Evaluation of different definitions of childhood obesity among Hong Kong Chinese children and adolescents.

• Other Authors: Wang, Lin
• Format: Others
• Language: English; Chinese; Chinese
• Published: 2012
• Subjects: Obesity in children; Obesity in children > China > Hong Kong; Obesity in adolescence; Obesity in adolescence > China > Hong Kong; Obesity; Obesity > Hong Kong; Child; Adolescent; Adolescent > Hong Kong
• Online Access: http://library.cuhk.edu.hk/record=b5549469; http://repository.lib.cuhk.edu.hk/en/item/cuhk-328432

本研究的目的在於以電阻抗法測定的脂肪比例作為標準，評價各種兒童肥胖定義診斷脂肪過量時的診斷準確性。爲了達到這項研究目的，本研究分為三個部份。首先，研究的第一部份評價了生物電阻抗法（Bio-electrical impedance analysis, BIA）在測量中國兒童青少年身體成份時的效度。隨後，幾種被廣泛應用的電子脂肪磅測量身體脂肪比例時的測量準確度被進行了評價。最後，在大規模人群測試中，利用精確的電子脂肪磅及幾種人體測量學指標決定的脂肪水平，不同兒童肥胖定義的診診斷脂肪過量的斷準確性被進行了評價。共有255名9至19歲的中國兒童青少年參與了第一和第二部份的測試，利用雙能X光骨密度儀（Dual-energy X-ray absorptiometry, DEXA）測量的體成份作為標準，比較了通過標準電阻抗法以及幾種被廣泛電子脂肪磅測量而來的體成份。研究結果顯示，生物電阻抗法是測量中國兒童青少年體成份有效的方法，然而，所有以前研究發展的BIA預測方程均未能準確估計受試者的去脂體重（Fat-free mass, FFM）。四種商業電子脂肪磅在測量受試者的體脂有類似的診斷準確度。隨後，另外的2,134名兒童青少年參與了本研究第三部份的測試，利用經過第二部份經過調整的脂肪磅測量受試者的體脂比例，同時測量人體測量學指標。研究結果發現，現有關於兒童肥胖的定義在診斷香港兒童、青少年脂肪過量的診斷準確性不夠。此外，受試者工作特徵曲綫分析（Receiver operating characteristics, ROC）發現利用基於身高和體重的指數來定義兒童肥胖在診斷肥胖兒童時較基於腰圍的指數有更好的敏感度和特異性。因此，研究者可以通過調整這些身高和體重的指數的劃分點來獲得對兒童肥胖更好的診斷準確性。 === 研究一（第3章）的目的在於評價生物電阻抗（Bio-electrical impendence analysis, BIA）方法在預測中國兒童青少年的去脂體重（Fat-free mass, FFM）中的效度，同時，也驗證已有的利用BIA方法發展的方程預測中國兒童、青少年FFM的效度。共有255名9-19歲健康的中國兒童、青少年（127名男生、128名女生）自願參與此次測試。利用傳統的單頻（50 kHz）的手腳電阻抗儀測試人體的電阻與電抗。利用雙能X光骨密度儀（Dual-energy X-ray absorptiometry, DEXA）作為測量FFM的標準。以DEXA測試的FFM為標準，對24個已有的BIA預測FFM的方程進行交互驗證。研究結果顯示，在本研究的人群中，利用24個已有的BIA預測方程所預測FFM與DEXA測量而來FFM高度相關。然而，這24個方程都未能準確預測本研究測試人群中由DEXA測量而來的FFM。利用多元線性回歸及交互驗證的方法，本研究發展了一個預測中國兒童、青少年FFM的預測方程：FFM (kg) = 1.613 + 0.742 × 身高(cm)²/電抗 (Ω) + 0.151 × 體重 (kg); R² = 0.95; SEE = 2.45kg; CV = 6.5%。本研究的研究結果表明，已有的BIA預測FFM的方程均不能通過交互驗證，從而準確的預測本研究人群的FFM。利用本研究人群發展而來的BIA預測FFM的方程，對預測中國兒童、青少年的FFM有良好的效度。本研究證明利用BIA方法可以準確的預測中國兒童、青少年的體成份。 === 研究二（第4章）旨在驗證4種利用BIA原理發展而來的商業電子脂肪磅在測量中國兒童、青少年體脂比例時的效度，以及判斷這些脂肪磅診斷身體脂肪過量時的診斷表現。參與研究一的255名中國兒童、青少年也參與本研究。利用DEXA作為測量體成份的標準測量。同時，4種電子脂肪磅（Model A, Biodynamics-310; Model B, Tanita TBF-543; Model C, Tanita BC-545 和 Model D, InBody 520）利用製造商內置的BIA方程測量受試者的體成份。研究結果表明，採用製造商內置的BIA方程測量體脂比例時，在男性中，Model B和Model C測量的體脂比例與DEXA測量的體脂比例有明顯差異 （p < 0.05）；Model C低估了脂肪過量兒童、青少年在在受試人群中的百分比（X²=10.714, p=0.001）。在女性中，Model B、Model C和Model D測量的體脂比例與DEXA測量的體脂比例有明顯差別，脂肪過量兒童在占受試人群中的比例被這三種脂肪磅所低估。由於在這四種脂肪磅和DEXA測量的%BF有著較高的相關性，回歸分析被用於調整這四種脂肪磅所測量的體脂比例。在調整后，在男性和女性中，所有調整的脂肪磅測試的體脂比例與DEXA測量的體脂比例之間沒有分別。在男性和女性受試者中，卡方檢驗結果顯示，所有校正的脂肪磅診斷的脂肪過量兒童、青少年占受試人群中的比例與DEXA的診斷比例之間沒有發現明顯區別。 此外，在女性中，與未校正的脂肪磅的對脂肪過量兒童、青少年診斷的敏感度比較，調整后的Model B 和 ModelC 脂肪磅對脂肪過量的兒童診斷的敏感度得到明顯改善 (Model B, X²=9.818,p=0.002; Model C, X²=4.615, p=0.032)。本研究的研究結果建議在利用電子脂肪磅內置方程測量中國青少年、兒童的體脂比例時，其效度需要進行驗證。因為在脂肪磅和DEXA測量之間有較高的相關性和較小的偏差，經調整的脂肪磅可以用於在進行大規模人群測試時，測量中國青少年、兒童的體脂比例，以及作為對脂肪過量兒童、青少年的診斷工具。 === 研究三（第5章）的目的在於調查基於身高體重的指數和基於腰圍的指數在診斷香港兒童、青少年脂肪過量時的總體診斷表現；以及判斷現有兒童肥胖定義診斷香港兒童、青少年脂肪過量時的診斷表現。共有來自香港多所中小學的2134名9-19歲的受試者，包括1135名男性、999名女性參與本研究。利用研究二調整的便攜式足對足電子脂肪磅測量受試者的體脂比例。基於體脂比例的兒童肥胖，即脂肪過量，被定義為體脂比例在男性高於25%，女性高於30%。四種基於BMI的兒童肥胖定義、兩種基於香港數據利用腰圍對兒童肥胖的定義、以及香港現有的兒童肥胖定義用於此研究。受試者工作特徵曲綫分析（Receiver operating characteristics, ROC）被用於分析身高體重指數（BMI）、標準身高體重（PWH）、腰圍（WC）和腰圍身高比（WHtR）在診斷香港兒童、青少年脂肪過量時的總體診斷表現。診斷試驗的指標被用來判斷現有兒童肥胖的定義診斷香港兒童、青少年脂肪過量時的診斷表現。本研究發現，在男性中，BMI、PWH、WC和WHtR在診斷香港兒童、青少年脂肪過量時有相同的良好的診斷表現[受試者工作特徵曲綫下面積(area under the curve of ROC, AUC) = 0.909-0.923]。然而，在女性中，WC（AUC=0.840）和WHtR（AUC=0.850）診斷香港兒童、青少年脂肪過量時的診斷表現差於BMI（AUC=0.900）和 PWH （AUC=0.903）。