You are here
Home > Human Fattening > Relative Fat Mass as an estimator of whole-body fat percentage among children and adolescents: A cross-sectional study using NHANES

Relative Fat Mass as an estimator of whole-body fat percentage among children and adolescents: A cross-sectional study using NHANES


Study population

The final sample for analyses included 10,390 girls and boys. Characteristics of the study participants are shown in Table 1. Mean body fat percentage measured by dual energy X-ray absorptiometry (DXA) was 32.2 ± 0.3% and 27.3 ± 0.2% in girls and boys 8 to 14 years of age, respectively, and 34.4 ± 0.2% and 22.8 ± 0.2% in girls and boys 15 to 19 years of age, respectively.

Table 1 Characteristics of children and adolescents included in the study.

Derivation of equations for pediatric populations

As shown in Supplementary Fig. 1, body weight, height, waist circumference, and body composition substantially differed across age categories. The RFM equation that we previously developed for adults17 may be used to estimate whole-body fat percentage among adolescents 15 to 19 years of age. For younger children and adolescents (8 to 14 years of age), we developed a modified RFM equation named as the RFM pediatric (RFMp). RFM equations for pediatric populations are as follows:

$$rmEquation,rmfor,rmgirls,8,rmto,14,rmyears,rmof,rmage:79-(22times rmheight/rmwaist)$$

(1)

$$rmEquation,rmfor,rmboys,8,rmto,14,rmyears,rmof,rmage:74-(22times rmheight/rmwaist)$$

(2)

or

$$rmRFMp,rmfor,rmgirls,rmand,rmboys,8,rmto,14,rmyears,rmof,rmage:74-(22times (rmheight/rmwaist))+(5times rmsex)$$

(3)

$$rmRFM,rmfor,rmgirls,rmand,rmboys,15,rmto,19,rmyears,rmof,rmage:64-(20times rmheight/rmwaist)+(12times rmsex)$$

(4)

In all equations, height and waist (circumference) are in the same units. U.S./English units or Metric units may be used. In (3) and (4), sex equals 0 for boys and 1 for girls. For comparison purposes, we also derived equations based on BMI, TMI, and WHtR using the same NHANES dataset. Note that for a fair comparison among indices, we derived sex-specific equations for two groups: 1) for children and adolescents 8 to 14 years of age and 2) for adolescents 15 to 19 years of age. All raw (non-rounded) equations are shown in Supplementary Table 1. Since the relationship of BMI and TMI with body fat percentage is nonlinear16, quadratic equations were developed for BMI and TMI. Plots of RFMp and RFM as a function of age are shown in Supplementary Fig. 2A,C for girls and boys 8 to 14 years of age and in Supplementary Fig. 2B,D for girls and boys 15 to 19 years of age. In younger boys, there is a progressive subtle decrease in DXA-measured body fat percentage as age increases, which is paralleled by RFMp but not BMI (Supplementary Fig. 2C).

Prediction of body fat percentage in children and adolescents 8 to 14 years of age

Compared with BMI, RFMp showed a better linear relationship with DXA whole-body fat percentage among girls (RFMp: R2, 0.74; 95% confidence interval (CI), 0.72−0.77; root mean squared error (RMSE): 3.83%; 95% CI, 3.70%-3.97%; BMI: R2, 0.65; 95% CI, 0.62−0.68; RMSE: 4.15%; 95% CI, 4.00%-4.29%) and boys (RFMp: R2, 0.77; 95% CI, 0.75−0.79; RMSE: 3.83%; 95% CI, 3.70%-3.97%; BMI: R2, 0.55; 95% CI, 0.52−0.57; RMSE: 5.34%; 95% CI, 5.19%-5.49%). RFMp also showed a better linear relationship with DXA whole-body fat percentage compared with TMI among girls (TMI: R2, 0.71; 95% CI, 0.69−0.73; RMSE: 3.77%; 95% CI, 3.64%-3.90%). RFMp was a better predictor than TMI among boys (TMI: R2, 0.69; 95% CI, 0.66−0.71; RMSE: 4.43%; 95% CI, 4.29%-4.56%) (Fig. 1). Body fat linear prediction showed some variability with age, regardless the index used (Supplementary Table 2). RFMp was better than BMI in predicting body fat percentage across ethnic groups in girls (Supplementary Fig. 3A–F) and boys (Supplementary Fig. 4A–F).

