Health & Fitness

7 pages
4 views

Insulin and obstructive sleep apnea in obese Chinese children

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
ObjectiveIn adults, obstructive sleep apnea (OSA) is associated with insulin resistance and dyslipidemia. We aimed to establish correlation between OSA, serum lipid profile, and insulin levels in obese snoring children.In adults, obstructive sleep
Transcript
  Pediatric Pulmonology 41:1175–1181 (2006) Insulin and Obstructive Sleep Apnea inObese Chinese Children Albert M. Li,  MB , 1 * Michael H.M. Chan,  MB , 2 Dorothy F.Y. Chan,  MB , 1 Hugh S. Lam,  MB , 1 Eric M.C. Wong,  MA , 3 H.K. So,  PhD , 1 Iris H.S. Chan,  PhD , 2 Christopher W.K. Lam,  MD , 2 and Edmund A.S. Nelson,  MD 1 Objective: In adults, obstructive sleep apnea (OSA) is associated with insulin resistance anddyslipidemia.WeaimedtoestablishcorrelationbetweenOSA,serumlipidprofile,andinsulinlevelsin obese snoring children.Methods: Consecutive obese children with habitual snoring were recruited. They underwentphysical examination, overnight polysomnography (PSG), and metabolic studies. OSA wasdiagnosed if apnea hypopnea index (AHI) > 1.0, and caseswere considered to have moderate tosevere OSA if AHI > 10.Results: Ninety-four obese subjects with habitual snoring were studied. Seventy-three subjectswere male and the median age of the studied group was 12.0 years (IQR 9.7–13.9). None of thesubjects had active cardiopulmonary disease, and the BMI values of our subjects were  > 95thpercentile using local reference charts. Sixty subjects had OSA, 47 being mild, and 13 beingmoderate to severe OSA. Multiple logistic regression analysis revealed that saturation nadir andinsulin levels were significantly associated with OSA.Conclusion: OSA is prevalent among obese children with habitual snoring and insulin isindependently associated with the condition. Its role in the cardiovascular complications ofchildhood sleep apnea is worthy of further exploration.  Pediatr Pulmonol. 2006; 41:1175–1181.   2006 Wiley-Liss, Inc. Key words: obstructive sleep apnea; insulin; lipids; child. INTRODUCTION A global increase in the prevalence of obesity inchildren has been noted. 1 A survey of 25,000 Hong Kongchildren in 1993 showed that 13.4% of boys and 10.5% of girls aged 6–18 years were obese (defined as  > 120%median weight-for-height using local reference ranges). 2 Recent unpublished data from the Department of Health,HongKongshowedthattheincidenceofobesityincreasedfrom 12.7% to 16.5% in primary school and 10.4% to13.7% in secondary school between the academic years1997/98 and 2002/03 (G Tong, personal communication).Parallel to this increase in childhood obesity, anincrease in the prevalence of cardiovascular risk factorsamong children has also been noted. In the Bogalusastudy, up to 60% of overweight children were found tohave at least one cardiovascular disease risk factor. 3,4 Among these risk factors, insulin resistance showed thestrongest correlation with body mass index (BMI), and ithas been suggested to be the important link in cardiovas-cular morbidityand mortality seen in obese subjects. 3,4 Inadults, obstructive sleep apnea (OSA) has been shown tobe a significant risk factor for insulin resistance. 5–7 Evenafter controlling for weight, evidence is accumulating todemonstrate that OSA is independently associated withaltered glucose metabolism and may predispose to theeventual development of type 2 diabetes mellitus. 7 Theactual mechanism is unclear but intermittent hypoxia cantrigger a cascade of pathophysiological events includingautonomic activation, alterations in neuroendocrine func-tion, and release of potent pro-inflammatory mediatorssuchastumornecrosisfactor- a andinterleukin-6andthus 1 Department of Paediatrics, Prince of Wales Hospital, The ChineseUniversity of Hong Kong, Shatin, Hong Kong SAR. 2 Department of Chemical Pathology, Prince of Wales Hospital, TheChinese University of Hong Kong, Shatin, Hong Kong SAR. 3 Centre for Epidemiology and Biostatistics, Prince of Wales Hospital, TheChinese University of Hong Kong, Shatin, Hong Kong SAR.Grant sponsor: Direct Grant for Research, CUHK (Reference no.2005.2.027).