 or B1 | Stage 2 | Axis attribution; damage signal |
| C1 / U2 / B0 or B1 | Stage 2 | That creatinine was only stage 1 |
| C2 / U2 / B0 or B1 | Stage 2 | Dual-axis involvement |
| C3 / any U / any B | Stage 3 | Axis attribution; damage signal |
| Any C / U3 / any B | Stage 3 | That creatinine may still be stage 1 |
Translating a 2012 stage to a 2026 code
This direction is lossy and cannot be completed from the stage alone. A 2012 “stage 2” does not record whether creatinine or urine output produced it, and it records nothing at all about biomarkers. You have to go back to the chart.
| 2012 stage | 2026 code | What you must retrieve |
|---|---|---|
| Stage 1 | C1 and/or U1, B unknown | Which criterion was met; whether a damage biomarker was drawn |
| Stage 2 | C2 and/or U2 (or C1/U2, or C2/U1), B unknown | Same |
| Stage 3 | C3 and/or U3, B unknown | Same, plus whether RRT initiation was the trigger |
| Not staged | May still be C0/U0/B1 | Whether a biomarker was ever sent |
Forward translation (2026 → 2012) is mechanical: max(C, U). Backward translation (2012 → 2026) requires the primary record. That asymmetry is why reading a 2012-defined trial and applying it to a 2026-coded patient is safe, while reading a 2026-coded patient and hunting for matching trial evidence is not — the trial never measured the axis you are staging on.
Worked examples
Case A. Post-op day 1 after aortic surgery. Creatinine 0.9 → 1.3 mg/dL (baseline 0.9). Urine output 0.4 mL/kg/h over the last 8 hours. Urinary TIMP-2 × IGFBP7 elevated. → 2026: C1 / U1 / B1. → 2012: stage 1. The 2012 label conceals dual-axis involvement and a positive damage signal in a patient the biomarker literature would flag for imminent progression [8][9].
Case B. Cirrhotic. Creatinine 0.7 → 0.9 mg/dL, urine output preserved, cystatin C 1.5 × baseline. → 2026: AKI met by the cystatin C functional criterion (Rec 1.1.1, 2B). → 2012: not AKI. Low muscle mass is the exact scenario where creatinine under-reads [11].
Case C. ICU day 3. Creatinine 1.0 → 3.4 mg/dL, urine output 0.2 mL/kg/h for 30 hours, no biomarker sent. → 2026: C3 / U3 / B unknown. → 2012: stage 3. Here the two systems agree and the 2012 label loses little. Severe AKI is where they converge.
Case D. Cardiac surgery. Creatinine unchanged, urine output 0.6 mL/kg/h, NGAL elevated at 6 hours. → 2026: C0 / U0 / B1 — subclinical AKI. → 2012: no AKI, no stage, no chart entry. This patient carries higher RRT and mortality risk than a doubly negative patient [7], and the 2012 system has no way to say so.
Part 4 — Which system to use when
| Setting | Use | Reason |
|---|---|---|
| Bedside teaching, UDPA lectures, this site’s content | 2026 | Andy’s editorial decision, 1 Aug 2026. The three-axis framework describes what you actually measure, and it names subclinical AKI and the AKD window that 2012 leaves unnamed. |
| Board and shelf exam answers | 2012 | Question banks key to published guidelines. The AKI/AKD guideline is not published. |
| Reading or citing a trial | 2012 | The trial’s endpoint is a 2012 stage. Read it as written; do not retro-code it. |
| Charting a stage in the record | 2012 stage, with the axis detail in the narrative | Coding, quality measures, and e-alerts run on 2012 logic. Nothing stops you writing “KDIGO stage 1 (C1/U1, TIMP-2×IGFBP7 positive)” — the stage satisfies the system, the parenthetical carries the information. |
| Deciding what to do next for a patient | Both, and neither alone | No stage in either system prescribes a treatment. A stage is a risk label. |
Chart the 2012 stage. Append the 2026 code. Stage 2 AKI (C2/U1/B1) is unambiguous to a coder, to a nephrologist, and to a student, and it survives the guideline transition without a rewrite.
Part 5 — Labelling the 2026 framework honestly
Any page on this site teaching the three-axis system carries an obligation to say what it is.
Verified status, 2 August 2026:
- The KDIGO 2026 AKI/AKD guideline is a public review draft dated March 2026. Comment closed 11 May 2026 after an extension from 27 April [14].
- No PubMed record exists for a published version. A title-field search for KDIGO AKI guidelines returns only 2012-era documents and their national commentaries.
- The three-axis staging system is a Practice Point (1.2.1) — the lowest tier of KDIGO’s own statement hierarchy, ungraded, not a GRADE recommendation. So are the AKI definition (PP 1.1.1) and the AKD definition (PP 1.3.1). The graded recommendation in that chapter is cystatin C use (Rec 1.1.1, 2B — a suggestion at moderate certainty).
- By contrast, the KDIGO 2026 Anemia in CKD guideline is fully published [16]. When a page says “KDIGO 2026” without naming the guideline, a reader cannot tell which of these two very different states applies.
