132 Original Research [1 4 7#1CHESTJ ANUARY2 0 1 5 ] Endothelial Function in Children With O SA and the Efects of Adenotonsillectomy Kate C. C.Chan ,MB ; Chun T. Au , MPhil ; PingChook,MD ; Dennis L. Y.Lee ,MB ;Hugh S.Lam ,MB ; Yun K.Wing ,MB ; and Albert MartinLi,MD BACKGROUND: Te association between childhood OSA and endothelial function as measured by fow-mediated dilation (FMD) and its response to OSA treatment are uncertain. Te objec-tive of this study was to compare FMD in children with OSA with nonsnoring control subjects and to examine its response to treatment. METHODS: Index cases were children aged 6 to 18 years with habitual snoring and polysom-nography(PSG)-confrmedOSA(obstructiveapneahypopneaindex[OAHI] . 1events/h). Eachcasewaspairedwithanage-,sex-,andBMI-matchednonsnoringcontrolsubject recruitedfromourpreviouscommunitygrowthsurvey.AllsubjectsunderwentFMDmea-surement in the morning afer overnight PSG. Adenotonsillectomy (AT) was ofered to sub-jects who satisfed predefned AT operation criteria. All cases underwent repeat PSG and FMD assessment 6 months later. RESULTS: A total of 63 case-control pairs were recruited. Te OSA group had a signifcantly higherOAHI(median,5.3events/h[interquartilerange(IQR),2.6-11.7]vs0.2events/h [IQR,0-0.5], P, .001)andlowerFMD(mean SD,7.9% 1.3%vs8.3% 0.8%; P5 .04) than the control group. Thirty-two case subjects underwent AT. A significant reduction in OAHI was documented in the AT group (2 8.8 events/h [IQR,2 13.7 to2 4.7];P, .001) accompaniedbyasignifcantincreaseinFMD(10.6%[IQR,0.4-1.4]; P, .001),whichwas not observed in subjects who did not undergo AT. CONCLUSIONS: Children with OSA had reduced FMD, which was reversible with treatment. CHEST 2015; 147(1):132- 139 [ Original ResearchSleep Disorders] Manuscript received June 1, 2014; revision accepted September 8, 2014; originally published Online First October 2, 2014. ABBREVIATIONS: AT 5 adenotonsillectomy;FMD 5 fow-mediated dilation; IQR 5 interquartile range; OAHI 5 obstructive apnea hypopnea index; PSG 5 polysomnography AFFILIATIONS: From the Department of Pediatrics (Drs Chan, Chook, Lam,andLi,andMrAu),theDepartmentofOtorhinolaryngology- HeadandNeckSurgery(DrLee),andtheDepartmentofPsychiatry (Dr Wing), Prince of Wales Hospital, Te Chinese University of Hong Kong, Shatin, Hong Kong . Part of this article was presented in abstract form at the Annual Scientifc Meeting of the Hong Kong College of Pediatricians, December 7, 2013, Hong Kong. FUNDING/SUPPORT: Tis study was supported by the Research Grants CounciloftheHongKongSpecialAdministrativeRegion,China [CUHK4508/06M ]. CORRESPONDENCETO:AlbertMartinLi,MD,Departmentof Pediatrics,PrinceofWalesHospital,TeChineseUniversityofHong Kong, 30-32 Ngan Shing St, Shatin, 000, Hong Kong; e-mail: albertmli@cuhk.edu.hk 2015 AMERICAN COLLEGE OF CHEST PHYSICIANS. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: 10.1378/chest.14-1307 Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 journal.publications.chestnet.org 133 Childhood OSA is a common sleep disorder, afecting around 5% of school-aged children. 1,2 Te condition is characterized by prolonged partial and/or intermittent complete upper airway obstruction. Repeated apneas and hypopneas during sleep result in intermittent hypoxemia and hypercapnia, cortical and sympathetic nervous system arousals, and sleep fragmentation. 2 In adults, OSA is established as an important and indepen-dent risk factor for cardiovascular adverse events. 3,4 However, the mechanisms underlying the association are not well defned. One of the postulated mechanisms is that OSA precipitates and/or accelerates atherosclerosis mediated via endothelial dysfunction. 3 Several studies have reported an association between OSA and high BP in children. 5-8 Intermittent hypoxia and the resulting oxidative stress in OSA can reduce nitric oxide bioavailability. 9 A recent study showed that children with OSA have more proinfammatory monocytes and reduced nitric oxide production in circulating mono-cytes, which are closely linked to endothelial dysfunc-tion. 10 Furthermore, OSA results in sleep fragmentation that causes systemic infammation and sympathetic acti-vation. 9,11,12 All these intermediate processes can lead to endothelial dysfunction, which has been shown to precede the formation of plaque and atherosclerosis. 3,9 Ultrasonographic assessment of endothelium-dependent flow-mediated dilation (FMD) of the brachial artery is the gold standard in assessing endo-thelial function. 13,14 FMD is predominately a result of endothelial nitric oxide release, and FMD of the brachial artery correlates well with both coronary endothelial function and the extent of coronary artery atherosclerosis. 