**2.1.1 Subjects and methods**

The subjects were type 2 diabetic patients seen regularly at the outpatient clinic of Toho University Hospital. We chose 130 patients, whose blood glucose control had been poor (more than 10% in HbA1c). Their medication was managed by diet only or with sulfonylurea, and patients under insulin therapy were excluded. The patient group consisted of 74 men and 56 women between the ages of 40-69yr. Age-matched normal subjects served as the control group. Informed consent was obtained from each subject before the study.

Blood samples were obtained from patients with type 2 diabetes mellitus and normal subjects between 9 and 10 a.m. after an overnight fast. From patients with type 2 diabetes mellitus, blood samples were obtained before and 6 months after the treatment. Serum levels of DHEA, DHEA-S and immunoreactive insulin (IRI), fasting plasma glucose (FPG) and HbA1c were measured. Steroid hormones were determined by the previously reported

DHEA and Impaired Glucose Tolerance Clinical and Basic Study 111

100 120 **male**

\*

**female**

DHEA-S (g/dL)

Fig. 1. Serum DHEA and DHEA-S levels in male and female type 2 diabetic patients before (stippled bars) and after (hatched bars) treatment and in age-matched normal subjects

In this study we demonstrated that serum DHEA and DHEA-S levels decreased markedly with poor control of type 2 DM and increased to age-matched normal values with the improvement of FPG and HbA1c after 6 months' treatment with diet and/or sulfonylurea. Barrett-Connor showed that DHEA and DHEA-S levels were also low in patients with noninsulin-dependent diabetes mellitus (Barrett-Connor, 1992), but she did not measure the changes of these steroid hormones after treatment. Markedly reduced levels of DHEA and DHEA-S in type 2 DM with poor therapeutic control with slightly increased plasma IRI are consistent with an association between DHEA synthesis and/or metabolism and insulin. Nestler et al. showed that insulin reduces serum DHEA and DHEA-S by increasing the metabolic clearance rate of DHEA in men or inhibiting their productin (Nestler,1992). The metablic clearance rate of DHEA is reported to be increased two- to fivefold in obesity and insulin-resistant, hyperinsulinemic state (Nestler,1995). The infusion of a high dose of insulin reduces serum DHEA levels suggesting the involvement of the inhibition of adrenal 17,20lyase activity. The administration of metformin which inhibits hepatic glucose production and enhances peripheral tissue sensitivity to insulin, to healthy normal weight men and to obese men with hypertension but without diabetes mellitus decreased serum insulin levels and increased serum DHEA-S levels in obese men with hypertension and in healthy controls (Nestler,1995). However, Yamauchi et al. reported that serum DHEA and DHEA-S are low even in patients with impaired glucose tolerance and low insulin response (Yamauchi,1996), and therefore the decrease in serum DHEA levels may not exclusively arise from the hyperinsulinemic state. Hyperglycemia may reduce 17,20-lyase activity and consequently serum DHEA may decrease. The improvement of plasma glucose control

40 yrs. 50 yrs. 60 yrs.

40 yrs. 50 yrs. 60 yrs.

\* \*

\* \* \*

DHEA-S (g/dL)

5 6 **male**

5 6

(opena bars).

**2.1.3 Discussion** 

\*

**female**

DHEA (ng/mL)

40 yrs. 50 yrs. 60 yrs.

40 yrs. 50 yrs. 60 yrs.

parallels the recovery of 17,20-lyase activity.

\* \*

\**P*< 0.05 compared with values after treatment and with normal values

\* \* \*

DHEA (ng/mL)

HPLC/RIA method(Ueshiba et al.,1991) except DHEA-S which was measured using RIA kit(Mitsubisi Chemical Co., Tokyo, Japan), FPG by glucose oxidase method, HbA1c by HPLC, IRI by commercial kits (Daiichi, Tokyo, Japan). Data are showed as mean ± SD. Variables were compared by Bonferroni's analysis and p-values less than 0.05 were considered to indicate statistical significance.
