Authors: Usman Ali, M.B.B.S., B.Sc., F.R.C.A. et al
Anesthesiology 8 2016, Vol.125, 419-420.
To the Editor:
We read with great interest the article by Wigmore et al.1 The authors have produced one of the largest studies published looking at cancer survival and anesthesia.
In an era where new cancer therapies are becoming ever more expensive to develop, their primary study finding, increase in mortality associated with inhalational anesthesia compared to intravenous anesthesia with an adjusted hazard ratio of 1.46 (1.29 to 1.69) after propensity score matching and multivariate analysis, is one of huge potential significance.
There are three points arising from the study we would like to discuss.
First, we agree with the authors that long-term survival is a key outcome measure in cancer surveillance. However, we noticed this study analyzed patients over a 3-yr period, with a further 18-month follow-up period, giving a maximum total potential follow-up time of 4.5 yr.
Widely accepted measures of long-term cancer survival are 5 and 10 yr.2,3 We feel that the use of the phrase “long term” in the study title could mislead some readers. Perhaps it would have been more appropriately titled, simply, “survival rates for patients undergoing volatile versus intravenous anesthesia for surgery.”
Second, we would like to highlight issues around the use of propensity scoring analyses and all-cause mortality data.
All-cause mortality is typically used as an outcome measure in prospective trials. Randomization helps account for unknown confounding factors affecting the outcome, thus facilitating the use of all-cause mortality as a primary outcome.
Usefulness of propensity scoring matching in retrospective studies is limited by the fact that remaining unmeasured confounding may still be present.4 This makes the all-cause mortality data presented in this study more difficult to evaluate.
Given this potential for causality combined with the limitations of propensity score matching, we suggest the study could be further refined by looking at cancer-related deaths only as opposed to all-cause mortality.
A possible suggestion would be obtaining mortality data from national cancer registries where cause of death would also be available, as an alternative to the National Health Service demographics service.
Finally, we noticed the Kaplan–Meier survival curves were for unmatched data only. We think that comparing the survival curves for the unmatched groups to those of the groups matched for known variables would add something to the study.
We acknowledge the authors recognized some of the inevitable shortcomings of retrospective studies and support their calls for urgent prospective work to corroborate their findings.
Wigmore, TJ, Mohammed, K, Jhanji, S Long-term survival for patients undergoing volatile versus IV anesthesia for cancer surgery: A retrospective analysis.. Anesthesiology. (2016). 124 69–79
Brenner, H Long-term survival rates of cancer patients achieved by the end of the 20th century: A period analysis.. Lancet. (2002). 360 1131–5
Brenner, H, Hakulinen, T Very-long-term survival rates of patients with cancer.. J Clin Oncol. (2002). 20 4405–9
Winkelmayer, WC, Kurth, T Propensity scores: help or hype?. Nephrol Dial Transplant. (2004). 19 1671–3
Buckley, A, Quaid, MS, Johnson, P, Buggy, D Serum from women undergoing breast cancer surgery, randomized to propofol-paravertebral anaesthetic technique, maintain natural killer cell anti-tumour activity compared with sevoflurane/opioid technique: ESAPC1-5.. Eur J Anaesthesiol. (2014). 31 2
Melamed, R, Bar-Yosef, S, Shakhar, G, Shakhar, K, Ben-Eliyahu, S Suppression of natural killer cell activity and promotion of tumor metastasis by ketamine, thiopental, and halothane, but not by propofol: Mediating mechanisms and prophylactic measures.. Anesth Analg. (2003). 97 1331–9