15-17 Te few studies that have investigated the association between childhood OSA and impaired endothelial func-tion have documented positive results. 18-22 In one study that assessed the efects of treatment, endothelial dysfunc-tion improved afer adenotonsillectomy (AT). 18 However, these studies were limited by their methodology, namely unmatched cases and control subjects, and the gold standard of measuring FMD induced by hyperemia was not used. 18-21 A publication that assessed FMD by hyper-emia yielded negative results; the authors failed to docu-ment a signifcant diference in FMD between children with a desaturation index of . 10/h and those with a desaturation index of 10/h. 23 Terefore, whether endo-thelial dysfunction is associated with childhood OSA remains unclear. In this study, we aimed to evaluate (1) FMD in children with OSA compared with normal control subjects and (2) its response to OSA treatment. We hypothesized that children with OSA would have lower FMD when compared with control subjects, and that the impair-ment would improve following AT. Materials and Methods Subjects and Study Design Children aged 6 to 18 years with habitual snoring ( 3 nights per week) were recruited from our sleep disorder clinic. Nonsnoring control sub-jects were recruited from participants in a community growth survey. Written informed consent and assent were obtained from parents and subjects,respectively.Teexclusioncriteriaincludedprevioustreat-mentofOSA;presenceofstructuralheartdisease;medicalhistoryof hypertension,dyslipidemia,diabetesmellitus,geneticsyndrome,cra-niofacial anomalies, congenital or acquired neuromuscular disease, prematurebirth,orintrauterinegrowthretardation;activesmoking; and acute illness within 4 weeks of recruitment. Weight,height,andBPofthesubjectsweremeasuredonthedayof the polysomnography (PSG). BP was measured using the oscillometric method (Datascope Accutorr Plus; Datascope Corp) after sitting for 5 min. Two readings were taken at 5-min intervals on the nondominant armattheheartlevelwiththeproper-sizedcufchosenaccordingto the length and circumference of the arm. Te average of two readings wasusedforanalysis.BMIwasconvertedtoBMIzscoreaccording to local reference. 24 Overweight was defned as a BMI z score . 1.036 (ie, . 85th percentile), and obesity was defned as a BMI z score . 1.645 (ie, . 95th percentile). All subjects underwent overnight PSG. A blood sample for fasting lipid profilewastakenthenextmorning,followedbyFMDevaluation. IndexcasesweresubjectswithhabitualsnoringandPSG-confrmed OSA with an obstructive apnea hypopnea index (OAHI) of . 1 event/h. Each index case was paired with an age-, sex-, and BMI-matched con-trolsubject.Forthecontrolgroup,therewasnoparentalreportof snoring.AllcontrolsubjectshadanormalPSG(OAHI 1event/h) and did not snore on the night of the PSG. All subjects found to have OSA were referred for upper airway assess-mentbyanotorhinolaryngologist.ATwasoferedtothosewithlarge tonsils (Brodsky grading 2) and/or large adenoids (adenoids 25% ofthepostnasalspace,asexaminedbynasoendoscopy). 25Forthose who refused surgical intervention, or in cases in which surgery was not indicated based on predefned criteria, intranasal corticosteroids (Mometasone, 100m g/day for 6 months ) or CPAP therapy were ofered. All OSA cases were invited to have a reassessment 6 months later. Tis study was conducted with the approval of the Joint Chinese University ofHongKong-NewTerritoriesEastClusterClinicalResearchEthics Committee (reference number 2005.356). Polysomnography AnovernightPSGwasperformedoneachsubjectusingtheSiesta ProFusion II PSG monitor (Compumedics Telemed PTY Ltd) to record thefollowingparameters:EEGfromfourleads(C3/A2,C4/A1), bilateral electrooculogram, and electromyogram of mentalis activity andbilateralanteriortibialis.Respiratorymovementsoftheribcage and abdomen were detected by piezo-based efort belts. ECG and heart rate were recorded continuously from two anterior chest leads. Arterial oxyhemoglobinsaturationwasmonitoredbyanoximeter(Ohmeda Biox3900PulseOximeter;Datex-Ohmeda ).Respiratoryairfowpres-suresignalsweremeasuredattheanteriornaresandconnectedtoa pressuretransducer.Anoronasalthermalsensorwasalsousedto Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 134 Original Research [1 4 7#1CHEST J ANUARY2 0 1 5 ] detecttheabsenceofairflow.Snoringwasassessedbyasnoring microphone placed near the throat. Body position was recorded by a body position sensor. 26 Computerized sleep data were edited visually by anexperiencedPSGtechnologistaccordingtostandardizedcriteria. 