Skip to main content
. 2021 Jul 23;23:75. doi: 10.1186/s13058-021-01454-3

Table 2.

Adjusted means of PMD across quartiles of PCaeC30:2 and SMC16:1, overall and by BMI

Quartiles of metabolites plasma concentrations
Q1 Q2 Q3 Q4
Metabolite Adjusted mean PMD1 (95 % CI) Adjusted mean PMD1 (95% CI) Adjusted mean PMD1 (95% CI) Adjusted mean PMD1 (95% CI) P trend2
SM C16:1 Overall 38.1 (32.2–44.0) 39.8 (34.1–45.5) 36.6 (30.9–42.2) 32.0 (26.3–37.8) < 0.01
BMI (kg/m2)
n = 286 < 27.4 43.7 (35.7–51.8) 41.7 (33.7–49.7) 39.6 (31.8–47.3) 35.1 (27.1–43.1) < 0.01
n = 287 ≥ 27.4 29.6 (19.1–40.1) 35.4 (25.3–45.6) 29.7 (19.8–39.5) 25.9 (15.7–36.1) 0.07
P interaction3 0.51
PC ae C30:2 Overall 40.7 (34.9–46.5) 37.6 (31.8–43.5) 34.5 (28.8–40.1) 35.0 (29.2–40.8) < 0.01
BMI (kg/m2)
n = 286 < 27.4 42.4 (34.4–50.4) 42.2 (34.1–50.2) 36.8 (29.0–44.6) 38.5 (30.4–46.7) 0.09
n = 287 ≥ 27.4 36.8 (26.2–47.3) 31.6 (21.4–41.9) 29.8 (20.0–39.6) 28.6 (18.3–38.8) < 0.01
P interaction3 0.31

Abbreviations: ae acyl-alkyl; BMI body mass index; CI confidence interval; PC phosphatidylcholine; PMD percent mammographic density; Q quartile; SM sphingomyelin

1Means were adjusted for age, BMI, age at menarche, family history of cancer, history of benign breast disease, use of oral contraceptive, number of full-term pregnancies, age at first full-term pregnancy, breastfeeding, alcohol intake, smoking status, socioeconomic status, and physical activity

2P value for test of linear trend across quartiles of metabolite, performed by assigning participants the median value in each quartile and modeling the corresponding variable as a continuous term

3P value for interaction term between residuals of log-transformed metabolites concentrations regressed on batch and dichotomized BMI (< 27.44/≥ 27.44)