Prevalence of low testosterone levels in men with type 2 diabetes mellitus: a cross-sectional study.

BACKGROUND: A high prevalence of low serum testosterone (LST) in men with type 2 diabetes have been reported worldwide. The aim of this study was to determine the prevalence and associated factors of LST in men with type 2 diabetes.
MATERIALS AND METHODS: This was a cross-sectional study, conducted among 1,089 men (aged 30-70 years) with type 2 diabetes who consecutively attended a major diabetes center in Amman, Jordan, between August 2008 and February 2009. The patients' demographic characteristics were collected using a prestructured questionnaire. Duration of diabetes, smoking habits, presence of retinopathy, neuropathy, and nephropathy were collected from the medical records. All participants were asked to complete the Androgen Deficiency in Ageing Male (ADAM) questionnaire. Venous blood sample was collected to test for total testosterone (TT), free testosterone (FT), sex hormone binding globulin (SHBG), follicle-stimulating hormone (FSH), luteinizing hormone (LH), prolactin (PRL), serum lipids, and glycosylated hemoglobin (HbA1c). LST was defined as TT <3 ng/ml.
RESULTS: Overall, 36.5% of patients with diabetes had TT level <3 ng/ml and 29% had symptoms of androgen deficiency. Of those with serum testosterone level <3 ng/ml, 80.2% had symptoms of androgen deficiency, 16.9% had primary hypogonadism (HG), and 83.1% had secondary HG. Univariate analysis showed a significant relationship between age, income, education, body mass index (BMI), smoking, duration of diabetes, diabetic nephropathy, diabetic neuropathy, and HbA1c. Multivariate logistic regression analysis indicated age, income, BMI, and diabetic neuropathy as the independent risk factors of LST.
CONCLUSIONS: The prevalence of LST among men with type 2 diabetes is high. Age, income, BMI, and diabetic neuropathy were found to be the independent risk factors for LST.

INTRODUCTION

Diabetes mellitus (DM) affects an estimated 285 million people worldwide. This number is expected to reach 438 million by the year 2030, with two-thirds of all cases occurring in low- to middle-income countries.[1] Asians develop diabetes at a younger age, at lower degrees of obesity, and at much higher rates given the same amount of weight gain compared with Western populations.[2] The prevalence of type 2 diabetes is high in Jordan and more than half of the patients with diabetes have unsatisfactory metabolic control.[3]

The association between low serum testosterone (LST) and DM has recently received substantial attention.[45] Studies have reported that men with type 2 DM (T2DM) have a high prevalence of LST.[678] Further, reduced total testosterone (TT) levels have been associated with insulin resistance and subsequent risk for developing T2DM.[910] The main symptoms of LST are reduced libido/erectile dysfunction, reduced muscle mass and strength, increased adiposity, osteoporosis/low bone mass, depressed mood, fatigue, low energy, and impaired quality of life.[1112]

Researchers have highlighted the potential metabolic consequences of testosterone decline on age-associated metabolic changes such as abdominal obesity, diabetes, and markers of prediabetes.[131415] hypogonadism (HG) is a clinical condition consisting of both symptoms and biochemical signs of testosterone deficiency. However, many studies in men with diabetes have defined LST solely on the basis of testosterone levels. Symptoms of HG have rarely been considered in combination with biochemical testosterone deficiency.[16] There is a growing interest in understanding the concurrence of symptoms of low testosterone and a low testosterone level since the clinical significance of a low testosterone level alone is unclear.[17] TT concentrations are determined, to a large extent, by the circulating sex hormone binding globulin (SHBG) concentrations. In the blood of normal men, 44% of TT is bound to SHBG, 2% is unbound (free testosterone (FT)), and 54% circulates bound to albumin and other proteins.[18] It is not known whether the lower testosterone levels in diabetics are associated with changes in luteinizing hormone (LH) and follicular stimulating hormone (FSH). Previously published data show that the commonest form of gonadal dysfunction was hypogonadotropic HG.[19] Ando et al.,[20] reported low TT and normal LH levels in diabetics; whereas, Ali et al.,[21] found that subjects with diabetic neuropathy had low testosterone, high LH and FSH levels. The prevalence of LST is largely unknown both in the diabetic population and in the general population in Arab countries, including Jordan. This study was conducted to determine the prevalence of LST in men with T2DM and its associated factors.

MATERIALS AND METHODS

A total of 1,089 men (aged 30-70 years) with type 2 diabetes who consecutively attended at a major diabetes center in Amman, Jordan, between August 2008 and February 2009 were included in the study. Patients with any inflammatory disease or infection and already receiving hormone replacement therapy were excluded. Informed verbal consent was obtained from all the patients to participate in the study, and the local research ethics committee approved the protocol. The sample size of 1,089 yielded a power of more than 80% at the confidence level of 95%, and a 5% margin of error.