Figure 1

Linear relationship between DXA-measured and estimated body fat percentage among children and adolescents 8 to 14 years of age. BFP, body fat percentage; BMI, body mass index (weight/height2); RFMp, Relative Fat Mass pediatric, which is based on height/waist circumference. R2, coefficient of determination; RMSE, root mean squared error; TMI, tri-ponderal mass index (weight/height3); WHtR, waist-to-height ratio. Data plots correspond to DXA imputation 1.

RFMp and WHtR showed similar predicting ability for whole-body fat percentage among girls (WHtR: R2, 0.74; 95% CI 0.72−0.76; RMSE: 3.60%; 95% CI, 3.48–3.71%) and boys (WHtR: R2, 0.76; 95% CI, 0.75−0.78; RMSE: 3.85%; 95% CI 3.72–3.99%). However, RFMp was a better predictor than WHtR among African-American girls (RFMp: R2, 0.76; 95% CI 0.72−0.79; RMSE: 3.80%; 95% CI, 3.61–3.99%; WHtR: R2, 0.73; 95% CI 0.69−0.77; RMSE: 4.00%; 95% CI, 3.78–4.21%).

Prediction of body fat percentage in adolescents 15 to 19 years of age

Compared with BMI, RFM equation for adults showed a better linear relationship with DXA whole-body fat percentage among boys (RFM: R2, 0.79; 95% CI, 0.76−0.81; RMSE: 3.35%; 95% CI, 3.22–3.49%; BMI: R2, 0.70; 95% CI, 0.67−0.73; RMSE: 3.97%; 95% CI, 3.83–4.12%) but not girls (RFM: R2, 0.72; 95% CI, 0.70−0.75; RMSE: 3.45%; 95% CI, 3.26–3.63%; BMI: R2, 0.73; 95% CI, 0.70−0.75; RMSE: 3.59%; 95% CI, 3.45–3.73%). RFM also showed a better linear relationship with DXA whole-body fat percentage compared with TMI among boys (TMI: R2, 0.69; 95% CI, 0.66−0.72; RMSE: 4.05%; 95% CI, 3.90–4.19%) but not girls (TMI: R2, 0.72; 95% CI, 0.69−0.74; RMSE: 3.64%; 95% CI, 3.51–3.78%) (Fig. 2). All indices showed variable prediction with age (Supplementary Table 3). RFM was better than BMI in predicting body fat percentage across ethnic groups among boys (Supplementary Fig. 4G–L) but not girls (Supplementary Fig. 3G–L). Highest prediction of body fat percentage by RFM was found among African-American adolescents (Supplementary Figs 3L and 4L).

Figure 2
figure2

Linear relationship between DXA-measured and estimated body fat percentage among adolescents 15 to 19 years of age. BFP, body fat percentage; BMI, body mass index; RFM, Relative Fat Mass. R2, coefficient of determination; RMSE, root mean squared error; TMI, tri-ponderal mass index; WHtR, waist-to-height ratio. Data plots correspond to DXA imputation 1.

WHtR appeared to have slightly lower predicting ability than RFM among girls (WHtR: R2, 0.70; 95% CI 0.68−0.73; RMSE: 3.75%; 95% CI, 3.58–3.91%) but not boys (WHtR: R2, 0.80; 95% CI, 0.78−0.82; RMSE: 3.25%; 95% CI 3.12–3.39%). However, RFM was a better predictor than WHtR among African-American girls (RFM: R2, 0.80; 95% CI 0.77−0.83; RMSE: 3.57%; 95% CI, 3.35–3.80%; WHtR: R2, 0.75; 95% CI 0.72−0.79; RMSE: 3.96%; 95% CI, 3.70–4.21%).

Agreement between DXA-measured and estimated whole-body fat percentage

Concordance analyses for RFMp and RFM linear equations were performed using their rounded and simplest expressions as indicated in (3) and (4) and compared with the raw quadratic equations for BMI and TMI (Supplementary Table 1). Among girls 8 to 14 years of age, the concordance correlation coefficients (ρc) were 0.85, 0.84 and 0.79 for RFMp, TMI and BMI equations, respectively. Among boys 8 to 14 years of age, RFMp appeared to show better agreement with DXA (ρc: 0.86) than did TMI equation (ρc: 0.80) or BMI (ρc: 0.70). Among girls 15 to 19 years of age, RFM, TMI and BMI equations showed good agreement with DXA. Concordance correlation coefficients were 0.83, 0.84 and 0.85, respectively. Among boys 15 to 19 years of age, RFM appeared to show better agreement with DXA (ρc: 0.86) than did TMI equation (ρc: 0.81) or BMI (ρc: 0.82). Overall, Bland-Altman plots showed good agreement between DXA-measured body fat percentage and RFMp-estimated body fat percentage among girls and boys 8 to 14 years of age of different ethnicities (Supplementary Fig. 5). Likewise, we found good agreement between DXA-measured body fat percentage and RFM-estimated body fat percentage among adolescents 15 to 19 years of age of different ethnicities (Supplementary Fig. 6).

