To evaluate the ability of ΔPP/ΔP [pulse pressure variations (ΔPP) adjusted by alveolar pressure variations (ΔP = Pplat-PEEPtot)] in predicting fluid responsiveness, to compare its accuracy to that of ΔPP used alone and to evaluate the influence of tidal volume (Vt) on these two indices.Design
A 22-bed general intensive care unit (ICU).Patients
Eighty-four surgical or medical ventilated patients requiring fluid challenge.Intervention
A 6 ml/kg colloid fluid challenge in 30 min.Measurements and results
Hemodynamic measurements taken before and after fluid challenge. Patients separated into responders and nonresponders according to a 15% increase in their cardiac output. Thirty-nine patients found to be responders and 45 nonresponders. ΔPP/ΔP and ΔPP were both higher in responders than in nonresponders. ΔPP/ΔP was a better predictor of fluid responsiveness than ΔPP, especially for patients ventilated with Vt ≥ 8 ml/kg [area under the curve (AUC) 0.88 (0.77–0.98) versus 0.75 (0.60–0.89), P < 0.01)]. In this population ΔPP/ΔP higher than 0.9 predicted fluid response with positive predictive value of 87% and negative predictive value of 78%. Overall ΔPP and ΔPP/ΔP reliability was poor for patients ventilated with Vt < 8 ml/kg [AUC 0.63 (0.45–0.81) and 0.72 (0.55–0.88), respectively].Conclusion
In this mixed ICU population ΔPP adjusted by ΔP is a simple index which outperforms ΔPP for patients ventilated with Vt ≥ 8 ml/kg. However, correcting ΔPP by ΔP still fails to predict fluid response reliably in patients ventilated with low tidal volume.