| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Department of Anesthesia, Intensive Care and Emergency, Poliambulanza Hospítal, Brescí, Italy.
Address correspondence and reprint requests to Giuseppe Natalini, MD, Unita’ di Terapia Intensiva – Intensive Care Unit Poliambulanza Hospital Via Bissolati 57, 25124 Brescia, Italy. Address e-mail to natalini-giuseppe{at}poliambulanza.it.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Up, Down, and changes in the preejection period on both arterial and photoplethysmographic waveforms were possible in 49 (86%) patients. The pulse pressure variation and preejection period were similar when calculated using both arterial blood pressure and photoplethysmography, whereas the other variables were different. Photoplethysmographic pulse variation >9% identified patients with arterial pulse pressure variation >13% (area under ROC curve = 0.85) or Down >5 mm Hg (area under ROC curve = 0.85). In hypotensive patients, photoplethysmographic pulse variation >9% remained the best threshold value (pulse pressure variation >13%: area under ROC curve = 0.90; Down >5 mm Hg: area under ROC curve = 0.93) for predicting fluid responsiveness. In conclusion, this study showed that pulse variations observed in the arterial pressure waveform and photoplethysmogram are similiar in response to positive pressure ventilation. Furthermore, photoplethysmographic pulse variation > 9% identifies patients with ventilation-induced arterial blood pressure variation that is likely to respond to fluid administration. |
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Down (systolic blood pressure decrease from the value during apnea) are more accurate indicators of fluid responsiveness than pulmonary artery occlusion pressure and left ventricular end-diastolic area (1,2,4). Another variable calculated by simultaneously recording invasive arterial blood pressure and electrocardiograph waveforms, the preejection period (PEP), showed correlation with changes of stroke volume index after fluid challenge (5). These data suggest that analysis of the arterial pressure waveform may be able to predict cardiac output changes after intravascular fluid administration. Discrimination between responders and nonresponders to intravascular volume expansion is also useful in anesthetized and mechanically ventilated patients without invasive monitoring. In these cases, the ability to predict fluid responsiveness without placing an arterial catheter would be useful and avoid the cost and potential complications of arterial catheter placement (6).
Pulse oximetry is a noninvasive monitoring technique routinely used to assess the oxygenation status of mechanically ventilated patients. This technique makes use of photoelectric plethysmography to detect changes in blood volume at the site of measurement (7). Although the photoplethysmographic waveform and the arterial waveform measure different physiologic changes in the arterial and venous vessels (7–9), a previous study showed that systolic blood pressure variation and Down correlated with similar indexes measured on the photoplethysmographic waveform after blood withdrawal (10).
Comparison, correlation, and agreement between ventilation-induced variation measured from arterial and photoplethysmographic waveform have never been performed.
This study was designed to 1) compare the ventilation-induced variation in arterial and photoplethysmographic waveforms in randomly selected patients in the operating room and intensive care unit undergoing controlled mechanical ventilation, 2) analyze the correlation and agreement between data collected from both the arterial pressure and photoplethysmogram waveforms, and 3) assess if photoplethysmographic-derived indexes of ventilation-induced arterial blood pressure variation can help to identify patients likely to respond to intravascular fluid administration.
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Invasive arterial blood pressure, electrocardiography, pulse oximetry, capnography, exhaled tidal volume, respiratory rate, peak airway pressure, positive end-expiratory pressure, and mean airway pressure were continuously monitored in all patients (AS3/CS3; Datex-Engstrom Division, Instrumentarium Corp., Helsinki, Finland). Patients were studied in the supine position, and zero pressure was measured at the midaxillary line. A photoplethysmographic waveform was obtained by applying the pulse oximeter probe to a finger or a toe. The location of the pulse oximeter was determined by selecting a site which yielded a high quality waveform. No attempt was made to relate the site of the pulse oximeter placement to the site of arterial catheter placement. A stable photoplethysmographic waveform was obtained before data acquisition.
