RESUMEN
BACKGROUND: In patients with acromegaly, sleep apnea-related hypoxemia results in considerable morbidity and mortality. AIMS: To evaluate the relative weight of pathogenic factors in predicting such hypoxemia. METHODS: In this cross-sectional study, 34 acromegaly patients were submitted to clinical evaluation, nocturnal oximetry, and nasolaryngeal airway tomography. GH, IGF-I, and its upper limit normal value were measured. Nocturnal hypoxemia was defined as >5 episodes of desaturation/h of sleep. Craniofacial abnormalities were expressed using a linear parameter index (LPI). Nocturnal hypoxemia was predicted using logistic regression, including the variables markers of craniofacial abnormality, hormonal alteration, and obesity. Coefficients were standardized in order to determine their effect magnitudes relative to the outcome. The best model included the variables gender, age, LPI, body mass index (BMI), and IGFI upper limit normal value. MAIN RESULTS: In the absence of the age and gender variables, the odds ratio for the LPI (1.60) was slightly higher than those found for BMI (1.49) and upper limit normal value (1.40). When the data were adjusted for age, the hormone upper limit normal value presented little alteration (1.49), although the decrease in the LPI was considerable (1.21), as was the increase in the BMI (2.18). The relative weight of the LPI was age-dependent. The gender variable did not alter the relevance of the others. CONCLUSIONS: The effects that craniofacial aspect, obesity, and hormonal alterations have on nocturnal hypoxemia are of similar magnitude.