Arterial Pressure Changes during the Valsalva Maneuver to Predict Fluid Responsiveness in Spontaneously Breathing Patients
ELECTRONIC SUPPLEMENTARY MATERIAL
M. Ignacio Monge, Anselmo Gil, J. Carlos Díaz
Servicio de Cuidados Críticos y Urgencias. Unidad de Investigación Experimental.
Hospital del SAS Jerez. Spain.
Materials and methods (supplementary material)
Hemodynamic and respiratory monitoring
Invasive pressure transducers were calibrated against atmosphere at the mid-chest level. All the patients received continuous electrocardiography, pulse oximetry, central venous pressure (CVP) and invasive blood pressure monitoring displayed on the bedside monitor (Datex CS/3, Datex-Ohmeda, Helsinki, Finland). Airway flow and pressure were continuously measured during the VM using a side-stream spirometry module (MCOVX, Datex-Ohmeda, Helsinki, Finland) integrated in the patient monitor. All the signals were simultaneously recorded on-line in a laptop computer using proprietary data acquisition software (S/5 Collect software, version 3.0; Datex-Ohmeda, Helsinki, Finland) for off-line analysis.
Cardiac output measurements
Cardiac output and stroke volume were calculated using the Flotrac/Vigileo® system. After zeroing the system against atmosphere, the arterial waveform signal fidelity was checked and the hemodynamic measurements initiated. This device allows calculating cardiac output and stroke volumefrom the real-time analysis of the arterial waveform over a period of 20 seconds without external calibration, using a proprietary algorithm based on the relationships between the arterial pulse pressure and stroke volume. Arterial compliance and vascular resistance contribution was estimated every minute based on individual patient demographic data (age, gender, body weight and height) and the arterial waveform analysis respectively. The method has been described in depth elsewhere [1].
Statistical analysis
Sample size was calculated according to a preliminary power analysis for the comparison of the area under the ROC curve (AUC) with the null hypothesis (∆VPP can´t predict fluid-responsiveness). We select a type I error of 0.05, a type II error of 0.2 and a desired AUC of 0.9, assuming that fluid responsiveness only occurs in 50% of UCI patients[2].
Arterial pressure responses to the Valsalva maneuver and Valsalva parameters calculation
Normal sinusoidal pattern on arterial pressure response to the Valsalva maneuver (Fig. E1)
Phase 1 starts with the beginning of the strain and is characterized by an abrupt raise in arterial pressure as a result of the transmission of a sudden increase in intrathoracic pressure to large vessels; phase 2 continues to a sharp decrease of both, systolic and diastolic pressure, and narrowing of arterial pulse pressure (early phase 2), caused by impairment of venous return and decreased left ventricular stroke volume, followed 5-10 seconds later by a progressive increase in heart rate and arterial pressure due to a sympathetic reflex and compensatory raise in peripheral vascular resistance (late phase 2); the sudden release of strain and drop in the intrathoracic pressure produce an acute drop in arterial pressure in phase 3; and finally, phase 4 is characterized by an overshoot of arterial pressure above baseline values accompanied with reflex bradycardia.
Valsalva pulse pressure variation (∆VPP)was calculated as the percent variation between the highest pulse pressure during phase 1 (PPmaxphase1) and the lowest pulse pressure during phase 2 (PPminphase2): ∆VPP (%) = 100 x (PPmaxphase1 – PPminphase2) / [(PPmaxphase1 + PPminphase2)/2].
Valsalva systolic pressure variation (∆VSP) was calculated as follows: ∆VSP (%) = 100 x (SPphase1 – SPphase2) / ([SPphase1 + SPphase2]/2), where SPphase1 is the higher systolic pressure during phase 1 and SPphase2 is the lowest systolic pressure during phase 2.
Figures E2 and E3 are examples of ∆VPP and ∆VSP calculation in two patients with normal sinusoidal response.
Abnormal arterial pressure response or square-wave response to the VM (Fig. E4)
Square-wave response is characterized by an unchanged or even an increased arterial pulse pressure throughout the strain period. Late overshoot phase can be solely absent, often associated with more moderated forms of left ventricular dysfunction [3].
