Self-helped detection of obstructive sleep apnea based on automated facial recognition and machine learning.

PURPOSE: The diagnosis of obstructive sleep apnea (OSA) relies on time-consuming and complicated procedures which are not always readily available and may delay diagnosis. With the widespread use of artificial intelligence, we presumed that the combination of simple clinical information and imaging recognition based on facial photos may be a useful tool to screen for OSA. METHODS: We recruited consecutive subjects suspected of OSA who had received sleep examination and photographing. Sixty-eight points from 2-dimensional facial photos were labelled by automated identification. An optimized model with facial features and basic clinical information was established and tenfold cross-validation was performed. Area under the receiver operating characteristic curve (AUC) indicated the model's performance using sleep monitoring as the reference standard. RESULTS: A total of 653 subjects (77.2% males, 55.3% OSA) were analyzed. CATBOOST was the most suitable algorithm for OSA classification with a sensitivity, specificity, accuracy, and AUC of 0.75, 0.66, 0.71, and 0.76 respectively (P < 0.05), which was better than STOP-Bang questionnaire, NoSAS scores, and Epworth scale. Witnessed apnea by sleep partner was the most powerful variable, followed by body mass index, neck circumference, facial parameters, and hypertension. The model's performance became more robust with a sensitivity of 0.94, for patients with frequent supine sleep apnea. CONCLUSION: The findings suggest that craniofacial features extracted from 2-dimensional frontal photos, especially in the mandibular segment, have the potential to become predictors of OSA in the Chinese population. Machine learning-derived automatic recognition may facilitate the self-help screening for OSA in a quick, radiation-free, and repeatable manner.

Systematic review and meta-analysis of the clinical outcomes of ACEI/ARB in East-Asian patients with COVID-19.

BACKGROUND: SARS-CoV-2 invades human cells and leads to COVID-19 by direct associating with angiotensin converting enzyme 2 (ACE2) receptors, the level of which may be increased by treatment with angiotensin-converting enzyme inhibitors (ACEIs) and/or angiotensin receptor blockers (ARBs). This meta-analysis aimed to explore the impact of ACEI/ARB treatment on the clinical outcomes of patients with COVID-19 infections among population in the East-Asia region. METHODS: We collected clinical data published from January 2000 to May 2022 in the English databases including PubMed, Embase, and the Cochrane Library. Two reviewers independently screened and identified studies that met the prespecified criteria. Review Manager 5.3 software was used to perform the meta-analysis. RESULTS: A total of 28 articles were included in this analysis. The results showed that patients who were prescribed with ACEI/ARB had a shorter duration of hospital stay [MD = -2.37, 95%CI (-3.59, -1.14), P = 0.000 2] and a lower mortality rate [OR = 0.61, 95% CI (0.52, 0.70), P<0.000 01] than patients who were not on ACEI/ARB. Furthermore, there was no statistically significant difference in disease severity [OR = 0.99, 95% CI (0.83, 1.17), P = 0.90] between individuals receiving ACEI/ARB or not. CONCLUSIONS: This meta-analysis suggested that the use of ACEI/ARB was not associated with adverse clinical outcomes in East-Asian Covid-19 patients and a reduced mortality and shorter duration of hospital stay among East-Asian population (especially for female subjects) was found. Thus, ACEI/ARB should be continued in patients infected by Covid-19.

Fast assessment of left ventricular systolic function in obstructive sleep apnea patients with automated function imaging: Comparison with mitral annular plane systolic excursion.