*Correspondence to: Albert M. Li, Clinical Lecturer, Department of Paediatrics, Prince of Wales Hospital, The Chinese University of HongKong, Shatin, Hong Kong SAR. E-mail: albertmli@cuhk.edu.hk Received 20 March 2006; Revised 16 June 2006; Accepted 23 June 2006.DOI 10.1002/ppul.20508Published online in Wiley InterScience(www.interscience.wiley.com).   2006 Wiley-Liss, Inc.  cause insulin resistance. 7 This potentially causal relation-ship between OSA and insulin resistance is furthersupported by results from interventional study on adultpatients with OSA. Improvements in insulin sensitivity asassessed by the hyperinsulinaemic euglycaemic clamp,were notedwithin 2 days after the initiation of continuouspositive airway pressure (CPAP) treatment, and thisimprovement was unrelated to changes in body weight. 8,9 The relationships between obesity, sleep apnea, andinsulin resistance in children however, are less welldefined with published studies giving conflictingresults. 10–12 de la Eva et al. 10 demonstrated that theseverity of OSA correlated positively with fasting insulinlevelsin62obesechildren,independentofBMI.However,inasubsequentstudyinvolvingasimilarnumberofobesechildren, Tauman et al. found that insulin resistance wasdetermined primarily by the degree of body adiposityrather than by the severity of OSA. The aim of this studywas to assess the relationship between fasting insulinlevels,dyslipidemia,andOSAinacohortofobesesnoringchildren in order to provide evidence to support or refutethe suggested relationship between these parameters. Wehypothesized that in obese children with snoring, severityof OSAwas correlated with insulin levels independent of the degree of obesity. METHODSSubjects Consecutive children with primary obesity, a history of snoring for more than three nights per week (habitualsnoring) and aged 7–18 years were recruited from ourSleep Disorder Clinic during the period between January2004 and June 2005. These children had been referred forassessment and counseling on healthy living by theirprimary health care physicians. Children were excludedfrom the study if they had intercurrent respiratory tractinfection, a neuromuscular disorder, Down syndrome,diabetes mellitus, or they were taking medications thatcould interfere with the interpretation of polysomnogra-phy (PSG) or glucose homeostasis. The study wasapproved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong and informedconsentfromthesubjectsandtheirparentswasobtainedatthe beginning of the assessment. All subjects had BMIabove the 95th percentile according to local sex and agespecific reference ranges. 2 Anthropometry Waistcircumferenceofthesubjectswasobtainedusingan inelastic cloth measuring tape around the area of greatest girth of the abdomen. Their weight and standingheight were measured with a calibrated weighingscale and stadiometer, respectively, by standard anthro-pometric methods. 13 BMI was calculated as weight/ height 2 (kg/m 2 ). Sleep Study AnovernightPSGwasperformedoneachsubjectusingSiesta ProFusion II PSG machine (Compumedics Tele-med PTY Ltd. Abbotsford, Australia) to record thefollowing parameters: electroencephalogram from fourleads (C 3  /A 2 , C 4  /A 1 ), bilateral electrooculogram, electro-myographic activity of the mentalis muscle, and bilateralanterior tibialis. Respiratory movements of the rib cageand abdomen were measured by pneumatic effort belt.Electrocardiogram and heart rate were continuouslyrecorded from two anterior chest leads. Arterial oxyhe-moglobinsaturation(SaO 2 )wasmeasuredbyanoximeter(Ohmeda Biox 3900 Pulse Oximeter) with finger probe.Respiratory airflow pressure signals were measured at theanterior