27 An obstructive apnea was defned as an absence of airflow with persis-tentrespiratoryeffortlastingatleasttwobaselinebreaths, irrespec-tiveofarterialoxygensaturationchanges.Anobstructivehypopnea was defned as a reduction of 50% in the amplitude of the pressure-transducedairfowsignal,withpersistentrespiratoryefortlastingat leasttwobaselinebreaths,accompaniedbyanoxygendesaturationof at least 3% and/or an arousal. Te OAHI was defned as the total number of obstructive apneas and hypopneas per hour of total sleep time. Te oxygendesaturationindexwasdefnedasthetotalnumberofdipsin arterialoxygensaturationofatleast3%perhourofsleep.Arousal wasdefnedasanabruptshifinEEGfrequencyduringsleep,which may have includedu ,a , and/or frequencies . 16 Hz but not spindles, with 3 to 15 s in duration. In rapid eye movement sleep, arousals were scoredonlywhenaccompaniedbyconcurrentincreasesinsubmental electromyogram amplitude. Te arousal index was defned as the total number of arousals per hour of total sleep time. Assessment of Endothelial Function Te assessment was carried out within 3 h afer awakening in a quiet, temperature-controlled room. All subjects abstained from food, including cafeine,foratleast6hbeforethestudy.Tediameterofthebrachial artery was measured on B-mode ultrasound images (1) at rest and (2) in response to reactive hyperemia, which was induced by inflation of a BPcufplacedaroundthelowerarmtoapressureof220mmHgfor 4 to 5 min, followed by rapid defation, using a linear array transducer (L10-5medianfrequency,7.5MHz)andtheAdvancedTechnology Laboratories3000ultrasoundsystem.Tominimizevariability,all measurementsweretakenatenddiastoleidentifedbythe Rwaveon electrocardiogram, and the average of three measurements taken along the vessel was calculated. The full procedure was described in detail previously. 28 FMDwasdefnedasthepercentageincreaseinvesseldiameterfrom baseline in response to reactive hyperemia. In arteries lined by healthy endothelium,increasedfowcausesdilatationofthevesselviarelease oftheendothelium-derivedrelaxingfactorand,therefore,FMDis endotheliumdependent.Hyperemiawascalculatedasthepercentage increaseinbloodfowafercufdefation,comparedwithbaseline.Te wholeprocedureproducedminimaldiscomfortandwaswelltolerated bythechildren. 28-30Allultrasonographicscanswereperformedby thesameinvestigator,whowasblindedtotheidentityandclinical characteristicsofthesubjects.Teaccuracy,reproducibility,andlow interobservererrorforthismeasurementhavebeendemonstrated previously. 28,30-32 Statistical Analysis Datawereexpressedasmean SDfornormallydistributeddataand median(interquartilerange[IQR])fornonnormallydistributeddata, unlessotherwisespecifed.Case-controlcomparisonsweretestedby independentStudentttestsandMann-WhitneyUtestsfornormally distributed and nonnormally distributed data, respectively. Te within-group changes over time for the AT and non-AT groups were tested by pairedStudentttestsandWilcoxonsigned-ranktestsfornormally distributedandnonnormallydistributeddata,respectively.Two-way repeated-measuresanalysisofvariancewasusedtoassessthedifer-ences in the changes in FMD between the AT and non-AT groups. All the statistical analyses were performed with SPSS 13.0 for Windows (IBM), and a two-tailedPvalue, .05 was considered statistically signifcant. Results Endothelial Function in Children With OSA A total of 63 sex-, age-, and BMI-matched case-control pairs were recruited ( Table 1 ). Te mean ages in the OSA and control groups were 10.3 2.9 years and 10.4 2.7 years, respectively. Tirty subjects in each group were overweight or obese. None of the recruited children had hypertension. Tere were no statistically signifcant diferences in the baseline BP and fasting lipid profles between the two groups. Te OSA group had a signifcantly higher OAHI (median, 5.3 events/h [IQR, 2.6-11.7 events/h] vs 0.2 events/h [IQR, 0-0.5 events/h]; P, .001) and lower FMD (mean, 7.9% 1.3% vs 8.3% 0.8%; P5 .04) than the control group ( Table 1 ). Responses to AT Forty-six out of 63 OSA cases were candidates for AT according to the predefned criteria, of whom 32 agreed to receive the operation. Of the 31 children who did not have AT, none agreed to use CPAP, 10 received 6-month intranasal corticosteroids therapy, and the remaining 21 subjects refused any treatment. Comparisons between groups are shown inTable 2 . Children who underwent surgical intervention had signifcantly higher OAHI at baseline when compared with those who did not undergo AT (10.5 events/h [IQR, 6.0-15.4 events/h] vs 3.3 events/h [IQR, 1.8-4.1 events/h]; P, .001) ( Table 2 ). No signifcant diferences in age, BMI, BP, FMD, hyper-emia, or fasting lipid profles were found between groups at baseline. At