Association of hyperandrogenemia and hyperestrogenemia with type 2 diabetes in Hispanic postmenopausal women.

Abstract

Accumulating evidence suggests that hyperandrogenemia may be a risk factor for coronary heart disease (CHD) in women. The present study was carried out to test the hypothesis that hyperandrogenemia is associated with type 2 diabetes in women and thus may contribute to the increased risk of CHD in women with type 2 diabetes.
Sex hormones, sex hormone-binding globulin (SHBG), and risk factors for CHD were measured in 20 postmenopausal women with type 2 diabetes and in 29 control subjects. All of the diabetic and control subjects were Hispanic women aged >55 years who were not taking hormone replacement therapy lipid-lowering drugs, or insulin and who were otherwise randomly chosen from a cohort of stroke-free subjects from the Northern Manhattan Stroke Study
Mean age, BMI, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, blood pressure, and smoking were not significantly different between cases and control subjects, but waist-to-hip ratio (WHR) was significantly higher in the diabetic subjects (P = 0.01). The mean levels of free testosterone (FT) (P = 0.01), dehydroepiandrosterone sulfate (P<0.04), and estradiol (P = 0.01) (controlled for WHR) were significantly higher in the diabetic subjects; with the statistical outliers removed, the testosterone (P = 0.05) and androstenedione (P = 0.002) levels (controlled for WHR) were also significantly higher in the diabetic subjects. The mean levels of estrone, cortisol, and SHBG were not significantly different. The results were similar in the 10 diabetic subjects treated with diet only Significant positive correlations (controlled for age and BMI) were observed between FT or testosterone and cholesterol, LDL cholesterol, and blood pressure.
Postmenopausal Hispanic women with type 2 diabetes had both hyperandrogenemia and hyperestrogenemia, and testosterone or FT correlated positively with risk factors for CHD. Hyperandrogenemia may be a link between diabetes and CHD in women.

Full text

Page 1

74DIABETESCARE, VOLUME 23, NUMBER 1, JANUARY 2000
C
women. An increase in waist-to-hip ratio
(WHR), which appears to be an andro g e n i c
p a t t e rn of obesity in women (1–6), has
been re p o rted to be associated in women
with CHD (1,7,8) and coro n a ry art e ry dis-
ease (CAD) (9–12). Hirsutism (10) and
polycystic ovary syndrome (PCOS) (13),
which are androgenic states, have also been
found to be associated with CAD. A posi-
tive relationship between free testostero n e
(FT) level and the degree of CAD in women
has recently been re p o rted (14). Support
for an androgen–CAD relationship and evi-
dence that it may be a cause-and-effect re l a-
t i o nship has been re p o rted in female
m o nkeys in which testosterone administra-
tion was found to exacerbate diet-induced
a t h e ro s c l e rosis independently of any eff e c t s
on risk factors for CHD (15).
A possible mechanism for an andro-
gen–CHD relationship is that andro g e n i c-
ity in women may underlie the expre s s i o n
of risk factors for CHD; this possibility is
suggested by the positive relationship of
a n d rogens to risk factors for CHD in
women (14,16). An increase in the levels
of testosterone or FT with the risk factors
h y p e rtension (17,18), smoking (19), and
i n c reased WHR (3–6) and a positive cor-
relation of testosterone or FT with blood
p re s s u re level (3,14,17,18), WHR (3,5),
BMI (6,14), and insulin level (3,5,14,20,
21) have been re p o rted in women. Fur-
t h e rm o re, the androgenic states PCOS
(22–24) and increased WHR (1,3,7–9,12,
25,26) have both been re p o rted to be as-
sociated with risk factors for CHD, and
a n d rogen administration to women has
been re p o rted to selectively increase vis-
ceral adipose (27), which in turn may
cause the expression of risk factors for
CHD (28) and/or serve as a marker for
them (14). Because type 2 diabetes is one
of the strongest risk factors for CHD in
women (29), the present study was car-
ried out to determine whether hyperan-
d rogenemia is also associated with type 2
onsiderable evidence suggests a re l a-
tionship between androgenicity and
c o ro n a ry heart disease (CHD) in
F rom the Departments of Medicine (G.B.P ., C.H.T., T. - Y.J.) and Neurology (B.B.-A., I.-F.L., R.L.S.), Colum-
bia University College of Physicians and Surgeons, St. Luke’s–Roosevelt Hospital Center (G.B.P ., T. - Y.J.);
and the Columbia-Presbyterian Medical Center (C.H.T., B.B.-A., I.-F.L., R.L.S.), New York.
A d d ress correspondence and reprint requests to Dr. Gerald B. Phillips, St. Luke’s–Roosevelt Hospital,
1000 10th Ave., New York, NY 10019.
Received for publication 19 May 1999 and accepted in revised form 14 September 1999.
A b b re v i a t i o n s : CAD, coro n a ry art e ry disease; CHD, coro n a ry heart disease; DHEA, dehydro e p i a n d ro s-
t e rone; DHEAS, dehydro e p i a n d ro s t e rone sulfate; FT, free testosterone; PCOS, polycystic ovary syndro m e ;
RIA, radioimmunoassay; SHBG, sex hormone–binding globulin; WHR, waist-to-hip ratio.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion
factors for many substances.