BACKGROUND: Early detection of left ventricular (LV) subclinical dysfunction is clinically relevant before developing irreversible impairment in obstructive sleep apnea (OSA) patients. Mitral annulus plane systolic excursion (MAPSE) is a fast tool for OSA due to high prevalent obesity; another quick but more comprehensive tool is LV global longitudinal stain (GLS) based on automated function imaging (AFI). We therefore aimed to compare the feasibility and reproducibility of AFI to MAPSE in OSA patients, as a good model in whom obesity is common. METHODS: A comprehensive echocardiographic examination was done in 186 consecutive patients having polysomnography for suspected OSA. MAPSE was measured by using M-mode to calculate excursion of mitral annulus. GLS was derived by offline analysis of three long-axis views that semi-automatically detects LV endocardial boundary, which is adjusted manually as necessary with AFI measurement. Variability of AFI and MAPSE were compared among the different subgroups. RESULTS: Despite a relatively high obesity rate (42.9%), the feasibility of AFI was 94% (175/186) and that of 100% in MAPSE. AFI showed excellent correlation (r = .882) superior to MAPSE (r = .819) between the Expert and Beginner. Intra- and inter- observer variability of AFI and MAPSE in Bland-Altman analysis were 5.5% and 6.5%; 6.2% and 8.8%, respectively. In repeated measurements, AFI showed higher intra-class correlation (ICC = .95) than MAPSE (ICC = .87) (p < 0.05). Furthermore, analysis showed that AFI was feasible even in more obese patients (BMI≥28 kg/m2 ). CONCLUSIONS: Even in obese patients with OSA, AFI-GLS is feasible and more reliable for less expert operators than MAPSE in detecting LV longitudinal dysfunction.

Association Between OSA and Quantitative Atherosclerotic Plaque Burden: A Coronary CT Angiography Study.

BACKGROUND: Limited evidence is available regarding the association between OSA and coronary plaque assessed by using quantitative coronary CT angiography. RESEARCH QUESTION: Are there any associations between OSA severity-related indexes and the presence and burden of coronary plaque? STUDY DESIGN AND METHODS: Cross-sectional data from 692 patients who underwent sleep monitoring and coronary CT angiography were used for this study. Of these patients, 120 (17.3%) underwent polysomnography, and 572 (82.7%) underwent respiratory polygraphy. Multivariable logistic and linear regression analyses were used to investigate the associations of OSA severity-related indexes with the presence, volume, and composition of plaque. RESULTS: In multivariable analyses, patients with moderate to severe OSA were more likely to have coronary plaques (P = .037), and plaques were more likely to contain a noncalcified plaque (NCP) component (P = .032) and a low-density NCP (LD NCP) component (P = .030). Furthermore, the apnea-hypopnea index and oxygen desaturation index as continuous variables were both associated with the presence of plaque, NCP, and LD NCP (all, P < .05). Multivariable linear regression models showed that moderate to severe OSA was associated with NCP volume (ß = 50.328; P = .042) and LD NCP volume (ß = 15.707; P = .011). Moreover, the apnea-hypopnea index (P = .015), oxygen desaturation index (P = .005), and percentage of nighttime with oxygen saturation < 90% (P = .017) were all significant predictors of LD NCP volume. Compared with those with no or mild OSA, patients with severe OSA had a significantly higher total plaque volume (P = .036), NCP volume (P = .036), and LD NCP volume (P = .013). INTERPRETATION: OSA was independently associated with the presence and burden of coronary plaque, which suggests an increased risk of coronary events. CLINICAL TRIAL REGISTRATION: Chinese Clinical Trial Registry; No. ChiCTR-ROC-17011027; http://chictr.org.cn.

Subclinical left ventricular systolic dysfunction detected in obstructive sleep apnea with automated function imaging and its association with nocturnal hypoxia.

BACKGROUND: Early detection of left ventricular (LV) dysfunction is crucial in obstructive sleep apnea (OSA) due to its close relationship with cardiovascular diseases. Global longitudinal strain (GLS) derived from automated function imaging (AFI) can precisely assess global longitudinal function. The aim of this study was to determine if LV GLS was reduced in patients with OSA and a normal LV ejection fraction (LVEF) and to assess any associated determinants. METHODS: Polysomnography (PSG) and echocardiography were done in consecutive patients with suspected OSA and normal LVEF in this prospective study. Patients were divided into two groups according to apnea-hypopnea index (AHI) (Group 1, normal or mild OSA: AHI < 15/h; Group 2, moderate-to-severe OSA: AHI ≥ 15/h). Clinical, PSG, and echocardiographic parameters were compared between the two groups and the associated factors were investigated. RESULTS: Of 425 consecutive patients, 244 were analyzed after exclusions. Patients in Group 2 had significantly worse GLS than those in Group 1 (p < 0.001). The prevalence of GLS reduction (defined as < - 19.7%) was 25% and 76%, respectively (χ2 = 34.19, p < 0.001). Nocturnal lowest pulse oxygen saturation (SpO2), AHI, body mass index (BMI), and gender were associated with GLS reduction (all p < 0.05). Further multivariate analysis showed that the lowest SpO2 (OR: 2.15), gender (OR: 2.45), and BMI (OR: 2.66) remained independent (all p < 0.05), and the lowest SpO2 was the most powerful determinant (χ2 = 33.0, p < 0.001) in forward regression analysis. The intra- and inter-operator variability for AFI and coefficient of repeatability was low even in those with relatively poor images. CONCLUSIONS: In patients with normal LVEF, more severe OSA was associated with a worse GLS. The major determinants were lowest nocturnal SpO2, gender, and obesity, but not AHI. GLS can be rapidly and reliably assessed using AFI.

