Association of ventilation volumes, pressures and rates with the mechanical power of ventilation in patients without acute respiratory distress syndrome: exploring the impact of rate reduction

Abstract

Introduction: High mechanical power is associated with mortality in patients who are critically ill and require invasive ventilation. It remains uncertain which components of mechanical power - volume, pressure or rate - increase mechanical power the most.

Methods: We conducted a post hoc analysis of a database containing individual patient data from three randomised clinical trials of ventilation in patients without acute respiratory distress syndrome. The primary endpoint was mechanical power. We used linear regression; double stratification to create subgroups of participants; and mediation analysis to assess the impact of changes in volumes, pressures and rates on mechanical power.

Results: A total of 1732 patients were included and analysed. The median (IQR [range]) mechanical power was 12.3 (9.3-17.1 [3.7-50.1]) J.min^-1. In linear regression, respiratory rate (36%) and peak pressure (51%) explained most of the increase in mechanical power. Increasing quintiles of peak pressure stratified on constant levels of respiratory rate resulted in higher risks of high mechanical power (relative risk 2.2 (95%CI 1.8-2.6), p < 0.01), while decreasing quintiles of respiratory rate stratified on constant levels of peak pressure resulted in lower risks of high mechanical power (relative risk 0.2 (95%CI 0.2-0.3), p < 0.01). Mediation analysis showed that a reduction in respiratory rate, with the increase in tidal volume, partially mediates an effect of reduction in mechanical power (average causal mediation effect -0.10, 95%CI -0.12 to -0.09, p < 0.01), but still with a direct effect of tidal volume on mechanical power (average direct effect 0.15, 95%CI 0.11-0.19, p < 0.01).

Discussion: In this cohort of patients without acute respiratory distress syndrome, pressure and respiratory rate were the most important determinants of mechanical power. The respiratory rate may be the most attractive ventilator setting to adjust when targeting a lower mechanical power.

Keywords: driving pressure; lung‐protective ventilation; mechanical power; mechanical ventilation; respiratory rate.

Figure 1

Study flowchart of patients included in this analysis. ARDS, acute respiratory distress syndrome.

Figure 2

Cumulative distribution plots of mechanical power in four groups. (a) Dark blue, low tidal volume (V_T) and low respiratory rate; green, high V_T and low respiratory rate; red, low V_T and high respiratory rate; light blue, high V_T and high respiratory rate. (b) Dark blue, low peak pressure (Ppeak) and low respiratory rate; red, low Ppeak and high respiratory rate; green, high Ppeak and low respiratory rate; light blue, high Ppeak and high respiratory rate. Vertical dotted lines represent a broadly accepted safety cut‐off for mechanical power of 17 J.min^‐1 and horizontal dotted lines show the median proportion of patients reaching this cut‐off.

Figure 3

Top panels: – stacked bar plots of stratified and resampled peak pressure (Ppeak) (dark grey) and respiratory rate (light grey); middle panels: – error bars of relative risk of mechanical power ≥ 17 J.min^‐1; and bottom panels: – error bars of relative risk of 28‐day mortality; each resampling produced subsamples (S1 to S5) with similar mean values for one ventilator variable but different values for the other variable; top panels show double stratification for (a) increasing Ppeak and matched respiratory rate; (b) increasing Ppeak and decreasing respiratory rate; and (c) matched Ppeak and decreasing respiratory rate. Lines at the top of the bars represent the standard deviation of the corresponding value of the subsamples. Middle panels, corresponding relative risks of high mechanical power ≥ 17 J.min^‐1 are shown, calculated for each subsample; error bars represent the 95%CI for (d) increasing Ppeak and matched respiratory rate; (e) increasing Ppeak and decreasing respiratory rate; and (f) matched Ppeak and decreasing respiratory rate. Bottom panels: corresponding relative risks of 28‐day mortality are shown, calculated for each subsample; error bars represent the 95%CI for (g) increasing Ppeak and matched respiratory rate; (h) increasing Ppeak and decreasing respiratory rate; and (i) matched Ppeak and decreasing respiratory rate.