To the Editor:—
In an article by Jungwirth et al. 1investigating influences of xenon application to rats undergoing cardiopulmonary bypass, it was reported that when rats received either xenon or nitrogen during ventilation on cardiopulmonary bypass, repeated air injections into the internal carotid artery were found only to negatively influence neurologic outcomes of animals exposed to cardiopulmonary bypass (CPB) procedures, regardless of the gas used for ventilation. Animals not undergoing CPB (sham groups) had significantly better neurologic outcomes. The kind of gas used for ventilation (xenon vs. nitrogen) did not influence the results; outcomes were influenced only by the total volume of injected air.1
Those findings contradict results published earlier by the same group.2Using the same model (injected air while on CPB), they found that the inhalation of xenon adversely affected the neurologic and histologic outcomes of the rats as compared with rats not being ventilated with xenon, but with nitrogen, during CPB.2
The results of the group published in 2006 were also at variance with experimental studies having found:
Only negligible small bubble expansion caused by xenon in a CPB circuit,
Preclinical findings of beneficial effects of xenon for rats reducing neurologic damages in a comparable CPB model,
And no adverse effect of xenon delivered on CPB to humans.3–7
The different findings in both studies, as published by other groups, deserve some attention regarding the experimental procedures. It would also be beneficial to allow a few questions concerning some descriptions of experimental details to erase possible misunderstandings. Finally, the interpretations of the results could be discussed briefly. For a complete assessment, we have not only studied the publications from Jungwirth et al. , but also both doctoral theses on which the publications were based and which are published electronically by the academic library of the University of Munich.*†
Establishing and Validation of the Model
The timeline of presenting and publishing all related papers is rather confusing to the author of this letter. The study published in 2006 is based on a doctoral thesis submitted by Carlsen JM to the veterinary faculty of the University of Munich. Its primary subject is neurologic outcome after xenon treatment in rats after CPB. The thesis submitted by Berkman was carried out to establish and validate the animal model,*†and was submitted in Munich on the same day as the treatment study, but was made available to a broader audience not before 2007, 1 yr after the results were published.
Title of the Studies
“Xenon impairs neurocognitive and histologic outcome after cardiopulmonary bypass combined with cerebral air embolism in rats” contains two speculations that were not investigated: The bubbles’“distance of travel” and “time to rest” in the brain’s vascular system are not known and were not investigated. Neither is it known whether persisting occlusions of the brain vessels were responsible for infarcts, found in histologic examinations. Any other factor known to negatively influence outcomes after CPB (e.g. , nonpulsatile flow, solid emboli, absence of cerebral autoregulation) at the least also could have been the cause of the troubles.
That the differences found between groups were caused by protective or destructive acting agents must take into account differences in adverse outcomes, which already were found in rats being treated with no significantly different agents. The causes of those varying outcomes in rats are extreme varieties of collateral brain perfusion. Another question concerns this detail: Cerebral infarct volumes were given as a mean ± SD of 82.7 ± 16.2 mm3in xenon animals and 33.2 ± 19.1 mm3in controls.2In the corresponding doctoral thesis, these values are given as 82.7 ± 51.3 mm3and 33.2 ± 60.4 mm3, respectively.*†Replacing the given SD with SEM values does not solve that problem.
Selection Criteria of Animals, Entering Calculations
“After CPB, animals demonstrating severe neurologic dysfunction were killed.”2*†“Animals not recovering from anesthesia after 3 h are subjected to brain death diagnostic and exsanguinated in deep isoflurane anesthesia. Animals exhibiting severe neurologic damage also are sacrificed during the first postoperative hours. In addition, euthanasia is carried out in animals showing clear signs of neurologic damage, e.g. , not being able to eat or drink.”*†It is described in Anesthesiology that “after CPB all neurologic, cognitive, and behavioral test procedures were performed by an investigator, blinded for the treatment.” It is not mentioned that prior to those test procedures an unknown number of animals presenting “severe neurologic dysfunctions” such as “not having recovered from anesthesia until 3 hours after CPB” were killed by the (unblinded) main investigators. It therefore cannot be excluded that this selection process was influenced by bias that makes all results questionable.
In the most recent paper,1the mean arterial pressures of sham animals were between 105 ± 18 mmHg after 45 min of CPB and 126 ± 20 mmHg after 90 min of CPB. CPB-treated animals had mean arterial pressures of 81 ± 16 mmHg after 45 min of CPB to a maximum of 86 ± 20 mmHg after 90 min of CPB. At the same time, neurologic impairments were found to be more severe in CPB-treated animals, with no differences between groups.
In the authors’ earlier paper,2the lowest blood pressures occurred during the first 45 min of CPB in the xenon groups, 71 ± 15 mmHg versus 88 ± 22 in the CPB nitrogen animals, and the values were 76 ± 11 mmHg in the xenon group versus 75 ± 14 mmHg in the nitrogen group at 90 min, the end of CPB. Sham animals during the same phase had mean arterial pressures between 127 ± 13 and 129 ± 17 mmHg during the same phases. Differences were found in neurologic outcomes comparing xenon to nitrogen animals submitted to CPB treatment. Duration of hypotensive phases was not given in any group. The authors only state that, “Hypotension is caused by the nature of CPB.” Can it therefore be asked whether the titles of both papers are slightly misleading in assuming that “cerebral air emboli” have caused any negative outcomes in rats? Cerebral air embolism is an event that was not investigated in the animals. Instead, “air bubbles” were injected into carotid arteries, and the authors speculated that these occluded cerebral vessels.
Differences in outcomes were found only between groups with different mean arterial blood pressures during CPB. Other investigators have described other external factors that can influence blood pressure: Low temperatures or elevated hydrostatic pressures (mean arterial pressure, intracranial pressure) can decrease bubble size, high flow velocities can destroy bubbles and lead to “foaming” effects, and low blood velocities can lead bubbles to form large entities.8–10In addition, low arterial blood pressures impair collateral perfusion of ischemic areas. Maintaining adequate blood pressure is one method to prevent neurologic damage during CPB and especially in cerebral embolism.11,12Hypotension, as remarked by the authors in both publications, may be caused by the “nature” of CPB. Should the “nature” of the anesthesiologist accept that fact without adequate reactions or therapeutic interventions, knowing that other investigators have published that the control of adequate levels of arterial blood pressures is mandatory during CPB with known gaseous embolic load?12
In summary, I would like to learn how severely the authors assessed the influences of differences in arterial blood pressures on the results and the interpretation of their studies. In addition, it would be interesting to know why animals displaying the most severe results were excluded from the study population, whether or not a blinded or nonblinded investigator carried out those exclusions, and how many animals in each group were excluded using that procedure.
Anesthesia Research, Medical Gas Group, Air Liquide Research Center, Jouy-en-Josas Cedex, France. email@example.com