| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles: Hormones and Reproductive Health |
Departments of Obstetrics and Gynecology (R.M.K., J.J., L.C.M.-P., J.S.T.) and Clinical Chemistry (A.R.), University Hospital of Oulu, 90220 Oulu; and Department of Medicine, University Hospital of Kuopio (I.V.), 70210 Kuopio, Finland
Address all correspondence and requests for reprints to: Dr. J. S. Tapanainen, Department of Obstetrics and Gynecology, University Hospital of Oulu, Kajaanintie 52 A, FIN-90220 Oulu, Finland. E-mail: juha.tapanainen{at}oulu.fi
Abstract
The prevalence of polycystic ovaries (PCO) and clinical, endocrine, and metabolic features were investigated in women with previous gestational diabetes (GDM). Thirty-three women with a history of GDM and 48 controls were studied. Glucose and insulin secretion capacity was evaluated by means of the oral glucose tolerance test (OGTT), and insulin action was determined by means of a euglycemic insulin clamp. Compared with control women, women with previous GDM more often had significantly abnormal OGTT, a higher prevalence of PCO (39.4% vs. 16.7%; P = 0.03), higher serum concentrations of cortisol, dehydroepiandrosterone, and dehydroepiandrosterone sulfate and a greater area under the glucose curve.
Women with previous GDM showed a lowered early phase insulin response to glucose and impaired insulin sensitivity, which was accounted for mainly by decreased glucose nonoxidation. They also demonstrated a significantly lower fasting serum C peptide/insulin ratio than the controls, indicating that women with previous GDM have impaired hepatic insulin extraction, which tended to be more marked among women with PCO. This may explain why women with PCO and previous GDM were significantly more hyperinsulinemic than women with normal ovaries.
In conclusion, our data demonstrate that women with previous GDM often have PCO and abnormal OGTT. They are insulin resistant as a result of lowered glucose nonoxidation and show inappropriately low insulin responses to glucose, reflecting impaired ß-cell function. They also have higher adrenal androgen secretion, which may be associated with abdominal obesity.
GESTATIONAL DIABETES mellitus (GDM) is defined as glucose intolerance that has its onset or is first recognized during pregnancy (1, 2). Gestational diabetes mellitus occurs in 2–5% of pregnancies (3). Some studies have indicated that insulin resistance is the primary cause of impaired glucose metabolism in GDM (4, 5, 6), whereas others have emphasized the role of defective insulin secretion (4, 6, 7, 8, 9, 10), and some have suggested that GDM is due to a combination of insulin resistance and diminished insulin secretion (5, 11, 12, 13, 14, 15). Women with GDM have a significantly increased risk of developing primarily type 2 diabetes later in life (13, 16, 17), and it has even been suggested that GDM and type 2 diabetes are the same disorder (18). Furthermore, GDM is thought to represent an early manifestation of the metabolic syndrome (or syndrome X), which is a cluster of abnormalities where the combination of insulin resistance and compensatory hyperinsulinemia predisposes individuals to develop high plasma triglyceride (Trigl) and low high density lipoprotein (HDL) cholesterol concentrations, high blood pressure, and coronary heart disease (19).
Women with polycystic ovary syndrome (PCOS) appear to share some of the same features as women with GDM. A large number of obese PCOS patients and a significant number of lean ones are also hyperinsulinemic (20, 21, 22). Some PCOS women with abnormal insulin secretion at the pregestational stage have been shown to develop impaired gestational glucose tolerance or GDM during pregnancy (23). When insulin sensitivity is assessed by the euglycemic hyperinsulinemic clamp technique, women with PCOS have profound peripheral insulin resistance of a magnitude similar to that seen in individuals with type 2 diabetes (24, 25). Similarly, there is evidence that women with polycystic ovaries (PCO) only, without clinical symptoms of PCOS, may be prone to develop various features of the insulin resistance syndrome (6, 26, 27).
Although there are data indicating the association between GDM and PCOS, the influence of PCOS or PCO on insulin sensitivity in subjects already known to be at risk for disturbed carbohydrate metabolism is not clear. The major objective of the present study was to evaluate ovarian appearance, insulin sensitivity, and insulin secretion in women with previous GDM with or without PCO. In addition, special attention was paid to ovarian and adrenal steroidogenesis and lipid metabolism, which are known to be altered in PCOS.
