Diagnosis of Obstructive Sleep Apnea Using Logistic Regression and Artificial Neural Networks Models

Abstract Background Understanding the causes and risk factors of metabolic syndrome is important for promoting population health. Oxidative stress has been associated with metabolic syndrome, and also obstructive sleep apnea. These are two diseases which have common prognostic characteristics for heart disease. The aim of this study was to examine the role of oxidative stress in the concurrent presence of metabolic syndrome and obstructive sleep apnea in a working population. Methods Participants were 163 artisan bakers in Shahroud, Iran, routinely exposed to significant heat stress and other oxidative stress indicators on a daily basis as part of their work. Using a cross-sectional design, data relevant to determining metabolic syndrome status according to International Diabetes Federation criteria, and the presence of obstructive sleep apnea according to the STOP-Bang score, was collected. Analyses included hierarchical binary logistic regression to yield predictors of the two diseases. Results Hierarchical binary logistic regression showed that oxidative stress – alongside obesity, no regular exercise, and smoking – was an independent predictor of metabolic syndrome, but not obstructive sleep apnea. Participants who were obese were 28 times more likely to have metabolic syndrome (OR 28.59, 95% CI 4.91–63.02) and 44 times more likely to have obstructive sleep apnea (OR 44.48, 95% CI 4.91–403.28). Participants meeting metabolic syndrome criteria had significantly higher levels of malondialdehyde (p < 0.05) than those who did not. No difference in oxidative stress index levels were found according to obstructive sleep apnea status. Conclusions Our findings suggest that oxidative stress contributes to the onset of metabolic syndrome, and that obstructive sleep apnea is involved in oxidative stress. Whilst obesity, exercise, and smoking remain important targets for reducing the incidence of metabolic syndrome and obstructive sleep apnea, policies to control risks of prolonged exposure to oxidative stress are also relevant in occupations where such environmental conditions exist.

Download Full-text

414 Deep Neural Networks: A Survey Tool for Obstructive Sleep Apnea Prediction

SLEEP ◽

10.1093/sleep/zsab072.413 ◽

2021 ◽

Vol 44 (Supplement_2) ◽

pp. A164-A164

Author(s):

Pahnwat Taweesedt ◽

JungYoon Kim ◽

Jaehyun Park ◽

Jangwoon Park ◽

Munish Sharma ◽

...

Keyword(s):

Machine Learning ◽

Neural Networks ◽

Obstructive Sleep Apnea ◽

Sleep Apnea ◽

Deep Neural Networks ◽

Support Vector ◽

Learning Models ◽

Obstructive Sleep ◽

Screening Questionnaires ◽

Machine Learning Models

Abstract Introduction Obstructive sleep apnea (OSA) is a common sleep-related breathing disorder with an estimation of one billion people. Full-night polysomnography is considered the gold standard for OSA diagnosis. However, it is time-consuming, expensive and is not readily available in many parts of the world. Many screening questionnaires and scores have been proposed for OSA prediction with high sensitivity and low specificity. The present study is intended to develop models with various machine learning techniques to predict the severity of OSA by incorporating features from multiple questionnaires. Methods Subjects who underwent full-night polysomnography in Torr sleep center, Texas and completed 5 OSA screening questionnaires/scores were included. OSA was diagnosed by using Apnea-Hypopnea Index ≥ 5. We trained five different machine learning models including Deep Neural Networks with the scaled principal component analysis (DNN-PCA), Random Forest (RF), Adaptive Boosting classifier (ABC), and K-Nearest Neighbors classifier (KNC) and Support Vector Machine Classifier (SVMC). Training:Testing subject ratio of 65:35 was used. All features including demographic data, body measurement, snoring and sleepiness history were obtained from 5 OSA screening questionnaires/scores (STOP-BANG questionnaires, Berlin questionnaires, NoSAS score, NAMES score and No-Apnea score). Performance parametrics were used to compare between machine learning models. Results Of 180 subjects, 51.5 % of subjects were male with mean (SD) age of 53.6 (15.1). One hundred and nineteen subjects were diagnosed with OSA. Area Under the Receiver Operating Characteristic Curve (AUROC) of DNN-PCA, RF, ABC, KNC, SVMC, STOP-BANG questionnaire, Berlin questionnaire, NoSAS score, NAMES score, and No-Apnea score were 0.85, 0.68, 0.52, 0.74, 0.75, 0.61, 0.63, 0,61, 0.58 and 0,58 respectively. DNN-PCA showed the highest AUROC with sensitivity of 0.79, specificity of 0.67, positive-predictivity of 0.93, F1 score of 0.86, and accuracy of 0.77. Conclusion Our result showed that DNN-PCA outperforms OSA screening questionnaires, scores and other machine learning models. Support (if any):

Download Full-text

Understanding The Role Of Oxidative Stress In The Incidence Of Metabolic Syndrome And Obstructive Sleep Apnea

10.21203/rs.3.rs-279301/v1 ◽

2021 ◽

Author(s):

Behnam KARGAR ◽

Zahra ZAMANIAN ◽

Majid Bagheri HOSSEINABADI ◽

Vahid Gharibi ◽

Mohammad Sanyar MORADI ◽

...

