Assessing fluid responsiveness in critically ill patients: False-positive pulse pressure variation is detected by Doppler echocardiographic evaluation of the right ventricle* Yazine Mahjoub, MD; Cyrille Pila, MD; Arnaud Friggeri, MD; Elie Zogheib, MD; Eric Lobjoie, MD; Franc ¸ ois Tinturier, MD; Claude Galy, MD; Michel Slama, MD, PhD; Herve ´ Dupont, MD, PhD F luid administration is used as first-line therapy in hypotensive critically ill patients (1) as inad- equate fluid administration can be detrimental. Respiratory variation of stroke volume assessed by pulse pressure variation (PPV) or aortic blood flow varia- tion has been demonstrated to predict ac- curately preload responsiveness in mechan- ically ventilated patients (2– 4). Although some studies have reported a good specific- ity of PPV (2), others show a lower speci- ficity for the prediction of fluid responsive- ness (5, 6). Despite a PPV of 12%, some patients do not respond to volume expan- sion (VE). As suggested by Vieillard-Baron et al (5), we formulated the hypothesis that these cases of false-positive PPV could be due to right ventricular (RV) dysfunction. RV dysfunction leads to a decrease of RV ejection during the inspiratory increase of RV afterload. In this situation, infused fluid cannot reach the pulmonary circulation and therefore does not increase left ventric- ular stroke volume. Despite the limited quantitative evaluation of RV function, Doppler echocardiography provides rele- vant parameters to assess RV dysfunction (7), including tissue Doppler tricuspid an- nular systolic velocity (Sta), which is easy to obtain (8 –11). The aim of this study was to demonstrate that Sta can be used to detect false-positive PPV. MATERIALS AND METHODS Patients Over a 6-mo period, all patients hospitalized for 24 hrs in our 16-bed intensive care unit presenting circulatory failure under mechanical ventilation, sedated, and with an arterial cathe- ter in place showing PPV of 12% (2) were eligible for inclusion in the study. The patient was formally included in the study when the attending physician decided that the patient re- quired VE. Circulatory failure was defined as sys- tolic blood pressure of 90 mm Hg or the need for infusion of vasoactive drugs (norepinephrine 0.10 g/kg/min or dopamine 10 g/kg/min). Patients with arrhythmias, spontaneous breath- ing effort (visualized by continuous monitoring of airway pressure and flow on the respirator screen), and major RV dilation, defined as ratio of right ventricular end-diastolic area to left ven- tricular end-diastolic area (RV-EDA/LV-EDA) 1, constituting a contraindication to fluid in- fusion by the French expert conference were excluded (12). Mechanical ventilation was per- formed in volume-controlled mode with a tidal volume of 5 to 12 mL/kg. External positive end- expiratory pressure was between 0 and 10 cm H 2 O. The plateau pressure was kept at 30 cm H 2 O. The following patient clinical charac- teristics were recorded: age, gender, weight, height, Simplified Acute Physiology Score II, diagnosis and dose of vasopressors, heart rate *See also p. 2662. From the Medical (YM, CP, AF, EL, FT, CG, HD) and Surgical Intensive Care Unit (EZ), Department of Anes- thesia and Intensive Care; and Medical Intensive Care Unit (MS), Department of Nephrology, Amiens Univer- sity Hospital, Amiens, France. The authors have not disclosed any potential con- flicts of interest. For information regarding this article, E-mail: mahjoub.yazine@chu-amiens.fr Copyright © 2009 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e3181a380a3 Objectives: To determine whether peak systolic velocity of tricuspid annular motion assessed by tissue Doppler echocardi- ography (Sta), a right ventricular function parameter, can discrim- inate patients with true- and false-positive pulse pressure variation. Pulse pressure variation is used to predict fluid responsiveness in mechanically ventilated patients. However, this parameter has been reported to be falsely positive, especially in patients with right ventricular dysfunction. Design: A prospective study. Setting: Medical and surgical intensive care unit of a university hospital. Patients: Thirty- five mechanically ventilated patients hospi- talized for >24 hrs with a pulse pressure variation of >12%. Interventions: Doppler echocardiography (including measure- ment of Sta and stroke volume) was performed before and after infusion of 500 mL of colloid solution. Patients were classified into two groups according to their response to fluid infusion: respond- ers (at least 15% increase in stroke volume) and nonresponders. Measurements and Main Results: Twenty-three patients (66%) were responders (true-positive group) and 12 (34%) were nonre- sponders (false-positive group). Before volume expansion, Sta was statistically lower in the nonresponder group (0.13 [0.04] vs. 0.20 [0.05], p  .0004). The area under the curve of the receiver operating characteristic curve was 0.87 (95% confidence interval, 0.74 –1). In patients with pulse pressure variation of >12%, a Sta cutoff value of 0.15 m/s discriminated between responders and nonresponders with a sensitivity of 91% (80 –100) and a speci- ficity of 83% (62–100). Conclusions: A Sta value of <0.15 m/s seems to be an accu- rate parameter to detect false-positive pulse pressure variation. Echocardiography should therefore be performed before fluid infusion in patients with pulse pressure variation of >12%. (Crit Care Med 2009; 37:2570 –2575) KEY WORDS: fluid responsiveness; pulse pressure variation; false-positive; right ventricle; tissue Doppler imaging 2570 Crit Care Med 2009 Vol. 37, No. 9