nares and connected to a pressure transducer.Snoring was recorded by a microphone placed on theupper neck. Body position was monitored via a bodyposition sensor. Obstructive apnea (OA) was defined asabsenceofairflowwithpersistentrespiratoryeffortlastinglonger than two baseline breaths, irrespective of SaO 2 changes. Obstructive apnea index (OAI) was defined asthe number of OA per hour of sleep. Central apnea (CA)was defined as absence of respiratory effort associatedwith absence of airflow. Those of greater than 20 sec withor without oxygen desaturation or arousals, and those of anydurationbutassociatedwithoxygendesaturationofatleast 4% and/or arousals were quantified. Hypopnea wasdefined as a reduction of 50% or more in the amplitude of theairflowsignal.Itwasonlyquantifiediflongerthantwobaseline breaths and associated with oxygen desaturationof at least 4% and/or arousals. Apnea hypopnea index(AHI) was defined as the total number of apneic andhypopneic episodes per hour of sleep. Oxygen saturationnadir and the percentage of total sleep timewhere oxygensaturation was below 90% were noted. Arousal wasdefined as an abrupt shift in EEG frequency duringsleep, which may include theta, alpha, and/or frequenciesgreater than 16 Hz but not spindles, with 3 to 15 sec induration. In rapid eye-movement (REM) sleep, arousalswere scored only when accompanied by concurrentincreases in submental EMG amplitude. We definedOSA as AHI > 1.0. Children with an AHI > 1.0 but < 10were considered to have mild OSA and those with anAHI  10 were considered to have moderate to severeOSA.We have not tried to separate our cases without OSAinto genuine habitual snoring and upper airway resistancesyndrome (UARS) for further comparison. We do notroutinely perform oesophageal pressure monitoring withPSGthusthediagnosisofUARSwouldhavebeendifficulttoverify.Besides,theissueofUARSisstillacontroversial 1176 Li et al.  topic with no agreed consensus in its definition, 14 and itwasnottheaimofourstudytolookatchildrenwithUARSseparately. Laboratory Study All subjects had blood samples taken in the morning,after an overnight fast, for the estimation of plasmaglucose (GLU), serum insulin (INS) concentrations, andlipid profile [total cholesterol (TC), triglyceride (TG),high-density lipoprotein cholesterol (HDL), and low-density lipoprotein cholesterol (LDL) concentrations].The homeostasis model assessment (HOMA) based onserum fasting insulin and glucose concentration wascalculated as INS inmIU/L  GLU in mmol/L dividedby22.5. This index has been shown to be a valid tool forassessing insulin sensitivity in both prepubertal andpubertal obese children. 15,16 PlasmaGLUwas measuredbyhexokinasemethod(DPModular Analytics, Roche Diagnostics Corp, Indianapo-lis, IN). The interassay coefficient of variation was 3% orless at all concentrations up to 41.6 mmol/L. Lipid profileincluding TC, HDL, and LDL concentrations weremeasured by cholesterol esterase/cholesterol oxidasecoupling Trinder’s reaction with appropriate pre-treat-ment steps; PEG modified enzyme and dextran sulfate forHDL and non-ionic detergent for LDL (DP ModularAnalytics, Roche Diagnostics Corp). The interassaycoefficients of variation for these assays were 3% or lessat levels up to 20.7, 3.1, and 14.2 mmol/L for TC, HDL,and LDL, respectively. TG concentration was measuredby lipoprotein lipase coupling Trinder’s reaction (DPModular Analytics). The interassay coefficient of varia-tion was 3% or less at all concentrations up to 11.4 mmol/ L. Serum insulin concentration was measured by micro-particle enzyme immunoassay (Imx analyser, AbbottDiagnostics, Illinois, IL) and the interassay coefficient of variation was 5% or less at all concentrations up to300 mIU/L. Statistical Analysis The subjects were divided into three groups (non-OSAsnoring group: AHI < ¼ 1, mild OSA group: AHI: 1–10,andmoderatetosevereOSAgroup:AHI > 10),accordingto the AHI score. The demographic