Update: Hypogonadotropic Hypogonadism in Type 2 Diabetes and Obesity

Abstract

Studies over the last few years have clearly established that at least 25% of men with type 2 diabetes have subnormal free testosterone concentrations in association with inappropriately low LH and FSH concentrations. Another 4% have subnormal testosterone concentrations with elevated LH and FSH concentrations. The Endocrine Society, therefore, now recommends the measurement of testosterone in patients with type 2 diabetes on a routine basis. The subnormal testosterone concentrations are not related to glycosylated hemoglobin or duration of diabetes, but are associated with obesity, very high C-reactive protein concentrations, and mild anemia. In addition, subnormal testosterone concentrations in these men are associated with a two to three times elevated risk of cardiovascular events and death in two early studies. Short-term studies of testosterone therapy in hypogonadal men with type 2 diabetes have demonstrated an increase in insulin sensitivity and a decrease in waist circumference. However, the data on the effect of testosterone replacement on glycemic control and cardiovascular risk factors such as cholesterol and C-reactive protein concentrations are inconsistent. As far as sexual function is concerned, testosterone treatment increases libido but does not improve erectile dysfunction and thus, phosphodiesterase inhibitors may be required. Trials of a longer duration are clearly required to definitively establish the benefits and risks of testosterone replacement in patients with type 2 diabetes and low testosterone.

Subnormal free testosterone concentrations in association with inappropriately low LH and FSH concentrations and a normal response to GnRH of LH and FSH in type 2 diabetes were first described in 2004 (1). These abnormalities were independent of the duration and severity of hyperglycemia [glycosylated hemoglobin (HbA1c)]. Magnetic resonance imaging in these hypogonadal patients showed no abnormality in brain or the pituitary (1). This association of hypogonadotropic hypogonadism (HH) with type 2 diabetes has now been confirmed in several studies and is present in 25–40% of these men (2–5). In this context, it is important that The Endocrine Society now recommends the measurement of testosterone in patients with type 2 diabetes on a routine basis (6). These observations were recently extended to younger patients with type 2 diabetes between the ages of 18 and 35 yr who had HH at a rate of 33% when the usual normal range for middle age was employed, whereas the rate was 58% when age-specific normal range for free testosterone for the young was employed (7). With the advent of more specific liquid chromatography tandem mass spectrometry assay for measuring total testosterone, the reference ranges for total and free testosterone have recently been revised downward. Using this methodology, in our most recent study, we have found that 29% of men with type 2 diabetes have subnormal free testosterone concentrations, as measured by equilibrium dialysis (8); 25% had HH, whereas 4% had hypergonadotropic hypogonadism.

Type 2 diabetic men with low testosterone levels have also been found to have a high prevalence of symptoms suggestive of hypogonadism such as fatigability and erectile dysfunction (2). In all of the above studies, total testosterone and free testosterone concentrations were inversely related to body mass index (BMI) and age. However, the presence of low testosterone concentration was not entirely dependent upon obesity because 25% of nonobese patients (31% of lean and 21% of overweight) also had HH (1). HH is relatively rare in type 1 diabetes and, therefore, is not a function of diabetes or hyperglycemia per se (9). Thus, in view of the inverse relationship between BMI and testosterone concentrations in both type 1 and type 2 diabetes, HH is probably related to insulin resistance (1, 4, 9). Previous studies have shown that hypogonadism is associated with upper abdominal adiposity, insulin resistance, and the metabolic syndrome (10, 11). Treatment of systemic insulin resistance by rosiglitazone leads to a modest increase in testosterone concentrations in men with type 2 diabetes (12), without the restoration of testosterone concentrations to normal.

A recent study investigated the prevalence of low testosterone concentrations in a large number of obese and diabetic men (mean age, 60 yr; range, 45–96 yr) (13); 44% of diabetic and 33% of age-matched nondiabetic men had subnormal free testosterone concentrations, respectively. Forty percent of obese men and 50% of obese diabetic men had subnormal free testosterone concentrations. Thus, obesity is associated with a high prevalence of hypogonadism, and the presence of diabetes adds to that risk.

