Table 4.
Association of ADBV with serum metabolites associated with childhood BMI Z-score (FDR < 0.20) and association of these metabolites with BMI Z-score
| Metabolite | Superclass | Percent difference in ADBV associated with 10% increase in serum metabolitea | Percent difference in serum metabolites associated with a unit increase in BMI z-scoreb | ||||
|---|---|---|---|---|---|---|---|
| Δ | 95% CI | P-value | Δ | 95% CI | P-value | ||
| X-16576 | Unknown | 2.91 | 1.25, 4.60 | 6.92E−04 | − 13.65 | − 21.53, − 4.99 | 3.01E−03 |
| X-12104 | Unknown | 3.81 | 1.24, 6.45 | 4.06E−03 | 9.66 | 3.48, 16.22 | 2.15E−03 |
| lysine | Amino acid | − 7.68 | − 12.76, − 2.31 | 6.26E−03 | 3.05 | 0.50, 5.65 | 1.98E−02 |
| X-24588 | Unknown | − 2.84 | − 5.05, − 0.57 | 1.53E−02 | 8.18 | 0.73, 16.19 | 3.21E−02 |
| 7-Methylguanine | Nucleotide | − 5.13 | − 9.63, − 0.39 | 3.57E−02 | 4.77 | 1.45, 8.20 | 5.16E−03 |
| Valine | Amino acid | − 5.28 | − 10.01, − 0.31 | 3.90E−02 | 4.13 | 1.65, 6.66 | 1.20E−03 |
| Homoarginine | Amino acid | − 2.71 | − 5.40, 0.06 | 5.64E−02 | 7.22 | 1.63, 13.12 | 1.16E−02 |
| X-17340 | Unknown | − 1.70 | − 3.51, 0.14 | 7.13E−02 | 28.52 | 17.89, 40.11 | 5.17E−08 |
| 5alpha-androstan-3alpha,17beta-diol monosulfate (1) | Lipid | 0.99 | − 0.08, 2.06 | 7.14E−02 | 37.90 | 17.09, 62.41 | 1.66E−04 |
| N-acetylvaline | Amino acid | − 4.55 | − 9.25, 0.39 | 7.21E−02 | 3.68 | 0.63, 6.83 | 1.87E−02 |
| Glycochenodeoxycholate 3-sulfate | Lipid | − 1.01 | − 2.10, 0.10 | 7.63E−02 | − 14.80 | − 24.58, − 3.75 | 1.09E−02 |
| Kynurenine | Amino acid | − 2.72 | − 5.76, 0.42 | 9.10E−02 | 6.20 | 1.74, 10.85 | 6.65E−03 |
aResults shown for metabolites associated with ADBV at P < 0.10. Estimates are from robust mixed effects multivariable linear regression models including ADBV as the dependent variable; childhood fixed effects – ln(metabolite level) (continuous), BMI z-score (continuous), age at BMI measurement (continuous), treatment group assignment, race (white/nonwhite) and menstrual cycle phase at blood collection (premenarche/luteal/follicular/unknown); adult fixed effects—BMI and BMI2 at breast density (continuous), college graduate (yes/no), duration hormone use (continuous), number live births (0/1 +), current smoker (yes/no); and DISC clinic as a random effect
bEstimates are from robust mixed effects multivariable linear regression models including ln(metabolite level) (continuous) as the dependent variable; childhood fixed effects—BMI z-score (continuous), age at BMI measurement (continuous), treatment group assignment, race (white/nonwhite) and menstrual cycle phase at blood collection (premenarche/luteal/follicular/unknown); and DISC clinic as a random effect