A health-based recommended occupational exposure limit for nitrous oxide using experimental animal data based on a systematic review and dose-response analysis.

Background: Nitrous oxide (NO) is a common inhalation anaesthetic used in medical, paramedical, and veterinary practice. Since the mid 1950's, concerns have been raised regarding occupational exposure to NO, leading to many epidemiological and experimental animal studies. Previous evaluations resulted in the classification of NO as a possible risk factor for adverse reproductive health outcomes based on animal data. Human data were deemed inadequate primarily because of simultaneous co-exposures to other risk factors for adverse reproductive and developmental outcomes, including other anaesthetic gases. Since previous evaluations, controversies regarding NO use remained and new approaches for dose response modelling have been adopted, calling for an update and re-evaluation of the body of evidence. This review aims to assess available animal evidence on NO reproductive and developmental outcomes to inform a health-based recommended occupational exposure limit (OEL) for NO with a benchmark dose-response modelling (BMD) approach.

Methods: Comprehensive searches in PubMed, EMBASE, and Web of Science were performed to retrieve all relevant studies addressing reproductive and developmental outcomes related to inhalation of NO in animals. The articles retrieved were screened based on title-abstract and full text by two independent reviewers. After data extraction, an overview of all studies was created for the different endpoints, namely foetal outcomes (e.g., resorption), female outcomes (e.g. implantations), and male outcomes (e.g. sperm count). A subset of studies reporting on exposure relevant to workplace settings and with a sufficient number of tested doses were included in dose-response modelling using the BMD approach.

Results: In total, 15.816 articles were retrieved, of which 47 articles were finally included while 4 of those were used for the quantitative data synthesis. The overall risk of bias was judged to be probably high (using OHAT risk of bias tool) and unclear (using SYRCLE's risk of bias tool). From eligible rat studies, three studies provided an acceptable result by fitting a Hill model to the dose-response data. The resulting benchmark dose lower bounds (BMDLs) from three studies converged to an average (±sd) exposure level of 925 ± 2 mg/m at an additional risk of one standard deviation of implantation losses above those observed in the control group (i.e. reduced number of live foetuses/mother). For extrapolation from rats to humans, an uncertainty factor of 10 was used and an additional factor of 5 was applied to account for interindividual variability within the population of workers.

Conclusion: With this systematic review, all available evidence for reproductive toxicity and adverse developmental outcomes in animals resulting from inhalation exposure to NO was used to derive a health-based OEL recommendation of 20 mg/m as 8-h time-weighted average.