1516 Diabetes Care Volume 41, July 2018

Sandeep Dhindsa,1,2 Husam Ghanim,1

Hypogonadotropic Hypogonadism Manav Batra,1 and Paresh Dandona1

in Men With Diabesity

Diabetes Care 2018;41:1516–1525 | https://doi.org/10.2337/dc17-2510

One-third of men with obesity or type 2 diabetes have subnormal free testosterone
concentrations. The lower free testosterone concentrations are observed in obese
men at all ages, including adolescents at completion of puberty. The gonadotropin
concentrations in these males are inappropriately normal; thus, these patients have
hypogonadotropic hypogonadism (HH). The causative mechanism of diabesity-
induced HH is yet to be defined but is likely multifactorial. Decreased insulin and
leptin signaling in the central nervous system are probably significant contributors.
Contrary to popular belief, estrogen concentrations are lower in men with HH. Men
with diabesity and HH have more fat mass and are more insulin resistant than
eugonadal men. In addition, they have a high prevalence of anemia and higher
mortality rates than eugonadal men. Testosterone replacement therapy results in
a loss of fat mass, gain in lean mass, and increase in insulin sensitivity in men with
diabesity and HH. This is accompanied by an increase in insulin-signaling genes in
adipose tissue and a reduction in inflammatory mediators that interfere with insulin
REVIEW

signaling. There is also an improvement in sexual symptoms, anemia, LDL cho-

lesterol, and lipoprotein (a). However, testosterone therapy does not consistently
affect HbA1c in men with diabetes. The effect of testosterone replacement on
cardiovascular events or mortality in men with diabesity is not known and remains
to be studied in prospective trials.

Approximately one-third of men with obesity, type 2 diabetes, or metabolic syndrome

have subnormal free testosterone concentrations (1–3). Testosterone concentrations
are inversely related to BMI and insulin resistance (3). One-third of obese young males
(14–35 years of age) also have subnormal free testosterone concentrations. The
gonadotropin concentrations in these males are inappropriately normal; thus, these
patients have hypogonadotropic hypogonadism (HH). This review will detail the met-
abolic consequences of HH and the effects of testosterone replacement in men with
obesity, type 2 diabetes, and metabolic syndrome. Because the pathophysiological 1
mechanisms of all of these conditions are closely related, we have decided to refer to Division of Endocrinology, Diabetes and Metab-
olism, State University of New York at Buffalo,
this group of patients as having “diabesity” in the rest of the review. Please note that and Kaleida Health, Buffalo, NY
this term is not limited to only those with the combination of diabetes and obesity. 2
Division of Endocrinology, Diabetes and Metab-
olism, Saint Louis University, St. Louis, MO
MEASUREMENT OF TESTOSTERONE CONCENTRATIONS IN THE OBESE Corresponding author: Paresh Dandona, pdandona@
Prior to a detailed discussion about subnormal testosterone concentrations in kaleidahealth.org.
obesity, it is important to clarify that obesity is associated with a decrease in both total Received 1 December 2017 and accepted 2 April
and free testosterone. Because testosterone is largely bound to sex hormone– 2018.
binding globulin (SHBG) (40–80%) and albumin (20–50%), total testosterone does not © 2018 by the American Diabetes Association.
reflect the bioavailability of testosterone at the cellular level. Insulin-resistant states Readers may use this article as long as the work
is properly cited, the use is educational and not
such as obesity and type 2 diabetes are known to be associated with low SHBG for profit, and the work is not altered. More infor-
concentrations, and, in fact, there are data demonstrating that low SHBG may be a mation is available at http://www.diabetesjournals
marker for the development of diabetes (4). Thus, there is a physiological lowering .org/content/license.
care.diabetesjournals.org Dhindsa and Associates 1517

of total testosterone concentrations in isnotyetclearwhetherfreeorbioavailable related to insulin resistance (2,8,10). Im-

obese men. Hence, free or bioavailable testosterone is a better marker of delivery provement of systemic insulin resistance
(non-SHBG–bound) testosterone mea- oftestosteronemoleculetothetissue.Nor by rosiglitazone leads to a modest in-
surement is essential in obese men to is a good test available to measure tissue crease in testosterone concentrations in
assess the gonadal status. The current androgen action. In the meantime, it is men with type 2 diabetes (11), but with-
gold standard method of measuring free prudent to measure free testosterone by out the restoration of testosterone con-
testosterone involves separation of free equilibrium dialysis/mass spectrometry centrations to normal.
testosterone by equilibrium dialysis and and follow the reference ranges provided
measurement by mass spectrometry (5). by the laboratory (5). Although these ASSOCIATION OF OBESITY AND
Although equilibrium dialysis is a tedious, assays are more widely available in the METABOLIC SYNDROME WITH
expensive, and time-consuming technique, U.S. than those for measurement of LOW FREE TESTOSTERONE
it is readily available at many commercial bioavailable testosterone, they usually Many studies have shown free testos-
laboratories in the U.S. Free testosterone require shipping of the sample by the terone concentrations to be low in the
measured by radioimmunoassay (an assay local laboratory to a central laboratory. obese, especially in those with a BMI
that is still commonly used) is unreliable When ordering the free testosterone $40 kg/m2 (12,13). Zumoff et al. (13)
(5). Bioavailable testosterone can be ac- test, the clinician must specify that free studied 48 healthy men with BMI ranging
curately measured by ammonium sulfate testosterone is to be measured by the from 21 to 95 kg/m2 and collected blood
precipitation. Free and bioavailable tes- above-mentioned technique so that the samples every 20 min for a period of 24 h.
tosterone can also be calculated from sample is adequately processed and not They found that 24-h mean free tes-
SHBG, albumin, and total testosterone measured by radioimmunoassay. tosterone and non-SHBG–bound testos-
concentrations. Although the calcula- terone were related inversely to BMI.
tion equation continues to be refined, ASSOCIATION OF TYPE 2 DIABETES Vermeulen et al. (12) compared 35 obese
this calculated free/bioavailable testos- WITH LOW FREE TESTOSTERONE men (mean BMI 41.1 kg/m2) with 54 lean
terone has been shown to correlate very It has been known for over two decades men. The free testosterone concentra-
well with directly measured free or bio- that males with type 2 diabetes fre- tions were 26% lower in the obese. The
available testosterone and is well suited quently have low total testosterone con- authors also compared LH pulsatility over
for epidemiological and clinical studies centrations (6). These studies did not 12 h in eight obese and lean men and
(5). Recent data indicate that linear mod- report free testosterone or gonadotropin found that the mean integrated LH levels
els used to estimate free and bioavailable concentrations, nor did they place low over12hweresignificantlylowerinobese
testosterone are based on binding affinity testosterone concentrations in a clinically men. Free testosterone concentrations
assumptions that are not entirely accurate relevant context. Subnormal free testos- correlated positively with the sum of LH
and result in a systematic bias. Multistep terone concentrations in association with pulse amplitudes in each individual (12).
ensemble allosteric models of testosterone inappropriately low luteinizing hormone Some recent studies have examined the
binding to SHBG provide estimates of (LH) and follicle-stimulating hormone (FSH) prevalence of hypogonadism in obesity
free testosterone levels that closely match concentrations in men with type 2 di- in a much larger number of men. A study
free testosterone measured by equilibrium abetes were first described by us in from the Netherlands in 160 obese men
dialysis (5). 2004 (1,2). These patients had normal (mean age 58 years) found a 36% prev-
Accuracyofcalculatedfree/bioavailable LH and FSH responses to gonadotropin- alence of HH (14). The largest study in this
testosterone is predicated upon the pre- releasing hormone (GnRH) stimulation regard found that 40% of obese men and
cise measurement of total testosterone. pointing to a hypothalamic defect. These 50% of obese men with diabetes had sub-
Liquid or gas chromatography/tandem abnormalities were independent of the normal free testosterone concentrations
mass spectrometry is the gold standard duration and severity of diabetes (HbA1c). (Fig. 1) (3). Thus, obesity is associated with
method to measure total testosterone (5). MRI in these hypogonadal patients showed a high prevalence of hypogonadism, and
The Endocrine Society has recently con- no abnormality in brain or the pituitary (2). the presence of diabetes adds to that risk
cluded a project to harmonize testoster- Free testosterone concentrations were in- only minimally (1).
one mass spectrometry assays across versely related to BMI. This association of The inverse relation of free testoster-
the U.S. This project calculated that the HH with type 2 diabetes has now been one concentrations with obesity is not
harmonized reference range (2.5–97.5 confirmed in several studies and is present restricted to middle-aged men (9,15,16).
percentiles) for total testosterone was in 25–40% of these men (7,8). We also We compared the testosterone, LH, and
264–916 ng/dL for nonobese men aged found that younger men with type 2 FSH concentrations of 25 lean and 25
19–39 years (5). Further work is now diabetes between the ages of 18 and obese boys in Tanner stage 4 and 5 (aged
needed to harmonize assays that accu- 35 years have a similarly high preva- 14–20 years) (16). The free testosterone
rately measure free or bioavailable tes- lence of HH (9). concentrations (measuredbyequilibrium
tosterone across all laboratories and to HH is relatively rare in type 1 diabetes, dialysis followed by mass spectrometry)
provide validated equations for calcula- unless the patients are obese. Therefore, of obese boys were 40% lower than those
tion of free/bioavailable testosterone. HH is not entirely a function of diabetes or of lean boys. In addition, 40% of obese
Free and bioavailable testosterone con- hyperglycemiaperse (10).Thus,in viewof adolescents had subnormal free testoster-
centrations correlate strongly with each the inverse relationship between BMI one concentrations (defined as ,5th per-
other, and measurement of only one of and testosterone concentrations in both centile of free testosterone concentrations
them is required for clinical purposes. It type 1 and type 2 diabetes, HH is probably of lean adolescents). The gonadotropin
1518 Hypogonadism in Men With Diabesity Diabetes Care Volume 41, July 2018