The respondents were purposely and conveniently selected according to their availability during their routine visit to the outpatient clinics. The patients’ demographic characteristics were collected using a prestructured questionnaire. Information on the duration of diabetes, presence of retinopathy, neuropathy, and nephropathy was collected from their medical records. Smoking habits were assessed by dividing men into categories of current, former, and never smokers. Study participants were asked to complete the androgen deficiency in ageing male (ADAM) questionnaire. This questionnaire has 88% sensitivity and 60% specificity.[22] A positive response is based on a decrease in libido or the strength of erections, or any three nonspecific questions that may include a decrease in muscle strength, fatigability, mood changes, and loss of height.

Body mass index (BMI) was computed by dividing the weight in kilograms by the square of height in meters. Blood pressure was measured using a standardized sphygmomanometer. A trained nurse performed the procedure while the subject was in a sitting position, with the arm at the level of the heart and after 5 min rest. Hypertension was defined as elevated systolic (≥140 mmHg) or diastolic (≥90 mmHg) blood pressure. A venous blood sample (20 ml) was drawn between 8:00 and 10:00 am after an overnight fast. Blood was withdrawn from the cubital fossa and/or dorsum of the hand veins from each participant, using a disposable syringe. It was injected into the plain complete blood count tube; the specimen immediately centrifuged, and serum was aliquoted and stored at –20°C to determine TT, FT, SHBG, FSH, LH, prolactin (PRL), glycosylated hemoglobin (HbA1c), total cholesterol, HDL cholesterol, LDL cholesterol, and triglyceride. Laboratory technicians were blinded to participants’ characteristics. TT was assessed using radioimmunoassay. FT was assessed using Axsym, which is based on a microparticle enzyme immunoassay.[16] SHBG was tested by means of an immunochemiluminometric assay. LH, FSH, and PRL were measured by chemiluminescent immunometric assays. HbA1c was analyzed using high performance liquid chromatography (HPLC) method (Bio-Rad). Information concerning HbA1c was adopted from tests made in the National Center for Diabetes, Endocrinology, and Genetics lab. Total cholesterol, triglyceride, HDL, and LDL were assayed through the automated spectrophotometer, enzymatic colorimetric method, COBAS INTEGRA using commercial kits supplied by Roche Diagnostics.

In this study, LST was defined as TT <3.0 ng/ml and TT used as reference parameter to define LST.[82324] Symptomatic androgen deficiency was defined as TT <3.0 ng/ml in addition to a positive response to ADAM questionnaire. Primary HG was defined as LH >10 MIU/ml with TT <3.0 ng/ml, while the secondary HG was defined as LH <2 MIU/ml with TT <3.0 ng/ml.[16]

The diagnosis of DM was based on the American Diabetes Association (ADA) criteria.[25] Nephropathy as diagnosed by the nephrologists (the presence of micro albuminuria which is defined as, the presence of micro albuminuria ≥30-299 mg/24 h urine collection sample or the presence of macro albuminuria ≥300 mg/24 h urine collection sample)[26] was obtained from the patient’s record. The diabetic neuropathies are heterogeneous, affect different parts of the nervous system, and present with diverse clinical manifestations. They may be focal or diffuse.[27] Retinopathy was defined according to the American Academy of Ophthalmology (AAO).[28] Patients were classified into nonproliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR).

Overweight was defined as BMI 25-29.9 kg/m2 , and obesity was defined as BMI ≥30 kg/m2 . Among type 2 diabetic patients, lipid levels were considered abnormal according to ADA criteria. Hypercholesterolemia referred to a total cholesterol level ≥200 mg/dl. HDL was considered low when the level was <40 mg/dl. LDL was considered high when the level ≥100 mg/dl. Hypertriglyceridemia was considered high when TG level was ≥150 mg/dl. Dyslipidemia was considered present when one or more of the previous abnormalities were found in serum lipids, or if the patient was receiving medication for any of the above conditions. Patients with HbA1c <7% were considered controlled.[25]

Statistical analysis was conducted using the Statistical Package for Social Sciences (SPSS version 15). In addition to descriptive analysis, Chi-square tests or independent t-test were done to find out associations. Multivariate logistic regression was carried out to assess the factors associated with low testosterone level. P < 0.05 was considered statistically significant.