Association of Hyperandrogenemia and
H y p e re s t rogenemia W ith Type 2
Diabetes in Hispanic Postmenopausal
Wo m e n
ORI GI NAL ART I C L E
O B J E C T I V E — Accumulating evidence suggests that hyperandrogenemia may be a risk
factor for coro n a ry heart disease (CHD) in women. The present study was carried out to test
the hypothesis that hyperandrogenemia is associated with type 2 diabetes in women and thus
may contribute to the increased risk of CHD in women with type 2 diabetes.
RESEA RCH DESIG N AND M ETHODS — Sex hormones, sex horm o n e – b i n d i n g
globulin (SHBG), and risk factors for CHD were measured in 20 postmenopausal women
with type 2 diabetes and in 29 control subjects. All of the diabetic and control subjects were
Hispanic women aged 55 years who were not taking hormone replacement therapy, lipid-
lowering drugs, or insulin and who were otherwise randomly chosen from a cohort of stro k e -
f ree subjects from the Nort h e rn Manhattan Stroke Study.
R E S U LT S — Mean age, BMI, total cholesterol, LDL cholesterol, HDL cholesterol, triglyc-
erides, blood pre s s u re, and smoking were not significantly diff e rent between cases and con-
t rol subjects, but waist-to-hip ratio (WHR) was significantly higher in the diabetic subjects
(P=0.01). The mean levels of free testosterone (FT) (P=0.01), dehydro e p i a n d ro s t e rone sul-
fate (P 0.04), and estradiol (P=0.01) (controlled for WHR) were significantly higher in the
diabetic subjects; with the statistical outliers removed, the testosterone (P = 0.05) and an-
d rostenedione (P = 0.002) levels (controlled for WHR) were also significantly higher in the
diabetic subjects. The mean levels of estrone, cortisol, and SHBG were not significantly diff e r-
ent. The results were similar in the 10 diabetic subjects treated with diet only. Significant pos-
itive correlations (controlled for age and BMI) were observed between FT or testosterone and
c h o l e s t e rol, LDL cholesterol, and blood pre s s u re .
C O N C L U S I O N S — Postmenopausal Hispanic women with type 2 diabetes had both hy-
p e r a n d rogenemia and hypere s t rogenemia, and testosterone or FT correlated positively with
risk factors for CHD. Hyperandrogenemia may be a link between diabetes and CHD in women.
Diabetes Care2 3 :7 4–79, 2000
GERALD B. PHILLIPS, MD
CATHERINE H. TUCK, MD
TIAN-YI JING, BS
BERNADETTE BODEN-ALBALA, MPH
I-FENG LIN, MS
NABILA DAHODWALA, BS
RALPH L. SACCO, MD
P a t h o p h y s i o l o g y / C o m p l i c a t i o n s

Page 2

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 75
Phillips and Associates
diabetes in women. To this end, the
s e rum levels of sex hormones and sex
h o rmone–binding globulin (SHBG) were
c o m p a red in 20 postmenopausal His-
panic women with type 2 diabetes and 29
healthy control subjects.
RESEA RCH DESIGN A ND
M ETHODS — A total of 20 post-
menopausal women with type 2 diabetes
and 29 nondiabetic control subjects were
studied. The diabetic and control subjects
w e re women from the stro k e - f ree contro l
population enrolled in the Nort h e rn Man-
hattan Stroke Study, a prospective commu-
nity-based study of stroke (30), who were
randomly chosen after the following crite-
ria were fulfilled: aged 55 years (at which
age essentially all women are expected to
be postmenopausal) (31); Hispanic ethnic-
ity; not taking hormone replacement ther-
a p y, insulin, or lipid-lowering drugs; no
major medical disorder other than type 2
diabetes or other risk factors for CHD; and
a blood sample had been drawn within the
past 2 years. Diabetes was defined as a fast-
ing blood glucose level 127 mg/dl, oral
hypoglycemic agent use, or a history of di-
abetes. Of the 20 subjects with diabetes, 10
w e re taking oral hypoglycemic agents, and
10 were treated with diet only. Of those
taking oral hypoglycemic agents, all 10
w e re taking a sulfonylurea; in addition, 3
w e re taking metformin, and 1 was taking
t ro g l i t a z o n e .
Venous blood samples were drawn in
the morning with the subject fasting and
the serum stored airtight at 70° for
years before analysis. Hormones were mea-
s u red by radioimmunoassay (RIA). Materi-
als for the RIA of testosterone (nonextrac-
2
tion coated-tube method), FT (nonpro t e i n -
bound testosterone), and dehydro e p i a n d ro s-
t e ro n esulfate (DHEAS) were obtained fro m
Diagnostic Products (Los Angeles, CA), and
for estradiol (third generation), estrone, an-
d rostenedione, cortisol, and SHBG fro m
Diagnostic Systems Laboratories (We b s t e r,
TX). Total cholesterol, LDL cholestero l ,
HDL cholesterol, and triglyceride levels
w e re measured using a Hitachi 705 auto-
mated spectrometer (Boehringer Mannheim,
Indianapolis, IN).
Statistical analyses were perf o rm e d
using SPSS Version 6.1 (SPSS, Chicago,
IL) on a Macintosh Perf o rma 6300 (Apple
C o m p u t e r, Cupertino, CA) computer to
calculate means, SEM, outliers (using box
plots), and partial correlation coeff i c i e n t s .