Keyword(s):

Oxidative Stress ◽

Metabolic Syndrome ◽

Obstructive Sleep Apnea ◽

Logistic Regression ◽

Sleep Apnea ◽

Prolonged Exposure ◽

International Diabetes Federation ◽

Stress Index ◽

Obstructive Sleep

Abstract Background: Understanding the causes and risk factors of metabolic syndrome is important for promoting population health. Oxidative stress has been associated with metabolic syndrome, and also obstructive sleep apnea. These are two diseases which have common prognostic characteristics for heart disease. The aim of this study was to examine the role of oxidative stress in the concurrent presence of metabolic syndrome and obstructive sleep apnea in a working population. Methods: Participants were 163 artisan bakers in Shahroud, Iran, routinely exposed to oxidative stress indicators on a daily basis as part of their work. Using a cross-sectional design, data relevant to determining metabolic syndrome status according to International Diabetes Federation criteria, and the presence of obstructive sleep apnea according to the STOP-Bang score, was collected. Analyses included hierarchical binary logistic regression to yield predictors of the two diseases. Results: Logistic regression showed that oxidative stress – alongside obesity, no regular exercise, and smoking – was an independent predictor of metabolic syndrome, but not obstructive sleep apnea. Participants who were obese were 28 times more likely to have metabolic syndrome (OR 28.59, 95% CI 4.91-63.02) and 44 times more likely to have obstructive sleep apnea (OR 44.48, 95% CI 4.91-403.28). Participants meeting metabolic syndrome criteria had significantly higher levels of malondialdehyde (p < 0.05) than those who did not. No difference in oxidative stress index levels were found according to obstructive sleep apnea status. Conclusions: Our findings suggest that oxidative stress contributes to the onset of metabolic syndrome, and that obstructive sleep apnea is involved in oxidative stress. Whilst obesity, exercise, and smoking remain important targets for reducing the incidence of metabolic syndrome and obstructive sleep apnea, policies to control risks of prolonged exposure to oxidative stress are also relevant in occupations where such environmental conditions exist.

Download Full-text

Comparison of Artificial Neural Networks with Logistic Regression for Detection of Obesity

Journal of Medical Systems ◽

10.1007/s10916-011-9711-4 ◽

2011 ◽

Vol 36 (4) ◽

pp. 2449-2454 ◽

Cited By ~ 13

Author(s):

Seyed Taghi Heydari ◽

Seyed Mohammad Taghi Ayatollahi ◽

Najaf Zare

Keyword(s):

Neural Networks ◽

Logistic Regression ◽

Artificial Neural Networks ◽

Artificial Neural

Download Full-text

Abstract TMP55: Polysomnographic and Sleep-related Variables Associated to Leukoaraiosis in Patients with Obstructive Sleep Apnea

Stroke ◽

10.1161/str.48.suppl_1.tmp55 ◽

2017 ◽

Vol 48 (suppl_1) ◽

Author(s):

Kevin R Duque ◽

Brian Villafuerte ◽

Fiorella Adrianzen ◽

Rodrigo Zamudio ◽

Andrea Mendiola ◽

...

Keyword(s):

