0021-972X/98/$03.00/0
Journal of Clinical Endocrinology and Metabolism
Copyright © 1998 by The Endocrine Society
Vol. 83, No. 2
Printed in U.S.A.
Evidence of High Circulating Testosterone in Women
with Prior Preeclampsia*
HANNELE LAIVUORI, RISTO KAAJA, EEVA-MARJA RUTANEN, LASSE VIINIKKA,
AND OLAVI YLIKORKALA
Department of Obstetrics and Gynecology, and Clinical Chemistry, Helsinki University Central
Hospital, Finland
ABSTRACT
Women with prior preeclampsia are characterized by hyperinsulinemia and a 2- to 3-fold excess risk of hypertension and ischemic
heart disease in later life. We therefore studied whether these women
present changes in pituitary, ovarian, and endothelial factors that
could also affect the risk of vascular disorders. Twenty-two women
with prior preeclampsia and 22 control women matched by age and
body mass index were studied an average of 17 yr after delivery.
Women with prior preeclampsia had elevated serum free testosterone
levels (20.6 6 2.2 vs. 15.0 6 1.3 pmol/L, mean 6 SE, P 5 0.03), an
elevated free androgen index (3.2 6 0.5 vs. 1.9 6 0.2, P 5 0.04), and
an elevated free testosterone estradiol ratio (0.089 6 0.017 vs. 0.046 6
0.006, P 5 0.02). The levels of insulin-like growth factor binding
protein-1 decreased as expected during a 3-h oral glucose tolerance
test without differences between the groups. Levels of FSH, LH,
androstenedione, dehydroepiandrosterone sulfate, and endothelin-1,
as well as urinary output of prostacyclin and thromboxane A2 metabolites, showed no difference between study groups. A history of
preeclampsia an average of 17 yr earlier thus appears to be associated
with elevated levels of testosterone, which may contribute to the
increased risk of vascular morbidity in such women. (J Clin Endocrinol Metab 83: 344 –347, 1998)
W
(n 5 2) during pregnancy (16) and 22 age- and body mass index (BMI)matched controls, each of whom had given birth at approximately the
same time (63– 4 months) after a normotensive pregnancy, were studied
17 yr after the first pregnancy (Table 1). All patients and controls had
been healthy before this index pregnancy. In the meantime, 15 women
in the patient group had given birth to 21 infants. Of the second pregnancies (n 5 15), five (33%) had been complicated by preeclampsia, and
four (27%) by hypertension without proteinuria. In their third pregnancies (n 5 4), two developed preeclampsia, and in their fourth pregnancies (n 5 2), one was hypertensive. In the control group, 19 women
women had given birth 27 times, and none had developed preeclampsia
or hypertension without proteinuria. Three women (one patient and two
controls) were excluded because of their use of hormone replacement
therapy or progestin-only contraception. Two women in the patient
group and one in the control group wearing levonorgestrel-releasing
intrauterine devices (Levonova, Leiras, Finland) were not excluded,
because such women ovulate and are thus endocrinologically representative (17). The patients and controls menstruated regularly except
for nine, of whom four patients and two controls were hysterectomized
and three used Levonova. All women were studied during the same
phase of their menstrual cycle, as reported earlier (11). No one presented
clinical features of androgen excess such as hirsutism, acne, or alopecia,
but the patient group exhibited hyperinsulinemia, one marker of insulin
resistance (11). The women came to the research center at 0800 h after
an overnight fast. After the physical examination (height, weight, and
blood pressure measurement), blood was drawn for measurement of
uric acid, lipids, baseline blood glucose, and serum insulin, and a urine
sample was collected. Thereafter, a 3-h oral glucose tolerance test was
started. The patients and controls were advised to refrain from using
aspirin or other nonsteroidal antiinflammatory drugs for 10 days before
the study.