RFMp performance among children and adolescents 8 to 14 years of age

RFMp linear equation showed higher accuracy than BMI quadratic equation to estimate whole-body fat percentage among girls (RFMp: 88.2%; 95% CI, 86.5–89.9%: BMI: 85.7%; 95% CI, 83.7–87.6%; P = 0.027) and boys (RFMp: 83.4%; 95% CI, 81.5–85.4%; BMI: 71.0%; 95% CI, 68.8–73.3%; P < 0.001). TMI quadratic equation had similar accuracy to RFMp among girls (TMI: 88.7%; 95% CI, 87.3–90.0%;P = 0.521) but was less accurate than RFMp among boys (TMI: 77.3%; 95% CI, 75.1–79.4%; P < 0.001). As indicated by a smaller interquartile range of the difference between DXA-measured and estimated whole-body fat percentage, RFMp was also more precise than BMI among girls (RFMp: 5.00%; 95% CI, 4.73–5.28%; BMI: 5.71%; 95% CI, 5.31–6.12%) and boys (RFMp: 5.01%; 95% CI, 4.67–5.35%; BMI: 7.16%; 95% CI, 6.72–7.59%). Likewise, RFMp was more precise than TMI among girls (TMI: 5.21%; 95% CI, 4.87–5.55%) and boys (TMI: 5.98%; 95% CI, 5.60–6.35%) (Supplementary Table 4). Overall, RFMp was more accurate than BMI across ethnic groups (Supplementary Table 4). RFMp also showed higher accuracy than BMI across body fat ranges among boys (Supplementary Fig. 7E) and across higher quartiles among girls (Supplementary Fig. 7B), but accuracy was lower in leaner individuals (Supplementary Fig. 7). Accuracy of RFMp was also more consistent than that of BMI across age groups (Supplementary Fig. 8).

RFM performance among adolescents 15 to 19 years of age

RFM linear equation showed higher accuracy than BMI quadratic equation among boys (RFM: 82.3%; 95% CI, 80.3–84.2% vs. BMI: 73.9%; 95% CI, 71.3–76.5%; P < 0.001) but was less accurate than BMI among girls (RFM: 89.0%; 95% CI, 86.7–91.2% vs. BMI: 92.6%; 95% CI, 91.1–94.1%; P = 0.002). RFM was less accurate than BMI among European-American girls (P = 0.002) and African-American girls (P = 0.015) but not Mexican-American girls (P = 0.384) (Supplementary Table 5). Conversely, RFM was more accurate than BMI across male ethnic groups (Supplementary Table 5) and across age groups among boys (Supplementary Fig. 8). TMI was less accurate than RFM among boys (TMI: 72.8%; 95% CI, 70.0–75.6%; P < 0.001) but more accurate than RFM among girls (TMI: 91.5%; 95% CI, 90.0–93.0%;P = 0.028), but only among European-American girls (RFM: 89.2%; 95% CI, 86.1–92.2%; TMI: 92.6%; 95% CI, 90.3–94.8%; P = 0.015). In fact, RFM was more accurate than TMI among African-American girls (RFM: 86.7%; 95% CI, 84.2–89.2%; TMI: 84.1%; 95% CI, 80.9–87.2%; P = 0.001) (Supplementary Table 5). Overall, RFM showed similar accuracy to BMI and TMI at higher body fat ranges but RFM accuracy was lower than BMI and TMI in leaner girls and higher in leaner boys (Supplementary Fig. 9B,E). RFM was also more precise than BMI among boys (4.51%; 95% CI, 4.21–4.80% vs. 5.09%; 95% CI, 4.79–5.39%) but not girls (4.64%; 95% CI, 4.35–4.93% vs. 4.78%; 95% CI, 4.52–5.05%) (Supplementary Table 5). Likewise, RFM was more precise than TMI among boys (TMI: 5.25%; 95% CI, 4.94–5.56%) but not girls (TMI: 4.76%; 95% CI, 4.43–5.08%).