Measurements and calculations were performed on the five consecutive respiratory cycles before an end-expiratory pause and mean values were used for statistical analysis. A representative patient is shown in Figure 1.
|
Systolic and diastolic blood pressures were measured on a beat-to-beat basis from the arterial pressure waveform, and pulse pressure was calculated as the difference between systolic and diastolic blood pressures. Maximal and minimal values for systolic blood pressure (SPmax and SPmin, respectively) and pulse pressure (PPmax and PPmin, respectively) were determined over a single respiratory cycle. The percentage change in pulse pressure was calculated as: PP(%) = 100 x {(PPmax – PPmin)/[(PPmax + PPmin)/2]} (11). Similarly the percentage change in systolic blood pressure was calculated as: SP(%) = 100 x {(SPmax – SPmin)/[(SPmax + SPmin)/2]} (2). Down values more than 5 mm Hg and PP(%) values more than 13% identify responders to intravascular volume expansion with accuracy more than 90% (1,2) and accordingly these threshold values were defined as "critical." Using the systolic blood pressure value during a period of 7-12 s of end-expiratory pause as a reference (1), the increase and decrease of systolic blood pressure during the respiratory cycle were defined, respectively, as Up and Down (12). The magnitude of Up and Down was expressed as a percentage of the systolic blood pressure value during apnea [Up(%) and Down(%), respectively] (1). The time interval between the beginning of the Q wave and the upstroke of the invasive radial arterial blood pressure curve was defined as the PEP. PEP values at the minimal and at the maximal arterial blood pressure over one respiratory cycle were defined as PEPe and PEPi, respectively. The percent PEP change over a respiratory cycle was calculated as: PEP(%) = 100 x {(PEPe – PEPi)/[(PEPe + PEPi)/2]} (5).
Photoplethysmographic waveforms were analyzed to obtain variables analogous to those obtained from the arterial pressure waveforms. The pulse amplitude of the photoplethysmographic wave was calculated as the difference between the systolic and diastolic value. Percentage changes over a single respiratory cycle of pulse [pulsepleth(%)] and systolic amplitude [systpleth(%)] were calculated similarly to PP(%) and SP(%), respectively. Uppleth(%) and Downpleth(%) were calculated on photoplethysmographic waveforms as Up(%) and Down(%), respectively. Finally, the time interval between the beginning of the Q wave and the upstroke of the photoplethysmographic trace was measured and PEPpleth(%) was calculated similarly to PEP (%). Hypotension was defined as a mean arterial blood pressure less than 70 mm Hg.
All patients’ lungs were mechanically ventilated in volume-controlled mode (Servo 300 Ventilator; Siemens-Elema AB, Solna, Sweden or ADU; Datex-Engstrom Division, Instrumentarium Corp.). Ventilatory variables were determined by the attending physician. The monitoring system was connected to a notebook (Omnibook 4150, Hewlett-Packard, Palo Alto, CA) and data were recorded at the sampling rate of 100 Hz and converted to ASCII files (Datex-Ohmeda S/5 Collect; Datex-Ohmeda Division, Instrumentarium Corp.). During data collection, the auto-gain function of the pulse oximeter was disabled. After recording 10 mechanical breaths, an end-expiratory pause lasting 12 s was performed and data acquisition was then stopped. Data were imported into a worksheet (Excel 2000; Microsoft Corporation, Redmond, WA) and instantaneous absolute values of arterial blood pressure, photoplethysmography data, electrocardiography, flow, and pressure at the airway opening were plotted against time, inspiration and expiration were identified, and spontaneous respiratory efforts or arrhythmias were excluded. All measurements were performed using the absolute numerical values reported in the worksheet at the times described above.
Preliminary data showed that 7, 11, and 48 patients were required for a 95% confidence interval on correlation coefficients comparing SP(%) versus syspleth(%), PP(%) versus pulsepleth(%), and Down(%) versus Downpleth(%), respectively.