In this condition, the proposed method to calculate ∆VPP and ∆VSP could be a bit difficult, especially in patients with severe heart failure.Applying strictly the definition of ∆VPP, we took the first pulse pressure after the beginning of strain as the PPmaxphase1, and the lowest pulse pressure before the arterial pressure drop in phase 3 as PPminphase2 (on some cases, it was virtually impossible to discriminate the early and late phase 2).
This response was observed in three patients before VE; all of them were nonresponders and had ischemic cardiac disease history in their clinical evaluation (Fig. E5 and E6). Their ΔVPP and ΔVSP was markedly low at baseline (6%, 10%, -5% and 9%, 10%, -10%, respectively) and in all three cases VE produced a worsening of SVi (decrease from 18% to 9% with respect to baseline levels) and cardiac index (decrease from 18% to 6% with respect to pre-VE values).
This response should be taken into account with special caution, so the mere presence of this arterial waveform profile during the Valsalva maneuver should alert us about a severe heart failure situation, and therefore, fluid administration may be harmful and should be avoided.
Discussion
One of the major strength of this study is that Valsalva maneuvers were performed in a similar way between responder and nonresponder patients, since the average airway pressure measured during strain was not significantly different in both groups. Thus, if we assume that airway pressure increase during the VM should be equal to pleural pressure increase (as transpulmonary pressure remained unchanged during strain), pleural pressure should be transmitted to cardiac chambers in similar degree in all patients. Therefore, changes observed in arterial pressure during the VM should be solely related to increases in total blood volume induced by fluid administration and differences in the preload dependence condition among patients.
ADDITIONAL FIGURES
Fig. E1 Normal sinusoidal pattern on the arterial pressure response to the Valsalva maneuver.
Fig. E2Normal sinusoidal pattern on arterial pressure response to the Valsalva maneuver. In this patient, ∆VPP at preinfusion time was 64%. The increase in stroke volume index after fluid administration was 21% from baseline (from 39 mL/m2 to 47 mL/m2). ∆VPP at postinfusion time was 32% (a decrease of 32% with respect to preinfusion value).
Fig. E3Normal sinusoidal pattern on arterial pressure response to the Valsalva maneuver.Another example of arterial pressure sinusoidal pattern. PPmaxphase1 and PPminphase2 are easily recognizable. SPphase1 (the higher systolic pressure during phase 1) and SPphase2 (the lowest systolic pressure during phase 2) are also clearly identified for Valsalva systolic pressure variation (∆VSP) calculation.
Fig. E4Abnormal arterial pressure response (square-wave response) to the Valsalva maneuver.
Fig. E5Abnormal arterial pressure response to the VM or square-wave response. In this concrete case, PPminphase2 was slightly higher than PPmaxphase1, probably due to an enhanced left ventricular function that sustains left ventricular stroke volume during the strain time, so the calculated ∆VPP and ∆VSP had negative values (-5% and -10% respectively).This patient had an ischemic cardiac disease and congestive heart failure history (NYHA III) in her clinical evaluation.
Fig. E6Abnormal arterial pressure response to the VM or square-wave response.Another example of an abnormal arterial pressure response. Valsalva pulse pressure variation (∆VPP) and Valsalva systolic pressure variation (∆VSP) at preinfusion time were 6% and 9% respectively. Stroke volume index decreased by 9% with respect to baseline level after fluid administration (500 cc of Voluven® over 30 minutes).This patient had a clinical history of ischemic cardiopathy and left ventricular heart failure.
References (supplementary material)
1.Manecke GR, (2005) Edwards FloTrac sensor and Vigileo monitor: easy, accurate, reliable cardiac output assessment using the arterial pulse wave. Expert Rev Med Devices 2: 523-527
2.Michard F, Teboul JL, (2002) Predicting fluid responsiveness in ICU patients: a critical analysis of the evidence. Chest 121: 2000-2008
3.Felker GM, Cuculich PS, Gheorghiade M, (2006) The Valsalva maneuver: a bedside "biomarker" for heart failure. Am J Med 119: 117-122
1