Possible Pathophysiological Mechanisms Underlying HH in Type 2 Diabetes

Role of estradiol

Because testosterone and androstenedione in the male can be converted to estradiol and estrone, respectively, through the action of aromatase in the mesenchymal cells and preadipocytes of adipose tissue, it has been suggested that excessive estrogen secretion due to aromatase activity in the obese may potentially suppress the hypothalamic secretion of GnRH (14). This hypothesis was examined in a recent study that compared the estradiol concentrations in 240 type 2 diabetic men with and without HH (8). Total estradiol concentrations were measured by immunoassay, and free estradiol concentrations were calculated using SHBG. Total and free estradiol concentrations in men with HH were significantly lower than in those without HH (8). To confirm these findings, total estradiol concentrations were measured in a subset of 102 men by the liquid chromatography tandem mass spectrometry assay, and free estradiol concentrations were measured by equilibrium dialysis. Estradiol concentrations were 25% lower in men with HH. Free estradiol concentrations were directly related to free testosterone concentrations, irrespective of age or BMI. The diminished availability of the substrate, testosterone, may therefore be the major determinant factor of estradiol concentrations in these men. A study in elderly men (European Male Ageing Study) has also found lower estradiol concentrations in hypogonadal men (15). Thus, it appears that the low testosterone concentrations in HH of diabetes, as in aging, are not the consequence of estradiol-dependent suppression of the hypothalamo-hypophyseal-gonadal axis. Furthermore, HH in type 2 diabetic men with a normal weight is not likely to be associated with increased estradiol concentrations (1).

Role of insulin resistance

The selective deletion of the insulin receptor from neurons in mice leads to a reduction in LH concentrations by 60–90% and low testosterone concentrations (16). These animals respond to GnRH challenge by normal or supranormal release of LH. In addition, these animals had atrophic seminiferous tubules with markedly impaired or absent spermatogenesis. In addition, it is known that the incubation of hypothalamic neurons with insulin results in the facilitation of secretion of GnRH (17, 18). Thus, insulin action and insulin responsiveness in the brain are necessary for the maintenance of the functional integrity of the hypothalamo-hypophyseal-gonadal axis.

Role of inflammatory mediators

TNF-α and IL-1β have been shown to suppress hypothalamic GnRH and LH secretion in experimental animals and in vitro (19, 20). It is therefore relevant that C-reactive protein (CRP) concentrations are markedly increased in hypogonadal type 2 diabetic men compared with men with type 2 diabetes and normal testosterone (6.5 vs. 3.2 mg/liter) (21). These data were confirmed by another study from Australia in which the median CRP concentration in type 2 diabetic patients with low total testosterone was 7.7 mg/liter compared with 4.5 mg/liter in men with normal testosterone (4). Free testosterone concentrations were inversely related to CRP concentrations (r = −0.27; P = 0.02). It is thus possible that inflammatory mediators may contribute to the suppression of the hypothalamo-hypophyseal axis and the syndrome of HH in type 2 diabetes. The presence of inflammation may also contribute to insulin resistance because several inflammation-related mediators, such as suppressor of cytokine signaling-3, IκB kinase β, and c-Jun N-terminal kinase-1 interfere with insulin signal transduction (22, 23) and contribute to insulin resistance. These mediators are also known to be increased in obesity (24).

In summary, it is likely that there are several interlinked causative mechanisms underlying HH in men with type 2 diabetes. It should also be noted that human chorionic gonadotropin-induced testosterone secretion by Leydig cells is inversely related to insulin sensitivity (as measured by hyperinsulinemic euglycemic clamp) among men with varying degrees of glucose tolerance (25). Thus, the lesion resulting in hypogonadism in obesity and type 2 diabetes may occur at several levels of the hypothalamic-pituitary-gonadal axis. However, the absence of an increase in gonadotropin concentrations indicates that the primary defect in type 2 diabetes and obesity is at the hypothalamo-hypophyseal level.

What Comes First: Hypogonadism or Type 2 Diabetes?

Because even young men with type 2 diabetes and patients with newly discovered type 2 diabetes have a high prevalence of HH and obesity is associated with HH, it is possible that HH precedes diabetes. Several epidemiological studies have shown that low testosterone at baseline approximately doubles the odds of development of type 2 diabetes (26–28). The data, however, are more consistent with total testosterone than with free testosterone (29). It is possible that low SHBG concentrations may mediate a portion of this association. SHBG polymorphisms that lead to lower SHBG concentrations are strongly predictive of the development of type 2 diabetes, whereas those that lead to higher SHBG concentrations are protective (30, 31).

