Dialysis-Associated Neurovascular Injury (DANI) in Acute Brain Injury

Introduction

Patients with acute brain injury are at heightened risk for dialysis-associated neurovascular injury (DANI), a clinical syndrome associated with worsening cerebral edema and impaired cerebral blood flow during dialysis (1). Acute brain injury itself is commonly associated with cerebral edema and disrupted cerebral blood flow. The volatility of cerebral blood flow and osmotic shifts during dialysis, combined with ongoing neurologic injury in acute brain injury, create a high-risk setting for secondary brain injury.

No consensus exists on neuromonitoring or intermittent dialysis modifications in these patients who are at risk. In this article, we use an illustrative patient who is clinical to discuss clinical decision making and prescription modification, specifically for intermittent hemodialysis and acute brain injury.

Case

A 65-year-old woman with kidney failure presented with acute aphasia and right-sided weakness. Imaging revealed a 40 cc acute left intraparenchymal hematoma. Imaging is stable on repeat imaging over 48 hours. She is hemodynamically stable, with a therapeutically hypernatremic serum sodium of 154 mEq/L. Nephrology is consulted for the reinitiation of hemodialysis.

• (1) For which pathophysiologic elements of DANI is the patient most at risk?

• (2) What options would you consider for RRT modification?

Dialysis-Associated Neurovascular Injury Pathophysiology

Cerebral Edema in Dialysis-Associated Neurovascular Injury

Cerebral edema during dialysis, sometimes referred to as dialysis disequilibrium syndrome, is a subset of DANI. Cerebral edema is most commonly thought to arise from faster removal of urea from plasma than brain parenchyma, which causes an osmotic gradient with water movement from plasma to the brain in a “reverse urea effect” (1). A separate hypothesis, less supported by current literature, assumes the osmotic gradient during dialysis is formed not by urea, but intracellular idiogenic osmoles created as an adaptive response to hyperosmolality from kidney failure.

Worsening cerebral edema, in turn, can cause increased intracranial pressure. Although these changes are most commonly associated with intermittent hemodialysis due to more rapid changes in plasma urea and osmolality, similar findings from slower modalities such as sustained low-efficiency dialysis and continuous RRT are being increasingly recognized among patients with acute brain injury (2).

Cerebral Blood Flow in Dialysis-Associated Neurovascular Injury

Cerebral blood flow can be impaired during RRT, leading to hypoperfusion and ischemia. Transcranial Doppler ultrasound is a noninvasive method for monitoring cerebral blood flow using intracranial vessel mean flow velocity as a marker for cerebral blood flow. During intermittent hemodialysis, the majority of transcranial Doppler studies have shown significant reductions in mean flow velocity, often correlated to ultrafiltration volume (3). In patients with acute brain injury, the impaired blood-brain barrier and decreased cerebral compliance from parenchymal injury may magnify the effects of intradialytic cerebral blood flow change. Patients with AKI requiring dialysis may be more at risk for intradialytic drops in cerebral blood flow (4,5).

In addition to decreased cerebral blood flow during treatment, increased blood viscosity, microemboli, and osmotic stress can create a low-flow, prothrombotic inflammatory environment, leading to intradialytic stroke. Uremia can impair endothelial function, independently compromising cerebral blood flow by impairing cerebrovascular autoregulation (1,3).

Case Continued

After tolerating 36 hours of continuous RRT with stable hemodynamics and unchanged neurologic examination, the patient is started on intermittent hemodialysis. Transcranial Doppler shows marked drops in left-hemisphere blood flow velocity through treatment. Postdialysis, the patient is newly somnolent and difficult to arouse. Repeat computed tomography head scan after dialysis is unchanged from before, and continuous electroencephalography monitoring shows left-sided slowing, but no seizures.

Neuro monitoring Options for Dialysis-Associated Neurovascular Injury

Invasive neuromonitoring by external ventricular drain or subarachnoid bolt can help assess changes in intracranial pressure as a marker for cerebral edema during dialysis, although not all patients may be candidates for invasive monitoring due to coagulopathies or the nature of neurologic disease.

Noninvasive neuromonitoring tools, such as near-infrared spectroscopy and transcranial Doppler ultrasound, can be helpful to establish trends in cerebral blood flow. Near-infrared spectroscopy measures frontal lobe tissue oxygenation. Reductions in these values during dialysis may suggest changes in intradialytic cerebral oxygen delivery, although thresholds for clinically relevant injury are unknown. Transcranial Doppler allows repeated, noninvasive monitoring of cerebrovascular hemodynamics during dialysis, which may be helpful to detect cerebral hypoperfusion in patients with limited neurologic examinations due to brain injury severity (3).