Overweight and obesity misclassification among children and adolescents 8 to 14 years of age

Analyses were performed with data from 5,395 children and adolescents 96 to 179 months old (2,355 girls). For TMI, overweight was defined as ≥17.0 kg/m3 and <19.8 kg/m3 for girls, and ≥16.2 kg/m3 and <19.2 kg/m3 for boys. Obesity was defined as a TMI ≥ 19.8 kg/m3 for girls and ≥19.2 kg/m3 for boys. For RFMp, overweight was defined as ≥40.0% and <44.0% for girls, and ≥34.5% and <39.3% for boys. Obesity was defined as an RFMp ≥ 44.0% for girls and ≥39.3% for boys. The aforementioned RFMp thresholds were arbitrarily chosen for comparison purposes only based on the commonly used 85th and 95th percentiles and are not intended to suggest RFMp cutoffs for the diagnosis of overweight and obesity.

RFMp showed similar false negative rate for overweight or obesity (defined as a DXA-measured body fat percentage ≥85th percentile) compared with BMI-for-age percentiles among girls (RFMp: 26.7%; 95% CI, 20.6–32.8%; BMI-for-age: 25.8%; 95% CI, 18.9–32.8%; P = 0.783) and boys (RFMp: 21.8%; 95% CI, 17.0–26.6%; BMI-for-age: 25.8%; 95% CI, 19.1–32.6%; P = 0.208) (Table 2). RFMp showed lower false positive rate among boys (RFMp: 3.7%; 95% CI, 2.8–4.7%; BMI-for-age: 4.7%; 95% CI, 3.7–5.7%; P = 0.045), but not girls (RFMp: 5.0%; 95% CI, 3.7–6.2%; BMI-for-age: 4.8%; 95% CI, 3.2–6.3%; P = 0.742) (Table 2).

Table 2 Positive and negative cases of DXA-diagnosed overweight or obesity for RFMp and RFM among children and adolescents 96–239 months old (n = 10,260).

RFMp and BMI-for-age percentiles had similar total misclassification error rates of overweight (girls: P = 0.685; boys: P = 0.322) and obesity (girls: P = 0.631; boys: P = 0.192) (Fig. 3). RFMp had lower misclassification error rate of overweight or obesity than BMI-for-age percentiles among boys (RFMp: 6.5%; 95% CI, 5.3–7.6%; BMI-for-age: 7.9%; 95% CI, 6.4–9.4%; P = 0.018) but not girls (RFMp: 8.2%; 95% CI, 6.8–9.6%; BMI-for-age: 7.9%; 95% CI, 6.0–9.8%; P = 0.681). RFMp and TMI had similar rates of misclassification of overweight (girls: P = 0.729; boys: P = 0.412) and obesity (girls: P = 0.657; boys: P = 0.887). Conversely, RFMp had lower misclassification error of overweight or obesity than TMI among boys (TMI: 7.7%; 95% CI, 6.2–9.2%; P = 0.016) but not girls (P = 0.444). WHtR had a slightly lower misclassification error rate of overweight or obesity than RFMp among boys (WHtR: 6.1%; 95% CI, 4.8–7.3%; P = 0.041) but not girls (WHtR: 8.3%; 95% CI, 6.9–9.7%; P = 0.785) (Fig. 3).

Figure 3
figure3

Comparison of total misclassification error rate of body adiposity between indices in children and adolescents 8 to 14 years of age. BMI, body mass index; RFMp, Relative Fat Mass pediatric; TMI, tri-ponderal mass index; WHtR, waist-to-height ratio. Bars show comparison of total misclassification of overweight only, obesity, and overweight or obesity diagnosed by DXA-whole-body fat percentage (≥85th percentile and <95th percentile for overweight and ≥95th percentile for obesity). Error bars are 95% confidence intervals. *P < 0.01; #P < 0.05; compared with BMI.

RFMp had lower total misclassification error of overweight or obesity than BMI-for-age percentiles among African-American girls (RFMp: 8.2%; 95% CI, 6.1–10.3%; BMI-for-age: 11.3%; 95% CI, 8.8–13.8%; P < 0.001) and African-American boys (RFMp: 6.5%; 95% CI, 4.7–8.2%; BMI-for-age: 8.7%; 95% CI, 6.6–10.7%; P = 0.009) (Supplementary Fig. 10). Among African-Americans girls and boys together, RFMp had lower total misclassification error than BMI-for-age percentiles (RFMp: 7.2%; 95% CI, 5.8–8.6%; BMI-for-age: 9.8%; 95% CI, 8.2–11.3%; P < 0.001).