Does Hypogonadism Matter? Possible Consequences of Hypogonadism in Type 2 Diabetes

It is well accepted that low testosterone concentrations are associated with symptoms such as fatigue, lack of libido, and erectile dysfunction. Recent studies have described pathophysiological effects of subnormal testosterone concentrations beyond those related to sexual health, as discussed below.

Symptoms of sexual dysfunction

Cross-sectional studies have found a high prevalence of low libido (64%), erectile dysfunction (74%), and fatigue (63%) in hypogonadal men with type 2 diabetes (2). However, the presence of these symptoms was similarly high in eugonadal men with type 2 diabetes as well (48, 65, and 57%, respectively). The treatment of erectile dysfunction with phosphodiesterase-5 inhibitors such as sildenafil in men with type 2 diabetes is known to be not as effective as that in nondiabetic subjects (32).

Cardiovascular disease

Recent evidence from longitudinal observational studies shows that low testosterone concentration is prospectively associated with an increase in the incidence of cardiovascular events. Laughlin et al. (33) prospectively followed 794 elderly men (mean age, 71 yr) for 20 yr in a community setting. The hazard ratio for men in the lowest quartile of bioavailable testosterone was 1.44 for all-cause mortality and 1.36 for cardiovascular mortality. Another prospective study [Osteoporotic Fracture in Men (MrOS) Swedish cohort (34)] that included 3014 men (mean age, 75 yr; mean follow-up, 4.5 yr) showed a 65% increased risk of mortality in men with low free testosterone (<6.1 ng/dl). Subnormal free testosterone concentrations are associated with a 69% increased risk of stroke or transient ischemic attack (35). Many cross-sectional, retrospective, case-control and smaller studies have also demonstrated an association of low testosterone with increased mortality (36–38). However, the relationship between cardiovascular mortality and low testosterone was not seen in two longitudinal studies (39, 40). These studies were done in relatively younger populations (mean ages, 52 and 55 yr) and had much lower mortality rates, which can possibly explain the lack of an association (39, 40).

A recent study in 930 men with coronary artery disease reported that a low testosterone at baseline was associated with increased mortality after 7 yr of follow-up (21 vs. 12%) (41). Only one study has looked at the association between subnormal testosterone concentrations and cardiovascular mortality specifically in men with type 2 diabetes (42): in 153 men with type 2 diabetes and known coronary artery disease, subnormal free testosterone concentration at baseline increased cardiovascular mortality by three times over 2 yr.

Insulin sensitivity

HH in men with type 2 diabetes is associated with a higher BMI (3–4 kg/m²), 12% more sc fat mass (measured by dual-energy x-ray absorptiometry), and higher waist-to-hip ratio compared with eugonadal men with type 2 diabetes (1, 2, 43). In one study involving type 2 diabetic men from the United Kingdom, 74% of hypogonadal men were obese compared with 54% of eugonadal men (2). As of yet, no study has measured visceral, im, or hepatic fat content in type 2 diabetic men with and without HH. Many studies have documented that hypogonadism is associated with insulin resistance (reviewed in Refs. 44 and 45). No study has compared the insulin resistance in type 2 diabetic men with subnormal or normal testosterone concentrations.

Hematocrit

Hypogonadal type 2 diabetic men have a lower hematocrit than those with normal testosterone concentrations (21). The prevalence of normocytic normochromic anemia in such patients is 38% compared with 3% in those with normal testosterone concentrations. A large study (464 men) also found a direct correlation between free testosterone concentrations and hemoglobin in men with type 2 diabetes and renal insufficiency (46). Testosterone regulates erythropoiesis (47). However, it has not yet been determined whether the association of anemia with hypogonadism in men with type 2 diabetes is causal or is secondary to other confounding factors such as inflammation. In these men, hemoglobin is positively related to testosterone but negatively related to CRP concentrations (21).

