Prostate carcinoma (PCa) is one of the most common malignancies in men. Androgen-deprivation therapy (ADT) is used frequently in the treatment of recurrent and metastatic PCa, rendering these men hypogonadal. Because male hypogonadism is associated with an unfavorable metabolic profile, and men with PCa have high cardiovascular mortality, the authors evaluated the effects of long-term ADT on fasting glucose levels, insulin levels, and insulin resistance.
To evaluate the long-term effects of ADT on fasting glucose and insulin resistance in men with PCa who received ADT and to determine whether these metabolic alterations are a result of hypogonadism, the authors conducted a cross-sectional study at a university-based research institution in the United States. In total, 53 men were evaluated, including 18 men with PCa who received ADT for at least 12 months prior to the onset of the study (the ADT group), 17 age-matched men with nonmetastatic PCa who had undergone prostatectomy and/or received radiotherapy and who were not receiving ADT (the non-ADT group), and 18 age-matched controls (the control group). None of the men had a known history of diabetes mellitus.
The mean age was similar in all 3 groups (P = 0.33). Serum total testosterone levels (P < 0.0001) and free testosterone levels (P < 0.0001) were significantly lower in the ADT group compared with the other groups. Men in the ADT group had a higher BMI compared with the other groups (overall P = 0.005). After adjustment for age and BMI, men in the ADT group had significantly higher fasting levels of the following parameters: 1) Glucose levels were 131.0 ± 7.43 mg/dL in the ADT group compared with 103.0 ± 7.42 mg/dL in the non-ADT group (P = 0.01) and 99.0 ± 7.58 mg/dL in the control group (P < 0.01). 2) Insulin levels were 45.0 ± 7.25 uU/mL in the ADT group compared with 24.0 ± 7.24 uU/mL in the non-ADT group (P = 0.05) and 19.0 ± 7.39 uU/mL in the control group (P = 0.02). 3) Leptin levels were 25.0 ± 2.57 ng/mL in the ADT group compared with 12.0 ± 2.56 ng/mL in the non-ADT group (P < 0.01) and 6.0 ± 2.62 ng/mL in the control group (P < 0.01). 4) The homeostatic model assessment for insulin resistance (HOMAIR) = 17.0 ± 2.78 in the ADT group compared with HOMAIR = 6.0 ± 2.77 in the non-ADT group (P < 0.01) and HOMAIR = 5.0 ± 2.83 in the control group (P = 0.01). There was a significant negative correlation between total and free testosterone levels with fasting glucose, insulin, leptin, and HOMAIR.
Prostate carcinoma (PCa) is one of the most common malignancies in men and has an increasing incidence.1 Local surgery and/or radiotherapy are the usual treatments for men with locally confined PCa. In men with recurrent or metastatic PCa, androgen-deprivation therapy (ADT) is used with bilateral orchiectomy or gonadotropin-releasing hormone (GnRH) agonists. Although the use of ADT has resulted in improved survival for men with advanced PCa,2, 3 the resulting hypogonadism has adverse consequences. These side effects include an increase in body mass index (BMI) and declines in lean body mass, muscle strength, bone density, sexual function, and quality of life compared with men who only underwent local surgery and/or received radiation therapy and compared with an age-matched control group.4
An under-appreciated consequence of male hypogonadism is the development of insulin resistance and diabetes mellitus. Epidemiological studies have shown that men with low testosterone levels are more prone to develop diabetes mellitus than their eugonadal counterparts.5, 6 This is supported by interventional studies showing that testosterone replacement in hypogonadal, obese men results in an improvement in insulin sensitivity.7
Data on the association between hypogonadism and insulin resistance in men with PCa undergoing ADT are very limited. A recent short-term study in 22 men with PCa who received ADT showed an increase in fasting insulin levels (without any changes in fasting glucose levels) after 3 months of treatment.8 However, that increase in insulin levels (a manifestation of insulin resistance) was attributed to the increase in fat mass. Because insulin resistance is a risk factor for cardiovascular disease (CVD), this may be another consequence of ADT in men with PCa. Indeed, recent evidence showed an increase in cardiovascular mortality in men with PCa.9 Because men who are receiving ADT provide an excellent model in which to study the adverse effects of hypogonadism (because they have castrate levels of testosterone), we decided to evaluate the long-term effects of ADT on fasting glucose levels, insulin levels, leptin levels, and insulin resistance. We also attempted to determine whether these metabolic alterations are either a result of changes in BMI or a direct consequence of hypogonadism.
