Table 2.
Association of childhood BMI Z-score with young adult breast density phenotypes
| Phenotype | Minimally adjusted modela | Fully adjusted modelb | ||||
|---|---|---|---|---|---|---|
| Δ/unit BMI | 95% CI | P-value | Δ/unit BMI | 95% CI | P-value | |
| Breast density (%) | − 3.60 | − 6.30, − 0.90 | 9.70E−03 | − 3.43 | − 6.04, − 0.82 | 1.09E−02 |
| Dense breast Volume (cm3) | − 24.36% | − 34.66%, − 12.43% | 2.52E−04 | − 24.44% | − 34.27%, − 13.14% | 1.18E−04 |
| Non-dense breast volume (cm3) | − 8.59% | − 17.02%, 0.69% | 7.05E−02 | − 8.57% | − 17.04%, 0.78% | 7.30E−02 |
aEstimates are from robust mixed effects multivariable linear regression models including breast density phenotype as the dependent variable; childhood BMI z-score (continuous), age at childhood BMI measurement (continuous), treatment group assignment, BMI and BMI2 at breast density measurement in young adulthood (continuous) as fixed effects; and DISC clinic as a random effect
bEstimates are from robust mixed effects multivariable linear regression models as described above under (1) plus fixed effects for race (white/nonwhite), college graduate (yes/no), duration hormone use (continuous), number live births (0/1+) and current smoker (yes/no)