Bone density

Hypogonadism is associated with a decrease in bone mineral density (BMD) and an increase in fracture rate (48, 49). Furthermore, trabecular bone architecture (measured by high-resolution magnetic resonance imaging) deteriorates much more in hypogonadal men compared with eugonadal men (50). Hypogonadal men usually have lower estradiol concentrations compared with eugonadal men because testosterone is the substrate for estradiol formation by aromatization (15). In epidemiological studies, estradiol concentrations correlate more robustly with BMD than testosterone concentrations in men (51). This is especially true of trabecular bone. However, testosterone appears to be an independent predictor of cortical bone density (52, 53). One study in men with type 2 diabetes has shown that free testosterone concentrations are positively associated with BMD in arms and ribs, but not with hip, spine, or total body BMD values (43). Another study has shown a positive relation of lumbar spine BMD with free testosterone concentrations in men with type 2 diabetes (54). No study has evaluated the relation between BMD and free estradiol concentrations in these men. It is possible that BMD in men with type 2 diabetes might relate more strongly to estradiol than to testosterone concentrations, as has been shown in elderly nondiabetic men. No data are available on the fracture rates of hypogonadal men with type 2 diabetes.

Prostate-specific antigen (PSA)

Type 2 diabetic men have 20% lower PSA concentrations than nondiabetic men (55). PSA concentrations are lower in hypogonadal than in eugonadal type 2 diabetic men (0.89 vs. 1.1 ng/ml) (56). It is interesting that the incidence of prostatic carcinoma is lower in men with diabetes. This is in contrast to the increased incidence of cancer in diabetics in various organs including the colon, the kidney, the breast, the endometrium, and the pancreas (57). The diminished incidence of prostate cancer in diabetics may receive a contribution from the high prevalence of HH and low testosterone concentrations. However, epidemiological studies do not support a causative role of testosterone in prostate cancer in nondiabetic populations (58).

Should Testosterone Be Measured in Every Patient with Type 2 Diabetes?

Because the frequency of subnormal free testosterone concentrations in type 2 diabetes is at least 25%, we believe that free testosterone concentration should be measured in every patient with type 2 diabetes. This is consistent with The Endocrine Society guidelines. The prevalence of hypothyroidism is between 5 and 8% in this population, and yet we screen every one for this condition. An Androgen Deficiency in Aging Male (ADAM) questionnaire should be administered in every patient with a low testosterone so that the presence of clinical hypogonadism can be established. One can argue that if the case for the replacement of testosterone in patients with HH is not proven, as discussed below, is there a case for measuring its concentrations in every patient with type 2 diabetes? We believe that there is because, like hypothyroidism, patients may slide gradually into this clinical state without any overt symptoms that may be revealed through direct questioning. “Asymptomatic” men may realize that they had been symptomatic only after a trial with testosterone. Such patients may potentially benefit from testosterone replacement therapy, as discussed below.

Should Men with Type 2 Diabetes and Low Testosterone Be Replaced with Testosterone? Issues to Be Considered in View of the Above Data

The Endocrine Society recommends that men with low testosterone and symptoms of androgen deficiency be considered for therapy with testosterone (6). The guidelines do not recommend treatment of asymptomatic men with low testosterone. The Institute of Medicine recommends that more short-term studies in selected populations should investigate the benefits and risks of testosterone therapy. Trials in men with type 2 diabetes and obesity are important in this regard because both are commonly associated with hypogonadism. A few studies on testosterone replacement in type 2 diabetic men with low testosterone have emerged and are described below.

