專科醫師陳光裕的部落格

真實世界的證據與量測因子以估算成人阻塞型睡眠呼吸中止症嚴重程度
Kuang -Yu Chen, Ming-Hui Hung, Kuo-Chin Chiu, Wei-Chun Lin, Fung-J Lin
Pulmonary and Critical Care Medicine, Poh-Ai Hosp, Yi-Lung County, Taiwan

Purpose:  
  Polysomnography (PSG) is still the current gold-standard for diagnosis of obstructive sleep apnea (OSA) but measurement of inspiratory flow limitation (IFL) may be associated to asthma or COPD. Therefore, pulmonary function test (PFT) is still recommended when a patient is supposed to be moderate to severe OSA. Flow-volume curve offered prediction of treatment outcome while mandibular advancement splint was considered (Biao Zeng et al, AJRCCM 2007;175:726-730) but, real-world evidence obtained from PFT to predict the apnea-hypopnea index (AHI) as the OSA severity is warranted.

Materials and methods: 
  A case-series study was performed on OSA patients (n=50) received PSG during 2015 Jan 1 to 2017 Sep 30. Those without PFT (n=3) or the time of their reports between PSG and PFT longer than 24 months (n=2) were both excluded. Body mass index, expiratory flow rate at 50% of vital capacity (i.e. MEF50), inspiratory flow rate at 50% of vital capacity (i.e. MIF50), maximal mid-expiratory flow (MMEF), Functional Residual Capacity were all recorded for statistical analysis. One-way ANOVA and logistic regression were for AHI, MMEF, MEF50:MIF50 ratio (MEI50R). ROC curves of AHI with AUC on neck circumference (NC), age and MEI50R were plotted. P-value less than 0.05 was associated with statistical significance. Standard deviation was shown as (SD) below.

Results: 
  N=45, Male: Female=28:17, mean Age=55.3 (15.6) years-old, BMI=27.9 (5.7) and NC=37.1 (4.0) cm. Mean FEV1% equaled to 93% (14%), MMEF= 78% (20%), RV%= 102% (19%), AHI=24.1 (3.4) and MEI50R equaled to 0.77 (0.3). There were statistical significances (p-value < 0.05) while 4 percentile subgroups of MEI50R were calculated in student’s t tests to NC>=38 cm and BMI>= 28. Compared with the 1st to 4th percentile MMEF revealed significantly difference, their mean MEI50R were 0.55 (0.26) and 0.99 (0.27). When ROC curve for AHI were plotted for different factors, the area under curve (AUC) for NC >=38 together with MEI50R < 0.8 revealed 0.773 (0.117) (95%CI 0.540 - 1.000) (p-value=0.035). For moderate OSA, combined MEI50R from PFT and NC offered PSG AHI (7.2) pre-test 81% and 80% and AHI (18.4) 64% and 100% for sensitivity and specificity.

Tables and figures

Figure 1a. Logistic regression analysis for MMEF by standardized MEI50R. Black dot  and the white circle  each represents case of NC ≥  38 cm and NC< 38 cm.

Figure 1b to 1e. ROC analysis for different measurement item to the detection of apnea-hypopnea index (AHI)(red line) and desaturation indext (DI)(black line). 1b: BMI ≥  28;  1c: MEI50R < 0.8;  1d: NC ≥  38 cm;  1e: MEI50R < 0.8 plus NC ≥  38 cm.

Table 4. Three logistic regression model of MEI50R by MEI50R < 0.8 plus NC ≥  38 cm wth BMI, age  and MMEF as constants. Model 3 R squared (R2) equaled to 0.648, SE 0.1160 and df1= 1, df2=7.

