no signiﬁcant differences were seen in testosterone,

estradiol, LH or FSH levels. AGD (r

0.20, p

0.03)

and PL (r

0.20, p

0.03) were signiﬁcantly asso-

ciated with serum testosterone levels (

ﬁg. 2

). Re-

moval of the largest AGD measures (3 or more SD

above mean) did not signiﬁcantly change the conclu-

sions. In contrast, total testis size did not show a

statistically signiﬁcant association with testoster-

one levels (r

0.17, p

0.07). While AGD and PL

showed no signiﬁcant association with LH, FSH or

estradiol, total testis volume was signiﬁcantly asso-

ciated with FSH (r

0.50, p

0.01) and LH (p

0.43,

0.01). All genital measurements seemed to be

correlated with each other, as AGD with penile

length (r

0.20, p

0.01), AGD with total testicular

volume (r

0.31, p

0.01) and penile length with

total testicular volume (r

0.24, p

0.01).

In the unadjusted and adjusted models AGD sig-

niﬁcantly correlated with serum testosterone levels.

In fact, for each 1 cm increase in AGD the serum

testosterone increased by 20.1 ng/dl (95% CI 1.8,

38.4; p

0.03). In contrast, no statistically signiﬁcant

correlation was seen with penile length, although a

trend was seen (

table 2

). A positive relationship was

also identiﬁed between the testosterone-to-LH ratio

and AGD as well as testis size but not penile length.

Moreover when stratifying men by several hypogo-

nadal testosterone levels between 200 and 300 ng/dl,

men with lower testosterone levels had signiﬁcantly

shorter anogenital distances than those with higher

levels (

table 3

Neither AGD nor penile length correlated with LH,

FSH or estradiol. On age adjusted models total testis

size was signiﬁcantly associated with FSH and LH

0.01). When stratifying by race, the direction and

magnitude of the measures of associations between

genital measurements and hormone values remained

similar. It should be noted that smaller sample sizes

limited the statistical signiﬁcance for white (p

0.1)

but not for black individuals (p

0.04).

DISCUSSION

The current study demonstrated an association be-

tween anogenital distance and serum testosterone

levels in a cohort of United States adult men evalu-

ated in an andrology practice. In addition, there was a

trend toward a similar association between testoster-

one levels and stretched penile length. We previously

conﬁrmed a relationship in this cohort between AGD

and semen quality, but to our knowledge the current

study represents the ﬁrst assessment of the relation-

ship between anogenital distance and serum testoster-

one levels.

In rodents in utero exposure to agents known to

disrupt androgen mediated pathways corrupts nor-

mal male genital development with a decrease in

genital lengths (ie phallus length, AGD) and im-

paired testosterone and sperm production.

18,21

humans Swan et al demonstrated that mothers ex-

posed to higher levels of endocrine disruptors gave

Table 1.

Demographic, anthropomorphic, hormonal

and genital characteristics of the cohort

Mean ht inches (SD)

70.4

(3.2)

Mean wt lb (SD)

210.3 (54.7)

Mean kg/m

body mass index (SD)

29.8

(7.5)

No. race (%):

White

(78.5)

Black

(13.8)

Asian

(7.8)

Mean hormone levels (SD):

Testosterone (ng/dl)

316.5 (131.6)

FSH (mIU/ml)

8.5

(8.4)

LH (mIU/ml)

4.7

(3.0)

Estradiol (ng/dl)

2.8

(2.9)

Testosterone/LH

88.1 (56.5)

Mean genital measurements (SD):

AGD (mm)

34.3 (13.3)

PL (mm)

108.5 (22.8)

Total testis vol (ml)

34.1

(8.2)

Figure 2.

Scatterplot displaying relationship between (

) anogenital distance and serum testosterone levels (r

0.20, p

0.03), and

(

) stretched penile length and serum testosterone levels (r

0.20, p

0.03).

ANOGENITAL DISTANCE AND REPRODUCTIVE HORMONES

596