Indications for Renal Replacement Therapy in the Critically Ill

Uremia, Blood Urea Nitrogen, Creatinine

Although the development of overt uremic symptoms such as pericarditis, neuropathy, or coma represents an obvious indication for initiation of RRT, the start of RRT rarely is delayed until the full-blown uremia develops in intensive care units (ICUs). On the other hand, early signs such as anorexia, nausea, vomiting, or changes in mental state are usually nonspecific and hardly may be discriminated from symptoms of other diseases present in critically ill patients. Consequently progressive azotemia frequently is used for indications to start RRT for critically ill patients developing AKI. However, up to now there is no generally accepted threshold for when exactly to start RRT.

The concept of prophylactic hemodialysis in AKI was established by Teschan et al. more than 50 years ago. Based on several retrospective case series between 1950 and 1970 and two prospective trials in the 1970s and 1980s, recommended threshold of blood urea nitrogen (BUN) for initiation of hemodialysis decreased from 165 mg/dL to more than 200 mg/dL to levels ranging from 60 to 100 mg/dL. In a retrospective trial investigating timing in CVVH and using a BUN of 60 mg/dL for defining “early” versus “late” initiation of RRT, Gettings et al. found significantly improved survival in the “early” group (average BUN of 43 mg/dL) when compared with the “late” group (average BUN of 94 mg/dL). Another retrospective analysis on 243 ICU patients with AKI from the PICARD (Program to Improve Care in Acute Renal Disease) study used the median BUN of 76 mg/dL for defining early versus late initiation of dialysis. The authors found that the higher degree (>76 mg/dL, mean BUN 114.8 mg/dL) of azotemia at initiation was associated with an increased relative risk of 1.85 for death. However, a more recent single-center study on 302 critically ill patients with AKI demonstrated that urea levels at the time of initiation of RRT do not predict mortality. Overall, BUN or urea levels in the low to median range (i.e., BUN < 110 mg/dL) do not appear to be a valid criterion for starting RRT in critically ill patients.

Creatinine is considered a better indicator of glomerular filtration rate (GFR) and consequently was adopted as a parameter for the definition of AKI in whatever guise. Indeed, serial measurements of creatinine demonstrating relatively small increases is an indicator for increased mortality. Like urea, creatinine is nontoxic, and changes in serum concentration may occur independently of the GFR through changes in volume status, altered production (e.g., in sepsis), reduced muscle mass (e.g., liver cirrhosis), or by drug effects on the tubular excretion of creatinine. Consequently, although changes in serum creatinine have been suggested for classifying and staging AKI, the rate or degree of increase in serum creatinine may not reflect adequately the level of decline of GFR and fail to indicate the optimal time point to start AKI.

Volume Overload, Oliguria

Volume overload resulting from salt and water retention is a frequent complication in AKI, occurring in 30% to 70% of the patients in the ICU. Although diuretics frequently are tried for antagonizing oliguria, their benefit has not been proven in this situation. Patients with volume overload exhibit greater risk for increased morbidity and mortality. This is supported by the fact that patients responsive to diuretics also show improved outcome, and restrictive fluid management has proven to be beneficial in surgical patients, in ARDS and septic shock. Consequently, in the presence of severe volume overload that does not respond to diuretic therapy, initiation of RRT appears indicated.

Moreover, in the intensive care setting initiation of RRT is guided more frequently by oliguria expected to result in volume overload than by increases in creatinine or BUN.

A few retrospective studies investigating early initiation of RRT compared oliguria to conventional criteria (BUN or creatinine) for starting RRT. Two studies investigating patients who underwent cardiac surgery started CRRT when urine output was less than 100 mL over 8 hours, a third study in patients with septic shock used oliguria present for more than 12 hours as criterion. All three studies showed significantly reduced hospital or 30-day mortality in patients started on RRT in the presence of oliguria instead of waiting for an increase in BUN or serum creatinine. The only prospective randomized study investigating timing of RRT specifically in patients with diuretic-resistant oliguria could not find a difference between “early” and “late” initiation. However, mortality was low in this study and the sample size very small.

Acute Kidney Injury Stage

The RIFLE (risk, injury, failure, loss of kidney function, and end-stage renal failure) criteria were introduced in an attempt to standardize the definition and staging of AKI. They underwent subsequent modifications by the Acute Kidney Injury Network (AKIN) and finally were framed by the current Kidney Disease: Improving Global Outcomes (KDIGO) and Acute Kidney Injury Network (AKIN) criteria. The obvious advantage of these is that they can combine increasing serum creatinine and decreasing urine output rather than considering just one “trigger” in isolation. Several retrospective analyses, mostly using only creatinine for AKI stage determination, showed contradictory results. In 2015 to 2016 three randomized controlled trials appeared that systematically investigated AKI stages according to KDIGO as indications to start RRT. The German single-center ELAIN trial randomized 231 predominantly cardiac surgical patients with AKI stage II and a serum neutrophil gelatinase-associated lipocalin (NGAL) of more than 150 ng/mL to early versus late RRT. Mortality of patients randomized to early treatment as defined by RRT within 8 hours of AKI stage II had a 35% lower mortality than patients randomized to late RRT (i.e., within 12 hours of reaching AKI stage III). Nearly 90% of the randomized patients received RRT. A multicenter Canadian trial included 100 severely ill patients with AKI stage II and whole-blood NGAL levels of at least 400 ng/mL. Patients were randomized to be either commenced within 12 hours of meeting the inclusion criteria or to be started on RRT when so-called “classic indications” (i.e., hyperkalemia [K ⁺ > 6 mmol/L], acidosis, severe respiratory failure [paO ₂ /FiO ₂ < 200]) were reached. There was no difference in mortality between the two groups, but 30% of the patients did not need RRT in the group in which classic indications were applied. The French multicenter AKIKI trial was the largest of these three randomized controlled trials, including 620 mainly medical ICU patients. The “early” treatment group received RRT within 6 hours of reaching AKI stage III, whereas “late” treatment was initiated only when absolute indications were fulfilled, including serum potassium exceeding 6 mmol/L, acidosis, oliguria for more than 72 hours, blood urea nitrogen (BUN) more than 112 mg/dL, or pulmonary edema. Mortality was around 50% with no significant difference between both groups. However, 60% of patients randomized to “late” treatment did not require RRT. Thus, based on current randomized controlled trials (RCTs), a specific AKI stage cannot be recommended for initiation of RRT. It appears that even when using AKI stage III as the only indication, too many patients are receiving RRT who would recover without it.