Respiratory Variations In Pulse Pressure Research Articles

Using generalized additive models to decompose time series and waveforms, and dissect heart–lung interaction physiology

Common physiological time series and waveforms are composed of repeating cardiac and respiratory cycles. Often, the cardiac effect is the primary interest, but for, e.g., fluid responsiveness prediction, the respiratory effect on arterial blood pressure also convey important information. In either case, it is relevant to disentangle the two effects. Generalized additive models (GAMs) allow estimating the effect of predictors as nonlinear, smooth functions. These smooth functions can represent the cardiac and respiratory cycles’ effects on a physiological signal. We demonstrate how GAMs allow a decomposition of physiological signals from mechanically ventilated subjects into separate effects of the cardiac and respiratory cycles. Two examples are presented. The first is a model of the respiratory variation in pulse pressure. The second demonstrates how a central venous pressure waveform can be decomposed into a cardiac effect, a respiratory effect and the interaction between the two cycles. Generalized additive models provide an intuitive and flexible approach to modelling the repeating, smooth, patterns common in medical monitoring data.

Evaluation of pulse pressure variation validity criteria in critically ill patients: a prospective observational multicentre point-prevalence study

Respiratory variation in pulse pressure (ΔPP) is commonly used to predict the fluid responsiveness of critically ill patients. However, some researchers have demonstrated that this measurement has several limitations. The present study was designed to evaluate the proportion of patients satisfying criteria for valid application of ΔPP at a given time-point. A 1 day, prospective, observational, point-prevalence study was performed in 26 French intensive care units (ICUs). All patients hospitalized in the ICUs on the day of the study were included. The ΔPP validity criteria were recorded prospectively and defined as follows: (i) mechanical ventilation in the absence of spontaneous respiration; (ii) regular cardiac rhythm; (iii) tidal volume ≥8 ml kg(-1) of ideal body weight; (iv) a heart rate/respiratory rate ratio >3.6; (v) total respiratory system compliance ≥30 ml cm H2O(-1); and (vi) tricuspid annular peak systolic velocity ≥0.15 m s(-1). The study included 311 patients with a Simplified Acute Physiology Score II of 41 (39-43). Overall, only six (2%) patients satisfied all validity criteria. Of the 170 patients with an arterial line in place, only five (3%) satisfied the validity criteria. During the 24 h preceding the study time-point, fluid responsiveness was assessed for 79 patients. ΔPP had been used to assess fluid responsiveness in 15 of these cases (19%). A very low percentage of patients satisfied all criteria for valid use of ΔPP in the evaluation of fluid responsiveness. Physicians must consider limitations to the validity of ΔPP before using this variable.

Respiratory Variations in Pulse Pressure Reflect Central Hypovolemia during Noninvasive Positive Pressure Ventilation

Background. Correct volume management is essential in patients with respiratory failure. We investigated the ability of respiratory variations in noninvasive pulse pressure (ΔPP), photoplethysmographic waveform amplitude (ΔPOP), and pleth variability index (PVI) to reflect hypovolemia during noninvasive positive pressure ventilation by inducing hypovolemia with progressive lower body negative pressure (LBNP). Methods. Fourteen volunteers underwent LBNP of 0, −20, −40, −60, and −80 mmHg for 4.5 min at each level or until presyncope. The procedure was repeated with noninvasive positive pressure ventilation. We measured stroke volume (suprasternal Doppler), ΔPP (Finapres), ΔPOP, and PVI and assessed their association with LBNP-level using linear mixed model regression analyses. Results. Stroke volume decreased with each pressure level (−11.2 mL, 95% CI −11.8, −9.6, P < 0.001), with an additional effect of noninvasive positive pressure ventilation (−3.0 mL, 95% CI −8.5, −1.3, P = 0.009). ΔPP increased for each LBNP-level (1.2%, 95% CI 0.5, 1.8, P < 0.001) and almost doubled during noninvasive positive pressure ventilation (additional increase 1.0%, 95% CI 0.1, 1.9, P = 0.003). Neither ΔPOP nor PVI was significantly associated with LBNP-level. Conclusions. During noninvasive positive pressure ventilation, preload changes were reflected by ΔPP but not by ΔPOP or PVI. This implies that ΔPP may be used to assess volume status during noninvasive positive pressure ventilation.

