Reference | Outcomes & definitions used | Urinary BPA categorisation | Results | Adjustment in model(s) used for review |
---|---|---|---|---|
In children | ||||
Prevalent overweight (4 publications) | ||||
Eng et al. 2013 [38] | Overweight: BMI ≥85th percentile for age/gender | BPA in quartiles (ng/mL): Q1: <1.3; Q2: 1.3–2.6; Q3: 2.6–4.9; Q4: >4.9 | Q2 vs. Q1: OR = 1.00 (0.74–1.36) | Age, gender, race/ethnicity, urine creatinine, poverty-to-income ratio, serum cotinine as a marker of smoking status, soda consumption |
Q3 vs. Q1: OR = 1.17 (0.89–1.54) | ||||
Q4 vs. Q1: OR = 1.07 (0.80–1.44) | ||||
Harley et al. 2013 [64] | Overweight: BMI ≥85th percentile at 5 and 9 years of age | Ratio of BPA-to-creatinine level as continuous (log2- transformed) at 5 years | OR = 1.07 (0.90–1.28) | Maternal prepregnancy BMI, household income, maternal education level, maternal years of residence in the United States, child’s environmental tobacco smoke exposure, soda intake, fast food intake, and sweet consumption |
Ratio of BPA-to-creatinine level in 3 tertiles at 5 years (μg/g): | T2 vs. T1: OR = 0.80 (0.45–1.42) | |||
T1: <LOD-2.4; T2: 2.4–4.5 μg/g; T3: 4.6–349.8 μg/g | T3 vs. T1: OR = 1.36 (0.75–2.47) | |||
Ratio of BPA-to-creatinine level as continuous (log2- transformed) at 9 years | OR = 1.06 (0.85–1.33) | |||
Ratio of BPA-to-creatinine level in 3 groups at 9 years (μg/g): | G2 vs. G1: OR = 3.08 (1.18–8.02) | |||
G1: <LOD (<0.4); G2: detectable < median (0.4–1.8); G3: detectable > median (1.8–22.5) | G3 vs. G1: OR = 4.20 (1.60–11.02) | |||
Li et al. 2013 [49] | Overweight: age- and gender-specific weight >90th percentile of the underlying population | BPA in 2 classes (ng/mL): Low BPA level: <2 (reference); high BPA level: ≥2 (2 mg/L is about the median urine BPA level in the U.S. population) | Girls | Age, gender, school, residence, paternal and maternal education and overweight, playing video games, unbalanced diet, eating junk food, vegetables or fruit, depression scores, sports/activities |
All: OR=1.29 (0.83–2.01) | ||||
Age 9-12: OR= 2.32 (1.15-4.65) | ||||
Age>12: OR= 0.90 (0.48-1.72) | ||||
 | Boys | |||
All: OR=0.82 (0.55–1.23) | ||||
Age 9-12: OR= 0.71 (0.34-1.45) | ||||
Age>12: OR= 0.87 (0.52-1.45) | ||||
Trasande et al. 2012 [47] | Overweight: BMI z-score ≥1.036 (85th percentile for age/sex) | BPA continuous (log-transformed) | OR = 1.04 (0.92–1.18) | Sex, caloric intake, television watching, poverty to income ratio, parental education, serum cotinine level, urinary creatinine level, age, race/ethnicity categories |
BPA in quartiles (ng/mL): Q1: ≤1.5; Q2: 1.5–2.7; Q3: 2.8–5.5; Q4: >5.6 | Q2 vs. Q1: OR = 1.26 (0.96–1.64) | |||
Q3 vs. Q1: OR = 1.28 (0.98–1.66) | ||||
Q4 vs. Q1: OR = 1.26 (0.86–1.82) | ||||
Prevalent obesity (3 publications) | ||||
