The quality of evidence is downgraded by study limitations (selective outcome reporting), indirectness (two of the studies were not recent), and by imprecise results (few outcome events and wide confidence intervals).
A Cochrane review included 5 studies with a total of 1 173 subjects with confirmed (2 studies) or suspected (3 studies) acute myocardial infarction (AMI). The intervention was inhaled oxygen at 4 to 8 L/min given by mask in 4 studies and by a nasal cannula in one study. The comparator was air in 4 studies and titrated oxygen delivered by nasal prongs or mask adjusting the flow-rate to achieve an oxygen saturation of 93% – 96% in 1 study.
There was no difference in all-cause mortality at hospital discharge in patients with confirmed AMI or in an intention-to-treat analysis including all participants, also those without confirmed AMI (table ). In the meta-analysis, the same number of people died (n = 16) in each group. Only one study reported all-cause mortality at 6 months: 9 participants out 318 died in oxygen group versus 13 out 320 in the air group (RR 0.39, 95% IC 0.14 to 1.07; 1 study, n=628). One study measured pain directly, and 2 others measured it by opiate usage. There was no effect for oxygen on pain relief when pain was directly measured nor when studies measured opiate use as a surrogate for pain. Recurrent ischaemia tended to be higher in the oxygen group compared to the air group but the difference was not statistically significant. There was no clear effect for oxygen on infarct size.
| Outcome | Relative effect (95%) | Assumed risk - Air or titrated oxygen | Correasponding risk - Oxygen (95% CI) | Participants (studies) |
|---|---|---|---|---|
| * including those without confirmed AMI; **statistical heterogeneity I2=80% | ||||
| All-cause mortality (participants with AMI) | RR 1.02 (0.52 to 1.98) | 36 per 1000 | 37 per 1000 (19 to 71) | 871 (4 studies) |
| All-cause mortality (all participants*) | RR 0.99 (0.50 to 1.95) | 28 per 1000 | 28 per 1000 (14 to 55) | 1 123 (4 studies) |
| All-cause mortality (all participants*) in studies done in the revascularisation era | RR 0.58 (0.24 to 1.39) | 27 per 1000 | 16 per 1000 (7 to 38) | 923 (3 studies) |
| Opiate use as a proxy for pain (all participants*) | RR 0.97 (0.78 to 1.20) | 583 per 1000 | 566 per 1000 (455 to 700) | 250 (2 studies) |
| Recurrent myocardial infarction (or ischaemia) | RR 1.67 (0.94 to 2.99)** | 64 per 1000 | 87 per 1000 (50 to 152) | 578 (2 studies) |
The results of this review do not mean that oxygen should be withheld from patients with obvious hypoxia.