10.1 AKI, RRT, and Renal Drug Dosing
Key Takeaways
- KDIGO defines AKI as a creatinine rise of at least 0.3 mg/dL in 48 hours, a 1.5-fold rise from baseline within 7 days, or urine output below 0.5 mL/kg/h for 6 hours, and stages severity by the worse of creatinine or oliguria.
- Separate prerenal hypoperfusion from ATN, contrast-associated injury, abdominal compartment oliguria, and mannitol or hypertonic-saline kidney injury before reflexively giving more fluid or more osmotherapy.
- Start renal replacement for AEIOU problems that will not reverse quickly: refractory acidosis, hyperkalemia, selected intoxications, pulmonary edema, and uremic pericarditis or encephalopathy.
- KDIGO suggests continuous RRT rather than intermittent hemodialysis when intracranial pressure is high, because IHD can drop mean arterial pressure and raise brain water.
- Levetiracetam and meropenem are small, unbound, and efficiently cleared on CRRT; do not leave a running circuit on once-daily end-stage-kidney dosing.
The kidney is how the neuro ICU keeps osmoles, potassium, and drug levels from becoming second brain injuries. This independent OpenExamPrep section covers how acute kidney injury (AKI) is staged, how to tell common mechanisms apart in a brain-injured patient, when renal replacement therapy (RRT) is indicated, why intermittent hemodialysis (IHD) can wreck cerebral perfusion pressure (CPP), and how to dose levetiracetam and key antibiotics while continuous renal replacement therapy (CRRT) is running. Kidney Disease: Improving Global Outcomes (KDIGO) published the staging system still used at the bedside in 2012. Name that source. Do not claim this guide is a substitute for a nephrology consult or for the package insert.
Why AKI items are written as physiology traps
A creatinine of 1.4 mg/dL is not a diagnosis. A 70 kg patient making 30 mL/h has already declared oliguria. Mannitol that “worked” on intracranial pressure (ICP) may have emptied the tank so mean arterial pressure (MAP) fell and CPP fell with it. Iodinated contrast for a CT angiogram is not the same insult as a clamped abdomen after damage-control laparotomy. The exam will hand you two numbers and a timeline and expect you to pick the mechanism and the next action, not recite a brand of dialysis filter.
KDIGO definition and stages
KDIGO defines AKI as any of the following:
- Serum creatinine increase of ≥0.3 mg/dL (≥26.5 µmol/L) within 48 hours, or
- Serum creatinine increase to ≥1.5 times baseline known or presumed to have occurred within 7 days, or
- Urine volume <0.5 mL/kg/h for 6 hours
Stage by whichever criterion is worse. Initiation of RRT is stage 3 even if the last creatinine you saw was modest, because the clinical decision already declared severe organ failure.
| Stage | Creatinine | Urine output |
|---|---|---|
| 1 | 1.5–1.9 × baseline, or ≥0.3 mg/dL rise | <0.5 mL/kg/h for 6–12 hours |
| 2 | 2.0–2.9 × baseline | <0.5 mL/kg/h for ≥12 hours |
| 3 | ≥3.0 × baseline, or creatinine ≥4.0 mg/dL, or RRT started (or pediatric eGFR <35 mL/min/1.73 m²) | <0.3 mL/kg/h for ≥24 hours, or anuria ≥12 hours |
Worked staging: baseline creatinine 0.9 mg/dL, now 1.3 mg/dL over 36 hours in a 80 kg adult making 35 mL/h for 10 hours. The creatinine rise is 0.4 mg/dL (stage 1 by creatinine). Urine output is 35 / 80 = 0.44 mL/kg/h for 10 hours (also stage 1 by urine). If the same patient later makes 15 mL/h for 14 hours, urine output is 0.19 mL/kg/h for ≥12 hours and the stage jumps even before the next chemistry panel. Do not wait for a “pretty” creatinine if the Foley bag is empty.
Creatinine lags. A muscular young patient can lose a large fraction of glomerular filtration rate (GFR) before creatinine doubles. A sarcopenic patient can look “stable” with a creatinine of 0.6 mg/dL while GFR is already poor. Urine output is the real-time vital sign. Cystatin C and novel biomarkers exist in research protocols; they are not required to stage AKI on this exam.
Prerenal hypoperfusion versus ATN
Prerenal AKI means the nephron is still working but delivery is not: hypovolemia after mannitol or diabetes insipidus, vasoplegia, high intra-abdominal pressure, or a MAP below that patient’s autoregulatory range. Typical clues are a bland sediment, urine sodium often <20 mEq/L, urine osmolality often >500 mOsm/kg, and fractional excretion of sodium (FENa) <1% if no diuretic was given. Restore preload and perfusion and creatinine often falls within a day.