Basal blood samples were assayed for sex steroids and relevant protein hormones, and in addition, urine samples were assayed for prostacyclin and thromboxane A2 metabolites with established methods
(Table 2). Free testosterone was calculated by using the formula: serum
total testosterone (in pmol/L)/100 3 [2.28 –1.38 3 logarithm serum
sex-hormone binding globulin (SHBG) in nmol/L/10] (19). The ratio
between serum total testosterone and SHBG that is commonly expressed
as the free androgen index was calculated from serum total testosterone
concentration in nmol/L 3 100 divided by serum SHBG concentration
in nmol/L (20). The prostanoid data are expressed against creatinine to
avoid possible errors caused by differences in urine dilution. The data
OMEN with a prior preeclamptic pregnancy have an
increased risk of cardiovascular disease in their subsequent life (1–7). This may imply that some harmful endocrine
or metabolic changes thought to be specific for preeclamptic
pregnancy may persist after pregnancy and perhaps predispose
these women to increased risk of vascular disorders. These
changes may involve endothelial function deficient in preeclampsia, as seen from reduced prostacyclin and/or elevated
endothelin-1 or thromboxane A2 production (8). However, no
long-term data exist on prostacyclin, thromboxane A2, and endothelin-1 after preeclamptic pregnancy. Another explanation
may involve metabolism that presents changes similar to those
in metabolic syndrome in preeclampsia (9, 10). We have reported that women with a preeclamptic first pregnancy show
hyperinsulinemia 17 yr later (11). Because, on the other hand,
hyperinsulinemia and possibly changes in the insulin-like
growth factor (IGF) system and hyperandrogenism are key
features in polycystic ovarian disease (PCO) (12), and because
women with PCO also show excess risk of preeclampsia (13, 14)
and subsequent cardiovascular morbidity (15), it seemed pertinent to study endothelial function and endocrine changes in
women with a prior preeclamptic pregnancy.
Subjects and Methods
Women who had suffered severe preeclampsia (n 5 20) (blood pressure constantly .160/110 mmHg; proteinuria .0.3 g/24 h) or eclampsia
Received July 23, 1997. Revision received October 8, 1997. Accepted
October 15, 1997.
Address all correspondence and requests for reprints to: Hannele
Laivuori, Department of Obstetrics and Gynecology, University Central
Hospital of Helsinki, Haartmaninkatu 2, FIN-00290 Helsinki, Finland.
E-mail: [email protected].
* This work was supported by grants from the Finnish Academy of
Science, the Obstetric and Gynecology Research Foundation, and the
Clinical Research Institute of the Helsinki University Central Hospital.
344
CIRCULATING TESTOSTERONE AFTER PREECLAMPSIA
are given as means 6 standard error (se). Comparisons between groups
were made by Student’s two-tailed unpaired t test (Statview II Program,
Abacus Concepts, Berkeley, CA). Linear regression analyses served to
assess the relationships between parameters.
Results
The study groups were comparable except for blood pressure, which was higher in the patient group than that in the
control group (Table 1). Women with prior preeclampsia
were characterized by elevated serum free testosterone levels
(P 5 0.03), free androgen index (P 5 0.04), and free testosterone/estradiol ratio (P 5 0.02) (Table 3), whereas androstenedione, dehydroepiandrosterone sulfate, estradiol, and
total testosterone were normal (Table 3). Free testosterone
correlated positively with basal insulin (reported previously,
Ref. 5) (r 5 0.52, P 5 0.016), systolic blood pressure (r 5 0.69,
P 5 0.001), diastolic blood pressure (r 5 0.62, P 5 0.004),
triglycerides (reported previously, Ref. 5) (r 5 0.55, P 5
0.009), and BMI (r 5 0.51, P 5 0.018) in women with prior
preeclampsia, but not in the controls. The study groups did
not differ with respect to SHBG, thyroxine, LH, or FSH concentrations or LH/FSH ratio (Table 3).
The oral glucose tolerance test caused a progressive fall in
IGF binding protein-1 (IGFBP-1 levels), but this response, as
345
well as the basal levels of IGFBP-1, did not differ between the
study groups (Table 3 and Fig. 1). Basal serum IGFBP-1 and
insulin were in negative relation to each other (r 5 20.49, P 5
0.001).
Plasma endothelin-1 concentrations, as well as the urinary
output of prostacyclin and thromboxane A2 metabolites,
were similar in the two groups and showed no correlations
with steroid hormones (Table 3).