data and laboratoryresults were expressed as median with interquartileranges.TheKruskal–WallistestwiththeMann–Whitneypost hoc test were used to explore the relationship of thefactors among these three groups. The two OSA groupswerecombinedintoasinglegroup(AHI > 1).Thevariousdemographic and laboratory parameters between thesingle OSA group and non-OSA snoring group werecompared by univariate logistic regression. All signifi-cantly different parameters between the two groups wereentered into logistic regression using a forward stepwiseselection strategy to pick out factors that were associatedwith OSA. When two or more potential risk factors werehighly correlated, the factor that was clinically importantwas selected for entry. SPSS for Windows (10.1, SPSS,Inc.,Chicago,IL)wasusedintheanalysis,andthelevelof significance was set at 5% for all comparisons.The BMI z-scores were calculated from the L, M, Scurves,usingvaluesappropriateforthechild’sageandsexand computed by the formula:Z  ¼ ½ Y ð t Þ = M ð t Þ L ð t Þ   1L ð t Þ S ð t Þ Y  ¼  BMIThe values of L, M, S curves were obtained fromReference 17. RESULTS A total of 94 obese subjects with habitual snoring werestudied.Seventy-threesubjectsweremaleand themedianage of the studied group was 12.0 years [interquartilerange, (IQR) 9.7–13.9]. None of the subjects had activecardiopulmonary disease, as evidenced by their normalphysical examination. Their median BMI was 28.9 (IQR26.1–32.2), and when compared to age-matched con-trols, 2 the BMI values of our subjects were  > 95thpercentile. The demographic, PSG, and laboratory dataof the subjects are shown in Table 1. TABLE 1—Characteristics of the Study Population (n ¼ 94) Median Interquartile rangeWeight (kg) 64.8 53.0–81.0Height (cm) 150.0 142.0–159.0Waist (cm) 85.8 77.5–93.5BMI (kg/m 2 ) 28.9 26.1–32.2BMI z-score 2.42 2.08–2.71Age (years) 12.0 9.7–13.9OAI (per hour) 0.3 0–1.0AHI (per hour) 1.6 0.8–4.3Oxygen saturation nadir (%) 82.0 73.0–89.0Percentage of total sleeptime with saturation < 90%0.0 0.0–0.1INS (mIU/L) 17.1 12.2–23.2HOMA 3.9 2.8–5.4GLU (mmol/L) 5.1 4.9–5.5TC (mmol/L) 4.6 4.0–5.1TG (mmol/L) 1.2 0.9–1.7HDL (mmol/L) 1.3 1.1–1.5LDL (mmol/L) 2.7 2.2–3.2BMI, body mass index; OAI, obstructive apnea index; AHI, apneahypopnea index; INS, serum insulin; HOMA, homeostasis modelassessment; GLU, plasma glucose; TC, total cholesterol; TG,triglyceride; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol. Insulin and Obstructive Sleep Apnea in Obese Chinese Children 1177  Among all studied subjects, 60 (64%) satisfied thediagnostic criterion for OSA, 47 and 13 had mild andmoderate to severe OSA, respectively. Subject character-istics according to their OSA severity are shown inTable 2. As obstructive events contributed to the majorityof the respiratory events, AHI reported in this studyincluded only obstructive apneic and hypopneicepisodes.Subjects with moderate to severe OSAwere significantlyolder, heavier, and had greater waist circumference thanthe mild OSA and non-OSA snoring groups. Mainly boyswere represented in the group with moderate to severeOSA. In addition, obese subjects with moderate to severeOSA were found to have significantly greater fastinginsulin levels and HOMA compared to the other twogroups.The median insulin level on the morning after anovernight PSG was 17.1 mIU/L (IQR 12.2–23.2).Twenty-two (22%) subjects from this study had a fastinginsulin level > 24 mIU/L suggestive of insulin resistance(unpublished data). We combined mild and moderate tosevereOSAgroupsintoasinglegroupandcomparedwiththe non-OSA snoring group. All significantly differentfactors between the two groups were entered into logisticregression using a forward stepwise selection strategy tofind out those parameters that were associated with OSA.Both saturation nadir and insulin levels were significantlyassociated with the condition (Tables 3 and 4). DISCUSSION The findings of this study provided evidence that inobese children with habitual snoring, insulin level was anindependent factor associated with OSA. In addition, wewereabletodemonstrateahighprevalenceofOSAamongobese snoring children. Out of the 94 children studied,64% met the diagnostic criteria for OSA.Insulin resistance is a state in which there is a less-than-normal biologic response to insulin. There isaccumulating evidence that it plays an important role inthe pathogenesis of hypertension and cardiovasculardisease, children and adults alike. 