Treatment Modifications for Dialysis-Associated Neurovascular Injury

As cerebral edema from both ischemia and hemorrhage may subacutely worsen, most commonly over the first 72 hours, we recommend a conservative delay in hemodialysis initiation when possible. Due to a damaged blood-brain barrier after acute brain injury, anticoagulation in at least the first 2 weeks postinjury should be discussed with neuro intensive care teams regarding risk of hemorrhage. Intermittent hemodialysis may be helpful, particularly to avoid anticoagulation, or in patients with pre-established vascular access. In addition to advocating for neuromonitoring during dialysis for patients with acute brain injury, we briefly review options for intermittent dialysis prescription modification.

Reduced Osmotic Gradients

Although there are no controlled trials in patients with acute brain injury to guide urea clearance or osmotic change, secondary neurologic injury can be mitigated with slower urea clearance and smaller changes in plasma osmolality to prevent cerebral edema. In patients with volume overload, we recommend ultrafiltration only, followed by brief hemodialysis.

• Changes in plasma osmolality: Although the ideal rate of plasma osmolal change is unknown, there is some evidence to target <25 mmol/kg per day plasma osmolality change, on the basis of pediatric studies noting cerebral edema after plasma osmolality changes at rates >30 mmol/kg per day, but not below (1). This is corroborated in continuous dialysis modalities. Hypertonic saline or mannitol administration may help maintain the therapeutic hyperosmolar state in acute brain injury. Smaller surface dialyzers, reduced blood volume processed, and shorter RRT time may all help minimize changes in plasma osmolality and subsequent osmotic gradients across the blood-brain barrier (6).

• Dialysate sodium: Hypertonic sodium dialysate or infusions of hypertonic saline have been shown to decrease intracranial pressures during dialysis (7). In patients first starting dialysis, sodium modeling (in which dialysate sodium is reduced from high to normal ranges during treatment) was associated with fewer neurologic complaints and electro-encephalographic changes (1,8). Patients with acute brain injury and therapeutic hypernatremia (145–155 meq/L) should receive constant high dialysate sodium in the acute phase of injury to limit serum sodium fluctuations. In patients transitioning to normonatremia targets with ongoing damaged blood-brain barrier, sodium modeling should be considered.

• Limited blood volume processed per session: Limited blood volume processed per session may curb the “reverse urea effect” by reducing the urea gradient between blood and cerebro-spinal fluid. Options may include small surface dialyzers, short daily dialysis sessions, and reduced volume removal. For patients transitioning to intermittent RRT, short daily intermittent hemodialysis treatments or combined isolated ultrafiltration (to maintain plasma urea concentration) with short periods of dialysis can be practical solutions (8).

Maintenance of Cerebral Perfusion

Cerebral blood flow changes in RRT are worsened by intradialytic hypotension. Cerebral perfusion can be maintained by high sodium dialysate, weight-based ultrafiltration rates, and cooled dialysate. When acute brain injury is associated with fluid overload and hypertension, we recommend continuing sodium modeling and cooled dialysate.

• Sodium modeling: In a multicenter trial in critically ill patients, high dialysate sodium (10 mEq above standard) intermittent hemodialysis was equivalent to continuous RRT regarding intradialytic hypotension (8).

• Graded flow rates: For critically ill patients, a blood volume monitor during treatment can help calculate relative blood volume changes and limit intradialytic hypotension. A study of hemodynamic tolerance for intermittent dialysis in critically ill patients found fewer hypotensive episodes by starting blood flow rates at 50 ml/min, and then increasing through treatment to a flow of 250 ml/min (9).

• Cooled dialysate (34–35.5°C) may both reduce intradialytic hypotension and increase intradialytic mean arterial pressure. In a randomized controlled trial, dialysate temperatures 0.5°C below core body temperature protected against white matter changes on brain magnetic resonance imaging at 1 year, compared with those receiving dialysate at 37°C (10).

Conclusions

Patients with acute brain injury undergoing RRT are at heightened risk for DANI, a syndrome of dialysis-associated secondary brain injury. RRT modifications in acute brain injury may include reduced blood and dialysate flow rates, combined treatments for limited clearance, and sodium modeling. Interdisciplinary collaboration with neurointensivists and applications of neuromonitoring for personalized RRT prescriptions are integral to DANI prevention in patients with kidney failure and acute brain injury.

Disclosures

All authors have nothing to disclose.

Funding

None.