baseline, those who received nasal spray (n5 10) had milder disease than did those who had AT (n5 32) (2.1 events/h [IQR, 1.6-3.5 events/h] vs 10.5 events/h [IQR, 6.0-15.4 events/h]; P, .001), and there was no signifcant diference in FMD (8.3% 1.1% vs 7.7% 1.5%;P5 .26). In the group that had undergone AT, a signifcant reduction in OAHI (2 8.8 events/h [IQR, 2 13.7 to 2 4.7 events/h]; P, .001) and a signifcant decrease in arousal index ( 2 6 events/h [2 15.0 to 0.5 events/h]; P, .001) were observed. Te amount of time in stage 1 also signifcantly decreased afer AT (2 0.6% [ 2 3.3% to 0.2%]; P5 .01). Te improvements were accompanied by a signifcant increase in FMD (1 0.6% [IQR, 0.4% to 1.4%]; P, .001). Tere was a signifcant diference in the change in FMD between the two groups even afer adjustment for age, sex, BMI z score, and brachial artery diameter at rest ( P, .001). Moreover, FMD afer AT was also found to be comparable with that of the non-OSA control sub-jects (8.5% 1.2% vs 8.3% 0.8%;P5 .26). On the Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 journal.publications.chestnet.org 135 other hand, treatment with nasal spray did not result in signifcant improvement in OAHI and FMD (baseline OAHI, 2.1 events/h [IQR, 1.6-3.5 events/h]; posttreat-ment OAHI, 2.2 events/h [IQR, 1.6-7.0 events/h], P5 .72; baseline FMD, 8.3% 1.1%; post-treatment FMD, 8.4% 0.6%; P5 .59). Discussion Our study showed that children with OSA had impaired endothelial function, as demonstrated by their signif-cantly lower FMD compared with nonsnoring control subjects, and in those with moderate to severe OSA, AT was able to reverse the impaired endothelial func-tion. Tese results provided robust evidence to support an association between childhood OSA and endothelial dysfunction. Our results were consistent with those of previous studies showing impaired endothelial function in chil-dren with OSA, although the methods adopted for assessing endothelial function were diferent. 18-22 Most of the previous studies measured the degree of vessel dilation in response to reactive hyperemia. Te associa-tion between reactive hyperemia and FMD was not linear and was dependent on several between-subject factors, including structural and genetic diferences. 33 Children with OSA in our study had signifcantly reduced FMD compared with the control subjects, but signifcant diferences in hyperemia between the groups were not observed. Similarly, a signifcant increase in FMD in the AT group was not accompanied by a signif-icant increase in hyperemia. In other words, efective treatment of OSA could increase the FMD response independent of the change in the magnitude of reactive hyperemia. Terefore, measuring reactive hyperemia solely, as carried out in previous studies, would not provide an accurate assessment of endothelial function. Although these studies demonstrated the blunted post-occlusive hyperemia in children with OSA and its revers-ibility with AT, our results further consolidated the association between childhood OSA and endothelial TABLE 1] Comparisons Between OSA Case Subjects and Control Subjects Characteristic Case Subjects (n 5 63) Control Subjects (n 5 63) PValueAge , y 10.3 2.9 10.4 2.7 .71Male, No. (%) 41 (65.1) 41 (65.1) 1.00Overweight, No. (%) 30 (47.6) 30 (47.6) 1.00BMI, kg/m 220.6 5.9 20.3 5.8 .72BMIz score 0.89 1.26 0.84 1.09 .81Sleep e ciency, % 88.8 8.6 85.7 9.8 .07Total sleep time, min 492 60 473 75 .12REM sleep, % 21.5 (18.8-25.0) 19.5 (16.6-21.9) .01Stage 1 sleep, % 6.0 (3.3-7.9) 4.0 (1.2-7.3) .02Stage 2 sleep, % 38.3 (32.7-41.9) 40.5 (34.3-45.8) .09Slow-wave sleep, % 33.5 (29.0-40.8) 34.4 (27.3-41.7) .70OAHI, events/h 5.3 (2.6-11.7) 0.2 (0-0.5), .001ODI, events/h 3.7 (1.1-11.0) 0.5 (0.1-1.0), .001Sp O 2 nadir, % 90 (85-93) 94 (93- 95) , .001Arousal index, events/h 16.4 (12.9-21.8) 9.5 (6.6-13.2) , .001Systolic BP, mm Hg 108 12 107 11 .61Diastolic BP, mm Hg 65 9 64 10 .43FMD, % 7.9 1.3 8.3 0.8 .04Hyperemia, % 508 141 527 116 .41TC, mM 4.3 0.6 4.3 0.8 .85LDL, mM 2.3 0.5 2.3 0.7 .86HDL, mM 1.5 0.4 1.6 0.4 .33TG, mM 1.2 0.8 1.0 0.5 .11 Data are presented as mean SD or median (interquartile range) unless indicated otherwise. FMD 5 ow-mediated dilation; HDL 5 high-density lipoprotein; LDL 5 low-density lipoprotein; OAHI 5 obstructive apnea hypopnea index; ODI 5 oxygen desaturation index; REM 5rapid eye movement; Sp O 2 5 oxygen saturation; TC 5 total cholesterol; TG 5 triglyceride. Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 136 Original Research [1 4 7#1CHEST J ANUARY2 0 1 5 ] TABLE 2 ] Comparisons Between AT and Non-AT Groups at Baseline and Reassessment CharacteristicAT (n532)Non-AT (n531)BaselineReassessment P Value a

