Non-invasive respiratory variations in arterial pulse pressure using infrared-plethysmography (PPVCNAP) are able to predict fluid responsiveness in mechanically ventilated patients. However, they cannot be continuously monitored. The present study evaluated a new algorithm allowing continuous measurements of PPVCNAP (PPVCNAPauto) (CNSystem, Graz, Austria). Thirty-five patients undergoing vascular surgery were studied after induction of general anaesthesia. Stroke volume was measured using the VigileoTM/FloTracTM. Invasive pulse pressure variations were manually calculated using an arterial line (PPVART) and PPVCNAPauto was continuously displayed. PPVART and PPVCNAPauto were simultaneously recorded before and after volume expansion (500 ml hydroxyethylstarch). Subjects were defined as responders if stroke volume increased by ≥15 %. Twenty-one patients were responders. Before volume expansion, PPVART and PPVCNAPauto exhibited a bias of 0.1 % and limits of agreement from -7.9 % to 7.9 %. After volume expansion, PPVART and PPVCNAPauto exhibited a bias of -0.4 % and limits of agreement from -5.3 % to 4.5 %. A 14 % baseline PPVART threshold discriminated responders with a sensitivity of 86 % (95 % CI 64-97 %) and a specificity of 100 % (95 % CI 77-100 %). Area under the receiver operating characteristic (ROC) curve for PPVART was 0.93 (95 % CI 0.79-0.99). A 15 % baseline PPVCNAPauto threshold discriminated responders with a sensitivity of 76% (95 % CI 53-92 %) and a specificity of 93 % (95 % CI 66-99 %). Area under the ROC curves for PPVCNAPauto was 0.91 (95 % CI 0.76-0.98), which was not different from that for PPVART. When compared with PPVART, PPVCNAPauto performs satisfactorily in assessing fluid responsiveness in hemodynamically stable surgical patients.< Réduire