Age and sex are associated with the plasma lipidome: findings from the GOLDN study

Table 3 Associations between total subclass intensity with age, sex, and their interaction

Subclass	Age β (SE)	Sex β (SE)	Age*Sex Interaction β (SE)
TG	0.0050 (0.0027)	−0.2035 (0.0600)^b	0.0098 (0.0037)^a
DG	0.0055 (0.0027)	−0.1708 (0.0612)^b	0.0082 (0.0037)
PC	0.0148 (0.0026)^c	0.3695 (0.0574)^c	0.0106 (0.0035)^a
PE	0.0057 (0.0026)^a	0.3079 (0.0584)^c	0.0116 (0.0036)^b
PI	0.0072 (0.0027)^a	0.2908 (0.0606)^c	0.0070 (0.0037)
PG	0.0103 (0.0026)^c	0.2559 (0.0569)^c	0.0023 (0.0035)
LPE	0.0059 (0.0028)^a	−0.1244 (0.0618)	0.0072 (0.0038)
LPC	0.0006 (0.0027)	−0.5470 (0.0606)^c	0.0094 (0.0037)^a
SM	0.0152 (0.0026)^c	0.3631 (0.0587)^c	0.0082 (0.0036)
Ceramide	0.0175 (0.0026)^c	−0.1335 (0.0579)^a	0.0046 (0.0035)
LCer	0.0109 (0.0028)^c	0.0242 (0.0624)	−0.0008 (0.0038)
GlcCer	0.0108 (0.0022)^c	0.1240 (0.0499)^a	0.0061 (0.0030)
GalGalCer	0.0000 (0.0027)	−0.0223 (0.0602)	−0.0130 (0.0037)^b
Cholesterol	−0.0123 (0.0027)^c	−0.2907 (0.0595)^c	0.0012 (0.0036)
CE	0.0101 (0.0028)^c	0.2637 (0.0623)^c	0.0002 (0.0038)

Rows include regression coefficients (β) and standard errors from separate linear regression models with outcome of standardized total subclass intensity. Each model includes age, sex, age*sex interaction, batch, and BMI as covariates. In all models, age is centered at the mean age of 48.3 years, and the reference category for sex is male. All P-values are adjusted for multiple testing using a Benjamini-Hochberg adjustment to control the false discovery rate among each set of 15 coefficients at 0.05. Age coefficients (β) can be interpreted as the expected change in standardized total subclass intensity for a one-year increase in age, among men, after adjustment for batch and BMI. Sex coefficients (β) can be interpreted as the expected difference in standardized total subclass intensity between women and men, at the mean age, after adjustment for batch and BMI (positive values indicate higher expected levels in women). Age*sex interaction coefficients (β) can be interpreted as the expected additional change in standardized total subclass intensity for a one-year increase in age, among women (on top of the age coefficient for men), after adjustment for batch and BMI. Put more simply, add the age coefficient and age*sex interaction coefficient together to find the expected change in standardized total subclass intensity for a one-year increase in age, among women, after adjustment for batch and BMI
^aadjusted P-value < 0.05
^badjusted P-value < 0.01
^cadjusted P-value < 0.001

ISSN: 1476-511X