CLINICAL QUESTION

For routine intravenous fluid administration in hospitalized patients, does lactated Ringer solution improve outcomes more than normal saline?

BOTTOM LINE

Hospital-wide administration of lactated Ringer solution vs normal saline for hospitalized patients did not significantly reduce 90-day mortality or hospital readmission. (Level of Evidence = 1b)

SYNOPSIS

In the cluster randomized, multicenter, open-label study from Canada, seven academic and community hospitals were assigned to hospital-wide lactated Ringer solution or normal saline for routine intravenous administration of fluids. Week 1 served as a run-in, the assigned study fluid was administered during weeks 2 to 13, and weeks 14 and 15 served as a washout period. After the washout period, hospitals switched to the other fluid for 12 weeks. Patients with an index admission during weeks 2 to 13 were included in the analysis, whether they received the study fluid or not. The exclusion criteria included patients younger than 1 month, those missing a birth date, or those not hospitalized in the past 90 days. Overall, 22,017 patients in the lactated Ringer solution group and 21,609 patients in the normal saline group were analyzed. Baseline characteristics in the groups were similar. Adherence to the study fluid was 78% in the lactated Ringer solution group and 94% in the normal saline group. There was no significant difference in the primary composite outcome of death or readmission within 90 days when comparing the groups (20.3% in the lactated Ringer solution group vs 21.4% in the normal saline group). Secondary outcomes, including emergency department visits, dialysis initiation, and discharge to a facility other than home, also did not differ. A subgroup analysis that excluded patients hospitalized for obstetric, elective surgical, or psychiatric conditions did not change the results. Because of the COVID-19 pandemic, only seven hospitals were recruited in this trial, rather than the targeted 16 sites, which may limit the study's power to detect small differences, if they exist.