eye movement sleep episode and remain

at that level until waking. Testosterone
production in men is related to rapid eye
movement sleep, sleep duration, and
sleep architecture (24). Obstructive sleep
apnea is common in men with obesity and
type 2 diabetes. A quarter of men with
diabesity have severe sleep apnea, de-
fined as apnea-hypopnea index of .30/h
(25).However,sleepapneadoesnotseem
to be a determinant of total, free, or
bioavailable testosterone concentrations
independently of BMI (24). Furthermore,
positive airway pressure therapy for sleep
apnea does not increase testosterone
concentrations (26). Thus, obstructive sleep
apnea per se is not a major contributor
to the lowering of free testosterone con-
centrations in obesity.
Figure 1—Prevalence of subnormal free testosterone concentrations in lean, overweight, and Neuronal Insulin Resistance
obese men (based on BMI) with and without diabetes. Mean age 60 years; range 45–96 years. A Obesity is associated with decreased
total of 44% of men with diabetes and 33% of age-matched men without diabetes had subnormal insulin signaling in the central nervous
free testosterone concentrations, respectively. Sample size: 275 lean, 687 overweight, and
489 obese men without diabetes; 36 lean, 135 overweight, and 227 obese men with diabetes.
system (27). This association may be
*P , 0.05 vs. obese men in the same group (with diabetes or without diabetes); #P , 0.05 vs. men bidirectional because increasing insulin
without diabetes. Adapted from data published in Dhindsa et al. (3). delivery to the central nervous system
decreases body weight in men (28).
Neuronal insulin receptor knockout
concentrations of lean and obese boys activity in the obese may potentially mice become obese and insulin resistant
were similar. The free testosterone con- suppress the hypothalamic secretion of (29). Interestingly, these mice also have a
centrationswereinverselyrelatedtoBMI, GnRH (1,20). However, we have shown reduction in LH concentrations by 60–
HOMA of insulin resistance (HOMA-IR), that this widely believed presumption 90% and low testosterone concentra-
and CRP concentrations. Another study is not true (21). Total and free estradiol tions. These animals respond to GnRH
has also shown lower free testosterone concentrations in men with HH are sig- challenge by normal or supranormal re-
concentrations in obese postpubertal nificantly lower than in those without HH lease of LH. In addition, it is known that
adolescents (17). (21) (Fig. 2). Population-based studies the incubation of hypothalamic neurons
The association of diabesity with HH is such as the European Male Ageing Study with insulin results in the facilitation of
not simply reflective of increased fat mass also showed that estradiol concentra- secretion of GnRH (30). However, it is
but also of adverse metabolic health. tions are lower in hypogonadal men as unlikely that the site of action of insulin
Some studies show that free testosterone compared with eugonadal men, regard- on reproductive axis in vivo is the GnRH
concentrations are inversely related to less of whether the hypogonadism is neuron. Isolated knockout of the insulin
HOMA-IR, triglyceride concentrations, primary or secondary (22). Clomiphene receptor in a GnRH neuron does not lead
inflammatory mediators, and subcutane- (estrogen antagonist) and aromatase to a decrease in LH concentrations or in
ous adipocyte cell size, independently inhibitors (which decrease estradiol fertility in either male or female mice
of BMI (16,18). In men with metabolic concentrations) have been shown to in- (31). Although the site (or sites) respon-
syndrome, free testosterone concen- crease testosterone concentrations in sible for hypogonadotropism in neuronal
trations decrease proportionately with obese men with low testosterone. This, insulin receptor knockout mice is not
increasing number of components of however, cannot be taken as evidence known, it is clear that insulin action and
metabolic syndrome (19). of estradiol as the cause of low testos- insulin responsiveness in the brain are
terone in those men. Estradiol has an necessary for the maintenance of the
POSSIBLE PATHOPHYSIOLOGICAL inhibitory effect upon gonadotropin se-
MECHANISMS UNDERLYING HH IN functional integrity of the hypothalamo-
cretion even in a normal physiological hypophyseal-gonadal axis.
DIABESITY
setting (23). Clearly, the suppression of
Estradiol LH, FSH, and testosterone in diabesity Leptin
Because testosterone and androstene- is not induced by circulating estradiol Leptin is known to play a permissive role
dione in the male can be converted to concentrations. in the regulation of reproductive axis.
estradiol and estrone, respectively, through Leptin appears to serve as a signal of energy
the action of aromatase in the mesen- Obstructive Sleep Apnea reserves to regulate the hypothalamo-
chymal cells and preadipocytes of adipose Serum testosterone concentrations be- pituitary-gonadal axis in relation to nutri-
tissue, it has been suggested that excessive gin to increase with the onset of sleep and tional status. Men and women with anorexia
estrogen production due to aromatase in young men they peak at the first rapid nervosa have hypogonadotropism and
care.diabetesjournals.org Dhindsa and Associates 1519

As mentioned above, obese boys have

lower free testosterone concentrations
than lean boys (16). Furthermore, bari-
atric surgery in morbidly obese patients
restores testosterone concentrations to
normal after marked weight loss (40).