Acute tubular necrosis (ATN) follows prolonged ischemia, sepsis, rhabdomyolysis, or direct toxins. Muddy-brown granular casts, renal tubular epithelial cells, urine sodium often >40 mEq/L, and FENa >2% are the classic cluster. Once ATN is established, dumping liters of crystalloid will not “open the kidneys”; it will flood the lungs and can worsen cerebral edema if osmolality falls. Support perfusion, stop the toxin, and plan for RRT if AEIOU features appear.
FENa is a tool, not a religion. Loop diuretics, mannitol, chronic kidney disease, contrast, and bicarbonaturia all scramble the number. If the patient received furosemide two hours ago, do not use FENa to “prove” ATN. Look at the volume exam, the Foley, the lactate, the abdomen, and the drugs.
Contrast-associated injury
Iodinated contrast can add a creatinine bump, typically beginning 24–72 hours after exposure and peaking around day 3–5. Intra-arterial injections (catheter angiography, some endovascular cases) carry more risk than a single intravenous CT dose, especially when GFR is already reduced, the patient is volume-down, or other nephrotoxins are stacked. The injury is often non-oliguric. Prevention that still matters is isotonic volume expansion before and after necessary angiograms in high-risk patients, holding extra NSAIDs, and not repeating contrast “because the last scan was almost diagnostic.” N-acetylcysteine is not a reliable shield. Isosmolar versus low-osmolar contrast debates should not delay a salvage thrombectomy; you treat the brain and then watch the creatinine.
Do not diagnose contrast nephropathy at hour 4 after a CTA. That early jump is more often prerenal, lab variation, or the original shock. Do not withhold a needed digital-subtraction angiogram for a ruptured aneurysm solely because creatinine is 1.6 mg/dL; control bleeding first and hydrate around the study.
Abdominal compartment syndrome
A tight abdomen after polytrauma, massive resuscitation, pancreatitis, or retroperitoneal hemorrhage can drop urine output even when MAP looks acceptable. Intra-abdominal hypertension is commonly defined as bladder pressure ≥12 mmHg. Abdominal compartment syndrome (ACS) is sustained intra-abdominal pressure >20 mmHg with new organ failure. The kidney sees compressed veins, increased renal vascular resistance, and falling abdominal perfusion pressure (MAP minus intra-abdominal pressure). Oliguria here is not an invitation for another 2 L bolus that will only raise abdominal pressure further.
Measure bladder pressure with a standardized protocol. Decompress the abdomen—open the fascia, evacuate hematoma, treat ascites, correct tight binders—rather than chasing a FENa. Neurologic overlap: ACS also raises intrathoracic pressure and can impede cerebral venous drainage, so ICP and airway pressures may climb together with the creatinine.
Mannitol, hypertonic saline, and the kidney
Mannitol is filtered and not reabsorbed. High cumulative doses, hypovolemia, and preexisting chronic kidney disease produce osmotic nephrosis (vacuolization of proximal tubular cells) and a climbing osmolar gap. Replace urine milliliter for milliliter unless the patient is already overloaded. An anuric patient cannot clear mannitol; further boluses then expand plasma osmolality without an exit ramp. Hold routine redosing if the osmolar gap is already wide (classic teaching around 20 mOsm/kg) unless herniation is active and surgery is minutes away.
Hypertonic saline (HTS) spares the osmotic diuresis but loads chloride. Hyperchloremia is associated with renal vasoconstriction and a non-anion-gap metabolic acidosis. Driving sodium above the mid-150s mEq/L without an ICP indication adds kidney and myelin risk for no perfusion gain. Prefer HTS over mannitol when the patient is already hypovolemic or oliguric, and prefer neither as a volume strategy in anuric end-stage kidney disease without a dialysis plan.
Rhabdomyolysis after prolonged seizure, immobilization, or compartment syndrome dumps myoglobin. Alkaline diuresis is sometimes used if urine is still flowing; once anuric, the treatment is RRT for potassium and volume, not more mannitol.
Indications for RRT: AEIOU
There is no creatinine number that by itself mandates dialysis in a viable patient. Start RRT for problems you cannot fix quickly:
| Letter | Problem | Neuro ICU examples |
|---|---|---|
| Acidosis | Refractory metabolic acidosis | Lactate from septic or ischemic gut; hyperchloremic acidosis after massive HTS if ventilation cannot compensate |
| Electrolytes | Refractory hyperkalemia | Succinylcholine after denervation, rhabdomyolysis, missed dialysis |
| Intoxications | Dialyzable poisons | Lithium, toxic alcohols, salicylate, valproate in selected overdoses |
| Overload | Pulmonary edema, refractory hypoxemia | Neurogenic or cardiogenic edema after SAH takedown of fluids |
| Uremia | Pericarditis, encephalopathy, bleeding | Platelet dysfunction that will not wait for a spontaneous GFR recovery |
Urea itself is an osmole. A very high blood urea nitrogen (BUN) that you crash with a first IHD session can drop plasma osmolality faster than brain urea, pulling water into the brain (dialysis disequilibrium). That is one reason slow continuous therapies are favored when ICP is already high.