Subjects and Methods
Subjects
A questionnaire was sent to 180 women who delivered at Oulu
University Hospital from 1990–1993 and who had had GDM for the first
time during pregnancy. One hundred and eighty age (±2 yr)-, parity
(nulliparous, 1–3 deliveries, or >3 deliveries)-, and delivery
date-matched women with uncomplicated pregnancy served as controls. All
of the women were 42 yr old or less at the time of this study (Table 1
). Gestational diabetes mellitus had
been verified by means of an oral glucose tolerance test (OGTT; 75
g) during the second trimester, and the indications for performing the
OGTT were glucosuria, previously diagnosed GDM, obesity, or a large for
gestational age fetus. The diagnosis of GDM was based on blood glucose
values (0 min, 
5.5; 60 min, 11; or 120 min, 9.0 mmol/L). Finally,
33 women (31–42 yr) with previous GDM (18 treated by diet and 15
treated with insulin) and 48 control women (27–42 yr) volunteered and
gave informed consent. The final selection of the women is shown in the
flow chart (Fig. 1). Because of
our strict inclusion criteria the number of case-control pairs was too
low, and the subjects were simply divided into 2 groups.
|
|
All of the women were examined on cycle days 1–6. Transvaginal
ultrasonography was performed on the same day as the OGTT by a single
investigator (R.K.), who did not know beforehand whether the subject
belonged to the control or the GDM group. Polycystic ovaries were
defined when there were 10 or more follicles 2–8 mm in diameter in one
plane in one ovary in association with increased ovarian stroma,
according to the criteria described by Adams et al.
(28). The criteria for PCOS were defined by Homburg
(29), i.e. PCO shown by transvaginal
ultrasonography and at least one of the following symptoms:
oligo/amenorrhea, clinical manifestations of hyperandrogenism such as
hirsutism scored according to Ferriman and Gallwey, acne, and/or
elevated serum testosterone concentration (2.7 nmol/L). Each ovary
was measured in three dimensions, ovarian volume was calculated using
the formula for an ellipsoid (volume = 0.5233 x dimension
1 x dimension 2 x dimension 3), and the results are
expressed as a sum of ovarian volumes divided by 2. Women with ovarian
cysts more than 30 mm in diameter were excluded with regard to
calculation of ovarian volume (three women, two with normal ovaries and
one with PCO). Women with a body mass index (BMI) of 27
kg/m2 or more were considered obese. Hirsutism
was graded using the Ferriman-Gallwey scoring system (30),
and the woman was considered hirsute if the score was 7 or more.
The subjects filled out a questionnaire concerning their gynecological history and family history of diabetes. Ovarian ultrasonography had not been performed before this study, and based on the inquiry none of the subjects had been diagnosed with PCOS previously. Menstrual cycles were considered irregular if the variation in menstrual intervals was more than 7 days from one period to another, and oligomenorrheic if the intermenstrual interval was 36 days or more. None of our patients had amenorrhea (intermenstrual interval >6 months). The study protocol was approved by the ethics committee of the University of Oulu.
OGTT
After an overnight fast of 10 h, all subjects underwent an OGTT (a load of 75 g glucose in 300 mL water). The glucose was ingested within 3 min under the supervision of the investigator. Venous blood samples for assay of blood glucose and serum insulin were collected before and 15, 30, 60, and 120 min after glucose ingestion. The glycemic response to the OGTT was defined according to the new American Diabetes Association criteria in 1997 (31).
Early phase insulin secretion (insulinogenic index) was calculated as the ratio of the increment in serum insulin 30 min after the oral glucose load to the blood glucose concentration 30 min after the glucose load [(30 min insulin - fasting insulin)/30 min glucose] (32). The insulinogenic index has previously been shown to correlate strongly with the first phase insulin response after an iv glucose tolerance test (r = 0.88) (33). Early phase C peptide secretion was calculated similarly [(30 min C peptide - fasting C peptide)/30 min glucose]. The formula [(insulin 30 min - insulin 0 min)/(glucose 30 min - glucose 0 min)] for insulinogenic index is more commonly used (34), but because the denominator became negative in some subjects when early phase insulin secretion and early phase C peptide secretion were calculated, this alternative formula was used (32).