Is Testicular Function Directly Altered

in Diabesity?
Human chorionic gonadotropin–induced
testosterone secretion by Leydig cells is
related inversely to insulin sensitivity
among men with varying degrees of
glucose tolerance (41). Decreased tes-
tosterone secretion in insulin-resistant
men after pharmacological stimulation
by high doses of human chorionic gonad-
otropin suggests that the lesion resulting
in HH of diabesity may occur at several
levels of the hypothalamic-pituitary-
gonadal axis. The absence, however, of
an increase in gonadotropin concen-
trations indicates that the primary major
Figure 2—Men with HH have lower free estradiol concentrations than eugonadal men. Estradiol defect in diabesity is at the hypothalamo-
concentrations in men with and without HH were 0.47 (0.35, 0.68) and 0.63 (0.46, 0.77) pg/mL, hypophyseal level (1).
respectively. Adapted from Dhindsa et al. (21). The specific metabolic insult in diabe-
sity that results in hypogonadotropism is
low levels of leptin. Absence of the lept- GnRH and LH secretion in experimental yet to be defined. It is very likely that the
in gene or leptin signaling in humans animals and in vitro (35). CRP concen- cause is multifactorial (Fig. 3). Decreased
results in HH (32). Although there are trations are markedly increased in hypo- insulin and leptin signaling in the central
extremely uncommon forms of human gonadal men as compared with men with nervous system appear to be the leading
obesity due to the lack of the leptin gene, normal testosterone concentrations (6.5 candidates. Direct evidence in humans
almost all obesity in humans is associ- vs. 3.2 mg/L) (8,36). It is thus possible that supporting or disproving this hypothesis
ated with leptin resistance and high inflammatory mediators may contribute is, however, lacking.
leptin concentrations. It is possible to the suppression of the hypothalamo-
that leptin resistance in the hypothala- hypophyseal axis and the syndrome of EFFECT OF WEIGHT LOSS ON
mus or in some other neurons is respon- HH in type 2 diabetes. The presence of TESTOSTERONE CONCENTRATIONS
sible for the hypogonadotropism seen in inflammation may also contribute to in-
Many studies have shown that weight
obesity. sulin resistance because inflammation-
loss increases total testosterone and
related mediators contribute to insulin
Kisspeptin SHBG concentrations (summarized in
resistance (37). These mediators are also
It is now known that kisspeptin, a hypo- Grossmann [42]). On average, 5% weight
known to be increased in obesity (38).
thalamic neuropeptide encoded by the loss increases total testosterone con-
KISS1 gene, and the presence of kisspep- What Comes First: HH or Diabesity? centrations by 50 ng/dL. Bariatric sur-
tin receptors on the GnRH neurons It is generally assumed that the associ- gery in morbidly obese adults has been
(G protein–coupled receptor 54) is obligate ation of low testosterone and obesity is shown to induce a reversal of the hypo-
for the release of GnRH (33). Intravenous bidirectional. Low testosterone predis- gonadal state and the restoration of nor-
injection of kisspeptin increases LH and poses to obesity and vice versa. Indeed, mal testosterone concentrations (40). A
testosterone concentrations in men with there are definite lines of evidence for meta-analysis of 29 studies of men un-
type 2 diabetes and HH (34). This suggests both directions of this association in dergoing bariatric surgery showed that
that the hypothalamic defect in men with different clinical settings. Men who are 64% of the men had subnormal total
HH and type 2 diabetes is either at kisspeptin rendered hypogonadal by GnRH agonist testosterone concentrations prior to
level or proximal to it. Kisspeptin neurons therapy for advanced prostate cancer surgery (40). The prevalence of subnormal
express both leptin and insulin receptors, increase their subcutaneous and visceral free testosterone concentrations was not
thus possibly accumulating evidence of fat mass by 15–20%. Consistent with this, ascertained in the analysis. After bariatric
metabolic health and translating it into re- they develop insulin resistance, and their surgery, total testosterone and SHBG
productive health. risk of incident type 2 diabetes increases concentrations increased by 233 ng/dL
by 30% (39). In the other direction, boys and 22 nmol/L, respectively, in a total
Inflammation who are obese prepubertally do not of 439 men across 16 studies. Eight
Tumor necrosis factor-a and interleukin-1b achieve normal testosterone concen- studies calculated free testosterone con-
have been shown to suppress hypothalamic trations at the completion of puberty. centrations in a total of 259 men: there was
1520 Hypogonadism in Men With Diabesity Diabetes Care Volume 41, July 2018

