Post hoc analyses of the effects of prone positioning (PP) on the occurrence of ventilator-associated pneumonia (VAP)
| Population | Sample size | VAP | ||||
| Supine | Prone | Supine | Prone | HR (95% CI) | p-value | |
| All | 199 | 199 | 8 (4) | 49 (24.6) | 1.64 (0.70–3.84) | 0.25 |
| Single day of PP | 109 | 109 | 7 (6.4) | 12 (11) | 1.11 (0.39–3.10) | 0.85 |
| ≥2 days of PP (time-dependent covariate) | 90 | 90 | 1 (1.1) | 37 (41.1) | 0.72 (0.32–1.65) | 0.44 |
| Minimum Pa,O2/FI,O2 ratio <120 within 2 days before PP | 32 | 72 | 2 (6.3) | 20 (27.8) | 0.69 (0.19–2.52) | 0.57 |
| SAPS II >50 on the first day of PP | 50 | 64 | 1 (2) | 11 (17.2) | 4.33 (0.70–26.65) | 0.11 |
| Only patients with Pa,O2/FI,O2 ratio <200 at admission | 153 | 153 | 7 (4.6) | 40 (26.1) | 1.03 (0.42–2.56) | 0.94 |
Data are presented as n or n (%), unless otherwise stated. Pa,O2/FI,O2: arterial oxygen tension/inspiratory oxygen fraction ratio; SAPS II: Simplified Acute Physiology Score. The marginal Cox model was used to compute the hazard ratio (HR) adjusted on imbalances between groups and risk factors for events.