CRRT versus IHD when ICP is high
CPP equals MAP minus ICP (or minus central venous pressure if that is higher). IHD removes solute and volume over a few hours. MAP commonly sags during ultrafiltration. Plasma urea and sodium can fall faster than intracellular brain osmoles, increasing brain water. Imaging studies have shown more brain edema after IHD than after continuous therapies. KDIGO suggests CRRT rather than intermittent RRT for patients with acute brain injury, raised ICP, or generalized brain edema.
If IHD is the only machine available, blunt the insult: slower blood flow, higher dialysate sodium, cooler dialysate, minimal ultrafiltration, lower bicarbonate rise, and osmotherapy standing by. Daily short sessions beat a heroic six-hour first run. None of those tricks makes IHD equivalent to CRRT in a herniating patient.
CRRT still has costs: immobilization, hypothermia, citrate-related hypocalcemia, filter clotting, and underdosing of drugs. Regional citrate anticoagulation is attractive after intracranial hemorrhage because you are not heparinizing the patient; watch ionized calcium and the total-to-ionized calcium ratio for citrate lock, especially in liver failure. Heparin circuits are simpler but raise bleeding risk.
Drug dosing on CRRT
Critical illness inflates volume of distribution for hydrophilic drugs. Give a full loading dose of levetiracetam, beta-lactams, and vancomycin even when GFR is near zero; the first dose fills the tank. Maintenance then follows residual urine output plus effluent rate (commonly on the order of 20–25 mL/kg/h in many protocols, higher in some units).
Levetiracetam is small, minimally protein-bound, and about two-thirds renally excreted. Intermittent hemodialysis patients are often dosed once daily. That schedule underdoses CRRT. After a status-epilepticus load (commonly 60 mg/kg, maximum 4.5 g, still given in AKI), many adults on CRRT need on the order of 750–1000 mg every 12 hours, and high-effluent or large patients may need more. Follow levels if seizures continue. When the filter clots for 12 hours, clearance falls—do not blindly keep the CRRT dose during a long downtime, and do not forget to raise it when the circuit restarts.
Meropenem is similarly unbound and CRRT-cleared. Meningitis and ventriculitis need high cerebrospinal-fluid (CSF) targets. While the circuit is running, many neuro ICUs keep meningitis-range dosing (often 2 g IV every 8 hours in adults without a reason to go lower) rather than the reduced end-stage-kidney every-24-hour schedule. Extended or prolonged infusions help time-above-MIC. Vancomycin needs a load and then levels or a continuous infusion; do not assume a random trough from yesterday still applies after effluent rate doubles.
Antibiotics with high protein binding or large volume of distribution (many azoles, ceftriaxone to a degree) are less CRRT-sensitive. Always re-dose after IHD for drugs that the session strips. Ask pharmacy to map the actual modality—CVVH, CVVHD, CVVHDF—and the current effluent, not a generic “renal dose” sticker.
Exam traps
Calling a 0.2 mg/dL creatinine rise “no AKI” while urine has been 0.3 mL/kg/h for 10 hours. Treating ATN with serial fluid challenges. Blaming contrast at hour 2. Giving more mannitol to an anuric patient. Starting IHD for a urea of 80 mg/dL in a patient with ICP 28 mmHg “because that is how the chronic dialysis unit does it.” Leaving levetiracetam at 500 mg daily on high-dose CRRT and wondering why seizures returned.
A 80 kg patient has a baseline creatinine of 1.0 mg/dL. Over 36 hours creatinine rises to 1.4 mg/dL and urine output has been 32 mL/h for 9 hours. Using KDIGO staging, which statement is most accurate?
A patient with ICP 26 mmHg and a functioning EVD now has refractory hyperkalemia and volume overload. Which RRT plan best protects cerebral perfusion?
An adult on CVVHDF at 25 mL/kg/h was loaded with levetiracetam 60 mg/kg for status epilepticus. Which maintenance plan is most appropriate while the circuit runs?
After massive abdominal resuscitation, bladder pressure is 24 mmHg, airway pressures are rising, and urine output is 10 mL/h despite a MAP of 80 mmHg. What is the best next renal maneuver?