RESULTS

This study included a total of 1,089 patients with type 2 diabetes, aged between 30 and 70 years with a mean age of 52 ± 11.2 years. The sociodemographic and clinical characteristics of participants are shown in Table 1. The mean BMI was 29.9 ± 5.2 kg/m2 and 43.8% of diabetics were obese, and 39.3% overweight. About 28.7% had diabetes for more than 10 years. More than half of the patients (58.7%) were on a combination of oral antidiabetic agents and insulin. Most patients (88.3%) were on statins. Two-thirds of the diabetic population (69.2%) was on antihypertensive treatment.

Table 1

Demographic, clinical, and relevant characteristics of study participants (n=1,089)

Table 2 shows the variations in the mean serum gonadal hormone levels by different age groups. It seems that total and FT levels decreased with increasing age. Overall, 36.5% of patients with diabetes had TT level <3 ng/ml and 29% of study participants had symptoms of androgen deficiency. Table 3 indicates the prevalence of LST level (TT <3 ng/ml) by groups of different variables. The prevalence of low total serum testosterone was 45.4% in age group 60-70 years, 37.8% in age group 50-59, and 34.3% in age group 40-49, and 19.1 in age group 30-39 [Table 3]. Of those with serum testosterone level <3 ng/ml, 80.2% had symptoms of androgen deficiency, about 16.9% had primary HG, and 83.1% had secondary HG. The univariate analyses indicated significant differences among the age groups (P = 0.005), income (P = 0.005), education (P = 0.005), BMI (P = 0.004), smoking (P = 0.004), duration of diabetes (P = 0.005), diabetic nephropathy (P = 0.005), diabetic neuropathy (P = 0.005), and HbA1c (P = 0.009).

Table 2

The mean (SD) values of serum gonadal hormones level among patients with diabetes according to age group

Table 3

Prevalence of low serum testosterone level (total testosterone <3 ng/ml) for patients with type 2 diabetes mellitus according to relevant characteristics

Multivariate analysis indicated age, income, BMI, and neuropathy as the factors significantly associated with a low total serum testosterone level [Table 4]. Compared to 30-39 years age group, patients aged 40-49 years (Odds ratio (OR = 1.89), 50-59 years (OR = 1.96), and 60-69 years (OR = 2.57) were more likely to have a low level of TT. Those who had a monthly income of less than 500 JD were more likely to have low TT levels (OR = 1.76) than those who had a monthly income of more than, or equal to 500 JD. Compared to patients with obesity, those who had BMI <25 kg/m2 , and those who were overweight; were less likely to have a low level of TT. Diabetic neuropathy increased the odds of a low TT level.

Table 4

Multivariate analysis of factors associated with low testosterone level

DISCUSSION

Reports that have clearly established that LST is common and that at least 25% of men with type 2 diabetes have LST with inappropriately low LH and FSH concentrations are on the increase. Another 4% have subnormal testosterone concentrations with higher LH and FSH concentrations.[10,2930] Due to a higher prevalence of low testosterone in diabetics, the possibility that LST might contribute to diabetes-related sexual dysfunction has recently been reevaluated.[31] The present study found that 36.5% of patients with T2DM had a TT level of <3 ng/ml, and 29% had symptoms of androgen deficiency.

The term secondary (hypogonadotropic) HG denotes an inadequate release of gonadotropin-releasing hormone (GnRH) and is characterized by low-normal or low levels of FSH, LH, and testosterone. The present study indicated that TT, FT, and SHBG decreased with age. Several cross-sectional studies and systemic analyses from various countries have reported that type 2 diabetes is associated with LST. However, these studies reported differences in LST levels with the varying ages of participants, cut-off points used to define HG, method of analysis, duration, and complication of diabetes are dissimilar. A study from Australia reported that 43% of type 2 diabetes patients had TT levels <10 nmol/l.[9] In the United Kingdom, a cross-sectional study of 355 men with type 2 diabetic aged >30 showed that 17% had HG with TT <8 nmol/1, and a further 25% had symptoms of HG associated with a TT level between 8 and 12 nmol/l.[16] A recent study from Egypt reported 33.2% HG in type 2 diabetes patients.[5] Another cross-sectional study from Brazil showed that FT and TT levels were subnormal in 46 and 34% of diabetics, respectively.[32] Considering the importance of the situation, The Endocrine Society recommends measuring the testosterone of patients with type 2 diabetes on a routine basis.[2933]

The present study reports that of the LST subjects, 16.9% had primary HG and 83.1% had secondary HG, suggesting that hypogonadotropic HG is the predominant type of HG in our diabetic subjects. This finding is not consistent with the findings revealed by Ali et al.,[21] which found high serum and urinary FSH and LH in diabetics with low total serum and serum FT levels. However, Tenover et al.,[34] found that the majority of hypogonadal men over the age of 60 had low or inappropriately normal LH levels. On the other hand, Chandel et al.,[35] found that LH and FSH concentrations in type 2 diabetic patients with low FT concentrations were in the normal range. Further, a recent study showed that high prevalence of HG and depressive symptoms were found in patients who had been newly diagnosed T2DM.[36]