BaselineReassessment P Value a

Age, y9.92.610.72.6 ,.00110.63.111.32.9 ,.001Male, No. (%)21 (65.6)20 (64.5)BMI, kg/m2 19.65.320.55.5.00321.86.321.86.1.78BMI z

score0.711.200.851.15.061.091.301.011.33.14Sleep e ciency, %89.28.390.58.1.4688.39.089.29.0.62Total sleep time, min5064750260.724786849264.29REM sleep, %19.2 (16.2-22.2)21.6 (18.0-25.1).0720.2 (17.2-21.9)22.2 (18.5-23.9).03Stage 1 sleep, %3.7 (1.0- 6.7)2.4 (1.1-3.5).014.0 (1.8-9.1)4.5 (1.5-8.5).66Stage 2 sleep, %42.7 (36.5-46.6)39.5 (34.3-43.3).0838.1 (33.7-45.5)39.3 (33.1-44.5).72Slow-wave sleep, %33.8 (27.2-40.0)37.1 (31.2-42.4).1835.5 (27.3-43.4)32.0 (26.5-40.8).19OAHI, events/h10.5 (6.0-15.4) b

0.6 (0.2-1.6) ,.0013.3 (1.8-4.1) b

3.5 (1.7-8.0).85ODI, events /h8.0 (2.4-13.5) b

0.4 (0.1-0.9) ,.0011.7 (0.8-4.7) b

2.2 (1.1-5.5).89Sp O2

nadir, %89 (83-94)94 (92-95) ,.00191 (86-93)91 (87-94).12Arousal index, events /h18.0 (14.8-25.7) b

13.1 (8.8-17.6) ,.00114.4 (12.5-18.1) b

15.8 (12.5-20.9).38Systolic BP, mm Hg1071111115.221091210816.58Diastolic BP, mm Hg659659.806510639.25FMD, %7.71.58.51.2 ,.0018.11.18.10.9.71Hyperemia, %496148513119.57521134537133.71TC, mM4.10.64.20.9.534.50.54.50.7.68LDL, mM2.20.52.20.7.822.40.52.40.5.34HDL, mM1.50.41.50.4.981.60.41.60.5.51TG, mM1.20.91.30.7.351.20.71.10.5.87 Data are presented as meanSD for normally distributed data and median (interquartile range) for nonnormally distributed data, unless indicated otherwise. See Table 1

legend for expansion of abbreviations. aP

values were obtained from paired t

tests and Wilcoxon signed rank tests for normally distributed and nonnormally distributed data, respectively. b Signi cant di erences between the pretreatment data of the two groups. Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 journal.publications.chestnet.org 137 dysfunction with the use of the gold standard FMD measurement. Te evidence linking childhood OSA to cardiovascular morbidities has been growing, and endothelial dysfunc-tion has been consistently found to be associated with OSA. 34 Te mechanisms leading to cardiovascular mor-bidities in patients with OSA were likely secondary to the complications associated with sleep-disordered breathing, namely disrupted sleep architecture, inter-mittent hypoxia, oxidative stress, hypertension, systemic infammation, maladaptive autonomic responses, meta-bolic imbalance, and so forth. 5,9,34-37 All these factors interacted and may also have placed an adverse infuence on endothelial function, platelets, and infammatory cells. 37,38 Endothelial homeostasis played a signifcant role in determining the risk of a cardiovascular event. With disturbances to the normal homeostasis, the endo-thelium would lose its ability to prevent abnormal vaso-constriction, platelet aggregation, and smooth muscle cell proliferation. 39 Endothelial dysfunction has indeed been shown to be an early risk marker preceding the process of atherogenesis. 3,4 FMD was an endothelial nitric oxide synthase-dependent response. Basically, the ultrasonographic assessment of FMD was a measure-ment of the compliance and the capacity of the blood vessels to respond to an increase in blood fow by vaso-dilation. 13 In adults, abnormal FMD predicted cardio-vascular morbidity independently of traditional cardiovascular risk factors, and interventions to reduce cardiovascular risk were accompanied by a parallel improvement in FMD. 14,40 FMD of the brachial artery correlated well with coronary artery endothelial func-tion, the impairment of which preceded the develop-ment of overt atherosclerosis and coronary artery disease. 16 Tus, FMD of the brachial artery was a useful surrogate refecting the risk of coronary artery athero-sclerosis. 17 Tis has signifed the importance of our fnd-ings that reduced FMD of brachial artery in the subjects with OSA might predict future cardiovascular morbid-ities and timely intervention might be able to reverse this risk factor. AT remains the frst-line treatment of children with moderate to severe OSA, yet the therapy for mild OSA is still controversial. In our study, only children with more severe disease underwent surgical treatment. Nevertheless, a reversal of endothelial dysfunction follow-ing surgical intervention was demonstrated. For those who had mild disease and refused surgery, a course of nasal spray corticosteroids was shown not to be efective in reversing endothelial dysfunction. All children with OSA were assessed by an otorhinolaryngologist, and those who satisfed the predefned criteria for surgery were ofered such intervention; however, only the par-ents of those with moderate to severe disease accepted surgery. Despite a thorough explanation of the possible complications associated with childhood OSA and our experience of likely progression to more severe disease in children with mild OSA and enlarged tonsils, most of the parents of children with mild disease refused surgery. 