Discussion
A 2- to 3-fold excess risk of hypertension and ischemic
heart disease in women with prior preeclampsia (1–7) implies that women with preeclampsia may have inherent endocrine or metabolic abnormalities expressed during preeclampsia; such abnormalities may persist, predisposing
these women to vascular disease. We have previously shown
that women with preeclampsia show changes similar to
those in metabolic syndrome (9), and that hyperinsulinemia
persists up to 17 yr after preeclamptic first pregnancy (11).
We now show that these women are characterized by mild
hyperandrogenism, as seen from increased free testosterone,
free androgen index, and free testosterone/estradiol ratio.
These changes could result from increased ovarian testos-
TABLE 1. Characteristics of study population
Characteristic
Women with prior
preeclampsia/eclampsia
(n 5 22)
Control women
(n 5 22)
P value
Age (yr)
Years since delivery
Subsequent full-term pregnancies/women
BMI (kg/ml2)
Smoking
Systolic blood pressure (mm Hg)
Diastolic blood pressure (mm Hg)
41.8 6 0.9
16.9 6 0.1
21/15
23.3 6 0.8
3
127 6 3
82 6 2
41.8 6 0.9
17.0 6 0.1
27/19
21.9 6 0.4
5
116 6 2
75 6 1
NS
NS
NS
NS
NS
P 5 0.004
P 5 0.01
Values are mean 6
SE;
NS, not significant.
TABLE 2. Characteristics of assays used
Factor
Androstenedione
Dehydroepiandrosterone
sulfate
Estradiol
Testosterone
FSH
LH
IGFBP 1
Sex-hormone binding
globulin
T4
Principle of assay
Source of reagents
RIA
Antiserum from ICN Biomedicals (Costa
Mesa, CA); tracer from Amersham
International (Little Chalfont, UK)
RIA
Coat-A-Count, Diagnostic Products Corp.;
CA
RIA
Sorin Biomedica Diagnostics, Saluggia,
Italy
RIA
Spectria Testosterone 125I-Coated Tube
RIA, Orion Diagnostica, Espoo, Finland
Fluoroimmunoassay DELFIA, Wallac, Turku, Finland
Fluoroimmunoassay DELFIA, Wallac, Turku, Finland
Immunoenzymometric Medix Biochemica, Kauniainen, Finland
Fluoroimmunoassay DELFIA, Wallac, Turku, Finland
Endothelin-1
Competitive
immunoassay
RIA
2,3 Dinor-6-keto-PG F1a
HPLC, RIA
2,3-Dinor thromboxane B2
HPLC, RIA
CV, Coefficient of variation.
ACSy T4, Chiron Diagnostics (E.
Walpole, MA)
Own antibody, tracer from Amersham
International
Own antibody, tracer from Amersham
International
Own antibody, tracer from Amersham
International
Intraassay CV
(percent)
Interassay CV
(percent)
6.0
8.5
7.4
10.2
5.7
5.7
5.3
8.9
2.9
1.8
2.9
1.5
3.9
3.4
6.8
4.0
3.2
6.4
5.7
Reference
(18)
(21)
(22)
,8
10.4 –14.1
(23)
,8
10.4 –14.1
(23)
346
JCE & M • 1998
Vol 83 • No 2
LAIVUORI ET AL.
