Variables influencing the prediction of fluid responsiveness: a systematic review and meta-analysis

INTRODUCTION: Prediction of fluid responsiveness in acutely ill patients might be influenced by a number of clinical and technical factors. We aim to identify variables potentially modifying the operative performance of fluid responsiveness predictors commonly used in clinical practice. METHODS: A sensitive strategy was conducted in the Medline and Embase databases to search for prospective studies assessing the operative performance of pulse pressure variation, stroke volume variation, passive leg raising (PLR), end-expiratory occlusion test (EEOT), mini-fluid challenge, and tidal volume challenge to predict fluid responsiveness in critically ill and acutely ill surgical patients published between January 1999 and February 2023. Adjusted diagnostic odds ratios (DORs) were calculated by subgroup analyses (inverse variance method) and meta-regression (test of moderators). Variables potentially modifying the operative performance of such predictor tests were classified as technical and clinical. RESULTS: A total of 149 studies were included in the analysis. The volume used during fluid loading, the method used to assess variations in macrovascular flow (cardiac output, stroke volume, aortic blood flow, volume‒time integral, etc.) in response to PLR/EEOT, and the apneic time selected during the EEOT were identified as technical variables modifying the operative performance of such fluid responsiveness predictor tests (p < 0.05 for all adjusted vs. unadjusted DORs). In addition, the operative performance of fluid responsiveness predictors was also influenced by clinical variables such as the positive end-expiratory pressure (in the case of EEOT) and the dose of norepinephrine used during the fluid responsiveness assessment for PLR and EEOT (for all adjusted vs. unadjusted DORs). CONCLUSION: Prediction of fluid responsiveness in critically and acutely ill patients is strongly influenced by a number of technical and clinical aspects. Such factors should be considered for individual intervention decisions. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s13054-023-04629-w.