3,4,18–20 The hyperinsu-linaemic euglycaemic clamp is the gold standard methodfor assessing insulin sensitivity. But it is too invasive aprocedure for children, and it is also very labor intensiveand technically demanding. 15,16 Previous validationstudies have found that in subjects without diabetes,fasting insulin and HOMA index are reliable surrogatemarkers of insulin sensitivity. 15,16 These markers havebeen used for this purpose in studies looking at childhoodOSA and insulin resistance. 10,12,21 Unpublished datainvolving Chinese obese children demonstrates an insulinlevel > 24mIU/Lissuggestiveofinsulinresistance.Inourstudy, we found 22% of our subjects had an insulinlevel > 24 mIU/L, a prevalence rate similar to what hasbeenreportedintheliterature. 22,23 InthestudybyTaumanet al. 12 12% of their studied population that comprisedobese and non-obese children had insulin resistance.In adults, OSA has been shown to be independentlyassociated with glucose intolerance, insulin resistance,and may eventually lead to type 2 diabetes mellitus. 7 Thiscomplex relationship is further supported by trialsdocumenting changes in glucose metabolism followingCPAP treatment of OSA. 9,24–26 In the largest studydocumenting a beneficial effect of CPAP to date, TABLE 2—Demographic, PSG, and Laboratory Data in 34 non-OSA Subjects, 47 With Mild OSA, and 13 With Moderate toSevere OSA Non-OSA Mild OSA Moderate to severe OSAWeight (kg) 63.0 (49.6–75.3) 60.7 (52.2–76.9) 86.4 (70.6–94.5)** , ***Height (cm) 150.0 (137.0–160.0) 147.5 (139.0–157.0) 156.5 (150.0–165.0)** , ***Waist (cm) 84.5 (77.0–87.5) 85.0 (75.5–94.0) 95.0 (91.9–107.9)** , ***BMI Z-score 2.34 (2.04–2.68) 2.45 (2.17–2.72) 2.48 (2.09–2.74)Age (years) 11.6 (9.0–13.4) 11.8 (9.5–13.5) 13.9 (11.1–14.8)** , ***Gender, male/female 27/7 35/12 11/2Oxygen saturation nadir (%) 89 (80.0–91.0) 80.0 (72.0–84.0)* 72 (56.0–82.0)** , ***Percentage of total sleep timewith saturation < 90%0.0 (0.0–0.1) 0.0 (0.0–0.1)* 0.0 (0.0–0.1)** , ***INS (mIU/L) 13.8 (11.7–18.9) 16.7 (12.1–23.1) 31.3 (26.7–40.8)** , ***HOMA 3.0 (2.6–4.4) 3.9 (2.6–5.3) 7.6 (6.2–10.1)** , ***GLU (mmol/L) 5.1 (4.8–5.4) 5.1 (4.9–5.6) 5.3 (5.0–5.6)TC (mmol/L) 4.6 (4.0–5.1) 4.6 (3.9–5.1) 4.7 (4.3–5.5)TG (mmol/L) 1.2 (0.9–1.7) 1.2 (0.9–1.6) 1.6 (1.2–2.0)HDL (mmol/L) 1.3 (1.1–1.5) 1.3 (1.1–1.4) 1.1 (1.0–1.4)LDL (mmol/L) 2.6 (2.1–3.0) 2.8 (2.1–3.0) 2.9 (2.5–3.4)Median (IQR).* P < 0.05 mild OSAversus non-OSA.** P < 0.05 moderate to severe OSAversus mild OSA.*** P < 0.05 moderate to severe OSAversus non-OSA. 1178 Li et al.  improvement in insulin sensitivity, as assessed by thehyperinsulinaemic euglycaemic clamp, was observedover a 3-month period after the initiation of CPAP. 26 Interestingly, insulin sensitivity increased within 2 daysof therapy, with further improvements occurring at the3-monthfollow-up.Ourfindingsprovidefurtherevidenceof an independent association between OSA and insulinlevels. The effect size obtained in the current study is notas great as that demonstrated invarious adult studies, anddifferences in the duration of disease and co-morbiditiesamong adult subjects could explain this discrepancy.Adult patients tend to have had prolonged disease prior todiagnosis and they typically have more severe OSAcompared to their pediatric counterparts. A positiveassociation between OSA and insulin levels has also beendemonstratedinanotherstudyinvolvingobesechildren. 