The ADRB3 rs4994 polymorphism increases risk of childhood and adolescent overweight/obesity for East Asia’s population: an evidence-based meta-analysis

Figures & data

Table 1. The characteristics of included studies in this meta-analysis

Author	Year	Ref ID	Country	Region	Genotyping method	Obese/Overweight definition	Case	Control	HWE	NOS
Aradillas-Garcia	2017	[29]	Mexico	Latin-America	TaqMan	BMI	348	698	0.23	7
Verdi	2015	[27]	Turkey	West Asia	PCR-RFLP	BMI	130	121	<0.01	5
Kuo	2015	[26]	China	East Asia	TaqMan	BMI	1924	3901	NA	5
Oguri	2013	[24]	Japan	East Asia	PCR-RFLP	BMI	73	59	0.83	5
Zhu	2013	[25]	China	East Asia	PCR-RFLP	BMI	92	71	0.56	5
Csernus	2013	[23]	Hungary	Europe	PCR-RFLP	BMI	703	634	NA	5
Chou	2012	[22]	China	East Asia	TaqMan	BMI	276	277	0.47	6
Peng	2010	[21]	China	East Asia	PCR-RFLP	WHO weight height chart	357	357	0.18	7
Wang	2008	[20]	China	East Asia	PCR-RFLP	BMI	151	85	0.10	5
Zhang	2008	[30]	China	East Asia	PCR-RFLP	Obesity index	95	85	NA	5
Li	2007	[19]	China	East Asia	PCR-RFLP	WHO weight height chart	100	100	0.91	6
Erhardt	2005	[18]	Hungary	Europe	PCR-RFLP	Body weight & body fat content	295	147	0.54	6
Ochoa	2004	[17]	Spain	Europe	PCR-RFLP	BMI	185	185	0.25	5
Mo	2001	[16]	China	East Asia	PCR-RFLP	NA	90	87	NA	6
Endo	2000	[15]	Japan	East Asia	PCR-RFLP	Obesity index	90	463	0.80	5
Hinney	1996	[14]	Germany	Europe	PCR-RFLP	BMI	238	80	0.43	5

HWE: Hardy–Weinberg equilibrium; NOS: Newcastle–Ottawa scale; NA: not available. BMI: body mass index, calculated according to the weight/height² (kg/m²) formula. Obesity index: calculated according to the (real weight-standard weight)/standard weight * 100 formula.

Figure 1. Systematic review flowchart for this meta-analysis

Table 2. Results of overall and subgroup analyses for the ADRB3 rs4994 polymorphism and risk of childhood and adolescent overweight/obesity

	Allele model			Homozygote model			Heterozygote model			Dominant model			Recessive model
Comparison	OR (95% CI)	P	Ph	OR (95% CI)	P	Ph	OR (95% CI)	P	Ph	OR (95% CI)	P	Ph	OR (95% CI)	P	Ph
Overall	1.23(1.10,1.38)	<0.01	0.24	1.36(0.90,2.06)	0.15	0.56	1.39(1.16,1.68)	<0.01	0.08	1.31(1.12,1.54)	<0.01	0.05	1.26(0.83,1.90)	0.28	0.64
Region
East Asia	1.47(1.25,1.71)	<0.01	0.80	1.97(1.09,3.56)	0.02	0.76	1.60(1.34,1.90)	<0.01	0.57	1.50(1.22,1.84)	<0.01	0.03	1.68(0.94,3.00)	0.08	0.73
Others	1.04(0.89,1.22)	0.62	0.74	0.92(0.50,1.69)	0.79	0.54	1.01(0.82,1.26)	0.90	0.46	1.01(0.82,1.24)	0.96	0.60	0.92(0.50,1.68)	0.79	0.51
HWE
Consistent	1.25(1.11,1.41)	<0.01	0.15	1.45(0.95,2.21)	0.09	0.63	1.28(1.11,1.47)	<0.01	0.12	1.34(1.11,1.62)	<0.01	0.09	1.34(0.88,2.04)	0.18	0.71
Others*	1.17(0.90,2.52)	0.25	0.76	0.33(0.03,3.23)	0.34	NA	2.16(1.32,3.53)	<0.01	0.66	1.11(0.99,1.25)	0.08	0.14	0.30(0.03,2.97)	0.31	NA
Sample Size
<200	1.82(1.16,2.84)	<0.01	0.84	1.27(0.30,5.32)	0.74	0.13	2.30(1.54,3.45)	<0.01	0.76	2.05(1.39,3.02)	<0.01	0.87	1.02(0.25,4.17)	0.98	0.10
≥200	1.20(1.07,1.35)	<0.01	0.30	1.37(0.89,2.11)	0.16	0.61	1.24(1.07,1.43)	<0.01	0.27	1.15(1.05,1.26)	<0.01	0.16	1.28(0.83,1.97)	0.26	0.74
Sex
male	1.31(1.03,1.67)	0.03	0.75	1.45(0.46,4.54)	0.53	0.93	1.38(1.05,1.83)	0.02	0.56	1.44(1.12,1.85)	<0.01	0.54	1.35(0.43,4.21)	0.61	0.91
female	1.74(1.29,2.35)	<0.01	0.63	2.47(0.54,11.18)	0.24	0.83	1.90(1.36,2.66)	<0.01	0.39	2.01(1.48,2.73)	<0.01	0.45	1.98(0.44,8.97)	0.37	0.81

HWE: Hardy–Weinberg equilibrium; OR: odds ratio; 95% CI: 95% confidence interval; NA: not available. * Studies inconsistent with HWE and unassessable were classified into subgroup ‘Others’.

Figure 2. Forest plot showing the association between the ADRB3 rs4994 polymorphism and risk of childhood and adolescent overweight/obesity. Odds ratio (OR) and 95% confidence interval (95% CI) were calculated under the dominant model. The random effects model was used to assess the pooled estimates. The included studies were stratified according to geographical regions. The grey squares represent the weight of the sample size under the random effects model. The black dot in the square represented the OR for each included study. The black horizontal lines showed the corresponding 95% CI. The black solid vertical line showed the null effect (OR = 1). The red dashed vertical line showed the pooled estimate. The hollow diamonds at the bottom represented the ORs and 95% CIs for the overall population and subgroups. The data sets drawn from ref[23] and ref[30] were not included because they did not provide genotype data for the dominant model

Figure 3. Filled funnel plot with imputed studies under the dominant model. Using the ‘trim-and-fill’ method, the pooled estimates are adjusted for possible missing data sets (squared circles) amongst published studies (hollow circles). The log odds ratio (OR) stands for the natural logarithm transferred OR of individual data sets. The standard error of the log OR represents the standard error of the natural logarithm transferred OR of individual data sets. The data sets drawn from ref[23] and ref[30] were not included because they did not provide genotype data for the dominant model

Figure 4. Sensitivity analysis for the pooled estimates under the dominant model. For each omitted data set listed on the left, summary statistics for the resulting pooled estimates are presented as odds ratio (OR, hollow circle) with 95% confidence interval (CI, horizontal line). The random effects model was used to assess pooled estimates. The data sets drawn from ref[23] and ref[30] were not included because they did not provide genotype data for the dominant model