Bottom line up front: A recent large, high quality meta-analysis reported a significant mortality effect with the use of liberal, rather than conservative, oxygen supplementation. The number needed harm for 30-day mortality was 1 in 126. A modest but relevant NNH given how common oxygen supplementation is in the ER. Consider titrating all forms of oxygen supplementation – whether it be NC, NPPV or invasive ventilation – to a SpO2 of 94-96%.
Background: Oxygen supplementation is ubiquitous in the practice of modern medicine. A recent large meta-analysis published in Lancet, the Improving Oxygen Therapy in Acute-illness (IOTA) review, called into question the use of liberal oxygen supplementation.
Methods: The study combined the findings of 25 RCTs comparing liberal v conservative O2 supplementation across a wide spectrum of illnesses (ACS, stroke, sepsis, critical illness and emergency surgery). Patients were excluded if they had chronic pulmonary disease, required hyperbaric therapy or ECMO. Primary study endpoints were mortality measured at various time points. Secondary endpoints included length of hospital stay and rates of hospital acquired infection.
Results: The 25 RCTs included a total of 16,037 patients. Median oxygen supplementation levels were fraction of inspired oxygen (FiO2) 0.52 vs. 0.21 (liberal vs. conservative). Authors found that liberal oxygen supplementation increased in-hospital mortality with a RR 1.21 (95% CI 1.03-1.43 ; NNH 1 in 137) and 30-day mortality with a RR 1.14 (95% CI 1.01-1.28 ; NNH 1 in 126). There were no significant differences in any of the secondary endpoints.
- Solely reliant on data from RCTs.
- Large number of patients over a wide spectrum of disease.
- Exclusion of pts with chronic pulmonary disease. We can’t blame these findings on inadvertent blunting of the hypoxic respiratory drive that we observe in COPDers.
- There are few study endpoints as meaningful as mortality.
- The authors demonstrated a dose-response relationship between SpO2 and increased mortality above SpO2 of 96%.
- Findings were robust to multiple sensitivity analyses.
- Authors fail to elicit or hypothesize the mechanism by which hyperoxemia causes harm. The authors allude to animal studies suggesting that oxidative stress increases risk for acute lung injury and excess oxygen promotes harmful vasoconstriction. However, it is equally plausible that liberal oxygen supplementation simply leads to delayed recognition of clinical decompensation.
- Authors provided limited commentary on subgroup analyses. The authors demonstrate some benefit to liberal oxygen therapy in reducing infection in the peri-operative setting, which is consistent with findings in prior papers. Unfortunately, the authors gloss over these findings in an effort to push forth their hypothesis. Furthermore, the authors provide limited commentary on the potential benefits of hyperoxemia in salvaging the penumbra for stroke patients.
- Authors do not differentiate between various types of O2 supplementation (eg, NC v NPPV v invasive ventilation).
Conclusion: Conservative O2 therapy > liberal O2 therapy. Titrate supplemental O2 to SpO2 94-96%. This is an active area of research and more papers are likely to come.
Chu DK, Kim L, et al. Mortality and morbidity in acutely ill adults treated with liberal versus conservative oxygen therapy (IOTA): a systematic review and meta-analysis. Lancet. Vol 391, Issue 10131, p 1693-1705, Apr 28, 2018.