Bhandari et al. 2013 [35] | Obesity: BMI ≥ 95th percentile for age/gender | BPA continuous (log-transformed) | OR = 1.25 (1.09–1.43) | Age, sex, race/ethnicity, education, moderate activity, urinary creatinine, serum cotinine |
BPA in quartiles (ng/mL): Q1: <1.5; Q2: 1.5–2.7; Q3: 2.8–5.4; Q4: >5.4 | Q2 vs. Q1: OR = 2.35 (1.56–3.53) |  | ||
Q3 vs. Q1: OR = 1.78 (1.13–2.79) | ||||
Q4 vs. Q1: OR = 2.55 (1.65–3.95) | ||||
p for trend = 0.002 | ||||
Stratified analyses by sex (p for interaction = 0.07): association of strong magnitude and statistically significant among boys | ||||
Eng et al. 2013 [38] | Obesity: BMI ≥95th percentile for age/gender | BPA in quartiles (ng/mL): Q1: <1.3; Q2: 1.3–2.6; Q3: 2.6–4.9; Q4: >4.9 | Q2 vs. Q1: OR = 1.73 (1.16–2.58) | Age, gender, race/ethnicity, urine creatinine, poverty-to-income ratio, serum cotinine as a marker of smoking status, soda consumption |
Q3 vs. Q1: OR = 1.63 (1.08–2.46) | ||||
Q4 vs. Q1: OR = 2.05 (1.38–3.04) | ||||
Trasande et al. 2012 [47] | Obesity: BMI z-score ≥1.64 (95th percentile for age/sex) | BPA continuous (log-transformed) | Continuous: OR = 1.24 (1.08–1.44) | Sex, caloric intake, television watching, poverty-to-income ratio, parental education, serum cotinine level, urinary creatinine level, age, race/ethnicity categories |
BPA in quartiles (ng/mL): Q1: ≤1.5; Q2: 1.5–2.7; Q3: 2.8–5.5; Q4: >5.6 | Q2 vs. Q1: OR = 2.24 (1.54–3.24) | |||
Q3 vs. Q1: OR = 2.08 (1.46–2.96) | ||||
Q4 vs. Q1: OR = 2.57 (1.72–3.83) | ||||
Prevalent elevated waist circumference | ||||
Eng et al. 2013 [38] | Abnormal WC: WC ≥90th percentile for age/gender Abnormal WC-to-height ratio: WC-to-height ratio ≥0.5 | BPA in quartiles (ng/mL): Q1: <1.3; Q2: 1.3–2.6; Q3: 2.6–4.9; Q4: >4.9 | Abnormal WC | Age, gender, race/ethnicity, urine creatinine, poverty-to-income ratio, serum cotinine as a marker of smoking status, soda consumption |
Q2 vs. Q1: OR = 1.33 (0.90–1.97) | ||||
Q3 vs. Q1: OR = 1.16 (0.75–1.81) | ||||
Q4 vs. Q1: OR = 1.40 (0.91–2.15) | ||||
Abnormal WC-to-height ratio | ||||
Q2 vs. Q1: OR = 1.37 (0.97–1.92) | ||||
Q3 vs. Q1: OR = 1.41 (1.07–1.87) | ||||
Q4 vs. Q1: OR = 1.56 (1.11–2.17) | ||||
Other | ||||
Braun et al. 2014 [63] | Change in BMI z-score between 2 and 5 years of age, as continuous | Ratio of BPA-to-creatinine continuous (log10-transformed) | Per 10-fold increase: β = −0.2 (−0.6, 0.1) | Maternal race, marital status, parity, age at delivery, household income, education, employment, insurance, BMI at 16 weeks of pregnancy, depressive symptoms at baseline, prenatal serum cotinine |
Ratio of BPA-to-creatinine in tertiles (μg/g creatinine): | T2 vs. T1: β = 0.0 (−0.3, 0.3) | |||
T1: 2.1–11; T2: 11–20; T3: 20–314 | T3 vs. T1: β = −0.2 (−0.5, 0.1) | |||
