Limits...
A novel quantitative body shape score for detecting association between obesity and hypertension in China.

Wang S, Liu Y, Li F, Jia H, Liu L, Xue F - BMC Public Health (2015)

Bottom Line: Totally 15,172 (6,939 male and 8,233 female) participants aged from 18 to 87 years old were included.PLSPM method illustrated the biggest path coefficients (95% confidence interval, CI) were 0.220(0.196, 0.244) for male and 0.205(0.182, 0.228) for female in model of BSS1.The area under receiver-operating characteristic curve (AUC(95% CI)) of BSS1(0.839(0.831,0.847)) was significantly larger than that of BSS2(0.834(0.825,0.842)) as well as the four single indices for female, and similar trend can be found for male.

View Article: PubMed Central - PubMed

Affiliation: Department of Epidemiology and Health Statistics, School of Public Health, Shandong University, PO Box 100, Jinan, 250012, China. wsk2001@sdu.edu.cn.

ABSTRACT

Background: Obesity is a major independent risk factor for chronic diseases such as hypertension and coronary diseases, it might not be only related to the amount of body fat but its distribution. The single body mass index (BMI), waist circumference (WC), waist to hip ratio (WHR) or waist to stature ratio (WSR) provides limited information on fat distribution, and the debate about which one is the best remained. On the other hand, the current classification of body shape is qualitative rather than quantitative, and only crudely measure fat distribution. Therefore, a synthetical index is highly desirable to quantify body shape.

Methods: Based on the China Health and Nutrition Survey (CHNS) data, using Lohmäller PLSPM algorithm, six Partial Least Squares Path Models (PLSPMs) between the different obesity measurements and hypertension as well as two synthetical body shape scores (BSS1 by BMI/WC/Hip circumference, BSS2 by BMI/WC/WHR/WSR) were created. Simulation and real data analysis were conducted to assess their performance.

Results: Statistical simulation showed the proposed model was stable and powerful. Totally 15,172 (6,939 male and 8,233 female) participants aged from 18 to 87 years old were included. It indicated that age, height, weight, WC, WHR, WSR, SBP, DBP, the prevalence of hypertension and obesity were significantly sex-different. BMI, WC, WHR, WSR, Hip, BSS1 and BSS2 between hypertension and normotensive group are significantly different (p < 0.05). PLSPM method illustrated the biggest path coefficients (95% confidence interval, CI) were 0.220(0.196, 0.244) for male and 0.205(0.182, 0.228) for female in model of BSS1. The area under receiver-operating characteristic curve (AUC(95% CI)) of BSS1(0.839(0.831,0.847)) was significantly larger than that of BSS2(0.834(0.825,0.842)) as well as the four single indices for female, and similar trend can be found for male.

Conclusions: BSS1 was an excellent measurement for quantifying body shape and detecting the association between body shape and hypertension.

Show MeSH

Related in: MedlinePlus

PLSPM-based models. The six models for detecting the association between obesity and hypertension with different manifest anthropometric measurements, with (a) using BMI/WC/Hip, (b) BMI/WC/WHR/WSR, (c) BMI; (d) WC; (e) WHR; (f) WSR.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4308906&req=5

Fig1: PLSPM-based models. The six models for detecting the association between obesity and hypertension with different manifest anthropometric measurements, with (a) using BMI/WC/Hip, (b) BMI/WC/WHR/WSR, (c) BMI; (d) WC; (e) WHR; (f) WSR.

Mentions: Figure 1 showed the model structure between hypertension and body shape under framework of PLSPM using different manifest anthropometric measurements, with (a) by BMI/WC/Hip, (b) by BMI/WC/WHR/WSR, (c) by BMI; (d) by WC; (e) by WHR; (f) by WSR. Figure 1(a) defined a latent quantitative measurement of body shape score (ξ1) extracted from BMI, WC and Hip (BSS1), while BSS2 from BMI, WC, WHR and WSR (Figure 1b). The latent score of blood pressure (BPS, ξ2) was summarized by binary variable of hypertension. We exemplified Figure 1(a) to explain the variables and their relationships : (1) The variables in the squared boxes were the manifest variables actually measured such as BP, age, BMI, WC and Hip; (2) The variables in the ellipses were the latent variables such as BPS(ξ2), BSS1(ξ1) and age(ξ3)calculated through the corresponding manifest variables, (3) The latent variable BSS1(ξ1) corresponded to the manifest variables BMI, WC and Hip, while the latent variable BPS(ξ2)to BP, and (ξ3)to the single variable age. Three types of parameters were included: (1) Latent variable scores (ξ) as combinations of their manifest variables obtained iteratively from an ordinary least squares (OLS)-type algorithm; (2) path coefficients (β) between dependent (ξ2) and independent latent variable (ξ1) by OLS or partial least squares (PLS); (3) loadings (λ) of each block of manifest variables with its latent variables by OLS, and the arrows was from the manifest variable to the latent variable when single manifest exist, otherwise the arrow direction is opposite. In this paper, the Lohmäller PLSPM algorithm was used [22].Figure 1


A novel quantitative body shape score for detecting association between obesity and hypertension in China.

