The aim of this chapter is to review the evidence surrounding the use of renal replacement therapy (RRT) in the intensive care unit (ICU) setting. It examines the conventional indications for emergency RRT and assesses the emerging evidence for earlier commencement of RRT and the expanded role of RRT in the management of sepsis and multiorgan failure (MOF).
What are the Conventional Indications for Commencing Renal Replacement in Acute Kidney Injury?
There is a paucity of consensus guidelines internationally with regard to RRT use in the ICU, and this has resulted in variable prescribing practices for continuous dialysis. However, some pathophysiologic states are generally considered absolute indications for this intervention ( Table 57-1 ).
|
|
|
|
|
|
|
|
Intravascular Volume Overload and Pulmonary Edema Refractory to Diuretic Therapy
The role of negative or neutral fluid balance in acute kidney injury (AKI) with pulmonary edema but without lung injury is unclear. Studies performed in critically ill children with AKI after cardiac surgery have suggested that early institution of continuous renal replacement therapy (CRRT) improves respiratory parameters with an associated improvement in multiple clinical outcomes. Randomized controlled trials (RCTs) in adults are lacking, although observational data indicate that a positive fluid balance in critically ill patients with AKI is independently associated with a higher 60-day mortality rate (hazard ratio [HR], 1.21; P < .001).
There is no evidence to support the common practice of trial of diuresis in AKI-associated pulmonary edema. Indeed, the use of diuretic therapy may increase the probability of nonrecovery of renal function. In addition, studies in animal models suggest that ultrafiltration is more effective than diuresis in reducing extravascular lung water in ARDS. In conclusion, RRT should be considered early in patients with AKI complicated by refractory pulmonary edema.
Metabolic Acidosis Refractory to Medical Management
Metabolic acidosis is a common complication of AKI, resulting from a combination of chloride-rich fluid resuscitation and the accumulation of lactate, phosphate, and unexcreted metabolic acids. RRT can be highly effective in correcting this acidosis. CRRT as a modality may be superior to intermittent hemodiafiltration (IHD) in terms of duration of treatment effect. Importantly, RRT avoids systemic administration of sodium bicarbonate therapy with its associated risk for exacerbating fluid overload and hypernatremia. The threshold pH or base deficit at which to commence RRT has not been established. Because a pH lower than 7.1 is associated with negative inotropic and metabolic effects, in general, one would consider intervening before this level is reached.
Hyperkalemia Refractory to Medical Management
No specific treatment threshold has been established for when to treat hyperkalemia with RRT. In general, myocardial toxicity is considered unlikely when the serum potassium concentration is less than 6.5 mmol/L. Potassium excretion by diuresis is generally ineffective in renal failure. For this reason, the threshold for commencing RRT in AKI might be lowered further, particularly if there is minimal response to initial emergency treatment (insulin-glucose, inhaled beta-agonist, exchange resins).
The Uremic State
Manifestations of the “uremic state” include encephalopathy, pericarditis, and bleeding diathesis. Mental status changes and bleeding propensity can be multifactorial in the septic, critically ill patient, and they can be difficult to attribute solely to renal failure. Uremic pericarditis requires urgent initiation of renal support once it is detected because it carries a high risk for intrapericardial hemorrhage and tamponade.
Intoxication with a Dialyzable Drug or Toxin
Toxins of low molecular weight residing in the extracellular space, which have little or no protein-binding properties, can be effectively removed by RRT. In general, IHD is preferable to CRRT for this purpose because it more rapidly clears solute. A review of the U.S. Poison Center’s “Toxic Exposure Surveillance System” records from 1985 to 2005 found that 19,351 cases received extracorporeal toxin removal over this time period. IHD was most commonly used for the treatment of lithium, ethylene glycol, salicylate, valproate, acetaminophen, methanol, ethanol, and theophylline poisoning, although some cases of IHD, used for removal of methotrexate and phenobarbital, were reported. Hemoperfusion techniques are used in the enhanced elimination of toxic levels of lipid-soluble or highly protein-bound substances when intervention will remove the substance more rapidly than endogenous clearance. An important consideration is the platelet-depleting effect of hemoperfusion.
Severe Electrolyte Derangements
AKI can be associated with an array of electrolyte disturbances, including hyponatremia, hypernatremia, hyperphosphatemia, hypercalcemia, hypocalcemia, and hypermagnesemia. CRRT may be helpful in the management of many of these disorders.