When using thresholds based on the highest Youden’s index, RFMp (threshold for girls and boys: 37.6% and 32.8%, respectively) and BMI-for-age had similar rates of misclassification of overweight or obesity among girls (RFMp: 11.9%; 95% CI, 10.1–13.7%; BMI-for-age: 11.6%; 95% CI, 9.5–13.7%; P = 0.796) and boys (RFMp: 8.2%; 95% CI, 6.9–9.6%; BMI-for-age: 9.3%; 95% CI, 7.9–10.7%; P = 0.109). RFMp had lower misclassification error of overweight or obesity than BMI-for age among African-American girls (RFMp: 12.5%; 95% CI, 10.3–14.7%; BMI-for-age: 17.6%; 95% CI, 14.7–20.4%; P < 0.001) and African-American boys (RFMp: 6.7%; 95% CI, 4.9–8.5%; BMI-for-age: 12.6%; 95% CI, 10.0–15.3%; P < 0.001). Overall, RFMp showed lower false negative rates and lower false positive rates for overweight and obesity compared with BMI among girls and boys (Supplementary Table 6).

Overweight and obesity misclassification among adolescents 15 to 19 years of age

Analyses were performed with data from 4,865 adolescents 180 to 239 months old (2,080 girls). For TMI, overweight was defined as ≥18.2 kg/m3 and <22.3 kg/m3 for girls, and ≥16.9 kg/m3 and <20.1 kg/m3 for boys. Obesity was defined as a TMI ≥ 22.3 kg/m3 for girls and ≥20.1 kg/m3 for boys. For RFM, overweight was defined as ≥42.3% and <46.2% for girls, and ≥28.6% and <32.9% for boys. Obesity was defined as an RFM ≥ 46.2% for girls and ≥32.9% for boys. The aforementioned RFM thresholds are not intended to suggest cutoffs for the diagnosis of overweight and obesity in this pediatric population.

RFM showed similar false negative rate for overweight or obesity compared with BMI-for-age percentiles among girls (RFM: 26.8%; 95% CI, 20.9–32.8%; BMI-for-age: 21.8%; 95% CI, 15.5–28.1%; P = 0.165) and boys (RFM: 22.9%; 95% CI, 16.9–28.9%; BMI-for-age: 24.3%; 95% CI, 18.1–30.4%; P = 0.564) (Table 2). Likewise, RFM showed similar false positive rate among girls (RFM: 4.7%; 95% CI, 3.5–5.9%; BMI-for-age: 4.0%; 95% CI, 2.8–5.1%; P = 0.253), and boys (RFM: 4.2%; 95% CI, 3.0–5.3%; BMI-for-age: 4.9%; 95% CI, 3.6–6.1%; P = 0.145) (Table 2).

RFM had lower total misclassification error rate of overweight than BMI-for-age percentiles among boys (P = 0.006) but not girls (P = 0.271). Likewise, RFM had lower total misclassification error rate of obesity than BMI-for-age percentiles among boys (P = 0.012) but not girls (P = 0.456) (Fig. 4). RFM and BMI-for-age percentiles had similar total misclassification error of overweight or obesity among girls (RFM: 8.0%; 95% CI, 6.4–9.7%; BMI-for-age: 6.6%; 95% CI, 5.2–8.0%; P = 0.076) and boys (RFM: 6.9%; 95% CI, 5.6–8.3%; BMI-for-age: 7.8%; 95% CI, 6.2–9.3%; P = 0.11). RFM had lower total misclassification error rate of overweight than TMI among boys (P = 0.006) but not girls (P = 0.526). Likewise, RFM had lower total misclassification error rate of obesity than TMI among boys (P = 0.005) but not girls (P = 0.557). RFM also had lower misclassification error of overweight or obesity than TMI among boys (TMI: 8.4%; 95% CI, 6.9–9.9%; P = 0.015) but not girls (TMI: 8.2%; 95% CI, 6.6–9.7%; P = 0.789). RFM and WHtR had similar rates of misclassification of overweight or obesity among girls (WHtR: 8.2%; 95% CI, 6.5–9.9%; P = 0.30) and boys (WHtR: 7.0%; 95% CI, 5.5–8.5%; P = 0.781) (Fig. 4).

