Abstract
Detection of nocturnal hypoxaemia, defined as a mean arterial oxygen saturation below 90%, in normoxic or mildly hypoxic chronic obstructive pulmonary disease (COPD) patients seems clinically relevant, since this feature may precede pulmonary hypertension. Nocturnal studies are expensive and time-consuming procedures. The current study investigates to what extent it is possible to predict nocturnal hypoxaemia from daytime parameters. Forty two COPD patients with a daytime arterial oxygen tension (PaO2) above 8 kPa participated. Nocturnal oxygenation, daytime blood gas values, and ventilatory responses to hypercapnia were measured. In 10 patients, enough desaturations occurred to qualify as nocturnal hypoxaemia. They had a significantly lower daytime PaO2 value, and a lower steady-state hypercapnic ventilatory response. They also smoked more often, and complained about daytime sleepiness. Multiple linear regression analysis demonstrated that daytime PaO2 (32%) was the best independent predictor. Sleepiness (12%), and number of cigarettes smoked (5%) also contributed independently, but in a minor way. Patients with a high daytime PaO2 (> 11 kPa) did not develop nocturnal hypoxaemia. The hypercapnic ventilatory response was used to distinguish nocturnal hypoxaemic from normoxaemic patients. Only patients with a low response (< 3.5 l.min-1.kPa-1) appeared to run a risk of developing nocturnal hypoxaemia. The sensitivity of this test was 80%, and the specificity 70%. It is concluded that daytime PaO2, hypercapnic ventilatory response and sleepiness are helpful in predicting nocturnal hypoxaemia.
