Our study aims to identify the variables associated with a high probability of Obstructive sleep apnea (OSA) to select the best candidates for respiratory polygraphy (RP) compared to SEPAR OSA Consensus? criteria.

We conducted a retrospective observational study of patients who underwent RP at our hospital from January to October 2022. Clinical variables, predictive scales and final diagnosis were collected. Logistic regression analysis was performed using SPSS Statistics, followed by multiple confusion matrices to study various patient selection models.

A total of 120 patients were studied (96.7% high initial suspicion of OSA), 73% were male with a mean age of 54 years. Diagnostic of OSA was positive in 89.2% of cases. Statistically significant variables associated with OSA included age (p=0.03), sex (p=0.026), and observed apneas (p=0.028). Only 21.7% of patients met the SEPAR criteria for intermediate-high probability, being 24 of them diagnosed with OSA. However, 84 out of 94 patients with low probability were also diagnosed with OSA (OR 1.42, P=0.66). After analyzing different patient selection models, the one that showed the highest statistical significance was having a high-risk STOP-Bang or intermediate-risk STOP- Bang being male or with observed apneas (C5 model). Of the 109 patients who met this criteria, 94.4% were ultimately diagnosed, and only 4 patients who did not meet the criteria were diagnosed (OR 30.04, P<0.0001).

We have observed that the criteria proposed to indicate RP by SEPAR have not been useful for reducing our workload. However, C5 model showed a highly significant association with the likelihood of having OSA and could be a predictive model to help select the best candidates for RP.