In order to determine the relationship between chronic hypercapnia and anthropomorphic data, pulmonary function tests and slopes of ventilatory responses to hypercapnia (HVCR) and hypoxia (HVR), we studied 55 patients with sleep apnea-hypopna syndrome (SAHS). Patients were divided into hypercapnic, PaCO2 > or = 45 mm Hg (Group I, n = 23, PaO2 = 61 +/- 10 and PaCO2 = 50 +/- 5 mm Hg, and [HCO3-] = 30 +/- 4 mEq/l [means +/- SD]) and normocapnic (or eucapnic), PaCO2 < 45 mm Hg (Group II, n = 32, PaO2 = 76 +/- 10 and PaCO2 = 39 +/- 4 mm Hg and [HCO3-] = 25 +/- 3 mEq/l [means +/- SD]) groups. When compared to the normocapnic group, hypercapnic patients were significantly heavier (with greater body surface area) and had significantly more severe restrictive and obstructive defects and impaired HVR and HCVR. The means (+/- SD) of some of the data follow (* indicates p < 0.05 when Group I is compared to Group II): [table: see text] When subgroups of hypercapnic and eucapnic patients with similar lung functions were compared, the subgroups differed significantly in their weights; conversely, in subgroups with comparable weights, lung function tests differed significantly. These data suggest that the mechanisms of chronic hypercapnia are multifactorial, and we hypothesize that, in the face of repetitive apneas and hypopneas, increased weight and abnormal lung function tests interact and contribute to the generation and maintenance of hypercapnia.