Progressive Azotemia or Oliguria Unresponsive to Fluid Administration
In the modern era, RRT is most often initiated before sufficient time has passed for the previously discussed scenarios to develop. Instead, the decision to commence treatment is made when urea and creatinine levels climb, or urine output falls, despite conservative measures. The threshold values of these parameters that should trigger a decision to commence RRT have not been established and are discussed later.
Should Renal Replacement Therapy be Initiated in Acute Kidney Injury Before Complications have Developed?
Although undisputed indications generally point to RRT as being a “rescue remedy” used when other measures have failed, several studies have examined the value of earlier commencement of therapy in improving patient outcomes ( Table 57-2 ).
Study | Mode | Design | Number of Patients | Group Definition | Survival | ||
|---|---|---|---|---|---|---|---|
| Early | Late | Early | Late | ||||
| Teschan, 1960 | IHD | Case series | 15 | <100 mg/dL | — | 33% | — |
| Parsons, 1961 | IHD | Single-arm (historical control) | 33 | BUN reaching 120-150 mg/dL | Clinical deterioration or BUN 200 mg/dL | 75% | 12% |
| Fischer, 1966 | IHD | Retrospective cohort study | 162 | Clinical deterioration or BUN increase to ∼150 mg/dL | Hyperkalemia BUN ∼200 mg/dL | 43% | 26% |
| Kleinknecht, 1972 | IHD | Retrospective cohort study | 500 | To maintain BUN <93 mg/dL (blood urea <200 mg/dL) | BUN >163 mg/dL (blood urea >350 mg/dL) or severe electrolyte disturbance | 73% | 58% |
| Conger, 1975 | IHD | RCT | 18 | BUN <70 mg/dL or SCr <5 mg/dL | BUN ∼150 mg/dL, SCr 10 mg/dL, or clinical indication | 64% | 20% |
| Gillum, 1986 | IHD | RCT | 34 | Maintenance of BUN <60 mg/dL | Maintenance of BUN ∼100 mg/dL | 41% | 53% |
| Gettings et al., 1999 | CRRT | Retrospective cohort study | 100 | BUN <60 mg/dL (mean 42.6 mg/dL) | BUN ≥60 mg/dL (mean, 94.5 mg/dL) | 39% | 20% |
| Bouman et al., 2002 | CVVH | RCT | 106 | Within 12 hours of developing UO <20 mL/h and Cr clearance <20 mL/min | Urea >40 mmol/L (BUN >112 mg/dL), SK >6.5 mEq/L (>6.5 mmol/L) or severe pulmonary edema | 69% | 75% |
| Demirkilic et al., 2004 | CVVHDF | Retrospective cohort study | 61 | UO <100 mL over 8 hours after surgery, despite furosemide bolus | SCr >5 mg/dL or SK >5.5 mEq/L | 77% | 45% |
| Elahi et al., 2004 | CVVH | Retrospective cohort study | 64 | UO <100 mL over 8 hours after surgery, despite furosemide infusion | BUN >84 mg/dL, SCr >2.8 mg/dL, or SK >6 mEq/L | 57% 78% | — — |
| Liu et al., 2006 | IHD, CRRT | Prospective cohort study | 243 | BUN <76 mg/dL | BUN >76 mg/dL | 65% | 59% |
It should be noted that there is no clear consensus on what is meant by “earlier” initiation of RRT; initiation at lower urea and creatinine levels, initiation closer to the time of renal injury, initiation sooner after urine output is noted to fall, and initiation sooner after admission to the ICU have all been studied (see Table 57-2 ). This makes study comparison and meta-analysis difficult. In addition, the effect of earlier initiation of RRT is likely to be influenced by the etiology of the AKI; thus, the heterogeneity of the populations studied renders meaningful meta-analysis even more difficult.
A small and retrospective study in post-traumatic AKI using a blood urea nitrogen (BUN) threshold for early initiation of RRT of 60 mg/dL demonstrated a significantly lower mortality rate for the early compared with the delayed RRT cohort (relative risk [RR] for death, 0.77; 95% confidence interval [CI], 0.58 to 1.0; P = .04). These results suggest that the BUN threshold for considering the initiation of RRT should be lowered to at least 60 mg/dL.
