Fluid Overload in Critically Ill Patients with Acute Kidney Injury

Fluid overload may occur in critically ill patients as a result of aggressive resuscitation therapies. In such circumstances, persistent fluid overload must be avoided since it does not benefit the patient while it may be harmful. In the septic patient, early volume expansion seems to be beneficial. Beyond that threshold, when organ failure develops, fluid overload has been shown to be associated with worse outcomes in multiple disparate studies. One well-designed randomized controlled trial showed the benefit of a conservative fluid management strategy based on limited fluid intake and use of furosemide in such patients. Use of diuretics should be only short term as long as it is effective, generally at high doses, while avoiding simultaneous utilization of nephrotoxins such as aminoglycosides. Multiple randomized controlled trials have not shown benefit in the use of diuretics, either to prevent AKI or to treat established AKI. If fluid overload (defined as fluid accumulation 1 10% over baseline) develops and the patient does not respond to diuretics, persistent use of these drugs will only lead to a delay in the initiaPublished online: February 19, 2010 Jorge Cerda, MD Capital District Renal Physicians 62 Hacket Blvd Albany, NY 12209 (USA) Tel. +1 518 434 2244, Fax +1 518 434 4659, E-Mail cerda @ nycap.rr.com © 2010 S. Karger AG, Basel 0253–5068/10/0294–0331$26.00/0 Accessible online at: www.karger.com/bpu D ow nl oa de d by : 54 .7 0. 40 .1 1 11 /1 9/ 20 17 7 :3 7: 04 A M Cerda /Sheinfeld /Ronco Blood Purif 2010;29:331–338 332 tional outcomes [6] and may be associated with worse patient and functional survival [7–9] . Their use should be limited to optimizing fluid management whenever possible, and to help control electrolyte disorders including hyperkalemia, hypermagnesemia and hypercalcemia [7] . Given the importance of prompt correction of fluid overload, when diuretics fail renal replacement therapies should be initiated without delay. Among critically ill patients with AKI, several observational studies initially performed in the pediatric population [10–13] suggest an association between fluid overload and increased mortality. An adverse effect of fluid overload on outcomes has been reported in other critically ill patients without AKI, including patients with septic shock and acute lung injury [14–19] . In these latter patient groups, aggressive fluid resuscitation is frequently provided to manage hypotension [20] . Unfortunately, this approach often results in significant fluid overload, especially when renal function is compromised [21] . On this subject, a ‘chicken and egg’ debate has occurred, one side arguing that severe fluid overload indicates a very sick patient who requires large volumes of fluid to maintain hemodynamic stability in the context of pronounced critical illness. On the other side, increasing evidence shows that fluid overload itself could be seen as a ‘biomarker’ with intrinsic toxic effects on organ function, contributing independently to the elevated mortality of critically ill patients [8, 22] . In this article, we will successively discuss the evidence showing the association between fluid overload and patient outcomes among hemodynamically unstable patients with and without AKI, and briefly review current knowledge on techniques available for volume control. Fluid Balance in Critically Ill Patients Critically Ill Patients without AKI In a landmark randomized controlled trial (RCT) performed by Rivers et al. [23] , a highly protocolized resuscitative approach (early goal-directed therapy; EGDT) was compared to conventional management in 263 emergency department patients admitted with sepsis-induced hemodynamic compromise. This treatment bundle formed the basis of the Surviving Sepsis Campaign guidelines [24] . Confirmatory EGDT studies are currently being carried out in a larger, multicentric 5-year RCT of alternative resuscitation strategies for septic shock (the ProCESS trial, ClinicalTrials.gov Identifier NCT00510835). This study is currently assessing the reproducibility of the initial findings. In the meantime, variations of this strategy are widely applied around the world [25] . The EGDT, conducted for 6 h in the emergency department, involved adjustments of cardiac preload, afterload, and contractility to balance oxygen delivery with oxygen demand. Central venous pressure, mean arterial pressure, and central venous oxygen saturation were corrected to predetermined targets. A major aspect of the protocol included initially aggressive volume administration. While the total amount of volume administered within the first 72 h was approximately the same in both arms (a mean of 13,357 8 7,729 vs. 13,443 8 6,390 ml), patients randomized to EGDT received the majority of volume within the first 6 h, and patients on standard therapy received the infusion between 7 and 72 h after admission. Rather than the amount of volume administered, the critical difference was the timing of the administration: early up-front aggressive resuscitation was effective, but late and excessive fluid administration was associated with worse survival at hospital discharge, 30 and 60 days. The survival difference was significant in patients with septic shock and severe sepsis but not in those with sepsis syndrome. The mean serum creatinine in both groups was 2.6 mg/dl at baseline, suggesting a high incidence of AKI even before fluid administration began, although inclusion of end-stage renal disease patients could partly explain the elevated mean serum creatinine at baseline. A significant percentage (more than 20%) of patients had pre-existing renal insufficiency, placing them at increased risk for the development of AKI. Rivers et al. [23] did not specifically report the percentage of patients with AKI at entry and the subsequent development of AKI (including the need for renal replacement therapy) after treatment interventions. As severe sepsis and septic shock have been shown to be risk factors for AKI [26, 27] and in view of the relatively high mean baseline serum creatinine, a reasonable assumption is that AKI progressed in a significant percentage, with many eventually requiring renal replacement therapy. Based on the volume of fluid used for resuscitation, the possibility that many patients, especially in the conventional management group, needed dialysis driven by AKI and fluid overload is concerning. The clinical implications of fluid overload in sepsis syndromes (see below) raise questions about the EGDT study. Alsous et al. [14] retrospectively evaluated the effect of fluid overload on outcome in septic shock in 36 patients. D ow nl oa de d by : 54 .7 0. 40 .1 1 11 /1 9/ 20 17 7 :3 7: 04 A M Fluid Overload in Acute Kidney Injury Blood Purif 2010;29:331–338 333 Mortality outcomes were adjusted for age, baseline APACHE II score, SOFA score on days 1 and 3, and the need for mechanical ventilation. For patients in whom a net negative fluid balance of at least 500 ml could be achieved within the first 3 days of ICU hospitalization, survival was 100%. Conversely, survival of patients in whom net negative fluid balance could not be attained was only 20%. During the 7-day analysis period, daily fluid balance on average was always more positive among patients who died than those who survived. From day 2 to day 7, surviving patients essentially attained neutral fluid balance on a daily basis, while nonsurviving patients had mean positive fluid balance ranging from 1 to 4 liters per day. The differences between surviving and nonsurviving patients were greatest (approximately 3.5– 4 liters) on days 2 and 3. Combining aspects of both the Rivers and Alsous [14, 23] studies, Murphy et al. [15] recently performed an observational analysis of 212 patients who developed acute lung injury in the setting of septic shock. These investigators assessed the effects of timing of fluid resuscitation and control of fluid balance on patient outcome. Adequate initial fluid resuscitation was achieved by administration of an initial fluid bolus of 6 20 ml/kg prior to and a central venous pressure of 6 8 mm Hg within 6 h after the onset of vasopressor therapy. Conservative late fluid management was defined as neutral or negative fluid balance on at least 2 consecutive days during the first 7 days after the onset of septic shock. Patients meeting both the initial resuscitation and fluid balance criteria showed the lowest hospital mortality, while patients meeting only one or neither of these criteria had significantly worse outcomes. Hospital mortality in patients meeting both criteria was 18.3%, compared with 77.1% among those meeting neither (p ! 0.001) ( fig. 1 ). Consistent with the Alsous study, daily fluid balance in nonsurviving patients was significantly greater (and always positive: +2,500 to +1,000 ml) than in surviving patients (+1,300 to –500 ml) on days 3–7, while surviving patients achieved zero daily fluid balance on day 4. Moreover, cumulative fluid balance differed significantly between nonsurviving and surviving patients during the same time period (day 7 values +13.7 and +8.0 liters, respectively). Multivariate regression analysis demonstrated that the inability to achieve conservative late fluid management (a likely surrogate for fluid overload) was an independent mortality risk factor (adjusted OR 6.13; p ! 0.001). In a multicentric observational cohort including 198 ICUs from 24 European countries (the Sepsis Occurrence in Acutely Ill Patients (SOAP) Trial), Vincent et al. [16] compared clinical parameters of 1,177 septic with 1,970 nonseptic patients. Fluid balance was assessed within the first 72 h of ICU admission, with mean cumulative fluid balance in the septic group during this period being +1.8 vs. +0.5 liters in the nonseptic group. Multivariate regression analysis demonstrated that mean fluid balance during this time period was a significant independent outcome predictor: each 1-liter increase in cumulative fluid balance was associated with a 10% increase in risk of death. The above three studies [14, 16, 23] conducted in septic patients, along with studies performed in other patient populations [17–19] , suggest that a positive fluid balance adversely impacts clinical outcomes. Unfortunately, as observational studies, these reports can only establish an association rather than a causal relationship. A recent RCT involving patients with acute lung injury provides substantive evidence of the deleterious effect of fluid overload in critically ill patients. In the Fluid and Catheter Treatment Trial (FACTT) performed by the ARDS Clinical Trial Network investigators [28] with a 2 ! 2 design, one intervention was the use of a pulmonary artery catheter or a central venous catheter for hemodynamic management. The other intervention was the use of a ‘conservative’ or ‘liberal’ fluid management strategy over a 1-week intervention period. The primary endpoint was 60-day mortality while secondary endpoints included ventilator-free days and organ failure-free days. Fluid administration and the use of vasoactive agents and fu0 H o sp it a lm o rt a li ty (% ) 10 20 30 40 50 60 70 80 90