Airway-derived mechanical power and postoperative pulmonary complications in obese adults undergoing laparoscopic bariatric surgery: a secondary analysis of a randomised trial.

PURPOSE: To determine the associations of participant-level mean airway-derived total and dynamic mechanical power with postoperative pulmonary complications (PPCs) after laparoscopic bariatric surgery in adults with obesity, and to characterise the intraoperative trajectories of power components. METHODS: This was a secondary analysis of a single-centre, parallel-group randomised trial in 160 adults with obesity undergoing laparoscopic bariatric surgery. Mechanical power was calculated at four predefined intraoperative timepoints (T1-T4). Mean dynamic mechanical power (MP_dynamic) was the primary exposure and mean total mechanical power (MP_total) was the key secondary exposure. For both exposures, the primary analysis used multivariable logistic regression across quartiles, with Q1 as the reference. Quartile-based modelling was selected after exploratory assessment indicated departure from linearity and because quartile categories provide clinically interpretable event rates and adjusted odds ratios. This analytic hierarchy was not prespecified in the parent-trial protocol and was defined post hoc after assessment of the observed exposure-outcome form. Restricted cubic spline analysis of mean MP_dynamic was supportive, and continuous per-standard-deviation models for both exposures were sensitivity analyses of a linear association. Models were adjusted for randomised group, age, sex, BMI, surgery duration, smoking status, diagnosed obstructive sleep apnoea (OSA), and asthma. RESULTS: The individualised strategy reduced mean MP_dynamic (6.42 ± 1.21 vs 6.99 ± 1.11 J/min; mean difference -0.57, 95% CI -0.93 to -0.21; p = 0.002) and increased mean MP_total (12.61 ± 2.48 vs 11.79 ± 1.96 J/min; mean difference +0.82, 95% CI +0.12 to +1.51; p = 0.022). PPCs occurred in 36/160 patients (22.5%). MP_dynamic quartiles showed no statistically significant association with PPCs (global p = 0.575). Mean MP_total showed an inverted U-shaped association: adjusted PPC odds were higher in Q2 (OR 4.74, 95% CI 1.32-17.01; p = 0.017) and Q3 (OR 3.91, 95% CI 1.12-13.60; p = 0.032) than in Q1, with no further increase in Q4 (OR 2.39, 95% CI 0.63-9.13; p = 0.201; global p = 0.061). Continuous sensitivity analyses showed no linear association for mean MP_dynamic (OR 1.10 per SD, 95% CI 0.73-1.65) or mean MP_total (OR 1.15 per SD, 95% CI 0.76-1.73). CONCLUSION: Mean MP_dynamic showed no statistically significant association with PPCs. In contrast, mean MP_total showed an inverted U-shaped association with PPC risk, with higher adjusted odds in the second and third quartiles and no further increase in the highest quartile. This pattern may partly reflect the opposing physiological implications of potentially injurious dynamic energy and PEEP-related energy delivered during lung-protective recruitment. MP_total combines components with different and potentially opposing physiological implications and should not be interpreted as a homogeneous measure of injurious energy exposure. Reporting dynamic and PEEP-related components provides clearer physiological interpretation than total MP alone. TRIAL REGISTRATION: ChiCTR2400085568 (Chinese Clinical Trial Registry).