The incremental insulin (AUCins) and glucose (AUCgluc) areas under the curve were calculated by the trapezoidal method. The fasting serum C peptide/fasting serum insulin ratio was calculated as an index of hepatic insulin extraction in the fasting state (35).
Euglycemic clamp
The degree of insulin resistance was evaluated using the euglycemic hyperinsulinemic clamp technique (36) in 15 women with previous GDM (7 women treated by diet and 8 treated with insulin) and 23 controls who volunteered for the clamp study. A priming dose of insulin infusion (100 IU/ml; Actrapid, Novo Nordisk, Gentofte, Denmark) was administered during the initial 10 min to acutely raise plasma insulin to the desired level, where it was maintained by continuous insulin infusion at a rate of 80 mU/m2 body surface area per minute. Blood glucose was clamped at 5.0 mmol/L for the next 180 min by adjusting the rate of 20% glucose infusion according to blood glucose measurements performed every 5 min using a photometric assay (HemoCue AB, Angelholm, Sweden). The M-value (amount of glucose infused, i.e. whole body glucose disposal, micromoles per kg/min) was calculated as the mean value for each 20-min interval during the last 60 min of the clamp. The insulin sensitivity index (M/I) was calculated by dividing the M-value by the mean steady state insulin concentration (picomoles per L) during the last 60 min of the clamp. Blood samples for assay of serum lactate, insulin, and free fatty acids (FFA) were drawn at 0, 120, 140, 160, and 180 min.
Calorimetry
Indirect calorimetry was performed with a computerized flow-through canopy gas analyzer system (DELTATRAC, TM Datex, Helsinki, Finland) in connection with the euglycemic clamp, as previously described (37). This device has a precision of 2.5% for O2 consumption and 1.0% for CO2 production. On the day of the experiment, gas exchange (O2 consumption and CO2 production) was measured after a 12-h fast before and during the last 30 min of the clamp. The values obtained during the first 10 min of both periods were discarded, and the mean values of the remaining 20 min of data were used for calculation. Protein, glucose, and lipid oxidation were calculated according to the method described by Ferrannini (38). Protein oxidation was calculated on the basis of the urinary nonprotein nitrogen excretion rate (38). The fraction of carbohydrate nonoxidation during the euglycemic clamp was estimated by subtracting the carbohydrate oxidation rate (determined by indirect calorimetry) from the glucose infusion rate (determined by the euglycemic clamp), and both values were adjusted for the prevailing insulin levels during the clamp by dividing the values by the mean steady state insulin concentration during the last 60 min of the clamp (glucose oxidation index and nonoxidation index, respectively).
Assays
Blood samples for hormone analyses were obtained in a standardized manner between 0700–0800 h after resting for 30 min in a sitting position. The concentrations of sex hormone-binding globulin (SHBG), LH, and FSH were analyzed by fluoroimmunoassays (Wallac, Inc., Turku, Finland), and RIAs were used for dehydroepiandrosterone (DHEA), dehydroepiandrosterone sulfate (DHEAS), androstenedione (A), C peptide (Diagnostic Products Corporation, Los Angeles, CA), cortisol (Orion Diagnostica, Turku, Finland), insulin (Pharmacia, Uppsala, Sweden), and leptin (Linco Research, Inc., St. Charles, MO) determinations. Serum testosterone (T) was determined using an automated chemiluminescence system (ACS-180, Ciba-Corning, Inc., Medfield, MA). The free androgen index (FAI) was calculated according to the equation: (T x 100)/SHBG. Serum insulin-like growth factor-binding protein-1 (IGFBP-1) concentrations were determined by immunoenzymometric assay (Medix Biochemica, Kauniainen, Finland). Serum total cholesterol, Trigl, HDL cholesterol, FFA, blood glucose, and lactate concentrations were determined by standard methods. The serum low density lipoprotein cholesterol (S-LDL) concentrations were calculated by the Friedewald formula if the serum Trigl (S-Trigl) level was less than 4.0 mmol/L; if the S-Trigl level was more than 4.0 mmol/L, it was precipitated by heparin in isoelectric point (LDL-cholesterol, Merck & Co., Darmstadt, Germany).