Assessment of changes in cardiac index and fluid responsiveness: a comparison of Nexfin and transpulmonary thermodilution

The Nexfin device uses non-invasive photoplethysmography to monitor cardiac output and respiratory variations in pulse pressure and stroke volume. The aim of this study was to compare rapid changes in cardiac index after fluid challenge between Nexfin and bolus transpulmonary thermodilution and the ability to predict fluid responsiveness of dynamic indices given by Nexfin. Simultaneous comparative cardiac index were collected from transpulmonary thermodilution and Nexfin before and after fluid challenge in 45 patients following conventional cardiac surgery. Correlations, Bland-Altman analyses and percentage errors were calculated. Pulse pressure variations and stroke volume variations before fluid challenge were collected to assess their discrimination in predicting fluid responsiveness. Eight (18%) patients were excluded. A weak positive relationship was found between rapid changes in cardiac index after fluid challenge given by both technologies (n = 37, r = 0.39, P = 0.019). Bias, precision and limits of agreements were 0.20 l/min/m(2) (95% confidence interval (CI) 0.02-0.40), 0.57 l/min/m(2) and ± 1.12 l/min/m(2) before fluid challenge, and 0.01 l/min/m(2) (95% CI -0.24 to 0.26), 0.74 l/min/m(2) and ± 1.45 l/min/m(2) after fluid challenge. Percentage errors between Nexfin and transpulmonary thermodilution were 55% and 58% before and after fluid challenge, respectively. Pulse pressure variations and stroke volume variations given by Nexfin were not discriminant to predict fluid responsiveness: areas under receiver operating characteristics curves 0.57 (95% CI 0.40-0.73) and 0.50 (0.33-0.67), respectively. The Nexfin cannot be used to measure rapid changes in cardiac index following fluid challenge and to predict fluid responsiveness after cardiac surgery.

New analyses of standard physiological signals: the old pipes give the sweetest smoke

In this issue of the Journal of Clinical Monitoring and Computing, Giraud et al. investigated the impact of hypovolemia on the respiratory variations in the QRS amplitude on a classical DII EKG lead (DRDII) [1]. In this animal study, they were able to demonstrate that bleeding and volume expansion induced significant changes in DRDII and that these changes mirrored those observed in the respiratory variations in pulse pressure (PPV). These observations suggest that this non-invasive and potentially widely available parameter could predict fluid responsiveness and detect hypovolemia in operating rooms and intensive care units. Studies previously published in the Journal of Clinical Monitoring and Computing have already demonstrated a significant relationship between DRDII and PPV (and between DRDII and stroke volume variation) in mechanically ventilated patients under general anesthesia [2, 3]. The manuscript by Giraud et al. builds on these previous studies by showing that DRDII is also sensitive to manually induced changes in ventricular preload. The remaining steps before acceptance of this new index in clinical practice will be to: (1) show that it can independently predict fluid responsiveness; and (2) that its calculation can be automated and monitored continuously. Thus, while the study by Giraud et al. does not yet confirm this new index as an accurate predictor of fluid responsiveness, it is nevertheless a significant step towards validation. This study exemplifies a new trend in the analysis of physiological signals in which an old and readily available stream of data is processed in a novel way to yield new information. In the present case, the electrocardiogram (EKG) was first described over 100 years ago and has been in clinical use for patient monitoring for decades [4]. Despite this long history, the possible use of the EKG waveform in predicting fluid responsiveness is only now becoming clear. Interestingly, the approach described in the paper by Giraud et al. is very similar to what happened in the analysis of the plethysmographic waveform: both signals have been clinically available for decades but the analysis of their respiratory variations (DRDII for the EKG and DPOP for the plethysmographic waveform [5]) is a new concept which may have significant clinical utility [6]. The beauty of these new indices (if proven accurate) is that the equipment needed to obtain them is already available at the bedside and that they require very little computing in order to be obtained continuously and automatically. Additionally, they are cheap, and could be used almost anywhere. These old signals may thus provide universal, non-invasive, and inexpensive ways of monitoring fluid responsiveness in a variety of environments (hence the Irish proverb that ‘‘the old pipe gives the sweetest smoke’’). A similar trend has been observed in cardiac output (CO) monitoring [7]. The traditional clinical approach to monitoring CO is, of course, transpulmonary thermodilution, and this is still the preferred approach in many settings. The highly invasive nature of pulmonary artery catheters and the associated risk, however, makes them undesirable for routine CO monitoring across the full spectrum of surgical and intensive care patients. Newer J. Rinehart M. Cannesson (&) Department of Anesthesiology & Perioperative Care, University of California Irvine, 101 S City Drive, Orange, CA 92868, USA e-mail: mcanness@uci.edu

