Abstract
Purpose: The efficacy of the 1-h bundle for emergency department (ED) patients with suspected sepsis, which includes lactate measurement, blood culture, broad-spectrum antibiotics administration, administration of 30 mL/kg crystalloid fluid for hypotension or lactate ≥ 4 mmol/L, remains controversial.Methods: We carried out a pragmatic stepped-wedge cluster-randomized trial in 23 EDs in France and Spain. Adult patients with Sepsis-3 criteria or a quick sequential organ failure assessment (SOFA) score ≥ 2 or a lactate > 2 mmol/L were eligible. The intervention was the implementation of the 1-h sepsis bundle. The primary outcome was in-hospital mortality truncated at 28 days. Secondary outcomes included volume of fluid resuscitation at 24 h, acute heart failure at 24 h, SOFA score at 72 h, intensive care unit (ICU) length of stay, number of days on mechanical ventilation or renal replacement therapy, vasopressor free days, unnecessary antibiotic administration, and mortality at 28 days. 1148 patients were planned to be analysed; the study period ended after 873 patients were included.Results: 872 patients (mean age 66, 42% female) were analyzed: 387 (44.4%) in the intervention group and 485 (55.6%) in the control group. Median SOFA score was 3 [1-5]. Median time to antibiotic administration was 40 min in the intervention group vs 113 min in the control group (difference - 73 [95% confidence interval (CI) - 93 to - 53]). There was a significantly higher rate, volume, and shorter time to fluid resuscitation within 3 h in the intervention group. There were 47 (12.1%) in-hospital deaths in the intervention group compared to 61 (12.6%) in the control group (difference in percentage - 0.4 [95% CI - 5.1 to 4.2], adjusted relative risk (aRR) 0.81 [95% CI 0.48 to 1.39]). There were no differences between groups for other secondary endpoints.Conclusions: Among patients with suspected sepsis in the ED, the implementation of the 1-h sepsis bundle was not associated with significant difference in in-hospital mortality. However, this study may be underpowered to report a statistically significant difference between groups.