- Top of page
- Research Methods and Procedures
Objective: The aim of this study was to investigate the effects of combined hypocaloric diet and metformin on circulating testosterone and leptin levels in obese men with or without type 2 diabetes.
Research Methods and Procedures: Twenty obese men with type 2 diabetes (mean body mass index [BMI]: 35.5 ± 1.1 kg/m2) and 20 nondiabetic obese men were enrolled in the study. We measured serum follicle-stimulating hormone, luteinizing hormone (LH), total testosterone (TT), free testosterone (FT), sex-hormone-binding globulin (SHBG), dehydroepiandrosterone sulfate (DHEAS), and plasma leptin levels before and 3 months after metformin treatment. Both groups were placed on a hypocaloric diet and 850 mg of metformin taken orally twice daily for 3 months.
Results: Metformin and hypocaloric diets led to decreases in BMI and waist and hip circumferences in both groups. A significant decrease in TT levels in the diabetic group and FT levels in the control group was found, whereas follicle-stimulating hormone, LH, and DHEAS levels were not changed significantly. A significant increase in SHBG levels was observed in the control group but not in the patient group. Leptin levels also decreased after treatment in both groups. Decreased testosterone levels were not correlated to changes in waist and hip circumference, waist-to-hip ratio, BMI, and levels of fasting blood glucose, leptin, SHBG, or DHEAS in the diabetic group. However, a decrease in FT was correlated to changes in the levels of SHBG (r = −0.71, p = 0.001) and LH (r = 0.80, p = 0.001) but not to other parameters.
Discussion: We conclude that metformin treatment combined with a hypocaloric diet leads to reduced FT levels in obese nondiabetic men and to reduced TT levels in obese men with type 2 diabetes. Increased SHBG levels may account for the decrease in FT levels in the former group.
- Top of page
- Research Methods and Procedures
Although metformin has been widely used in the treatment of type 2 diabetes, the glucose-lowering mechanisms are still not clear. Some studies have shown that metformin exerts its glucose-lowering effect by suppressing hepatic glucose production (1) (2). Recent studies on type 2 diabetic subjects have demonstrated that the glucose-lowering effects of metformin may be attributable to decreased release of free fatty acids from adipose tissue (3) (4) (5). Recent evidence has shown that metformin increases glucose uptake and inhibits leptin secretion from cultured adipocytes (6) (7).
Previous reports have shown inhibition of food intake (8) and significant weight loss associated with metformin treatment in type 2 diabetes, compared with sulfonylureas or placebo (9) (10) (11). Recently, the U.K. Prospective Diabetes Study has shown that metformin is particularly effective in overweight type 2 diabetic subjects (12). Weight loss during metformin treatment was primarily attributed to loss of adipose tissue (1), and this was explained by differential effects of metformin on adipose tissue and muscle. Whereas metformin improves insulin sensitivity in muscle, it does not affect the antilipolytic action of insulin on adipose tissue (13). The overall effect of metformin on body weight is attributed to a reduction in caloric intake (11) (14) rather than to an increase in energy expenditure (1) (15) (16). Because reduction in body weight reduces insulin resistance, this may also represent a mechanism by which metformin improves insulin resistance (13). Furthermore, metformin treatment can lead to a decrease in androgen levels in women with polycystic ovary syndrome (17) (18). However, little is known about the effects of metformin treatment combined with a hypocaloric diet on testosterone levels in obese men with or without diabetes. Therefore, this study was carried out to evaluate the effects of metformin treatment combined with hypocaloric diet on plasma testosterone and leptin levels in such men.
- Top of page
- Research Methods and Procedures
Metformin treatment was well-tolerated by all subjects. None of the subjects suspended the therapy due to side effects, although some experienced transient diarrhea and flatulence during the first month of treatment.
Clinical and laboratory features of the patient and control groups are given in Table 1. In basal conditions, there was no significant difference in age, BMI, LH, TT, and FT in both groups. However, body weight (p = 0.004), waist circumference (p = 0.009), hip circumference (p = 0.001), plasma leptin (p = 0.008), and FSH (p = 0.014) were significantly lower and SHBG (p = 0.001) and WHR (p = 0.001) were significantly higher in diabetic subjects compared with nondiabetic obese men. In addition, despite having similar BMIs, diabetics had higher WHRs and higher SHBG levels.