mass. We measured the fat mass and lean

mass by DEXA scans in 138 men with
type 2 diabetes. Men with HH had a
higher BMI (by 3 to 4 kg/m2), 12% more
subcutaneous fat mass, and higher waist-
to-hip ratio (2,7,48) as compared with
eugonadal men. Men with HH also have
higher visceral and subcutaneous fat
content as compared with eugonadal
men (44). After testosterone replace-
ment, there is a decrease in fat mass
and increase in lean mass (Table 1). We
randomized 44 men with type 2 diabetes
and HH to intramuscular therapy with
testosterone or placebo injections for
6 months (44). Men randomized to TRT
lost ;3 kg of subcutaneous fat and gained
;3 kg of lean mass, without a net change
in body weight. There was no change in
visceral or hepatic fat mass after TRT.
Similar to the effects of TRT in men with
type 2 diabetes, a decrease in subcuta-
neous fat and an increase in lean mass
after TRT have been shown to occur in
obese men without diabetes and men
Figure 3—Interplay of different factors in obesity associated HH. The thickness of arrows is
proportional to the strength of available evidence supporting the mechanism. with metabolic syndrome (49–51). The
loss of fat after testosterone therapy is
an increase of 2.2 ng/dL on average. LH and to determine whether the etiology of attributed to aromatization of testoster-
FSH concentrations increased by 1.0 and symptoms is hypogonadism or one of the one to estradiol (52). It is of interest that
1.8 IU/L, respectively. other comorbidities. However, the prev- although caloric reduction induced
It is tempting to speculate that the alence of sexual symptoms is higher in weight loss is due to reduction of both
reduction in aromatase activity due to fat men with HH as compared with eugonadal fat mass and lean mass, weight loss due
loss is the major mechanism underlying men (44). In a randomized trial of trans- to TRT in combination with very low-
the increase in testosterone concentra- dermal testosterone gel or placebo gel calorie diet is exclusively due to fat loss
tions following weight loss. However, as for 1 year in patients with type 2 diabetes (53). TRT also reduces circulating leptin
we have discussed before, elevated es- and hypogonadism, there was an im- concentrations (44).
trogen concentrations are not the cause provement in sexual desire, but other
of obesity-associated HH. Weight loss symptoms did not change (45). Trials Insulin Resistance
may restore neuronal leptin sensitivity of testosterone replacement in hypogo- Many studies have shown that hypogo-
and thus increase gonadotropin release. nadal men with type 2 diabetes have also nadism is associated with insulin resis-
However, exercise, even in the absence shown an improvement in sexual desire tance. Almost two decades ago, Mårin
of weight loss, increases testosterone (44–46). In contrast to consistent effect et al. (49,50,54) demonstrated an im-
concentrations (43). It is possible that in- of testosterone replacement therapy provement in insulin sensitivity (as mea-
sulin sensitization at the hypothalamic level (TRT) on improving libido, its effect on sured by euglycemic clamp) with oral and
mediates the increase in free testoster- erectile dysfunction is modest and in- transdermal T treatment in obese men
one concentrations following weight loss. consistent among studies (44,45). Hy- without diabetes. We have recently
pogonadal men with erectile dysfunction shown that men with type 2 diabetes
CLINICAL CHARACTERISTICS OF are much more likely to have symptomatic and HH are less insulin sensitive than
HH IN MEN WITH DIABESITY AND those without HH and that TRT increases
benefit from phosphodiesterase-5 inhib-
THE EFFECTS OF TESTOSTERONE insulin sensitivity (as measured by glu-
itors than from TRT. Furthermore, addition
TREATMENT cose infusion rate during hyperinsuline-
of TRT to phosphodiesterase-5 inhibitors
Sexual Symptoms does not improve the erectile response mic-euglycemic clamps) in men with HH
It is well accepted that subnormal tes- as compared with phosphodiesterase-5 (Fig. 4) (44). HOMA-IR also decreased
tosterone concentrations are associated inhibitors alone (47). by 40%. Jones et al. (45) showed a 16%
with sexual symptoms in elderly men. decrease in HOMA-IR after 1 year of
Cross-sectional studies have found that Body Composition treatment with transdermal testosterone
approximately two-thirds of men with It is well known that muscle mass is in 220 hypogonadal men with type 2
type 2 diabetes and HH have symptoms of inversely related to testosterone concen- diabetes. However, two trials did not
low libido, erectile dysfunction, or fatigue trations in men and that testosterone show a change in HOMA-IR after replace-
(7). In clinical practice, it is often difficult treatment results in an increase in muscle ment with long-acting intramuscular
care.diabetesjournals.org Dhindsa and Associates 1521

type 2 diabetes with transdermal testos-

Table 1—Effect of TRT on nonsexual parameters in men with diabesity terone also showed a decrease in HbA1c
Parameter Change References from 7.5 to 6.3% over a period of 1 year
Fat mass ↓ (44,49–51,53,54,56,61) (61). This was in conjunction with diet
Lean mass ↑ (44,53,56) and exercise, but no hypoglycemic med-
BMI ↔ (44,45,55,56) ications. However, some randomized trials
Insulin sensitivity ↑ (44,49,50,54) did not show an effect of testosterone
Glycemic control ↔ (44–46,55,56,61) replacement on HbA1c (44,45,55,56). Fur-
Hemoglobin ↑ (44,62) thermore, it is possible that increased
Bone density Not studied
erythropoiesis after testosterone re-
Lipids ↓ in total cholesterol, LDL, and (44,45,56)
placement may affect HbA 1c indepen-
lipoprotein(a),minordecreaseinHDL, dently of changes in glycemia. Studies
↔ in triglycerides with careful collection of self-monitored
Cardiovascular disease Not studied glucose concentrations and measures
PSA ↔ (44,45,51,54) of glycemia other than HbA1c should
be designed to evaluate the effect of
We focused on randomized controlled trials in men with obesity, type 2 diabetes, or metabolic
syndrome. If there was disagreement among studies on a parameter, we chose to depict it as increased testosterone on HbA1c independently of
(↑), decreased (↓), or no change/variable results (↔) based on the totality of evidence. glycemia.
Anemia
Men with HH have a higher prevalence of
testosterone given for 30–40 weeks in also induced a reduction in free fatty acid
normocytic normochromic anemia (38%)
hypogonadal men with type 2 diabetes concentrations, probably through the
as compared with eugonadal men (3%)
(55,56). HOMA-IR may be inadequate as suppression of lipolysis. Free fatty acids
(36). This anemia is associated with an
an index of insulin resistance, especially are known to induce oxidative and in-
increase in hepcidin and a decrease in
in patients with type 2 diabetes, because flammatory stress and interfere with
ferroportin and transferrin receptor ex-
b-cell loss and inadequate insulin se- insulin signal transduction (60). The im-
pression. Testosterone therapy increases
cretion can lead to inappropriately low provement in insulin sensitivity following
erythropoiesis. We found that hemoglobin
insulin concentrations and HOMA-IR. In testosterone treatment was also associ-
and hematocrit concentrations increased
contrast, HOMA-IR is assessed under ated with suppression of other inflam-
by 0.54 g/dL and 2.3%, respectively, after
fasting conditions and therefore a dif- matory mediators that interfere with
6 months of testosterone therapy in men
ferent index of insulin sensitivity than insulin signaling. These included IKKb,
with type 2 diabetes (44). TRT in hypo-
hyperinsulinemic clamp, which largely SOCS-3, and PTEN in mononuclear cells
gonadal men with type 2 diabetes leads
measures glucose uptake by muscle. and TLR-4 and PTP-1B in adipose tissue.
to a significant increase in plasma eryth-
Because TRT increases muscle mass, TRT Although SOCS-3 interferes with insulin
ropoietin concentration and enhances
may have a more prominent effect on signal transduction by causing the ubiq-
the mobilization of iron from its stores
muscle glucose uptake rather than HOMA- uitination and proteasomal degradation
by suppressing hepcidin and increasing
IR. In our study, there was no change in of IRS-1, IKKb induces serine phosphor-
the expression of ferroportin and trans-
glucose uptake during clamps at an ylation of IRS-1 and thus prevents insulin
ferrin on mononuclear cells (62).
early time point of 3 weeks (44). Thus, signal transduction though IRS-1. PTP-1B
it appears that the insulin sensitization dephosphorylates the insulin receptor Expression of Androgen Receptor in HH
of testosterone is not an immediate after the autophosphorylation by tyro- It is intuitive to expect that the state of HH
effect and may be mediated by changes sine kinase and thus limits insulin signal- in diabesity would result in a compen-
in body composition. The results of TRT ing. The expression of PTEN, another satory increase in the expression of an-
on insulin sensitization are inconsistent protein that interferes with insulin signal- drogen receptor. However, we found a
in hypogonadal men who are not obese ing, was also suppressed in mononuclear decreased expression of androgen re-
(57). In addition, TRT replacement in cells following testosterone treatment ceptor in mononuclear cells and adipose
trials composed largely of men with low (44). However, we cannot state whether tissue in men with HH and diabetes as
normal testosterone (instead of subnor- they are the direct actions of testosterone compared with eugonadal men (63). The
mal) concentrations fail to show a change or are indirectly mediated by changes in expression of estrogen receptor-a and
in insulin sensitivity (58). adiposity. aromatase was also reduced in adipose
In parallel with an increase in glucose tissue. Thus, rather than there being a
uptake during clamps, we found a sig- Glycemic Control compensatory increase, there is a de-
nificant increase in the expression of Some studies have evaluated glycemic crease in androgen receptor, estrogen
insulin receptor b, insulin receptor sub- control by measuring HbA1c and fasting receptor,andaromataseexpression.Thus,
strate-1 (IRS-1), Akt-2, and GLUT-4 in glucose after testosterone therapy (Table the hypogonadal state in diabetes is also
adipose tissue. These data are consistent 1). Kapoor et al. (46) showed a decrease in associated with diminished responsive-
with similar effects found after TRT in fasting glucose (28 mg/dL) and HbA1c ness to testosterone and estradiol. TRT
mice (59). This provides a mechanistic (0.37%) as compared with placebo with restored the expression of androgen re-
explanation for the increase in insulin 3 months of testosterone replacement in a ceptor, estrogen receptor, and aromatase
sensitivity. Testosterone administration small trial. A trial in men with new-onset to normal.
1522 Hypogonadism in Men With Diabesity Diabetes Care Volume 41, July 2018