Dynamic variables of fluid responsiveness during pneumoperitoneum and laparoscopic surgery

Few data exist on dynamic variables predicting fluid responsiveness during laparoscopic surgery. The aim of this study was to explore the effects of laparoscopy on four dynamic variables: respiratory variations in pulse pressure (ΔPP), stroke volume variation by Vigileo/FloTrac (SVV (Vigileo) ), pleth variability index (PVI) and respiratory variations in pulse oximetry plethysmography waveform amplitude (ΔPOP), and their relation to fluid challenges during laparoscopic surgery. ΔPP, SVV (Vigileo) , PVI and ΔPOP were studied in 20 adult patients before and during pneumoperitoneum (10-12 mmHg). During ongoing laparoscopic surgery, relations between the dynamic variables and changes in stroke volume oesophageal Doppler, (SV(OD) ) after fluid challenges (250 ml colloid) were evaluated. Pneumoperitoneum changed the dynamic variables as follows {mean [95% confidence interval (CI)]}: ΔPP 0.5 (-1.3, 2.3)%, P = 0.53; SVV (Vigileo) 0.6 (-1.3, 2.5)%, P = 0.52; PVI 2.9 (0.4, 5.3)%, P = 0.025. For ΔPOP, median difference (95% CI) was 2.5 (-0.15, 6.7)%, P = 0.058. During laparoscopic surgery, areas under receiver operating characteristics curves (95% CI) were ΔPP 0.53 (0.31-0.75), SVV (Vigileo) 0.74 (0.51-0.90), PVI 0.61 (0.38-0.81), ΔPOP 0.63 (0.40-0.82). Correlation coefficients (P-values) between changes in dynamic variables and changes in SV(OD) were ΔPP r = -0.65, P = 0.009; SVV (Vigileo) r = -0.73, P = 0.002; PVI r = -0.22, P = 0.44; ΔPOP r = -0.32, P = 0.24. ΔPP and SVV (Vigileo) did not change as pneumoperitoneum was established, whereas PVI increased and ΔPOP tended to increase. All four dynamic variables predicted fluid responsiveness relatively poor during ongoing laparoscopic surgery. ΔPP and SVV (Vigileo) tracked changes in stroke volume induced by fluid challenges during ongoing laparascopic surgery, whereas ΔPOP and PVI did not.

Photoplethysmographic and pulse pressure variations during abdominal surgery

Respiratory variations in pulse pressure (ΔPP) predict fluid responsiveness during mechanical ventilation. Variations in pulse oximetry plethysmography amplitude (ΔPOP) are proposed as a non-invasive alternative. Large variations in ΔPOP and poor agreement between ΔPP and ΔPOP are found in intensive care unit patients. General anaesthesia is suggested to reduce variability of ΔPOP and improve agreement between the variables. We evaluated the variability of the agreement between and the diagnostic values of ΔPP and ΔPOP during ongoing open abdominal surgery. The variability of diagnostic methods in specific clinical conditions is important, as this reflects the stability over time during which clinical decisions are made. Observational study during open abdominal surgery in general anaesthesia. ΔPP and ΔPOP were calculated semi-automatically from recording periods of approximately 5 min both before and after fluid challenges. Fluid responsiveness was evaluated by changes in stroke volume (oesophageal Doppler) after 250 ml colloid. Thirty-four fluid challenges were performed in 25 patients. Variance both within registration periods and between patients were significantly larger for ΔPOP than for ΔPP (54.1% vs. 22.1% and 69.6% vs. 22.6%, respectively, both P < 0.001). Limits of agreement with a regression-based correction were ± 13.9%. Areas under receiver operating characteristics curves for fluid responsiveness were 0.67 for ΔPP and 0.72 for ΔPOP. Analysis of raw signals during open abdominal surgery documents that the variance of ΔPOP is larger than of ΔPP, with wide limits of agreement between ΔPP and ΔPOP. The diagnostic values of ΔPP and ΔPOP are relatively poor.

