Sensitivity and specificity of four screening sleep-disordered breathing tests in patients with and without cardiovascular disease

Abstract

Objectives: Polysomnogram is the gold standard for the diagnosis of sleep-disordered breathing (SDB); a sensitive and specific alternative strategy would be ideal, due to its low availability, and screening patients at high risk of OSA is very important. This study aimed to determine the operating characteristics of screening tests in patients with and without cardiovascular disease (CVD). Material and Methods: Epworth sleepiness scale (ESS), Berlin, STOP-bang and Pittsburgh sleep quality index (PSQI) were applied in adults with and without cardiovascular disease in three Colombian cities, as well as anthropometric measurements and a polysomnogram. Operating characteristics were calculated for each test and the best cut-off values in patients with and without CVD were obtained. Results: 964 patients (median age: 58), 662 with and 302 without CVD were included. The prevalence for SDB (AHI ≥5) were 43.4 % (OSA), 16.2% (central apnea), and 12.4 % (other). In patients without CVD, the highest sensitivity for OSA and central apnea was for PSQI (80-85%). The highest specificity was for STOP-bang (68%) and Berlin (78.6%). In CVD the best sensitivity was for PSQI (81.9%) followed by Berlin (71.9%) and the best specificity for STOP-bang (82.1%). No isolated questionnaire showed good diagnostic performance (AUC≤0.6) and the cut-off values had no variations except for ESS. Conclusion: Screening tests showed low operating characteristics for the diagnosis to SDB, but better performance in patients with CVD. They are not recommended as the only diagnostic test, but they can be useful to guide the initial diagnostic process.

Insomnia as a predictor of recurrent cardiovascular events in patients with coronary heart disease

“…We assessed OSA risk with the Berlin Questionnaire. The screening properties of the Berlin Questionnaire in a CHD population are not known and even though its properties for the accuracy of identifying OSA have been questioned [ 31 , 32 ], it has still been regarded as a valuable tool in epidemiological studies [ 33 ]. However, OSA was not assessed by polysomnography and we cannot exclude OSA as a potential confounding factor for the association between clinical insomnia and MACE.…”

The value of nurse-led anthropometric and oropharyngeal measurements combined with STOP-Bang questionnaire in screening for obstructive sleep apnea in patients with acute coronary syndrome: a prospective cohort study

“…Current screening strategies, such as the ESS and the SBQ, rely primarily on self-reported patient symptoms and provide mixed results. The sleep questionnaire alone has low diagnostic accuracy and is not recommended as the only diagnostic test [ 14 , 18 , 19 ]. The ESS is used frequently to measure excessive daytime sleepiness in research and clinical settings.…”

Machine Learning Identification of Obstructive Sleep Apnea Severity through the Patient Clinical Features: A Retrospective Study

“…The strong connection between the severity of obstructive apnea identified through standard diagnostic tests, such as nocturnal polysomnography or pulse oximetry, and the associated cardiovascular or neurodegenerative risks are established in the literature [ 3 , 4 ]. Conversely, the different clinical features and scoring systems, such as age, BMI, the anatomical scores of palatal collapses, or validated questionnaires, although significantly correlating with pathology, do not possess adequate sensitivity or specificity to substitute instrumental diagnostics tools [ 5 ].…”

“…Artificial intelligence (AI) is rapidly gaining importance in medicine, opening promising new perspectives [ 6 , 7 , 8 ]. Through a deep learning system, different types of data can be exploited, such as clinical scores, questionnaires, or diagnostics [ 2 , 4 , 5 ].…”