restricted to treatment that achieves

supranormal concentrations of testoster-
one (for example, during abuse of an-
drogens for body building). Testosterone
has not been shown to cause any mean-
ingful change in HDL concentrations in
studies in which testosterone is replaced
to normal levels (5). Some studies have
shown a reduction in LDL or total cho-
lesterol concentrations after TRT in men
with diabetes (45,56). Most studies find
no change or a small decrease in HDL
cholesterol and triglycerides after TRT
(44,45,56).

Cardiovascular Disease
Epidemiological studies have shown that
elderly men with low testosterone are
more likely to die of a major cardiovas-
cular event (65). Inverse association of
mortality with endogenous testosterone
concentrations has also been observed in
men with diabetes (66). However, no
randomized control trials have been con-
ducted to examine the question: “Does
TRT change cardiovascular outcomes in
men?” Cardiovascular outcomes have
been sporadically reported in random-
ized trials of TRT designed for other end
points (such as muscle strength and
glucose control), but these trials were
underpowered to look at cardiac events.
Meta-analyses of these trials do not
find a consistent effect of TRT on car-
diovascular events (67). A recently con-
cluded randomized placebo-controlled
trial in elderly men ($65 years of age,
790 subjects) by Snyder et al. (68) found
no difference in cardiovascular events
between men who received TRT or pla-
cebo for 1 year. A prospective cohort
Figure 4—A: Insulin sensitivity measured by hyperinsulinemic-euglycemic insulin clamps in 94 men
with type 2 diabetes (44 men had HH, and 50 were eugonadal). Men with HH had greater BMI (40 vs. study in an endocrine clinic investigated
34 kg/m2; P , 0.001) but similar age (55 vs. 52 years; P = 0.08) and HbA1c (7.0 vs. 7.1%; P = 0.7) as the effect of TRT in 238 hypogonadal men
compared with eugonadal men. The glucose infusion rate during hyperinsulinemic-euglycemic with HH and type 2 diabetes on all-cause
insulin clamps was 36% lower in men with HH as compared with eugonadal men. The glucose mortality (66). Sixty-four hypogonadal
infusion rate between men with and without HH was, however, not different once adjusted for
men received testosterone (mean du-
BMI difference in the two groups. This suggests that obesity is the predominant determinant of
insulin resistance in men with HH. B: A total of 44 men with HH were randomized to intramuscular ration 42 6 20 months), and 174 men
testosterone or placebo injections every 2 weeks for 6 months. Insulin sensitivity (glucose uptake during were not treated. The mortality rate in
clamps) increased by 32% after 6 months. Data adapted from Dhindsa et al. (44). untreated hypogonadal men was 20%,
whereas hypogonadal men treated with
Bone Density bone density or fractures in men with HH testosterone had a mortality rate of
In epidemiological studies, estradiol con- and diabesity. The effect of TRT on bone 9.4% (P = 0.002). Most retrospective
centrations correlate more robustly with density or fracture rates in men with epidemiological studies have shown a
bone mineral density than testosterone diabesity has also not been studied. benefit on cardiovascular events from
concentrations in men (64). Free testos- long-term testosterone use in elderly
terone concentrations are positively asso- Lipid Profile men (69–71). In one of the largest studies
ciated with bone density in arms, ribs, and It used to be believed that testosterone conducted on this issue, Sharma et al. (70)
lumbar spine in men with type 2 diabetes treatment adversely affects cardiovascu- showed a 56% reduction in total mortality
(48). No study has evaluated the relation lar risk because it lowers HDL cholesterol and 24% reduction in myocardial infarc-
between free estradiol concentrations and concentration. However, that effect is tion with the use of TRT. Some reports
care.diabetesjournals.org Dhindsa and Associates 1523