The intra- and interassays coefficients of variation were 1.3% and 5.1% for SHBG, respectively; 4.9% and 6.5% for LH; 3.8% and 4.3% for FSH; 6.5% and 7.9% for DHEA; 5.3% and 7.0% for DHEAS; 5.0% and 8.6% for A; 5.3% and 7.2% for C peptide; 4.0% and 4.3% for cortisol; 5.3% and 7.6% for insulin; 5.0% and 6.0% for leptin; 4.0% and 5.6% for T; 3.4% and 7.4% for IGFBP-1; 1.5% and 2.3% for blood glucose; 2.2% and 3.8% for lactate; and 3.8% and 5.5% for FFA.
Statistical analysis
All calculations were performed with the SPSS for Windows
program (SPSS, Inc., Chicago, IL). Clinical data are shown
as the mean ± SD or as the number and percentage of
subjects. Hormonal and metabolic data are shown as the arithmetic mean
[95% confidence intervals (CI) in parentheses], marked
with an asterisk, if the distribution was normal, or as the
geometric mean if the distribution was skewed. Student’s two-tailed
t test was used for comparisons of variables, which were
normally distributed either without or after log transformation, and
the Mann-Whitney U test was used for variables that still showed a
skewed distribution after log transformation. Analysis of covariance
was used to correct for confounding variables [age, BMI, or waist to
hip ratio (WHR)] when needed. If the degree of significance changed
after these corrections, the P values are shown. The
2 test was used for comparisons between
category variables. Analysis of correlation between parameters was
performed using Pearson’s bivariate correlation coefficient.
Results
Women with previous GDM vs. control subjects
Clinical and biochemical characteristics. Table 1
shows the
clinical and metabolic characteristics of the study groups. Women with
previous GDM were younger than the control women (P =
0.01). The BMI tended to be higher (P = 0.1), and the
WHR was significantly higher (P = 0.002) in the GDM
group than in the control group. Fifteen of 33 women (45.5%) in the
GDM group and none in the control group have had a pregnancy with GDM
after the index pregnancy. A family history of type 2 diabetes mellitus
among first degree relatives was more frequent in the GDM group than in
the control group (P < 0.001).
Women with previous GDM had PCO significantly more often than women in
the control group (13 of 33, 39.4% vs. 8 of 48, 16.7%;
P = 0.03; Table 1). If the subjects with PCO and
irregular cycles (n = 9) were excluded, the frequency of PCO and
regular cycles was 21.2% (7 of 33) in the GDM group and 10.4% (5 of
48) in control subjects (P = 0.1). The majority of the
subjects (7 of 9) were classified as having irregular menstruation
because of the large variation in menstrual intervals (21–35 days).
All of these women had become pregnant without any treatment, and none
of them was hyperandrogenic. Thus, it is very likely that all of these
women had ovulatory cycles despite menstrual irregularity, and
therefore they could not be classified as PCOS. Only 2 subjects (of 9)
classified as having irregular menstruation had oligomenorrhea
(menstrual cycle >36 days), hyperandrogenism, and most likely
anovulation, and they fulfilled the criteria of PCOS. If these women
were excluded, the frequency of PCO was 35.5% (11 of 31) in GDM group
and 16.7% (8 of 48) in control women (P = 0.05).
Exclusion of these 2 women with oligomenorrhea did not affect the
endocrinological and metabolic results, and therefore they were
included. There were no differences in parity [2.7 ± 1.0
(±SD) vs. 3.4 ± 2.5;
P = 0.5] or in the time since the index pregnancy
(5.4 ± 1.6 vs. 6.6 ± 2.7 yr; P =
0.3) between the GDM and the control group.
The GDM group had higher fasting blood glucose (P =
0.008) and fasting serum insulin levels (P = 0.003)
than the control group. The difference in fasting serum insulin levels
disappeared after correction for WHR (P = 0.1). Fasting
serum C peptide levels were comparable between the study groups. The
GDM group had lower hepatic insulin extraction than the control group
(P = 0.01). This difference persisted after correction
for age (P = 0.02), BMI (P = 0.01), and
WHR (P = 0.008; Table 1).
Glucose tolerance. Table 2 shows the results of OGTTs.
Abnormal glucose tolerance was more frequent in the women with previous
GDM (57.6%) than in the control group (12.5%; P <
0.001). Three subjects (two treated by diet and one treated with
insulin during the pregnancy) in the GDM group and one subject in the
control group had diabetes mellitus, and they were excluded from the
analysis of endocrinological and metabolic variables.
|
depicts the blood glucose
and serum insulin responses during the OGTT. The GDM group had
significantly higher fasting blood glucose as well as blood glucose
levels at 30 and 60 min after oral glucose than the control group.