Each of the variables measured was found
to be normally distributed when using the
K o l m o g o ro v - S m i rnov test. Analysis of
variance (with WHR as a covariate) was
used to compare the means of the vari-
ables, and Fisher’s exact test was used to
c o m p a re the number of smokers between
the control and diabetic groups. The sig-
nificance of the diff e rence between the
slopes of two re g ression lines was deter-
mined by comparing the means of the
s q u a red Studentized re s i d u a l s .
R E S U LT S — A comparison of the non-
h o rmonal variables determined in the dia-
betic and control subjects is shown in
Table 1. Mean age, BMI, and other risk fac-
tors for CHD were not significantly diff e r-
ent, but mean WHR was significantly
higher in the diabetic subjects.
Table 2 shows comparisons (con-
t rolled for WHR) of the mean levels of the
h o rmones measured and SHBG between
diabetic and control subjects. FT, DHEAS,
and estradiol were significantly higher and
a n d rostenedione and estrone were mar-
ginally higher in the 20 diabetic subjects.
With the statistical outliers re m o v e d ,
t e s t o s t e rone and androstenedione were
significantly higher and SHBG was mar-
ginally lower in the diabetic subjects. FT,
estradiol, and cortisol each had one, and
t e s t o s t e rone and androstenedione each
had three statistical outliers. All of the out-
liers were in the control group; DHEAS
had no outliers. The mean level of cort i s o l
was not significantly diff e rent between di-
abetic and control subjects.
To determine whether the oral hypo-
glycemic agents taken by 10 of the diabetic
subjects may have confounded the re s u l t s ,
the mean values for the variables measure d
w e re compared between the 10 diabetic
subjects controlled by diet only and the 29
c o n t rol subjects (Table 2). Although the
mean age and BMI were not significantly
d i ff e rent between these two groups, the
mean WHR was significantly higher (P =
0.02) in the diabetic subjects; thus, the
comparisons were controlled for WHR.
Notwithstanding the decreased statistical
power resulting from the smaller number
of diabetic subjects, this group of diabetic
subjects still showed hyperandro g e n e m i a
and hypere s t rogenemia. Thus, these find-
ings could not be attributed to an effect of
the oral hypoglycemic agents.
To determine whether FT or testos-
t e rone correlated significantly with risk fac-
tors for CHD other than diabetes, corre l a-
tions (controlled for age and BMI) were
c a rried out in the diabetic and control sub-
jects separately. In the diabetic subjects, FT
correlated with diastolic blood pressure
(r= 0.48, P= 0.04). In the control subjects,
t e s t o s t e rone correlated with systolic blood
p re s s u re (r= 0.50, P
t e rol (r= 0.41, P0.04), and LDL choles-
t e rol (r = 0.38, P= 0.05), whereas FT cor-
related with systolic blood pre s s u re (r =
0.49, P 0.01) and total cholesterol (r =
0.41, P 0.04). SHBG correlated inversely
with FT/testosterone in both the diabetic
(r= 0.72, P= 0.001) and control subjects
(r= 0.46, P 0.02); the slopes of the re-
g ression lines showed no significant diff e r-
ence between these two groups (P= 0.95)
0.01), total choles-
CONCLUSIONS — In the pre s e n t
s t u d y, the mean levels of the androgens FT
and DHEAS were significantly higher in
the diabetic subjects than in the contro l
subjects. When the statistical outliers were
Table 1—Comparison of variables between diabetic and control subjects
C o n t rol subjects Diabetic subjectsP
n
Age (years)
BMI (kg/m2)
W H R
Systolic blood pre s s u re (mmHg)
Diastolic blood pre s s u re (mmHg)
Total cholesterol (mg/dl)
LDL cholesterol (mg/dl)
HDL cholesterol (mg/dl)
Triglyceride (mg/dl)
C u rrent cigarette smokers
Duration of diabetes (years)
Data are n, means ± SEM, or n( % ) .
2 92 0—
66.8 ± 0.8
28.9 ± 0.81
0.858 ± 0.017
147 ± 3.4
85.4 ± 2.0
206 ± 6.1
131.6 ± 5.9
50.8 ± 2.68
116 ± 11.7
5 (17)
—
65.6 ± 0.9
28.9 ± 0.86
0.929 ± 0.022
154 ± 4.7
90.5 ± 2.8
201 ± 9.7
126.7 ± 8.8
47.9 ± 3.24
134 ± 15.5
1 (5)
6.6 ± 1.3
0 . 3 4 5
0 . 9 8 0
0 . 0 1 3
0 . 2 1 1
0 . 1 3 7
0 . 7 0 2
0 . 6 3 1
0 . 4 9 5
0 . 3 4 9
0 . 3 7 9
—

Page 3

76DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000
Hyperandrogenemia in type 2 diabetes
excluded, the mean levels of andro s t e n e-
dione and total testosterone also were
significantly higher. Thus, the diabetic
subjects exhibited hyperandro g e n e m i a
with elevation of the main andro g e n ,
t e s t o s t e rone, and two weaker andro g e n s
(DHEAS and androstenedione). All of the
comparisons were controlled for WHR,
the mean value of which was significantly
higher in the diabetic subjects. An in-
c rease in the levels of dehydro e p i a n d ro s-
t e ro n e(DHEA) (32), DHEAS (33), and an-
d rostenedione (32) has been re p o rt e d
p reviously in women with type 2 diabetes,
while one study re p o rted a decrease in
DHEA (5), and another re p o rted no signif-
icant diff e rence in testosterone or FT (34).
