We observed a lower diurnal Pa02 in patients with OSA than in patients without OSA. Because only a small proportion of OSA patients had an associated bronchial obstruction (6/40, 15%), this resting hypoxemia may be in part explained by high BMI. Indeed, several studies conducted in predominantly obese populations found values of Pa02 similar to those of our patients. Among them, Gold et al also found a higher PaC02 in sleep apnea patients than in control subjects, which was not the case in our study; this discrepancy is likely related to a higher proportion of overlap syndromes in their population, because their patients with OSA had lower FEV1 and FVC than patients without OSA, which we did not find. cialis professional 20 mg
Some studies have even suggested that lower lung volumes and increased airway resistance contribute to the severity of OSA. However, AHI did not correlate to indices of bronchial obstruction (ie, FEV1 and FEV1/FVC) in our patients. Other studies also suggested that the development of hypercapnia in OSA patients requires the presence of an associated bronchial obstruction. However, this was not found in a study of 111 patients in which the only predictive factors of hypercapnia were Pa02 and female sex, although the sex-related difference in PaC02 did not reach significance in a subsequent analysis of this population. In our study, there was no trend toward a higher PaC02 in women than in men, despite a higher BMI.
As in other studies of flow-volume curves and UAO indices, we found that the saw-tooth pattern and the FEF50/FIF50 ratio are not useful for OSA case finding. Conversely, we found that the FEV1/ FEV0 5 ratio, which has been shown to detect UAO when > 1.5, was a predictor of OSA in the logistic regression analysis when > 1.3. However, there was a great overlap between patients with OSA and patients without OSA (Fig 2).