Most of the calculated variables were not normally distributed as assessed by the Kolmogorov-Smirnov test. Data not normally distributed were analyzed by nonparametric tests. Results were expressed as mean ± sd or median and 10th and 90th percentiles as appropriate. Differences between groups were assessed by Wilcoxon signed ranks test or by Student’s t-test and 95% confidence intervals of the difference. Linear correlations were tested using Pearson’s product moment or Spearman rank method. Agreement between measurements and repeatability of measurements on the same subject were assessed by the procedure devised by Bland and Altman (13).
To assess the ability of variables calculated on photoplethysmographic waveform to identify patients with "critical" Down or "critical" PP(%), receiver operating characteristic (ROC) curves were generated, varying the discriminating threshold of each parameter. The optimal threshold value (the values that maximizes the sum of the sensitivity and specificity) was also determined. The areas under the ROC curves were calculated for each parameter and compared (14). For all comparisons, a P value <0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
There was no correlation between arterial blood pressure (mm Hg) and photoplethysmographic amplitude (arbitrary unit), for either systolic values (r = –0.148; P = 0.309) or pulse values (r = –0.04; P = 0.787). Differently from arterial blood pressure, photoplethysmographic systolic and pulse values were similar in hypotensive and nonhypotensive patients (Table 2).
|
Variables calculated for arterial and photoplethysmographic waveforms are shown in Table 3. Pulse variation and PEP were similar when calculated for arterial blood pressure and photoplethysmography, whereas the other variables were different.
|
A significant correlation was shown between PP(%) and pulsepleth(%) (r = 0.621; P < 0.001) and between SP(%) and syspleth(%) (r = 0.473; P = 0.001), whereas there was no correlation between Down(%) and Downpleth(%) (r = 0.219; P = 0.13), Up(%) and Uppleth(%) (r = –0.085; P = 0.562) and PEP (%) and PEPpleth(%) (r = 0.264; P = 0.066). In hypotensive patients a significant correlation was confirmed only between PP(%) and pulsepleth(%) (r = 0.573; P = 0.013) and between SP(%) and syspleth(%) (r = 0.602; P = 0.008).
Agreement between arterial and photoplethysmographic pulse variation and between arterial and photoplethysmographic systolic variation are shown in Figure 2. Because systolic differences varied in a systematic manner over the range of measurement, the logarithmic transformation of mean and difference of arterial and photoplethysmographic systolic variations is shown. The mean difference between PP(%) and pulsepleth(%) was 0% and the standard deviation of the differences was 6% (limits of agreement: –12% and 12%). The repeatability of PP(%) and pulsepleth(%) was 8% and 6%, respectively. The limits of agreement between logSP(%) and logsyspleth(%) were –2.25 and –0.16, predicting that for approximately 95% of cases SP(%) will be between 0.11 and 0.86 times syspleth(%). Repeatability of SP(%) and syspleth(%) was 7% and 21%, respectively. Mean differences and standard deviation of differences were similar in hypotensive patients.
|
Seventeen patients showed critical PP(%) values (>13%) and 23 patients had critical Down values (>5 mm Hg). The pulsepleth(%) threshold value of 9% allowed the best discrimination between patients with critical and noncritical PP(%) (sensitivity 100%, specificity 59%, area under ROC curve = 0.85) and between patients with critical and noncritical Down (sensitivity 91%, specificity 65%, area under ROC curve = 0.85). syspleth(%) discriminated less precisely than pulsepleth(%) between patients with critical and noncritical Down (area under ROC curve = 0.66, P = 0.003) and between patients with critical and noncritical PP(%) (area under ROC curve = 0.73, P = 0.041). When only hypotensive patients were considered, the pulsepleth(%) value of 9% remained the best threshold value, and it allowed discrimination between patients with critical and noncritical PP(%) (sensitivity 100%, specificity 75%, area under ROC curve = 0.90) and between patients with critical and noncritical Down (sensitivity 82%, specificity 100%, area under ROC curve = 0.93).