Obstructive Sleep Apnea ◽

Logistic Regression ◽

Sleep Apnea ◽

Multinomial Logistic Regression ◽

Sleep Time ◽

Cross Sectional Study ◽

Apnea Hypopnea Index ◽

Cross Sectional ◽

Obstructive Sleep ◽

Medical Histories

Introduction: Obstructive sleep apnea (OSA) is a biological plausible risk factor for leukoaraiosis (LA). We tested the hypothesis that polysomnographic (PSG) and sleep-related variables are associated to LA in OSA patients. Methods: Cross-sectional study in which PSG records, medical histories and brain 1.5T MRI were collected from all consecutive patients who had attended a Sleep Medicine Center between 2009-2014. LA was graded from 0 to 9 with the ’Atherosclerosis Risk In Communities’ study scale. OSA was defined by The International Classification of Sleep Disorders, 2014, and its severity categorizing according to apnea-hypopnea index (AHI, <15 mild, 15 to <30 moderate, 30 to <45 severe and ≥45 very severe). A multinomial logistic regression was performed to describe the association between OSA severity and LA (divided into 2 groups: mild-to-moderate LA and non-to-minimal LA). The covariates for all regression models were age, gender, BMI, hypertension, ischemic stroke, myocardial infarction, diabetes and pack-year of smoking. Results: From 82 OSA patients (77% male; mean age 58±9 years, range 19-91), 54 (66%) had LA. Mild-to-moderate LA was found in 13 patients (8 mild and 5 moderate LA) and non-to-minimal LA in 69 (41 minimal and 28 non LA). Spearman’s correlation coefficient between AHI and LA grade was 0.41 (p<0.001). Furthermore, the higher OSA severity, the higher LA severity (p<0.001, for Jonckheere-Terpstra test for ordered alternatives). In the multinomial logistic regression model adjusted for cofounders, severe OSA patients had higher risk for mild or moderate LA (HR 12.8, 95% IC 1.2-141) compared to mild-to-moderate OSA patients. Additionally, self-reported habitual sleep duration from 7 to 9 hours (HR 0.36, 90% IC 0.14-0.90) and proportion of time in apnea/hypopnea over total sleep time (HR 1.04 for one unit increase, 90% IC 1.01-1.08) could be associated with the presence of LA (adjusted only for age and gender). In a multiple regression analysis with all the aforementioned variables, age (p=0.002), diabetes (p=0.003), and OSA severity (p=0.04) were predictors of the presence of LA. Conclusion: Patients with severe OSA had higher risk for mild to moderate LA when compared to patients with mild or moderate OSA.

Download Full-text

0705 High Incidence Of Inadequate Health Literacy In Adults With Obstructive Sleep Apnea Is Not Associated With Positive Airway Pressure Use

SLEEP ◽

10.1093/sleep/zsaa056.701 ◽

2020 ◽

Vol 43 (Supplement_1) ◽

pp. A268-A269

Author(s):

A J Watach ◽

B Saconi ◽

A M Sawyer

Keyword(s):

Obstructive Sleep Apnea ◽

Logistic Regression ◽

Health Literacy ◽

Sleep Apnea ◽

Airway Pressure ◽

Positive Airway Pressure ◽

Lead Author ◽

Obstructive Sleep ◽

The Relationship ◽

Inadequate Health Literacy

Abstract Introduction Inadequate health literacy (HL) is associated with 1.5 - 3 times increased risk for poor health outcomes, nonadherence and lack of skills needed to manage one’s own health. Inadequate HL prevalence in adults with obstructive sleep apnea (OSA) may be as high as 30%. The relationship between HL and positive airway pressure (PAP) adherence has been rarely examined. Methods A secondary analysis of prospective observational data was conducted to: 1) examine the prevalence of inadequate HL in adults with newly-diagnosed OSA and 2) determine if inadequate HL is associated with 1-wk and 1-mo PAP use. HL was measured using a 3-item Health Literacy Screening Questionnaire. Descriptive statistics, multiple linear regression, and logistic regression were used. Results Participants (n=67) were white (85%), males (52%) and females (48%), middle-aged (50±12 yrs), 64.2% had a middle to high school education and severe OSA (mean AHI 38.2±21 events/hr). Mean PAP use was 4.62±2.43 hrs/night at 1-wk and 4.33±2.27hrs/night at 1-mo. Using a threshold of ≥4 hrs/night, 64% were adherent at 1-wk and 60% at 1-mo. Sixty two percent (n=42) screened positive for inadequate HL. A positive screen for inadequate HL (by individual screening items or by cumulative number of items screened positive) was not associated with PAP use (mean hr/night) at 1-wk or 1-mo (not retained in the final model). HL was also not associated with PAP non-adherence (<4hrs/night) or PAP failure (<2hrs/night) by logistic regression. Conclusion Inadequate HL may be prevalent in adults with OSA. OSA and PAP patient education content and design should align with HL abilities and skills. Disease and treatment education are influential on PAP adherence. Future research should consider adequacy of three generalized items to assess HL and disease-specific HL as more robust measures are available. Larger, heterogeneous sample sizes are needed to precisely estimate the relationship between HL and PAP adherence. Support Lead author receives support from NIH/NHLBI Award T32 HL07953.

Download Full-text