Stewardship has not become as standard an approach to fluids as it has to antibiotics in U.S. ICUs, as shown by a session at the annual meeting of the Society of Critical Care Medicine, held in Orlando in February.
“How many of you in the room have an actual protocol for deresuscitation of your patients in the ICU?” asked speaker Ty H. Kiser, PharmD, a professor of clinical pharmacology at the University of Colorado School of Pharmacy and Pharmaceutical Sciences in Aurora. No attendees raised their hands.
Yet there's growing evidence that physicians should be concerned about fluid balance in patients treated for sepsis, explained Timothy J. Ellender, MD, another session speaker and assistant professor of clinical emergency medicine at Indiana University School of Medicine in Indianapolis.
An early piece of evidence was the Fluids and Catheters Treatment Trial (FACTT), published by the New England Journal of Medicine in 2006, which found improvements in lung function and duration of mechanical ventilation and ICU care in patients randomized to conservative fluid management.
“The FACTT trial probably did the most . . . to show that, in fact, a conservative management strategy and actually taking fluids away from patients had a major impact,” said Dr. Ellender. More recent evidence has shown worse outcomes associated with higher fluid balances in ICU patients with acute renal failure and septic shock, he noted.
“Not only for seven days or 14 days, but in fact fluid balance over the first three to four to five days had an impact,” Dr. Ellender said, citing a prospective study of acute lung injury patients that was published in the Journal of Intensive Care Medicine in 2009. “When we began to impact that balance and drive patients toward a negative fluid balance, we were able to show mortality benefit.”
Based on these and other findings, the two speakers made their case for physicians to deresuscitate patients after they've been successfully resuscitated from septic shock. “I'm going to try to give you a practical approach to managing your fluids and diuretics,” said Dr. Kiser.
How to do it
The process starts with understanding that fluid resuscitation is a careful balance. “Hypovolemia is associated with risk of complications, and hypervolemia is associated with risk of complications,” said Dr. Kiser, citing a study published by Pharmacotherapy in January. “In this U-shaped curve, positive fluid balance early on—somewhere between two and four liters, which is somewhere around that 30 cc's per kilo, in most patients seems to be reasonable. But as you get over six liters, you're starting to run into trouble.”
That means it's still appropriate to follow existing guidelines for fluid resuscitation of patients with sepsis. But the experts recommended starting to think about deresuscitation shortly afterward. “Like most things in the ICU, the sooner we address this, the better,” Dr. Kiser said. “The first 24 hours, you can't really see a huge difference in death based upon your early resuscitation efforts and fluid balance, so adequately resuscitating patients is still very important. But as you get into that 72-hour window, that's when we really start to see major differences.”
Starting on the third day of treatment, too much fluid appears to be associated with worse outcomes, according to an observational trial that compared fluid balance and mortality in 1,800 sepsis patients and was published in the March 2017 Critical Care Medicine. “Survivors are getting a pretty even balance of intake and output by day three, whereas in the nonsurvivors, there's a substantial positive fluid balance,” Dr. Kiser reported.
Although earlier is better, deresuscitation should wait until the patient is stable. “Stabilization does not always mean that you're completely off vasopressors. The patient has been adequately fluid resuscitated. They're stable, and the pressors are going down,” he said.
For some patients, passive deresuscitation will be sufficient. “If the patient has adequately functioning kidneys, then they may be able to keep up with the diuresis themselves,” Dr. Kiser said. In these patients, clinicians' main task will be to reduce the amount of fluid going in. Obviously, that means the fluids with which they were resuscitated, but also think more broadly.
“What is going into the patient besides those continuous infusions? We miss this quite a bit unless someone's looking very closely,” he said.
For example, a substantial percentage of fluid given to ICU patients comes in the form of medication diluents, electrolyte replacements, and volume to keep venous access open, according to a study published by Intensive Care Medicine in 2018. “It translates to 25% of your IV volume is coming from your maintenance and carrier fluids, 63% of your IV volume is coming from medication diluents, on average,” cited Dr. Kiser.
