Abstract Rationale Assessing fluid responsiveness (FR) in acute respiratory distress syndrome (ARDS) remains challenging because of altered cardiopulmonary interactions, right-ventricular dysfunction, and protective ventilation strategies that reduce the reliability of conventional hemodynamic indices. We conducted a systematic review and meta-analysis exclusively in ARDS to evaluate the diagnostic performance of dynamic FR predictors and identify factors modifying their accuracy. Methods A PRISMA-DTA-compliant meta-analysis (PROSPERO CRD42024574837) was conducted through systematic searches in PubMed, EMBASE, Web of Science, and Google Scholar from January 1999 to October 2024. Eligible studies prospectively evaluated at least one FR predictor in adult ARDS patients (AECC 1994, Berlin 2012, or Global 2024 definitions) using a predefined reference standard such as a fluid challenge, passive leg raising (PLR), or end-expiratory occlusion test (EEOT). Two reviewers independently performed screening, data extraction, and quality assessment (QUADAS-2); certainty of evidence was graded with GRADE. Random-effects bivariate models pooled sensitivity, specificity, and diagnostic odds ratio (DOR), generating hierarchical summary ROC curves. Prespecified sensitivity, subgroup (preload assessment, posture, ECMO status), and meta-regression analyses (tidal volume, PEEP, compliance) explored heterogeneity. Results Ten studies (n = 489 patients; 599 fluid challenges) provided 21 operative data points. Four predictors were pooled: pulse pressure variation (PPV), stroke volume variation (SVV), EEOT, and tidal volume challenge (VTC). PPV showed pooled sensitivity 0.72 (95% CI 0.66-0.78), specificity 0.77 (0.63-0.87), DOR 7.76 (4.56-13.12), AUC 0.75 (I²=21%), indicating moderate accuracy with a threshold of ≥ 7.5% (Figure A). SVV yielded sensitivity 0.78 (0.38-0.95), specificity 0.70 (0.29-0.93), DOR 8.21 (2.89-23.28), AUC 0.81 (I²=0%), suggesting good discrimination near ≥11.2% (Figure B). EEOT performed best (sensitivity 0.84 0.31-0.98, specificity 0.94 0.87-0.98, DOR 71.14 21.81-232.04, AUC 0.95 I²=0%) when cardiac index increased ≥6% after EEOT (Figure C). VTC showed high sensitivity (0.93 0.71-0.99) and moderate specificity (0.69 0.32-0.92) with DOR 18.04 (0.26-1237.83) and AUC 0.89 but substantial heterogeneity (I²=90%) (Figure D). Subgroup analyses found no differences by preload assessment, position, or ECMO. Meta-regression identified tidal volume as a significant determinant of PPV accuracy (p = 0.02). Conclusion Among ventilated ARDS patients, PPV, SVV, and especially EEOT show consistent diagnostic value for predicting FR, whereas VTC displays variable performance. Practical thresholds (PPV ≥7.5%, SVV ≥11.2%, EEOT ≥6% CI increase, VTC ≥6.2% PPV rise) may guide bedside fluid management but require contextual interpretation by ventilatory settings and patient phenotype. This abstract is funded by: None
Alvarado et al. (Fri,) studied this question.