Furthermore, the oral glucose response, expressed as an incremental
glucose area under the curve (AUCgluc), was
higher in the GDM group (829 mmol/L·h; 95% CI, 776–891) than
in the control group (673 mmol/L·h; 95% CI, 631–708;
P < 0.001). In addition, serum insulin levels at 60
and 120 min in the OGTT were significantly higher in the GDM group than
in the control group. The difference disappeared at 120 min after
correction for BMI (P = 0.2). The serum insulin
response, expressed as an incremental insulin area under the curve
(AUCins), did not differ significantly between
the groups [579 pmol/L·h (95% CI, 428–729) for the GDM group
vs. 433 pmol/L·h (95% CI, 339–527) for the
controls].
|
), and this difference persisted after
corrections for BMI and WHR. In addition, serum C peptide
concentrations at 30 min in the OGTT were significantly lower in the
GDM group than in the control group [1130 (95% CI, 911-1350)
vs. 1424 (95% CI, 1229–1618) pmol/L; P =
0.05] even after corrections for BMI and WHR.
|
]. This difference persisted after correction for BMI, but
disappeared after correction for WHR (P = 0.5). The
lower M/I in the GDM group was due to a lower glucose nonoxidation
index [0.040 (95% CI, 0.026–0.051) vs. 0.054 (95% CI,
0.049–0.061) µmol/kg BW·min/pmol/L; P = 0.05].
The glucose oxidation index during the clamp was similar in both groups
[0.02 (95% CI, 0.014–0.029) vs. 0.02 (95% CI,
0.018–0.024) µmol/kg BW·min/pmol/L; P = 0.7; Fig. 3]. Plasma lactate concentrations did not differ between the groups in
the fasting state [0.72 (95% CI, 0.56–0.87) vs. 0.75
(95% CI, 0.57–0.93) mmol/L] or during the euglycemic clamp [1.0
(95% CI, 0.88–1.11) vs. 1.1 (95% CI, 0.94–0.12)
mmol/L].
Endocrine parameters and lipids. The GDM group had
significantly higher serum cortisol, DHEA, DHEAS, and A
concentrations than the control group (Table 3). Serum cortisol (r = 0.18;
P = 0.03), DHEA (r = 0.27;
P = 0.02), and A (r = 0.30; P =
0.01) concentrations correlated positively with WHR, but not with BMI.
Furthermore, the GDM group had significantly lower serum HDL
concentrations than the control group [1.3 (95% CI, 1.2–1.4)
vs. 1.5 (95% CI, 1.4–1.6) mmol/L; P <
0.01]. There were no significant differences in total serum
cholesterol, S-LDL, or S-Trigl concentrations between the groups (data
not shown).
|
Clinical characteristics. During the pregnancy 18 women with GDM were treated by diet (GDM-diet) and 15 were treated with insulin (GDM-insulin). The women in the GDM-diet group were more obese than the women in the GDM-insulin group (BMI, 28.2 ± 5.9 vs. 24.2 ± 4.0 kg/m2; P < 0.05), and they also had a significantly higher WHR (0.84 ± 0.04 vs. 0.81 ± 0.05; P = 0.02).
Glucose tolerance. Women in the GDM-insulin group more often
had abnormal glucose tolerance in the OGTT (86.7%) than women in the
GDM-diet group (33.3%; P = 0.002) or the control group
(12.5%; P = 0.001). Furthermore, abnormal glucose
tolerance in the OGTT tended to be more frequent in the GDM-diet group
than in the control group (P = 0.09; Table 2).
There was no difference in the fasting glucose level between the GDM-insulin (5.4 mmol/L; 95% CI, 5.0–5.8) and GDM-diet (5.1 mmol/L; 95% CI, 4.8–5.4) groups. The GDM-insulin group had higher blood glucose levels at 60 min [8.9 (95% CI, 7.5–10.3) vs. 6.3 (95% CI, 5.7–7.0) mmol/L; P = 0.001] and at 120 min after oral glucose [6.9 (95% CI, 6.2–7.7) vs. 5.3 (95% CI, 4.7–5.9) mmol/L; P = 0.02] as well as a higher AUCgluc [935.7 (95% CI, 697.0–829.2) vs. 763.1 (95% CI, 846.3–1025.0) mmol/L·min; P = 0.001] than the GDM-diet group. Both groups had a significantly higher AUCgluc than the controls. The GDM-insulin group had higher serum fasting insulin levels [93.4 (95% CI, 67.6–119.2) vs. 62.0 (95% CI, 49.9–74.2) pmol/L; P = 0.02] and higher insulin levels at 120 min in the OGTT [545.4 (95% CI, 311.1–779.8) vs. 285.5 (95% CI, 204.5–366.4) pmol/L; P = 0.02] than the GDM-diet group, but these differences disappeared after adjusting for BMI and WHR.
Insulin secretion and insulin sensitivity. Early phase insulin and C peptide secretion and the insulin sensitivity index were comparable in the GDM-insulin and GDM-diet groups.
Control women: comparison between women with normal ovaries and those with PCO
Clinical characteristics. The clinical characteristics of the
subjects are shown in Table 4. The mean
age, BMI, and WHR were similar in control women with normal ovaries and
those with PCO. The women with PCO had irregular periods more often (3
of 8, 37.5% vs. 9 of 40, 22.5%; P = 0.04)
and, in accordance with the definition of PCO, a greater mean ovarian
volume [11.0 (95% CI, 8.7–13.3) vs. 9.7 (95% CI,
7.0–12.3) cm3; P = 0.04] than
the women with normal ovaries.
|
and
Table 4) or in insulin sensitivity (Table 4).
|
Women with previous GDM: comparison between women with normal ovaries and those with PCO
Clinical characteristics. The clinical characteristics of the
subjects are shown in Table 4. Women with PCO and previous GDM were
younger, had a higher mean ovarian volume [12.4 (95% CI, 9.4–15.6)
vs. 8.4 (95% CI, 6.1–9.0) cm3;
P < 0.004], and more often had irregular cycles (6 of
13, 46.2% vs. 2 of 20, 10%; P < 0.001)
than women with normal ovaries and previous GDM. However, only 2 women
in the PCO group fulfilled the criteria of PCOS.
Glucose tolerance. The frequency of abnormal OGTT results was
similar among the women with PCO and those with normal ovaries (53.8%
vs. 60%, respectively). There were no significant
differences in fasting blood glucose, serum insulin, serum C peptide,
or the AUCgluc between the groups (Table 4).
However, women with PCO had higher serum insulin levels at 30 min and
higher AUCins [776.9 (95% CI, 446.7–1107.1)
vs. 463.8 (95% CI, 320.1–607.6 pmol/L·h);
P = 0.03] in the OGTT than women with normal ovaries
(Fig. 5). The difference in
AUCins did not disappear after correction for BMI
(P = 0.03).
|
).
Insulin sensitivity. The insulin sensitivity index was
comparable in these groups (Table 4). Women with normal ovaries,
however, tended to have a lower glucose oxidation index [0.027 (95%
CI, 0.01–0.043) vs. 0.018 (95% CI, 0.008–0.027) µmol/kg
BW·min/pmol/L; P = 0.1] and a lower glucose
nonoxidation index [0.044 (95% CI, 0.01–0.077) vs. 0.037
(95% CI, 0.023–0.05) µmol/kg BW·min/pmol/L; P =
0.2] during the clamp than the women with PCO.
Endocrine parameters and lipids. Serum concentrations of leptin [PCO, 19.8 (95% CI, 12.9–30.9) ng/mL; normal ovaries, 13.3 (95% CI, 10.2–17.4) ng/mL; P = 0.09], T [1.4 (95% CI, 1.0–1.8) vs. 1.39 (95% CI, 1.1–1.6) nmol/L; P = 0.9] and SHBG [42.7 (95% CI, 30.9–58.9) nmol/L vs. 53.1 (95% CI, 42.7–66.1) nmol/L; P = 0.2] and FAI [3.9 (95% CI, 1.7–6.1) vs. 2.9 (95% CI, 1.9–3.9); P = 0.3] did not differ significantly between women with PCO and normal ovaries.