A strong relationship between respiratory variations in pulse pressure (PPV) and airway pressure in fluid nonresponders: a potential explanation for false positive PPV values

Respiratory variations in pulse pressure (PPV) during mechanical ventilation predict fluid responsiveness when the tidal volume is >8 ml/kg [1]. The effect of airway pressure on the ability of PPV to predict fluid responsiveness is less explored. In patients undergoing major abdominal surgery, we found low specificity of PPV and therefore explored the relation between peak airway pressure (Paw) and PPV in fluid challenge nonresponders.

Impact of norepinephrine on the relationship between pleth variability index and pulse pressure variations in ICU adult patients

IntroductionPleth Variability Index (PVI) is an automated and continuous calculation of respiratory variations in the perfusion index. PVI correlates well with respiratory variations in pulse pressure (ΔPP) and is able to predict fluid responsiveness in the operating room. ICU patients may receive vasopressive drugs, which modify vascular tone and could affect PVI assessment. We hypothesized that the correlation between PVI and ΔPP and the ability of PVI to identify patients with ΔPP > 13% is dependent on norepinephrine (NE) use.Methods67 consecutive mechanically ventilated patients in the ICU were prospectively included. Three were excluded. The administration and dosage of NE, heart rate, mean arterial pressure, PVI and ΔPP were measured simultaneously.ResultsIn all patients, the correlation between PVI and ΔPP was weak (r2 = 0.21; p = 0.001). 23 patients exhibited a ΔPP > 13%. A PVI > 11% was able to identify patients with a ΔPP > 13% with a sensitivity of 70% (95% confidence interval: 47%-87%) and a specificity of 71% (95% confidence interval: 54%-84%). The area under the curve was 0.80 ± 0.06. 35 patients (53%) received norepinephrine (NE(+)). In NE(+) patients, PVI and ΔPP were not correlated (r2 = 0.04, p > 0.05) and a PVI > 10% was able to identify patients with a ΔPP > 13% with a sensitivity of 58% (95% confidence interval: 28%-85%) and a specificity of 61% (95% confidence interval:39%-80%). The area under the ROC (receiver operating characteristics) curve was 0.69 ± 0.01. In contrast, NE(-) patients exhibited a correlation between PVI and ΔPP (r2 = 0.52; p < 0.001) and a PVI > 10% was able to identify patients with a ΔPP > 13% with a sensitivity of 100% (95% confidence interval: 71%-100%) and a specificity of 72% (95% confidence interval: 49%-90%). The area under the ROC curve was 0.93 ± 0.06 for NE(-) patients and was significantly higher than the area under the ROC curve for NE(+) patients (p = 0.02).ConclusionsOur results suggest that in mechanically ventilated adult patients, NE alters the correlation between PVI and ΔPP and the ability of PVI to predict ΔPP > 13% in ICU patients.

Respiratory pulse pressure variation fails to predict fluid responsiveness in acute respiratory distress syndrome

IntroductionFluid responsiveness prediction is of utmost interest during acute respiratory distress syndrome (ARDS), but the performance of respiratory pulse pressure variation (ΔRESPPP) has scarcely been reported. In patients with ARDS, the pathophysiology of ΔRESPPP may differ from that of healthy lungs because of low tidal volume (Vt), high respiratory rate, decreased lung and sometimes chest wall compliance, which increase alveolar and/or pleural pressure. We aimed to assess ΔRESPPP in a large ARDS population.MethodsOur study population of nonarrhythmic ARDS patients without inspiratory effort were considered responders if their cardiac output increased by >10% after 500-ml volume expansion.ResultsAmong the 65 included patients (26 responders), the area under the receiver-operating curve (AUC) for ΔRESPPP was 0.75 (95% confidence interval (CI95): 0.62 to 0.85), and a best cutoff of 5% yielded positive and negative likelihood ratios of 4.8 (CI95: 3.6 to 6.2) and 0.32 (CI95: 0.1 to 0.8), respectively. Adjusting ΔRESPPP for Vt, airway driving pressure or respiratory variations in pulmonary artery occlusion pressure (ΔPAOP), a surrogate for pleural pressure variations, in 33 Swan-Ganz catheter carriers did not markedly improve its predictive performance. In patients with ΔPAOP above its median value (4 mmHg), AUC for ΔRESPPP was 1 (CI95: 0.73 to 1) as compared with 0.79 (CI95: 0.52 to 0.94) otherwise (P = 0.07). A 300-ml volume expansion induced a ≥2 mmHg increase of central venous pressure, suggesting a change in cardiac preload, in 40 patients, but none of the 28 of 40 nonresponders responded to an additional 200-ml volume expansion.ConclusionsDuring protective mechanical ventilation for early ARDS, partly because of insufficient changes in pleural pressure, ΔRESPPP performance was poor. Careful fluid challenges may be a safe alternative.