Although 90% of DHEAS and 50% of
a n d rostenedione and testosterone derive
f rom the adrenal in women (35), the cort i-
sol level was found to be normal in the di-
abetic subjects in the present study. An in-
c rease in cortisol, DHEA, and DHEAS
levels has been re p o rted in pre m e n o p a u s a l
women with PCOS, obesity , and diabetes
c o m p a red with women with PCOS and
obesity without diabetes (36).
That FT was more strongly related to
diabetes than testosterone in the pre s e n t
study implicates SHBG, whose concentra-
tion appears to affect the ratio of FT to
t e s t o s t e rone (37). Indeed, the ratio of FT
to testosterone correlated inversely with
SHBG in both the diabetic and contro l
subjects. But the mean SHBG level was
only marginally (P = 0.055) decreased in
the diabetic subjects and only after re m o v-
ing the statistical outliers. Two pre v i o u s
studies, however, have re p o rted a de-
c reased SHBG level and an incre a s e d
t e s t o s t e rone/SHBG ratio (which was used
as an index of FT level) in postmenopausal
women with type 2 diabetes (5) or im-
p a i red glucose tolerance (38). The pre s e n t
study appears to be the first to re p o rt an
i n c rease in directly measured FT in post-
menopausal women with type 2 diabetes.
A low SHBG level has also been re p o rt e d
to be an independent risk factor for type 2
diabetes in women (39,40). The SHBG
level has been re p o rted to be elevated in
postmenopausal women with insulin-
t reated diabetes (41).
The mean estradiol level was also
found to be significantly increased in the
diabetic subjects in the present study . An
i n c rease in the estradiol and estrone levels
has been re p o rted previously in post-
menopausal women with insulin-tre a t e d
diabetes (41), and an increase in the es-
t rone level but not the estradiol level has
been re p o rted in postmenopausal women
with non-insulin-treated diabetes (5). The
main source of estrogen in postmenopausal
women appears to be the aromatization of
plasma androstenedione to estrone (42).
Estradiol is produced both by the interc o n-
version of estrone (43) and the aro m a t i z a-
tion of testosterone (44). Although this pe-
ripheral aromatization of androgens to
e s t rogens has been re p o rted to corre l a t e
positively with adiposity (45), the BMI of
the diabetic subjects in the present study
did not differ significantly from that of the
c o n t rol subjects. Although the WHR of the
diabetic subjects was significantly gre a t e r
than that of the control subjects, the hyper-
e s t rogenemia was observed after statisti-
cally controlling for WHR. Another possi-
ble mechanism for the hypere s t ro g e n e m i a
is that the increased levels of andro s t e n e-
dione and testosterone in the diabetic sub-
jects may lead through mass action to an
i n c rease in estrogen product. The hyper-
t e s t o s t e ronemia that occurs in patients
with other disorders, such as PCOS
(46,47) and hirsutism (6), however, has
been re p o rted to be associated with norm a l
e s t rogen levels. The concurrence of hyper-
t e s t o s t e ronemia and hypere s t ro g e n e m i a
has been re p o rted in women with in-
c reased WHR (4), hypert ension (17), and
re f r a c t o ry depression (48). That hyperan-
d rogenemia and hypere s t rogenemia were
also observed in the diabetic subjects
t reated with diet only indicates that these
a b n o rmalities were not effects of the oral
hypoglycemic agents.
The finding of hyperandrogenemia in
the diabetic subjects in the present study
is consistent with the hypothesis that hy-
p e r a n d rogenemia may underlie and link
risk factors for CHD in women (14,16).
But whether the hyperandrogenemia and
diabetes are causally related and, if so,
which is the cause are unclear. The associ-
ation between hyperandrogenemia and
diabetes in premenopausal women with
PCOS (20,21), together with a positive
c o rrelation between testosterone and in-
sulin (3,20,21) in these women, support s
the possibility of a cause-and-effect re l a-
tionship. The positive correlation of
t e s t o s t e rone with insulin (5,14) and glu-
cose (49) in postmenopausal women pro-
vides further support. That testostero n e
may underlie diabetes by producing in-
sulin resistance is suggested by re p o rt s
that administration of testosterone esters
(50) or methyltestosterone (51) to healthy
p remenopausal women induced insulin
resistance. Furt h e rm o re, it has been re-
p o rted that lowering the androgen level in
h y p e r a n d rogenemic women with each of
t h ree diff e rent drugs resulted in a decre a s e
in insulin resistance (52). Thus, these ob-
s e rvations suggest that, in women, hyper-
a n d rogenemia may lead to insulin re s i s-
tance and hyperinsulinemia.
Table 2—Comparison of hormone and SHBG levels between diabetic and control subjects
Diabetic subjects
t reated with
diet only
Pw i t h o u t
o u t l i e r s *
Pw i t h o u t
o u t l i e r s *C o n t rol subjectsDiabetic subjectsP*P*
n
E s t rone (pg/ml)
Estradiol (pg/ml)
Te s t o s t e rone (ng/ml)
FT (pg/ml)
SHBG (nmol/l)
DHEAS (µg/dl)
A n d rostenedione (ng/ml)
C o rtisol (µg/dl)
Data are nor means ± SEM. *Controlled for WHR.
