The Effect of Fluid Initiation Timing on Sepsis Mortality: A Meta-Analysis

Abstract

Objective: Current guidelines suggest the immediate initiation of crystalloid for sepsis-induced hypoperfusion but note that supporting evidence is low quality. The aim of this study is to examine the effect of timing of fluid initiation on mortality for adults with sepsis.

Data sources: Two authors independently reviewed relevant articles and extracted study details from PubMed, Scopus, Cochrane, Google Scholar, and previous relevant systematic reviews from 1-1-2000 to 1-6-2022. Registered with PROSPERO (CRD42021245431) and bias assessed using CLARITY.

Study selection: A minimum of severe sepsis (Sepsis-2) or sepsis (Sepsis-3) for patients ≥18 years old. Fluid initiation timing ranging from prehospital to 120 min within sepsis onset defined as "early" initiation.

Data extraction: Included studies providing mortality-based odds ratios (or comparable) adjusting for confounders or prospective trials.

Data synthesis: From 1643 citations, five retrospective cohort studies were included (n = 20,209) with in-hospital mortality of 21.8%. A pooled analysis (odds ratio = OR [95% CI]) did not observe an impact on mortality for the early initiation of fluids among all patients, OR = 0.79 [0.62-1.02]; heterogeneity: I² = 86% [70-94%], but when studies analyzed cases of hypotension where available, a survival benefit was observed, OR = 0.74 [0.61-0.90]. Initiation of fluids in two prehospital studies did not impact mortality, OR = 0.82 [0.27-2.43]. However, both prehospital cohorts observed benefit among hypotensive patients individually, although heterogenous results precluded significance when pooled, OR = 0.50 [0.21-1.18]. Three hospital-based studies with initiation stratified at 30, 100, and 120 min, observed survival benefit both individually and when pooled, OR = 0.78 [0.63-0.97]. No differences were observed between prehospital versus hospital subgroups.

Conclusion: This meta-analysis supports the guideline recommendations for early fluid initiation once sepsis is recognized, especially in cases of hypotension. Findings are limited by the small number, heterogeneity, and retrospective nature of available studies. Further retrospective investigations may be worthwhile as randomized studies on fluid initiation are unlikely.

Keywords: fluids; initiation; meta-analysis; sepsis; septic; shock; timing.