Ability of respiratory pulse pressure variation to predict fluid responsiveness in ARDS: still an unanswered question?

We read with interest the study of Lakhal and colleagues [1] suggesting that respiratory pulse pressure variation (PPV) is not an accurate predictor of fl uid responsiveness in patients with acute respiratory distress syndrome (ARDS) ventilated with low tidal volume (Vt). Th ey stated that the poor performance of PPV is attributed to small changes in respiratory pleural pressure related to low Vt, such that PPV is low even in the case of marked cardiac preload dependency. Th ese fi ndings are not surprising since in this study [1] the mean static compliance was not very low (40 cmH 2 O); therefore, the percentage of transmission of alveolar pressure to the pleural space was not really altered. Th us, the ventilation with low Vt induced small variations in pleural and transpulmonary pressures. It must be stressed, however, that patients with ARDS usually exhibit a severe decrease in lung compliance and high plateau pressure. Moreover, in ARDS patients ventilated with low Vt, application of relatively high levels of positive end expiratory pressure (between 10 and 15 cmH 2 O) is now recommended [2]. Consequently, respira tory changes in transpulmonary pressure should remain greater than normal, and in spite of reduced lung compliance, cyclic changes in intrathoracic pressure may still be high enough for PPV to predict fl uid responsiveness [3]. Th e fi ndings of Lakhal and colleagues [1] perhaps provide insuffi cient support for the assertion that PPV is unable to predict volume responsiveness in ARDS patients ventilated with low Vt. Additional studies in severe ARDS patients (compliance <30 cmH 2 O) are necessary to investigate whether or not PPV could be used in such cases.

Respiratory Variation in Pulse Pressure and Plethysmographic Waveforms

Dynamic variables are the best predictors of fluid responsiveness in patients under general anesthesia and mechanical ventilation; namely, respiratory variations in pulse pressure and in the plethysmographic waveform. However, these variables have potential limitations. Our aim was to evaluate their intraoperative applicability. We extracted clinical data from all anesthesia procedures performed at our institution in 2009 and identified the number of cases that presented predetermined conditions of application. Among the 12,308 procedures, 39% met the criteria for the noninvasive monitoring of variations in the plethysmographic waveform of which 23% had arterial lines and met the criteria for the invasive monitoring of variations in pulse pressure.

Poor Agreement between Respiratory Variations in Pulse Oximetry Photoplethysmographic Waveform Amplitude and Pulse Pressure in Intensive Care Unit Patients

To identify fluid responsiveness, a correlation between respiratory variations in pulse pressure (DeltaPP) and respiratory variations in pulse oximetry photoplethysmographic waveform amplitude (DeltaPOP) in mechanically ventilated patients has been demonstrated. To evaluate the agreement between the two methods, knowledge about the repeatability of the methods is imperative. However, no such data exist. Based on knowledge of slow oscillation in skin blood flow, the authors hypothesized that the variability of DeltaPOP would be larger than that of DeltaPP when calculations were performed continuously over a long recording period. Respiration, continuous invasive blood pressure, pulse oximetry, and skin microcirculation were recorded in 14 mechanically ventilated intensive care unit patients. No intravenous fluid challenges were given, and no other interventions were performed during the measurements. Seventy consecutive comparisons between DeltaPP and DeltaPOP were calculated for each of the 14 patients. For all patients, DeltaPOP was 13.7 +/- 5.8% and DeltaPP was 5.8 +/- 2.6% (P < 0.001). There was a larger intraindividual (8.94 vs. 1.29; P < 0.001) and interindividual (26.01 vs. 5.57; P < 0.001) variance of DeltaPOP than of DeltaPP. In six patients, there was no significant correlation between DeltaPP and DeltaPOP. A Bland-Altman plot showed poor agreement between the two methods. A large variability of DeltaPOP and a poor agreement between DeltaPP and DeltaPOP limits DeltaPOP as a tool for evaluation of fluid responsiveness in intensive care unit patients. This is in contrast to DeltaPP, which shows a small variability.