Table 1. Clinical and laboratory features of diabetic and nondiabetic obese groups
| ||Diabetic group||Obese group||T or z||p Value|
|Age (years)||41.55 ± 8.03||41.50 ± 7.78||−0.014||NS|
|Waist (cm)||108.35 ± 6.10||114.45 ± 7.79||−2.758||0.009|
|Hip (cm)||119.40 ± 5.91||136.05 ± 10.69||−6.097||0.001|
|WHR||0.91 ± 0.04||0.84 ± 0.04||5.626||0.001|
|Weight (kg)||104.07 ± 7.49||105.32 ± 7.76||−1.8||NS|
|BMI (kg/m2)||35.50 ± 1.15||35.85 ± 1.57||−1.684||NS|
|Glucose (mM)||9.49 ± 2.76||5.58 ± 0.406||−5.29||0.001|
|Leptin (ng/mL)||16.32 ± 3.70||20.45 ± 5.42||−2.820||0.008|
|FSH (IU/liter)||2.75 ± 1.72||5.25 ± 3.38||−2.450||0.014|
|LH (IU/liter)||4.85 ± 2.55||5.50 ± 2.01||−1.272||NS|
|TT (ng/dL)||353.40 ± 133.77||365.85 ± 106.14||−0.189||NS|
|FT (pg/mL)||15.60 ± 5.23||28.51 ± 21.00||−1.650||NS|
|SHBG (nmol/mL)||26.36 ± 11.15||15.49 ± 8.67||−3.316||0.001|
|DHEAS (μg/mL)||226.28 ± 145.80||201.73 ± 121.52||−0.149||NS|
|Calculated FT (%)||15.96 ± 8.84||28.03 ± 11.7||−3.192||0.001|
Effects of metformin and the hypocaloric diet on clinical and laboratory parameters in both groups are given in Table 2. BMI and waist and hip circumferences were significantly decreased after metformin treatment combined with a hypocaloric diet. FSH, LH, and DHEAS levels were not significantly influenced by treatment in both groups. Although calculated FT levels did not change significantly in the diabetics, a significant decrease was evident in TT in diabetics but in FT in controls. SHBG did not change in the diabetic group, but it increased significantly in obese control subjects. Plasma leptin levels decreased significantly in both groups after treatment.
Table 2. Clinical and laboratory features before and after metformin treatment in diabetic obese and nondiabetic obese groups
| ||Diabetic group||Nondiabetic obese group|
| ||Before||After||Change (%)||T or z||p Value||Before||After||Change (%)||T or z||p Value|
|Age (years)||41.55 ± 8.03|| || || || ||41.50 ± 7.78|| || || || |
|Waist (cm)||108.35 ± 6.10||106.80 ± 5.67||−1.43||4.72||0.001||114.45 ± 7.79||112.35 ± 7.92||−1.83||5.80||0.001|
|Hip (cm)||119.40 ± 5.91||116.65 ± 5.76||−2.30||8.11||0.001||136.05 ± 10.69||132.80 ± 10.54||−2.38||9.58||0.001|
|Weight (kg)||104.07 ± 7.49||102.1 ± 7.27||−1.89||6.26||0.001||111.32 ± 7.76||109.1 ± 7.9||−1.99||7.06||0.001|
|BMI (kg/m2)||35.50 ± 1.15||34.85 ± 1.18||−1.83||−3.61||0.001||35.85 ± 1.57||35.10 ± 1.59||−2.09||7.55||0.001|
|Leptin (ng/mL)||16.32 ± 3.70||15.58 ± 3.87||−4.53||6.39||0.001||20.45 ± 5.42||19.13 ± 4.87||−6.45||5.78||0.001|
|Glucose (mM)||9.49 ± 2.76||7.19 ± 0.44||24.2||−4.23||0.001||5.58 ± 0.406||5.45 ± 0.29||−2.32||−1.23||0.38|
|FSH (IU/liter)||2.75 ± 1.72||3.20 ± 1.57||16.36||−1.74||0.08||5.25 ± 3.38||5.27 ± 3.58||−0.38||−0.45||0.65|
|LH (IU/liter)||4.85 ± 2.55||4.29 ± 3.00||11.5||−0.77||0.44||5.50 ± 2.01||6.17 ± 2.86||12.18||−1.91||0.06|
|TT (ng/dL)||353.40 ± 133.77||314.60 ± 124.78||−10.9||−2.84||0.01||365.85 ± 106.14||325.45 ± 117.39||−11.04||−1.74||0.08|
|FT (pg/mL)||15.60 ± 5.23||13.53 ± 4.34||−13.2||−0.78||0.43||28.51 ± 21.00||14.77 ± 5.17||−48.1||−2.58||0.01|
|SHBG (nmol/mL)||26.36 ± 11.15||23.56 ± 10.00||−10.6||−1.49||0.14||15.49 ± 8.67||21.81 ± 6.27||40.8||−3.36||0.001|
