Background

The wide 95% confidence interval for S(a)O2 measured by pulse oximetry (S(P)O2) and the inherent characteristics of the oxyhaemoglobin dissociation curve can lead to modest but significant decreases in P(a)O2 (deltaP(a)O2 > or = 5 mmHg) that may be under-appreciated.

Aim

To avoid missing potentially significant deltaP(a)O2 by using S(P)O2, this study establishes a threshold of deltaS(P)O2 to detect deltaP(a)O2 by examining the correlation between deltaS(P)O2 and deltaP(a)O2.

Methods

We enrolled 29 elderly patients with moderate to severe chronic obstructive pulmonary disease as assessed by lung function testing. Arterial blood gases and S(P)O2 measurements were carried out during maximal exercise testing. The patients were assigned to groups based on P(a)O2 measurements: group 1 had P(a)O2 at peak exercise (P(a)O2peak) > or = 60 mmHg without a deltaP(a)O2; group 2 had P(a)O2peak > or = 60 mmHg with a deltaP(a)O2; group 3 had P(a)O2peak < 60 mmHg without a deltaP(a)O2; and group 4 had P(a)O2peak < 60 mmHg with a deltaP(a)O2.

Results

The study population was evenly distributed between groups 1, 2 and 4. However, group 3 did not have any patients enrolled in this study that met group 3 criteria. The sensitivity of pulse oximetry required to detect S(a)O2 below 90% was 19%. DeltaS(P)O2 of 3% may increase the low sensitivity of S(P)O2 and was shown by a 92% positive predictive value for deltaP(a)O2 > or = 5 mmHg.

Conclusion

This study suggests that important changes in oxygenation may be avoided if using deltaS(P)O2 rather than absolute values of S(P)O2 in patients with chronic obstructive pulmonary disease undergoing exercise testing to detect exercise-induced hypoxaemia.