Eng et al. 2013 [38] | Prevalent abnormal body fat: body fat ≥85th percentile for age/gender | BPA in quartiles (ng/mL): Q1: <1.3; Q2: 1.3–2.6; Q3: 2.6–4.9; Q4: >4.9 | Q2 vs. Q1: OR = 4.85 (0.80–21.4) | Age, gender, race/ethnicity, urine creatinine, poverty-to-income ratio, serum cotinine as a marker of smoking status, soda consumption |
Q3 vs. Q1: OR = 5.36 (0.71–43.3) | ||||
Q4 vs. Q1: OR = 2.10 (0.24–17.8) | ||||
Harley et al. 2013 [64] | Incident overweight: BMI ≥85th percentile at 9 years of age | Ratio of BPA-to-creatinine level at 5 years as continuous (log2- transformed) | OR = 1.02 (0.84–1.23) | Maternal prepregnancy BMI, household income, maternal education level, maternal years of residence in the United States, child’s environmental tobacco smoke exposure, soda intake, fast food intake, and sweet consumption at age 5 years |
Ratio of BPA-to-creatinine level at 5 years in tertiles (μg/g): | T2 vs. T1: 0.91 (0.48–1.73) | |||
T1: <LOD-2.4; T2: 2.4–4.5; T3: 4.6–349.8 | T3 vs. T1: 1.28 (0.65–2.51) | |||
Wang et al. 2012b [50] | Prevalent BMI as continuous (kg/m2) | BPA continuous (log-transformed and corrected for specific gravity) | β = 0.017 (0.002–0.032) | Age, sex |
Wells et al. 2013 [48] | Prevalent waist-to-height ratio as continuous | BPA in quartiles (ng/mL): Q1: <1.2; Q2: 1.2–2.6; Q3: 2.6–5.1; Q4: >5.1 | Change in waist-to-height ratio: | Urinary creatinine, age, sex, race/ethnicity, education, smoking status based on serum cotinine, caloric intake |
Q2 vs. Q1: β = 0.011 (0.001–0.020) | ||||
Q3 vs. Q1: β = 0.010 (0.001–0.019) | ||||
Q4 vs. Q1: β = 0.016 (0.007–0.026) | ||||
Significant associations among boys but not girls. | ||||
In adults | ||||
Prevalent overweight (3 publications) | ||||
Carwile & Michels 2011 [36] | Overweight: 25 ≤ BMI < 30 kg/m2 (reference: BMI <25 kg/m2) | BPA in quartiles (ng/mL): Q1: ≤1.1; Q2: 1.2–2.3; Q3: 2.4–4.6; Q4: ≥4.7 | Q2 vs. Q1: OR = 1.66 (1.21–2.27) | Age, gender, race/ethnicity, education, smoking status, urinary creatinine |
Q3 vs. Q1: OR = 1.26 (0.85–1.87) | ||||
Q4 vs. Q1: OR = 1.31 (0.80–2.14) | ||||
Kim et al. 2011 [55] | Overweight: BMI = 23-24.9 kg/m2, according to the WHO definitions for the Asian populations (reference: BMI <18.5 kg/m2) | BPA continuous (log-transformed) | Adjusted proportional change (95 % CI) = 1.01 (0.78–1.31) | Age, gender, education, income, cigarette smoking status, place of residence, urinary creatinine |
Wang et al. 2012a [52] | Generalized overweight: 24 ≤ BMI < 28 kg/m2, according to Chinese criteria (reference = BMI <24 kg/m2) | BPA in quartiles (ng/mL): Q1: ≤0.47; Q2: 0.48–0.81; Q3: 0.82–1.43; Q4: >1.43 | Q2 vs. Q1: OR = 1.23 (0.97–1.57) | Age, sex, urinary creatinine, smoking, alcohol drinking, education levels, systolic blood pressure, HDL cholesterol, LDL cholesterol, total cholesterol, TG, hsCRP, fasting plasma glucose, fasting serum insulin, serum ALT and GGT |
