Reply to Chousterman et al.: Delaying Renal Replacement Therapy Could Be Harmful in Patients with Acute Brain Injury.

From the Authors:

We thank Chousterman and colleagues for their positive appreciation of our work (1). However, we believe that their contention is mainly speculative, as it is based on anecdotal reports that provide no or little detail on the renal replacement therapy (RRT) modalities that were supposed to be responsible for neurological deterioration.

More important, we feel that the authors miss several points. They reason as if RRT were not associated with any risk except the increase in intracranial pressure. They fail to incorporate in their thinking process the different regulators of cerebral blood flow: arterial blood pressure, intracranial pressure, and cerebrovascular resistance (2). The first component, the cardiovascular component, has been highlighted for over a century (3). Hemodynamic instability is a frequent issue in brain-injured patients, and even more so in cases involving multiple trauma. Thus, RRT-associated hemodynamic instability, which occurs frequently and within the first minute of RRT (unlike disorders linked to osmolal changes, which are rare and have a delayed onset) may have catastrophic consequences on an injured brain. Starting RRT in a patient with recent head injury (especially in the context of polytrauma) may likely affect hemodynamics. In addition, the authors fail to consider that a delayed strategy has been shown to allow the avoidance of RRT in one-third to one-half of patients (4, 5). Obviously, the best way to avoid RRT-associated osmolal brain changes is to avoid RRT. The application of an early RRT strategy potentially increases the risk of hemodynamic fluctuation (which may decrease cerebral perfusion and contribute to acute brain injury) for all patients. In this regard, the remedy they propose (starting RRT early in all acute kidney injury patients with brain injury) may be worse than the disease. Finally, a careful reading of case reports and case series cited by Chousterman and colleagues (6) shows that in most cases, patients received “aggressive” intermittent RRT. For instance, in one case blood urea nitrogen decreased from 141 to 54 mg/dl in one session, which is not desirable even in a patient without brain injury. Several ways to avoid acute osmotic shifts exist (7) but were not discussed: slow and gentle initial hemodialysis (time <2 h and low blood flow rate), increasing dialysate sodium level, or administration of osmotically active substances (e.g., intravenous manitol).

In our era of evidence-based medicine, we must point out that stating “we suggest not using the delayed RRT initiation strategy in patients at risk of elevated intracranial pressure” is not supported by data. Similarly, stating that “the best strategy for RRT modalities and initiation in this subset of patients remains to be determined” means that one has to carefully weigh the actual (and proven) risk of undue RRT against that of delaying RRT in brain-injured patients. We suggest that before issuing so strong a warning without firm evidence, it would be necessary to conduct a randomized clinical trial on this particular population.