Construcción y validación de una escala que determine la presencia de síndrome de apnea obstructiva del sueño en pacientes con sospecha clínica basada en medidas antropométricas y síntomas clínicos.

Abstract

Introducción: El síndrome de apnea obstructiva del sueño se cataloga como un problema de salud pública, debido a su alta prevalencia la cual se sitúa entre un 2% a 26 %, en la población en general , con el agravante que es sub -diagnosticada al no existir métodos de tamizaje suficientemente válidos y fiables, es así como los métodos existentes se han basado en cuestionarios o modelos clínicos que cuantifican la presencia de síntomas clínicos, dentro de los más difundidos esta la escala de Epworth que valora la somnolencia diurna, el cuestionario de Berlín que identifica los paciente con sospecha de SAOS en atención primaria y la escala de NAMES producto de la construcción de un modelo a partir de los síntomas clínicos y medidas antropométricas.

Introduction: Obstructive sleep apnea had classified as a public health problem due to its high prevalence, which is between 2% to 26% of the general population 1. The other way, the absence methods sufficiently valid and reliable screening produce sub - diagnostic. There is methods that based on questionnaires or clinical models that quantify the presence of clinical symptoms. The most widespread is the Epworth scale that assesses daytime sleepiness, the Berlin questionnaire identifies the patient with suspected OSA in primary care and the scale of product NAMES building a model based, clinical symptoms and anthropometric measurements. Objective: To develop a scale in the prediction of OSA, it based on clinical and anthropometric characteristics in a population of patients with clinical suspicion of OSA. Methods: A cross-sectional observational study was conducted adult subjects referred for clinical suspicion of OSA who underwent overnight polysomnographic study, diagnosis of OSA is made from hiccupped apnea index (AHI) was included> 5 events / hour. Variables that were associated with the presence of OSA, and who had a biological plausibility were included in the building of the model. Yielding six models by the method of Stepwise Forward, the most parsimonious model included the variables age, sex, BMI, neck circumference and snoring, the scale scores were assigned considering the OR values obtained from each of the variables, this was validated from the point of view of discrimination and calibration in a new population. Results: Among 2553 subjects with OSAS prevalence of 91.4% was found while in 1386 the population of subjects where the scale was validated, the prevalence was 96.2%, the variables associated with the presentation of OSAS were: neck circumference, BMI, presence or absence of snoring, age and sex, no statistically significant association with the Epworth scale found. The scale showed adequate discrimination ability with area under the ROC curve of 0.76 (95% CI 0.67 to 0.83) and a proper calibration for a given value of �ℎ�! = 27.2 p< 0.001 in the Hosmer-Lemeshow test. Conclusions: The present scale showed adequate discriminatory and calibration capacity of predict the occurrence of OSA in this population. However due to the high prevalence of OSAS, the predictive power is high even with scores of 0, being within a range probabilities between 0.89 to 0.99, it is important to validate this scale in a population whose prevalence of OSA is less in order to evaluate the discrimination ability.