TABLE 3. Concentrations of hormones, related compounds, and PG metabolites in women with and without prior preeclampsia/eclampsia
Androstenedione (nmol/L)
Dehydroepiandrosterone sulfate (mmol/L)
Estradiol (nmol/L)
Free androgen Indexa
Free testosterone (pmol/L)
Free testosterone/estradiol ratio
Total testosterone (nmol/L)
FSH (IU/L)
LH (IU/L)
LH/FSH ratio
IGFBP-1 (mg/L)
Sex-hormone binding globulin (nmol/L)
T4 (nmol/L)
Endothelin-1 (pmol/L)
2-3-Dinor-6-keto-PGF1a (ng/mmol creatinine)
2-3-Dinor thromboxane B2 (ng/mmol
creatinine)
Women with prior
preeclampsia
(n 5 21)
Control women
(n 5 20)
P value
6.7 6 0.6
4.9 6 0.4
0.35 6 0.05
3.2 6 0.5
20.6 6 2.2
0.089 6 0.017
1.7 6 0.1
8.0 6 2.6
8.1 6 1.2
1.3 6 0.1
5.4 6 0.6
70.1 6 6.5
97.0 6 3.5
0.73 6 0.03
20.6 6 2.6
25.9 6 3.0
5.8 6 0.4
4.1 6 0.4
0.44 6 0.06
1.9 6 0.2
15.0 6 1.3
0.046 6 0.006
1.4 6 0.1
5.9 6 1.4
8.5 6 1.7
1.6 6 0.2
6.1 6 0.7
84.2 6 6.7
94.6 6 2.8
0.73 6 0.04
21.3 6 1.8
27.4 6 2.9
NS
NS
NS
0.04
0.03
0.02
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Serum testosterone concentration (nmol/L) 3 100 divided by serum SHBG concentration (nmol/L).
Values are mean 6 SE; NS, not significant.
a
FIG. 1. IGFBP-1 levels in serum before and during oral glucose tolerance test (75 mg) in women with and without prior preeclampsia/
eclampsia (mean 6 SE). Differences between groups were not statistically significant.
terone production or decreased circulating SHBG levels, or
both, but our data do not allow us to deduce the initial cause
of these changes. Regardless of the cause of testosterone
changes and of the presence of normal androstenedione levels, we feel that our patients had slight ovarian hyperandrogenism. To the best of our knowledge, this is the first evidence to suggest an association between a history of
preeclampsia and ovarian hyperandrogenism.
At the moment we do not know which of the two major
abnormalities, hyperinsulinemia or high levels of testosterone, is the primary change. Insulin stimulates the production
of testosterone by ovarian tissue in vitro (24), which suggests
that hyperinsulinemia could be the primary change that triggered the increased release of testosterone. However, hyperinsulinemia should also stimulate the production of ad-
renal androgens (25), but this was not seen in our patients.
On the other hand, androgens are known to decrease both
hepatic removal of insulin and peripheral sensitivity to insulin (26), which suggests that hyperandrogenism could lead
to hyperinsulinemia. Similar coexistence of hyperinsulinemia and hyperandrogenism is present in PCO (12), and
these patients appear to be at increased risk of preeclampsia
(13, 14). This suggests that hyperinsulinemia and hyperandrogenism could precede the onset of preeclampsia. Although our patients had significantly elevated levels of free
testosterone and free androgen index, they had no clinical
signs of hyperandrogenism, menstruated normally, and had
normal BMI, IGFBP-1, SHBG and LH/FSH ratio. Thus it is
unlikely that our patients suffered from classic PCO, although their blood pressure was higher than that in the
controls, a feature typical of PCO patients (27). It was unfortunate that our study design did not include the use of
ultrasound to assess presence or absence of multiple ovarian
follicular cysts that could have been of help in the diagnosis
of PCO (28).
Prostacyclin and thromboxane A2 are important in pregnancy physiology and in preeclampsia (29), in which
endothelin-1 production can also be disturbed (30 –32). We
present the first long-term follow-up data on prostacyclin,
thromboxane A2, and endothelin-1 in women who have had
a preeclamptic first pregnancy. These data show that the
prostacyclin deficiency and/or thromboxane A2 or endothelin-1 dominance characterizing preeclamptic pregnancy (29 –
32) had vanished within 17 yr after pregnancy. That these
endothelial factors contributed to increased vascular morbidity in these subjects is thus very unlikely (1–7).
Combining our present data with the previous data (11),
we can state that women with a prior preeclamptic pregnancy are characterized by hyperinsulinemia and mild hyperandrogenism for up to 17 yr after delivery. Epidemiological observational studies have linked hyperinsulinemia
to increased risk of occlusive vascular disorders in men (33,
34). There is also abundant evidence that women with androgen excess are at increased risk of cardiovascular disease,
CIRCULATING TESTOSTERONE AFTER PREECLAMPSIA
although we do not have any clear-cut threshold values for
definitively vasotoxic levels of androgens in women (35, 36).
