ObjectiveTo evaluate the reliability of stroke volume variation (SVV) for predicting responsiveness to fluid therapy in patients undergoing cardiac and thoracic surgery.DesignSystematic review and meta-analysis.Data sourcesPubMed, EMBASE, Cochrane Library, Web of Science up to 9 August 2020.MethodsQuality of included studies were assessed with the Quality Assessment of Diagnostic Accuracy Studies-2 tool. We conducted subgroup analysis according to different anaesthesia and surgical methods with Stata V.14.0, Review Manager V.5.3 and R V.3.6.3. We used random-effects model to pool sensitivity, specificity and diagnostic odds ratio with 95% CI. The area under the curve (AUC) of receiver operating characteristic was calculated.ResultsAmong the 20 relevant studies, 7 were conducted during thoracic surgery, 8 were conducted during cardiac surgery and the remaining 5 were conducted in intensive critical unit (ICU) after cardiac surgery. Data from 854 patients accepting mechanical ventilation were included in our systematic review. The pooled sensitivity and specificity were 0.73 (95% CI: 0.59 to 0.83) and 0.62 (95% CI: 0.46 to 0.76) in the thoracic surgery group, 0.71 (95% CI: 0.65 to 0.77) and 0.76 (95% CI: 0.69 to 0.82) in the cardiac surgery group, 0.85 (95% CI: 0.60 to 0.96) and 0.85 (95% CI: 0.74 to 0.92) in cardiac ICU group. The AUC was 0.73 (95% CI: 0.69 to 0.77), 0.80 (95% CI: 0.77 to 0.83) and 0.88 (95% CI: 0.86 to 0.92), respectively. Results of subgroup of FloTrac/Vigileo system (AUC=0.80, Youden index=0.38) and large tidal volume (AUC=0.81, Youden index=0.48) in thoracic surgery, colloid (AUC=0.85, Youden index=0.55) and postoperation (AUC=0.85, Youden index=0.63) in cardiac surgery, passive leg raising (AUC=0.90, Youden index=0.72) in cardiac ICU were reliable.ConclusionSVV had good predictive performance in cardiac surgery or ICU after cardiac surgery and had moderate predictive performance in thoracic surgery. Nevertheless, technical and clinical variables may affect the predictive value potentially.