Table 2.

Associations between cognitive function and individual body composition and hormonal measures (N=53)

	Executive Function		Verbal Learning and Memory		Processing Speed		Visual Memory		Visuospatial Skills#
	partial correlation coefficient	p-value	partial correlation coefficient	p-value	partial correlation coefficient	p-value	partial correlation coefficient	p-value	partial correlation coefficient	p-value
BODY COMPOSITION
BMI (log)	−0.09	0.56	−0.05	0.73	0.03	0.86	−0.25	0.08	−0.27	0.055
Total fat % (DXA)	−0.17	0.24	−0.00	0.98	−0.05	0.74	−0.36	0.012	−0.48	0.0005*
Total Lean mass (DXA)	−0.05	0.72	−0.05	0.65	0.00	0.96	0.03	0.82	−0.07	0.61
HORMONES
Glucose homeostasis
HOMA-IR (log)	−0.12	0.40	−0.07	0.63	−0.09	0.54	−0.26	0.06	−0.27	0.05
Gonadal steroids
Total testosterone	−0.29	0.043*	−0.29	0.042*	−0.10	0.51	−0.06	0.69	0.17	0.24
Free testosterone	−0.05	0.71	−0.12	0.39	−0.01	0.92	0.06	0.68	0.22	0.14
SHBG (log)	−0.34	0.015*	−0.24	0.09	−0.10	0.52	−0.15	0.29	0.03	0.86

Model: Regression models adjusting for age and education. Primary outcomes are underlined. Partial correlation coefficients and p-values are reported. Standard scores were analyzed for the first 4 domains, and raw scores for visuospatial skills domain.

Bolded values represent associations for which the p-value was ≤ 0.05.

^#

For the visuospatial skills domain, when a logistic regression model was used, with outcome a score above or below the 5^th percentile for test norms, the following results were obtained:

• Estimate for 1% increase in total fat: −0.22, p=0.002
• Estimate for 1 log increase in HOMA-IR: −1.10, p=0.05.

^*Association remains significant (p<0.05) when model further adjusted for BMI.

Abbreviations: AUC, area under the curve; BMI, body mass index; DXA, dual-energy X-ray absorptiometry; HOMA-IR, homeostasis model assessment – insulin resistance; SHBG, sex hormone binding globulin.