Postoperative AKI — staging, bundle, fluids, dialysis timing and limits

Examiner probes

1. Stage him — exact KDIGO bands for his creatinine and his urine, and which one governs?^[1]
2. Bundle or score — what are the four planks, and which two trials with which numbers defend them?^[18][19]
3. Fluids and pressure — RELIEF numbers, balanced-versus-saline numbers, and why neither restriction nor pressor substitution is the answer?^[20][21]
4. Dialyse today or watch — STARRT-AKI numbers, the five AKIKI triggers, the AKIKI-2 floor, and the honest place of ELAIN?^{[38][39][40][42]}
5. Albumin, vancomycin, follow-up — which albumin harms and which helps, what changes in vancomycin dosing, and why does a recovered stage-1 patient still need surveillance?^{[24][32][10][4]}

Model responses

1. Stage 1 by both doors — quote the bands, take the worse, they agree. Creatinine 1.8-fold sits inside 1.5–1.9 (stage 1) and 9-hour oliguria below 0.5 sits inside 6–12 hours (stage 1) — against the definition's three doors (0.3 mg/dL in 48 hours, 1.5-fold in 7 days, 6-hour oliguria) and the stage-2/3 cutoffs above them (2.0–2.9-fold with 12-hour oliguria; 3.0-fold or 4.0 mg/dL or dialysis with 0.3-rate oliguria for 24 hours or 12-hour anuria).^[1]

2. The bundle prevents; the score only stratifies. Four planks — haemodynamic optimisation with monitoring, volume restoration, nephrotoxin/contrast avoidance, no hyperglycaemia — proved first in PrevAKI (55.1 versus 71.7%, single-centre, hypothesis-generating)^[18] and confirmed in BigpAK-2 across 34 hospitals (14.4 versus 22.3%, OR 0.57, NNT 12, no harm).^[19] Scores stratify at best (validated AUROC 0.71–0.75 after recalibration) and never substitute for the bundle.^[12]

3. Balanced liberal fluids to a pressure floor — both extremes injure. RELIEF: 3.7 versus 6.1 L raised AKI 8.6 versus 5.0% with no survival gain — restriction refused.^[20] Balanced fluids beat saline (mortality OR 0.84) and starch (AKI OR 0.80) across 58 trials — saline chosen.^[21] Overload still harms and pressor-for-volume substitution tracks more AKI — so resuscitate controlled crystalloids, hold MAP 65–70, give volume before squeezing.^[20]

4. Watch today — triggers written, floor respected, ELAIN quarantined. STARRT-AKI: identical 90-day death (43.9 versus 43.7%) with more dependence (10.4 versus 6.0%) and harm — no trigger, no dialysis.^[38] AKIKI's five triggers (hyperkalaemia, acidosis, oedema, BUN above 112, 72-hour oliguria) end the wait, and half the delayed arm escaped dialysis entirely.^[39] AKIKI-2 sets the floor: past 72-hour oliguria/BUN 112, further delay to BUN 140 gained nothing and raised death (HR 1.65).^[40] ELAIN's early benefit (39.3 versus 54.7%) stays single-centre, NGAL-enriched and contradicted — quoted only to dismiss.^[42][43]

5. Hyperoncotic albumin harms, iso-oncotic resuscitates; AUC vancomycin; every survivor watched. Post-cardiac 20% albumin infusion increased AKI (adjusted RR 1.12) — never routine — while iso-oncotic albumin as a resuscitation fluid beats saline and starch.^[24][21] Vancomycin moves to AUC-guided dosing (OR 0.625 for AKI versus trough).^[32] Follow-up is essential: even mild AKI predicts 1-year death (adjusted HR 2.96), creatinine dips overestimate recovery, and kidney health assessment after discharge is essential.^[10][4]