Intensive care physicians often face the conundrum of deciding when to consider renal replacement therapy (RRT) in acute kidney injury (AKI). RRT may be commenced for the early correction of metabolic complications and prevention of volume overload. However, an early strategy may entail unnecessary therapy for some patients who might recover renal function otherwise. Even more concerning is the perpetuation of damage that may occur due to the hemodynamic instability that often accompanies RRT. The proximal tubules located at the corticomedullary junction are particularly vulnerable to ischemic insult. The partial pressure of oxygen at this region is less than 40 mm Hg, which increases the susceptibility of the tubular cells to RRT-induced damage (1). Hence it is critically important to decide on the requirement for RRT and to initiate therapy at the optimal time.
Renal replacement therapy: benefit vs. possible harm
The likelihood of benefit compared to the possibility of harm must be carefully weighed prior to the commencement of RRT. Particularly, in patients with no life-threatening complications and those with a reasonable likelihood of recovery, RRT may worsen renal function and may be independently associated with increased mortality (2,3). More than three decades ago, Conger et al. proposed that the potential benefits of early, aggressive RRT may be offset episodic intra-dialytic hypotension that may perpetrate recurrent ischemic insult and affect recovery from renal failure (4). Although less common compared to intermittent therapies, hypotensive episodes may still occur with continuous therapies (5). In the BEST trial, hypotension occurred in 19% of patients who underwent CRRT (6). Other complications of RRT include those related to vascular access and inappropriate antibiotic dosage, especially considering the lack of adequate information regarding optimized dosing (7). Besides, excessive loss of electrolytes and essential nutrients may occur (8). The dialytic membrane and the extracorporeal circuit may evoke an inflammatory response and lead to potential harm (9). The poor predictive ability to discern the possibility of harm from delayed initiation of RRT adds further complicity to the clinical scenario.
Many of the older studies that support early RRT have been observational (10); besides, the overall management of critically ill patients has changed considerably over the years. Hence, it is appropriate to reconsider the impact of the timing of RRT, particularly with new information available from the more recent randomized controlled studies (RCTs) (Table 1).
Table 1. RCTs evaluating the timing of RRT
Early vs. late: the evidence
The ELAIN trial
The ELAIN trial involved 231 postoperative patients (47% cardiac surgical) with AKI from a single center in Germany (11). The early group was randomized to receive RRT within 8 h of the diagnosis of AKI, KDIGO stage 2; the late group was randomized to undergo RRT within 12 h of reaching KDIGO stage 3. Besides, at least one of the following conditions had to be present, including severe sepsis, catecholamine use, refractory fluid overload, and new-onset or worsening of non-renal organ dysfunction. All 112 patients randomized to the early group received RRT, compared to 108/119 patients in the late group. The 90-d mortality, the primary outcome, was significantly lower in the early group. Approximately 75% of all patients had evidence of fluid overload according to the pre-defined criteria. This raises the question of whether RRT may have been delayed among patients who had evidence of fluid overload but were randomized to the late group, leading to worse outcomes in this group of patients. Besides, the fragility index for the primary outcome was only 3 patients (the results would have been non-significant if 3 extra patients had suffered 90-day mortality in the early group). Furthermore, it is hard to explain a mortality benefit at 90 d, with no difference in the 28 and 60 d mortality.
The AKIKI trial
Patients in KDIGO stage 3 were randomized to receive early or late RRT in the AKIKI trial (12). In the early group, RRT was initiated immediately after randomization. In the late group, RRT was performed if any of the following criteria were present: K+ > 6.0 mmol/l (> 5.5 mmol/l after corrective measures), BUN > 112 mg/dl, pH < 7.15, and acute pulmonary edema due to fluid overload). In the early group, 305/311 (98.1%) received RRT, while only 157/308 (51%) patients in the late group underwent RRT. There was no significant difference in mortality between groups on day 60, which was the primary outcome of the study. Thus, nearly half of all patients who were randomized to the late group did not require RRT and suffered no adverse outcomes.