Standard label for site pages using the three-axis system:
KDIGO 2026 AKI/AKD Clinical Practice Guideline — public review draft, March 2026. Comment period closed 11 May 2026; not yet published in final form. The three-axis staging system is a Practice Point, not a graded recommendation, and may change before publication. KDIGO 2012 remains the classification used in the published trial literature and in most institutional protocols.
Teaching summary
- 2012 is one axis, three stages, staged by the worse criterion. That last clause is the tested point.
- 2026 is three axes — creatinine, urine output, damage biomarker — staged independently and reported as measured.
- Forward translation is easy: max(C, U) is the 2012 stage. Backward translation needs the chart.
- C0/U0/B1 is the cell with no counterpart. Subclinical AKI exists in 2026 and is invisible in 2012.
- AKD is the new window — up to 3 months — that 2012 left unnamed, and it is where the AKI-to-CKD risk [13] accumulates.
- The AKI/AKD guideline is a draft. The anemia guideline is published. Name the guideline every time.
References
[1] Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179-84. PMID: 22890468 — the citable summary of the full guideline (Kidney Int Suppl. 2012;2(1):1-138, which carries no PubMed record).
[2] Bellomo R, Ronco C, Kellum JA, et al. Acute renal failure — definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care. 2004 Aug;8(4):R204-12. PMID: 15312219
[3] Mehta RL, Kellum JA, Shah SV, et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care. 2007;11(2):R31. PMID: 17331245
[4] Chertow GM, Burdick E, Honour M, et al. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol. 2005 Nov;16(11):3365-70. PMID: 16177006
[5] Kellum JA, Sileanu FE, Murugan R, et al. Classifying AKI by urine output versus serum creatinine level. J Am Soc Nephrol. 2015 Sep;26(9):2231-8. PMID: 25568178
[6] Uchino S, Bellomo R, Bagshaw SM, et al. Transient azotaemia is associated with a high risk of death in hospitalized patients. Nephrol Dial Transplant. 2010 Jun;25(6):1833-9. PMID: 20054022
[7] Haase M, Devarajan P, Haase-Fielitz A, et al. The outcome of neutrophil gelatinase-associated lipocalin-positive subclinical acute kidney injury: a multicenter pooled analysis of prospective studies. J Am Coll Cardiol. 2011 Apr 26;57(17):1752-61. PMID: 21511111
[8] Kashani K, Al-Khafaji A, Ardiles T, et al. Discovery and validation of cell cycle arrest biomarkers in human acute kidney injury. Crit Care. 2013 Feb 6;17(1):R25. PMID: 23388612
[9] Bihorac A, Chawla LS, Shaw AD, et al. Validation of cell-cycle arrest biomarkers for acute kidney injury using clinical adjudication. Am J Respir Crit Care Med. 2014 Apr 15;189(8):932-9. PMID: 24559465
[10] Ostermann M, Zarbock A, Goldstein S, et al. Recommendations on acute kidney injury biomarkers from the Acute Disease Quality Initiative Consensus Conference: a consensus statement. JAMA Netw Open. 2020 Oct 1;3(10):e2019209. PMID: 33021646
[11] Shlipak MG, Matsushita K, Ärnlöv J, et al. Cystatin C versus creatinine in determining risk based on kidney function. N Engl J Med. 2013 Sep 5;369(10):932-43. PMID: 24004120
[12] Chawla LS, Bellomo R, Bihorac A, et al. Acute kidney disease and renal recovery: consensus report of the Acute Disease Quality Initiative (ADQI) 16 Workgroup. Nat Rev Nephrol. 2017 Apr;13(4):241-257. PMID: 28239173
[13] Coca SG, Singanamala S, Parikh CR. Chronic kidney disease after acute kidney injury: a systematic review and meta-analysis. Kidney Int. 2012 Mar;81(5):442-8. PMID: 22113526
[14] KDIGO 2026 Clinical Practice Guideline for the Evaluation and Management of Acute Kidney Injury and Acute Kidney Disease. Public Review Draft, March 2026. Kidney Disease: Improving Global Outcomes. Comment period 31 March – 11 May 2026. No PubMed record — not yet published. Local copy in DEVONthink: x-devonthink-item://D6C12DCC-1952-4773-9506-3639A0C81E4A. Chapter 1 content in this note was read directly from that PDF (Practice Points 1.1.1-1.5.1; Tables 1, 6, 7, 9, 10).
[15] Evans RDR, Sharma SK, Claure-Del Granado R, et al. Identification and outcomes of acute kidney disease in patients presenting in Bolivia, Brazil, South Africa, and Nepal. PLoS Med. 2024 Nov;21(11):e1004495. PMID: 39541400
[16] Kidney Disease: Improving Global Outcomes (KDIGO) Anemia Work Group. KDIGO 2026 Clinical Practice Guideline for the Management of Anemia in Chronic Kidney Disease (CKD). Kidney Int. 2026 Jan;109(1S):S1-S99. PMID: 41485812 — published, unlike the AKI/AKD guideline. Executive summary: Babitt JL, Berns JS, Bozkurt B, et al. Kidney Int. 2026 Jan;109(1):44-56. PMID: 41485807
Verification note. Every PMID above was located by an NCBI E-utilities esearch query and confirmed against the esummary record for title, first author, journal, and year on 2 August 2026. Reference [14] has no PMID because the document is unpublished; its content was read from the primary PDF rather than from any summary.