Noninvasive assessment of arterial pressure

To briefly review recent advances in the noninvasive assessment of arterial pressure (indirect methods) in the field of critical care. Automated oscillometric measurements underestimate intraarterial systolic blood pressure. Digital photoplethysmography has led to conflicting results, although the obtained respiratory pulse pressure variation correlates with the fluid-challenge-induced changes in stroke volume. The pulse oximetry photoplethysmographic signal recorded at the digital or ear level may be useful in monitoring respiratory arterial pressure variations, although technical improvements and clarifications are needed. Arterial tonometry is increasingly used in the cardiovascular field to reconstruct central aortic pressure. A recent study has shown that radial artery tonometry is feasible in hemodynamically stable patients and that peripheral pulse pressure reflects the combined influences of arterial stiffness and stroke volume, especially in elderly patients. The limitations of this technique include the potential bias related to the use of a generalized transfer function and the difficulty in obtaining reliable recordings in hemodynamically unstable patients. Intraarterial blood pressure must be preferred over noninvasive blood pressure recordings when critical decisions are required. In hemodynamically stable patients, valuable information may be obtained by using noninvasive techniques, amongst which arterial tonometry seems especially promising.

Respiratory Variations in Pulse Oximetry Plethysmographic Waveform Amplitude to Predict Fluid Responsiveness in the Operating Room

Respiratory variations in pulse oximetry plethysmographic waveform amplitude (DeltaPOP) are related to respiratory variations in pulse pressure (DeltaPP) and are sensitive to changes in preload. The authors hypothesized that DeltaPOP can predict fluid responsiveness in mechanically ventilated patients during general anesthesia. Twenty-five patients referred for cardiac surgery were studied after induction of general anesthesia. Hemodynamic data (cardiac index, central venous pressure, pulmonary capillary wedge pressure, DeltaPP, and DeltaPOP) were recorded before and after volume expansion (500 ml hetastarch, 6%). Fluid responsiveness was defined as an increase in cardiac index of 15% or greater. Volume expansion induced changes in cardiac index (2.0+/-0.4 to 2.3+/-0.5 mmHg; P<0.05), DeltaPP (11+/-7 to 6+/-5%; P<0.05), and DeltaPOP (12+/-9 to 7+/-5%; P<0.05). DeltaPOP and DeltaPP were higher in responders than in nonresponders (17+/-8 vs. 6+/-4 and 14+/-7 vs. 6+/-4%, respectively; P<0.05 for both). A DeltaPOP greater than 13% before volume expansion allowed discrimination between responders and nonresponders with 80% sensitivity and 90% specificity. There was a significant relation between DeltaPOP before volume expansion and percent change in cardiac index after volume expansion (r=0.62; P<0.05). DeltaPOP can predict fluid responsiveness noninvasively in mechanically ventilated patients during general anesthesia. This index has potential clinical applications.

Passive leg raising predicts fluid responsiveness in the critically ill*

Passive leg raising (PLR) represents a "self-volume challenge" that could predict fluid response and might be useful when the respiratory variation of stroke volume cannot be used for that purpose. We hypothesized that the hemodynamic response to PLR predicts fluid responsiveness in mechanically ventilated patients. Prospective study. Medical intensive care unit of a university hospital. We investigated 71 mechanically ventilated patients considered for volume expansion. Thirty-one patients had spontaneous breathing activity and/or arrhythmias. We assessed hemodynamic status at baseline, after PLR, and after volume expansion (500 mL NaCl 0.9% infusion over 10 mins). We recorded aortic blood flow using esophageal Doppler and arterial pulse pressure. We calculated the respiratory variation of pulse pressure in patients without arrhythmias. In 37 patients (responders), aortic blood flow increased by > or =15% after fluid infusion. A PLR increase of aortic blood flow > or =10% predicted fluid responsiveness with a sensitivity of 97% and a specificity of 94%. A PLR increase of pulse pressure > or =12% predicted volume responsiveness with significantly lower sensitivity (60%) and specificity (85%). In 30 patients without arrhythmias or spontaneous breathing, a respiratory variation in pulse pressure > or =12% was of similar predictive value as was PLR increases in aortic blood flow (sensitivity of 88% and specificity of 93%). In patients with spontaneous breathing activity, the specificity of respiratory variations in pulse pressure was poor (46%). The changes in aortic blood flow induced by PLR predict preload responsiveness in ventilated patients, whereas with arrhythmias and spontaneous breathing activity, respiratory variations of arterial pulse pressure poorly predict preload responsiveness.