THE number of novel noninvasive hemodynamic monitoring gizmos currently available and being developed is amazing. Independent of their accuracy in measuring what they say they measure, how they are used in clinical decision making is potentially more important because monitoring devices will only improve outcome if coupled to a treatment that itself improves outcome. For example, the continuous measure of cardiac output did little to improve patient outcomes until it was used as part of the measure of whole body oxygen delivery (Do²) as a targeted resuscitation algorithm in high-risk surgical patients before surgery. This approach, called preoptimization  , reduces postoperative morbidity1 and mortality2 and also reduces overall hospital cost.3 Although several modifications of the same approach have been used, all employ a volume challenge until cardiac output no longer increases and then add inotropes and/or vasopressors to reach their targeted Do².4 However, although volume loading is the traditional means to start resuscitation, fully half of all critically ill patients are not volume responsive, making this approach less efficient5 and potentially dangerous. If one could determine volume responsiveness before therapy, an appropriate and effective treatment algorithm could be used to drive these proven therapeutic approaches.6 Such a sensitive and specific parameter is known: assessment of arterial pulse pressure variation (PPV) during positive-pressure ventilation.7 A PPV of greater than 13%, when averaged over 3–6 breaths of a tidal volume of 5–8 ml/kg, is predictive of an increase in cardiac output of greater than 15% in response to a volume challenge of 250 ml colloid.8 However, such analysis requires the insertion of an arterial catheter, which itself is both time-consuming and associated with measurable morbidity. If a noninvasive means could be used to measure the same effect on PPV, it would be very useful clinically. Until recently, the only device available to make such measurements was a finger plethsomgraph.9 However, this device is relatively expensive, is not universally available in operating rooms and other acute care monitoring environments, and may not maintain accuracy as vascular tone varies.10 However, pulse oximetry is noninvasive, universally available, easy, and inexpensive to use. Furthermore, the pulse oximetry plethysmographic waveform amplitude is an essential variable in calculating pulse oximetry saturation and reflects the pulsatile change in tissue density during the cardiac cycle. Therefore, it is not surprising that the pulse oximetry plethysmographic waveform should resemble the arterial pulse pressure waveform in both shape and amplitude variation. However, there is no relation between absolute arterial pressure and the pulse oximeter signal, only in their variation over the ventilatory cycle. It follows, therefore, that if this relationship is not only qualitative but quantitative, it may be used as a surrogate measure of arterial pulse pressure variation and thus define preload responsiveness. Solus-Biguenet et al.  11 first described this phenomenon. They showed that pulse oximeter plethysmographic waveform amplitude variation (POV) predicted preload response in patients undergoing hepatic resection. In this issue of Anesthesiology, Cannesson et al.  12 extends these observations to show that POV can predict fluid responsiveness in a fashion similar to PPV across a wide range of surgical patients. These findings, when coupled to the preoptimization resuscitation protocol approach, may represent a highly cost-effective means to reduce anesthesia stress and decrease mortality, morbidity, and cost of surgery. Although such a prospective study must be performed, there are still specific issues with the use of POV that must be considered.

First, the pulse oximeter plethysmographic waveform displayed on the monitoring screen and reported by all commercially available pulse oximeters is a highly processed pulse density signal. The displayed pulse density signal is not really the absolute pulse density change but a time-averaged and mean-adjusted signal wherein the actual mean density is held constant but the dynamic changes in density are reported for quality control purposes. If no pulsatile signal is sensed, the pulse oximeter is unable to calculate oxygen saturation measured by pulse oximetry. The raw plethysmographic signal is much more variable. Therefore, the findings of both Solus-Biguenet et al.  11 and Cannesson et al.  12 must be validated in the setting of other pulse oximeter devices and different patient groups. Furthermore, the manufacturers of the various pulse oximeters must reintroduce the graphic display of POV as part of their usual output both onto the screen and into recoverable data logs.

Second, pulse oximeter plethysmographic density will be a function of tissue (nonchanging signal) and blood (changing signal) inputs, and its pulsatility will be primarily a function of changing blood density. Therefore, one must ask: What determines the blood density change over the sensing region? Clearly this will be a function of both perfusion pressure and vasomotor tone. As upstream vasomotor tone increases, for example, pulse oximeter plethysmographic changes would decrease for the same pulse pressure, and vice versa  with vasodilation. Accordingly, it would be interesting to see the relation between PPV and POV as cardiovascular conditions are varied by pharmacologic intervention and disease. Clearly, this new use of pulse oximetry is exciting and potentially very important. Let us define its value carefully and, if it is proven to be useful, apply this new use of an established monitor broadly to help both monitor and guide resuscitation.

Department of Critical Care Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania.