A clinical prediction formula for apnea-hypopnea index

OSApredictor: A tool for prediction of moderate to severe obstructive sleep apnea-hypopnea using readily available patient characteristics

Sleep apnea is a common sleep disorder. The availability of an easy-to-use sleep apnea predictor would provide a public health benefit by promoting early diagnosis and treatment. Our goal was to develop a prediction tool that used commonly available variables and was accessible to the public through a web site. Using data from polysomnography (PSG) studies that measured the apnea-hypopnea index (AHI), we built a machine learning tool to predict the presence of moderate to severe obstructive sleep apnea (OSA) (defined as AHI ≥15). Our tool employs only seven widely available predictor variables: age, sex, weight, height, pulse oxygen saturation, heart rate and respiratory rate. As a preliminary step, we used 16,958 PSG studies to examine eight machine learning algorithms via five-fold cross validation and determined that XGBoost exhibited superior predictive performance. We then refined the XGBoost predictor by randomly partitioning the data into a training and a test set (13,566 and 3392 PSGs, respectively) and repeatedly subsampling from the training set to construct 1000 training subsets. We evaluated each of the resulting 1000 XGBoost models on the single set-aside test set. The resulting classification tool correctly identified 72.5 % of those with moderate to severe OSA as having the condition (sensitivity) and 62.8 % of those without moderate to-severe OSA as not having it (specificity); overall accuracy was 66 %. We developed a user-friendly publicly available website (https://manticore.niehs.nih.gov/OSApredictor). We hope that our easy-to-use tool will serve as a screening vehicle that enables more patients to be clinically diagnosed and treated for OSA.

Validation of STOP, STOP-BANG, STOP-BAG, STOP-B28, and GOAL screening tools for identification of obstructive sleep apnea in patients with Parkinson disease

STUDY OBJECTIVES

Obstructive sleep apnea (OSA) is common in Parkinson disease (PD). Questionnaires can be used as screening tools and have been used as a surrogate definition of OSA in large-scale research. This study aimed to validate the performance of STOP, STOP-BANG, STOP-BAG, STOP-B28, and GOAL and OSA predictors as tools to identify OSA in PD.

METHODS

Data were analyzed from a PD cohort study in which OSA was diagnosed using laboratory polysomnography. We calculated sensitivity and specificity of each questionnaire for OSA using different definitions and performed receiver operating characteristics curve analysis. Linear regression was used to assess adjusted associations between questionnaires and outcomes: Montreal Cognitive Assessment, Epworth Sleepiness Scale, and Movement Disorder Society revision of the Unified Parkinson Disease Rating Scale.

RESULTS

Questionnaire data were available for 68 PD patients (61.8% male, mean age 64.5 [standard deviation 9.9] years, and Hoehn and Yahr score 2.1 [0.8]). OSA (apnea-hypopnea index ≥ 15 events/h) occurred in 69.4% of participants. STOP-B28 ≥ 2 presented a higher sensitivity for OSA than STOP ≥ 2 (0.76 vs 0.65, respectively) and slightly lower specificity (0.65 vs 0.70, respectively). GOAL ≥ 2 had the highest sensitivity but poor specificity. Loud snoring had sensitivity 0.63 and specificity 0.65. STOP and snoring were significantly associated with Montreal Cognitive Assessment, Epworth Sleepiness Scale, and Movement Disorder Society revision of the Unified Parkinson Disease Rating Scale (total, motor, and nonmotor); STOP-BANG, STOP-BAG, and STOP-B28 showed associations with most outcomes, but the GOAL showed none.

CONCLUSIONS

The STOP-B28 followed by STOP and presence of loud snoring alone seem to have the best overall properties to identify PD patients with OSA, whose clinical characteristics differ from the general population with OSA.

CITATION

Gomes T, Benedetti A, Lafontaine A-L, Kimoff RJ Robinson A, Kaminska M. Validation of STOP, STOP-BANG, STOP-BAG, STOP-B28, and GOAL screening tools for identification of obstructive sleep apnea in patients with Parkinson disease. J Clin Sleep Med. 2023;19(1):45-54.

Review of Application of Machine Learning as a Screening Tool for Diagnosis of Obstructive Sleep Apnea

Obstructive sleep apnea syndrome (OSAS) is a pervasive disorder with an incidence estimated at 5-14 percent among adults aged 30-70 years. It carries significant morbidity and mortality risk from cardiovascular disease, including ischemic heart disease, atrial fibrillation, and cerebrovascular disease, and risks related to excessive daytime sleepiness. The gold standard for diagnosis of OSAS is the polysomnography (PSG) test which requires overnight evaluation in a sleep laboratory and expensive infrastructure, which renders it unsuitable for mass screening and diagnosis. Alternatives such as home sleep testing need patients to wear diagnostic instruments overnight, but accuracy continues to be suboptimal while access continues to be a barrier for many. Hence, there is a continued significant underdiagnosis and under-recognition of sleep apnea in the community, with at least one study suggesting that 80-90% of middle-aged adults with moderate to severe sleep apnea remain undiagnosed. Recently, we have seen a surge in applications of artificial intelligence and neural networks in healthcare diagnostics. Several studies have attempted to examine its application in the diagnosis of OSAS. Signals included in data analytics include Electrocardiogram (ECG), photo-pletysmography (PPG), peripheral oxygen saturation (SpO2), and audio signals. A different approach is to study the application of machine learning to use demographic and standard clinical variables and physical findings to try and synthesize predictive models with high accuracy in assisting in the triage of high-risk patients for sleep testing. The current paper will review this latter approach and identify knowledge gaps that may serve as potential avenues for future research.

Helping early obstructive sleep apnea diagnosis with machine learning: A systematic review (Preprint)

Background

American Academy of Sleep Medicine guidelines suggest that clinical prediction algorithms can be used to screen patients with obstructive sleep apnea (OSA) without replacing polysomnography, the gold standard.

Objective

We aimed to identify, gather, and analyze existing machine learning approaches that are being used for disease screening in adult patients with suspected OSA.

Methods

We searched the MEDLINE, Scopus, and ISI Web of Knowledge databases to evaluate the validity of different machine learning techniques, with polysomnography as the gold standard outcome measure and used the Prediction Model Risk of Bias Assessment Tool (Kleijnen Systematic Reviews Ltd) to assess risk of bias and applicability of each included study.

Results

Our search retrieved 5479 articles, of which 63 (1.15%) articles were included. We found 23 studies performing diagnostic model development alone, 26 with added internal validation, and 14 applying the clinical prediction algorithm to an independent sample (although not all reporting the most common discrimination metrics, sensitivity or specificity). Logistic regression was applied in 35 studies, linear regression in 16, support vector machine in 9, neural networks in 8, decision trees in 6, and Bayesian networks in 4. Random forest, discriminant analysis, classification and regression tree, and nomogram were each performed in 2 studies, whereas Pearson correlation, adaptive neuro-fuzzy inference system, artificial immune recognition system, genetic algorithm, supersparse linear integer models, and k-nearest neighbors algorithm were each performed in 1 study. The best area under the receiver operating curve was 0.98 (0.96-0.99) for age, waist circumference, Epworth Somnolence Scale score, and oxygen saturation as predictors in a logistic regression.

Conclusions

Although high values were obtained, they still lacked external validation results in large cohorts and a standard OSA criteria definition.

Trial Registration

PROSPERO CRD42021221339; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=221339

A review on the anesthetic management of obese patients undergoing surgery

There has been an observed increase in theprevalence of obesity over the past few decades. The prevalence of anesthesiology related complications is also observed more frequently in obese patients as compared to patients that are not obese. Due to the increased complications that accompany obesity, obese patients are now more often requiring surgical interventions. Therefore, it is important that anesthesiologists be aware of this development and is equipped to manage these patients effectively and appropriately. As a result, this review highlights the effective management of obese patients undergoing surgery focusing on the preoperative, perioperative and postoperative care of these patients.

Is there a relationship between tonsil volume and the success of pharyngeal surgery among adult patients with obstructive sleep apnea?

OBJECTIVES

Hypertrophic palatine tonsils play a role in the blockage of the upper airway, one of the known causes of Obstructive Sleep Apnea (OSA). Therefore, it is possible that there is an association between tonsil size and the success of pharyngeal surgery during OSA treatment. The main objective of this study was to evaluate the relationship between tonsil grade and volume, as well as to establish whether a relationship exists between tonsil size and the success rate of pharyngeal surgery (tonsillectomy and pharyngoplasty with barbed sutures).

