Oxygen after surgery: review questions WHO advice that high levels reduce infection

BMJ 2019; 364 doi: https://doi.org/10.1136/bmj.l777 (Published 18 February 2019)Cite this as: BMJ 2019;364:l777

World Health Organization guidelines on using high levels of oxygen during and after surgery have been questioned after claims that research by an Italian surgeon lacks credibility.

Mario Schietroma of the University of L’Aquila in central Italy is the author of many surgical studies, including several that found that liberal oxygen use reduced the risk of infection after surgery. Two papers by his group formed part of the evidence WHO considered in preparing its 2016 guideline.

The guideline states: “Adult patients undergoing general anaesthesia with endotracheal intubation for surgical procedures should receive 80% fraction of inspired oxygen intraoperatively and, if feasible, in the immediate post-operative period.”

Anaesthetists were frankly sceptical. Thomas Volk of the University of Saarland in Germany and colleagues wrote an editorial in Der Anaesthetist in 2017 saying that the guideline directly contradicted the results of many trials and did not provide useful guidance.1 No anaesthetists served on the WHO guideline group, they noted.

Retracted paper

In March 2018 one of Schietroma’s papers cited by the WHO group was retracted from the American Journal of Surgery after recalculation of the P values showed that none of them came close to significance despite claims made in the paper. In December WHO downgraded its guidance from strong to conditional.

A new analysis of 40 of Schietroma’s papers has now questioned the integrity of 38 of them. The lead author, Paul Myles of the Alfred Hospital in Melbourne, Australia, told the Guardian newspaper that he had been prompted to investigate because the WHO guideline ran counter to long established views among anaesthetists.

Among the coauthors of the review published in Anaesthesia2 is John Carlisle, a consultant anaesthetist at Torbay Hospital in Devon, UK, who has pioneered statistical studies of the baseline data of participants in trials that can show whether they are inconsistent with participants chosen randomly, as they should be. The team also checked the calculations of P values and found that a high proportion had been miscalculated—324 of 825.

“Authors seem not to realise that we can re-run the statistical analyses,” Carlisle told the Guardian. “I was a bit staggered that some of the errors were so obvious.”

If the Schietroma papers are excluded from WHO’s guideline the advice it contains no longer reaches statistical significance. The team concluded, “Pooled data from all trials by authors other than Schietroma et al do not support the use of liberal inspired oxygen to reduce surgical site infection rates,” calling for studies by Schietroma’s group to be critically examined by his employers.

Carlisle said that he had no confidence in any of the papers and that they should be retracted. Schietroma told the Guardian that his papers had been reviewed by an independent statistician and been subject to peer review. Almost all of his data and records of informed consent had been lost in an earthquake in 2009, he said.

The university made no comment despite undertaking a review of Schietroma’s research after being alerted more than a year ago by one of the journals published by the American Medical Association, which found data duplicated from a previously retracted article in the Journal of the American College of Surgeons.

References

↵
Volk T,
1. Peters J,
2. Sessler DI. The WHO recommendation for 80% perioperative oxygen is poorly justified. Anaesthesist2017;66:227-9. doi:10.1007/s00101-017-0286-4. pmid:28246753CrossRef PubMed Google Scholar
↵
1. Myles PB,
2. Carlisle JB,
3. Scarr B. Evidence for compromised data-integrity in studies of liberal peri-operative inspired oxygen. Anaesthesia2019. doi:10.1111/anae.14584.CrossRef Google Scholar