Further support for a strategy of earlier initiation of RRT was provided by retrospective studies in the postoperative coronary artery bypass graft (CABG) patient population. These studies used reduced urine output (<100 mL within 8 hours consecutively after surgery, despite frusemide administration) as their criterion for early initiation of CRRT. The attainment of specified BUN, serum creatinine, or potassium thresholds was the trigger for late commencement of therapy. The first of these studies examined the outcomes of 64 patients with a high baseline prevalence of class 3 or 4 heart failure and chronic kidney disease. It reported a survival rate of 78% in the early initiation group compared with 57% in the late initiation group ( P < .05). The early initiation group was also found to have had a significantly shorter ICU stay (12.5 vs. 8.5 days; P < .05), shorter hospital stay (20.9 vs. 15.4 days; P < .05), and lower rate of MOF (19% vs. 29%; P = .01). The second study, a retrospective analysis of post-CABG AKI using a historical control group, again showed significantly improved survival (77% vs. 45%; P = .016), shorter length of ICU stay (12 vs. 8 days ; P = .0001), and shorter length of hospital stay (30 vs. 15 days) in the early treatment group.
Clinical benefit of early initiation of RRT was also reported in a secondary analysis of a prospectively collected AKI database. Despite there being, on average, more failed organ systems in the early intervention group, the RR for death associated with delayed initiation was 1.85 (95% CI, 1.16 to 2.96) after covariate adjustment for age, hepatic failure, sepsis, thrombocytopenia, serum creatinine, study site, and initial dialysis modality.
Although these observational studies generally support earlier commencement of RRT, available higher-level evidence is less convincing. In a prospective RCT of 106 patients examining the effects of timing of initiation of dialysis and dose of dialysis on 28-day survival rates in AKI, there was no survival advantage to early initiation of RRT (survival 69% in the early low-volume group vs. 75% in the late low-volume group, nonsignificant). In addition, and of particular interest, the authors did not find a survival advantage to higher-dose therapy compared with lower-dose therapy (survival 74% in the high-volume group vs. 69% in the low-volume group, nonsignificant). In this trial, patients were randomized to three different treatment groups: an early high-volume hemofiltration group, an early low-volume hemofiltration group, and a late low-volume hemofiltration group. “Early treatment” was defined by treatment initiation within 12 hours of meeting the study’s AKI definition, whereas “late treatment” was initiated only when the patient’s BUN was higher than 112 mg/dL or hyperkalemia (>6.5 mmol/L) or pulmonary edema developed. Mean BUN in the early treatment group was 48 mg/dL, compared with a mean BUN of 105 mg/dL in the late treatment group. Unfortunately, this study was underpowered to detect a clinically significant treatment effect; six patients in the late group did not require dialysis because they recovered renal function or died.
A recent meta-analysis evaluated the evidence for and against early initiation of RRT in AKI. Two main questions were asked: (1) Does early RRT improve survival? and (2) Is early initiation of RRT associated with improved renal recovery? Marked heterogeneity was noted among study groups in terms of population settings, baseline disease severity, cutoff value definitions of early compared with late initiation, dialysis technique, and duration of study follow-up. The overall study method quality scores were low, and most trials (78%) were observational in nature. Primary analysis of the five included randomized trials that concluded that early RRT was associated with a 36% mortality risk reduction (not significant, P = .08). A secondary analysis of nonrandomized trials supported this hypothesis (26% mortality risk reduction; P < .001). The meta-analysis of renal recovery included two RCTs and five comparative cohort studies—there was no significant difference in outcomes.
There is clearly a need for a large, multicenter RCT to confirm or refute these hypotheses. The development of novel biomarkers that might estimate the severity of renal injury more accurately than current methods (creatinine, urea, urine output) and better predict likelihood of spontaneous renal recovery would assist greatly in informing the decision to commence early RRT.
Until such time as more definitive evidence is available to confirm the role of earlier initiation of RRT in improving outcome, clinicians must perform a risk-to-benefit analysis for each patient on a case-by-case basis. Decisions can be aided by expert-derived management guidelines, such as the U.K. Renal Association Clinical Practice Guidelines, relating to timing of initiation of renal replacement treatment in AKI ( Table 57-3 ).
Related posts:
What Lessons Have Intensivists Learned During the Evidence-Based Medicine Era?
What Is the Role of Invasive Hemodynamic Monitoring in Critical Care?
Should Fever Be Treated?
Are Computerized Algorithms Useful in Managing the Critically Ill Patient?
Inhaled Vasodilators in ARDS: Do They Make a Difference?
How Can We Monitor the Microcirculation in Sepsis? Does It Improve Outcome?
Full access? Get Clinical Tree