METHODS

This retrospective study includes forty-four adult patients who underwent tonsillectomy and pharyngeal surgery with barbed sutures for the treatment of simple snoring and OSA between January 2016 and September 2019. Patients who had been previously tonsillectomized or those for whom tonsil volume measurement was lacking were excluded. All patients underwent a pre-operative physical exploration at the clinic exam room and a sleep study. Prior to surgery a Drug Induced Sleep Endoscopy (DISE) was performed. Tonsil volume was measured intraoperatively using the water displacement method. The same sleep study was repeated six months following surgery.

RESULTS

A significant correlation was found between tonsil grade and volume and between such measurements and the blockage observed at the level of the oropharynx during the DISE. Moreover, an association was observed between tonsil volume, but not tonsil grade, and the success of tonsillectomy and pharyngoplasty with barbed sutures. A tonsil volume greater than 6.5 cm³ was linked to success during pharyngeal surgery.

CONCLUSION

A correlation exists between tonsil grade and tonsil volume. A bigger tonsil volume is associated with a greater success rate of oropharyngeal surgery during treatment of OSA.

LEVEL OF EVIDENCE

Level 3, non-randomized cohort study.

Explainable fuzzy neural network with easy-to-obtain physiological features for screening obstructive sleep apnea-hypopnea syndrome

OBJECTIVE/BACKGROUND

Recently, several tools for screening obstructive sleep apnea-hypopnea syndrome (OSAHS) have been devised with varied shortcomings. To overcome these drawbacks, we aimed to propose a self-estimation method using an explainable prediction model with easy-to-obtain variables and evaluate its performance for predicting OSAHS.

PATIENTS/METHODS

This retrospective, cross-sectional study selected significant easy-to-obtain variables from patients, suspected of having OSAHS by regression analysis, and fed these variables into the proposed explainable fuzzy neural network (EFNN), a back propagation neural network (BPNN) and a stepwise regression model to compare the screening performance for OSAHS.

RESULTS

Of the 300 participants, three easily available features, such as waist circumference, mean blood pressure (BP) at the end of polysomnography and the difference in systolic BP between the end and start of polysomnography, were obtained from regression analysis with a five-fold cross-validation scheme. Feeding these three variables into the prediction models showed that the average prediction differences for apnea-hypopnea index (AHI) when using the EFNN, BPNN, and regression model were respectively 1.5 ± 18.2, 3.5 ± 19.1 and 0.1 ± 19.3, indicating none of the tested methods had good efficacy to predict the AHI values. The performance as determined by the sensitivity + specificity-1 value for screening moderate-to-severe OSAHS of the EFNN, BPNN and regression model were respectively 0.440, 0.414 and 0.380.

CONCLUSIONS

When fed with easy-to-obtain physiological features, the understandable EFNN should be the preferred method to predict moderate-to-severe OSAHS.

Predicting sleep apnea from three-dimensional face photography

STUDY OBJECTIVES

Craniofacial anatomy is recognized as an important predisposing factor in the pathogenesis of obstructive sleep apnea (OSA). This study used three-dimensional (3D) facial surface analysis of linear and geodesic (shortest line between points over a curved surface) distances to determine the combination of measurements that best predicts presence and severity of OSA.

METHODS

3D face photographs were obtained in 100 adults without OSA (apnea-hypopnea index [AHI] < 5 events/h), 100 with mild OSA (AHI 5 to < 15 events/h), 100 with moderate OSA (AHI 15 to < 30 events/h), and 100 with severe OSA (AHI ≥ 30 events/h). Measurements of linear distances and angles, and geodesic distances were obtained between 24 anatomical landmarks from the 3D photographs. The accuracy with which different combinations of measurements could classify an individual as having OSA or not was assessed using linear discriminant analyses and receiver operating characteristic analyses. These analyses were repeated using different AHI thresholds to define presence of OSA.