Passive Prescription of Secondary Prevention Medical Therapy during Index Hospitalization for Acute Myocardial Infarction Is Prevalent and Associated with Adverse Clinical Outcomes.

Secondary prevention therapy reduces death and reinfarction after acute myocardial infarction (AMI), but it is underutilized in clinical practice. Mechanisms for this therapeutic gap are not well established. In this study, we have explored and evaluated the impact of passive continuation compared to active initiation of secondary prevention therapy for AMI during the index hospitalization. For this purpose, we have analyzed 1083 consecutive patients with AMI to a tertiary referral hospital in Hong Kong and assessed discharge prescription rates of secondary prevention therapies (aspirin, beta-blockers, statins, and ACEI/ARBs). Multivariate analysis was used to identify independent predictors of discharge medication, and Kaplan-Meier survival curve was used to evaluate 12-month survival. Overall, prescription rates of aspirin, beta-blocker, statin, and ACEI/ARBs on discharge were 94.8%, 64.5%, 83.5%, and 61.4%, respectively. Multivariate analysis showed that prior use of each therapy was an independent predictor of prescription of the same therapy on discharge: aspirin (odds ratio (OR) = 4.8, 95% CI = 1.9-12.3, P < 0.01), beta-blocker (OR = 2.5, 95% CI = 1.8-3.4, P < 0.01); statin (OR = 8.3, 95% CI = 0.4-15.7, P < 0.01), and ACEI/ARBs (OR = 2.9, 95% CI = 2.0-4.3, P < 0.01). Passive continuation of prior medication was associated with higher 1-year mortality rates than active initiation in treatment-naïve patients (aspirin (13.7% vs. 5.7%), beta-blockers (12.9% vs. 5.6%), and statins (11.0% vs. 4.6%); all P < 0.01). Overall, the use of secondary prevention medication for AMI was suboptimal. Our findings suggested that the practice of passive continuation of prior medication was prevalent and associated with adverse clinical outcomes compared to active initiation of secondary preventive therapies for acute myocardial infarction during the index hospitalization.

Obstructive sleep apnoea, intermittent hypoxia and heart failure with a preserved ejection fraction.

Obstructive sleep apnoea (OSA) is recognised to be a potent risk factor for hypertension, coronary heart disease, strokes and heart failure with a reduced ejection fraction. However, the association between OSA and heart failure with a preserved ejection fraction (HFpEF) is less well recognised. Both conditions are very common globally.It appears that there are many similarities between the pathological effects of OSA and other known aetiologies of HFpEF and its postulated pathophysiology. Intermittent hypoxia induced by OSA leads to widespread stimulation of the sympathetic nervous system, renin-angiotensin-aldosterone system and more importantly a systemic inflammatory state associated with oxidative stress. This is similar to the consequences of hypertension, diabetes, obesity and ageing that are the common precursors to HFpEF. The final common pathway is probably via the development of myocardial fibrosis and structural changes in collagen and myocardial titin that cause myocardial stiffening. Thus, considering the pathophysiology of OSA and HFpEF, OSA is likely to be a significant risk factor for HFpEF and further trials of preventive treatment should be considered.

Validation of a portable monitoring device for the diagnosis of obstructive sleep apnea: electrocardiogram-based cardiopulmonary coupling.