Wang S, Liu Y, Li F, Jia H, Liu L, Xue F - BMC Public Health (2015)

PLSPM-based models. The six models for detecting the association between obesity and hypertension with different manifest anthropometric measurements, with (a) using BMI/WC/Hip, (b) BMI/WC/WHR/WSR, (c) BMI; (d) WC; (e) WHR; (f) WSR.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4308906&req=5

Fig1: PLSPM-based models. The six models for detecting the association between obesity and hypertension with different manifest anthropometric measurements, with (a) using BMI/WC/Hip, (b) BMI/WC/WHR/WSR, (c) BMI; (d) WC; (e) WHR; (f) WSR.
Mentions: Figure 1 showed the model structure between hypertension and body shape under framework of PLSPM using different manifest anthropometric measurements, with (a) by BMI/WC/Hip, (b) by BMI/WC/WHR/WSR, (c) by BMI; (d) by WC; (e) by WHR; (f) by WSR. Figure 1(a) defined a latent quantitative measurement of body shape score (ξ1) extracted from BMI, WC and Hip (BSS1), while BSS2 from BMI, WC, WHR and WSR (Figure 1b). The latent score of blood pressure (BPS, ξ2) was summarized by binary variable of hypertension. We exemplified Figure 1(a) to explain the variables and their relationships : (1) The variables in the squared boxes were the manifest variables actually measured such as BP, age, BMI, WC and Hip; (2) The variables in the ellipses were the latent variables such as BPS(ξ2), BSS1(ξ1) and age(ξ3)calculated through the corresponding manifest variables, (3) The latent variable BSS1(ξ1) corresponded to the manifest variables BMI, WC and Hip, while the latent variable BPS(ξ2)to BP, and (ξ3)to the single variable age. Three types of parameters were included: (1) Latent variable scores (ξ) as combinations of their manifest variables obtained iteratively from an ordinary least squares (OLS)-type algorithm; (2) path coefficients (β) between dependent (ξ2) and independent latent variable (ξ1) by OLS or partial least squares (PLS); (3) loadings (λ) of each block of manifest variables with its latent variables by OLS, and the arrows was from the manifest variable to the latent variable when single manifest exist, otherwise the arrow direction is opposite. In this paper, the Lohmäller PLSPM algorithm was used [22].Figure 1

Bottom Line: Totally 15,172 (6,939 male and 8,233 female) participants aged from 18 to 87 years old were included.PLSPM method illustrated the biggest path coefficients (95% confidence interval, CI) were 0.220(0.196, 0.244) for male and 0.205(0.182, 0.228) for female in model of BSS1.The area under receiver-operating characteristic curve (AUC(95% CI)) of BSS1(0.839(0.831,0.847)) was significantly larger than that of BSS2(0.834(0.825,0.842)) as well as the four single indices for female, and similar trend can be found for male.

View Article: PubMed Central - PubMed

Affiliation: Department of Epidemiology and Health Statistics, School of Public Health, Shandong University, PO Box 100, Jinan, 250012, China. wsk2001@sdu.edu.cn.

ABSTRACT

Background: Obesity is a major independent risk factor for chronic diseases such as hypertension and coronary diseases, it might not be only related to the amount of body fat but its distribution. The single body mass index (BMI), waist circumference (WC), waist to hip ratio (WHR) or waist to stature ratio (WSR) provides limited information on fat distribution, and the debate about which one is the best remained. On the other hand, the current classification of body shape is qualitative rather than quantitative, and only crudely measure fat distribution. Therefore, a synthetical index is highly desirable to quantify body shape.

Methods: Based on the China Health and Nutrition Survey (CHNS) data, using Lohmäller PLSPM algorithm, six Partial Least Squares Path Models (PLSPMs) between the different obesity measurements and hypertension as well as two synthetical body shape scores (BSS1 by BMI/WC/Hip circumference, BSS2 by BMI/WC/WHR/WSR) were created. Simulation and real data analysis were conducted to assess their performance.

Results: Statistical simulation showed the proposed model was stable and powerful. Totally 15,172 (6,939 male and 8,233 female) participants aged from 18 to 87 years old were included. It indicated that age, height, weight, WC, WHR, WSR, SBP, DBP, the prevalence of hypertension and obesity were significantly sex-different. BMI, WC, WHR, WSR, Hip, BSS1 and BSS2 between hypertension and normotensive group are significantly different (p < 0.05). PLSPM method illustrated the biggest path coefficients (95% confidence interval, CI) were 0.220(0.196, 0.244) for male and 0.205(0.182, 0.228) for female in model of BSS1. The area under receiver-operating characteristic curve (AUC(95% CI)) of BSS1(0.839(0.831,0.847)) was significantly larger than that of BSS2(0.834(0.825,0.842)) as well as the four single indices for female, and similar trend can be found for male.

Conclusions: BSS1 was an excellent measurement for quantifying body shape and detecting the association between body shape and hypertension.

Show MeSH
Related in: MedlinePlus