RESULTS

Relative to linear measurements, geodesic measurements of craniofacial anatomy improved the ability to identify individuals with and without OSA (classification accuracy 86% and 89% respectively, P < .01). A maximum classification accuracy of 91% was achieved when linear and geodesic measurements were combined into a single predictive algorithm. Accuracy decreased when using AHI thresholds ≥ 10 events/h and ≥ 15 events/h to define OSA although greatest accuracy was always achieved using a combination of linear and geodesic distances.

CONCLUSIONS

This study suggests that 3D photographs of the face have predictive value for OSA and that geodesic measurements enhance this capacity.

Risk factors and clinical prediction formula for the evaluation of obstructive sleep apnea in Asian adults

Obstructive sleep apnea is a highly prevalent cyclic repetitive hypoxia-normoxia respiratory sleep disorder characterized by intermittent upper-airway collapse. It is mainly diagnosed using in-laboratory polysomnography. However, the time-spatial constraints of this procedure limit its application. To overcome these limitations, there have been studies aiming to develop clinical prediction formulas for screening of obstructive sleep apnea using the risk factors for this disorder. However, the applicability of the formula is restricted by the group specific factors included in it. Therefore, we aimed to assess the risk factors for obstructive sleep apnea and develop clinical prediction formulas, which can be used in different situations, for screening and assessing this disorder. We enrolled 3,432 Asian adult participants with suspected obstructive sleep apnea who had successfully undergone in-laboratory polysomnography. All parameters were evaluated using correlation analysis and logistic regression. Among them, age, sex, hypertension, diabetes mellitus, anthropometric factors, Berlin questionnaire and Epworth Sleepiness Scale scores, and anatomical tonsil and tongue position were significantly associated with obstructive sleep apnea. To develop the clinical formulas for obstructive sleep apnea, the participants were divided into the development (n = 2,516) and validation cohorts (n = 916) based on the sleep laboratory visiting date. We developed and selected 13 formulas and divided them into those with and without physical examination based on the ease of application; subsequently, we selected suitable formulas based on the statistical analysis and clinical applicability (formula including physical exam: sensitivity, 0.776; specificity, 0.757; and AUC, 0.835; formula without physical exam: sensitivity, 0.749; specificity, 0.770; and AUC, 0.839). Analysis of the validation cohort with developed formulas showed that these models and formula had sufficient performance and goodness of fit of model. These tools can effectively utilize medical resources for obstructive sleep apnea screening in various situations.

Clinical Prediction Model for Obstructive Sleep Apnea among Adult Patients with Habitual Snoring

OBJECTIVE

To identify standard clinical parameters that may predict the presence and severity of obstructive sleep apnea/hypopnea syndrome (OSA).

DESIGN

Case series with chart review.

SETTING

Tertiary academic medical center.

SUBJECTS AND METHODS

A total of 325 adult patients (274 men and 51 women; mean age, 44.2 years) with habitual snoring completed comprehensive polysomnography and anthropometric measurements, including modified Mallampati grade (also known as updated Friedman's tongue position [uFTP]), tonsil size grading, uvular length, neck circumference, waist circumference, hip circumference, and body mass index (BMI).

RESULTS

When the aforementioned physical parameters were correlated singly with the apnea/hypopnea index (AHI), we found that sex, uFTP, tonsil size grading, neck circumference, waist circumference, hip circumference, thyroid-mental distance, and BMI grade were reliable predictors of OSA. When all important factors were considered in a multiple stepwise regression analysis, an estimated AHI can be formulated by factoring sex, uFTP, tonsil size grading, and BMI grade as follows: -43.0 + 14.1 × sex + 12.8 × uFTP + 5.0 × tonsil size + 8.9 × BMI grade. Severity of OSA can be predicted with a receiver operating characteristic curve. Predictors of OSA can be further obtained by the "OSA score."

CONCLUSION

This study has distinguished the correlations between sex, uFTP, tonsil size, and BMI grade and the presence and severity of OSA. An OSA score might be beneficial in identifying patients who should have a full sleep evaluation.

Tonsil size and outcome of uvulopalatopharyngoplasty with tonsillectomy in obstructive sleep apnea

OBJECTIVES/HYPOTHESIS

To investigate the relationship of tonsil volume and grade on outcomes of uvulopalatopharyngoplasty (UPPP) with tonsillectomy in patients with obstructive sleep apnea (OSA).

STUDY DESIGN

Retrospective cohort analysis.

METHODS

Data of 70 consecutive patients undergoing UPPP with tonsillectomy between 2015 and 2018 were analyzed. Patients with an apnea-hypopnea index (AHI) <10/hr or concomitant surgery other than nasal surgery were excluded. Tonsil volume was measured intraoperatively. Preoperatively and 3 months after surgery we assessed the AHI using respiratory polygraphy, daytime sleepiness using the Epworth Sleepiness Scale (ESS), and a visual analog scale for the snoring index (SI).

RESULTS

Tonsil grade and volume both showed a significant correlation with preoperative AHI. Postoperative AHI was not significantly different between grades and volume. The AHI reduction after surgery increased significantly with larger volume and higher tonsil grade. For all grades, the postoperative ESS was significantly reduced compared to the preoperative value, but was not significantly correlated with tonsil volume. Preoperative and postoperative SI was not significantly correlated between tonsil grade or volume. In all grades, SI was significantly reduced after surgery.

CONCLUSIONS

In our study, we found that large tonsils are responsible for higher preoperative AHI values, and their removal leads to greater reduction of initial AHI. However, the postoperative effect on daytime sleepiness and snoring reduction is not significantly correlated with tonsil size and volume, indicating that these parameters are mainly influenced by other factors. The knowledge of the significance of tonsil size and volume is important for ear, nose, and throat physicians when counseling OSA patients.

LEVEL OF EVIDENCE

2c Laryngoscope, 129:E449-E454, 2019.

Correlation between Brodsky Tonsil Scale and Tonsil Volume in Adult Patients

Purpose

To evaluate the value of Brodsky tonsil scale in predicting the objective tonsil volume and to identify the potential factors that might interfere with the accuracy of prediction.

Methods

A total of 87 adult patients who underwent single tonsillectomy or uvulopalatopharyngoplasty (UPPP) procedure including tonsillectomy in our hospital between Jan 2015 and Dec 2016 were included. The data of Brodsky tonsil scale evaluated preoperatively and objective tonsil volume evaluated postoperatively were collected for analysis.

Results

Among the 87 adult patients included, 85 patients underwent bilateral tonsillectomy, while only 2 underwent unilateral procedure. Therefore, a total of 172 tonsils were included. Significant positive correlations were established between Brodsky scale and objective volume for either right (R = 0.647), left (R = 0.664), or overall tonsils (R = 0.654) (all p < 0.001). However, volume overlaps could be found between 2+ and 3+ tonsils. Age [odds ratio (OR) = 4.053, p = 0.003] and body mass index (BMI; OR=1.740, p = 0.044) were found to be independent factors that could influence the consistency between the Brodsky scale and objective volume. As a result, a formula “Index = -1.409+1.399×age+0.554×BMI” was constructed for the evaluation of the consistency.

Conclusion

Tonsil grading was significantly correlated with tonsil volume; preoperative tonsil grading that reflected the real tonsil volume was regarded as the protocol for the evaluation of the tonsil size. Age and BMI were independent factors that could affect the consistency between tonsil grade and tonsil volume. A mathematical model was estimated to predict the consistency accurately.

Can Statistical Machine Learning Algorithms Help for Classification of Obstructive Sleep Apnea Severity to Optimal Utilization of Polysomno graphy Resources?

Objectives: The goal of this study is to evaluate the results of machine learning methods for the classification of OSA severity of patients with suspected sleep disorder breathing as normal, mild, moderate and severe based on non-polysomnographic variables: 1) clinical data, 2) symptoms and 3) physical examination.

Methods: In order to produce classification models for OSA severity, five different machine learning methods (Bayesian network, Decision Tree, Random Forest, Neural Networks and Logistic Regression) were trained while relevant variables and their relationships were derived empirically from observed data. Each model was trained and evaluated using 10-fold cross-validation and to evaluate classification performances of all methods, true positive rate (TPR), false positive rate (FPR), Positive Predictive Value (PPV), F measure and Area Under Receiver Operating Characteristics curve (ROC-AUC) were used.

Results: Results of 10-fold cross validated tests with different variable settings promisingly indicated that the OSA severity of suspected OSA patients can be classified, using non-polysomnographic features, with 0.71 true positive rate as the highest and, 0.15 false positive rate as the lowest, respectively. Moreover, the test results of different variables settings revealed that the accuracy of the classification models was significantly improved when physical examination variables were added to the model.

Conclusions: Study results showed that machine learning methods can be used to estimate the probabilities of no, mild, moderate, and severe obstructive sleep apnea and such approaches may improve accurate initial OSA screening and help referring only the suspected moderate or severe OSA patients to sleep laboratories for the expensive tests.

Liver Enlargement Predicts Obstructive Sleep Apnea-Hypopnea Syndrome in Morbidly Obese Women

Obstructive sleep apnea-hypopnea syndrome (OSAHS) is frequently present in patients with severe obesity, but its prevalence especially in women is not well defined. OSAHS and non-alcoholic fatty liver disease are common conditions, frequently associated in patients with central obesity and metabolic syndrome and are both the result of the accumulation of ectopic fat mass. Identifying predictors of risk of OSAHS may be useful to select the subjects requiring instrumental sleep evaluation. In this cross-sectional study, we have investigated the potential role of hepatic left lobe volume (HLLV) in predicting the presence of OSAHS. OSAHS was quantified by the apnea/hypopnea index (AHI) and oxygen desaturation index in a cardiorespiratory inpatient sleep study of 97 obese women [age: 47 ± 11 years body mass index (BMI): 50 ± 8 kg/m²]. OSAHS was diagnosed when AHI was ≥5. HLLV, subcutaneous and intra-abdominal fat were measured by ultrasound. After adjustment for age and BMI, both HLLV and neck circumference (NC) were independent predictors of AHI. OSAHS was found in 72% of patients; HLLV ≥ 370 cm³ was a predictor of OSAHS with a sensitivity of 66%, a specificity of 70%, a positive and negative predictive values of 85 and 44%, respectively (AUC = 0.67, p < 0.005). A multivariate logistic model was used including age, BMI, NC, and HLLV (the only independent predictors of AHI in a multiple linear regression analyses), and a cut off value for the predicted probability of OSAHS equal to 0.7 provided the best diagnostic results (AUC = 0.79, p < 0.005) in terms of sensitivity (76%), specificity (89%), negative and positive predictive values (59 and 95%, respectively). All patients with severe OSAHS were identified by this prediction model. In conclusion, HLLV, an established index of visceral adiposity, represents an anthropometric parameter closely associated with OSAHS in severely obese women.

The effect of acetazolamide on sleep apnea at high altitude: a systematic review and meta-analysis

Background:

Acetazolamide has been investigated for treating sleep apnea in newcomers ascending to high altitude. This study aimed to assess the effect of acetazolamide on sleep apnea at high altitude, determine the optimal therapeutic dose, and compare its effectiveness in healthy trekkers and obstructive sleep apnea (OSA) patients.

Methods:

PubMed, Embase, Scopus, Cochrane Library, and Airiti Library databases were searched up to July 2015 for randomized controlled trials (RCTs) performed above 2500 m in lowlanders and that used acetazolamide as intervention in sleep studies. Studies including participants with medical conditions other than OSA were excluded.

Results:

Eight studies of 190 adults were included. In healthy participants, the pooled mean effect sizes of acetazolamide on Apnea–Hypopnea Index (AHI), percentage of periodic breathing time, and nocturnal oxygenation were 34.66 [95% confidence interval (CI) 25.01–44.30] with low heterogeneity (p = 0.7, I² = 0%), 38.56% (95% CI 18.92–58.19%) with low heterogeneity (p = 0.24, I² = 28%), and 4.75% (95% CI 1.35–8.15%) with high heterogeneity (p < 0.01, I² = 87%), respectively. In OSA patients, the pooled mean effect sizes of acetazolamide on AHI and nocturnal oxygenation were 13.18 (95% CI 9.25–17.1) with low heterogeneity (p = 0.33, I² = 0%) and 1.85% (95% CI 1.08–2.62%) with low heterogeneity (P = 0.56, I² = 0%).

Conclusions:

Acetazolamide improves sleep apnea at high altitude by decreasing AHI and percentage of periodic breathing time and increasing nocturnal oxygenation. Acetazolamide is more beneficial in healthy participants than in OSA patients, and a 250 mg daily dose may be as effective as higher daily doses for healthy trekkers.