PURPOSE: We aimed to evaluate the validity of the cardiopulmonary coupling (CPC) device, a limited-channel portable monitoring device for obstructive sleep apnea (OSA) screening in one single-center cohort, in particular in those with some cardiovascular diseases since the cardiopulmonary coupling might be different from those without. METHODS: Consecutive patients referred to the sleep medical center for assessment of possible OSA were enrolled in this study. Patients were examined with standard polysomnography (PSG) and CPC evaluation simultaneously. The results of the two examinations were compared in all subjects and in those with or without cardiovascular abnormalities. RESULTS: A total of 179 subjects suspected with OSA were finally analyzed. According to OSA severity degree based on AHI, the area under ROC curve for the CPC device in the whole cohort patients was 0.79 (mild), 0.79 (moderate), and 0.86 (severe OSA), respectively (all p < 0.001). For patients with cardiovascular disease with different OSA severity, the area under the ROC curve was 0.86 (mild), 0.73 (moderate), and 0.83 (severe OSA), respectively (all p < 0.0001), and 0.74 (mild), 0.85 (moderate), and 0.91 (severe OSA), respectively in patients without cardiovascular disease (all p < 0.0001). CONCLUSIONS: The overall performance of CPC technique was acceptable to assess OSA in subjects with clinical suspicion of OSA, and thus it might act as a fast tool to screen OSA patients. However, the sensitivity of CPC technology for patients with cardiovascular disease was relatively insufficient. Therefore, CPC technology should be carefully interpreted in OSA screening in those with cardiovascular disease.

Dietary Fiber Intake, Myocardial Injury, and Major Adverse Cardiovascular Events Among End-Stage Kidney Disease Patients: A Prospective Cohort Study.

INTRODUCTION: Dialysis patients are frequently advised to restrict fruit and vegetable intake due to their high potassium content. This study aimed to evaluate the association between dietary fiber intake and major adverse cardiovascular events (MACE) among dialysis patients. METHODS: A total of 219 prevalent dialysis patients were prospectively recruited from a major university teaching hospital and regional dialysis center in Hong Kong. Dietary fiber intake estimated using a 7-day locally validated food frequency questionnaire was examined in relation to a primary composite outcome of MACE over a follow-up period of 4 years. RESULTS: A total of 127 patients were complicated with 1 or more MACE. In the multivariable Cox regression analysis, every 1 g higher fiber intake, and every 1 g/d per 1000 kcal higher fiber intake density were associated with an 11% (95% confidence interval [CI]: 0.81-0.97) and a 13% lower risk of MACE (95% CI: 0.77-0.99), respectively, independent of clinical, demographic, biochemical, hemodynamic, adequacy parameters, dietary protein, energy intake, inflammatory, and cardiac markers. Patients in the lower tertile of fiber intake density showed an increased hazard for MACE (adjusted hazard ratio: 1.78; 95% CI: 1.13-2.80) than those in the upper tertile. CONCLUSION: Higher fiber intake and higher fiber intake density may be associated with less inflammation, less myocardial hypertrophy, injury, and lower risk of MACE in dialysis patients. These data form an important basis for a randomized controlled trial to examine fiber supplementation on cardiovascular outcomes in the dialysis population.

Prognostic value of acoustic cardiography in patients with chronic heart failure.

BACKGROUND: Rapid risk stratification in patients with heart failure is critically important but challenging. The aim of our study is to ascertain whether acoustic cardiography can identify heart failure (HF) patients at high risk for mortality. METHODS: A total of 474 HF patients were enrolled into our study (76±11years old). Acoustic cardiographic parameters included S3 score (ie, third heart sound exists) and systolic dysfunction index (SDI) (correlated closely with left ventricular systolic dysfunction). The event-free survival curves were plotted by Kaplan-Meier method. Cox regression analysis was used to identify independent predictors for all-cause mortality. RESULTS: During a mean follow-up of 484days, 169 (35.7%) patients died and 126 (26.6%) were due to cardiac causes. After controlling for age, systolic blood pressure, hemoglobin, blood urea nitrogen, albumin, as well as ACEI and beta-blocker treatment in multivariate Cox regression analysis, SDI ≥5 and S3 score ≥4 were both independent predictors for all-cause mortality. Kaplan-Meier analysis showed that HF patients with SDI ≥5 or S3 score ≥4 had a significantly lower survival (52.2% vs. 69.2%, Log-rank χ(2)=18.07, P<0.001; 56.8% vs. 68.6%, Log-rank χ(2)=10.58, P=0.001, respectively) than those with lower SDI or S3 score. CONCLUSIONS: Acoustic cardiography could serve as a cost-effective and time-efficient tool to identify HF patients at high risk for mortality who might benefit from aggressive monitoring and intervention. It may improve assessment and initial disposition decisions in HF management.