The IDEAL-ICU trial
The IDEAL-ICU investigators, in a multicentric French study, randomized patients in early-stage septic shock who were in the failure stage (stage F) of the RIFLE classification (13). The early group received RRT within 12 h of documentation of stage F; the late group underwent RRT after 48 h if renal recovery had not occurred in the meantime. The primary outcome was 90-d mortality. The study was stopped for futility after a second interim analysis. No difference was observed in the 90-d mortality between the early and late groups (58% vs. 54%, p = 0.38). In the late group, 38% of patients did not receive RRT.
The STARRT-AKI trial
Critically ill patients with AKI from 168 sites across 15 countries were randomized to receive an accelerated or standard strategy of renal replacement therapy in the STARRT-AKI trial. Patients were included if they had reached KDIGO stage 2 or 3 and the treating physician considered that there was sufficient equipoise in choosing between an accelerated compared to a standard strategy. In the accelerated strategy group, RRT was commenced within 12 h of randomization. In the standard strategy group, RRT was initiated in the presence of metabolic acidosis (pH <7.2, or HCO3 <12 mmol/l), hyperkalemia (K+ >6 mmol/l), fluid overload with hypoxemia, or persistent AKI lasting >72 h after randomization. The primary outcome was the 90-d mortality; there was no difference between the accelerated compared with the standard strategy of RRT (43.9% vs. 43.7%; RR: 1, [95% CI 0.93-1.09], P = 0.92). RRT dependence at 90 days after randomization was significantly higher in the accelerated compared to the standard strategy group (10.4% vs 6%), RR 1.74 [95% CI 1.24-2.43]). Adverse events, including hypotension and hypophosphatemia, were more common with the accelerated strategy. No differences were observed between groups in the composite of all-cause mortality and RRT-dependence at 90 days. The ICU, 28-d, and hospital mortality were also similar between the two groups. The number of RRT-free days at 90 days, and the ventilation-free days, and vasopressor-free days at 28 days were also not significantly different between the two groups. This study demonstrated no improvement in the 90-d survival with an accelerated strategy of RRT; besides, it suggested increased RRT-dependence at 90 days compared to a standard strategy.
The AKIKI-2 trial
The AKIKI 2 trial compared two delayed strategies in critically ill patients who were on mechanical ventilation or catecholamine infusion and in KDIGO stage 3 of AKI. Patients were oliguric or anuric (urine output <0·3 ml/kg/h or <500 ml/day for > 72 h) or had a blood urea nitrogen level between 112–140 mg/dl. RRT was carried out within 12 h after meeting the above criteria (delayed group) or postponed until the blood urea nitrogen concentration level reached 140 mg/dl, or an urgent requirement for RRT arose in the meantime (more delayed group). The primary outcome was the number of RRT-free days until day 28; there was no significant difference between the delayed and more-delayed strategies (12 vs. 10 days; P = 0.93). On multivariable analysis, one of the risk factors associated with 60-d mortality was a more-delayed strategy; other risk factors were a higher SAPS III score and the requirement for mechanical ventilation. RRT-dependence at 60 days and complications potentially resulting from acute kidney injury or RRT were similar between groups.
Apart from the ELAIN trial, other RCTs suggest that in the absence of life-threatening complications, a delayed strategy may be appropriate in most patients with AKI. Such a strategy may facilitate avoidance of RRT, with its associated complications. These studies reinforce the view that initiation of RRT should not be based on the AKI stage; it may be more appropriate to be guided by the presence of complications, including fluid overload and metabolic abnormalities. Furthermore, a strategy of watchful waiting may enhance the possibility of renal recovery. In contrast, RRT-related hemodynamic instability may lead to worsening or perpetuation of renal damage and adversely impact recovery.
It is also important to assess the trajectory of the underlying disease process and the presence of failure of other organs while considering RRT. If the patient is clinically improving with the resolution of other organ failures, perhaps a waiting strategy would be more appropriate. However, if the metabolic demands continue to remain high with a continued requirement for fluid resuscitation, it may be appropriate to consider early RRT. In such situations, the metabolic and fluid demands far exceed the capacity of the kidney, necessitating prompt supportive therapy. The “frusemide test” was used to identify patients who are likely to require renal replacement therapy in a previous study (14). It involves the administration of intravenous frusemide, 1.0 mg/kg to frusemide-naïve patients and 1.5 mg/kg to those already on frusemide. A urine output of less than 200 ml in the ensuing 2 h was highly predictive of patients who required RRT.