Both cross-sectional and longitudinal studies have indicated decreased testosterone levels in the older population compared to the younger.[91137] Our study found a significant association between TT and age. A higher prevalence of low TT was seen in men aged between 60 and 70 years, which is in agreement with the previous finding that HG was frequently associated with T2DM, at least in the 6th decade.[31] In this study, there was a lower prevalence in age group 30-39. The frequency of diabetes with HG in those aged less than 40 years could be due to the fact that most diabetic patients in this age group have by type I, and not T2DM.[31]

Surprisingly, in the present study, there was a statistically significant association between monthly income and LST. This finding is consistent with the work done by Hall et al.[38] A possible explanation for this association is that a lower monthly income may function as a marker for poorer access to health, increased stress, adverse health behaviors, and impoverished neighborhood environment. Wong et al.,[39] found that a lower personal income was independently associated with the increased risk of having androgen deficiency.

Many studies have addressed the relationship between BMI and LST of patients with T2DM. The outcomes of these studies are controversial. Some studies have shown a significant association between BMI and LST level.[1540] In contrast, there was a study that reported no relationship between TT and BMI.[19] However, the present study observed a significant association between BMI and LST levels. A recent study that examined the prevalence of LST in obese and diabetic men found that around 44% of diabetic and 33% of age-matched nondiabetic men had subnormal FT concentrations. Forty percent of obese men and 50% of obese diabetic men had subnormal FT concentrations. Thus, obesity is associated with a high prevalence of HG, and the presence of diabetes adds to that risk.[41] However, a study reported that the presence of LST was not entirely dependent upon obesity because 25% of nonobese patients (31% of lean and 21% of overweight) also had HG.[13]

DM is strongly associated with microvascular complications such as retinopathy, nephropathy, and neuropathy resulting in organ and tissue damage in approximately one-third to one-half of people with the disease.[4243] Our study also found that more than one-third of the diabetes subjects had a microvascular complication, and no significant differences were observed in the study on the prevalence of LST among the subjects with retinopathy and nephropathy. However, a significant association was found in diabetic neuropathy subjects. This is in agreement with other studies which found that a low total serum testosterone level was associated with diabetic neuropathic patients.[92144]

The prevalence of LST was higher in subjects with a longer duration of DM. However, multivariate analysis showed no association with duration of diabetes which is in agreement with previous studies.[45]

Most cross-sectional studies have shown a positive association between smoking and total or FT levels.[4647] Further, cessation of smoking decreased the testosterone levels of men followed for idiopathic infertility.[47] The mechanisms by which cigarette smoking may increase testosterone levels are uncertain.[48] Multivariate analysis indicated no significant association between smoking and total or FT levels, which is in agreement with previously published data.[4449]

Research shows that serum testosterone levels have been reported to be lower in men with hypertension.[50] The present study also found a higher prevalence of LST in the diabetic subjects with hypertension. However, total serum testosterone level and hypertension did not indicate any significant association in multivariate analysis. These findings are consistent with those reached by others.[51]

The present study reports that subjects with dyslipidemia had a higher prevalence of LST than subjects without dyslipidemia. However, serum lipids were not significantly associated with low TT in multivariate analysis, which is in agreement with the findings of several other investigators who reported no significant relationship between serum lipids and LST.[4452] In contrast, some other studies showed that lower HDL cholesterol and higher triglyceride levels were significantly correlated with LST in diabetics.[95153]

In the present study, multivariate analysis did not find an association between the serum testosterone level and HbA1c concentration, which is consistent with the results obtained by some studies,[954] while it is in contrast with those by Kapoor et al.[16] Our findings also contradict those of the study undertaken by Fukui et al.,[55] in which TT concentrations correlated positively with HbA1c concentrations.

There are few limitations in the study. Firstly, this was a cross-sectional design, which made it impossible to determine whether diabetes preceded or followed the decline in serum testosterone level. Secondly, FT was assessed using Axsym, which is based on microparticle enzyme immunoassay. Thirdly, the study examined a limited number of risk factors in a single center.

In conclusion, given the large number of individuals with diabetes, the number of diabetic patients with LST is undoubtedly enormous. What is required is an urgent implementation of early universal screening programs, irrespective of the symptoms of androgen deficiency, in order to detect those with a low total serum testosterone level at an early stage and supplement testosterone accordingly. We recommend the screening of all type 2 diabetics for androgen levels in healthcare settings for referral to endocrinologists. Establishing a hormonal base line for patients with T2DM is also vital so that a comparison of the situation could be made in the future.