Importance of chronotropic response and left ventricular long-axis function for exercise performance in patients with heart failure and preserved ejection fraction.

BACKGROUND: We evaluated the relationship between the degree of chronotropic incompetence and left ventricular (LV) impairment during exercise with severity of exercise intolerance in patients with heart failure and preserved ejection fraction (HFPEF). METHODS: All patients underwent exercise echocardiography during bicycle Ergometer exercise with the acquisition of long-axis tissue Doppler imaging (TDI). Peak heart rates during exercise were also recorded and the percentages of maximal age-predicted heart rate (%MPHR) and heart rate reserve (%HRR) were calculated thereby. Besides, cardiopulmonary exercise testing was performed with peak oxygen consumption (VO2) measuring averaged at the highest 30-second during exercise. RESULTS: Forty HFPEF patients (aged 65±9 years; 75% male) were divided into two groups according to the median of peak VO2: patients with peak VO2<16.5 and ≥16.5 ml/kg/min, respectively. Patients with lower peak VO2 had decreased peak heart rates, %MPHR, %HRR, stroke volume and cardiac indices (LVSI and LVCI) than those with higher peak VO2 (all p<0.05). The LV long-axis functions (TDI Sm, Em, s' and e') were reduced in patients with lower peak VO2 (all p<0.05). Moreover, peak VO2 correlated with the following parameters: peak heart rates, %MPHR, %HRR, LVSI, LVCI, TDI Sm, Em, s' and e'(all p<0.05). CONCLUSIONS: The degree of blunted chronotropic response and impaired LV long-axis function were more profound in HFPEF patients with poor exercise performance.

Factors related to outcome in heart failure with a preserved (or normal) left ventricular ejection fraction.

Heart failure with a preserved ejection faction (HFpEF) is a growing and expensive cause of heart failure (HF) affecting particularly the elderly. It differs in substantial ways in addition to the normal left ventricular ejection fraction, from the more easily recognized form of heart failure with a reduced ejection fraction (HFrEF or 'systolic heart failure') and unlike HFrEF there have been little advances in treatment. In part, this relates to the complexity of the pathophysiology and identifying the correct targets. In HFpEF, there appears to be widespread stiffening of the vasculature and the myocardium affecting ventricular function (both systolic and diastolic), impeding ventricular suction, and thus early diastolic filling leading to breathlessness on exertion and later atrial failure and fibrillation. Left ventricular ejection fraction tends to gradually decline and some evolve into HFrEF. Most patients also have a mixture of several co-morbidities including hypertension, diabetes, obesity, poor renal function, lack of fitness, and often poor social conditions. Therefore, many factors may influence outcome in an individual patient. In this review, the epidemiology, possible causation, pathophysiology, the influence of co-morbidities and some of the many potential predictors of outcome will be considered.

Micro-RNA and mRNA myocardial tissue expression in biopsy specimen from patients with heart failure.

AIMS: There is increasing evidence that changes in microRNA (miRNA) expression occur in chronic heart failure and these may be involved in the pathogenesis. In this study we have explored the expression of selected myocyte and fibroblast-related microRNAs and messenger RNAs (mRNAs) that are associated with hypertrophy, apoptosis and fibrosis in biopsy specimens from patients with relatively new onset heart failure compared to a group of patients without heart failure. METHODS AND RESULTS: Myocardial biopsy specimens taken from Chinese patients presenting with recent heart failure were compared with a group of patients without heart failure undergoing routine cardiac surgery (n=34). miRNAs (miR-1, -21, -23, -29, -30, -130, -133, -195, -199, -208, and -320) and corresponding mRNA expression were measured by real-time quantitative-PCR method. miR-1, -21, -23, -29, -130, -195 and -199 were significantly up-regulated in the heart failure group when compared to those without heart failure (all p<0.01). However, miR-30, -133, -208 and -320 were not significantly different. Related mRNAs (casp3, coll I, coll III and TGF) were also significantly up-regulated (all p<0.05) in the heart failure group. CONCLUSION: Certain selected microRNAs involved in apoptosis, hypertrophy and fibrosis are up-regulated in the myocardium of patients with a clinical history of heart failure compared to those without. These specific miRNAs may be the most suitable for circulating biomarkers in the early stages of chronic heart failure and possibly future therapeutic targets.