41 Tis was not unexpected because the accep-tance of surgery among Chinese parents is generally low. Similar to the experience of other centers, the accep-tance of and compliance with home noninvasive venti-lation is also low in our locality; in fact, none of the subjects in our cohort with moderate to severe disease received noninvasive ventilation. Hyperlipidemia and obesity are both important, independent determining factors of endothelial dysfunction. 22,28,29,42 Terefore, BMI-matched control subjects were recruited in this study to minimize the cofounding efects of body weight. Moreover, there were no signifcant diferences in fasting lipid profle between the OSA and control groups or between children who received AT and those who did not. In the AT group, the fasting lipid profles were similar before and afer AT. Tis further supports the fnding that the impairment of endothelial function and its reversibility were associated with OSA and AT, respectively. Our study had several limitations. FMD has been dem-onstrated not to be entirely nitric oxide mediated, but afected by a complex interplay of vasodilator and vasoconstrictor stimuli. 43,44 Factors including other vasoactive substances, wall shear stress, and the activity of the sympathetic nervous system also con-tribute to the total FMD response. 43,44 Potential biased interference on endothelial function by the scaling properties of the FMD index is another concern. 45 However, to date, the methodology adopted in our study remains the most widely accepted standard assess-ment of endothelial function. We successfully demon-strated the association between childhood OSA and endothelial dysfunction; however, whether this would contribute to the development of cardiovascular com-plications in adulthood remained uncertain. However, it has been found that the disease process of atheroscle-rosis could begin as early as the frst decade of life. 14,28,29 Impairment of the endothelial function in early life may initiate abnormal reactions among the vessel wall, platelets, neutrophils, and macrophages, and may result in atherogenesis. Further prospective study Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 138 Original Research [1 4 7#1CHEST J ANUARY2 0 1 5 ] would be needed to clarify the long-term implications of endothelial dysfunction in children. In our study, only selected subjects in the cohort underwent AT; therefore, our results were more applicable to those with moderate to severe disease. Further studies are needed to confrm whether a reversal of impaired endothelial function could also be seen with treatment in children with milder disease. Conclusions Our study supported an association between childhood OSA and endothelial dysfunction as refected by reduced FMD. Te impairment in the endothelial func-tion was reversible with AT. Tese results suggest that childhood OSA exerts a signifcant adverse efect on the endothelial function and therefore, early diagnosis of childhood OSA and timely intervention are vital. Acknowledgments Author contributions: Dr Chan is the guarantor of the manuscript and takes responsibility for the integrity of the data and the accuracy of the data analysis. K. C. C. C., C. T. A., and A. M. L. contributed to the project planning and data interpretation; C. T. A. contributed to the interpretation of PSG; K. C. C. C. and A. M. L. contributed to the recruitment of subjects; C. T. A. contributed to the data analysis; K. C. C. C. contributed to the preparation of the manuscript; P. C. contributed to the assessment of endothelial function; D. L. Y. L. contributed to the upper airway assessment and surgical treatment of the subjects; and K. C. C. C., C. T. A., P. C., D. L. Y. L., H. S. L., Y. K. W., and A. M. L. contributed to the revision and approval of the fnal manuscript . Financial/nonfnancial disclosures:Te authors have reported to CHEST the following conficts of interest: Dr Wing has received honoraria for serving as a part-time consultant for Renascence Terapeutics Ltd. Drs Chan, Chook, Lee, Lam, and Li and Mr Au have reported that no potential conficts of interest exist with any companies/organizations whose products or services may be discussed in this article . Role of sponsors: Te design, execution, data collection, and analysis of the study were carried out solely by the research team without the involvement of the funding body. Other contributions: This work was performed at the Department of Pediatrics, Prince of Wales Hospital, Te Chinese University of Hong Kong, Shatin, Hong Kong. We are grateful for the cooperation and participation of all the children and their parents. References 1 . LiAM ,SoHK ,AuCT ,et al .Epidemiology of obstructive sleep apnoea syndrome in Chinese children: a two-phase community study .Torax .2010 ; 65 ( 11 ): 991 - 997 . 