Those other fluids can also contribute to chloride and sodium imbalances, he noted. Reducing them should be a priority during deresuscitation. “Day one, yes, resuscitation fluids are a big part. But as you go to day two, days three through seven, these other fluids, this fluid creep is really where the bulk of your fluids are coming from,” he said.
For some patients, restriction will be enough to get them back to fluid balance, but many will need additional interventions. “About 40% to 50% of your patients are going to need some diuretic assistance in addition to stopping all the intravenous intake you can. Their urine output will just not quite keep up,” said Dr. Kiser.
The typical strategy is furosemide at a starting dose of 20 mg, but the plan may often need to be tailored, he said. “As your serum creatinine goes up, not only do you use more of a diuretic, but you also may be combining your diuretics together to get a kind of synergistic approach.”
A combination of clinicians may also be needed. “The key here, and I think something we don't do very well, is when you're using diuretics, we need to be giving a dose and setting goals for our nurses, who are excellent at achieving whatever input-to-output ratio we want. And they can come back to us every four to six hours and tell us whether we're hitting that balance,” Dr. Kiser said.
In general, the goal will be a balance of input and output on day 3, with more fluid coming out than going in after that. “We want to work our way negative as tolerated between days three and seven,” he said.
Toleration is an important point, as there are risks to deresuscitation. “Things to look out for are going to be hypotension, electrolyte abnormalities, worsening kidney function, and metabolic alkalosis,” said Dr. Kiser.
There's also the opposing risk that diuretics won't get the job done. “More and more patients are getting renal replacement therapy in order to address the fluid imbalance,” he said. “We know our ICU patients are very heterogeneous, so it's all going to be very individualized.”
It would be helpful to have some markers to better individualize deresuscitation, said Dr. Ellender. “We've talked eloquently about the need to identify the resuscitative phase versus the deresuscitative phase, but just where does my patient fall? And how do I measure basic endpoints to decide when to say when?”
He offered a few examples of potential endpoints to measure. “Pitting edema, fluid balance, cumulative body weight—those things give us a superficial gestalt for what the patient looks like and perhaps where our fluid balance may lie.”
It would be useful to also have biochemical markers, but most of the available ones, such as albumin-to-creatinine ratios and lactate changes, are more helpful during resuscitation than deresuscitation. “Things like osmolar units can be used but are very hard to use at the bedside in any meaningful way,” said Dr. Ellender.
Biometric measures, including the passive leg raise and stroke volume variation, may work better. “Last but not least, imaging, especially ultrasound, can help us tell more about extravascular lung water, intra-abdominal fluid, [and] hepatic and renal congestion, so we can use these as adjuncts with our bedside examination to tell us more data,” he said.
Physicians should also be keeping an eye on organ function during deresuscitation, using markers like cerebral performance. “In this phase, there are no great mediators for endpoints, but there are perhaps indices,” Dr. Ellender said.
He urged particular caution in deresuscitation of patients who didn't handle the resuscitation phase well, for example, by swelling up dramatically during treatment. “These are the people where fluid overload itself may contribute to end-organ injury, and here we need to avoid enthusiastic fluid removal, because we can also drive additional hypovolemia, which begins the cycle again,” he said.
Dr. Ellender also offered a more general caveat on the lack of definitive proof that there's a causal relationship between fluid imbalance and mortality rather than just correlation. “Is this really just a marker of illness, perhaps in those that are severely ill? Or is it a contributor to illness? We don't yet know that, but ongoing studies hope to tell us more,” he said.
The Role of Active Deresuscitation After Resuscitation-2 (RADAR-2) trial is one such study, a randomized trial of deresuscitation in critically ill patients. “These patients are all being resuscitated according to appropriate strategies,” explained Dr. Ellender. After that, they're treated with either usual care or a conservative fluid strategy, including restriction of fluids and diuretic therapy. “The primary outcome is fluid balance achieved within day three, change in fluid balance, and then a host of different mortality, morbidity, and clinical outcome endpoints,” he said.
If those endpoints are significantly better in the conservative group, fluid stewardship may soon become as common as antibiotic stewardship. But Dr. Kiser doesn't want physicians resuscitating sepsis patients to wait on deresuscitation until then. “At least put it on your radar to keep track of,” he said.