Figure 4
figure4

Comparison of total misclassification error rate of body adiposity between indices in adolescents 15 to 19 years of age. BMI, body mass index; RFM, Relative Fat Mass; TMI, tri-ponderal mass index; WHtR, waist-to-height ratio. Bars show comparison of total misclassification of overweight only, obesity, and overweight or obesity diagnosed by DXA-whole-body fat percentage (≥85th percentile and <95th percentile for overweight and ≥95th percentile for obesity). Error bars are 95% confidence intervals. *P < 0.05; #P < 0.01; compared with BMI.

RFM had lower total misclassification error of overweight or obesity than BMI-for-age percentiles among African-American boys (RFM: 6.3%; 95% CI, 4.5–8.2%; BMI-for-age: 8.0%; 95% CI, 6.3–9.8%; P = 0.024) but not girls (RFM: 9.2%; 95% CI, 7.1–11.3%; BMI-for-age: 10.4%; 95% CI, 7.6–13.1%; P = 0.378) (Supplementary Fig. 11). Among African-Americans girls and boys together, RFM had lower total misclassification error than BMI-for-age percentiles (RFM: 7.6%; 95% CI, 6.1–9.0%; BMI-for-age: 9.0%; 95% CI, 7.5–10.6%; P < 0.001).

When using thresholds based on the highest Youden’s index, RFM (threshold for girls and boys: 38.9% and 26.2%, respectively) had lower total misclassification error of overweight or obesity than BMI-for-age among boys (RFM: 10.1%; 95% CI, 8.5–11.8%; BMI-for-age: 12.7%; 95% CI, 11.1–14.2%; P < 0.001) and higher misclassification error among girls (RFM: 13.2%; 95% CI, 11.2–15.2%; BMI-for-age: 9.6%; 95% CI, 8.1–11.0%; P < 0.001). RFM had lower misclassification error of overweight or obesity than BMI-for age among African-American boys (RFM: 7.8%; 95% CI, 6.1–9.4%; BMI-for-age: 12.6%; 95% CI, 10.2–15.0%; P < 0.001) but not girls (RFM: 16.4%; 95% CI, 13.4–19.4%; BMI-for-age: 17.5%; 95% CI, 14.5–20.5%; P = 0.181). Compared with BMI, RFM showed lower false negative rates and lower false positive rates for overweight among boys and lower false positive rates for obesity among boys (Supplementary Table 6).

Diagnostic accuracy of overweight or obesity

Among children and adolescents 8 to 14 years of age, compared with BMI, RFMp had a better diagnostic accuracy of overweight or obesity among girls (C-statistic: 0.95; 95% CI, 0.94−0.96 vs. 0.93; 95% CI, 0.91−0.95; P = 0.001) and boys (C-statistic: 0.97; 95% CI, 0.96−0.98 vs. 0.95; 95% CI, 0.94−0.96; P < 0.001). Among adolescents 15 to 19 years of age, compared with BMI, RFM had a better diagnostic accuracy of overweight or obesity among boys (C-statistic: 0.97; 95% CI, 0.96−0.98 vs. 0.95; 95% CI, 0.93−0.96; P < 0.001) but not girls (C-statistic: 0.96; 95% CI, 0.95−0.97 vs. 0.96; 95% CI, 0.95−0.97; P = 0.588).

Among children and adolescents 8 to 19 years of age, compared with BMI, height/waist ratio (the basis of both RFMp and RFM equations) showed better diagnostic accuracy of overweight or obesity among girls (C-statistic: 0.95; 95% CI, 0.94−0.96 vs. 0.92; 95% CI, 0.90−0.94; P < 0.001) and boys (C-statistic: 0.97; 95% CI, 0.96−0.97 vs. 0.93; 0.92−0.93; P < 0.001).

Association of RFMp and RFM with biomarkers for cardiometabolic disease

Among girls 8 to 14 years of age, RFMp showed the strongest association with log-insulin (r = 0.59; P < 0.001) and log-(triacylglycerol-HDLc-ratio) (r = 0.31; P < 0.001) compared with other biomarkers. Among boys 8 to 14 years of age, RFMp also showed the strongest association with log-insulin (r = 0.63; P < 0.001) and log-(triacylglycerol-HDLc-ratio) (r = 0.48; P < 0.001) (Supplementary Table 7). Among girls 15 to 19 years of age, RFM showed the strongest association with log-insulin (r = 0.48; P < 0.001) and log-HDLc (r = −0.34; P < 0.001). Among boys 15 to 19 years of age, RFM showed the strongest association with log-insulin (r = 0.59; P < 0.001) and log-(triacylglycerol-HDLc-ratio) (r = 0.45; P < 0.001) (Supplementary Table 8).



Source link

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Top