In Reply

We would like to thank Mikhael et al. for their interest and comments related to our article¹  and appreciate the opportunity to respond to the raised questions. Our study was by no means designed to prove the usefulness of cutaneous mitochondrial oxygen tension (mito-Po₂) as physiologic transfusion trigger, but had the more modest goal of investigating the effects of hemodilution on this parameter. Because of the surprisingly useful behavior of mito-Po₂ in our model, i.e., a steep decline when the individual critical hemoglobin threshold during hemodilution is neared, mito-Po₂ seems a very promising parameter for use in transfusion medicine. Obviously, as also pointed out in the accompanying editorial,²  both technical and physiologic questions remain and the work has only just begun. In their letter, Mikhael et al. raise some of such remaining questions.

For example, the influence of anesthetic agents on mitochondrial function and their effect on mito-Po₂ remains to be investigated. Mito-Po₂ reflects the balance between cellular oxygen supply and demand,³  and any drug that alters mitochondrial respiration could therefore have an effect on this parameter. On the supply side, drugs inducing vasodilation or vasoconstriction will alter cellular oxygen supply and might therefore also affect mito-Po₂. So, the authors are correct in suggesting that the use of specific anesthetic agents might influence mito-Po₂ measurements. However, our choice for intravenous agents was not because of any potential advantageous hemodynamic profile, but because it is the standard anesthesia regimen in our laboratory.⁴  The animal ventilator used for the protocol was not equipped for the administration of inhalation agents.

As the authors state, heating the measuring site to 38°C might improve local regional blood flow, which potentially counteracts adaptive peripheral vasoconstriction. For us, the primary reason for heating the skin was not because of such vasoactive effects. The calibration constants needed for conversion of a measured triplet state lifetime to oxygen tension,⁵  and thus for the mito-Po₂ measurements, are temperature dependent, as is cellular metabolism. Therefore, we chose to standardize the skin temperature in our protocol. It would definitely be interesting to alternatively measure skin temperature and correct the mito-Po₂ measurement accordingly. Such an algorithm is currently under development and evaluation.

For our study design, we followed the concept of critical hemoglobin level (Hb_crit) to be defined as the Hb below which total body oxygen consumption starts depending on oxygen delivery.⁶  Surely this definition of Hb_crit does not guarantee that certain tissues do not become hypoxic at an earlier stage during hemodilution. Key in the idea to use mito-Po₂ in the skin as the parameter in transfusion-related decision making is the “canary” hypothesis. As pointed out by O’Brien and Schmidt,²  “This ‘canary’ hypothesis—that cutaneous mitochondrial PO₂ changes foretell changes in other vital organs—while plausible, remains unproven.” Therefore, as the authors rightfully suggest, efforts to gather information about the oxygenation status of other organs (like the brain) in relation to cutaneous mito-Po₂ under various pathophysiologic circumstances are important to take into account in future studies.

The authors state that inclusion of mixed venous oxygen saturation and arterial oxygen tension in the original publication could have been valuable and more informative. We understand that these data might be of help for some readers and therefore provide the values in this reply. For the control group at the nine time points from baseline to hemodilution step 8, the values for arterial oxygen tension were (mean ± SD) 199 ± 34, 199 ± 33, 199 ± 34, 199 ± 34, 196 ± 33, 196 ± 34, 196 ± 34, 194 ± 34, 196 ± 33 mmHg and for mixed venous oxygen saturation 72 ± 14%, 72 ± 17%, 68 ± 17%, 69 ± 18%, 66 ± 18%, 67 ± 17%, 66 ± 17%, 66 ± 15%, 66 ± 14%. For the experimental group, these values were (mean ± SD) 207 ± 8, 207 ± 10, 206 ± 5, 206 ± 7, 203 ± 11, 199 ± 12, 196 ± 24, 196 ± 14, 168 ± 56 mmHg and 72 ± 10%, 72 ± 7%, 73 ± 7%, 70 ± 8%, 65 ± 12%, 59 ± 16%, 55 ± 18%, 59 ± 17%, and 53 ± 15%, respectively.

We believe that autonomic metabolic reprogramming and mitochondrial dysfunction are important contributors to the pathophysiology of critical illness. Therefore, we aim at developing our technology to enable measurement of both mitochondrial oxygenation and respiration in vivo.^7–10  Since the ultimate goal of erythrocyte transfusion is maintaining a sound oxygen balance and preventing cell death by sustaining cellular respiration, mitochondrial respiration itself might prove to be an important transfusion threshold. In any case, the proof of the pudding will be in the eating, and more experimental and clinical studies are needed to fully understand the potential of mito-Po₂ measurements. Very recently a Conformité Européenne-marked clinical monitoring system enabling mito-Po₂ measurements at the bedside became available.¹¹  This COMET monitor, with COMET being an acronym for Cellular Oxygen METabolism, paves the way for clinical research and firm testing of the proposed concepts.