Resumen

BORSINI, Eduardo et al. Utilidad de los componentes del cuestionario Stop-Bang para identificar pacientes con apneas del sueño. Rev. am. med. respir. [online]. 2014, vol.14, n.4, pp.382-403. ISSN 1852-236X.

Purpose: The questionnaires used to estimate the probability of suffering from obstructive sleep apnea (OSA) have variable utility. The ability of the STOP-BANG questionnaire has not been evaluated in our high risk population. Aims: The aim of this study was to evaluate the ability of the STOP-BANG assessment tool to predict sleep hourly apnea-hypopnea index (AHI) in patients with high clinical suspicion compared to a self-administered home level III respiratory polygraphy (RP). Methods: We conducted a longitudinal study in patients referred to RP (level III) over fourteen months. The ability of STOP-BANG questionnaire to identify patients with OSA for each severity grade was validated against the results of RP using AHI. The relationships between symptoms (STOP), anthropometrics parameters (BANG) and the combination (STOP-BANG) and AHI (>5 and ≥ 30/hour) were evaluated using multiple logistic regression linear models expressing Odds Ratio (OR) with 95% confidence intervals (CI) for each of the components. For each model, we studied the discrimination power by calculating the area under ROC curve and the fitness using the Hosmer-Lemershow test. Results: 299 patients were studied. 194 were male (64.9%), average age was 52.77 years (SD: 14.67) and body mass index (BMI) was 32.49 (SD: 7.67). 161 cases (53.8%) showed BMI > 30 (obesity). The frequency of identifying AHI >5/hour (area under ROC curve) for each measured component were; STOP: 0.58, BANG: 0.66, and STOP-BANG: 0.66. The best relationship between sensitivity (S) and specificity (Sp) for identifying AHI > 5/h was found by using three STOP components in any possible combination (S: 52.97%; Sp: 60%) with two BANG components (S: 79%; Sp: 43.75%). For an AHI ≥ 30/h the area under ROC curve for each combination were; STOP: 0.67, BANG: 0.67 and STOP-BANG: 0.73. The best relation including S-Sp has been obtained with two STOP components (S: 79%-Sp: 43.75%). Similarly, 3 BANG components reached S of 61% and Sp of 65.48%. Five components of STOP-BANG (in each combination) reached S of 60.73% and Sp of 65.00% (RV+: 1.73 - RV-: 0.60). Finally, we used an automatic selector of variables for the eight STOP-BANG components and we found a model to predict AHI ≥ 30/hour formed by; observed apneas (O): OR: 3.62 (CI 95%: 1.69-7.77); p = 0.001, IMC > 30 (B): OR: 2.51 (CI 95%: 1.19 - 5.28); p = 0.015 and male sex (G): OR: 6.63 (CI 95%: 2.39 -18.3); p = 0.0001 (Area under the curve; 0.75. Goodness of fit). Conclusions: The STOP-BANG questionnaire shows different results for AHI >5 and AHI ≥ 30/hour when RP has been used. The STOP combination shows low capacity to discriminate for AHI > 5/hour and this result differs from the results reported with polisomnography in the sleep laboratory. The anthropometric variables (BANG) show good discriminating capacity evaluated by the area under curve of the model for both cutoff in the analyzed AHI. Five STOP-BANG components in any combination have a high diagnostic sensitivity to identify patients with sleep respiratory disturbance in severe grade. Three anthropometric variables showed good performance as predictors (BMI, age and male sex); the last one was the most important to identify pathologic AHI (> 5/hour) or severe high AHI (≥30/h). In our population the prediction model O-G-B had the best performance.

Palabras clave : STOP BANG; Respiratory Polygraphy; Predictors; SAHOS.

        · resumen en Español     · texto en Español     · Español ( pdf )