Discussion
The present study confirms that women with previous GDM have impaired insulin sensitivity (4, 5, 6), and in addition it demonstrates that this is mainly due to decreased glucose nonoxidation reflecting a defect in glycogen storage. A similar phenomenon has been shown in first degree relatives of type 2 diabetic patients (39, 40, 41, 42). Note, however, that the women with previous GDM were more obese, especially abdominally, than the control women, and further, correction for WHR (but not for BMI) abolished the difference in insulin sensitivity between these groups. This observation strongly suggests that the impaired insulin sensitivity in women with previous GDM is at least in part due to abdominal obesity. Indeed, it has been shown that especially abdominal fat tissue has high lipolytic activity that may lead to increased flux of FFA from the adipose tissue, with subsequent competition between FFA and glucose for oxidation in the muscle cells (43). This, in turn, may lead to impaired insulin sensitivity.
Besides impaired insulin sensitivity, the women with previous GDM had defective early phase insulin secretion, reflected in the serum C peptide response and a delayed serum insulin response to an oral glucose load. These observations reflect impaired ß-cell function in these women. They also had a low fasting serum C peptide/insulin ratio that indicates impaired hepatic insulin extraction. It is well known that hepatic insulin extraction is lowered in many conditions such as obesity (44), and thus it is possible that the women with previous GDM had lowered hepatic insulin extraction that was at least in part due to obesity.
Women with previous GDM significantly more often had an abnormal OGTT result (57.6%) than the controls (12.5%). This supports the concept that women with GDM are at risk of developing diabetes mellitus later in life (9, 13, 45). Women who had insulin treatment during pregnancy more often had abnormal glucose tolerance in the OGTT than diet-treated and control women. The use of insulin in pregnancy, which reflects the severity of GDM, has been shown to be strongly associated with subsequent diabetes (46). The high prevalence of abnormal OGTT results in diet-treated women (33.3%), which was similar to the incidence reported in earlier studies (29.7–34.3%) (10, 13), underlines the importance of regular postpartum assessment of glucose tolerance in these women.
Besides having metabolic defects typical of type 2 diabetes, the women with previous GDM also revealed a high prevalence of PCO, which was 2-fold higher than that reported in healthy premenopausal women (27, 47, 48). This confirms observations in two recent studies on women with GDM (6, 49). It is not known how many subjects in the present study had previously had PCO, but due to slow progression and chronic nature of PCO, it is very likely that a significant number of women had PCO before the index pregnancy. Interestingly, although women with PCO and previous GDM showed insulin sensitivity, insulin secretory capacity, and blood glucose responses in the OGTT comparable with those in the women with normal ovaries and previous GDM, they had more marked hyperinsulinemia, which was not explained by obesity. This may indicate that these women have a lowered MCR of insulin. Notably, they also had a low serum C peptide/insulin ratio in the fasting state, implying that the lowered MCR of insulin was at least partly due to its impaired hepatic extraction. The pathogenic mechanism behind this novel finding, however, remains largely unclear. Recently, it has been shown that serum FFA levels may alter hepatic insulin extraction (50), but in the present study the serum FFA levels were comparable among GDM women with and without PCO (data not shown). On the other hand, the women with PCO tended to have a higher lipid oxidation rate in the fasting state than the women with normal ovaries (data not shown), implying that hyperinsulinemia in women with PCO and previous GDM may, indeed, be associated with altered adipose tissue metabolism.
Clinical symptoms typical of PCOS were scarce in the women with PCO. However, these women had subtle clinical, endocrine and metabolic features characteristic of PCOS. Women with full-blown PCOS have been shown to have a higher risk of developing impaired glucose metabolism either before conception or during pregnancy (23, 51, 52). Our findings suggest that the ultrasonographic appearance of PCO may be a predictive factor as regards abnormal glucose tolerance during and after pregnancy.
