This study describes the characteristics of women who used self-administered nitrous oxide during labor compared with nonusers, to show patterns of use of nitrous oxide as well as other pain medications and to identify differences in maternal and neonatal outcomes of women who used nitrous oxide compared with those who did not. Retrospective electronic health record review. This Institutional Review Board–approved study was conducted in an upper Midwestern, urban, magnet-designated, 75-bed quaternary perinatal center with greater than 5,000 births annually. The sample included two groups: 400 nitrous oxide users during a 26-month period (January 2015–March 2017) and a comparison group of 6,733 women who met nitrous oxide eligibility but who did not use it, during the same period. Women in the nitrous oxide group and the comparison group did not differ significantly with regard to age, race, ethnicity, preferred language, marital status, gestational diabetes, body mass index (on admission), or provider type at birth. Women in the nitrous oxide group were slightly more likely to have private insurance relative to the comparison population (81% vs. 75%, p < .05). Nitrous oxide users were also more likely to be nulliparous (61%) than women in the comparison group (43%, p < .001). Among women in the nitrous oxide group, most used intermittent administration, and 17% experienced side effects, the most common being dizziness and nausea. Half (48%) of women initiated nitrous oxide at 0–4-cm dilation. In the nitrous oxide–only group, there was a 50.5% conversion to epidural. There were no significant differences in instrumentation, lacerations, 5-minute Apgar scores < 7, special care nursery or NICU transfers, or shoulder dystocia between the two groups. Adjusted analysis also found no difference in first or second stage duration of labor or duration of stay from birth to discharge. Parity was associated with nitrous oxide use in labor in this study, with more nulliparous women choosing this option. Outcomes did not differ generally by user group after adjustment for covariates, and there were no safety concerns.