Aims: Recreational nitrous oxide (N 2O) misuse is a growing problem. It can cause functional hydroxycobalamin (B 12) deficiency and severe neurological complications. Despite increasing harms, the dose-response relationship and clinical value of biomarkers remain unclear. This study aimed to characterise the demographics, patterns of use, outcomes, and biomarker utility in patients hospitalised with N 2O toxicity, and to examine whether a dose-response relationship exists between exposure and toxicity. Methods: This was a retrospective cohort study of hospitalised individuals with N 2O toxicity between 1 January 2020 and 31 July 2025 across six Sydney hospitals. Weekly and cumulative N 2O exposure volumes were estimated. Exposure volumes were compared with clinical outcomes and biomarker results to assess correlations and biomarker utility. Results: Eighty-one individuals across 100 presentations were included. Higher cumulative volumes of N 2O exposure were associated with worse neurological impairment, including subacute combined degeneration of the spinal cord (SCD)( p=0.006) and peripheral neuropathy (PN)( p=0.036), and showed moderate correlation with greater neuropathy severity (Spearman’s ⍴=0.43, p=0.001). Median homocysteine was 50 μmol/L(IQR: 32-114; normal 5-15) with a median half-life of 14.59 hours. Median methylmalonic acid was 0.68 μmol/L (IQR: 0.18-3.16; normal <0.27). Regarding diagnosis of SCD, B 12 and holotranscobalamin had high specificity (87% and 98%) but low sensitivity (22% and 5%); while homocysteine and methylmalonic acid were more sensitive (95% and 93%) but less specific (13% and 45%). Conclusions: Cumulative N 2O exposure is the strongest predictor of severe neurological outcomes in a dose-dependent manner. Homocysteine and methylmalonic acid are good screening tests for neurotoxicity but poor confirmatory tests.