Fluid responsiveness using non-invasive predictors during major hepatic surgery
Lyon, France
* E-mail: maxime_cannesson{at}hotmail.com
Editor—We read with great interest the study by Solus-Biguenet and colleagues1 regarding the evaluation of fluid responsiveness using non-invasive predictors during major hepatic surgery. The authors compared the value of several indices of fluid responsiveness and concluded that the respiratory variations in arterial pulse pressure obtained from the invasive (PPVart) and FinapressTM (PPVfina) arterial pressure curves were the most accurate predictors of response to volume expansion. Moreover, they concluded that respiratory variations in the pulse oximetry waveform (PPVsat) were greater in responders than in non-responders to volume expansion even if the predictive value of this parameter was weaker than PPVart and PPVfina. These results are extremely interesting, as pulse oximeters are non-invasive, inexpensive, and are daily used in the operating theatre. Thus, the results from this study suggest that PPVsat may be a useful predictor of fluid responsiveness in the operating theatre. However, in our opinion, technical description regarding the way the pulse oximetry waveform was acquired in this study is not clear enough to sustain the hypothesis that PPVsat is a weaker predictor of fluid responsiveness than PPVart and PPVfina. The software generates a signal that is substantially filtered, amplified, and smoothed before display. However, some data acquisition software allows disengaging the automatic gain incorporated in the pulse oximeter in order to avoid the potential influence of the signal processing on the displayed waveform. Using this technique, PPVsat has been shown to be strongly related to PPVart with far lower limits of agreement than those described in the present study.2 We can postulate that Solus-Biguenet and colleagues did not control this limiting factor as no mention is made concerning the gain and as agreement between PPVart and PPVsat was lower than those previously reported.
As some recently published studies are suggesting, PPVsat is strongly influenced by the site of measurements (ear, finger, forehead).3 Consequently, it is of major importance to mention the site of measurement and to consistently use the same site between patients, as an up to 10-fold variation can occur. Mixing the sites may induce an important bias.
The authors postulate that the sensitivity of the plethysmographic signal to humoral and neurogenic factors may explain the poor predictive value of PPVsat. However, we can postulate that these potentially confounding factors are constant throughout a single respiratory cycle and that they do not impact on the minimal and maximal pulse oximeter waveform amplitudes during the same respiratory cycle. On the other hand, peripheral vasoconstriction may alter the pulse oximeter signal quality. Most monitors display a signal quality index or a perfusion index providing informations regarding the quality of the curve. These data should have been controlled before pulse oximeter waveform recording and analysis in order to avoid confounding factors related to poor signal quality.
Pulse oximeter waveform is influenced by outside light absorption. Thus, the pulse oximeter should be wrapped in order to prevent outside light from interfering with the signal. This is not mentioned in this study.
In conclusion, pulse oximetry waveform has been shown to be strongly related to respiratory cycles in previously published studies.4 These variations have been shown to be related to PPVart2 and to loading conditions.5 However, this waveform depends on signal processing, site of measurement, peripheral vasoconstriction, and outside light absorption. This study1 shows promising results regarding the ability of PPVsat to predict fluid responsiveness. We can postulate that further studies without automatic gain control, standardized site of measurements, and adequate signal quality index and recording will improve the predictive value of this new index.
Lille, France
* E-mail: btavernier{at}chru-lille.fr
Editor—We appreciate Dr Cannesson's interest in our recent publication.1 While we found that the respiratory variation in the pulse oximeter waveform (quantified by the PPVsat) predicted fluid responsiveness less accurately than either the respiratory variation in radial artery pulse pressure (PPVart) or the non-invasive PPV obtained using the FinapresTM device (PPVfina), Dr Cannesson suggests that technical problems in acquiring or processing the signal from the pulse oximeter waveform may be responsible for this lack of reliability.
Dr Cannesson first states that avoiding the potential influence of the signal processing on the displayed waveform by disengaging the automatic gain incorporated in the pulse oximeter may allow obtaining a stronger relation between PPVsat and PPVart, and postulates that we did not control this limiting factor. This, however, cannot be an explanation for our results, because we actually used a manual gain control in our study. Moreover, a change in gain between two sequences of measurements would not change the value of PPV, because PPV is a relative indice (the difference between minimum and maximum pulse waveforms within a respiratory cycle divided by the mean of the two values). This can be easily verified by changing the gain between two successive waveform recordings. In addition, it should be mentioned that auto-centering algorithms, which are not deactivated when automatic gain is disengaged, rather than gain itself, are likely to alter the effects of ventilation on the waveform.