Insulin resistance

Three studies have shown a decrease in insulin resistance after testosterone therapy in hypogonadal men with type 2 diabetes. Kapoor et al. (59) studied the effects of treatment with im testosterone for 3 months in 24 hypogonadal type 2 diabetic men in a placebo-controlled, double-blind, crossover trial. Homeostasis model assessment for insulin resistance (HOMA)-IR decreased by 1.73 after testosterone therapy compared with placebo. In another trial, 32 men with the metabolic syndrome and newly diagnosed type 2 diabetes with total testosterone concentration of less than 350 ng/dl (12 nmol/liter) were prescribed diet and exercise (60). Half of them were also given transdermal testosterone for 1 yr. Testosterone therapy resulted in greater improvements in insulin sensitivity (measured by HOMA-IR; −0.9) compared with diet and exercise alone. A prospective, randomized, double-blind multicenter trial of transdermal testosterone (3 g metered-dose 2% gel for 1 yr) therapy in 220 hypogonadal men with type 2 diabetes or metabolic syndrome has recently been published [Testosterone Replacement in Hypogonadal Men with Either Metabolic Syndrome or Type 2 diabetes study (TIMES2) (61)]. The primary endpoint of the study was a change in insulin sensitivity, as measured by HOMA-IR. Patients were evaluated every 3 months. A total of 136 men in the study had type 2 diabetes, 176 men had metabolic syndrome, and 92 men had both. Testosterone therapy resulted in a 15% (P = 0.01) decrease in HOMA-IR at 6 months and at 1 yr time-points in men with type 2 diabetes as well as in those with metabolic syndrome. One study in lean hypogonadal type 2 diabetic men with a mean BMI of 24 kg/m² did not show any change in insulin sensitivity after treatment with low-dose im testosterone (100 mg every 3 wk) for 3 month (62). This dose is inadequate and may account for the lack of effect. It is, however, possible that the change in insulin sensitivity due to testosterone therapy occurs only in obese, and presumably insulin-resistant, men. Thus, it appears that insulin resistance improves with testosterone therapy in obese men with type 2 diabetes. These studies have calculated HOMA-IR to measure insulin resistance. This needs to be confirmed by trials that use hyperinsulinemic-euglycemic clamp methodology. It is also not clear whether the effect is due to a change in body composition or independently of it.

Glycemic control

In three of the above-mentioned studies, glycemic control was also evaluated by measuring HbA1c and fasting glucose. The small study by Kapoor et al. (59) showed a decrease in fasting glucose (28 mg/dl) and HbA1c (0.37%) compared with placebo with 3 months of testosterone replacement. The trial in men with new onset type 2 diabetes with transdermal testosterone did show a decrease in HbA1c from 7.5 to 6.3% over a period of 1 yr (60). This was in conjunction with diet and exercise, but no hypoglycemic medications. The comparison group in this study was a diet and exercise group. There was a decrease in HbA1c from 7.5 to 7.1% in this group. The mean fasting glucose decreased by 34 and 29 mg/dl in the testosterone and diet/exercise groups, respectively (P = 0.06 for comparison among groups). However, the larger trial (TIMES2) did not show a clear effect of testosterone replacement on HbA1c (61). Medication changes were not allowed for the first 6 months of the study. Patients with type 2 diabetes showed a trend toward improvement in HbA1c at 1 yr (−0.4%; P = 0.057) but not at 6 months (P = 0.6). Although no changes were made in patient's medications for the first 6 months, the study protocol allowed medication changes between 6 and 12 months; therefore, no clear conclusions can be made regarding the effect of testosterone therapy on glycemic control from this trial. There were no changes in fasting glucose or insulin. Thus, there appears to be a mild decrease in HbA1c with testosterone therapy in men with type 2 diabetes, but the data are inconsistent and currently testosterone replacement cannot be recommended for glycemic control.

Symptoms and sexual dysfunction

In the TIMES2 trial, there was an improvement in the International Index of Erectile Function score in the testosterone replacement group, mainly due to an increase in sexual desire, but other symptoms did not change. The smaller trial of im testosterone by Kapoor et al. (59) in hypogonadal men with type 2 diabetes showed an improvement in symptoms as measured by the ADAM questionnaire. Although there are no specific studies assessing the effect of testosterone replacement on the effectiveness of phosphodiesterase IV inhibitors like sidenafil, studies in hypogonadal nondiabetics do show this benefit (63).

Body composition and abdominal adiposity

Heufelder et al. (60) showed a decrease in waist circumference of 14 cm in men with new onset type 2 diabetes treated for 1 yr with transdermal testosterone, diet, and exercise. The control group that was prescribed only diet and exercise lost 5 cm. Kapoor et al. (59) showed a decrease by 1.63 cm in waist circumference after im testosterone treatment. In the TIMES2 trial, there was a small but statistically significant decrease in waist circumference (0.8 cm) in type 2 diabetic men treated with testosterone. Significantly, BMI did not change in any of the studies despite the decrease in abdominal girth.