have shown evidence of harm with short- especially in young men, may partially are visual symptoms, headaches, or other
term testosterone use. However, these explain the high prevalence of infertility pituitary hormone deficits. MRI is also
studieshadanumberofshortcomings(67). or oligospermia in these men. Studies advisable if the free testosterone is very
Large-scale prospective randomized con- comparing sperm parameters of men low (,50% of the lower limit of normal).
trolled trials on testosterone therapy, who are obese or have diabetes with
How Should Men Receiving TRT Be
focusing on cardiovascular benefits and and without HH have not been con-
Followed?
risks, are clearly needed. ducted. TRT cannot be used in those
SHBG concentrations are decreased by
who desire fertility because it decreases
Prostate-Specific Antigen TRT. Hence, free testosterone should be
spermatogenesis.
Men with type 2 diabetes have 20% lower used to titrate testosterone dosing during
prostate-specific antigen (PSA) concen- PRACTICAL CONSIDERATIONS TRT. Hemoglobin and prostate should be
trations than men without diabetes. PSA FOR A CLINICIAN monitored as per guidelines (5).
concentrations are lower in hypogonadal These recommendations are opinions
Should Every Man With Diabesity Have
than in eugonadal men with diabetes based on the authors’ clinical experience.
His Testosterone Concentration
(0.89 vs. 1.1 ng/mL) (72). It is interesting
Checked?
that the incidence of prostatic carcinoma CONCLUSIONS
The answer to this question is yes. The
is lower in men with diabetes. This is in HH is found in 25–33% of men with
high prevalence justifies screening for
contrast to the increased incidence of diabesity. Although the underlying mech-
HH. Limiting testing only to men who
cancer in diabetes in various organs in- anism is not known, neuronal insulin and
report symptoms is likely to result in
cluding the colon, kidney, breast, endo- leptin resistance may play a role at the
false negatives. Hypogonadal patients
metrium, and pancreas. Similar to type 2 hypothalamic level. Low free and bioavail-
may slide gradually into this clinical state
diabetes, obesity is also associated with able testosterone concentrations in these
without any overt symptoms, which may
10–30% lower PSA concentrations and men are associated with an increased
be revealed through direct questioning.
lower prostate cancer incidence (73). prevalence of sexual symptoms, obe-
Asymptomatic men may realize that they
Obese men are half as likely to have PSA sity, highCRPconcentrations,mildanemia,
had been symptomatic only after a trial
concentrations .4 mg/L. The lower PSA insulin resistance, and decreased bone
with testosterone. However, a complete
concentrations may not be a result of mineral density. In addition, these men
discussion of risks and benefits between
lower testosterone concentrations in obe- may have an elevated risk of cardiovascular
thepatientandphysicianshouldprecedea
sity and type 2 diabetes, but may also be events and death. Short-term studies of
trial of TRT, and the decision about treat-
due to the larger plasma volumes and testosterone therapy have demonstrated
ment should be weighed carefully by the
hence hemodilution(74).Prostatecancer an increase in libido, insulin sensitivity, and
physician.
progression and mortality, however, are lean body mass and a reduction in inflam-
increased in obese men, possibly related What Test Should Be Used to Make a mation and fat mass. Men on TRT should be
to later detection due to lower PSA Biochemical Diagnosis of Testosterone monitored for the development of poly-
concentrations (75). Deficiency in a Man With Diabesity? cythemia or prostate complications. Trials
The PSA concentrations are lower in Free testosterone should be checked in of a longer duration are clearly required to
hypogonadal men than in eugonadal men. the morning (fasting) by accurate meth- definitively establish the benefits and risks
TRT may result in a modest increase in PSA odology, preferably equilibrium dialysis/ of TRT in these men.
concentrations (;30%). In most studies mass spectrometry. Studies have shown
of TRT replacement in men with diabesity, that testosterone has a diurnal variation,
there is no change in PSA concentrations and its concentrations also decline after Funding. This work was supported by the
after TRT (44,45,51,54). In this context, eating. Total testosterone should not be National Institute of Diabetes and Digestive
it is important that the replacement of used to make diagnostic or therapeutic and Kidney Diseases (grant R01-DK-075877 to
testosterone in hypogonadal patients in decisions. P.D.) and American Diabetes Association (Junior
Faculty grant 110JF13 to S.D.).
general does not lead to an increased risk Duality of Interest. P.D. is on the speaker panel
What Workup Should Be Undertaken
of prostatic carcinoma, although the trials for and provides research support to AbbVie. No
in a Man With Diabesity and Subnormal
have been too limited in duration and other potential conflicts of interest relevant to
Free Testosterone? this article were reported.
number of patients (5).
Subnormal free testosterone needs to be Author Contributions. S.D. and P.D. wrote the
Spermatogenesis confirmed (at least once). Men with one manuscript and researched data. H.G. contributed
Some (but not all) studies have shown normal value and one low value are not to discussion and reviewed and edited the man-
uscript. M.B. reviewed and edited the manuscript.
that BMI is inversely related to sperm hypogonadal. Studies have generally
counts, sperm morphology, and sperm shown a lack of benefit of TRT in men
References
motility. Data are most consistent about with low normal testosterone concen- 1. Dandona P, Dhindsa S. Update: hypogonado-
the impact of morbid obesity on sper- trations. A man with confirmed hypogo- tropic hypogonadism in type 2 diabetes and
matogenesis. In a meta-analysis of 21 nadism should have LH concentrations obesity. J Clin Endocrinol Metab 2011;96:
studies, BMI .40 kg/m2 was associated measured. The majority have HH. MRI 2643–2651
with a doubling of risk of oligospermia of the pituitary is not beneficial in the 2. Dhindsa S, Prabhakar S, Sethi M, Bandyopadhyay
A, Chaudhuri A, Dandona P. Frequent occurrence
(,40 million spermatozoa per ejaculate) majority of these men due to low in- of hypogonadotropic hypogonadism in type 2
or azoospermia (76). The high prevalence cidence of anatomical pathology. MRI of diabetes. J Clin Endocrinol Metab 2004;89:
of HH in obesity and type 2 diabetes, the pituitary should be done if there 5462–5468
1524 Hypogonadism in Men With Diabesity Diabetes Care Volume 41, July 2018

