Effect of early adjunctive vasopressin initiation for septic shock patients: a target trial emulation

Abstract

Background: In septic shock, the optimal timing of adjunctive vasopressin initiation shock is unknown. We aimed to assess the effect of its early initiation for patients with septic shock.

Methods: We conducted a multicenter target trial emulation to estimate the intensive care unit (ICU) mortality effect of early (≤ 6 h) adjunctive vasopressin compared with usual care. Eligible patients had septic shock diagnosed within 6 h of ICU admission. The primary outcome of this study was 30-day ICU mortality. Subgroup analyses were conducted to test the interaction of early vasopressin start with peak norepinephrine-equivalent dose (NED) at 6 h, APACHE score, peak lactate at 6 h and invasive mechanical ventilation. Secondary outcomes were the impact of delayed vasopressin introduction on 30-day ICU mortality and effect of NED at vasopressin start on 30-day ICU mortality. We used the parametric g-formula to emulate a target trial.

Results: Overall, 3,105 patients fulfilled the inclusion criteria. Mean age was 62 years and mean APACHE III score was 83. In the first six hours of vasopressor therapy, 1,864 (60%) patients were invasively ventilated. Estimated 30-day ICU mortality was 19.34% (95%CI, 17.0 to 21.68) in the no vasopressin group and 18.45% (95%CI, 16.26 to 20.63) in the early vasopressin group; relative risk 0.95 (95%CI, 0.93 to 0.98). The estimated 30-day ICU mortality effect of starting vasopressin was particularly strong at lower norepinephrine doses (< 0.25 µg.kg^-1.min^-1) and significant at lower norepinephrine doses than recommended by the Surviving Sepsis Campaign Guidelines. Vasopressin administration progressively increased over the study period, from 35.2% (95%CI, 30.0 to 40.5) in 2015 to 45.1% (95%CI, 40.7 to 49.6) in 2021 (ß = + 1.3% per year; 95%CI, + 0.46 to + 2.16, p = 0.011). Patients had progressively lower norepinephrine equivalent dose (ß = - 0.05 µg.kg^-1.min^-1 per year; 95%CI, - 0.09 to - 0.002, p = 0.038) and lower total SOFA score (ß = - 0.1 point per year; 95%CI, - 0.18 to - 0.07, p < 0.001) at vasopressin start.

Conclusions: In this emulation of a hypothetical target trial, patients with septic shock benefited from early vasopressin administration. These findings can help design prospective randomised-control trials of early adjunctive vasopressin use in septic shock.

Keywords: Critical care; Hypotension; Sepsis; Shock; Vasodilation; Vasopressin.

Fig. 1

Estimated 30-day potential drug-induced mortality risk under a treatment strategy that allows to start vasopressin in the first 6 h versus an intervention that never allows for vasopressin start

Fig. 2

Forest plot of risk difference of 30-day ICU mortality for each of the interventions tested against an intervention that never allows for vasopressin start

Fig. 3

Forest plot of risk difference of 30-day ICU mortality for each of the interventions that allows to start vasopressin in the first 6 h versus an intervention that never allows for vasopressin start according to subgroups

Fig. 4

Forest plot of risk difference of 30-day ICU mortality according to norepinephrine equivalent dose threshold for vasopressin introduction versus an intervention that never allows for vasopressin start

Fig. 5

Evolution over the study period of a vasopressin administration; b noradrenaline equivalent dose at vasopressin start time; c Total SOFA score at vasopressin start time; d time from vasopressor start to first vasopressin administration