Q3 vs. Q1: OR = 1.28 (1.01–1.63) | ||||
Q4 vs. Q1: OR = 1.24 (0.97–1.59) | ||||
Prevalent obesity (4 publications) | ||||
Carwile & Michels 2011 [36] | Obesity: BMI ≥30 kg/m2 (reference: BMI <25 kg/m2) | BPA in quartiles (ng/mL): Q1: ≤1.1; Q2: 1.2–2.3; Q3: 2.4–4.6; Q4: ≥4.7 | Q2 vs. Q1: OR = 1.85 (1.22–2.79) | Age, gender, race/ethnicity, education, smoking status, urinary creatinine |
Q3 vs. Q1: OR = 1.60 (1.05–2.44) | ||||
Q4 vs. Q1: OR = 1.76 (1.06–2.94) | ||||
Kim et al. 2011 [55] | Obesity: BMI ≥25 kg/m2, according to the WHO definitions for the Asian populations (reference = BMI <18.5 kg/m2) | BPA continuous (log-transformed) | Adjusted proportional change (95 % CI) = 0.96 (0.75–1.23) | Age, gender, education, income, cigarette smoking status, place of residence, urinary creatinine |
Shankar et al. 2012 [44] | Obesity: BMI ≥30 kg/m2 (reference = BMI <30 kg/m2) | BPA in quartiles (ng/mL): Q1: <1.10; Q2: 1.10–2.10; Q3: 2.11–4.20; Q4: >4.20 | Q2 vs. Q1: OR = 1.40 (1.10–1.76) | Age, gender, race/ethnicity, education categories, smoking, alcohol consumption, physical inactivity, diabetes, hypertension, total cholesterol |
Q3 vs. Q1: OR = 1.59 (1.25–2.02) | ||||
Q4 vs. Q1: OR = 1.69 (1.30–2.20) | ||||
p for trend < 0.0001 | ||||
Associations still significant in analyses stratified by sex. | ||||
Wang et al. 2012a [52] | Generalized obesity: BMI ≥28 kg/m2, according to Chinese criteria (reference: BMI <28 kg/m2) | BPA in quartiles (ng/mL): Q1: ≤0.47; Q2: 0.48–0.81; Q3: 0.82–1.43; Q4: >1.43 | Q2 vs. Q1: OR = 1.14 (0.87–1.50) | Age, sex, urinary creatinine, smoking, alcohol drinking, education levels, systolic blood pressure, HDL cholesterol, LDL cholesterol, total cholesterol, TG, hsCRP, fasting plasma glucose, fasting serum insulin, serum ALT and GGT |
Q3 vs. Q1: OR = 1.19 (0.90–1.57) | ||||
Q4 vs. Q1: OR = 1.50 (1.15–1.97) | ||||
Prevalent elevated waist circumference (4 publications) | ||||
Carwile & Michels 2011 [36] | Elevated WC: WC ≥102 cm in men and WC ≥88 cm in women | BPA in quartiles (ng/mL): Q1: ≤1.1; Q2: 1.2–2.3; Q3: 2.4–4.6; Q4: ≥4.7 | Q2 vs. Q1: OR = 1.62 (1.11–2.36) | Age, gender, race/ethnicity, education, smoking status, urinary creatinine |
Q3 vs. Q1: OR = 1.39 (1.02–1.90) | ||||
Q4 vs. Q4: OR = 1.58 (1.03–2.42) | ||||
Ko et al. 2014 [57] | Abdominal obesity: WC ≥90 cm in men and WC ≥85 cm in women | BPA in quartiles (μg/mL) Q1: <0.853; Q2: 0.853–1.407; Q3: 1.407–2.594; Q4: >2.594 | Q2 vs. Q1: 1.117 (0.757–1.649) | Age, sex, urinary creatinine, education, income, alcohol consumption, smoking status |
Q3 vs. Q1: 1.337 (0.908–1.967) | ||||
Q4 vs. Q1: 1.938 (1.314–2.857) | ||||
p for trend = 0.01 | ||||
