Study into the Reversal of Septic Shock with Beta Blockade (Landiolol): the STRESS-L randomised trial

Background Autonomic dysfunction and tachycardia are associated with poor outcomes in septic shock. Noradrenaline is recommended to maintain blood pressure but has been associated with harm. Beta-blockade has been associated with improved survival. We hypothesised that beta1 receptor antagonism with landiolol would benefit patients with septic shock through myocardial protection and immunomodulation. Objectives To assess the efficacy, safety and mechanisms of landiolol in patients with septic shock and tachycardia requiring > 24 hours of treatment with noradrenaline. Design A randomised, open-label, phase IIb trial comparing usual treatment with usual treatment plus landiolol. This pragmatic trial was scheduled to recruit 340 patients. Setting Forty-one United Kingdom intensive care units. Participants The study randomised adult intensive care unit patients (≥ 18 years) with septic shock, who, having received adequate fluid resuscitation (for > 24 hours), were treated with ≥ 0.1 mcg/kg/minute noradrenaline and had a heart rate of ≥ 95/minute. Interventions Continuous intravenous landiolol infusion; the dose was titrated to achieve a heart rate (80–94 bpm). Main outcome measures The primary outcome was mean Sequential Organ Failure Assessment score over the first 14 days after randomisation. Secondary end points included patient-centred, safety and exploratory mechanistic outcomes. Results The study was terminated prematurely on the advice of the independent Data Monitoring Committee that landiolol was unlikely to demonstrate benefit and was a signal for possible harm. A total of 4137 patients were screened and 126 patients were randomised: 63 landiolol and 63 standard care. The mean (standard deviation) Sequential Organ Failure Assessment score over the 14 days was 8.8 (3.9) for the landiolol group compared with 8.1 (3.2) for the standard care group. There was no statistical difference between groups [mean difference 0.75 (95% confidence interval −0.49 to 2.0); p = 0.24]. Mortality at day 28 was 37.1% (23 of 62) in the landiolol group and 25.4% (16 of 63) for standard care [absolute difference, 11.7% (95% confidence interval, −4.4% to 27.8%); p = 0.16]. The Cox proportional hazards model to day 28 demonstrated no difference in survival [unadjusted hazard ratio, 1.64 (95% confidence interval, 0.87 to 3.10); p = 0.13]. There were reports of adverse events in 17.5% (10 of 63) in landiolol and 12.7% (8 of 63) in standard care ( p = 0.80). However, 25.4% (16 of 63) of landiolol patients versus 6.4% (4 of 63) in the standard care group ( p = 0.006) experienced serious adverse events. There were no detectable differences between groups in measured cytokine or metabolomic analysis. Limitations The primary outcome was selected because it had been previously used in other septic shock trials but does not deal well with deaths and discharges from intensive care unit. Data are lacking on cardiac function. By stopping prematurely, the trial may not have sufficient power to describe clinically important effects. Conclusions The STRESS-L was stopped after recruiting 126 of 340 patients because it was unlikely to demonstrate benefit and there was a signal of possible harm. Mechanistic studies did not reveal significant pathways through which landiolol may be exerting effect. Future work A planned post hoc analysis will attempt to identify phenotypes of patients who may have benefited or been harmed during STRESS-L to inform future beta-blocker trials.