此外，現有兒童肥胖定義在診斷香港兒童、青少年脂肪過量時診斷敏感度較低，敏感度在男性中為0.325-0.761；女性中為0.128-0.588。利用最好診斷準確度的方法來確定的BMI、PWH、WC和WHtR的劃分點在診斷脂肪過量時，在男性中有相似的診斷表現，診斷敏感度在0.816-0.868之間，診斷特異度在0.803-0.869之間；對於女性，BMI和PWH 較WC和WHtR有較高的敏感度和特異度。本研究的發現表明，在診斷脂肪過量時，基於身高和體重的指數總體診斷表現好於基於腰圍的指數；現有兒童肥胖定義在診斷香港兒童、青少年脂肪過量時的診斷表現不佳。進一步，通過調整以身高和體重為基礎指數的劃分點可以使其在診斷脂肪過量時有更好的診斷表現。 === The purpose of this study was to evaluate the diagnostic accuracy of various childhood obesity definitions using bioelectrical impedance analysis (BIA) as a criterion measure of body fat. To achieve such an objective, the study involved three phases. First, the validity of BIA in measuring body composition in Chinese children and adolescents was evaluated. Then, examination on measuring accuracy of several popular BIA scales was performed. Finally, mass testing of body fat levels using an accurate BIA scale, as well as other anthropometric measures, was conducted to evaluate the diagnostic accuracy of childhood obesity definitions. A total of 255 healthy Chinese children and adolescents aged 9 years to 19 years participated in the first and second phases of the study. Body composition was measured from BIA using a criterion device (Biodynamics 310) and four popular commercial BIA scales and compared with measurements from dual-energy X-ray absorptiometry (DEXA). BIA provided excellent predictions of body composition among Chinese children and adolescents. However, all of the previously developed BIA equations yielded biased estimation. The four commercial BIA scales had similar diagnostic accuracy in measuring body fat. Subsequently, another 2,134 boys and girls were recruited to take part in body fat measurements using the best BIA scale identified in the second phase with adjusted equations, as well as other anthropometric measurements. The diagnostic accuracy of all existing definitions of childhood obesity was poor in both genders. Moreover, Receiver Operating Characteristics (ROC) analysis found that childhood obesity definitions using weight-and-height based indices [Body mass index (BMI) and weight-for-height) had superior sensitivity and specificity in identifying obese children compared with waist circumference-based indices. Therefore, with adjusted cut-off criteria for weight-and-height indices, the diagnostic accuracy of childhood obesity would be improved. === Study I : Validity of BIA method in predicting FFM === The first study (Chapter 3) aimed to examine the validity of the BIA method in predicting fat-free mass (FFM) in Chinese children and adolescents and of various published BIA equations in estimating FFM in this particular group. A total of 255 healthy Chinese children and adolescents (127 boys and 128 girls) aged 9 years to 19 years participated in this study. BIA variables (e.g., resistance and impedance) were measured at 50 kHz between the hand and foot using a traditional BIA device. The criterion of FFM measurement was also assessed using DEXA. FFM estimated from 24 published BIA equations were cross-validated against the criterion measure from DEXA. FFM estimated from the 24 published BIA equations yielded high correlations with the directly measured FFM from DEXA. However, none of the 24 equations was statistically equivalent with the DEXA-measured FFM. Using multiple linear regression and cross-validation against the DEXA measurement, an alternative prediction equation was developed as follows: FFM (kg) = 1.613 + 0.742 × height (cm)²/impedance (Ω) + 0.151 × body weight (kg); R² = 0.95; SEE = 2.45 kg; CV = 6.5%. None of the previously developed BIA equations was able to cross-validate the FFM estimates of the present sample. An alternative BIA equation, with evidence of validation and cross-validation, was thus proposed. The method based on the BIA principle provides a valid estimation of body composition among Chinese children and adolescents. === Study II: Validity of popular BIA consumer scales in measuring body fat === The second study (Chapter 4) examined the validity in measuring body fat and the diagnostic performance of four different consumer BIA scales. The 255 Chinese children and adolescents from the first study also participated in this study. DEXA was used as the criterion measurement for %BF, which was also assessed using four BIA scales [Model A (Biodynamics 310), Model B (Tanita TBF-543), Model C (Tanita BC-545), and Model D (InBody 520)]. The validity in measuring body fat and the diagnostic performance in screening excess body fat of these BIA scales was first examined. In boys, differences in %BF between Models B, C, and DEXA were significant (p < 0.05). In girls, significant differences in %BF were observed between Models B, C, D, and DEXA (p < 0.05). The prevalence of overfat in boys was underestimated by Model C (X² = 10.714, p = 0.001). For girls, the prevalence of overfat was underestimated in Models B, C, and D. Because of the high correlation between the BIA scales and DEXA (r = 0.770.94), regression analysis was used to adjust the BIA scales in measuring %BF in this sample. After adjustment, the paired t-tests did not show differences in %BF between the adjusted BIA scales and the DEXA measurement in both genders. All adjusted BIA scales except Model A in girls showed substantial agreement with the DEXA measurement. In addition, compared with overfat classification using DEXA measurement, X² test showed that the prevalence of overfat in the present samples was classified correctly by all adjusted BIA scales in both genders. Compared with the original BIA measurements, the use of the adjusted Models B and D offered significant improvements in sensitivity for girls. These findings suggest that the embedded equations in BIA scales should be validated in assessing body compositions among Chinese children and adolescents. The adjusted BIA scales can be used in large population surveys due to the high correlation and small mean bias between the BIA scales and the DEXA measurements. In conclusion, the adjusted BIA scales can serve as diagnostic tools to classify overfat Chinese children into relevant subgroups. === Study III: Diagnostic accuracy of definitions of childhood obesity === The third study (Chapter 5) aimed to investigate the diagnostic performance of weight-and-height-based indices and waist circumference (WC-based indices as diagnostic tests to screen excess body fat in Hong Kong children and adolescents and to investigate the diagnostic accuracy of existing definitions of childhood obesity in Hong Kong. A total of 2,134 participants (1,135 boys and 999 girls) were recruited from local schools. The adjusted foot-to-foot BIA scale (Model B) in the second study was applied to assess %BF. The criterion of childhood obesity (i.e., overfat) was defined as over 25%BF for boys and over 30%BF for girls. Childhood obesity was also determined from four BMI-based references, two WC-based references, and the 1993 HK reference. The diagnostic accuracy of the existing definition for childhood obesity in screening excess