Nevertheless, our present data on hyperandrogenism in
women with prior preeclamptic pregnancy can provide one
explanation as to why these women have a 2- to 3-fold excess
of cardiovascular morbidity (1–7).
References
1. Chesley LC. 1976 Remote prognosis after eclampsia. In: Lindheimer MD, Katz
AL, Zuspan FP, eds. Hypertension in pregnancy. New York: John Wiley &
Sons; 31– 40.
2. Mann JI, Doll R, Thorogood M, Vessey MP, Waters WE. 1976 Risk factor for
myocardial infarction in young women. Br J Prev Soc Med. 30:94 –100.
3. Sibai BM, E-Nazer A, Gonzales-Ruiz A. 1986 Severe preeclampsia-eclampsia
in young primigravid women: subsequent pregnancy outcome and remote
prognosis. Am J Obstet Gynecol. 155:1011–1016.
4. Croft P, Hannaford PC. 1989 Risk factors for acute myocardial infarction in
women: evidence from the Royal College of General Practioners’ oral contraceptive study. Br Med J. 298:165–168.
5. Sibai BM, Sarinoglu C, Merecr BM. 1992 Eclampsia VII. Pregnancy outcome
after eclampsia and long-term prognosis. Am J Obstet Gynecol. 166:1757–1763.
6. Jónsdóttir LS, Arngrimsson R, Geirsson RT, Sigvaldason H, Sigfússon N.
1995 Death rates from ischemic heart disease in women with a history of
hypertension in pregnancy. Acta Obstet Gynecol Scand. 74:772–776.
7. Hannaford P, Ferry S, Hirsch S. 1997 Cardiovascular sequelae of toxemia of
pregnancy. Heart. 77:154 –158.
8. Dekker GA, van Geijn HP. 1994 Pathophysiology of preeclampsia and fetal
outcome. In: van Geijn HP and Copray FJA, eds. A critical appraisal of fetal
surveillance. Amsterdam: Elsevier Science B.V.; 79 –94.
9. Kaaja R, Tikkanen MJ, Viinikka L, Ylikorkala O. 1995 Serum lipoproteins,
insulin, and urinary prostanoid metabolites in normal and hypertensive pregnant women. Obstet Gynecol. 85:353–356.
10. Hubel CA, McLaughlin MK, Evans RW, Hauth BA, Sims CJ, Roberts JM.
1996 Fasting serum triglycerides, free fatty acids, and malondialdehyde are
increased in preeclampsia, are positively correlated, and decrease within 48
hours post partum. Am J Obstet Gynecol. 174:975–982.
11. Laivuori H, Tikkanen MJ, Ylikorkala O. 1996 Hyperinsulinemia 17 years after
preeclamptic first pregnancy. J Clin Endocrinol Metab. 81:2908 –2911.
12. Homburg R. 1996 Polycystic ovary syndrome—from gynaecological curiosity
to multisystem endocrinopathy. Hum Reprod. 11:29 –39.
13. Diamant YZ, Rimon E, Evron S. 1982 High incidence of preeclamptic toxemia
in patients with polycystic ovarian disease. Eur J Obstet Gynec Reprod Biol.
14:199 –204.
14. Gjønnæs H. 1989 The course and outcome of pregnancy after ovarian electrocautery in women with polycystic ovarian syndrome: the influence of bodyweight. Br J Obstet Gynaecol. 96:714 –719.
15. Dahlgren E, Janson PO, Johansson S, Lapidus L, Oden A. 1992 Polycystic
ovary syndrome and risk for myocardial infarction. Evaluated from a risk
factor model based on a prospective population study of women. Acta Obstet
Gynecol Scand. 71:599 – 604.
16. National High Blood Pressure Education Program Working Group. 1990
Report on High Blood Pressure in Pregnancy. Am J Obstet Gynecol.
163:1689 –1712.
347
17. Nilsson CG, Lähteenmäki PLA, Luukkainen T. 1984 Ovarian function in
amenorrheic and menstruating users of a levonorgestrel-releasing intrauterine
device. Fertil Steril. 41:52–55.
18. Schnakenburg v. K, Bidlingmaier F, Knorr D. 1980 17-Hydroxyprogesterone,
androstenedione and testosterone in normal children and in prepubertal patients with congenital adrenal hyperplasia. Eur J Pediatr. 133:259 –267.
