Exhaled carbon monoxide levels in obstructive sleep apnoea.

BACKGROUND

Obstructive sleep apnoea (OSA) is characterised by chronic intermittent hypoxia, which enhances airway inflammation and oxidative stress. Exhaled carbon monoxide (eCO), a marker for oxidative stress, has been investigated in OSA. However, previous studies could be biased as they did not differentiate patients with OSA based on smoking history, a known factor influencing eCO levels. The aim of this study to investigate eCO levels in patients with OSA and non-OSA controls and compare evening to morning results.

METHODS

Exhaled carbon monoxide concentration was measured in the evening and in the morning following an in-hospital cardiorespiratory polygraphy in 60 never-smoker OSA patients, 14 ex-smoker OSA patients, 39 current-smoker OSA patients, 10 never-smoker asthmatic patients with OSA, 16 COPD patients with OSA and 20 never-smoker non-OSA controls. OSA was diagnosed based on the apnoea-hypopnoea index (AHI > 5/h).

RESULTS

There was no difference between the never-smoker controls and never-smoker patients with OSA either in the evening (1.98 ± 1.00 ppm versus 1.95 ± 1.28 ppm, p = 0.57, OSA versus controls, respectively) or morning (1.95 ± 0.96 ppm versus 1.80 ± 0.95 ppm, p = 0.42), however there was a weak correlation between eCO and AHI in the evening (r = 0.31, p = 0.01). Accordingly, patients with severe OSA had higher eCO levels in the evening (2.43 ± 1.12 ppm) compared to mild OSA patients (1.57 ± 0.87 ppm, p < 0.01). Ex-smoker (3.07 ± 2.23 ppm), current-smoker (13.13 ± 11.35 ppm), asthmatic (2.70 ± 1.16 ppm) and COPD (18.25 ± 18.60 ppm) patients with OSA had higher levels of eCO compared to the non-smoker OSA group.

CONCLUSION

Exhaled carbon monoxide is elevated only in severe never-smoker OSA suggesting accelerated oxidative stress. Previous smoking history is a major influencing factor which may explain differences between our findings and those of previous studies. Although our results show some impact of OSA on eCO measurements, the bias is small, and it does not significantly affect the clinical utility of eCO to monitor smoking cessation.