body fat was evaluated using diagnostic indices. The Receiver Operating Characteristics (ROC) analysis was used to evaluate the general performance of BMI, PWH, WC, and WHtR in detecting overfat. In boys, ROC analysis showed no significant difference among the four indices in screening overfat [(area under the curve of ROC, AUC) = 0.909-0.923]. In girls, BMI and PWH performed better in detecting overfat than WC and WHtR (AUC of BMI = 0.900; AUC of PWH = 0.903; AUC of WC = 0.840 and AUC of WHtR = 0.850). All definitions for obesity showed low sensitivity (boys, 0.325-0.761; girls, 0.128-0.588) in detecting overfat. Cut-offs derived for best accuracy showed similar diagnostic performance in each index in boys but not in girls. In boys, the cut-offs of BMI, WC, WHtR, and PWH can provide similar sensitivity (0.816-0.868) and specificity (0.803-0.869) in screening overfat. In girls, BMI and PWH can provide higher sensitivity and specificity than WC and WHtR. This study’s findings demonstrate that the diagnostic performance of all existing definitions for obesity is poorer than expected in both genders. BMI and PWH are superior to use as proxy measures in screening overfat among Hong Kong Chinese children and adolescents for both genders. Moreover, the diagnostic performance of these indices can be improved by adjusting the existing cut-offs. === Detailed summary in vernacular field only. === Detailed summary in vernacular field only. === Detailed summary in vernacular field only. === Detailed summary in vernacular field only. === Wang, Lin. === Thesis (Ph.D.)--Chinese University of Hong Kong, 2012. === Includes bibliographical references (leaves 166-197). === Abstract and appendixes also in Chinese. === ABSTRACT --- p.i === 摘要 --- p.v === ACKNOWLEDGEMENT --- p.ix === PUBLICATIONS --- p.xi === LIST OF TABLES --- p.xv === LIST OF FIGURES --- p.xvii === ABBREVIATIONS --- p.xviii === Chapter CHAPTER 1. --- INTRODUCTION --- p.1 === Chapter 1.1. --- Background --- p.1 === Chapter 1.2. --- Purposes and significance --- p.4 === Chapter 1.3. --- Operational definitions --- p.5 === Chapter 1.3.1. --- Childhood overweight and obesity --- p.5 === Chapter 1.3.2. --- Body composition --- p.6 === Chapter 1.3.3. --- Bioelectric impedance analysis (BIA) --- p.6 === Chapter 1.3.4. --- Validity --- p.6 === Chapter 1.3.5. --- Diagnostic performance --- p.6 === Chapter 1.4. --- Hypothesis --- p.6 === Chapter 1.5. --- Limitations --- p.7 === Chapter 1.6. --- Delimitations --- p.7 === Chapter CHAPTER 2. --- LITERATURE REVIEW --- p.9 === Chapter 2.1. --- The epidemic and health consequence of childhood obesity --- p.10 === Chapter 2.1.1. --- Prevalence of childhood obesity --- p.10 === Chapter 2.1.2. --- Health consequence of childhood obesity --- p.15 === Chapter 2.2. --- Contributors of Childhood obesity --- p.16 === Chapter 2.2.1. --- Genetics --- p.17 === Chapter 2.2.2. --- Dietary intake --- p.17 === Chapter 2.2.3. --- Physical activity --- p.18 === Chapter 2.2.4. --- Sedentary behaviors --- p.18 === Chapter 2.3. --- Definitions of overweight and obesity in children and adolescents --- p.19 === Chapter 2.3.1. --- General definition of obesity --- p.19 === Chapter 2.3.2. --- Definitions and prevalence of childhood obesity in different countries --- p.21 === Chapter 2.3.3. --- Methods and current practices for identifying childhood obesity --- p.21 === Chapter 2.4. --- Methods for assessing body composition --- p.42 === Chapter 2.4.1. --- Body composition models --- p.42 === Chapter 