While point measurement of resting arterial partial pressure of oxygen (P(a)O(2)) is the traditional gold-standard for assessment of oxygenation in chronic obstructive pulmonary disease (COPD), 24-h oximetry may identify further patients with clinically significant hypoxaemia. We aimed to describe the relationship between these two parameters and identify other correlated variables.

All patients registered with the Barwon Health Hospital Admission Risk Program from 1 March to 31 October 2008 for the diagnosis of COPD were identified. The main inclusion criteria were obstructive spirometry, clinical stability and moderate resting hypoxaemia (P(a)O(2) 56-70 mmHg). All patients underwent 24-h oximetry, arterial blood gas, spirometry, anthropometry and telephone questionnaire, and 23 patients also completed polysomnography.

Inclusion criteria were met in 35 of 287 patients. Mean recording time was 23.5 h, representing 97% of intended oximetry time. Nineteen patients (54%) spent greater than 30% of recorded oximetry time below 90%. There was a moderate inverse correlation between time below 90% saturations and P(a)O(2) (r=-0.40, P= 0.02), with body mass index (BMI) the only other independent predictor of the primary outcome identified (r= 0.39, P= 0.02). Correlations were similar for waking hours considered separately. However, for sleeping oximetry, BMI and age were the only independent predictors of time below 90%. Polysomnography demonstrated a high prevalence of rapid eye movement-related hypoventilation and obstructive sleep apnoea syndrome.

Many patients with moderate hypoxaemia on resting P(a)O(2) desaturate significantly on ambulatory oximetry. The correlation between P(a)O(2) and proportion of saturations below 90% is moderate and similar to BMI, but this pattern does not hold during sleeping hours.