• Figure 1b, AUC (mean (SD)) in ROC by BMI ≥ 28: AHI: 0.672 (0.086) (95%CI 0.503, 0.840) (p=0.053); DI: 0.691 (0.084) (95%CI 0.526, 0.857) (p=0.031) 
• Figure 1c, AUC (mean (SD)) in ROC by MEI50R < 0.8: AHI 0.625 (0.092) (95%CI 0.444, 0.806) (p=0.153); DI: 0.595 (0.092) (95%CI 0.414, 0.776) (p=0.278) 
• Figure 1d, AUC (mean (SD)) in ROC by NC ≥ 38 cm: AHI: 0.707 (0.081) (95%CI 0.548, 0.865) (p=0.021); DI: 0.738 (0.077) (95%CI 0.586, 0.890) (p=0.008)
• Figure 1e, AUC (mean (SD)) in ROC by MEI50R < 0.8 plus NC ≥ 38 cm: AHI 0.773 (0.117) (95%CI 0.540, 1.000) (p=0.035); DI: 0.786 (0.112) (95%CI 0.561, 1.000) (p=0.027)
• AUC (mean (SD)) in ROC by BMI ≥ 28 plus NC ≥ 38 cm: AHI 0.780 (0.092) (95%CI 0.600, 0.960) (p=0.016); DI: 0.798 (0.088) (95%CI 0.625, 0.970) (p=0.010)(no figure in this abstract)
• While the measurement by MEI50R < 0.8 plus NC ≥ 38 cm as a cut-off level to detect AHI equaled to 7.2, the sensitivity  would be 81% and the specificity would be 80%. The same measurement for detect AHI equaled to 18.4, the sensitivity would be 64% and specificity would be 100%. Comparing to measurement by BMI ≥ 28 plus NC ≥ 38 cm as a clinical cut-off criteria to dectect AHI equaled to 16.8, the sensitivity  would be 83% and the specificity would be 57%, and when it was applied to detect AHI equaled to 29.0, the sensitivity would be 67% and specificity would be 86%. 
• Pulmonary function test and the measurement of each FEV1, FVC or MMEF offered little in clinical sleep disorder survey, however, MEF50:MIF50 ratio (MEI50R) had been studied to be correlated to mandibular advancement splint treatment benefit in sleep apnea patient. According to recent evidence of inspiratory flow limitation which might  develop in asthma or COPD at sleep apnea patient, decision of either nasal CPAP or mandibular advancement splint still depend on how the pulmonary function would be proper or not. Therefore, this clinical cohort study offered some real-world evidence for our colleagues to predict a moderate OSA severity. 
• From the point of view from pulmonary physiology, MMEF or MEF50 might be associated to small airways’ function in distal bronchioles with diameter of 2mm in size to balance the forces from both lung parenchyma (dilatation force) and airway smooth muscles (constriction force). Inspiration flow limitation, if it could be represented by MIF50, our measurement of MEI50R might off a proper estimation index for individual pulmonary recoverability to upper airway collapse during nocturnal sleeping stages. Our cohort study revealed a lower MEI50R (i.e. < 0.8) associated with a larger NC (i.e.  ≥ 38 cm) would be sensitive enough to detect a moderate obstructive sleep apnea severity without BMI or age differences. But if BMI and NC both elevated and closed to very obese, the severity of AHI would be too high to applied an mandibular advancement splint for success of treatment. 
• Be awareness of pre-test pulmonary function for PSG study, especially introduction of MEF50:MIF50 ratio (MEI50R), and further study on real-world response from treatment model selection would be continued.

  Combined MEF50:MIF50 ratio in flow-volume curve and neck circumference as real-world measurement for severity evaluation in adult obstructive sleep apnea is adequate for both diagnosis and treatment.

References:

1. Sushmita Pamidi et al on behalf of the American Thoracic Society Ad Hoc Committee on Inspiratory Flow Limitation (THIS OFFICIAL WORKSHOP REPORT OF THE AMERICAN THORACIC SOCIETY (ATS) WAS APPROVED BY THE ATS BOARD OF DIRECTORS, MARCH 2017) An Official American Thoracic Society Workshop Report: Noninvasive Identification of Inspiratory Flow Limitation in Sleep Studies Ann Am Thorac Soc 2017 Jul;14(7):1076-1085
2. Biao Zeng et al Am J Respir Crit Care Med 2007;175:726-730 Use of Flow–Volume Curves to Predict Oral Appliance Treatment Outcome in Obstructive Sleep Apnea
3. Whitney Mostafiz et al Chest 2011;139(6):1331-1339 Influence of Oral and Craniofacial Dimensions on Mandibular Advancement Splint Treatment Outcome in Patients With Obstructive Sleep Apnea
4. Daniel L. Stadler et al Abdominal Compression Increases Upper Airway Collapsibility During Sleep in Obese Male Obstructive Sleep Apnea Patients Sleep 2009;32(12):1579-1587