Nocturnal polysomnography is the standard diagnostic test for sleep apnea syndrome (SAS) but is both expensive and time-consuming. We developed a predictive index for SAS based on pulmonary function data, including respiratory resistance determined by the forced oscillation technique, from 168 obese snorers with suspected SAS. Our model used logistic regression to obtain case-by-case predictions of the probability of SAS, defined as an apnea-hypopnea index (AHI) > or = 15 during overnight polysomnography. We then tested our model in a prospective group of 101 similar patients. Specific respiratory conductance and daytime oxygen saturation contributed significantly to the model. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the index computed from these parameters were 98%, 86%, 90%, and 97%, respectively. In the prospective group, the model proved repeatable, with 100% sensitivity, 84% specificity, 86% PPV, and 100% NPV. The high NPV may help to identify obese snorers with a SAS risk that is so low as to make polysomnography unnecessary. Based on the 50% prevalence of SAS in our study and on the fact that polysomnography is required in all patients with daytime somnolence, we calculated that using our model would have obviated the need for polysomnography in 38% of our patients.