MATERIALS AND METHODS
The study design was cross-sectional. We studied 3 groups of men: 1) 18 men with PCa who were receiving ADT (for recurrent or metastatic disease) for at least 12 months prior to the onset of the study (the ADT group); 2) 17 age-matched men with nonmetastatic PCa who had undergone prostatectomy and/or received radiotherapy but had not received ADT and were eugonadal (the non-ADT group); and 3) 18 age-matched, normal, eugonadal men (serum testosterone levels > 280 ng/dL) who had normal prostate-specific antigen levels (the control group). The average duration of ADT (ADT group) was 45 months (range, 12–101 mos). Three patients in the ADT group had undergone orchiectomy, and the remaining 15 men were receiving GnRH agonists. None of the patients in any of the three groups had a known history of diabetes mellitus, they were not on any special diet, and they were not receiving any special diabetic medications.
The evaluation of men who were receiving ADT provided an opportunity to determine the effects of hypogonadism on the metabolic profile, the evaluation of men in the non-ADT group allowed us to account for the influence of PCa on these metabolic parameters, and the control group was recruited to account for the insulin resistance that occurs with aging.10 Men in the ADT and non-ADT groups were recruited from the Kimmel Cancer Center at Johns Hopkins. The control group was recruited from a data base of age-matched, eugonadal men at the Johns Hopkins Hospital Clinical Trials Unit. All men signed informed consent that was approved by the Institutional Review Board of the Johns Hopkins Medical Institutions.
Men were excluded from the study if they had any of the following: liver function tests or serum creatinine levels > 2 times the upper limit of normal, glucocorticoid use in the previous 3 months, a history of thyroid disease, a history of any form of hypogonadism prior to the diagnosis of PCa (both the ADT group and the non-ADT group), and men who had undergone or were currently undergoing cytotoxic chemotherapy.
Total and free testosterone levels were measured commercially (Quest Diagnostics). The normal range for total testosterone was 241–827 ng/dL, and the normal range for free testosterone was 8–24 ng/dL. Serum insulin and leptin levels were measured by using an enzyme-linked immunosorbent assay (ALPCO Diagnostics; kit 008-10-113-01; Linco Company; kit EZH4-80SK, respectively). The calculation of homeostatic model assessment for insulin resistance (HOMAIR) was performed by using the method described by Matthews et al.11
All data were analyzed using SAS software (version 9.1; SAS Institute Inc., Cary, NC). Standard methods were used to compute means, standard errors, and linear regression models. One-way analyses of variance were used to compare confounders among the three groups. Bonferroni multiple comparison, post-hoc tests were used to compare the mean values. Patient age and BMI adjusted for mean values were compared by using analyses of covariance. All significance tests for the comparisons were 2-sided, and P values < 0.05 were considered statistically significant.
The mean age was similar in the three groups. Serum total testosterone (P < 0.0001) and free testosterone (P < 0.0001) levels were significantly lower in the ADT group compared with the other 2 groups (Table 1). Men in the non-ADT and control groups were eugonadal. Men in the ADT group had a significantly higher BMI compared with the BMI in other 2 groups (P = 0.005). Fasting leptin levels also were significantly higher in the ADT group, indicating increased fat mass (P < 0.001). Similarly, fasting glucose and insulin levels were significantly higher in the ADT group compared with the other 2 groups (P = 0.002 and P = 0.002, respectively). Men in the ADT group also were more insulin-resistant, as reflected by their HOMAIR values (P < 0.001).
Table 1. Demographic, Hormone, and Unadjusted Metabolic Parameters by Study Groupa
ADT group (n = 18 men)
Non-ADT group (n = 17 men)
Control group (n = 18 men)
Overall P value
ADT: androgen-deprivation therapy; HOMAIR: homeostatic model assessment for insulin resistance.
Data are presented as the mean ± standard error.
Values within a row with different superscript letters are significantly different (P < 0.05).
After adjustment for age and BMI, men in the ADT group had significantly higher levels of fasting serum glucose (131.0 ± 7.43 mg/dL) compared with men in the non-ADT group (103.0 ± 7.42 mg/dL; P = 0.01) and in the control group (99.0 ± 7.58 mg/dL; P < 0.01; overall P = 0.01) (Fig. 1). Similarly, fasting serum insulin levels were significantly higher in the ADT group (45.0 ± 7.25 uU/mL) compared with the non-ADT group (24.0 ± 7.24 uU/mL; P = 0.05) and the control group (19.0 ± 7.39 uU/mL; P = 0.02; overall P = 0.04). Hypogonadal men were more insulin resistant (HOMAIR = 17 ± 2.78) compared with men in the non-ADT group (HOMAIR = 6.0 ± 2.77; P < 0.01) and the control group (HOMAIR = 5.0 ± 2.83; P = 0.01; overall P = 0.01). It is noteworthy that, even after adjustment for BMI, men in the ADT group had significantly higher leptin levels (25.0 ± 2.57 ng/mL) compared with men in the non-ADT group (12.0 ± 2.56 ng/mL; P < 0.01) and the control group (6.0 ± 2.62 ng/mL; P < 0.01; overall P = 0.0005), suggesting that hypogonadism is directly responsible for hyperleptinemia.
