MD, FASN, FISN
Professor of Medicine, Universidad Mayor de San Simon School of Medicine;
Associated Clinical Researcher, IIBISMED;
Director, AKI & CRRT Program, Division of Nephrology, Hospital Obrero No. 2 – Caja Nacional de Salud, Cochabamba, Bolivia;
Chair, ISN Fellowship Program;
Member, KDIGO
Executive Committee; Co-Chair, SLANH Acute Kidney Injury Committee.
The patient should not be viewed simply as “not yet AKI” because creatinine is near baseline. If urine output of 0.35 mL/kg/h over six hours is reliable, he already meets the U1 urine-output component. With an elevated urinary [TIMP-2] × [IGFBP7], he is also biomarker-positive (B1). Therefore, the practical question is not whether the alert is “diagnostic” by itself, but whether the combined C/U/B profile and clinical context trigger a proportional kidney-protection pathway.
(Click / Tap on Questions to Reveal Content)
Early, but proportionately. KDIGO 2026 explicitly incorporates functional criteria (SCr and urine output) and structural/stress biomarkers into AKI staging. In this vignette, the combination of sepsis, vasopressor use, oliguria, and [TIMP-2] × [IGFBP7] positivity should trigger immediate kidney-protective actions: reassess perfusion and fluid responsiveness, avoid hypotension and fluid overload, review nephrotoxins, adjust renally cleared drugs, and monitor SCr, urine output, electrolytes, and acid-base status more closely. This is not “treating a biomarker”; it is acting on a high-risk, B1-positive phenotype before creatinine catches up.
A single model may support broad triage, but KDIGO 2026 emphasizes use of externally validated models in the setting where they will be deployed. Sepsis-associated AKI, cardiac surgery–associated AKI, contrast-associated AKI, and drug-associated AKD have different susceptibilities, exposures, timing, and modifiable levers. The safest approach is either etiology-specific models or broad models with strong external validation, subgroup calibration, and continuous performance monitoring in each clinical phenotype.
Explainability should be action-oriented. The clinician should see the top three to five risk drivers, whether each is modifiable, and the linked action. For example: “oliguria + vasopressor use + sepsis + nephrotoxin exposure = high risk; recommended response: hemodynamic reassessment, nephrotoxin stewardship, renal dosing, and repeat kidney assessment.” KDIGO 2026 is clear that risk models are adjuncts to clinical judgment, not replacements for individualized assessment.
The key KDIGO 2026 message is that alerts alone are not enough. Interruptive creatinine-based alerts should not be delivered in isolation without a structured response pathway. A high-risk alert should be linked to a named responsible clinician, a brief checklist, documentation of actions taken, escalation criteria, and audit of outcomes. The “loop” must include accountability, multidisciplinary workflow, and reassessment of model performance over time.
The bundle should follow a multicomponent kidney-protection strategy: optimize hemodynamics and volume status; target adequate perfusion, generally MAP >65 mmHg while individualizing for chronic hypertension or baseline physiology; avoid or minimize nephrotoxins; ensure renal dose adjustment and therapeutic drug monitoring when appropriate; maintain appropriate glycemic control; use contrast only when needed and with preventive measures in high-risk patients; and intensify monitoring of SCr, urine output, electrolytes, acid-base status, and fluid balance. In sepsis, the bundle must not delay source control or appropriate antimicrobial therapy; instead, it should make antimicrobial use safer through dosing, monitoring, and stewardship.
KDIGO 2026 explicitly recognizes that simplified AKI risk scores can support early identification in community and low-resource settings. The LMIC translation is not “low-cost AI”; it is risk-stratified care using symptoms, exposures, urine dipstick, point-of-care creatinine when available, fluid status, blood pressure, medication review, and timely referral. Biomarkers may be useful where available, but the scalable principle is the same: identify risk early, apply a simple kidney-protection pathway, and reassess.
Yes, because AKI risk is dynamic. Sepsis trajectory, vasopressor dose, urine output, fluid balance, nephrotoxin burden, ventilation, and laboratory trends can change within hours. KDIGO 2026 supports dynamic risk prediction when it improves targeting of prevention, but it also stresses local validation, calibration, terminology harmonization, and reassessment over time to prevent model drift. A changing risk score should prompt reassessment, not automatic treatment escalation.
Probably, but the immediate clinical advance in KDIGO 2026 is the integration of functional markers, structural/stress biomarkers, and validated risk models. Multi-omics and continuous monitoring may eventually identify individual susceptibility and AKI endotypes before overt injury, but they must prove clinical utility, equity, affordability, and actionability. Prediction will change outcomes only when it reliably triggers prevention.
AKI prediction changes outcomes only when it is embedded in a prevention pathway. The future is not an alert that says, “this patient is high risk”; it is a system that reliably answers, “what should we do now, who will do it, and how will we know it worked?”