2 92 0—— 1 0——
19.6 ± 1.5
22.0 ± 1.2
0.27 ± 0.04
0.65 ± 0.10
138 ± 17.4
55.6 ± 4.3
1.30 ± 0.12
11.7 ± 1.0
25.5 ± 3.1
27.5 ± 1.7
0.29 ± 0.02
1.00 ± 0.10
97 ± 19.5
81.0 ± 9.2
1.58 ± 0.13
12.0 ± 1.6
0 . 0 7 8
0 . 0 1 3
0 . 4 8 3
0 . 0 1 1
0 . 3 7 5
0 . 0 3 7
0 . 0 8 3
0 . 8 6 0
0 . 2 6 7
0 . 0 0 3
0 . 0 5 3
0 . 0 0 3
0 . 0 5 5
0 . 0 3 7
0 . 0 0 2
0 . 8 6 2
28.5 ± 5.6
30.8 ± 2.7
0.30 ± 0.03
1.07 ± 0.17
96 ± 32.6
76.5 ± 10.1
1.57 ± 0.13
10.9 ± 1.2
0 . 0 4 1
0 . 0 0 2
0 . 5 4 3
0 . 0 2 2
0 . 5 5 7
0 . 1 6 2
0 . 1 4 6
0 . 4 5 8
0 . 3 9 8
0 . 0 0 1
0 . 1 1 3
0 . 0 0 9
0 . 0 0 9
0 . 1 6 2
0 . 0 0 3
0 . 5 9 3

Page 4

DIABETESCARE, VOLUME 23, NUMBER 1, JANUARY 2000 77
Phillips and Associates
T h e re is also evidence, however, that
hyperinsulinemia could give rise to hyper-
a n d rogenemia. Marked hyperinsulinemia
s e c o n d a ry to severe insulin resistance of
d i ff e rent etiologies appears to be associated
with hyperandrogenemia (53). Also, a de-
p ressing effect of insulin on the SHBG level
(54) might increase the FT/testosterone
ratio (37) and thus the androgenicity inas-
much as FT may be the biologically active
component of total testosterone. Decre a s-
ing both the insulin level (55) and insulin
resistance (47,56,57) in women with
PCOS by using drugs has been re p o rted to
d e c rease testosterone (55–57), FT (56,57),
non–SHBG-bound testosterone (47,55),
DHEAS (47,56), and androstenedione (56)
levels, whereas decreasing the andro g e n
level in women with PCOS by using a go-
n a d o t ro p i n - releasing hormone agonist did
not appear to affect insulin resistance in
t h ree studies (58–60) but did relieve mild
insulin resistance in another study (61).
Thus, these observations suggest that, in
women, hyperinsulinemia could lead to
h y p e r a n d rogenemia. However, insulin in-
fusion in premenopausal women, either
healthy or with PCOS or marked insulin
resistance, has been re p o rted to incre a s e
the androstenedione (62,63) but not the
t e s t o s t e rone (62–64) or DHEAS level (64);
in fact, decreases in testosterone (63),
non–SHBG-bound testosterone (63), and
DHEAS (64) and an increase in estradiol
(63) have been re p o rted. Hyperinsuline-
mia from glucose challenge similarly ap-
pears, if anything, to lower rather than
raise androgen levels (65,66). That PCOS
has been re p o rted to occur without insulin
resistance or hyperinsulinemia (67) sug-
gests that the insulin resistance and hyper-
insulinemia may not be responsible for the
h y p e rt e s t o s t e ronemia of PCOS. In sum-
m a ry, while there is evidence in women
that hyperandrogenemia may underlie hy-
perinsulinemia and insulin resistance,
t h e re is also evidence that hyperinsuline-
mia and insulin resistance may underlie
h y p e r a n d ro g e n e m i a .
Whether the hypere s t rogenemia ob-
s e rved in the diabetic subjects in the pre s-
ent study has a causal relationship to dia-
betes is also unclear. Lower fasting glucose
and insulin levels in postmenopausal
women taking estrogen suggest a favorable
e ffect of estrogen on glucose metabolism
(68), but ethinyl estradiol administere d
orally to premenopausal women appears
to decrease insulin sensitivity (69). The ef-
fect of estrogen may depend on the estro-
gen preparation, dose, and route of ad-
ministration. In postmenopausal women,
oral equine estrogens have been re p o rt e d
to increase insulin sensitivity at a lower
dose but decrease it at a higher dose (70),
w h e reas transdermal estrogen at a dose
similar to the higher dose increased in-
sulin sensitivity (71).
In conclusion, type 2 diabetes in post-
menopausal women was found to be
a ssociated with hyperandrogenemia and
h y p e re s t rogenemia. The hypere s t ro g e n e-
mia may have been secondary to the hy-
p e r a n d rogenemia. Evidence that hyper-
a n d rogenemia is associated with CHD
and its risk factors in women raises the
possibility that hyperandrogenemia may
be a link between diabetes and CHD in
women and may underlie both. If this is
the case, lowering the androgen level in
women may have a preventive effect on
both diabetes and CHD.