19. Anderson DC, Thorner MO, Fisher RA, Woodham JP, Goble HL, Besser GM.
1975 Effect of hormonal treatment on plasma unbound androgen levels in
hirsute women. Acta Endocrinol Suppl (Copenh) 199:224.
20. Carter GD, Holland SM, Alaghband-Zadeh J, Rayman G, Dorrington-Ward
P, Wise PH. 1983 Investigation of hirsutism: testosterone is not enough. Ann
Clin Biochem. 20:262–263.
21. Rutanen E-M, Kärkkäinen T, Stenman U-H, Yki-Järvinen H. 1993 Aging is
associated with decreased suppression of insulin-like growth factor binding
protein-1 by insulin. J Clin Endocrinol Metab. 77:1152–1155.
22. Ylikorkala O, Orpana A, Puolakka J, Pyörälä T, Viinikka L. 1995 Postmenopausal hormonal replacement decreases plasma levels of endothelin-1. J Clin
Endocrinol Metab. 80:3384 –3386.
23. Tulppala M, Viinikka L, Ylikorkala O. 1991 Thromboxane dominance and
prostacyclin deficiency in habitual abortion. Lancet. 337:879 – 881.
24. Barbieri RL, Makris A, Randall RW, Dianiels G, Kistner RW, Ryan KJ. 1986
Insulin stimulates androgen accumulation in incubations of ovarian stroma
obtained from women with hyperandrogenism. J Clin Endocrinol Metab.
62:904 –910.
25. Martikainen H, Salmela P, Nuojua-Huttunen S, et al. 1996 Adrenal steroidogenesis is related to insulin in hyperandrogenic women. Fertil Steril.
66:564 –570.
26. Peiris AN, Mueller RA, Struve MF, Smith GA, Kissebah AH. 1987 Relationship of androgenic activity to splanchnic insulin metabolism and glucose
utilization in premenopausal women. J Clin Endocrinol Metab. 64:162–169.
27. Holte J, Gennarelli G, Berne C, Bergh T, Lithell H. 1996 Elevated ambulatory
day-time blood pressure in women with polycystic ovary syndrome: a sign of
a pre-hypertensive state? Hum Reprod. 11:23–28.
28. Adams J, Polson DW, Franks S. 1986 Prevalence of polycystic ovaries in
women with anovulation and idiopathic hirsutism. Br Med J. 293:355–359.
29. Ylikorkala 0, Viinikka L. 1992 The role of prostaglandins in obstetrical disorders. Baillierés Clin Obstet Gynaecol. 6:809 – 827.
30. Taylor R, Varma M, Teng NNH, Roberts J. 1990 Women with preeclampsia
have higher plasma endothelin levels than women with normal pregnancies.
J Clin Endocrinol Metab. 71:1675–1677.
31. Florijn KW, Derkx FHM, Visser W, et al. 1991 Elevated plasma levels of
endothelin in pre-eclampsia. J Hypertens. 9[Suppl 6]:S166 –S167.
32. Friedman SA, Lubarsky SL, Ahokas RA, Nova A, Sibai BM. 1995 Preeclampsia and related disorders. Clinical aspects and relevance of endothelin and
nitric oxide. Clin Perinatol. 22:343–355.
33. Pyörälä K, Savolainen E, Kaukola S, Haapakoski J. 1985 Plasma insulin as
coronary heart disease risk factor: relationship to other risk factors and predictive value during 9 1/2-year follow-up of the Helsinki Policemen Study
population. Acta Med Scand Suppl. 701:38 –52.
34. Fontbonne A, Charles MA, Thibult N, et al. 1991 Hyperinsulinemia as a
predictor of coronary heart disease mortality in a healthy population: the Paris
Prospective Study, 15-year follow-up. Diabetologia. 34:356 –361.
35. Wild RA. 1995 Obesity, lipids, cardiovascular risk, and androgen excess. Am J
Med. 98[Suppl 1A]:27S–32S.
36. Björntorp P. 1996 The android woman—a risky condition. J Int Med. 239:
105–110.