There are no dragons: Low-flow anaesthesia with sevoflurane is safe

In this issue of Anaesthesia and Intensive Care, McGain et al. ¹ report the results of a survey of inhalational anaesthesia practice in Australia and New Zealand. The authors found that sevoflurane was preferred by 72% of respondents, propofol by 16%, desflurane by 12% and isoflurane by 1%. Based primarily on the high cost, low potency and profound greenhouse gas effect of desflurane, the authors advocated for switching from desflurane to sevoflurane to minimise the cost and environmental impact of inhalation anaesthesia. Although fresh gas flows (FGFs) of under 1 L/min were acceptable for 80% of desflurane users, only 50% would use sevoflurane at these FGFs. While reducing desflurane use would lead to a significant reduction in the cost and environmental impact of inhalational agents, we believe that additional worthwhile gains can also be made by reducing FGFs with sevoflurane.

The primary reason identified in the study for not using FGF <1 L/min with sevoflurane was the recommendation against low flow rates on the drug data sheets or package inserts because of concern for compound A production. ² It appears that low flow practices per se are not seen as a risk because most desflurane users did use low flows, and 59% of those sampled were using machines that automate low-flow vapour delivery. We surmise the rationale for avoiding low flows with sevoflurane by 50% of the respondents is concern for compound A production. We contend that concerns about compound A production during sevoflurane anaesthesia are not founded in fact and should be abandoned, both in data sheets and in the minds of anaesthetists.

Compound A (fluoromethyl-2,2-difluoro-1-(trifluoromethyl)vinyl ether) is produced by the interaction between sevoflurane and Ca(OH)₂-based carbon dioxide (CO₂) absorbents in the presence of KOH and (less so) NaOH (“soda lime”). Compound A has been shown to be nephrotoxic in rats, but not in humans. Renal damage markers did not differ after three or nine hours of sevoflurane or isoflurane using low-flow anaesthesia.^3,4 In a 2003 letter, Kharasch et al. pointed out that 120 million sevoflurane anaesthetics had been administered without evidence of human nephrotoxicity whereas both halothane hepatitis and methoxyflurane nephrotoxicity were recognised within a few years of the introduction of these agents. ⁵ If harm is truly absent, no body of science or any number of studies will be able to prove there is no harm. Quoting the late Ted Eger on halothane hepatitis: ‘One cannot disprove the existence of dragons’. ⁶ Why is it so hard for clinicians to ‘let go’ of practices that are not supported by facts?

The data sheets required by regulatory agencies have a big impact on prescribing practice despite inconsistent guidance. According to the Australian Therapeutic Goods Administration, “The concentrations of Compound A, measured in the anaesthesia circuit when sevoflurane is used clinically are not known to be deleterious to humans …. a level of Compound A exposure at which clinical nephrotoxicity might be expected to occur has not been established, it is prudent to consider all of the factors leading to Compound A exposure in humans, … Because of limited clinical experience with sevoflurane in low – flow systems, fresh gas flow rates below 2 L/min in a circle absorber system are not recommended”. ²

The provenance of information in data sheets is worth considering. These data sheets are a mix of information submitted by the drug companies and required by regulators, with the process varying by jurisdiction. In New Zealand the original supplier of sevoflurane, Abbott (now AbbVie), removed any mention of minimum flow rates by 1996. ⁷ Two other companies supply sevoflurane in the New Zealand market – one makes no mention of flow rates, the other recommends against low flow rates. In Australia, the AbbVie data sheet uses bold letters to warn about compound A and low flows. Unfortunately, the McGain and colleagues survey does not tell us whether there are differences in flow rate practice between Australia and New Zealand. Changes to data sheets typically need to be initiated by the company. In the mid-1990s, having flow rate restrictions removed was in Abbott’s interest because it made use of their new, more expensive agent more acceptable. Nowadays, in a highly competitive market with a drug that is off patent there is little if any incentive to initiate changes that will reduce utilisation of the drug.

If there is lingering, albeit unjustified, concern over compound A toxicity in humans, the data on modern absorbent formulations should put that concern to rest. When desflurane and sevoflurane were introduced into practice, the potential for interaction with CO₂ absorbents to produce carbon monoxide (CO) and compound A respectively was recognised. CO₂ absorbents in common use at the time contained significant percentages of the strong bases KOH and NaOH. Stabernak et al. studied commercially available absorbents with varying concentrations of KOH and NaOH and found that the quantity of compound A produced was directly related to the concentration of strong base, with minimal to no compound A produced if the strong base was limited in concentration or eliminated entirely. ⁸ Kobayashi and colleagues compared Draegersorb Free (NaOH <2%) to Amsorb (Ca(OH)₂ alone) and found no difference in the amount of compound A produced. ⁹ Keijzer et al. also showed that production of compound A levels in the circuit is reduced by using CO₂ absorbents with reduced amounts of strong bases or eliminated by those without NaOH. Interestingly, they found that desiccation reduced the compound A produced when NaOH was present whereas it increased with Ca(OH)₂ alone. ¹⁰ Most of these products were not available at the time flow rate guidelines were established, but even today the sales of non–compound A-producing absorbents remain only a fraction of total sales (personal communication), suggesting that concern about compound A is considered clinically irrelevant by many.

Reducing FGFs to avoid wasting inhaled anaesthetics has both economic and environmental advantages. Sevoflurane costs decrease more rapidly than absorbent costs increase and total cost and environmental impact will still be minimised as flows approach a closed circuit condition. ¹¹

Volatile anaesthetic agents are unique in that it is possible to significantly reduce the mass of drug delivered to the circuit (by lowering FGFs) without altering the amount that actually enters the patient. While the cost (dollar and environmental) of these agents may be low for any one individual, when multiplied over the several million anaesthetics delivered each year in Australia and New Zealand, the potential savings from reducing FGF are considerable. Recent data suggest that a 1 L/min change in FGF with sevoflurane results in a change of between 4¹² and 18 mL/hour ¹³ in consumption of liquid agent. McGain et al. ¹ suggest the most common anaesthetic in Australia and New Zealand is sevoflurane, with a 1–2 L/min FGF. Even though some aspects such as the added cost of absorbents and the ecological footprint of CO₂ absorbent utilization require further study, a FGF reduction by 500 mL/min is likely to save millions of dollars and help to reduce the environmental footprint of anaesthesia.

We believe there is no rationale for avoiding FGFs <1 L/min when using sevoflurane. The data from the McGain study on desflurane indicate that practitioners in Australia and New Zealand are comfortable using those flows and much of the equipment is designed to automate low-flow anaesthetic delivery. Compound A has never been found to be toxic to humans, and clinicians with lingering concerns have access to absorbent formulations proven to minimise or eliminate compound A production. Indeed, hundreds of millions of sevoflurane anaesthetics have been administered, many at low flow rates, with no evidence of poorer outcome. It is time to abandon concerns about low flow rates and sevoflurane, and to request manufacturers and regulators to remove any restrictions from their drug data sheets. We owe it to our profession to base our practice on data rather than beliefs. We owe it to society to be good stewards of the healthcare dollar and work to minimise our carbon footprint.