10 de la Eva et al showed that significant correlations werepresent between AHI and fasting insulin levels indepen-dent of BMI. Tauman et al. 12 on the other hand wereunable to demonstrate a positive correlation betweeninsulin resistance and OSA. In their study, insulinresistance and dyslipidemia were determined primarilyby the degree of body adiposity. Their study had 70 obesechildrenconsistingof41boysandtheirmeanrelativeBMIand insulin levels were 1.9 times and around 13.0 mIU/L,respectively. Compared to our study where the genderratio greatly favored boys and the median relative BMIusing local reference data was 2.5 times. Besides, ourmedianinsulinlevelswerealsogreaterat17.0mIU/L.Thedifferences in gender distribution and degree of obesitycould have accounted for the different results.The actual mechanisms by which OSA may disruptinsulin sensitivity are not well defined. Potential inter-mediates include an activation of the sympathetic path-way, alteration in the hypothalamic-pituitary-adrenalhormonal axis, release of inflammatory mediators, andthe direct effects of hypoxemia on glucose regulation. 7 Data from previous studies have provided support thatchildren with OSA exhibit elevated levels of sympatheticactivity. 27,28 In addition, there is growing evidence thatchildren with OSA have higher blood pressure comparedto controls, 28 likely as a result of sympathetic activation.How this complex interaction between childhood OSA,sympathetic activity, and glucose homeostasis leads toinsulin resistance will require further investigation. Theincidence of OSA has been suggested to increase amongobese children. 29 Our study has added further support tothishypothesis,where64%ofourcohortofobesesnoringchildren actually had OSA. By demonstrating a positiveassociationbetweenOSAandinsulinlevels,andknowingthe adverse effects associated with insulin resistance, ourstudyresultswouldhaveimportantclinicalimplicationsinthe management of childhood obesity.There are a few potential limitations to our study. First,even though our subjects’ puberty was age appropriate,their Tanner staging was not recorded at the time of clinical assessment. Second, our study was a cross-sectional assessment at a single time point. Longitudinaldata including interventions would have provided moreinteresting data for the association and complex interac-tion between OSA and insulin levels. Thirdly, we haveselected a biased sample as we included only obesesnoring children and they were recruited from attendantsto our Sleep Disorder Clinic. We did not have obese and TABLE 3—Univariate Logistic Regression Non-OSA OSA OR (95%CI)  P -valueWeight (kg) 63.0 (49.6–75.3) 66.0 (54.2–82.5) 1.02 (0.99–1.04) 0.180Height (cm) 150.0 (137.0–160.0) 149.0 (144.9–159.0) 1.00 (0.97–1.03) 0.937Waist (cm) 84.5 (77.0–87.5) 87.0 (77.8–94.5) 1.05 (1.00–1.09) 0.053BMI z-score 2.34 (2.04–2.68) 2.48 (2.16–2.73) 1.13 (0.59–2.20) 0.710Age (years) 11.6 (9.0–13.4) 12.0 (10.5–13.9) 1.06 (0.91–1.24) 0.461Oxygen saturation nadir (%) 89 (80.0–91.0) 79 (71–84) 0.00 (0.00–0.05) 0.002Percentage of total sleep time withsaturation  < 90%0.0 (0.0–0.1) 0.0 (0.0–0.1) 0.00 (0.00–0.00) 0.022INS (mIU/L) 13.8 (11.7–18.9) 18.7 (12.5–26.7) 1.10 (1.03–1.17) 0.004HOMA 3.0 (2.6–4.4) 4.1 (3.0–6.4) 1.51 (1.15–2.00) 0.003GLU (mmol/L) 5.1 (4.8–5.4) 5.2 (4.9–5.6) 2.16 (0.83–5.62) 0.116TC (mmol/L) 4.6 (4.0–5.1) 4.6 (4.1–5.2) 1.34 (0.80–2.26) 0.274TG (mmol/L) 1.2 (0.9–1.7) 1.2 (0.9–1.7) 1.17 (0.57–2.40) 0.665HDL (mmol/L) 1.3 (1.1–1.5) 1.2 (1.1–1.4) 0.52 (0.13–2.02) 0.343LDL (mmol/L) 2.6 (2.1–3.0) 2.8 (2.3–3.2) 1.68 (0.87–3.23) 0.123 TABLE 4—Logistic Regression Analysis, FactorsAssociated With OSA Coefficient SE Odds ratio (95% CI)  P -valueInsulin levels 0.074 0.033 1.077 (1.011–1.048) 0.022Saturation nadir   6.738 2.718 0.001 (0–0.244) 0.013 Insulin and Obstructive Sleep Apnea in Obese Chinese Children 1179
Related Documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks