Patient–Ventilator Asynchrony Burden, Phenotypes, and Clinical Outcomes in Critically Ill Adults: A Systematic Review and Meta-analysis - 17/11/26

Highlights |
• | Overall PVA burden was associated with higher crude ICU mortality. |
• | Nonmortality associations were less robust under small-sample inference. |
• | Adjusted estimates varied by PVA phenotype, timing, and statistical model. |
• | Residual confounding limits PVA as an independent prognostic target. |
• | PVA is better viewed as a heterogeneous signal than a treatment threshold. |
Abstract |
Background |
Patient–ventilator asynchrony (PVA) is common during invasive mechanical ventilation, but its prognostic significance remains uncertain because studies differ in the type of asynchrony measured, exposure definitions, monitoring windows, and statistical adjustment. We examined associations between PVA and clinical outcomes, distinguishing overall asynchrony burden from phenotype-specific exposures and unadjusted from adjusted evidence.
Methods |
We searched PubMed, Embase, Web of Science, and the Cochrane Library through July 15, 2026 for observational prognostic studies of adults receiving invasive mechanical ventilation in intensive care units (ICUs). The asynchrony index (AI) and overall asynchrony index (OAI) were used to describe aggregate burden; the ineffective triggering index (ITI) and other phenotype-specific exposures were examined separately. Comparable unadjusted associations were pooled using random-effects models; modified Hartung–Knapp confidence intervals were used for primary inference when at least three studies were available. Adjusted estimates were synthesized descriptively. Risk of bias and certainty were assessed using QUIPS and GRADE. PROSPERO: CRD420261451479.
Results |
Twenty-one reports were included. Higher overall AI/OAI burden, defined using study-specific thresholds around 10%, was associated with greater unadjusted ICU mortality (7 studies, 498 patients; OR 3.17; modified Hartung–Knapp 95% CI 1.50–6.72; conventional 95% CI 1.74–5.79; I² = 22%; low-certainty evidence). The association persisted after excluding two short manual assessments (5 studies, 401 patients; OR 3.75; modified Hartung–Knapp 95% CI 1.21–11.60). Overall AI/OAI associations with ventilation duration (MD 3.39 days; modified Hartung–Knapp 95% CI −0.90 to 7.68) and ICU length of stay (MD 2.91 days; 95% CI −0.91 to 6.73) were less certain. ITI-related ICU mortality and other phenotype-specific associations were based on limited, heterogeneous evidence. Adjusted findings differed by exposure and model; confounding was the principal risk-of-bias concern.
Conclusions |
Higher overall AI/OAI burden was associated with greater unadjusted ICU mortality, including in sensitivity analyses, whereas associations with ventilation duration and ICU length of stay were less certain. The prognostic implications of PVA vary by phenotype, exposure definition, and monitoring strategy, and the 10% AI/OAI threshold is not established as a treatment target.
Le texte complet de cet article est disponible en PDF.Keywords : asynchrony index, mechanical ventilation, mortality, patient–ventilator asynchrony, prognosis
Abbreviations : AI, aHR, aOR, CI, EAdi, GRADE, HR, ICU, ITI, MD, MV, OAI, OR, PEEP, PRISMA, PVA, QUIPS, RT, RT-BS, sHR, SOFA
Plan
Bienvenue sur EM-consulte, la référence des professionnels de santé.