Dr Cannesson cites the crucial importance of using only one site of measurement, because this has been shown to strongly influence the value of PPVsat.3 Dr Cannesson even states that PPVsat may be more than 10 times stronger in the region of the head when compared with the finger. This is an over-interpretation of the results of a study which quantified the amplitude of the power spectrum at the respiratory frequency following spectral analysis of the waveforms, not PPVsat itself, which quantifies only one component of the waveform variations induced by ventilation. Interestingly, this study3 strongly supports the notion that pulse oximetry cannot be recommended to accurately assess the respiratory variation in arterial pressure. Nevertheless, it is true that the whole waveform, including the respiratory change in its pulsatile component, and thus PPVsat, may vary in a given patient with the site of measurement. We have anecdotally observed in patients with septic shock that the effect of ventilation on the pulse oximeter waveform could even be dramatically different between two fingers of the same hand (data not shown). In our study, measurements in a given patient were always performed at the same finger, and never at the ear or forehead location.
Dr Cannesson also postulates that because they are constant throughout a single respiratory cycle, humoral and neurogenic factors should not significantly impact the minimal and maximal waveform amplitudes, and thus PPVsat, during the same respiratory cycle. We agree that this hypothesis may be true, although, at this time, it remains to be demonstrated. In fact, the clinical evidence for this to be true comes from our results obtained not with PPVsat, but with PPVfina, since the FinapresTM device also measures arterial pressure at a distal site and predicted fluid responsiveness as accurately as PPVart.
Finally, Dr Cannesson proposes that the pulse oximeter waveform should be used only when perfusion is good enough (as attested by the signal quality index of the monitor), the gain adequately controlled, the site of measurement unique, and the probe wrapped (to prevent outside light from interfering with the signal). In our study, poor signal or perfusion was explicitly recognized as a potential limitation of both PPVfina and PPVsat. Measurements were performed throughout major surgery, and reduced finger perfusion may in part account for the lack of agreement between PPVsat and PPVart. We feel that one of the most interesting results in our study was precisely that, in similar experimental conditions (of peripheral perfusion, temperature, light exposure...), PPVfina was a better predictor than PPVsat. These conditions are indeed different of those advocated by Dr Cannesson, but likely represent ‘real life’, when anaesthetists wonder whether fluids should be given to their patient in the operating theatre. We agree with Dr Canesson that PPVsat, in these conditions, may be of some value and, accordingly, our study showed that PPVsat was a better predictor than the classical static pressure measurements. Nevertheless, in accordance with other reports,3 6 our results also suggest that PPVsat should be used with caution in this indication. From a theoretical point of view, it is conceivable that modifications in device algorithms may allow commercial pulse oximeters to provide results similar to those obtained with the FinapresTM.
References
1 Solus-Biguenet H, Fleyfel M, Tavernier B, et al. (2006) Non-invasive prediction of fluid responsiveness during major hepatic surgery. Br J Anaesth 97:808–16.
2 Cannesson M, Besnard C, Durand PG, et al. (2005) Relation between respiratory variations in pulse oximetry plethysmographic waveform amplitude and arterial pulse pressure in ventilated patients. Crit Care 9:.
3 Shelley KH, Jablonka DH, Awad AA, et al. (2006) What is the best site for measuring the effect of ventilation on the pulse oximeter waveform? Anesth Analg 103:372–7.
4 Shelley KH, Awad AA, Stout RG, Silverman DG. (2006) The use of joint time frequency analysis to quantify the effect of ventilation on the pulse oximeter waveform. J Clin Monit Comput 20:81–7.[Medline]
5 Shamir M, Eidelman LA, Floman Y, Kaplan L, Pizov R. (1999) Pulse oximetry plethysmographic waveform during changes in blood volume. Br J Anaesth 82:178–81.
6 Michard F. (2005) Changes in arterial pressure during mechanical ventilation. Anesthesiology 103:419–28.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||