Women in the GDM group had significantly higher serum adrenal steroid concentrations (cortisol, DHEA, DHEAS, and A) than control women, which may have been due to insulin resistance, hyperinsulinemia, and abdominal fat accumulation (i.e. increased WHR). Multiple endocrine perturbations have been associated with abdominal fat accumulation, including elevated serum cortisol and androgen levels in women and, in particular, hyperinsulinemia secondary to pronounced insulin resistance (53, 54, 55). It is possible that androgens are overproduced by the adrenals as a component of the consequences to the hypersensitive hypothalamo- pituitary-adrenal axis (56, 57, 58). With regard to the role of insulin in adrenal steroid secretion, obese hyperinsulinemic women with PCOS have been found to have excessive secretion of adrenal androgens (59). Furthermore, a positive relationship has been shown between serum insulin and DHEAS levels in healthy obese women (60). However, there are studies in which improvement of insulin resistance and hyperinsulinemia achieved by way of pharmacological therapy has not altered circulating DHEA or DHEAS levels in women (61). The possible stimulatory effect of insulin on adrenal steroidogenesis may be mediated via insulin receptors and/or the insulin growth factor (IGF) system. In our study the lower serum IGFBP-1 concentration in the GDM group reflects the higher free IGF-I concentrations that may contribute to adrenal steroid secretion. It has been shown that IGF-I enhances the steroidogenic response to ACTH and the expression and activity of cytochrome P450C17 in adult human adrenal cortical cells (62).
In conclusion, our data demonstrate that women with previous GDM often have abnormal glucose tolerance, they are insulin resistant, mainly as a result of lowered glucose nonoxidation in the peripheral tissues, and, further, they show inappropriately low insulin responses to glucose, which is due to impaired ß-cell function. The prevalence of PCO in women with previous GDM is twice as high as that reported for premenopausal women overall. This observation suggests that women with PCO, although not fulfilling the diagnosis of PCOS, are at higher risk of developing GDM and impaired glucose tolerance after pregnancy, and they should be checked regularly. Women with previous GDM also have higher adrenal androgen secretion than control women. The clinical importance of this observation and whether it is due to abdominal obesity, insulin resistance, and/or hyperinsulinemia remain to be studied.
Acknowledgments
We thank the nurses Mirja Ahvensalmi, Anni Jurvakainen, Pirjo Ylimartimo, and Ritva Vasala of the Research Unit of the Department of Obstetrics and Gynecology, and Anja Heikkinen, of the Clinical Chemistry Laboratory, Oulu University Hospital, for their expertise in the technical work and the care of the patients. We are grateful for statistical advice from Ms. Sirkka Pramila, who passed away during the study.
Footnotes
1 This work was supported by grants from the Sigrid Juselius
Foundation, the Academy of Finland, and the Finnish Association of
Obstetrics and Gynecology. 
Received July 13, 2000.
Revised January 30, 2001.
Revised February 26, 2001.
Accepted March 20, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  K. E. S. Salley, E. P. Wickham, K. I. Cheang, P. A. Essah, N. W. Karjane, and J. E. Nestler POSITION STATEMENT: Glucose Intolerance in Polycystic Ovary Syndrome A Position Statement of the Androgen Excess Society J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4546 - 4556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Lo, S. L. Feigenbaum, G. J. Escobar, J. Yang, Y. M. Crites, and A. Ferrara Increased Prevalence of Gestational Diabetes Mellitus Among Women With Diagnosed Polycystic Ovary Syndrome: A population-based study Diabetes Care, August 1, 2006; 29(8): 1915 - 1917. [Full Text] [PDF]
|
|
|
|
 |
|
 M.A. Checa, A. Requena, C. Salvador, R. Tur, J. Callejo, J.J. Espinos, F. Fabregues, J. Herrero, and (Reproductive Endocrinology Interest Group of the Insulin-sensitizing agents: use in pregnancy and as therapy in polycystic ovary syndrome Hum. Reprod. Update, July 1, 2005; 11(4): 375 - 390. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. O. Goodarzi, S. Erickson, S. C. Port, R. I. Jennrich, and S. G. Korenman {beta}-Cell Function: A Key Pathological Determinant in Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., January 1, 2005; 90(1): 310 - 315. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. De Leo, A. la Marca, and F. Petraglia Insulin-Lowering Agents in the Management of Polycystic Ovary Syndrome Endocr. Rev., October 1, 2003; 24(5): 633 - 667. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) and in women with previous GDM (•). *, P <
0.05; **, P < 0.01; ***, P <
0.001 (compared with controls).
)
and in women with previous GDM (
). The third panel
shows the insulin sensitivity index (glucose oxidation and
nonoxidation) assessed by hyperinsulinemic euglycemic clamp. *,
P < 0.05 (glucose nonoxidation only); **,
P < 0.01 (compared with controls).
