When should renal replacement therapy be initiated?
There are accepted urgent indications for renal replacement therapy (RRT) in patients with acute kidney injury (AKI) and generally include: refractory fluid overload, hyperkalemia >6 mEq/L or rapidly rising potassium levels, signs of uremia, severe metabolic acidosis, and certain alcohol and drug intoxications.
Although the maintenance of serum creatinine and blood urea nitrogen (BUN) concentrations below arbitrarily set levels is usually a reference for starting dialysis treatment, neither creatinine nor BUN should be used to absolutely determine when to initiate dialysis.
BUN reflects factors not directly associated with kidney function, such as catabolic rate and volume status.
Serum creatinine is influenced by age, race, muscle mass, and catabolic rate, and its volume of distribution varies in fluid overloaded patients.
Other factors such as fluid balance control, nutrition needs, severity of the underlying disease, and acid base and electrolyte balance should guide the decision to start dialysis, as has been suggested by KDIGO AKI guidelines.
RRT initiation is favored prior to the development of severe electrolyte disturbances in patients with severe metabolic acidosis (pH < 7.2) despite optimal medical management and with no signs that kidney function or metabolic acidosis is improving; and in patients with positive fluid balance despite aggressive use of diuretics, predominantly if they have increasing oxygen requirements.
Therefore it is better to use RRT as a supportive therapy rather than a rescue therapy for late manifestations of AKI.
Since the optimal timing of RRT initiation is controversial, any presumed benefit from early dialysis needs to be balanced by the safety concerns from dialysis including:
• Risk for infection from an indwelling dialysis catheter
• Hypotension
• Potential for delayed kidney recovery
• Leukocyte activation from contact with dialysis membranes
Randomized, controlled trials that have compared strategies of early versus delayed initiation of RRT (in the absence of absolute indications) have yielded conflicting results.
In fact, three of the largest trials have not demonstrated a benefit with earlier initiation of RRT. The best evidence comes from a randomized, controlled trial, which included 620 critically ill patients that had severe AKI (KDIGO stage 3) and required either or both mechanical ventilation and vasopressors (AKIKI Study).
Patients were assigned to early RRT (within 6 hours after AKI was identified) or to delayed RRT; the delayed strategy required RRT initiation after the onset of severe hyperkalemia, metabolic acidosis, pulmonary edema, increase in BUN levels >112 mg/dL, or the development of oliguria for more than 72 hours after allocation.
The choice of the method of RRT (intermittent or continuous technique, duration and interval between sessions, device setting, and anticoagulation method) was left to the discretion of each study site.
The study showed no difference in terms of mortality between patients who received early RRT with a median time of initiation of 2 hours (patients with AKI stage 3 KDIGO without BUN levels >112 mmol/L, pH < 7.15, K levels > 6 mmol/L, and pulmonary edema due to fluid overload) and patients who received late RRT with a median time of initiation of 57 hours.
An interesting finding in this study was that kidney recovery marked by increased diuresis was more rapid, and catheter-related infections were lower in the delayed-strategy RRT.
Another study was the ELAIN study: a single center, randomized, controlled trial that enrolled 231 critically ill patients with AKI KDIGO stage 2.
All patients had severe sepsis, required vasopressors or catecholamines, or had refractory volume overload. In this study, early RRT (continuous venovenous hemodialfiltration) was started within 8 hours of diagnosis of AKI KDIGO stage 2, and delayed RRT was started within 12 hours of stage 3 AKI, or developed an absolute indication for RRT initiation (serum urea level >100 mg/dL, potassium >6 mEq/L, serum magnesium >8 mEq/L, urine output 200 mL over 12 hours, or diuretic-resistant edema).
Compared with delayed or no initiation, early RRT initiation reduced 90-day mortality.
In addition, more patients recovered kidney function in the early versus delayed group by 90 days, and both the duration of RRT and the hospital stay were shorter in the early initiation group.
The timing of RRT, a modifiable factor, might exert an important influence on patient survival.
We favor utilizing an approach that recognizes that the strategy in treating AKI is to minimize and avoid uremic and volume overload complications.
Thus, it is not necessary to wait for progressive uremia to initiate dialytic support; it is better to use RRT as a supportive therapy in the presence of progressive azotemia and oliguria, rather than a rescue therapy for the late manifestation of AKI.
Our recommendations to initiate RRT with worsening AKI are listed below:
• In patients with K >5.8 to 6.0 mEq/L or severe metabolic acidosis despite optimal medical management in patients that demonstrate no sign that kidney function or metabolic acidosis is improving
• In patients with positive fluid balance despite aggressive attempts of increasing diuresis, particularly if they have increasing oxygen support requirements
Careful surveillance is mandatory when deciding to wait for RRT initiation in patients with severe AKI, so that any complication will be adequately detected and RRT will be started without delay.
