Recorded body height and sex differences in lung-protective ventilation among adults with ARDS: an 11-year single-centre cohort study

Background Lung-protective ventilation (LPV) reduces mortality in ARDS, but adherence remains incomplete and women and shorter patients disproportionately receive tidal volumes above predicted-body-weight (PBW)-based targets. The relative associations of recorded sex and body height with this disparity remain uncertain. Methods Retrospective single-centre cohort study (January 2013 to December 2023, RWTH Aachen University Hospital surgical ICU). Adults (≥ 18 years) with Berlin-definition ARDS receiving mechanical ventilation for ≥ 72 h were included. Patient-level adherence to lung-protective targets was described, and repeated-measure PBW-normalised tidal volume and mechanical power were analysed with linear mixed-effects models containing a patient-level random intercept and fixed effects for recorded sex, centred recorded height, age, BMI, SAPS II, SOFA, ARDS aetiology, linear relative ICU day and the lowest same-day PaO₂/FiO₂ ratio. Results Among 798 adults (263 women [33.0%]), unadjusted analyses showed that women received higher tidal volumes per kg PBW than men (median, 7.6 vs. 6.8 mL/kg PBW; p < 0.001), with lower VT adherence (55.3% vs. 76.1%; p < 0.001), higher plateau and driving pressures, and higher PBW-normalised mechanical power. In 527 patients contributing 58,802 observations, the simultaneous adjusted VT model yielded a female-sex coefficient of 0.051 mL/kg PBW (95% CI, −0.273 to 0.376; p = 0.756), whereas recorded height was inversely associated with VT/kg PBW (β per cm, −0.0342; 95% CI, −0.0498 to −0.0185; p < 0.001); the sex-by-height interaction was not supported (p = 0.971). In the corresponding mechanical-power model (410 patients; 44,725 observations), the female-sex coefficient was −0.0158 J·min⁻ 1 ·kg PBW⁻ 1 (95% CI, −0.0512 to 0.0196; p = 0.384). Conclusions Unadjusted analyses showed lower lung-protective ventilation adherence in women. In the simultaneous adjusted model, recorded height, but not recorded sex, was independently associated with VT/kg PBW. Because PBW incorporates both sex and height and recorded height was not validated against a reference measurement, the study does not establish mediation or measurement error as the underlying mechanism. Standardised direct height measurement and PBW-based decision support warrant prospective evaluation.