210

R.P.

Smith

et

al.

/

Maturitas

74 (2013) 208–

212

Based

on

the

substantial

heterogeneity

within

these

studies,

the

evidence

for

seasonal

variation

of

androgens

can

only

be

character-

ized

as

inconsistent

at

this

time.

3.

Seasonal

variations

in

testosterone:

inﬂuence

of

light

exposure,

melatonin

and

sleep–wake

cycles

Seasonal

changes

in

the

duration

of

the

light–dark

cycle

are

known

to

cause

variations

in

the

activity

of

human

neuroendocrine

systems.

Light

exposure

and

its

potential

impact

on

reproductive

hormone

variability

was

recently

evaluted

in

a

series

of

men

from

the

Arctic

Circle.

Ruhayel

et

al.

[28]

examined

how

extremes

of

temperature

and

light–dark

cycles

might

inﬂuence

these

hormone

ﬂuctuations.

The

authors

reported

on

two

cohorts

of

Norwegian

men

divided

by

those

living

north

(Tromsø)

and

south

(Oslo)

of

the

Arctic

circle

to

maximize

variations

in

photoperiod

and

tem-

perature.

These

205

men

underwent

serial

blood

measurements

over

a

12-month

period

including

during

the

study

period

daylight

maximum

and

minimum.

Serum

concentrations

of

FSH,

LH,

TT,

E2,

SHBG,

and

urinary

excretion

of

the

major

melatonin

metabolite,

6-sulfatoxymelatonin

(aMT6s)

were

assessed.

aMT6s

levels

were

evaluated

based

on

the

association

of

melatonin

with

sleep–wake

cycles

and

circadian

rhythm

[28]

.

Both

locations

demonstrated

declining

LH

levels

during

the

early

winter;

however,

seasonal

variation

of

LH

was

only

signi±-

cant

in

Tromsø.

The

seasonal

variation

of

TT

concentrations

was

signi±cant

in

Oslo,

but

not

in

Tromsø,

and

neither

location

showed

signi±cant

seasonal

variation

of

FT

levels

[28]

.

One

would

have

expected

that

seasonal

patterns

in

sex

hormones,

if

inﬂuenced

by

photoperiod

and

temperature,

would

have

been

most

apparent

in

this

population

where

climatic

inﬂuences

are

signi±cantly

diver-

gent.

However,

aside

from

LH,

patterns

of

change

in

other

sex

steroids

were

non-seasonal,

possibly

due

to

confounding

factors

such

as

lifestyle

and

variation

in

genetic

backgrounds

[28]

.

As

a

correlate

to

ﬂuctuations

in

photoperiod,

Svartberg

et

al.

[7,24]

hypothesized

that

changes

in

sleep

patterns

could

also

have

inﬂuenced

the

discrepancy

between

the

presence

of

androgen

vari-

ation

in

their

original

Norwegian

study

in

comparison

to

their

San

Diego

study.

Phase

shifts

in

hormone

secretion

have

been

shown

to

be

causally

related

to

concomitant

shifts

in

the

onset

of

sleep;

how-

ever,

in

the

study

by

Ruhayel

et

al.,

which

did

not

show

signi±cant

variation

in

T,

the

recorded

sleep

duration

did

not

differ

between

summer

and

winter

amongst

the

groups

[28]

.

The

authors

could

not

exclude

its

inﬂuence,

however.

Given

the

central

regulation

of

melatonin

and

its

effect

on

light–dark

cycles

and

sleep,

Svartberg,

Ruhayel

and

oth-

ers

have

postulated

that

melatonin

may

likewise

inﬂuence

the

hypothalamic–pituitary–testicular

(HPT)

axis

[7,23,28,29]

.

Mela-

tonin

secretion

and

subsequent

urinary

excretion

of

aMT6s

has

been

shown

to

correspond

to

changes

in

the

light–dark

cycle

and

sleep

patterns

[28,30]

.

In

the

prior

study

by

Ruhayel

et

al.

[28]

,

sea-

sonal

changes

in

T

concentration

exhibited

signi±cant

correlation

with

the

changes

in

excretion

of

aMT6s,

a

melatonin

by-product;

however,

this

correlation

was

only

seen

in

the

Oslo

population.

Conversely,

seasonal

changes

in

LH

concentration

demonstrated

signi±cant

correlation

with

the

changes

in

aMT6s

in

Tromsø

alone.

Urinary

aMT6s

concentrations

were

lowest

during

early

summer

in

both

locations

indicating

possible

suppression

due

to

longer

day-

light

exposure.

The

correlations

between

changes

in

T

and

LH

and

excretion

of

aMT6s,

although

signi±cant,

differed

in

the

two

loca-

tions

indicating

this

relationship

was

likely

circumstantial

[28]

.

Martikainen

and

Huhtaniemi

[23,30]

similarly

examined

the

associations

between

melatonin

and

gonadotropins

in

a

popula-

tion

of

24

men.

