Failure to cite leads to ignoble end for xenon paper, and a correction

anaesthesiaXenon may be an inert gas, but that doesn’t mean papers about the molecule aren’t subject to change.

Indeed, the journal Anaesthesia has retracted a 2010 article about xenon-based anesthesia, and corrected a 2005 article by some of the same researchers, for what appears to be a case of wurst slicing.

The 2005 paper, “Comparison of xenon-based anaesthesia compared with total intravenous anaesthesia in high risk surgical patients,” came from a group at the Department of Anaesthesiology and Intensive Care Medicine at University Hospital Schleswig-Holstein, in Kiel, Germany. It has been cited 10 times, according to Thomson Scientific’s Web of Knowledge.

According to the abstract:

Xenon, a noble gas with anaesthetic and analgesic properties, has gained renewed interest due to its favourable physical properties which allow a rapid emergence from anaesthesia. However, high costs limit its use to a subset of patients who may benefit from xenon, thereby offsetting its costs. To date, there are only limited data available on the performance of xenon in high risk patients. We studied 39 patients with ASA physical status III undergoing aortic surgery. The patients were randomly assigned to either a xenon (Xe, n = 20) or a TIVA (T, n = 19) group. Global cardiac performance and myocardial contractility were assessed using transoesophageal echocardiography, and myocardial cell damage with troponin T and CK-MB. Echocardiographic measurements were made prior to xenon administration, following xenon administration, and after clamping of the abdominal aorta, after declamping and at corresponding time points in the TIVA group. Laboratory values were determined repeatedly for up to 72 h. Data were analysed using two-way anova factoring for time and anaesthetic agent or with ancova comparing linear regression lines. No significant differences were found in global myocardial performance, myocardial contractility or laboratory values at any time during the study period. Mean (SEM) duration of stay on the ICU (xenon: 38 +/- 46 vs. TIVA 25 +/- 15 h) or in hospital (xenon: 14 +/- 12 vs. TIVA 10 +/- 6 days) did not differ significantly between the groups. Although xenon has previously been shown to exert superior haemodynamic stability, we were unable to demonstrate an advantage of xenon-based anaesthesia compared to TIVA in high risk surgical patients.

But the correction indicates that the authors had failed to mention a previous study — also in Anaesthesia — based on the same patients:

In the article ‘Comparison of xenon-based anaesthesia compared with total intravenous anaesthesia in high risk surgical patients’ [1], the authors omitted to refer to their previous paper [2] describing the same patients. The authors apologise for this oversight.

The previous paper, “Effect of xenon anaesthesia on accuracy of cardiac output measurement using partial CO2 rebreathing,” found that:

Cardiac output (CO) determination based on partial CO(2) rebreathing has recently been introduced into clinical practice. The determination of flow is crucial for exact CO readings and the physical properties of xenon, i.e. high density and viscosity, may influence flow readings. This study compared echocardiography-derived CO measurements with the partial rebreathing method during total intravenous (TIVA) vs. xenon-based anaesthesia. Thirty-nine patients ASA physical status III undergoing aortic reconstruction were randomly assigned to receive either xenon (Xe, n = 20) or TIVA (T, n = 19) based general anaesthetic. Paired measurements were taken before xenon administration, after xenon administration, before and after clamping of the abdominal aorta and after declamping and at corresponding time points in the TIVA group. Data were analysed with a Bland-Altmann plot. Bias and precision were acceptable and comparable before xenon administration (T 0.54 +/- 0.92 l.min(-1) vs. Xe 0.11 +/- 1.1 l.min(-1)), but after xenon administration CO was largely overestimated by partial CO(2) rebreathing (T 0.04 +/- 0.91 l.min(-1) vs. Xe -4.0 +/- 2.1 l.min(-1)). In the TIVA group, bias and precision after declamping increased significantly (P < 0.01) compared to all time points except baseline. In its current application, the NICO cardiac output monitor appears to be inappropriate for determination of CO during xenon based anaesthesia.

For the retraction, the same. That article, “Differences between bispectral index and spectral entropy during xenon anaesthesia: a comparison with propofol anaesthesia,” has been cited three times and reported that:

We enrolled 114 patients, aged 65-83 years, undergoing elective surgery (duration > 2h) into a randomised, controlled study to evaluate the performance of bispectral index and spectral entropy for monitoring depth of xenon versus propofol anaesthesia. In the propofol group, bispectral index and state entropy values were comparable. In the xenon group, bispectral index values resembled those in the propofol group, but spectral entropy levels were significantly lower. Mean arterial blood pressure was higher and heart rate was lower in the xenon group than in the propofol group. Bispectral index and spectral entropy considerably diverged during xenon but not during propofol anaesthesia. We therefore conclude that these measures are not interchangeable for the assessment of depth of hypnosis and that bispectral index is likely to reflect actual depth of anaesthesia more precisely compared with spectral entropy.

But per the retraction notice:

The above article from Anaesthesia, published online on 16 April 2010 in Wiley Online Library ( and in Volume 65, No. 6, pp. 595–600, has been retracted by agreement between the journal Editor-in-Chief, Steve Yentis, the authors and John Wiley & Sons Ltd. This article is being retracted because it did not refer to a previous paper arising from the same work on the same patients [1], and different power calculations and primary outcomes were presented in these papers as if they were separate studies. In addition, we are informed by the Research Ethics Committee Chairman that the Committee was not informed that the protocol and sample size had been changed after approval, nor that the study had finished, both of which were required by the Committee.

That referenced paper appeared in Anesthesiology, an American specialty publication. Here’s its abstract:


Postoperative cognitive dysfunction (POCD) in elderly patients after noncardiac surgery is a common problem. The noble gas xenon has been demonstrated to exert substantial neuroprotective properties in animal studies. Therefore, this study was designed to assess POCD after xenon anesthesia in comparison to propofol in elderly patients undergoing major noncardiac surgery.


After approval of the local ethical committee was obtained, 101 patients (American Society of Anesthesiologists physical status I-III; age, 65-83 yr) undergoing elective abdominal or urologic surgery (duration, > 2 h) were enrolled into this randomized, double-blinded controlled pilot study. Patients received anesthesia with sufentanil and either propofol or xenon and were assessed before treatment and 1, 6, and 30 days after treatment using a neuropsychological test battery based on previous studies investigating POCD.


There were no significant differences in terms of age, American Society of Anesthesiologists status, education, duration of surgery, administered analgetics, and preoperative neurocognitive status between study groups. POCD as classified was present in 22 patients (44%) of the xenon group versus 25 patients (50%) of the propofol group 1 day after treatment, in 6 xenon patients (12%) versus 9 propofol patients (18%) 6 days after treatment, and in 3 xenon patients (6%) versus 6 propofol patients (12%) 30 days after treatment. These differences were not statistically significant.


Postoperative impairment of neurocognitive function was observed in a substantial proportion of elderly patients even 30 days after treatment. Xenon-based anesthesia was not associated with decreased incidence of POCD in comparison to propofol.

Anaesthesia and its editor, Steve Yentis, was one of the journals that aggressively prosecuted the editorial case against retraction leader Yoshitaka Fujii. Indeed, it was Yentis who published the statistical analysis of Fujii’s papers that demonstrated conclusively that the Japanese anesthesiologist must have faked his results.

So we were somewhat surprised to find Yentis with only this to say about the xenon papers:

I’m afraid I can’t give you any details about the deliberations leading up to our retraction notice, sorry.

2 thoughts on “Failure to cite leads to ignoble end for xenon paper, and a correction”

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.