The bottom line
- Previous observational studies of AKI seemed to suggest that an early-RRT strategy may lead to more favorable clinical outcomes.
- Recent randomized controlled trials reinforce the view that a delayed strategy may be appropriate in the absence of life-threatening complications of AKI.
- Apart from the complications directly related to therapy, hemodynamic perturbations during RRT may adversely impact the recovery of renal function.
- The decision to perform RRT will also depend on the trajectory of the underlying illness and the metabolic demand.
- Fastidious conformance to physiological or biochemical parameters is unlikely to be helpful with the timing of initiation of RRT.
- A holistic, individualized strategy based on clinical, physiological, and biochemical parameters is more likely to be beneficial compared to a “one-size-fits-all” approach.
1. Gaudry S, Quenot J-P, Hertig A, Barbar SD, Hajage D, Ricard J-D, et al. Timing of Renal Replacement Therapy for Severe Acute Kidney Injury in Critically Ill Patients. Am J Respir Crit Care Med [Internet]. 2019 Feb 20 [cited 2019 Mar 21]; Available from: https://www.atsjournals.org/doi/10.1164/rccm.201810-1906CP
2. Clec’h C, Gonzalez F, Lautrette A, Nguile-Makao M, Garrouste-Orgeas M, Jamali S, et al. Multiple-center evaluation of mortality associated with acute kidney injury in critically ill patients: a competing risks analysis. Crit Care Lond Engl. 2011;15(3):R128.
3. Elseviers MM, Lins RL, Van der Niepen P, Hoste E, Malbrain ML, Damas P, et al. Renal replacement therapy is an independent risk factor for mortality in critically ill patients with acute kidney injury. Crit Care Lond Engl. 2010;14(6):R221.
4. Conger JD. Does Hemodialysis Delay Recovery from Acute Renal Failure? Semin Dial. 1990;3(3):146–8.
5. Shawwa K, Kompotiatis P, Jentzer JC, Wiley BM, Williams AW, Dillon JJ, et al. Hypotension within one-hour from starting CRRT is associated with in-hospital mortality. J Crit Care. 2019 Dec;54:7–13.
6. Uchino S, Kellum JA, Bellomo R, Doig GS, Morimatsu H, Morgera S, et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA. 2005 Aug 17;294(7):813–8.
7. Fissell WH. Antimicrobial dosing in acute renal replacement. Adv Chronic Kidney Dis. 2013 Jan;20(1):85–93.
8. Finkel KW, Podoll AS. Complications of continuous renal replacement therapy. Semin Dial. 2009 Apr;22(2):155–9.
9. Gutierrez A, Alvestrand A, Wahren J, Bergström J. Effect of in vivo contact between blood and dialysis membranes on protein catabolism in humans. Kidney Int. 1990 Sep;38(3):487–94.
10. Bagshaw SM, Uchino S, Bellomo R, Morimatsu H, Morgera S, Schetz M, et al. Timing of renal replacement therapy and clinical outcomes in critically ill patients with severe acute kidney injury. J Crit Care. 2009 Mar;24(1):129–40.
11. Zarbock A, Kellum JA, Schmidt C, Van Aken H, Wempe C, Pavenstädt H, et al. Effect of Early vs Delayed Initiation of Renal Replacement Therapy on Mortality in Critically Ill Patients With Acute Kidney Injury: The ELAIN Randomized Clinical Trial. JAMA. 2016 May 24;315(20):2190.
12. Gaudry S, Hajage D, Schortgen F, Martin-Lefevre L, Pons B, Boulet E, et al. Initiation Strategies for Renal-Replacement Therapy in the Intensive Care Unit. N Engl J Med. 2016 Jul 14;375(2):122–33.
13. Barbar SD, Clere-Jehl R, Bourredjem A, Hernu R, Montini F, Bruyère R, et al. Timing of Renal-Replacement Therapy in Patients with Acute Kidney Injury and Sepsis. N Engl J Med. 2018 Oct 11;379(15):1431–42.
14. Lumlertgul N, Peerapornratana S, Trakarnvanich T, Pongsittisak W, Surasit K, Chuasuwan A, et al. Early versus standard initiation of renal replacement therapy in furosemide stress test non-responsive acute kidney injury patients (the FST trial). Crit Care Lond Engl. 2018 Apr 19;22(1):101.