Abstract
Background:
Low oxygen saturation measured by pulse oximetry(SpO2) has been shown to predict mortality in emergency care, and in patients with lung diseases.
Aims and objectives:
We wanted to see whether low SpO2 values are associated with mortality in a general adult population.
Methods:
Pulse oximetry was carried out in 5152 participants in a cross-sectional survey in Tromsø, Norway, in 2001/2002. Follow-up for all-cause mortality until 31.12.2010 were linked to the National Population Register. The SpO2 categories ≤92% and 93-95% were assessed as predictors of mortality in Cox Proportional Hazard regression models correcting for age, sex, smoking and self-reported cardiovascular disease(CVD). Further analysis also included body mass index(BMI), C-reactive protein(CRP) and forced expiratory volume in 1 s(FEV1)% predicted.
Results:
Mean age was 65.8 years, 56% were women, SpO2≤92% and 93-95% were recorded in 53(1.0%) and 537(10.4%) individuals, respectively. During follow-up, 910(17.7%) died. The adjusted Hazard Ratio(HR) was 1.31(p<0.005) for SpO2 93-95% and 2.16(p<0.001) for ≤92% compared to the reference group(>95%). By adding BMI, CRP and FEV1% predicted as covariates in the model HR for 93-95% was 1.14(p=0.17) and for ≤92% 1.56(p<0.05).
Conclusion:
We found that low oxygen saturation is associated with mortality. When including FEV1% predicted in the analysis the strength of the association weakens and is only statistically significant for SpO2≤92%.
- © 2013 ERS
