We thank Drs. Lichtor and Takizawa et al.  for their interest in our study and thought-provoking observations. Dr. Lichtor points out that in settings of hypotensive extremis from blood loss i) perfusion to the brain can be compromised to the point of significantly altering electrical activity in the brain and ii) delivering an anesthetic that results in brain isoelectricity is clearly excessive. It is interesting to note that in our experimental protocol, we performed a series of pilot studies directed at characterizing the impact of our hemorrhage and resuscitation protocol on a processed measure of brain electrical activity, the Bispectral Index score, in the presence of isoflurane. Although hemorrhage was severe (42 ml/kg) and resuscitation restored hemodynamic parameters to near-baseline levels, we observed no significant change in the Bispectral Index (i.e. , no Bispectral Index decrease beyond that initially produced by isoflurane). We point out, however, that in our protocol, the mean arterial blood pressure never dropped below 40 mmHg and resuscitation was initiated just at the onset of cardiovascular decompensation. The clinical correlates here are twofold: 1) although hypotensive, compensatory mechanisms most likely maintained adequate cerebral perfusion to sustain brain electrical activity, and 2) resuscitation was initiated before going beyond a time that loosely coincides with the end of the “golden hour.” Once in a severely hypotensive (i.e. , less than a mean arterial blood pressure of 40 mmHg) or in a decompensated cardiovascular state (e.g. , beyond the “golden hour”), brain electrical activity may be dramatically altered, as suggested by the studies referenced in Dr. Lichtor’s letter.

Dr. Lichtor also points out that although the dose response of blood loss to changes in the Bispectral Index is not well defined, Bispectral Index can be useful in titrating an anesthetic. We concur with his recommendations that monitoring brain electrical activity during surgeries associated with excessive hemorrhage may offer a pragmatic approach to titrating the appropriate dose of anesthetic when the consequences of overdosing can be unpredictable.

In a previous study, we reported that blood loss alone led to a dramatic change in the pharmacokinetics and pharmacodynamics of propofol. In a follow-up study, we explored whether or not resuscitation would reverse these shock-induced changes in propofol kinetics and dynamic behavior. The most compelling finding was that despite resuscitation with crystalloid to near-baseline hemodynamics, the pharmacologic behavior of propofol was still altered when compared with controls. We concluded that hemorrhagic shock followed by resuscitation with lactated Ringer’s solution nearly restored the pharmacokinetic profile of propofol to a pre-hemorrhage state, but that resuscitation did not reverse the pharmacodynamic changes.

From a pharmacokinetic analysis standpoint, our work was primarily observational. We measured plasma propofol levels and estimated volumes and clearances using compartmental models and used these estimates to make comparisons between study groups. We did not make measurements that would allow us to discover how drug distribution and clearance were altered by hemorrhagic shock and resuscitation. For example, we did not measure or estimate i) plasma protein content, ii) propofol-plasma protein binding, or iii) unbound propofol levels throughout our experimental protocol. Furthermore, we did not measure how the initial distribution and subsequent redistribution of propofol was altered following blood loss and resuscitation or how altered plasma pH levels may have impacted unbound propofol availability. Finally, we did not explore to what extent the clearance of unbound propofol by metabolic organs would be compromised by our experimental protocol (e.g. , a comparison of hepatic extraction ratios for propofol between control and bled-resuscitated animals).

As suggested by Dr. Takizawa et al. , what we reported as an increase in end-organ sensitivity to propofol may be, at least in part, attributable to an unrecognized increase in unbound propofol. Thus our reported leftward shift in the C50of propofol may represent an undetected pharmacokinetic difference between groups. Although the measured plasma propofol levels were comparable between the control and hemorrhage-resuscitation groups, the amount of unbound propofol available to exert a pharmacologic effect may have been increased. After removing more than 50% of the estimated blood volume and replacing it with crystalloid, plasma protein content was most likely decreased. Furthermore, alterations in organ blood flow (as manifest by a change in systemic vascular resistance), capillary wall integrity, and plasma pH may have influenced the levels of unbound propofol.

* University of Utah, Salt Lake City, Utah. ken.johnson@hsc.utah.edu