Table 1 A summary of the randomized controlled trials studying the effects of O₂ therapy in patients with suspected or confirmed myocardial infarctions

Rawles et al. [45] (1976)

Double blind. Inclusion: Suspected MI. Patients randomized to O₂ or air.

IS increased in patients treated with O₂ as measured by AST. No significant differences were shown between the arms in discussing mortality, malignant arrythmias and use of analgesics.

Only those with suspected MI was included,why it is uncertain how many who in fact did had a MI.

The study was conducted pre-PCI era.

IS was measured by AST.

No description of how the randomization sequence was conducted.

Wilson et al. [42] (1997)

Open label. Inclusion: Confirmed MI.

Patients randomized to O₂ or air.

No significant differences were shown between the arms in discussing arrhythmias as well as ST segment changes in the ECG.

The study was conducted pre-PCI era.

IS was measured by AST.

16% of those initially included, fell out and was thus not analyzed in the final analysis cohort.

Ukholkina et al. [46] (2005)

Open label. Inclusion: Confirmed MI. Patients randomized to O₂ or air.

MaR, IS and arrhythmias were significantly lower in the O₂ group.

The randomization process is unclear.

Many have been excluded without any discussion.

IS was measured by CKMB and through ECG mapping.

Ranchord et al. [47] (2012)

Open label. Inclusion: STEMI/LBBB. Patients randomized to O₂ or titrated O₂.

No significant differences between the two arms in discussing IS as measured with cTn and MRI, as well as 30-day mortality.

Data is lacking for a considerable amount of the patients in regard to mortality.

MRI was performed in a subgroup of patients surviving more than 30 days, thus giving rise to a possible selection bias.

Stub et al. [48] (2015)

Open label. Inclusion: STEMI. Patients randomized to O₂ or air.

Patients in the O₂ group had a significantly higher mean peak CK but not cTn, increased IS as measured with MRI, and a higher rate of arrythmias as well as recurrent MI.

CK is not specific for MI.

MRI was conducted in only some patients, thus giving rise to a possible selection bias.

MRI showed increased IS measured in grams of the LV, but not as a percentage of the LV.

Nehme et al. [49] (2016)

Sub study. The main study was conducted by Stub et al. (2015).

For every 100 L of O₂ given to a patient, both cTnI as well as CK, increased with 1.4% and 1.2% respectively.

See limitations for Sub et al. (2015).

A little over 8% of the patients were excluded since they had no cTnI measurements.

Khoshnood et al. [51] (2017)

Single blind. Inclusion: STEMI. Patients randomized to O₂ or air.

No significant differences between the two groups in discussing MSI, MaR and IS.

MRI was conducted in only some patients, thus giving rise to a possible selection bias.

Khoshnood et al. [52] (2017)

Sub study. The main study was conducted by Khoshnood et al. (2017; ref. 44).

No significant differences between the groups in discussing WMSI, LVEF as well as NT-proBNP.

A considerable number of patients were excluded because they, among others, denied participation after that they were initially included. This may be a source of bias.

Khoshnood et al. [53] (2018)

Sub study. The main study was conducted by Khoshnood et al. (2017; ref. 44).

Before the randomization, patients in the O₂ group had a significantly higher VAS and also received significantly more morphine.

No significant differences between the two groups in regard to VAS at the start of the PCI or median VAS decrease from randomization to PCI.

A considerable amount of the patients missed VAS rates and were therefore excluded. This may be a source of bias.

Hoffman et al. [55] (2017)

Open label. Inclusion: Suspected MI. Patients randomized to O₂ or air.

No significant differences between the groups on all-cause mortality at 1 year.

The study may have been underpowered.

Hoffman et al. [56] (2018)

Sub study. The main study was conducted by Hoffman et al. (2017).

No significant differences between the groups in discussing all-cause mortality at 1 year, or adverse cardiac events like MI rehospitalization or cardiogenic chock.

See limitations for Hoffman et al. (2017).

Sparv et al. [57] (2018)

Sub study. The main study was conducted by Hoffman et al. (2017).

No significant differences between the groups in discussing analgesic effect, or the use of both sedatives and opiates during PCI.

Some of the included patients received opiates in the ambulance, why it may have decreased pain at the PCI.