2 . MarcusCL ,BrooksLJ ,DraperKA ,et al ; American Academy of Pediatrics . Diagnosis and management of childhood obstructive sleep apnea syndrome . Pediatrics .2012 ; 130 ( 3 ): 576 - 584 . 3 . KatoM ,AdachiT ,KoshinoY ,Somers VK .Obstructive sleep apnea and cardio-vascular disease.Circ J .2009 ; 73 ( 8 ): 1363 - 1370 . 4 . Snchez-de-la-TorreM ,Campos-RodriguezF ,BarbF .Obstructive sleep apnoea and cardiovascular disease. Lancet Respir Med .2013 ; 1 ( 1 ): 61 - 72 . 5 . LiAM ,AuCT,SungRY,et al .Ambulatory blood pressure in children with obstruc-tive sleep apnoea: a community based study. Torax .2008 ; 63 ( 9 ): 803 - 809 . 6 . AminR ,SomersVK ,McConnellK ,et al . Activity-adjusted 24-hour ambulatory blood pressure and cardiac remodeling in children with sleep disordered breathing. Hypertension .2008 ; 51 ( 1 ): 84 - 91 . 7 . BixlerEO ,VgontzasAN ,LinHM ,et al . Blood pressure associated with sleep-disordered breathing in a population sample of children.Hypertension .2008 ; 52 ( 5 ): 841 - 846 . 8 . NisbetLC ,YiallourouSR ,BiggsSN ,et al . Preschool children with obstructive sleep apnea: the beginnings of elevated blood pressure? Sleep .2013 ; 36 ( 8 ): 1219 - 1226 . 9 . JelicS ,PadelettiM ,KawutSM ,et al . Infammation, oxidative stress, and repair capacity of the vascular endothelium in obstructive sleep apnea.Circulation . 2008 ; 117 ( 17 ): 2270 - 2278 .10 . Kheirandish-GozalL ,WangY ,Duggan RC ,et al .Nitric oxide production by monocytes in children with OSA and endothelial dysfunction.Clin Sci (Lond) . 2014 ; 127 ( 5 ): 323 - 330 .11 . HakimF,GozalD ,Kheirandish-GozalL . Sympathetic and catecholaminergic alterations in sleep apnea with particular emphasis on children . Front Neurol . 2012 ; 3 .12 . KimJ ,HakimF ,Kheirandish-GozalL , GozalD .Infammatory pathways in chil-dren with insuf cient or disordered sleep . Respir Physiol Neurobiol .2011 ; 178 ( 3 ): 465 - 474 .13 . CorrettiMC ,AndersonTJ ,Benjamin EJ ,et al ;International Brachial Artery Reactivity Task Force. Guidelines for the ultrasound assessment of endothelial-dependent fow-mediated vasodilation of the brachial artery: a report of the International Brachial Artery Reactivity Task Force.J Am Coll Cardiol .2002 ; 39 ( 2 ): 257 - 265 .14 . UrbinaEM ,WilliamsRV ,AlpertBS , et al ;American Heart Association Atherosclerosis, Hypertension, and Obesity in Youth Committee of the Council on Cardiovascular Disease in the Young. Noninvasive assessment of subclinical atherosclerosis in children and adolescents: recommendations for standard assessment for clinical research: a scientifc statement from the American Heart Association.Hypertension .2009 ; 54 ( 5 ): 919 - 950 .15 . JoannidesR ,HaefeliWE ,LinderL ,et al . Nitric oxide is responsible for fow-dependent dilatation of human periph-eral conduit arteries in vivo.Circulation . 1995 ; 91 ( 5 ): 1314 - 1319 .16 . AndersonTJ ,UehataA ,GerhardMD , et al .Close relation of endothelial func-tion in the human coronary and periph-eral circulations.J Am Coll Cardiol .1995 ; 26 ( 5 ): 1235 - 1241 .17 . SchroederS ,EnderleMD ,OssenR , et al .Noninvasive determination of endothelium-mediated vasodilation as a screening test for coronary artery disease: pilot study to assess the predictive value in comparison with angina pectoris, exercise electrocardiography, and myo-cardial perfusion imaging.Am Heart J . 1999 ; 138 ( 4 pt 1 ): 731 - 739 .18 . GozalD ,Kheirandish-GozalL ,SerperoLD , Sans CapdevilaO ,DayyatE .Obstructive sleep apnea and endothelial function in school-aged nonobese children: efect of adenotonsillectomy.Circulation .2007 ; 116 ( 20 ): 2307 - 2314 .19 . Kheirandish-GozalL ,BhattacharjeeR , KimJ ,ClairHB ,GozalD .Endothelial progenitor cells and vascular dysfunc-tion in children with obstructive sleep apnea.Am J Respir Crit Care Med . 2010 ; 182 ( 1 ): 92 - 97 .20 . KimJ ,BhattacharjeeR ,SnowAB , CapdevilaOS ,Kheirandish-GozalL , GozalD .Myeloid-related protein 8/14 levels in children with obstructive sleep apnoea. Eur Respir J .2010 ; 35 ( 4 ): 843 - 850 . 21 . BrunettiL ,FrancavillaR ,ScicchitanoP , et al. Impact of sleep respiratory disor-ders on endothelial function in children. Scientifc World Journal.2013 :719456. 