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Down and Up components was totally unrelated to the correspondent variables calculated using the arterial trace, making any apnea maneuver useless. Intermittent positive-pressure ventilation induces cyclic changes in left ventricular stroke volume. Because left ventricular stroke volume is a major determinant of systolic and pulse pressures, respiratory-induced changes in arterial blood pressure reflect respiratory changes in left ventricular stroke volume (15,16). Ventilation-induced stroke volume and/or arterial blood pressure changes predicted fluid responsiveness in several different clinical settings, including hypovolemic shock (3,12), sepsis (1,2), brain surgery (17), cardiac surgery (4,18), acute lung injury (11), and myocardial dysfunction (19). In all the different clinical and experimental settings the greatest ventilation-induced arterial blood pressure or stroke volume change was associated with the largest cardiac output increase after intravascular volume expansion.
In this study we assessed if photoplethysmography could noninvasively provide useful information on arterial blood pressure changes induced by mechanical ventilation in surgical and intensive care patients and identify those patients likely to respond to intravascular fluid administration. The photoplethysmogram is a highly processed signal developed to measure oxyhemoglobin saturation rather than blood volume. The data presented in this manuscript should be interpreted with a clear understanding of the nature of the photoplethsymogram. The unprocessed photoplethysmographic signal is generated both by the static component and the pulsatile component of the blood volume in the path of the light. Both systolic and diastolic values are positive, but during signal processing, auto-centering removes the static component, it amplifies and inverts the dynamic component (20). The result of the signal processing is negative diastolic values and pulse variation more than systolic variation. Background light, motion artifact, vasodilation, and vasoconstriction all influence photoplethysmographic signal generation and processing (7,20,21). Provided that these factors are constant throughout a respiratory cycle, ventilation should be the only cause of photoplethysmographic amplitude variation. Cautions and limitations in photoplethysmography amplitude evaluation always have to be considered when this technique is used.
The comparison of arterial and photoplethysmographic waveform needs some consideration. Quantifying the signal is the main problem in photoplethysmographic waveform analysis (22). As previously suggested, all photoplethysmographic variables were expressed as a percentage to allow comparisons between patients despite the lack of units (10). Similarly, we expressed variables derived from arterial waveform analysis as percentage values, as previously recommended in hypotensive patients (23). Furthermore, the origin of the photoplethysmographic and arterial waves is different. Ventilation-induced photoplethysmographic changes are caused by blood volume changes both in arterial and in venous bed (7,8). Moreover, variation in photoplethysmographic waveform amplitude depends on the intravascular pulse pressure as well as on the distensibility of the vascular wall (7).
Ventilation-induced pulse changes on both arterial and photoplethysmographic traces had never been investigated. Previous studies correlated systolic variations of arterial blood pressure with similar changes in the photoplethysmographic waveform (10,24), but their results were inconclusive. An observational study on a small sample (12 patients) had relevant methodological limitations (24); a randomized controlled trial compared arterial and photoplethysmographic ventilation-induced systolic variations in 12 patients undergoing instrumental posterior spine fusion involving hemodilution. Systolic variation and Down showed a significant correlation between arterial and plethysmographic calculations during hypovolemia but not during intravascular volume replacement (10).
The definition of critical values of pulse pressure variation and Down has some limitations. First, in a heterogeneous population we used threshold values derived from septic patients (1,2). Nevertheless, a recent paper showed a similar threshold value after coronary artery bypass grafting (4): in septic patients PP% >13% predicted cardiac index increases more than 15% (2), whereas after coronary artery bypass grafting PP% >11% predicted cardiac index increases more than 12% (4). Second, systolic blood pressure variation is affected both by the intravascular status and by the depth of tidal volume (25). The tidal volume used in study patients ranged from 5 to 12 mL/kg, whereas in the above-reported studies it ranged from 8 to 11 mL/kg (1) and from 8 to 12 mL/kg (2). In these ranges of tidal volume, pulse pressure variation does not correlate with tidal volume (11). Therefore, both these possible limitations should have a negligible impact on the study results.