Correlation with Sex Hormones
There was a significant overall negative correlation of total and free testosterone with fasting glucose, insulin, leptin, and HOMAIR (Fig. 2A,B). We determined the number of men in the 3 groups who had fasting glucose levels ≥ 126 mg/dL, which is among the criteria for the diagnosis of diabetes mellitus. Eight of 18 men (44%) in the ADT group had fasting glucose levels ≥ 126 mg/dL compared with 2 of 17 men (12%) in the non-ADT group and 2 of 18 men (11%) in the control group (Fig. 3).
Epidemiologic data suggest that male hypogonadism is associated with the development of insulin resistance and diabetes mellitus,12–14 and testosterone replacement in men results in a significant improvement in insulin sensitivity.7 Because men with PCa who are receiving ADT have castrate levels of testosterone, they provide an excellent model to study the association between testosterone and insulin resistance. We found that men in the ADT group had higher levels of fasting glucose, insulin, and leptin and were more insulin resistant compared with the other two groups, even after adjustment for age and BMI. Furthermore, testosterone was correlated negatively with all of those parameters. These observations suggest that ADT in men with PCa leads to the development of insulin resistance and diabetes mellitus.
In contrast to women, in whom hyperandrogenism is associated with insulin resistance and diabetes,15 hypogonadism in men leads to an adverse metabolic profile. Indeed, low testosterone levels in men are associated with elevated insulin levels, even after adjustment for BMI,16 a key finding of the current study. Furthermore, epidemiologic studies have shown that hypotestosteronemia predicts the development of diabetes in nondiabetic men.12 These observations have been confirmed in clinical trials in which testosterone replacement in men resulted in an improvement in insulin sensitivity.7
A review of the literature suggests that both physiologic aging and elevated BMI are associated with the development of insulin resistance.10, 17 We found that, even after controlling for age and BMI,4 fasting glucose levels, insulin levels, and HOMAIR values were significantly higher in the ADT group than in the other two groups. Furthermore, a significant negative correlation of testosterone with these metabolic parameters suggests that hypogonadism itself leads to insulin resistance and hyperglycemia. Because men in the ADT group continued to have significant insulin resistance and hyperglycemia, even after adjusting for age and BMI, it appears that muscle mass may have a role in maintaining insulin sensitivity in these men. Animal studies have shown that castration in rats leads to the development of insulin resistance at the level of the muscle by inhibition of glycogen synthesis, and this resistance is overcome by testosterone replacement.18 Indeed, testosterone replacement in men results in an improvement in insulin sensitivity by increasing glucose disposal in the muscle.7 Furthermore, testosterone administration in men results in hyperplasia of Type 1 skeletal muscle fibers, which are responsible for glucose uptake.19, 20 Because men in the ADT group had decreased lean body mass compared with the other groups,4 it is conceivable that lower number of these fibers may have resulted in reduced glucose uptake by the muscles. Male hypogonadism also results in an increase in visceral fat, which is a known risk factor in the development of insulin resistance. We did not measure visceral fat in the current study. Therefore, it is conceivable that increased visceral fat in the ADT group also may have contributed to insulin resistance.
Another important finding of this study was that, even after adjusting for BMI, men in the ADT group had higher levels of leptin compared with men in the other two groups, and there was a negative correlation between testosterone levels and leptin levels. This indicates that low testosterone levels were responsible directly for high leptin levels. Indeed, it has been shown that testosterone directly inhibits leptin production.21 Furthermore, a recent longitudinal study showed that the elevation in leptin levels with aging is a direct consequence of decline in testosterone levels and not because of changes in BMI.22
Only a few studies have evaluated the metabolic alterations because of ADT in men with PCa. A recent prospective study of 22 men with PCa who were receiving ADT showed significant increases in insulin levels after 3 months of treatment compared with baseline; however, in that study, there was no significant change in plasma glucose levels.8 Similarly, another prospective, 3-month study showed that ADT resulted in a 63% increase in fasting insulin levels without any changes in fasting glucose levels.23 Those observations suggest that insulin resistance (manifested by hyperinsulinemia) develops within a few months of starting ADT; however, this hyperinsulinemia is sufficient to prevent the development of hyperglycemia. In our study, men in the ADT group not only were insulin-resistant (elevated insulin levels), they had significantly higher levels of fasting glucose than the other two groups. Indeed, 44% of men in the ADT group had fasting glucose levels >126 mg/dL (which is among the criteria for the diagnosis of diabetes mellitus) compared with to 12% and 11% in the non-ADT and control groups, respectively (Fig. 3). The reason for the higher prevalence of hyperglycemia in the ADT group may be related to the longer duration of ADT in our study than in previous reports (mean duration, 45 mos; range, 12–101 mos). Therefore, it may be concluded that insulin resistance develops within a few months after the initiation of ADT, but the resulting hyperinsulinemia prevents the development of hyperglycemia. However, β cells eventually fail to control glucose levels in men who receive prolonged treatment, resulting in hyperglycemia.