2.4.2. --- Measurements methods to estimate body composition --- p.44 === Chapter 2.4.3. --- Specific issues of body fat in childhood --- p.60 === Chapter 2.5. --- Diagnostic accuracy of different definitions of childhood obesity --- p.61 === Chapter CHAPTER 3. --- STUDY I: VALIDITY OF BIOELECTRICAL IMPEDANCE MEASUREMENT IN PREDICTING FAT-FREE MASS OF CHINESE CHILDREN AND ADOLESCENTS --- p.67 === Chapter 3.1. --- Introduction --- p.67 === Chapter 3.2.1. --- Participants --- p.69 === Chapter 3.2.2. --- Measurements --- p.70 === Chapter 3.2.3. --- Data reduction and statistic analysis --- p.72 === Chapter 3.3. --- Results --- p.78 === Chapter 3.3.1. --- Descriptive statistics --- p.78 === Chapter 3.3.2. --- Cross-validation of published BIA equations --- p.79 === Chapter 3.3.3. --- Development of alternative BIA equations --- p.84 === Chapter 3.4. --- Discussion --- p.86 === Chapter 3.5. --- Conclusion --- p.92 === Chapter CHAPTER 4. --- STUDY II: VALIDITY OF FOUR COMMERCIAL BIA SCALES IN MEASURING BODY FAT AMONG CHINESE CHILDREN AND ADOLESCENTS --- p.93 === Chapter 4.1. --- Introduction --- p.93 === Chapter 4.2. --- Methods --- p.97 === Chapter 4.2.1. --- Participants --- p.97 === Chapter 4.2.2. --- Anthropometrics measurement --- p.97 === Chapter 4.2.3. --- Measurement of body composition --- p.97 === Chapter 4.2.4. --- Statistical analysis --- p.99 === Chapter 4.3. --- Results --- p.102 === Chapter 4.3.1. --- Characteristics of participants --- p.102 === Chapter 4.3.2. --- Reliability of BIA analysis --- p.103 === Chapter 4.3.3. --- Comparison of measuring body composition between BIA commercial devices (manufacturers’ equations) and DEXA measurement-Step 1 --- p.103 === Chapter 4.3.4. --- Comparison of measuring body composition between BIA commercial devices (adjusted equations) and DEXA measurement-Step 2 --- p.110 === Chapter 4.4. --- Discussion --- p.116 === Chapter 4.5. --- Conclusion --- p.123 === Chapter CHAPTER 5. --- STUDY III: DIAGNOSTIC ACCURACY OF DIFFERENT DEFINITIONS OF CHILDHOOD OBESITY IN IDENTIFYING OVERFAT AMONG CHINESE CHILDREN AND ADOLESCENTS --- p.124 === Chapter 5.1. --- Introduction --- p.124 === Chapter 5.2. --- Methods --- p.126 === Chapter 5.2.1. --- Participants --- p.126 === Chapter 5.2.2. --- Anthropometrics measurement --- p.127 === Chapter 5.2.3. --- Body fat measurement --- p.127 === Chapter 5.2.4. --- Definition of excess fatness --- p.128 === Chapter 5.2.5. --- Classification of participants --- p.128 === Chapter 5.2.6. --- Data reduction and statistic analysis --- p.129 === Chapter 5.3. --- Results --- p.131 === Chapter 5.3.1. --- Characteristics of participants --- p.131 === Chapter 5.3.2. --- Age-adjusted correlation among the different indices of obesity --- p.133 === Chapter 5.3.3. --- Prevalence rates of overweight/obesity --- p.134 === Chapter 5.3.4. --- Diagnostic agreement in assessing excess fat between %BF and anthropometric-based definitions --- p.137 === Chapter 5.3.5. --- Sensitivity and specificity --- p.138 === Chapter 5.3.6. --- Diagnostic performance of anthropometric indices in assessing excess fat --- p.141 === Chapter 5.3.7. --- Cut-offs of the anthropometric indices for screening excess fat --- p.143 === Chapter 5.4. --- Discussion --- p.148 === Chapter 5.5. --- Conclusion --- p.157 === Chapter CHAPTER 6. --- GENERAL DISCUSSION AND CONCLUSION --- p.158 === REFERENCE --- p.166 === APPENDIX A --- p.198 === APPENDIX B --- p.202 === APPENDIX C --- p.204 === APPENDIX D --- p.206 === APPENDIX E --- p.210