Practice Variation in the Prescription of Volume Removal with Acute KRT

The approach to fluid management in the intensive care unit (ICU) has evolved over the last several decades from the adage that patients must swell to get well to an increasing appreciation that volume overload is a marker and likely a mediator of poor outcomes in critical illness. Along with several large multicenter randomized controlled trials (RCTs) demonstrating no benefit to early goal-directed therapy, dozens of observational studies have shown that fluid overload in critical illness is independently associated with higher rates of mortality, development and/or persistence of AKI, and other organ dysfunction. Consequently, critical care medicine has shifted focus to studying fluid restriction, but there are limited prospective data demonstrating its benefit. The 2006 Fluid and Catheter Treatment Trial (FACTT) is possibly the most compelling RCT supporting fluid restriction because it demonstrated that a conservative fluid strategy in patients with acute respiratory distress syndrome resulted in shorter duration of mechanical ventilation and ICU length of stay, with a trend toward lower need for kidney replacement therapy (KRT).¹ However, while FACTT helped trigger a swing of the pendulum away from the routine aggressive use of intravenous fluids in the ICU, it was a negative study regarding the primary mortality end point. Subsequent large multicenter RCTs analyzing fluid restriction in the ICU have repeatedly yielded no benefit or even signals for harm.² Although additional multicenter trials are ongoing (NCT04569942, NCT05179499), large RCTs with positive outcomes demonstrating how to best use intravenous fluids in the ICU are lacking.

Despite the strong association between fluid overload and adverse outcomes, even fewer data exist to guide clinicians on the optimal approach to deresuscitation, which involves the active removal of fluid by means of diuretics or KRT after the initial resuscitation phase of critical illness. In the case of fluid removal with KRT, observational studies have shown that higher overall achieved ultrafiltration rates and more negative fluid balances are associated with improved outcomes.³ However, other observational studies suggest that higher weight-adjusted net ultrafiltration rates (>1.75 ml/kg per hour) with continuous KRT (CKRT) are independently associated with adverse outcomes,⁴ suggesting that the effects of ultrafiltration in critically ill patients are complex. While prompt normalization of volume status with net ultrafiltration may mitigate harm associated with volume overload, this potential benefit must be balanced with the risk of inducing hemodynamic compromise and end-organ injury. For example, studies have shown that both intermittent hemodialysis and CKRT for AKI can induce myocardial stunning,⁵ as previously demonstrated in end-stage kidney disease. The pathophysiology of intradialytic hypotension in critically ill patients is complicated, with mechanistic studies suggesting that only some hypotensive episodes are caused by volume removal.⁶ This implies that some patients may benefit from ongoing net ultrafiltration despite experiencing intradialytic hypotension. Without prospective data to guide clinicians on the optimal approach to ultrafiltration in the ICU, the best way to remove excess fluid with KRT in critically ill patients remains unclear.

How nephrologists and intensivists deal with this realm of complex pathophysiology and limited data are the subject of an international survey conducted by Ledoux-Hutchinson and colleagues, the results of which are published in this issue of CJASN.⁷ Over 750 critical care and nephrology providers and trainees from 96 countries were surveyed to evaluate practice patterns related to fluid removal with KRT in critically ill patients. The key—if unsurprising—finding of the survey was the marked variation in the approach to ultrafiltration, not only between providers of different geographic regions and specialties but also across different providers within these groups. The survey identified several key elements of practice variation, including the maximal daily net ultrafiltration rate considered safe, the extent to which providers feel comfortable using vasopressor support to achieve ultrafiltration goals, the tools preferentially used to guide volume assessment, and the interventions used to improve hemodynamic tolerance of ultrafiltration. Few physicians prescribed ultrafiltration adjusted for body weight and even fewer worked in ICUs with fluid stewardship policies or protocols guiding interruptions in ultrafiltration in response to hypotension or increased vasopressor needs. There was consensus among the survey respondents about the importance of promptly correcting volume overload whenever feasible, but equipoise existed regarding whether mitigating KRT-induced hypotension should be prioritized over the goal of correcting volume overload and whether moderate degrees of volume overload are acceptable in the absence of respiratory compromise. Survey respondents largely agreed that this is an important topic for future research and that conducting RCTs in this area would be ethical.

The limitations of this study are those largely inherent to the methodology of an internet-based self-report survey, including the risk of response bias. Surveying physicians about their practice patterns is not equivalent to directly observing their practices. In addition, as the authors note, the overrepresentation of physicians from academic centers may somewhat limit generalizability, although this is offset by the inclusion of respondents from almost 100 countries. Furthermore, the probability that practice variation would be consistently lower in nonteaching hospitals seems low. Moreover, the investigators used a reasonably rigorous process to identify and validate survey items.

This study is an important addition to a growing body of literature that demonstrates significant practice variation worldwide in fluid management with KRT in the ICU.^8,9 The next methodologic step could be an international observational study directly measuring how ultrafiltration is delivered in real-world practice, as has been previously performed for practice patterns related to other aspects of CKRT delivery. However, although such a study could be valuable, there are already sufficient data to argue that RCTs are necessary to better inform the provision of ultrafiltration with KRT in the ICU.

The authors propose four questions to be addressed in future RCTs: (1) What tools should be used to guide fluid removal? (2) When should fluid removal be initiated? (3) How quickly should fluid be removed? (4) How should vasopressor support be used for fluid removal? Small studies have examined the usefulness of various hemodynamic monitoring devices and volume assessment tools—including inferior vena cava collapsibility index by ultrasound, bioimpedance analysis, passive leg raise, and/or calibrated pulse contour analysis devices—in setting ultrafiltration goals or predicting intradialytic hypotension in critically ill patients. However, the results of these trials have been rather mixed, possibly because all these tools were originally developed to guide fluid administration rather than fluid removal, and it remains unclear whether the same methods used to evaluate fluid responsiveness or fluid tolerance can be applied in reverse to the assessment of ultrafiltration tolerance. Pilot trials are underway comparing different rates of net ultrafiltration with CKRT and comparing protocolized fluid removal strategies with usual care (e.g., GO NEUTRAL [NCT04801784], RELIEVE-AKI [NCT05306964], Probe-Fluid [NCT05473143]). The role of vasopressors in achieving ultrafiltration goals has not been studied. Additional potential areas for future research include (1) how to best maintain volume status during the transition from CKRT to intermittent HD and the potential role of prolonged intermittent KRT in managing this transition and (2) optimal approaches to prevention or treatment of intradialytic hypotension in the ICU, especially because recent randomized trials in other settings have casted doubt on the effectiveness of commonly used strategies such as oral vasopressors¹⁰ and dialysate cooling.¹¹

Furthermore, defining optimal outcomes for future critical care nephrology RCTs is crucial, given the recent repeated failures of critical care RCTs to achieve primary mortality end points. Trialists must use more sophisticated organ-specific end points that are meaningful to nephrologists, intensivists, and patients. For RCTs evaluating the optimal approach to ultrafiltration with KRT, kidney-specific metrics such as long-term KRT dependence or newer composite end points such as major adverse kidney events—a composite of death, KRT dependence, and persistent kidney dysfunction—should be considered as primary study end points alongside more traditional outcomes such as ventilator-free days or mortality.

As we experience a golden age of critical care trials, with large multicenter critical care RCTs seemingly being published monthly, the critical care nephrology community must advocate that we harness our trial infrastructure to evaluate the optimal approach to volume removal with KRT in critically ill patients. As data continue to accumulate on how to best provide or restrict intravenous fluids in critical illness, the comparative paucity of data to guide optimal fluid removal becomes increasingly stark. In the absence of such trial data, the lack of consensus on how to best prescribe ultrafiltration in KRT-dependent patients in the ICU—clearly documented by Ledoux-Hutchinson and colleagues—will only persist.

Disclosures

J.P. Teixeira reports having consultancy agreements with and serving on a speakers bureau for Outset Medical; having ownership interest using current or previous stocks or options in Novo Nordisk A/S; and receiving research funding from Astute Medical/bioMérieux, La Jolla Pharmaceutical Company, Pfizer, Rediscovery Life Sciences LLC, and Sentien Biotechnologies Inc. A.J. Tolwani reports having consultancy agreements with and serving on a speakers bureau for Baxter Healthcare, having consultancy agreements with Outset, having a patent on 0.5% trisodium citrate solution for CRRT anticoagulation (the license has been purchased by Baxter), serving on the editorial boards of CJASN and Kidney International, and receiving honoraria from UpToDate.

Funding

None.

Author Contributions

Writing – original draft: J. Pedro Teixeira.

Writing – review & editing: Ashita J. Tolwani.