A17-08 Driving Pressure, Respiratory Rate, and Mortality in Acute Respiratory Failure

Abstract Rationale Low tidal volume (VT) mechanical ventilation is a core lung-protective strategy in acute respiratory failure, but the mechanism by which VT reduction influences driving pressure (ΔP), respiratory rate (RR), and associated mortality is complex and requires mechanistic clarification. Since ΔP and RR are physiologically coupled and influenced by changes in VT, their effects on the VT-mortality relationship might heavily depend on the hypothesized causal framework. Methods We analyzed data from the Toronto Intensive Care Observational Registry (2014-2022), including adult patients who received mechanical ventilation for ≥ 4 hours. The exposure was a population-level average reduction of baseline VT from 10 to 5 mL per kilogram of predicted body weight (PBW). The outcome was 30-day intensive care unit mortality. Three causal mediation analyses were performed to evaluate dynamic ΔP (peak inspiratory pressure minus positive end-expiratory pressure) and RR, including scenarios where they acted as joint mediators, and where one was a primary mediator and the other a mediator-outcome confounder influenced by the exposure. Four-way effect decomposition with G-computation quantified mediation and interaction effects, adjusted for key confounders. Results Among 22,337 patients (median age 62 years IQR, 50-73; 37% female), median baseline ventilation parameters were VT 6.9 IQR, 6.1-8.1 mL/kg PBW, dynamic ΔP 12 IQR, 8-16 cmH2O, and RR 20 IQR, 17-25 breaths per minute. The 30-day mortality rate was 17.7% (95% CI, 17.3-18.2). Simulated VT reduction from 10 to 5 mL/kg PBW reduced dynamic ΔP from 14 IQR, 13-14 to 12 IQR, 11-12 cmH2O, increased RR from 20 IQR, 19-20 to 23 IQR, 23-23 breaths per minute. Lower VT was consistently associated with reduced mortality (total-effect odds ratio, 0.82-0.86; p 0.01). Approximately 95% of this benefit was mediated through ΔP reduction (proportion mediated, 0.95; p 0.0001). The concomitant rise in RR was associated with increased mortality (natural indirect-effect odds ratio, 1.08; p 0.0001), attenuating 28% of the overall mortality benefit from VT reduction. Findings were consistent in analyses using static ΔP (plateau pressure minus PEEP) instead of dynamic ΔP. Conclusions Lowering VT reduces mortality primarily by decreasing ΔP, but compensatory increases in RR attenuates this benefit by up to 28%. Our results highlight that mechanistic conclusions are highly dependent on the hypothesized causal framework. Ventilatory strategies that jointly minimize both ΔP and RR are most likely to optimize survival during critical illness. This abstract is funded by: Canadian Institutes of Health Research