|DHEAS (μg/mL)||226.28 ± 145.80||217.26 ± 99.34||−3.98||−0.52||0.60||201.73 ± 121.52||201.74 ± 92.51||0.004||−0.60||0.55|
|WHR||0.91 ± 0.04||0.92 ± 0.03||1.10||−2.38||0.03||0.84 ± 0.04||0.85 ± 0.04||1.19||1.56||0.14|
|Calculated FT (%)||15.96 ± 8.84||16.11 ± 10.72||0.93||−0.52||0.6||28.03 ± 11.7||15.2 ± 4.3||−45.7||−3.50||0.001|
Of interest, significant decreases in glucose (24.2%), leptin (4.53%), TT (10.9%), and FT (13.2%) levels were noted in the diabetic group. These changes are higher than changes in BMI (−1.83%) and WHR (−1.1%). In the diabetic group, a decrease in TT levels was not correlated to changes in waist and hip circumference, WHR, BMI, fasting blood glucose, leptin, SHBG, or DHEAS. Changes in FT or TT levels were correlated to changes in FSH (r = −0.64, p = 0.002; r = −0.60, p = 0.005, respectively), but not to those observed in LH. Changes in calculated FT were correlated to changes in the levels of DHEAS (r = 0.45, p = 0.043), FSH (r = −0.49, p = 0.026), SHBG (r = −0.94, p = 0.001), and TT (r = 0.69, p = 0.001), but not with other parameters. Decreased leptin levels were associated with reduced BMI (r = 0.75, p = 0.001) and hip circumference (r = 0.56, p = 0.01), but not correlated to changes in other parameters, such as fasting blood glucose, FT or TT, calculated FT, WHR, waist circumference, and SHBG in diabetic obese men.
As shown in Table 2, a marked decrease in calculated FT (45.7%) and an increase in SHBG (40.8%) were noted in nondiabetic obese patients. In the obese group, changes in FT were correlated to changes in SHBG levels (r = −0.72, p = 0.001), FSH (r = 0.54, p = 0.012), and LH (r = 0.80, p = 0.001), but not to those observed in other parameters. Changes in TT were correlated to changes in FSH (r = 0.54, p = 0.012), LH (r = 0.44, p = 0.051), DHEAS (r = 0.45, p = 0.048), leptin (r = 0.58, p = 0.007), and SHBG (r = −0.75, p = 0.001), but not to changes in other parameters. Changes in calculated FT were correlated to changes in FSH (r = 0.51, p = 0.022), LH (r = 0.57, p = 0.007), leptin (r = 0.54, p = 0.013), SHBG(r = −0.88, p = 0.001), and TT (r = 0.88, p = 0.001), but not to changes in other parameters. Decreased leptin levels were associated with changes in TT (r = 0.58, p = 0.007) and changes in waist circumference (r = 0.59, p = 0.006), but not with the other parameters evaluated.
- Top of page
- Research Methods and Procedures
Metformin treatment combined with a hypocaloric diet leads to reduced FT or calculated FT levels in obese nondiabetic men and to reduced TT levels in obese men with type 2 diabetes. In contrast, Lima et al. (20) recently reported that weight loss by diet plus dexfenfluramine causes significant increase in TT or FT in moderately obese men. Because weight loss (or diet) in obese men has been reported to increase depressed testosterone levels (21), the decrease in testosterone levels observed in this study suggests an implication for metformin rather than for weight loss. Supporting this view, we could not find any relationship between changes in FT or TT and changes in BMI. Thus, our data suggest that metformin, but not weight loss, may be responsible for the decrement in TT levels after treatment.
