We appreciate the interest in our report1by Drs. Rozet and Lam and by Dr. Gallart et al. , and we welcome the opportunity to address the important issues raised in their letters. We agree with them that our results are limited by the retrospective study design and require a prospective study for confirmation, as we stated in our report.

The other concerns of Drs. Rozet and Lam seem to derive from their criticism that “the study did not include any patient who developed frank propofol infusion syndrome.” Although it is true that none of our patients exhibited severe metabolic acidosis, the purpose of our study was not to detect fulminant propofol infusion syndrome, which is rare and unlikely to be detected in the size population we studied, but to determine whether we could detect “initially reversible and usually benign” metabolic acidosis in patients receiving propofol.1Drs. Rozet and Lam’s assertion that propofol infusion syndrome includes only “unexplained severe metabolic, usually lactic, acidosis and is associated with high-dose or long-duration propofol infusion, usually both,” implies that the mild syndrome either never occurs or never progresses to become severe and clinically evident. These assertions are incorrect (see references 12, 14–17, and 19 and the associated discussion in our article1) and highlight the need to address the question we posed in our study.

As Drs. Rozet and Lam suggest, defining metabolic acidosis for our study was challenging. They suggest that a definition of base excess (BE) of −3 mEq/l or less would be more appropriate than BE of −2 or less, which we chose, to “meet the generally accepted criterion” given in their reference 2. However, that reference and others we are aware of refer to stratifying critically ill intensive care unit patients, with the acidotic population in their reference 2 having a mortality of 45%.2In contrast, we were trying to detect early metabolic acidosis in a relatively healthy population where no mortality was expected. As we noted in our report, “in three well-documented cases of the syndrome, the initial negative BE was −2 to −3,”1so we prospectively selected BE of −2 or less before data abstraction. Furthermore, it is clear from inspecting our data (table 1 and fig. 1) that excluding the few patients with maximal negative BE of −2 would cause only minor changes in the incidence of metabolic acidosis in our study population and would not change the statistical comparison with our comparator group.

Our original submission did contain data on time and propofol dose for each arterial blood gas sampling in all patients who had multiple arterial blood gases drawn. This was deleted in the interest of shortening the article, but did not appreciably affect our analysis.

We look forward to publication of the details of the retrospective study of Drs. Rozet and Lam. We congratulate them on the collection of more detailed parameters of acid–base status, including lactate and chloride, than were available to us in our study. It will be important for the report to include data on calculated BE as well as the other parameters listed, because it is the parameter most indicative of metabolic acidosis independent of ventilator adjustments.

We appreciate Dr. Gallart et al.  making us aware of the article by their group where arterial blood gases and lactate levels were obtained during propofol anesthesia while studying almitrine and nitric oxide during one-lung ventilation.3Both in terms of total dose and time administered, their patients’ propofol exposure was less than a third of ours, suggesting a possible lower threshold for propofol metabolic acidosis. However, their study, although well designed for the questions it was addressing, may be limited by variables besides propofol: Almitrine is known to affect metabolic acidosis, ventilation was adjusted after institution of one-lung ventilation while BE was not reported, and cardiac index decreased and then increased significantly during the study.

It is important to elucidate the mechanism and incidence of propofol infusion syndrome to better prevent and treat it. We appreciate the interest of Drs. Rozet and Lam and Dr. Gallart et al.  in addressing this area of concern.

*Mayo Clinic College of Medicine, Rochester, Minnesota. johnson.michael@mayo.edu

Cravens GT, Packer DL, Johnson ME: Incidence of propofol infusion syndrome during noninvasive radiofrequency ablation for atrial flutter or fibrillation. Anesthesiology 2007; 106:1134–8
Gunnerson KJ, Saul M, He S, Kellum JA: Lactate versus  non-lactate metabolic acidosis: A retrospective outcome evaluation of critically ill patients. Crit Care 2006; 10:R22
Silva-Costa-Gomes T, Gallart L, Valles J, Trillo L, Minguella J, Puig MM: Low- versus  high-dose almitrine combined with nitric oxide to prevent hypoxia during open-chest one-lung ventilation. Br J Anaesth 2005; 95:410–6