Ventilation-induced arterial blood pressure changes can be used to correctly predict fluid responsiveness in more than 90% of patients (1,2,4). In the present study, pulsepleth(%) > 9% correctly identified approximately 90% of hypotensive patients with critical values of PP% or Down. It is reasonable to suppose that pulsepleth(%) could correctly predict fluid responsiveness in approximately 80% of hypotensive patients. This speculation suggests that pulsepleth could be more accurate than central venous pressure, pulmonary artery occlusion pressure, or left ventricular end-diastolic area to predict cardiac output changes after intravascular volume expansion. Future controlled clinical trials should assess if photoplethysmographic pulse variation predicts stroke volume increase after intravascular volume expansion. Because photoplethysmography is a volumetric monitoring tool (7), it could be equivalent to, or even better than blood pressure monitoring to detect beat-to-beat ventilation-induced stroke volume changes.
Data analyzed in this study did not show any relationship between arterial pressure value and photoplethysmogram amplitude (Table 2). These findings underscore the different meaning and modality of blood pressure and photoplethysmographic monitoring. Accordingly, the attempt to relate photoplethysmography-derived variables to arterial blood pressure values would give unsatisfying results (26) or it could be questionable from a theoretical point of view (27).
In conclusion, this study showed that arterial pulse pressure variation in response to positive pressure ventilation is strongly reflected by pulse variation of the photoplethysmographic waveform. Moreover, plethysmographic pulse variation more than 9% identifies hypotensive patients who are likely to increase cardiac output after intravascular volume expansion, as judged by pulse pressure variation or Down. Despite some limitations, the ability of photoplethysmographic pulse variation to predict critical thresholds of fluid responsiveness suggests a role in the initial steps of the decision-making process concerning intravascular volume expansion in surgical or critically ill patients.
|
ACKNOWLEDGMENT |
|---|
|
Footnotes |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. H. Shelley Photoplethysmography: Beyond the Calculation of Arterial Oxygen Saturation and Heart Rate Anesth. Analg., December 1, 2007; 105(6S_Suppl): S31 - S36. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Cannesson, O. Desebbe, and J.-J. Lehot Fluid Responsiveness Assessment Using the Pulse Oxymeter Waveform: Not Yet Ready for Prime Time Anesth. Analg., June 1, 2007; 104(6): 1598 - 1599. [Full Text] [PDF]
|
|
|
|
|
|
B. Tavernier, M.-S. Destandau, and B. Vallet Variations in Photoplethysmographic Waveform During Mechanical Ventilation Anesth. Analg., June 1, 2007; 104(6): 1599 - 1599. [Full Text] [PDF]
|
|
|
|
|
|
F. Michard Using Pulse Oximetry Waveform Analysis to Guide Fluid Therapy: Are We There Yet? Anesth. Analg., June 1, 2007; 104(6): 1606 - 1607. [Full Text] [PDF]
|
|
|
|
|
|
G. Natalini Variations in Photoplethysmographic Waveform During Mechanical Ventilation Anesth. Analg., June 1, 2007; 104(6): 1599 - 1600. [Full Text] [PDF]
|
|
|
|
 |
|
 C. Runcie, B. Tavernier, H. Solus-Biguenet, M. Fleyfel, and B. Vallet Predicting fluid responsiveness in theatre Br. J. Anaesth., April 1, 2007; 98(4): 545 - 547. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ercole Assessing fluid responsiveness: the role of dynamic haemodynamic indices Trauma, January 1, 2007; 9(1): 13 - 19. [Abstract] [PDF]
|
|
|
|
|
|
G. Natalini, A. Rosano, M. Taranto, B. Faggian, E. Vittorielli, and A. Bernardini Arterial Versus Plethysmographic Dynamic Indices to Test Responsiveness for Testing Fluid Administration in Hypotensive Patients: A Clinical Trial Anesth. Analg., December 1, 2006; 103(6): 1478 - 1484. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Feldman Can Clinical Monitors Be Used as Scientific Instruments? Anesth. Analg., November 1, 2006; 103(5): 1071 - 1072. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