22 . BhattacharjeeR ,KimJ ,AlotaibiWH , Kheirandish-GozalL ,CapdevilaOS , GozalD .Endothelial dysfunction in children without hypertension: potential contributions of obesity and obstructive sleep apnea.Chest .2012 ; 141 ( 3 ): 682 - 691 .23 . DubernB ,AggounY ,BoulM ,FaurouxB , BonnetD ,TounianP .Arterial alterations in severely obese children with obstruc-tive sleep apnoea . Int J Pediatr Obes . 2010 ; 5 ( 3 ): 230 - 236 .24 . LeungSS ,ColeTJ ,TseLY ,LauJT .Body mass index reference curves for Chinese children.Ann Hum Biol .1998 ; 25 ( 2 ): 169 - 174 .Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015 journal.publications.chestnet.org 139 25 . NgSK ,LeeDL ,LiAM ,WingYK ,Tong MC .Reproducibility of clinical grading of tonsillar size .Arch Otolaryngol Head Neck Surg .2010 ; 136 ( 2 ): 159 - 162 .26 . LiAM ,NgC ,NgSK ,et al .Adipokines in children with obstructive sleep apnea and the efects of treatment . Chest . 2010 ; 137 ( 3 ): 529 - 535 .27 . IberC ,Ancoli-IsraelS ,ChessonALJr , QuanSF ; for theAmerican Academy of Sleep Medicine .Te AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifcations . 1st ed.Westchester, IL :American Academy of Sleep Medicine ;2007 . 28 . CelermajerDS ,SorensenKE ,Gooch VM ,et al .Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis.Lancet . 1992 ; 340 ( 8828 ): 1111 - 1115 .29 . WooKS ,ChookP,YuCW,et al . Overweight in children is associated with arterial endothelial dysfunction and intima-media thickening . Int J Obes Relat Metab Disord .2004 ; 28 ( 7 ): 852 - 857 .30 . WooKS ,ChookP ,YuCW ,et al .Efects of diet and exercise on obesity-related vascular dysfunction in children . Circulation .2004 ; 109 ( 16 ): 1981 - 1986 .31 . SorensenKE ,CelermajerDS , SpiegelhalterDJ ,et al .Non-invasive measurement of human endothelium dependent arterial responses: accuracy and reproducibility.Br Heart J .1995 ; 74 ( 3 ): 247 - 253 .32 . WooKS ,ChookP ,LolinYI ,Sanderson JE ,MetreweliC ,CelermajerDS .Folic acid improves arterial endothelial function in adults with hyperhomocys-tinemia.J Am Coll Cardiol .1999 ; 34 ( 7 ): 2002 - 2006 .33 . WidlanskyME .Shear stress and fow-mediated dilation: all shear responses are not created equally.Am J Physiol Heart Circ Physiol .2009 ; 296 ( 1 ): H31 - H32 .34 . Kheirandish-GozalL ,BhattacharjeeR , GozalD .Autonomic alterations and endothelial dysfunction in pediatric obstructive sleep apnea.Sleep Med . 2010 ; 11 ( 7 ): 714 - 720 .35 . ChanJY ,LiAM ,AuCT ,et al .Cardiac remodelling and dysfunction in children with obstructive sleep apnoea: a commu-nity based study.Torax .2009 ; 64 ( 3 ): 233 - 239 .36 . Kheirandish-GozalL ,CapdevilaOS , TaumanR ,GozalD .Plasma C-reactive protein in nonobese children with obstructive sleep apnea before and afer adenotonsillectomy.J Clin Sleep Med . 2006 ; 2 ( 3 ): 301 - 304 . 37 . GozalD ,Kheirandish-GozalL . Cardiovascular morbidity in obstructive sleep apnea: oxidative stress, infamma-tion, and much more. Am J Respir Crit Care Med .2008 ; 177 ( 4 ): 369 - 375 .38 . KimJ ,BhattacharjeeR ,Kheirandish-GozalL ,SpruytK ,GozalD .Circulating microparticles in children with sleep disordered breathing . Chest .2011 ; 140 ( 2 ): 408 - 417 .39 . DeanfeldJE ,HalcoxJP ,RabelinkTJ . Endothelial function and dysfunction: testing and clinical relevance.Circulation . 2007 ; 115 ( 10 ): 1285 - 1295 . 40 . ChanSY ,ManciniGB ,KuramotoL , SchulzerM ,FrohlichJ ,IgnaszewskiA . Te prognostic importance of endothe-lial dysfunction and carotid atheroma burden in patients with coronary artery disease.J Am Coll Cardiol .2003 ; 42 ( 6 ): 1037 - 1043 .41 . LiAM ,AuCT ,NgSK ,et al .Natural history and predictors for progression of mild childhood obstructive sleep apnoea. Torax .2010 ; 65 ( 1 ): 27 - 31 .42 . MartinoF ,LofredoL ,CarnevaleR ,et al . Oxidative stress is associated with arterial dysfunction and enhanced intima-media thickness in children with hypercholesterol-emia: the potential role of nicotinamide-adenine dinucleotide phosphate oxidase . Pediatrics .2008 ; 122 ( 3 ): e648 - e655 .43 . RenemanRS ,ArtsT ,HoeksAP .Wall shear stressan important determi-nant of endothelial cell function and structurein the arterial system in vivo. Discrepancies with theory.J Vasc Res . 2006 ; 43 ( 3 ): 251 - 269 .44 . GreenDJ ,DawsonEA ,GroenewoudHM , JonesH ,TijssenDH .Is fow-mediated dilation nitric oxide mediated?: A meta-analysis.Hypertension .2014 ; 63 ( 2 ): 376 - 382 .45 . AtkinsonG ,BatterhamAM .Te per-centage fow-mediated dilation index: a large-sample investigation of its appro-priateness, potential for bias and causal nexus in vascular medicine.Vasc Med . 2013 ; 18 ( 6 ): 354 - 365 .Downloaded From: http://journal.publications.chestnet.org/ by a Addenbrookes Hospital Useron 08/13/2015