Although PCa is a leading cause of death in men, approximately 50% of men who are diagnosed with PCa die of other, unrelated causes. Of these, CVD is one of the most common causes of death. In 1998, a report showed that, after deaths that were attributable directly to PCa (and related complications), CVD was the second leading cause of death (responsible for 27% of deaths).9 It was reported recently that non-PCa-related deaths now exceed PCa-related mortality, with CVD the single most common cause of non-PCa-related deaths.24 In fact, the proportion of deaths from PCa was similar to the proportion of deaths from cardiovascular causes.
Based on the findings of our current study, it is possible that the adverse metabolic profile observed in men who received long-term ADT may be responsible for higher cardiovascular mortality in this population. In addition to these metabolic alterations, ADT also is associated with a decrease in systemic arterial compliance (resulting in an increase in aortic stiffness), which adds further to the cardiovascular risk in this population.23 These findings are in agreement with evidence suggesting that male hypogonadism is associated with higher cardiovascular mortality25 and that testosterone replacement in hypogonadal men results in coronary vasodilation26 and improvements in angina27 and lipid profiles.28 Based on existing evidence, it is conceivable that hypogonadism in the ADT group may have been responsible for the higher cardiovascular mortality.
Our study also provided some insights into the influence of insulin and leptin on the natural history of PCa. Often, after an initial response to ADT, the majority of these tumors become androgen-independent and aggressive.29 Epidemiologic studies have shown that higher serum levels of insulin and insulin resistance are associated with an increased risk of PCa, even after adjustment for BMI, body fat, levels of sex hormone, and insulin-like growth factor 1 (IGF-1).30, 31 Because insulin is a known growth factor, there is a possibility that insulin even may be responsible for the stimulation of PCa cells. Furthermore, insulin may increase the risk of PCa by stimulating IGF-1 synthesis,32 which is known to stimulate growth of the prostate.33 Similarly, high leptin levels in men have been associated with an increased risk of PCa, and in vitro studies have shown that the addition of leptin leads to proliferation of androgen-independent PCa cell lines.34, 35 Leptin may promote the growth and survival of PCa cells by several mechanisms. Leptin up-regulates the signaling of signal transducer and activator of transcription 3,36 which has an antiapoptotic role and is important for the growth and survival of PCa cells.37 Furthermore, it has been shown that leptin promotes angiogenesis,38 which plays a crucial role in PCa metastases.39 This finding is supported by the fact that high-grade PCa lesions on pathology show strong immunoreactivity for leptin receptor.34 Thus, we propose that, although ADT is beneficial initially, by causing hyperinsulinemia and hyperleptinemia, ultimately, it leads to the growth of PCa. Based on these findings, we propose that insulin-sensitizing agents may be of benefit in men who are receiving ADT.
The current study had a few limitations. First, this was a cross-sectional study, and we recommend long-term prospective studies in men who are receiving ADT. Second, the majority of men in the three groups were Caucasians. Future studies should include individuals from diverse ethnic backgrounds to determine whether the risk and degree of these complications vary according to ethnicity. Third, we did not measure visceral fat in this study. Although our sample size was comparable to most studies done in this field,8, 23 it still was relatively small, and future studies should include more men who are receiving ADT. Our study also had several strengths. First, we studied men who had been receiving ADT for a long time (range, 1–9 yrs). Second, we included two different groups of men to compare with the ADT group. Evaluation of the non-ADT group allowed us to account for the influence of PCa on these metabolic parameters, and evaluation of the control group provided us the opportunity to account for the insulin resistance that occurs with normal aging. This provide us with an opportunity to analyze the metabolic changes that were likely to be a direct result of ADT, independent of patient age and disease status (PCa). To the best of our knowledge, this is the only study that has evaluated these metabolic parameters in three such groups.
In conclusion, men with PCa who receive long-term ADT develop insulin resistance and hyperglycemia. This adverse metabolic profile is independent of age and BMI and is a direct result of hypogonadism. These complications of ADT impart an increased cardiovascular risk and may be responsible for the increased cardiovascular mortality seen in men with PCa. Long-term studies are needed to determine the timing of the onset of these metabolic complications and the role of insulin-sensitizing agents, such as thiazolidinediones, in this patient population. In the meantime, we recommend that men with PCa who have received ADT for at least 12 months (the minimum duration of ADT in our patients) should be screened for hyperglycemia.
The authors sincerely thank Dr. Milena Braga-Basaria for her thoughtful comments.