A c k n o w l e d g m e n t s— This study was sup-
p o rted in part by Mr. and Mrs. Chang Yue Va n ;
the Myron C. Patterson, MD, Fund; and grants
f rom the National Institute of Neurological Dis-
o rders and Stroke (R01-NS-29993) and the
General Clinical Research Center (2-M01-
RR00645).
R e f e re n c e s
1 . Vague J: The degree of masculine diff e re n-
tiation of obesities: a factor determ i n i n g
p redisposition to diabetes, athero s c l e ro s i s ,
gout, and uric acid calculous disease. Am J
Clin Nutr 4:20–34, 1956
2 . Evans DJ, Hoffmann RG, Kalkhoff RK,
Kissebah AH: Relationship of andro g e n i c
activity to body fat topography, fat cell
m o r p h o l o g y, and metabolic aberrations in
p remenopausal women. J Clin Endocrinol
M e t a b57:304–310, 1983
3 . Hauner H, Ditschuneit HH, Pal SB, Mon-
cayo R, Pfeiffer EF: Fat distribution, en-
docrine and metabolic profile in obese
women with and without hirsutism. M e-
t a b o l i s m37:281–286, 1988
4 . Kirschner MA, Samojlik E, Drejka M,
Szmal E, Schneider G, Ertel N: Andro g e n -
e s t rogen metabolism in women with
upper body versus lower body obesity.
J Clin Endocrinol Metab70:473–479, 1990
5 . Andersson B, Marin P , Lissner L, Ve r-
meulen A, Bjorntorp P: Te s t o s t e rone con-
centrations in women and men with
NIDDM. Diabetes Care17:405–411, 1994
6 . B e rnasconi D, Del Monte P , Meozzi M,
Randazzo M, Marugo A, Badaracco B,
M a rugo M: The impact of obesity on hor-
monal parameters in hirsute and nonhir-
sute women. M e t a b o l i s m45:72–75, 1996
7 . Lapidus L, Bengtsson C, Larsson B, Pen-
n e rt K, Rybo E, Sjostrom L: Distribution of
adipose tissue and risk of card i o v a s c u l a r
disease and death: a 12 year follow up of
p a rticipants in the population study of
women in Gothenburg, Sweden. Br Med J
289:1257–1261, 1984
8 . Gillum RF: The association of body fat dis-
tribution with hypertension, hypert e n s i v e
h e a rt disease, coro n a ry heart disease, dia-
betes and cardiovascular risk factors in
men and women aged 18–79 years.
J Chronic Dis40:421–428, 1987
9 . Hauner H, Bognar E, Blum A: Body fat dis-
tribution and its association with meta-
bolic and hormonal risk factors in women
with angiographically assessed coro n a ry
a rt e ry disease: evidence for the presence of
a metabolic syndrome. A t h e ro s c l e ro s i s1 0 5 :
209–216, 1994
1 0 . Wild RA, Grubb B, Hartz A, Van Nort JJ,
Bachman W, Bartholomew M: Clinical signs
of androgen excess as risk factors for coro-
n a ry art e ry disease. F e rtil Steril 5 4 : 2 5 5 – 2 5 9 ,
1 9 9 0
1 1 . H a rtz A, Grubb B, Wild R, Van Nort JJ,
Kuhn E, Freedman D, Rimm A: The asso-
ciation of waist hip ratio and angiographi-
cally determined coro n a ry art e ry disease.
Int J Obes 14:657–665, 1990
1 2 . Thompson CJ, Ryu JE, Craven TE, Kahl
FR, Crouse JR III: Central adipose distri-
bution is related to coro n a ry athero s c l e ro-
sis. A rterioscler Throm b 11:327–333, 1991
1 3 . B i rdsall MA, Farquhar CM, White HD: As-
sociation between polycystic ovaries and
extent of coro n a ry art e ry disease in
women having cardiac catheterization.
Ann Intern Med126:32–35, 1997
1 4 . Phillips GB, Pinkernell BH, Jing TY: Rela-
tionship between serum sex horm o n e s
and coro n a ry art e ry disease in post-
menopausal women. A rterioscler Thro m b
Vasc Biol 17:695–701, 1997
1 5 . Adams MR, Williams JK, Kaplan JR: Eff e c t s
of androgens on coro n a ry art e ry athero-
sclerosis and atherosclerosis-related im-
p a i rment of vascular responsiveness. A rt e -
rioscler Thro mb Vasc Biol15:562–570, 1995
1 6 . Phillips GB: Relationship of serum sex hor-
mones to coro n a ry heart disease. S t e ro i d s
58:286–290; discussion 291–292; re f u t a-
tion 554–555, 1993
1 7 . Hughes GS, Mathur RS, Margolius HS: Sex
s t e roid hormones are altered in essential hy-
p e rtension. J Hypertension 7:181–187, 1989
1 8 . Phillips GB, Jing TY, Laragh JH: Serum sex
h o rmone levels in postmenopausal women
with hypertension. J Hum an Hypertens 1 1 :
523–526, 1997
1 9 . Friedman AJ, Ravnikar VA, Barbieri RL:
S e rum steroid hormone profiles in post-
menopausal smokers and nonsmokers.
F e rtil Steril 47:398–401, 1987
2 0 . B u rghen GA, Given SJ, Kitabchi AE: Cor-
relation of hyperandrogenism with hyper-

Page 5

78DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000
Hyperandrogenemia in type 2 diabetes
insulinism in polycystic ovarian disease.
J Clin Endo crinol Metab50:113–116, 1980
2 1 . Shoupe D, Lobo RA: The influence of an-
d rogens on insulin resistance. F e rtil Steril
41:385–388, 1984
2 2 . D a h l g ren E, Johansson S, Lindstedt G,
Knutsson F , Oden A, Janson PO, Mattson
LA, Crona N, Lundberg PA: Women with
polycystic ovary syndrome wedge re s e c t e d
in 1956 to 1965: a long-term follow-up fo-
cusing on natural history and circ u l a t i n g
h o rmones. F e rtil Steril 57:505–513, 1992
2 3 . Conway GS, Agrawal R, Betteridge DJ, Ja-
cobs HS: Risk factors for coro n a ry art e ry
disease in lean and obese women with the
polycystic ovary syndrome. Clin Endo crinol
37:119–125, 1992
2 4 . Talbott E, Guzick D, Clerici A, Berga S,
D e t re K, Weimer K, Kuller L: Coro n a ry
h e a rt disease risk factors in women with
polycystic ovary syndrome. A rt e r i o s c l e r
T h ro m b15:821–826, 1995
2 5 . Soler JT, Folsom AR, Kaye SA, Prineas RJ:
Associations of abdominal adiposity, fasting
insulin, sex hormone binding globulin,
and estrone with lipids and lipoproteins in
post-menopausal women. Atherosclerosis
79:21–27, 1989
2 6 . K a l k h o ff RK, Hartz AH, Rupley D, Kissebah
AH, Kelber S: Relationship of body fat dis-
tribution to blood pre s s u re, carbohydrate
tolerance, and plasma lipids in healthy
obese women. J Lab Clin Med 1 0 2 : 6 2 1 – 6 2 7 ,
1 9 8 3
2 7 . Lovejoy JC, Bray GA, Bourgeois MO, Mac-
chiavelli R, Rood JC, Greeson C, Part i n g t o n
C: Exogenous androgens influence body
composition and regional body fat distri-
bution in obese postmenopausal women: a
clinical research center study. J Clin En -
docrinol Metab 81:2198–2203, 1996
2 8 . B j o rntorp P: “Portal” adipose tissue as a
generator of risk factors for card i o v a s c u l a r
disease and diabetes. A rt e r i o s c l e rosis 1 0 :
493–496, 1990
2 9 . Kannel WB: Lipids, diabetes, and coro n a ry
h e a rt disease: insights from the Framing-
ham Study. Am Heart J 1 1 0 : 1 1 0 0 – 1 1 0 7 ,
1 9 8 5
3 0 . Sacco RL, Boden-Albala B, Gan R, Chen X,
K a rgman DE, Shea S, Palk MC, Hauser WA :
S t roke incidence among white, black, and
Hispanic residents of an urban community.
Am J Epidem io l 147:259–268, 1998
3 1 . G a m b rell RD Jr: The menopause: benefits
and risks of estro g e n - p rogestogen re p l a c e-
ment therapy. F e rtil Steril 3 7 : 4 5 7 – 4 7 4 ,
1 9 8 2
3 2 . Szpunar WE, Blair AJ Jr, McCann DS:
Plasma androgen concentrations in diabetic
women. D i a b e t e s26:1125–1129, 1977
3 3 . B a rrett-Connor E, Ferrara A: Dehy-
d ro e p i a n d ro s t e rone, dehydro e p i a n d ro s-
t e rone sulfate, obesity, waist-hip ratio, and
noninsulin-dependent diabetes in post-
menopausal women: the Rancho Bern a rd o
S t u d y. J Clin Endocrinol Metab 8 1 : 5 9 – 6 4 ,
1 9 9 6
3 4 . Stamataki KE, Spina J, Rangou DB, Chlou-
verakis CS, Piaditis GP: Ovarian function
in women with non-insulin dependent dia-
betes mellitus. Clin Endocrinol4 5 : 6 1 5 – 6 2 1 ,
1 9 9 6
3 5 . Abraham GE: Ovarian and adrenal contri-
bution to peripheral androgens during the
m e n s t rual cycle. J Clin Endocrinol Metab
39:340–346, 1974
3 6 . B u ffington CK, Givens JR, Kitabchi AE:
Enhanced adre n o c o rtical activity as a con-
tributing factor to diabetes in hyperandro-
genic women. M e t a b o l i s m 4 3 : 5 8 4 – 5 9 0 ,
1 9 9 4
3 7 . Andersson DC: Sex-horm o n e - b i n d i n g
globulin. Clin Endocrinol 3:69–96, 1974
3 8 . Larsson H, Ahren B: Androgen activity as a
risk factor for impaired glucose tolerance
in postmenopausal women. Diabetes Care
19:1399–1403, 1996
3 9 . Lindstedt G, Lundberg PA, Lapidus L,
L u n d g ren H, Bengtsson C, Bjorntorp P:
Low sex-hormone-binding globulin con-
centration as independent risk factor for
development of NIDDM: 12-yr follow-up
of population study of women in Gothen-
b u rg, Sweden. D i a b e t e s40:123–128, 1991
4 0 . H a ffner SM, Valdez RA, Morales PA ,
Hazuda HP , Stern MP: Decreased sex hor-
mone-binding globulin predicts nonin-
sulin-dependent diabetes mellitus in
women but not in men. J Clin Endocrinol
M e t a b77:56–60, 1993
4 1 . Nyholm H, Djursing H, Hagen C, Agner T,
Bennett P , Svenstrup B: Androgens and es-
t rogens in postmenopausal insulin-tre a t e d
diabetic women. J Clin Endocrinol Metab
69:946–949, 1989
4 2 . G rodin JM, Siiteri PK, MacDonald PC:
S o u rce of estrogen production in post-
menopausal women. J Clin Endocrinol
M e t a b36:207–214, 1973
4 3 . Longcope C, Tait JF: Validity of metabolic
clearance and interconversion rates of es-
t rone and 17 -estradiol in normal adults.