Cardiovascular outcomes

A recent meta-analysis of testosterone therapy trials ranging from 3 months to 3 yr did not show any change in the rates of death, myocardial infarctions, revascularization procedures, or cardiac arrhythmias compared with placebo/nonintervention groups (64). However, none of these trials was powered to show a difference. Surprisingly, a recent trial of testosterone replacement therapy designed to study the effects of testosterone replacement for 6 months on muscle mass and strength in elderly men (>65 yr old) with limited mobility had to be discontinued prematurely due a higher incidence (22 vs. 5%) of cardiovascular-related adverse events in the testosterone treatment arm compared with the placebo arm (65). This trial was not included in the previously mentioned meta-analysis. The study population had a high prevalence of chronic conditions, and it is possible that the results could have been due to chance alone. However, other studies in elderly populations have not shown an increase in cardiac events after testosterone replacement (66–68). The TIMES2 trial (61) reported that cardiovascular events occurred less commonly with testosterone than with placebo (4.6 vs. 10.7%; P = 0.095); however, this effect was short of significance. A recent study presented at the British Endocrine Societies meeting is of interest (69). This study investigated the effect of baseline testosterone concentrations and testosterone replacement therapy in hypogonadal men with type 2 diabetes on all-cause mortality. A total of 578 men with type 2 diabetes with a mean age of 61 yr were followed for 5.8 ± 1.3 yr; 338 men had normal testosterone concentrations at baseline; 240 were hypogonadal, of which 58 men received testosterone replacement therapy; and 72 men (12%) died during follow-up. The mortality rate in eugonadal men and untreated hypogonadal men was 9 and 20%, respectively. Hypogonadal men treated with testosterone had a mortality rate of 8.6%, significantly lower than that in the untreated hypogonadal group. Testosterone replacement in the setting of heart failure has also recently been reported to have beneficial effects on exercise capacity, muscle strength, and HOMA-IR (70).

One study showed a decrease of 15 mg/dl in total cholesterol but no change in low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, or triglycerides after testosterone therapy for 3 months (59). In the TIMES2 trial, men with metabolic syndrome had a 15% decline in lipoprotein(a) and a 7% decline in total and low-density lipoprotein concentrations. Men with type 2 diabetes had similar trends, but the results were not significant. There was, however, a 6% decline in high-density lipoprotein concentrations in both the metabolic syndrome and type 2 diabetes groups (61). No changes have been seen in blood pressure after testosterone treatment (59, 61).

Heufelder et al. showed a decrease in CRP concentrations (−0.5 mg/dl) and an increase in adiponectin (0.9 μg/ml) after testosterone therapy (60). However, CRP, IL-6, resistin, and TNF-α concentrations did not change after im testosterone replacement for 3 months in a trial by Kapoor et al. (71). There was also a decrease in adiponectin after testosterone therapy. The reasons for the discrepancies between studies are not clear but could be related to the differences in study design, route of testosterone administration, and duration of therapy.

Safety issues

The TIMES2 trial (61) did not show an increase in age-adjusted PSA values. PSA concentrations exceeded normal limits in four subjects at 12 months (three in the testosterone treatment arm and one in placebo). Mean PSA concentrations did not change after 1 yr of therapy in the study by Heufelder et al. (60) either. In this context, it is important that the replacement of testosterone in hypogonadal patients in general does not lead to an increased risk of prostatic carcinoma, although the trials have been too limited in duration and number of patients (64).

Conclusions

HH is found in 25% of men with type 2 diabetes. An additional 4% have hypergonadotropic hypogonadism. Low testosterone concentrations in men with type 2 diabetes are associated with an increased prevalence of symptoms of hypogonadism, obesity, very high CRP concentrations, mild anemia, and decreased BMD. In addition, these men have an elevated risk (two to three times) of cardiovascular events and death in two small studies. Short-term studies of testosterone therapy have demonstrated an increase in libido. In addition, there is an increase in insulin sensitivity. Some, but not all studies, have shown an improvement in glycemia, body composition, and cardiovascular risk factors such as cholesterol and CRP concentrations. Trials of a longer duration are clearly required to definitively establish the benefits and risks of testosterone replacement in patients with type 2 diabetes and HH.