Anaesthetic gas emissions in a tertiary hospital in Pakistan: Behavioural drivers versus technological solutions

Background Anaesthetic gases contribute disproportionately to healthcare-related greenhouse gas emissions. End-tidal control (ETC), originally developed to reduce cognitive load and improve dosing accuracy, has been promoted as a potential strategy to lower emissions. However, its benefit must be considered alongside established drivers such as agent choice and flow. This study examined predictors of emissions, focusing on clinician behaviour versus technological modifications such as ETC. Methods We conducted a six-month prospective audit at Aga Khan University Hospital, Karachi. Data were extracted from Carestation™ Insights. The outcome was emissions (kgCO₂e/min), with predictors including ETC use, fresh gas flow, case duration, and flow-normalised consumption of isoflurane, sevoflurane, and nitrous oxide. Multivariable regression models were run for all cases and stratified by agent. Results A total of 7231 cases were analysed. In the combined regression, nitrous oxide usage was the strongest predictor of emissions (β=1.56, p < 0.001), followed by isoflurane (β=2.76, p = 0.040), sevoflurane (β=1.54, p = 0.009), and mean fresh gas flow (β=0.37, p < 0.001). ETC and case duration were not significant. In agent-specific models, gas usage and flow rate were the dominant predictors. ETC had a small but significant protective effect in isoflurane models, the opposite effect in sevoflurane models, and no effect in nitrous oxide models. Conclusions Anaesthetic gas emissions are driven mainly by clinician behaviours, agent choice, concentration, and flow, rather than by end-tidal control. Nitrous oxide is the largest contributor and should be targeted for elimination. Low-cost behavioural interventions offer the most effective and scalable pathway for sustainable anaesthesia in resource-limited settings.