J Clin Endocrinol Metab 32:481–490, 1971
4 4 . Longcope C, Kato T, Horton R: Conver-
sion of blood androgens to estrogens in
n o rmal adult men and women. J Clin In -
v e s t48:2191–2201, 1969
4 5 . Longcope C, Baker R, Johnston CC Jr: An-
d rogen and estrogen metabolism: re l a t i o n-
ship to obesity. M e t a b o l i s m 3 5 : 2 3 5 – 2 3 7 ,
1 9 8 6
4 6 . Franks S: Polycystic ovary syndrome.
N Engl J Med333:853–861, 1995
4 7 . Dunaif A, Scott D, Finegood D, Quintana
B, Whitcomb R: The insulin-sensitizing
agent troglitazone improves metabolic and
re p roductive abnormalities in the polycys-
tic ovary syndrome. J Clin Endocrinol Metab
81:3299–3306, 1996
4 8 . Vogel W, Klaiber EL, Bro v e rman DM: Roles
of the gonadal steroid hormones in psychi-
atric depression in men and women. P ro g
N e u ro - P s y c h o p h a rm a c o l 2:487–503, 1978
4 9 . Khaw KT, Barrett-Connor E: Fasting
plasma glucose levels and endogenous an-
d rogens in non-diabetic postmenopausal
women. Clin Sci 80:199–203, 1991
5 0 . P o l d e rman KH, Gooren LJG, Asscheman
H, Bakker A, Heine RJ: Induction of insulin
resistance by androgens and estrogens.
J Clin Endocrino l Metab79:265–271, 1994
5 1 . Diamond MP , Grainger D, Diamond MC,
S h e rwin RS, DeFronzo RA: Effects of
m e t h y l t e s t o s t e rone on insulin secre t i o n
and sensitivity in women. J Clin Endocrino l
Metab 83:4420–4425, 1998
5 2 . Moghetti P , Tosi F, Castello R, Magnani
CM, Negri C, Brun E, Furlani L, Caputo
M, Muggeo M: The insulin resistance in
women with hyperandrogenism is par-
tially reversed by antiandrogen tre a t m e n t :
evidence that androgens impair insulin ac-
tion in women. J Clin Endocrinol Metab
81:952–960, 1996
5 3 . Taylor SI, Dons RF , Hernandez E, Roth J,
G o rden P: Insulin resistance with andro-
gen excess in women with autoantibodies
to the insulin re c e p t o r. Ann Intern Med
97:851–855, 1982
5 4 . Nestler JE: Sex hormone-binding globulin:
a marker for hyperinsulinemia and/or in-
sulin resistance? J Clin Endocrinol Metab
76:273–274, 1993
5 5 . Nestler JE, Barlascini CO, Matt DW, Stein-
gold KA, Plymate SR, Clore JN, Blackard
WG: Suppression of serum insulin by dia-
zoxide reduces serum testosterone levels
in obese women with polycystic ovary
s y n d rome. J Clin Endocrinol Metab 6 8 :
1027–1032, 1989
5 6 . Velazquez EM, Mendoza S, Hamer T, Sosa
F, Glueck CJ: Metformin therapy in poly-
cystic ovary syndrome reduces hyperinsu-
linemia, insulin resistance, hyperandro-
genemia, and systolic blood pre s s u re ,
while facilitating normal menses and pre g-
n a n c y. M e t a b o l i s m43:647–654, 1994
5 7 . E h rmann DA, Schneider DJ, Sobel BE, Cav-
aghan MK, Imperial J, Rosenfield RL, Polon-
sky KS: Troglitazone improves defects in in-
sulin action, insulin secretion, ovarian
s t e roidogenesis, and fibrinolysis in women
with polycystic ovary syndrome. J Clin En -
docrinol Metab 82:2108–2116, 1997
5 8 . G e ffner ME, Kaplan SA, Bersch N, Golde
D W, Landaw EM, Chang RJ: Persistence of
insulin resistance in polycystic ovarian dis-
ease after inhibition of ovarian steroid se-
c retion. F e rtil Steril 45:327–333, 1986
5 9 . Dunaif A, Green G, Futterweit W, Dobrjan-
sky A: Suppression of hyperandro g e n i s m
does not improve peripheral or hepatic in-
sulin resistance in the polycystic ovary syn-
d rome. J Clin Endo crinol Metab 7 0 : 6 9 9 – 7 0 4 ,
1 9 9 0
6 0 . Dale O, Tanbo T, Djoseland O, Jervell J,
Abyholm T: Persistence of hyperinsuline-