System-Level Reduction in Anesthetic Greenhouse Gas Emissions: Central Nitrous Oxide Deactivation, Clinical Decision Support, and Secular Change in a Multicenter Interrupted Time-Series Study

BACKGROUND: Volatile anesthetic agents and nitrous oxide (N2O) are greenhouse gases and a modifiable source of health care-related carbon dioxide equivalent (CO2e) emissions; losses from central N2O pipeline systems are a major contributor. We evaluated anesthetic gas use and assessed system-level emission-reduction strategies across a large public health system. METHODS: We conducted a multicenter observational interrupted time-series study of case-level data aggregated to monthly group-level observations from 11 NYC Health + Hospitals facilities between April 1, 2022, and December 31, 2025. After exclusions, 148,762 of 158,881 screened general anesthetic cases (93.6%) were analyzed. Fresh gas flow (FGF) rate and time were recorded, and hourly CO2e emission rate from N2O and volatile agents was calculated. Two interventions were evaluated: a clinical decision support Advisory prompting FGF reduction and deactivation of central N2O pipelines at three hospitals. RESULTS: During Phase 1 (pre-intervention), median FGF was 3.90 L·min-,1 N2O was used in 27.2% of cases, and sevoflurane in 97.8%. N2O generated an annualized CO2e impact equivalent to 77% of all volatile agents combined. System-level hourly CO2e emission rate declined over the study period by 0.143 kg CO2e·h-1 per month (95% confidence interval [CI], 0.11-0.18; P < .001). The system-wide Initial Advisory was associated with step reductions in FGF (0.163 L·min-1; 95% CI, 0.11-0.21; P < .001) and in total hourly CO2e emission rate (0.78 kg CO2e·h-1; 95% CI, 0.24-1.32; P = .006). Central N2O pipeline deactivation was the dominant structural effect: complete deactivation was associated with a 1.67 kg CO2e·h-1 reduction in N2O-attributable hourly emission rate at the group level (95% CI, 1.23-2.12; P < .001). At Bellevue, 94.6% of centrally supplied N2O was lost before reaching patients, generating approximately 2136 metric tons of CO2e annually, 2.58 times the combined annual emissions from all clinically administered anesthetic gases across the 11-hospital system. CONCLUSIONS: Central N2O pipeline deactivation was associated with the largest reduction in the N2O-attributable hourly CO2e emission rate; the association with total emissions in the primary model was not statistically significant (P = .069). Concurrent secular declines in N2O and desflurane use contributed additionally, whereas the clinical decision support Advisory was associated with a modest but statistically significant reduction in FGF rate and total hourly CO2e emission rate. These findings support the 2024 American Society of Anesthesiologists recommendation to deactivate centrally piped N2O to reduce health care emissions.