3. Dhindsa S, Miller MG, McWhirter CL, et al. 19. Corona G, Mannucci E, Schulman C, et al. 35. Watanobe H, Hayakawa Y. Hypothalamic
Testosterone concentrations in diabetic and non- Psychobiologic correlates of the metabolic syn- interleukin-1 beta and tumor necrosis factor-
diabetic obese men. Diabetes Care 2010;33: drome and associated sexual dysfunction. Eur alpha, but not interleukin-6, mediate the endo-
1186–1192 Urol 2006;50:595–604; discussion 604 toxin-induced suppression of the reproductive
4. Ding EL,Song Y, Manson JE, et al.Sex hormone- 20. Pitteloud N, Dwyer AA, DeCruz S, et al. axis in rats. Endocrinology 2003;144:4868–4875
binding globulin and risk of type 2 diabetes in The relative role of gonadal sex steroids and 36. Bhatia V, Chaudhuri A, Tomar R, Dhindsa S,
women and men. N Engl J Med 2009;361:1152– gonadotropin-releasing hormone pulse frequency Ghanim H, Dandona P. Low testosterone and high
1163 in the regulation of follicle-stimulating hormone C-reactive protein concentrations predict low
5. Bhasin S, Brito JP, Cunningham GR, et al. secretion in men. J Clin Endocrinol Metab 2008;93: hematocrit in type 2 diabetes. Diabetes Care
Testosterone therapy in men with hypogonadism: 2686–2692 2006;29:2289–2294
an Endocrine Society clinical practice guideline. 21. Dhindsa S, Furlanetto R, Vora M, Ghanim H, 37. Dandona P, Aljada A, Bandyopadhyay A.
J Clin Endocrinol Metab 2018;103:1715–1744 Chaudhuri A, Dandona P. Low estradiol concen- Inflammation: the link between insulin resis-
6. Barrett-Connor E, Khaw KT, Yen SS. Endoge- trations in men with subnormal testosterone tance, obesity and diabetes. Trends Immunol
nous sex hormone levels in older adult men with concentrations and type 2 diabetes. Diabetes 2004;25:4–7
diabetes mellitus. Am J Epidemiol 1990;132: Care 2011;34:1854–1859 38. Ghanim H, Aljada A, Daoud N, Deopurkar R,
895–901 22. Tajar A, Forti G, O’Neill TW, et al.; EMAS Chaudhuri A, Dandona P. Role of inflammatory
7. Kapoor D, Aldred H, Clark S, Channer KS, Jones Group. Characteristics of secondary, primary, and mediators in the suppression of insulin receptor
TH. Clinical and biochemical assessment of hy- compensated hypogonadism in aging men: ev- phosphorylation in circulating mononuclear cells
pogonadism in men with type 2 diabetes: cor- idence from the European Male Ageing Study. J of obese subjects. Diabetologia 2007;50:278–285
relationswithbioavailabletestosteroneandvisceral Clin Endocrinol Metab 2010;95:1810–1818 39. Grossmann M, Cheung AS, Zajac JD. Andro-
adiposity. Diabetes Care 2007;30:911–917 23. T’Sjoen GG, Giagulli VA, Delva H, Crabbe P, De gens and prostate cancer; pathogenesis and
8. Grossmann M, Thomas MC, Panagiotopoulos Bacquer D, Kaufman JM. Comparative assess- deprivation therapy. Best Pract Res Clin Endo-
S, et al. Low testosterone levels are common and ment in young and elderly men of the gonad- crinol Metab 2013;27:603–616
associated with insulin resistance in men with otropin response to aromatase inhibition. J Clin 40. Escobar-Morreale HF, Santacruz E, Luque-
diabetes. J Clin Endocrinol Metab 2008;93:1834– Endocrinol Metab 2005;90:5717–5722 Ramı́rez M, Botella Carretero JI. Prevalence of
1840 24. Wittert G. The relationship between sleep ‘obesity-associated gonadal dysfunction’ in se-
9. ChandelA,DhindsaS,TopiwalaS,ChaudhuriA, disorders and testosterone in men. Asian J Androl verely obese men and women and its resolution
Dandona P. Testosterone concentration in young 2014;16:262–265 after bariatric surgery: a systematic review and
patients with diabetes. Diabetes Care 2008;31: 25. Foster GD, Sanders MH, Millman R, et al.; meta-analysis. Hum Reprod Update 2017;23:
2013–2017 Sleep AHEAD Research Group. Obstructive sleep 390–408
10. Tomar R, Dhindsa S, Chaudhuri A, Mohanty P, apnea among obese patients with type 2 diabetes. 41. Pitteloud N, Hardin M, Dwyer AA, et al.
Garg R, Dandona P. Contrasting testosterone Diabetes Care 2009;32:1017–1019 Increasing insulin resistance is associated with
concentrations in type 1 and type 2 diabetes. 26. Zhang XB, Jiang XT, Du YP, Yuan YT, Chen B. a decrease in Leydig cell testosterone secretion
Diabetes Care 2006;29:1120–1122 Efficacy of continuous positive airway pressure inmen.J ClinEndocrinolMetab2005;90:2636–2641
11. Kapoor D, Channer KS, Jones TH. Rosiglita- on testosterone in men with obstructive sleep apnea: 42. Grossmann M. Low testosterone in men
zone increases bioactive testosterone and re- a meta-analysis. PLoS One 2014;9:e115033 with type 2 diabetes: significance and treatment.
duces waist circumference in hypogonadal men 27. KullmannS,HeniM,HallschmidM,FritscheA, J Clin Endocrinol Metab 2011;96:2341–2353
with type 2 diabetes. Diab Vasc Dis Res 2008;5: Preissl H, Häring HU. Brain insulin resistance at the 43. Kraemer WJ, Ratamess NA. Hormonal re-
135–137 crossroads of metabolic and cognitive disorders in sponses and adaptations to resistance exercise
12. Vermeulen A, Kaufman JM, Deslypere JP, humans. Physiol Rev 2016;96:1169–1209 and training. Sports Med 2005;35:339–361
Thomas G. Attenuated luteinizing hormone (LH) 28. Hallschmid M, Benedict C, Schultes B, Fehm 44. Dhindsa S, Ghanim H, Batra M, et al. Insulin
pulse amplitude but normal LH pulse frequency, HL,BornJ,KernW.Intranasalinsulinreducesbody resistance and inflammation in hypogonado-
and its relation to plasma androgens in hypo- fat in men but not in women. Diabetes 2004;53: tropic hypogonadism and their reduction after
gonadism of obese men. J Clin Endocrinol Metab 3024–3029 testosterone replacement in men with type 2
1993;76:1140–1146 29. Brüning JC, Gautam D, Burks DJ, et al. Role of diabetes. Diabetes Care 2016;39:82–91
13. Zumoff B, Strain GW, Miller LK, et al. Plasma brain insulin receptor in control of body weight 45. Jones TH, Arver S, Behre HM, et al.; TIMES2
free and non-sex-hormone-binding-globulin-bound and reproduction. Science 2000;289:2122–2125 Investigators. Testosterone replacement in hypo-
testosterone are decreased in obese men in pro- 30. Salvi R, Castillo E, Voirol MJ, et al. Gonad- gonadal men with type 2 diabetes and/or met-
portion to their degree of obesity. J Clin Endocrinol otropin-releasing hormone-expressing neurons abolicsyndrome(theTIMES2study).DiabetesCare
Metab 1990;71:929–931 immortalized conditionally are activated by in- 2011;34:828–837
14. Hofstra J, Loves S, van Wageningen B, sulin: implication of the mitogen-activated pro- 46. Kapoor D, Goodwin E, Channer KS, Jones TH.
Ruinemans-Koerts J, Jansen I, de Boer H. High tein kinase pathway. Endocrinology 2006;147: Testosterone replacement therapy improves in-
prevalence of hypogonadotropic hypogonadism 816–826 sulin resistance, glycaemic control, visceral adi-
inmenreferred for obesitytreatment.Neth J Med 31. Divall SA, Williams TR, Carver SE, et al. Di- posity and hypercholesterolaemia in hypogonadal
2008;66:103–109 vergent roles of growth factors in the GnRH men with type 2 diabetes. Eur J Endocrinol 2006;
15. Nielsen TL, Hagen C, Wraae K, et al. Visceral regulation of puberty in mice. J Clin Invest 2010; 154:899–906
and subcutaneous adipose tissue assessed by 120:2900–2909 47. Shabsigh R, Kaufman JM, Steidle C, Padma-
magnetic resonance imaging in relation to circu- 32. Farooqi IS, Matarese G, Lord GM, et al. Nathan H. Randomized study of testosterone gel
lating androgens, sex hormone-binding globulin, Beneficial effects of leptin on obesity, T cell hypo- as adjunctive therapy to sildenafil in hypogonadal
and luteinizing hormone in young men. J Clin responsiveness, and neuroendocrine/metabolic men with erectile dysfunction who do not re-
Endocrinol Metab 2007;92:2696–2705 dysfunction of human congenital leptin deficiency. spond to sildenafil alone. J Urol 2004;172:658–
16. Mogri M, Dhindsa S, Quattrin T, Ghanim H, J Clin Invest 2002;110:1093–1103 663
Dandona P. Testosterone concentrations in 33. Seminara SB, Messager S, Chatzidaki EE, et al. 48. Dhindsa S, Bhatia V, Dhindsa G, Chaudhuri A,
young pubertal and post-pubertal obese males. The GPR54 gene as a regulator of puberty. N Engl J Gollapudi GM, Dandona P. The effects of
Clin Endocrinol (Oxf) 2013;78:593–599 Med 2003;349:1614–1627 hypogonadism on body composition and bone
17. VandewalleS,Taes Y, Fiers T, et al. Sex steroids 34. George JT, Veldhuis JD, Tena-Sempere M, mineral density in type 2 diabetic patients. Di-
in relation to sexual and skeletal maturation in Millar RP, Anderson RA. Exploring the patho- abetes Care 2007;30:1860–1861
obese male adolescents. J Clin Endocrinol Metab physiology of hypogonadism in men with type 2 49. Mårin P, Holmäng S, Jönsson L, et al. The
2014;99:2977–2985 diabetes: kisspeptin-10 stimulates serum testos- effects of testosterone treatment on body com-
18. Bekaert M, Van Nieuwenhove Y, Calders P, terone and LH secretion in men with type 2 position and metabolism in middle-aged obese
et al. Determinants of testosterone levels in diabetes and mild biochemical hypogonadism. men. Int J Obes Relat Metab Disord 1992;16:991–
human male obesity. Endocrine 2015;50:202–211 Clin Endocrinol (Oxf) 2013;79:100–104 997
care.diabetesjournals.org Dhindsa and Associates 1525

