Distinction between induction and maintenance dosing in continuous renal replacement therapy

In their excellent review of dosing continuous renal replacement therapy (CRRT), Dr Prowle and colleagues concluded that patients should be prescribed 20 to 25 ml/kg/h [1]. However, by averaging CRRT dose over time, studies in this area obfuscate the benefits of appropriately higher dose therapy early in the course of illness, potentially misguiding clinicians into blindly adopting a 'one-size-fits-all' approach and consequently prescribing inadequate doses in life-threatening emergencies. To take a crude example, it would be inappropriate to prescribe 20 ml/kg/h CRRT in a patient with serum potassium 9 mmol/L. Rather, the highest possible dose of CRRT should be initially prescribed to maximize solute clearance. This depends on the maximum circuit flow permitted by the access catheter, which in turn determines the maximum dose, assuming that the countercurrent flow to blood flow ratio should be <0.3 with diffusive CRRT, or a filtration fraction with convective therapy <0.2 [2]. As the potassium level falls, the dose can be lowered to more conventional levels. 
 
Parallels could be drawn to general anaesthesia, where induction and maintenance are two distinct phases with different requirements. CRRT prescription could similarly be conceptualized as 'induction', where life-threatening abnormalities are corrected quickly with high-dose therapy, then 'maintenance' where solute clearance is achieved with more temperate doses (for example, 20 to 25 ml/kg/h) to avoid complications such as hypophosphataemia. 
 
It seems unlikely that this issue will be the subject of prospective research. Yet the principle that faster correction of life-threatening abnormalities leads to better patient outcomes seems both practical and intuitive.

In their excellent review of dosing continuous renal replacement therapy (CRRT), Dr Prowle and colleagues concluded that patients should be prescribed 20 to 25 ml/kg/h [1]. However, by averaging CRRT dose over time, studies in this area obfuscate the benefi ts of appropriately higher dose therapy early in the course of illness, potentially misguiding clinicians into blindly adopting a 'one-size-fi ts-all' approach and consequently prescribing inadequate doses in life-threatening emergencies. To take a crude example, it would be inappropriate to prescribe 20 ml/kg/h CRRT in a patient with serum potassium 9 mmol/L. Rather, the highest possible dose of CRRT should be initially prescribed to maximize solute clearance. Th is depends on the maximum circuit fl ow permitted by the access catheter, which in turn determines the maximum dose, assuming that the counter current fl ow to blood fl ow ratio should be <0.3 with diff usive CRRT, or a fi ltration fraction with convective therapy <0.2 [2]. As the potassium level falls, the dose can be lowered to more conventional levels.
Parallels could be drawn to general anaesthesia, where induction and maintenance are two distinct phases with diff erent requirements. CRRT prescription could simi lar ly be conceptualized as 'induction' , where life-threatening abnor malities are corrected quickly with high-dose therapy, then 'maintenance' where solute clearance is achieved with more temperate doses (for example, 20 to 25 ml/kg/h) to avoid complications such as hypo phos phataemia.
It seems unlikely that this issue will be the subject of prospective research. Yet the principle that faster correction of life-threatening abnormalities leads to better patient outcomes seems both practical and intuitive.

John R Prowle and Rinaldo Bellomo
We thank Dr MacLaren for his comments. We agree that such 'induction' therapy is mandatory in situations of severe hyperkalemia (>8.0 mmol), especially when ongoing potassium release is taking place (for example, rhabdomyolysis). Although such reasoning and therapeutic adjustment seemed obvious to us, we agree that making them explicit is important.
We would also like to emphasize several important aspects of adjustments to unique cases. First, avoidance of often ineff ective and delayed interventions in critically ill patients with severe acute kidney injury [3] and early CRRT as applied in the RENAL (Randomised Evaluation of Normal versus Augmented Level of Replacement Th erapy) trial [4][5][6] are key to preventing such lifethreaten ing events in the fi rst place. In this regard, in the RENAL trial, only 6 out of 1,454 (0.4%) patients with baseline electrolyte values had a potassium level >8 mmol/L. Second, faster solute removal can be achieved by the application of modifi ed dialytic techniques that deliver full equilibration between plasma fl ow and dialysate fl ow and much greater solute clearance [7]. Th ird, the concept of 'induction' applies to volume control as well: in a patient with pulmonary edema or receiving large amounts of blood products, the intensity of volume removal should be adjusted accordingly. Finally, however, also a word a caution: in the RENAL trial, three patients suff ered from disequilibrium syndrome because of too rapid solute removal. In patients with very high concentrations of azotemic markers, overly intensive solute removal can be detrimental. Similarly, in patients with marked hyper-or hyponatremia, rapid normali zation can be dangerous and induce cerebral edema or demyelination.