To compare three different oxygen therapy methods in primary headaches.

Design: A randomized placebo-controlled clinical trial was conducted between January 2016 and October 2017.

Setting: The emergency department of a university-affiliated urban hospital in Tehran, Iran.

Participants: Adult patients (aged 18 years and above) with moderate and severe primary headaches (VAS score of 4 or more).

Interventions: Participants were allocated to one of four groups. Group A (n = 34) received 30 mg of intravenous ketorolac plus oxygen at 15 l/min (min) through a non-rebreather mask (NRB), group B (n = 34) received 30 mg of intravenous ketorolac plus 7 l/min of oxygen through a 60% venturi mask, group C (n = 34) received 30 mg of intravenous ketorolac plus 4 l/min of oxygen through a nasal cannula and group D (n = 34) received 30 mg of intravenous ketorolac and room air.

Main outcomes measured: Pain was assessed using the visual analog scale (VAS) at 0, 15, 30 and 60 min after admission.

Altogether, 136 patients were included. The most significant VAS change occurred in the NRB group at 30 min (p-value = 0.001). At this point, pain reduction in the NRB group was clinically higher than for the venturi and nasal cannula groups, but this effect had disappeared at 60 min.

Although the non-rebreather mask was significantly more effective at 30 min, after 60 min, none of the groups met the endpoint criterion of a 1.3-cm difference on the VAS scale.