50. Mårin P, Krotkiewski M, Björntorp P. Andro- levels. J Clin Endocrinol Metab. 24 January 2018 Association of Clinical Endocrinologists and
gen treatment of middle-aged, obese men: ef- [Epub ahead of print]. DOI: 10.1210/jc.2017-02545 American College of Endocrinology position
fects on metabolism, muscle and adipose tissues. 59. Kelly DM, Jones TH. Testosterone and obe- statement on the association of testosterone
Eur J Med 1992;1:329–336 sity. Obes Rev 2015;16:581–606 and cardiovascular risk. Endocr Pract 2015;21:
51. Kalinchenko SY, Tishova YA, Mskhalaya GJ, 60. Tripathy D, Mohanty P, Dhindsa S, et al. 1066–1073
Gooren LJ, Giltay EJ, Saad F. Effects of testos- Elevation of free fatty acids induces inflammation 68. Snyder PJ, Bhasin S, Cunningham GR, et al.;
terone supplementation on markers of the meta- andimpairsvascularreactivityinhealthysubjects. Testosterone Trials Investigators. Effects of tes-
bolic syndrome and inflammation in hypogonadal Diabetes 2003;52:2882–2887 tosterone treatment in older men. N Engl J Med
men with the metabolic syndrome: the double- 61. Heufelder AE, Saad F, Bunck MC, Gooren L. 2016;374:611–624
blinded placebo-controlled Moscow study. Clin Fifty-two-week treatment with diet and exercise 69. Cheetham TC, An J, Jacobsen SJ, et al. As-
Endocrinol (Oxf) 2010;73:602–612 plus transdermal testosterone reverses the met- sociation of testosterone replacement with car-
52. Finkelstein JS, Lee H, Burnett-Bowie SA, et al. abolicsyndromeandimprovesglycemiccontrolin diovascularoutcomes among men with androgen
Gonadal steroids and body composition, strength, men with newly diagnosed type 2 diabetes and deficiency. JAMA Intern Med 2017;177:491–499
and sexual function in men. N Engl J Med 2013;369: subnormal plasma testosterone. J Androl 2009; 70. Sharma R, Oni OA, Gupta K, et al. Normal-
1011–1022 30:726–733 ization of testosterone level is associated with
53. NgTangFuiM,PrendergastLA,DupuisP,etal. 62. Dhindsa S, Ghanim H, Batra M, et al. Effect of reduced incidence of myocardial infarction and
Effects of testosterone treatment on body fat and testosterone on hepcidin, ferroportin, ferritin and mortality in men. Eur Heart J 2015;36:2706–2715
lean mass in obese men on a hypocaloric diet: iron binding capacity in patients with hypogona- 71. Corona G, Maseroli E, Rastrelli G, et al.
a randomised controlled trial. BMC Med 2016; dotropic hypogonadism and type 2 diabetes. Clin Cardiovascular risk associated with testoster-
14:153 Endocrinol (Oxf) 2016;85:772–780 one-boosting medications: a systematic review
54. Mårin P, Holmäng S, Gustafsson C, et al. 63. Ghanim H, Dhindsa S, Abuaysheh S, et al. and meta-analysis. Expert Opin Drug Saf 2014;13:
Androgen treatment of abdominally obese men. Diminishedandrogenandestrogenreceptorsand 1327–1351
Obes Res 1993;1:245–251 aromatase levels in hypogonadal diabetic men: 72. Dhindsa S, Upadhyay M, Viswanathan P,
55. Hackett G,Cole N, BhartiaM, KennedyD, Raju reversal with testosterone. Eur J Endocrinol 2018; Howard S, Chaudhuri A, Dandona P. Relationship
J, Wilkinson P; BLAST Study Group. Testosterone 178:277–283 of prostate-specific antigen to age and testos-
replacement therapy improves metabolic param- 64. Khosla S, Melton LJ III, Atkinson EJ, O’Fallon terone in men with type 2 diabetes mellitus.
eters in hypogonadal men with type 2 diabetes WM, Klee GG, Riggs BL. Relationship of serum sex Endocr Pract 2008;14:1000–1005
but not in men with coexisting depression: the steroid levels and bone turnover markers with 73. Werny DM, Thompson T, Saraiya M, et al.
BLAST study. J Sex Med 2014;11:840–856 bone mineral density in men and women: a key Obesity is negatively associated with prostate-
56. Gianatti EJ, Dupuis P, Hoermann R, et al. role for bioavailable estrogen. J Clin Endocrinol specific antigen in U.S. men, 2001-2004. Cancer
Effect of testosterone treatment on glucose Metab 1998;83:2266–2274 Epidemiol Biomarkers Prev 2007;16:70–76
metabolism in men with type 2 diabetes: a ran- 65. Araujo AB, Kupelian V, Page ST, Handelsman 74. Ba~nez LL, Hamilton RJ, Partin AW, et al.
domized controlled trial. Diabetes Care 2014;37: DJ, Bremner WJ, McKinlay JB. Sex steroids and all- Obesity-related plasma hemodilution and PSA
2098–2107 cause and cause-specific mortality in men. Arch concentration among men with prostate cancer.
57. Magnussen LV, Glintborg D, Hermann P, Intern Med 2007;167:1252–1260 JAMA 2007;298:2275–2280
Hougaard DM, Højlund K, Andersen M. Effect 66. Muraleedharan V, Marsh H, Kapoor D, Channer 75. Calle EE, Rodriguez C, Walker-Thurmond K,
of testosterone on insulin sensitivity, oxidative KS, Jones TH. Testosterone deficiency is associated ThunMJ.Overweight,obesity,andmortalityfrom
metabolism and body composition in aging men with increased risk of mortality and testosterone cancer in a prospectively studied cohort of U.S.
with type 2 diabetes on metformin monotherapy. replacement improves survival in men with type 2 adults. N Engl J Med 2003;348:1625–1638
Diabetes Obes Metab 2016;18:980–989 diabetes. Eur J Endocrinol 2013;169:725–733 76. Sermondade N, Faure C, Fezeu L, et al. BMI in
58. Huang G, Pencina KM, Li Z, et al. Long-term 67. Goodman N, Guay A, Dandona P, Dhindsa S, relation to sperm count: an updated systematic
testosterone administration on insulin sensitivity Faiman C, Cunningham GR; AACE Reproductive review and collaborative meta-analysis. Hum
in older men with low or low-normal testosterone Endocrinology Scientific Committee. American Reprod Update 2013;19:221–231