Four-Leafed CLOVERS or Otherwise?
By: Steven Q. Simpson, MD, FCCP
September 28, 2018
How much IV fluid should we give to a septic, hypotensive patient—or, for that matter, to any hypotensive patient—before we initiate a vasopressor? It engenders a wry smile that this question remains unanswered because it precedes my entry into the profession. I recall being told as a third- year medical student, “You have to fill the tank before you squeeze.” Otherwise, you will “kill the kidneys.”(That was in 1981, by the way. One could still visit the last of the dinosaurs at a few select zoos around the world.) However, if one asked the residents and attendings just how much fluid this should take, they were met with a rainbow of opinion, ranging from “some” to “it’s not a resuscitation, until it’s (liters) in the double digits.”
That the question remains unanswered is a sociomarker of how physicians, including critical care physicians, can often leave simple, fundamental questions unanswered. Often, it is because we accept the current standard, or nonstandard, as it were—as good enough. Sometimes, we think there are more important places to focus our intention, such as on a sepsis drug or on goal-directed therapy, as a whole. But probably most importantly, it is just plain not easy to marshal the ideas, the personnel, and the study design to answer a tough question, even when it affects every one of our patients and the decision is taken every day—especially when physicians have disparate opinions on the correct approach.
The CLOVERS Trial
The ongoing, NIH-sponsored Crystalloid Liberal or Vasopressors Early Resuscitation in Sepsis (CLOVERS) trial approaches this question by comparing higher volumes of fluid administration before initiation of vasopressors with lower volumes on the hypothesis that lower volumes will lead to improved outcomes. Hypotensive septic patients will be assessed for study entry after receiving one liter of crystalloid and may receive no more than two additional liters before entering the study. Once randomized, patients are given either norepinephrine infusion or additional fluids.
In the liberal fluid arm of the study, two liters are given immediately upon study entry, and additional fluid boluses are given, up to a total of 5 liters, before norepinephrine is initiated. Additional 500 mL boluses can be given by the managing physicians in both arms, according to physician discretion. The primary outcome is in-hospital all-cause mortality, with secondary outcomes of ventilator and dialysis-free days, ICU and hospital length of stay, duration of vasopressor use, changes in SOFA and KDIGO scores, and variations on all of those. Given the recent literature, mostly retrospective, that demonstrates an association of high volumes of IV fluids with mortality, this is an important randomized trial to undertake.
The Critique: Right or Wrong, Standard or Variable
Interestingly, this study has been vociferously criticized as unethical and has been reported in the New York Times and on National Public Radio. The criticisms span the spectrum of what is unethical, from too little fluid to too much fluid, and that physicians are restricted from choosing a middle ground. The vociferous opponents to the study assert that neither arm represents “standard of care.” Most of them are not intensivists, especially those who lead the charge. While any given intensivist may not choose either of these pathways, it is clear that some do. It is likewise clear that the two arms of the study represent, perhaps, the upper and lower bounds of fluid administration practices in the ICU.
The critics of this trial seem to be neglecting several important issues. First, there is no known scientifically demonstrated answer to this question, and the physician judgment that the critics say will be lacking has clearly not been demonstrated to be effective. It is an illusion to believe that something about physician judgment in a situation with no clearly demonstrated gold standard is safer for patients. The study was proposed and developed by numerous accomplished physician scientists who are experienced ICU docs. The study was reviewed by an NIH study section, also accomplished physician scientists in this area of research. It was approved by a national human subjects protection committee and by multiple local institutional review boards (IRBs). Patients are protected both morally and legally.
My personal concern with the study is not the administration of high or low IV fluid volumes, per se, but that it ignores modern technology that can individually right-size the amount of fluid. I would rather see a protocol arm that compares stroke volume or echocardiographic techniques for volume management and optimizes that parameter on an individual patient basis to either or both of the arms in the CLOVERS trial. Additionally, the allowance for anywhere from one to three liters of fluid before study initiation is likely to lead to baseline variability, as is the lack of accounting for patient size, which could be accommodated if fluids were given on a mL/kg basis. Nevertheless, it is a step in the right direction to finally be investigating this fundamental issue that affects us all every day. The data will very likely contribute to saving lives of patients with septic shock.
Steven Q. Simpson, MD, FCCP, FACP, is a sepsis expert and researcher, a CHEST member, and a Professor of Medicine in the Division of Pulmonary and Critical Care Medicine at the University of Kansas. Follow Dr. Simpson on Twitter @sqsimp.