The

participants

were

followed

over

a

period

of

13

months

in

northern

Finland,

where

day

length

is

22

h

in

the

summer

and

3.5

h

in

the

winter.

Serum

melatonin

levels

showed

two

annual

peaks

in

December

and

May

with

a

nadir

in

August.

The

melatonin

peak

in

May

was

associated

with

signi±cant

increases

in

LH.

The

authors

found

no

signi±cant

differences

in

FSH,

LH,

or

T

levels

between

individual

months

leading

them

to

conclude

that

extreme

seasonal

changes

in

daylight

do

not

play

an

obvious

role

in

the

regulation

of

testicular

androgens

[23,31]

.

The

lack

of

consistent

variation

in

the

levels

of

T

and

LH

in

these

studies

weakens

the

argu-

ment

for

a

substantial

relationship

between

melatonin

secretion,

sleep,

and

variations

in

sex

hormones.

4.

Seasonal

variations

in

testosterone:

the

role

of

the

hypothalamic–pituitary–testicular

axis

and

vitamin

D

Several

investigators

have

sought

to

determine

the

interre-

lationship

between

hormones

of

the

HPT

axis,

vitamin

D,

and

seasonality.

A

recent,

randomized,

placebo-controlled

trial

by

Pilz

et

al.

suggested

that

vitamin

D

may

increase

the

production

of

T

in

men

[32]

.

Additional

studies

have

previously

given

scienti±c

support

to

the

relationship

between

vitamin

D

and

reproductive

hormones

[33,34]

.

In

the

European

Male

Aging

Study,

a

cross-

sectional

survey

of

3369

men

was

performed

with

T,

E2,

DHT,

LH,

FSH,

SHBG,

and

vitamin

D

levels

assayed

[35]

.

FT

levels

were

lower

and

LH

levels

were

higher

in

men

with

vitamin

D

de±ciency.

Vita-

min

D

was

positively

associated

with

TT

and

FT

and

negatively

with

E2

and

LH.

After

adjustments

were

made

for

potential

health

and

lifestyle

confounders,

no

statistically

signi±cant

associations

were

found

between

vitamin

D

and

individual

sex

hormones.

Notably,

seasonal

variation

was

seen

only

for

Vitamin

D

[35]

.

Despite

a

lack

of

a

correlation

in

seasonal

patterns,

reinforcement

of

a

relationship

between

vitamin

D

and

androgens

has

important

clinical

implica-

tions

as

both

vitamin

D

de±ciency

and

hypogonadism

have

been

associated

with

morbidity

and

mortality

[36,37]

.

A

cross-sectional

study

by

Wehr

et

al.

examined

vitamin

D,

T,

and

SHBG

levels

in

2299

men

referred

for

coronary

angiography

[38]

.

Men

with

suf±cient

vitamin

D

levels

had

statistically

signif-

icant

higher

levels

of

T

compared

with

vitamin

D

de±cient

men.

When

adjusting

for

confounders,

a

signi±cant

association

remained

between

vitamin

D

and

T.

In

contrast

to

the

study

by

Lee

[35]

,

T

showed

a

seasonal

pattern

with

a

nadir

in

March

and

a

peak

in

August.

Wehr

et

al.

concluded

that

androgen

levels

and

vitamin

D

levels

are

associated

in

men

and

reveal

concordant

seasonal

vari-

ation.

The

population

examined,

however,

was

restricted

to

older

men

with

atherosclerosis

and,

therefore,

the

data

could

not

be

gen-

eralized

[38]

.

A

population-based

study

was

therefore

needed

to

con±rm

the

seasonal

patterns

in

vitamin

D

and

T

seen

in

the

Wehr

study.

Nimptsch

and

colleagues

investigated

a

population

of

1362

men,

aged

40–75,

from

the

Health

Professionals

Follow-up

Study

[39]

.

This

included

assays

of

Vitamin

D,

T,

and

FT.

Vitamin

D

was

posi-

tively

associated

with

TT

and

FT

levels;

however,

unlike

for

vitamin

D,

the

authors

did

not

observe

any

seasonal

variation

of

T

con-

centrations

[39]

.

While

the

association

of

vitamin

D

with

T

levels

appears

founded,

these

studies

lend

further

uncertainty

as

to

the

evidence

for

seasonal

patterns

of

T.

5.

Seasonal

variations

in

testosterone:

lifestyle

inﬂuences?

The

potential

exists

that

seasonal

changes

in

androgens

may

be

due

to

other

factors

that

vary

during

the

year.

For

instance,

physical

activity

increases

signi±cantly

during

the

summer

months

[40,41]

,

and

exercise

has

been

shown

to

be

a

potent

stimulus

for

the

pro-

duction

of

T

[42]

.

A

study

examining

seasonal

variation

in

waist

circumference

and

BMI

found

both

to

be

greater

in

winter

than

in

summer

months

[43]

.

One

may

postulate

that

seasonal

changes

in