Controversy in Nephrology: Has Continuous Kidney Replacement Failed Its Promise?

More than 35 years ago, continuous arteriovenous hemofiltration (CAVH) was introduced by Kramer and colleagues (1) in order to optimize volume in hemodynamically compromised individuals with insufficient urine output. The successful treatment of congestive heart failure, despite cardiogenic shock, was heralded as a major advance, but soon limitations in solute clearance and complications of critical limb ischemia had clinicians looking for better solutions. This ushered in an era of multiple continuous dialytic techniques, including slow continuous ultrafiltration (SCUF), continuous venovenous hemofiltration (CVVH), continuous venovenous hemodiafiltration (CVVHDF), and sustained low-efficiency dialysis (SLED), which have become commonplace in intensive care units (ICUs). Benefits of these continuous renal replacement therapies (CRRTs) were quickly acknowledged. Clinicians could achieve excellent solute control throughout the day and remove substantial amounts of volume to keep up with or exceed input, and hemodynamic consequences appeared minimal. In fact, the blood pressure frequently stabilized.

Shortly thereafter, intensivists were galvanized by studies suggesting that critically ill patients could benefit from numerous “goal-directed” therapies targeting optimal organ perfusion by careful adjustment of numerous parameters, including mean arterial pressure, central venous pressure, mixed venous O2 saturation, and serum lactate levels (2). Thus, the combination of rigorous control of volume and solutes in patients with acute kidney injury (AKI) or fluid overload was immediately appealing, and CRRTs seemed a perfect fit to the goal-directed approach. However, subsequent studies have demonstrated limited efficacy in aggressively pursuing targets, as compared to more general efforts assuring adequate volume support and early use of antibiotics (3). Still, the notion of early and aggressive replacement of kidney function in a highly efficient manner, one that takes advantage of the full day while limiting hemodynamic consequences, is quite attractive.

Nonetheless, several research findings from randomized controlled trials (RCTs) now seem generalizable to most patients in the ICU. First, “early” initiation of kidney replacement therapy does not improve care or outcomes in the ICU and may subject many patients, who will never require it, to dialysis. “Early” is before there are absolute indications (uremia or unmanageable hyperkalemia or fluid overload) or relentless advanced azotemia. This finding, on the basis of multiple RCTs (4, 5), likely sheds light on the fact that despite improved volume and solute control afforded by kidney replacement therapy, these potential gains are offset by potential risks of kidney replacement therapy. Secondly, for a “typical” hemodynamically compromised patient in the ICU, continuous therapies do not result in better outcomes of interest than does intermittent hemodialysis. That is to say that despite evidence for greater hemodynamic stability and greater ability to control volume status, key outcomes, such as mortality, ventilator days, ICU length of stay, and even kidney recovery rates, in no way are improved (6−9). This is regardless of using a hemofiltration or hemodialysis approach or higher or lower doses of kidney replacement therapy. Thus, it has been stated that for typical lCU patients with hemodynamic instability and pressor requirements, the most reasonable approach is to manage them with intermittent hemodialysis. The advantages are that the patients can be more flexible for other procedures and be off the dialysis machine for parts of the day. Additionally, the associated costs tend to be more acceptable for intermittent dialysis as compared to continuous kidney replacement.

Perhaps these findings should not be so surprising. One of the major findings regarding kidney replacement therapy over the past decade has been that dose and frequency, beyond a certain minimum, do not improve outcomes. That is to say that once those solutes are under control and not causing immediate harm, further therapy does not appear to be useful. Thus, the evaluation of the improved stability of 6 times per week dialysis versus 3 sessions, the improved time, and the improved volume control did not equate to improved survival or improved rates of kidney recovery. Similarly, for continuous therapies, increasing the clearance rates did not result in improved survival (10).

Understanding exactly why this occurs is difficult, as trials have not been able to incorporate tests for associated components of the dialytic technique. Is it possible that rigorous control of blood pressure in the ICU is not meaningful and that patients with frequent exacerbations of hypotension may do as well as those with well-controlled stable blood pressure? Can it be that volume control that is episodic in nature and ultimately less efficient in getting to goal is none the worse for severely ill ICU patients? Perhaps these are true. However, it seems more likely that there is a tradeoff. That is to say that exposure to dialysis, whether intermittent or continuous is, by its very nature, hazardous. There are changes in many electrolytes and solutes and in volume status that may not be as controlled as we would hope. There is the extracorporeal circuit with risks for contamination, infection, and direct allergic and immunologic reactions to the artificial membranes. There are the hazards of being stuck on and requiring various forms of anticoagulation. Additionally, there could be harm from aggressive therapy that does not allow recognition of the earliest stages of kidney recovery, wherein dialysis could be discontinued and its potential for further harm averted.

Further research into dialytic techniques may revolutionize the practice of intensive care nephrology. However, for now, it would appear that optimal communication is necessary among all of those who take care of critically ill patients at risk for, or suffering from, AKI. The large number of completed trials together suggests that kidney replacement therapy should be utilized only when it is absolutely needed and no earlier and should be used as sparingly as tolerated to achieve manageable volume, solute, and electrolyte control. Early kidney replacement, continuous therapy, and ever-increasing desires to normalize every laboratory and physiological measure have not lived up to their promise to improve patient outcomes. More understanding is certainly needed and essential.