The Heritability of Sex Steroid Concentrations in Men
The Heritability of Sex Steroid Concentrations in Men
A total of 3367 men contributed to the analysis, representing 543 pedigrees; these were comprised of 1066 father/son pairs, 1284 brother pairs and additional extended family relationships. Participants had mean (standard deviation) age of 49 (14) years and BMI 28 (5) kg/m; 16% were current smokers and 7·5% had T2DM. The distributions of hormones and SHBG are described in Table 1. Aside from being older, subjects from the Offspring cohort exhibited greater rates of T2DM (14% vs 3%) and included a slightly lower percentage of current smokers (13% vs 16%) than their Generation 3 counterparts, while BMI, at 29 (5) vs 28 (5) kg/m, was similar in the two cohorts.
Estimates of age-adjusted heritability are presented in Table 1. In age-adjusted analyses, TT, E1 and SHBG exhibited substantial heritability – estimated h (standard error; SE) was approximately 0·40 (0·5) in each case, whereas E2 and cFT had estimated h (SE) of 0·30 (0·05) and 0·19 (0·04), respectively. Additional adjustment for BMI, smoking and T2DM produced no substantial reduction in estimates of heritability (Fig. 1). For TT, however, further adjustment for SHBG produced a substantial reduction in estimated heritability, h = 0·18 (0·04), such that the SHBG-adjusted TT heritability estimated was similar to the heritability estimate for cFT, 0·19 (0·04). For E1 and E2, by contrast, further adjustment for SHBG produced no meaningful difference in results. For E2, adjustment for age, BMI, smoking and T2DM yielded estimated h = 0·28 (0·05); the addition of SHBG yielded estimated h = 0·26 (0·05).
(Enlarge Image)
Figure 1.
Change in estimates of heritability (h) with successive adjustment for covariates. Moving left to right, each independent analysis adds new covariates as indicated along the horizontal axis. When total testosterone is adjusted for sex hormone binding globulin (along with all other covariates), the heritability estimate is substantially reduced, so that it is similar to that exhibited by free testosterone.
Paired analyses of hormones and SHBG adjusted for covariates demonstrated significant genetic correlation between TT and cFT (ρG = 0·68), between TT and SHBG (ρG = 0·87), between E2 and E1 (ρG = 0·48) and between TT and E2 (ρG = 0·38). There was somewhat lesser genetic correlation between cFT and SHBG (ρG = 0·24) and between E2 and SHBG (ρG = 0·26).
The estimated environmental correlation (ρE) between each pair of measurements was generally comparable with its genetic counterpart, with the exception of the TT and SHBG pairing, where ρE (0·28) was substantially lower than ρG.
Results
A total of 3367 men contributed to the analysis, representing 543 pedigrees; these were comprised of 1066 father/son pairs, 1284 brother pairs and additional extended family relationships. Participants had mean (standard deviation) age of 49 (14) years and BMI 28 (5) kg/m; 16% were current smokers and 7·5% had T2DM. The distributions of hormones and SHBG are described in Table 1. Aside from being older, subjects from the Offspring cohort exhibited greater rates of T2DM (14% vs 3%) and included a slightly lower percentage of current smokers (13% vs 16%) than their Generation 3 counterparts, while BMI, at 29 (5) vs 28 (5) kg/m, was similar in the two cohorts.
Estimates of age-adjusted heritability are presented in Table 1. In age-adjusted analyses, TT, E1 and SHBG exhibited substantial heritability – estimated h (standard error; SE) was approximately 0·40 (0·5) in each case, whereas E2 and cFT had estimated h (SE) of 0·30 (0·05) and 0·19 (0·04), respectively. Additional adjustment for BMI, smoking and T2DM produced no substantial reduction in estimates of heritability (Fig. 1). For TT, however, further adjustment for SHBG produced a substantial reduction in estimated heritability, h = 0·18 (0·04), such that the SHBG-adjusted TT heritability estimated was similar to the heritability estimate for cFT, 0·19 (0·04). For E1 and E2, by contrast, further adjustment for SHBG produced no meaningful difference in results. For E2, adjustment for age, BMI, smoking and T2DM yielded estimated h = 0·28 (0·05); the addition of SHBG yielded estimated h = 0·26 (0·05).
(Enlarge Image)
Figure 1.
Change in estimates of heritability (h) with successive adjustment for covariates. Moving left to right, each independent analysis adds new covariates as indicated along the horizontal axis. When total testosterone is adjusted for sex hormone binding globulin (along with all other covariates), the heritability estimate is substantially reduced, so that it is similar to that exhibited by free testosterone.
Paired analyses of hormones and SHBG adjusted for covariates demonstrated significant genetic correlation between TT and cFT (ρG = 0·68), between TT and SHBG (ρG = 0·87), between E2 and E1 (ρG = 0·48) and between TT and E2 (ρG = 0·38). There was somewhat lesser genetic correlation between cFT and SHBG (ρG = 0·24) and between E2 and SHBG (ρG = 0·26).
The estimated environmental correlation (ρE) between each pair of measurements was generally comparable with its genetic counterpart, with the exception of the TT and SHBG pairing, where ρE (0·28) was substantially lower than ρG.