Shankar et al. 2012 [44] | Abdominal obesity: WC ≥102 cm in men and WC ≥88 cm in women | BPA in quartiles (ng/mL): Q1: <1.10; Q2: 1.10–2.10; Q3: 2.11–4.20; Q4: >4.20 | Q2 vs. Q1: OR = 1.63 (1.20–2.22) | Age, gender, race/ethnicity, education categories, smoking, alcohol consumption, physical inactivity, diabetes, hypertension, total cholesterol |
Q3 vs. Q1: OR = 1.66 (1.28–2.14) | ||||
Q4 vs. Q1: OR = 1.59 (1.21–2.09) | ||||
p for trend = 0.0009 | ||||
Associations still significant in analyses stratified by sex | ||||
Wang et al. 2012a [52] | Abdominal obesity: WC ≥90 cm in men and WC ≥85 cm in women | BPA in quartiles (ng/mL): Q1: ≤0.47; Q2: 0.48–0.81; Q3: 0.82–1.43; Q4: >1.43 | Q2 vs. Q1: OR = 1.26 (1.02–1.57) | Age, sex, urinary creatinine, smoking, alcohol drinking, education levels, systolic blood pressure, HDL cholesterol, LDL cholesterol, total cholesterol, TG, hsCRP, fasting plasma glucose, fasting serum insulin, serum ALT and GGT |
Q3 vs. Q1: OR = 1.28 (1.03–1.59) | ||||
Q4 vs. Q1: OR = 1.28 (1.03–1.60) | ||||
Other | ||||
Galloway et al. 2010 [58] | Prevalent WC as continuous Prevalent weight as continuous | Daily BPA excretion (μg/day) as a continuous variable | WC: β = 0.0062 (0.0016–0.0108) | Age, sex, study site |
Weight: β = 0.0064 (0.0023–0.0104) | ||||
Kim et al. 2011 [55] | Prevalent normal weight: BMI = 18.5–22.9 kg/m2, according to the WHO definitions for the Asian populations (reference: BMI < 18.5 kg/m2) | BPA continuous (log-transformed) | Adjusted proportional change (95 % CI) = 0.92 (0.72–1.17) | Age, gender, education, income, cigarette smoking status, place of residence, urinary creatinine |
Song et al. 2014 [62] | Weight change rate (WCR) during follow-up (kg/year) | BPA in quartiles (nmol/L): | Â | Age at baseline, urinary creatinine concentration, cohort origin, menopausal status, smoking, alcohol consumption, physical activity, alternative healthy eating index, total energy intake |
Q1: median (IQR) = 3.6 (2.6–4.5) | Q2 vs. Q1: WCR = 0.15 (0.00–0.31) | |||
Q2: median (IQR) = 6.4 (5.8–7.3) | Q3 vs. Q1: WCR = 0.18 (0.03–0.34) | |||
Q3: median (IQR) = 10.5 (9.0–12.1) | Q4 vs. Q1: WCR = 0.23 (0.07–0.38) | |||
Q4: median (IQR) = 21.9 (16.8–35.7) | p for trend = 0.02 | |||
Zhao et al. 2012 [53] | Fat mass, fat-free mass, body weight, BMI, WC, hip circumference, waist-hip ratio (all variables as continuous) | BPA continuous | Fat mass: r = 0.35 (p < 0.001) | Age |
Fat-free mass: r = 0.186 (p = 0.009) | ||||
Body weight: r = 0.24 (p = 0.001) | ||||
BMI: r = 0.298 (p < 0.001) | ||||
WC: r = 0.296 (p < 0.001) | ||||
Hip circumference: r = 0.27 (p < 0.001) | ||||
Waist-hip ratio: r = 0.149 (p = 0.035) | ||||
With additional adjustment for age and BMI, BPA was still significantly associated with fat mass (r = 0.193, p = 0.006) but not with fat-free mass. |