PRECISE Trial: Improving Sepsis Treatment with Balanced Fluids using Machine-Learning Algorithm
An innovative PRECISE trial has discussed the potential benefits and significance of Precision Resuscitation With Crystalloids in Sepsis (PRECISE). This trial addressed the critical clinical question of improving the sepsis treatment by personalizing the fluid resuscitation using a machine-learning-based system, especially for a specific subphenotype known as group D, who may benefit more from balanced crystalloids than from normal saline. The trial was published in the journal JAMA Network Open.
Sepsis is the most common cause of death globally. This can be better managed using intravenous fluids, which can have a major global health impact. Intravenous fluids play an essential part in resuscitation and are most commonly prescribed to hospitalized patients, especially in patients with sepsis. Despite the abundance of intravenous fluids in sepsis, there is uncertainty about the use of balanced crystalloids or normal saline, the 2 major classes of crystalloid fluids. Hence researchers conducted a trial to identify patients with group D sepsis and categorize them based on the need for crystalloids or saline using a machine-learning algorithm approach. The trial tested whether the targeted use of balanced crystalloids through an electronic health record (EHR) alert system can reduce 30-day inpatient mortality in group D sepsis patients compared to usual care (normal saline).
The PRECISE trial is a parallel-group, multihospital, single-blind, pragmatic randomized clinical trial conducted in 6 hospitals within the Emory Healthcare system. Patients suspected of group D sepsis in the emergency department (ED) or intensive care unit (ICU) will be randomized to receive either the usual care (normal saline) or be nudged towards balanced crystalloids via an EHR alert. The primary intervention is an EHR alert prompting clinicians to use balanced crystalloids instead of normal saline. The primary outcome of measurement is 30-day inpatient mortality while the secondary outcomes included ICU admission, in-hospital mortality, the need for vasoactive drugs, kidney replacement therapy, or mechanical ventilation, which are counted only if they occur after randomization and within the 30-day study period.
The trial is notable for its precision medicine approach, utilizing a machine learning algorithm based on routine vital signs to identify patients with the group D sepsis subphenotype. This is one of the first trials to subcategorize sepsis. If successful, this trial could lead to a shift in international guidelines, establishing precision resuscitation practices that tailor fluid management in sepsis to individual patient subphenotypes. This trial helps to customize crystalloid prescriptions to individual needs instead of giving one type of fluid to all. The implications could be especially impactful in low-resource settings, as the necessary data for the machine learning algorithm (vital signs) are widely available.
Upon completion, the findings will be shared in a peer-reviewed journal, potentially influencing future resuscitation protocols.
Further reading: Bhavani SV, Holder A, Miltz D, et al. The Precision Resuscitation With Crystalloids in Sepsis (PRECISE) Trial: A Trial Protocol. JAMA Netw Open. 2024;7(9):e2434197. doi:10.1001/jamanetworkopen.2024.34197
