Both Too Little and Too Much Respiratory Drive and Effort Predict Worse Outcomes on Mechanical Ventilation: Insights from a Prospective Toronto Cohort

Highlight

– In 1,186 mechanically ventilated adults from a prospective registry, both low and high respiratory drive (P0.1) and effort (Pocc) were non-linearly associated with higher ICU mortality and lower rates of ICU discharge.
– Associations were strongest when oxygenation was impaired (PaO2:FiO2 ≤150 mmHg); high dynamic transpulmonary driving pressure (ΔPL,dyn) predicted worse outcomes, and the negative effect of ventilator-delivered driving pressure (ΔPaw,dyn) was magnified by greater patient effort.
– Findings support the clinical relevance of monitoring and titrating respiratory drive and effort to reduce potential lung- and diaphragm-injury during assisted spontaneous breathing.

Background

Management of the patient who breathes spontaneously while mechanically ventilated requires balancing two competing hazards. Insufficient respiratory drive and effort can compromise ventilation, delay liberation from the ventilator, and contribute to diaphragm disuse atrophy. Conversely, excessive drive or effort may amplify lung stress and cause patient self-inflicted lung injury (P-SILI), and can injure the diaphragm through excessive load and fatigue. Physiologic and animal data support both mechanisms, but there has been limited high-quality human outcome data linking breath-by-breath measures of drive and effort to clinically meaningful endpoints.

Study design

This prospective, registry-based cohort study by Dianti and colleagues used the Toronto Intensive Care Observational Registry to include all adults receiving mechanical ventilation in the mixed medical–surgical ICU at Toronto General Hospital between June 25, 2019, and April 1, 2022 (no exclusions). For the first 10 days of mechanical ventilation, daily physiologic measurements were recorded: airway occlusion pressure at 100 ms (P0.1, a surrogate of respiratory drive), expiratory occlusion pressure (Pocc, a bedside index of inspiratory muscle effort), dynamic transpulmonary driving pressure during spontaneous breathing (ΔPL,dyn; an index of lung stress produced during patient effort), and ventilator-delivered driving pressure (ΔPaw,dyn). The primary endpoints were daily hazards of ICU death and discharge alive. The investigators used Cox proportional hazards models with time-varying covariates, adjusted for changing severity of illness, to quantify associations between physiologic variables and outcomes. Interaction by oxygenation (PaO2:FiO2 dichotomized at 150 mmHg) was assessed.

Key findings

Population and outcomes: 1,186 patients met inclusion criteria; 298 (25%) died during follow-up. The cohort included patients across a spectrum of disease severity and ventilatory modes, reflecting real-world practice in a tertiary ICU.

Non-linear (U-shaped) associations for drive and effort

Both P0.1 and Pocc demonstrated a non-linear association with the hazards for death and for discharge alive (statistical tests for non-linearity p≤0.024). Practically, this followed a U-shaped pattern: very low drive/effort and very high drive/effort were each associated with worse outcomes (higher mortality and lower rates of ICU discharge) compared with intermediate levels.

Effect modification by oxygenation (PaO2:FiO2)

Oxygenation materially modified the relationship between drive/effort and outcome (interaction p<0.0001). In patients with more severe hypoxemia (PaO2:FiO2 ≤150 mmHg), both low and high P0.1 and Pocc were associated with a reduced rate of ICU discharge — that is, either extreme was detrimental. By contrast, in patients with PaO2:FiO2>150 mmHg, higher drive and effort were associated with faster discharge alive. These findings imply that the balance of risks changes with underlying lung injury: when oxygenation is poor, high effort may worsen lung stress and impede recovery; when oxygenation is better, some increased effort may facilitate ventilatory recovery.

Ventilator and transpulmonary driving pressures

High dynamic transpulmonary driving pressure (ΔPL,dyn) during spontaneous breathing was independently associated with a lower rate of ICU discharge (p<0.0001), and this adverse association was most pronounced in patients with PaO2:FiO2 <150 mmHg. Importantly, the investigators found an interaction between patient effort and ventilator-delivered driving pressure: higher effort magnified the harmful association between ΔPaw,dyn and delayed discharge alive (interaction p=0.0052). In other words, the combination of high ventilator pressures and high patient effort appeared especially deleterious.

Clinical effect sizes and statistical considerations

The study used time-varying Cox models to relate daily physiologic measures to daily hazards of death or discharge, accounting for evolving illness severity. The reported non-linear associations and interactions were statistically robust. As an observational study, causality cannot be established definitively, but the temporal granularity (daily serial measures during the first 10 ventilator days) strengthens the inference that contemporaneous drive/effort relate to near-term outcomes.

Expert commentary and physiologic plausibility

The findings fit a coherent physiologic narrative. Prior mechanistic and animal studies have demonstrated that excessive inspiratory effort can increase transpulmonary pressures and regional lung stress, promoting volutrauma and P-SILI. Conversely, prolonged underuse of the diaphragm accelerates disuse atrophy, which has been linked to prolonged weaning and worse outcomes. The U-shaped relationship supports the concept of an optimal