Interestingly, we have observed increased SHBG levels in the presence of higher WHR in the diabetic group when compared with the control group before treatment. TT levels in diabetic subjects but FT levels in nondiabetic obese subjects were decreased after treatment. However, no change in SHBG levels in diabetic subjects was observed, whereas it was increased in nondiabetic obese subjects. It is possible that type 2 diabetes had reduced β-cell function and had lowered insulin after treatment, explaining the increased SHBG levels in the non-obese group. Previous studies have demonstrated that plasma testosterone concentrations are lower in men with type 2 diabetes than in normoglycemic men (22) (23) (24). Birkeland et al. (25) showed a strong correlation between insulin sensitivity and SHBG and a moderate correlation between insulin sensitivity and TT and WHR in 23 men with type 2 diabetes. Haffner et al. (26) showed a strong correlation between total and non-oxidative glucose disposal with SHBG. Thus, the association between SHBG and insulin levels is much weaker in men than in women and is probably inverse (27). However, in women, upper body obesity is associated with increased FT and decreased SHBG levels (28). Increased SHBG levels in diabetics vs. obese men may also be due to their higher WHR and this might be consistent with decreased insulin secretion observed in these patients, because insulin is a potent inhibitor of SHBG synthesis (29) (30). In fact, it could also explain why no alterations in SHBG levels were found in diabetics. Supporting this view, we have observed that all relationships with changes of several hormones and those of SHBG in diabetic group are clearly related to baseline values (particularly of SHBG). It is also conceivable that, because SHBG and leptin are negatively correlated in these patients, leptin may affect, at least in part, testosterone concentrations indirectly by inducing an alteration in SHBG levels. However, one study found no relationship between leptin and SHBG in obese men (31). In contrast, recently, Soderberg et al. (32) reported a negative association between leptin and SHBG in the low tertile of waist circumference in men. A decrease in the amplitude of spontaneous LH pulses can also account for decreased FT levels observed after treatment in nondiabetic obese patients. Although changes in FT were correlated with changes in FSH and LH, the differences in FSH and LH before and after treatment were not significant. Thus, it is unlikely that the decrements in FT levels are due to changes in LH release.
Our findings demonstrate that metformin treatment and a hypocaloric diet lead to a decrease in plasma leptin levels. Leptin, the product of the obese gene, is primarily produced by adipocytes, and its circulating levels correlate significantly with BMI or fat mass (33) (34). Because a decrease in BMI was observed after treatment, it is most likely that this is responsible for decreased leptin levels after treatment. Supporting this view, we have also found a correlation between BMI and leptin levels. Previous studies have also reported similar results (35). In support of our findings, Halle et al. (36) reported that weight loss in obese individuals with type 2 diabetes leads to a reduction in serum leptin levels. However, Haffner et al. (37) reported elevated leptin levels after sulfonylurea treatment in type 2 diabetics, because their patients had gained weight after treatment. In contrast, we could not find any correlation between changes in leptin and changes in glucose, although some studies have reported a positive relation between leptin and glycemic control or hemoglobin A1c (38) (39), whereas others found a negative relationship (40) or no relationship at all (41). Recent studies suggest a complex interaction between leptin and insulin or insulin resistance (6). In contrast, Mueller et al. (6) and Mick et al. (7) recently demonstrated that metformin also inhibits leptin secretion from cultured adipocytes. A limitation of our study is the lack of placebo group. Thus, it is not clear whether metformin directly reduces plasma leptin levels (through yet to be defined mechanisms) or whether the combination of metformin and a hypocaloric diet reduces plasma leptin levels by inducing weight loss.
We conclude that metformin treatment combined with a hypocaloric diet lead to a decrease in FT levels in nondiabetic obese men and to a decrease in TT levels in obese men with type 2 diabetes. Furthermore, a significant increase in SHBG levels was observed in the control group, and this may explain the decrease in FT in obese control subjects. These findings suggest that metformin treatment plus a hypocaloric diet lead to a decrease in TT levels in obese men with type 2 diabetes, whereas lowered FT levels are observed in nondiabetic obese men.