Testosterone threshold – does one size fit all?

Abstract

Numerous studies have sought to examine the relationship between hypogonadal symptoms and total serum testosterone levels in men in order to identify a threshold at which symptoms occur and at which treatment is beneficial. However, the search for a discrete threshold may be futile given emerging evidence. Recent studies suggest that testosterone threshold varies by symptoms and among individuals. In addition, thresholds may vary between young and old men. Therefore, initiation of treatment should rely more on symptoms and less on a discrete numerical threshold.

• Hypogonadism
• symptoms
• testosterone

Introduction

Male hypogonadism is a clinical syndrome that results from failure of the testis to produce physiological levels of testosterone due to disruption at one or more levels of the hypothalamic–pituitary–testes axis [1]. Hypogonadism can result not only from disruption of the testes (primary) or the hypothalamus/pituitary (secondary) but also from target organ resistance and age-related decline. In addition, hypogonadism has been shown to occur from certain medication use including glucocorticoids and opioids [2,3]. Hypogonadism is associated with symptoms such as low libido, erectile dysfunction, hot flushes, fatigue, irritability, depressed mood, poor concentration, reduced physical performance and sleep disturbances [1]. We reviewed recent studies that have attempted to identify a specific total testosterone threshold at which the symptoms of androgen deficiency manifest and a threshold at which testosterone supplementation could be beneficial.

Historical evidence for testosterone threshold

A testosterone threshold would be a biochemical level of total serum testosterone at which no hypogonadal symptoms would exist. While a definitive threshold has never been established, the historical threshold for hypogonadism of ∼300 ng/dL (10.4 nmol/L) has been suggested on the basis of a consensus by the Endocrine Society [1].

One of the initial trials of testosterone replacement found that in 15 men with hypogonadism, men with serum testosterone less than 200 ng/dL all benefited from replacement, while only some of those with testosterone levels between 200 and 450 ng/dL benefited [4]. A study of testosterone supplementation in 11 men given a gonadotropin-releasing hormone (GnRH) agonist found that men treated with lower dose testosterone (4 mg/dL) who achieved a mean testosterone level of 10.5 nmol/L (303 ng/dL) had similar erectile function and libido when compared to men receiving a higher dose of testosterone (8 mg/dL) who had a mean testosterone level of 26 g nmol/L (756 ng/dL) [5].

A total testosterone level of 300 ng/dL, currently recommended by the Endocrine society as the threshold for treatment with testosterone therapy, is based on uncontrolled non-randomized studies.

Testosterone threshold in older men

Correlating testosterone levels and hypogonadal symptoms in older men poses a difficult clinical challenge for two primary reasons. First, older men, especially those with comorbidities, may have symptoms similar to those of hypogonadism. For example, with age, men experience decline in bone mineral density, muscle mass, strength and sexual function [6]. Second, testosterone levels decline with age. In general, testosterone levels tend to peak in men between ages 20 and 30 years and gradually decline [7] at a rate of 1–2% annually [8].

Several studies have attempted to further explore hypogonadism in the aging population. The European Male Aging Study (EMAS) by Wu et al. reported that classic symptoms of hypogonadism were not associated with decreased testosterone levels in older men [9]. Interestingly, three sexual symptoms (poor morning erection, low sexual desire and erectile dysfunction) were consistently associated with low testosterone (<11 nmol/L or 317 ng/dL) in older men. However, psychological symptoms had little association with testosterone levels, a finding also supported by other studies [10,11]. On the basis of these results, the authors recommend using the presence of three sexual symptoms with a testosterone level <11 nmol/L as a means of diagnosing late-onset hypogonadism. Using their criteria for diagnosing late-onset hypogonadism, the authors identified an incidence of 2.1%, which is lower than the incidence of biochemical hypogonadism quoted in other studies [8]. Diagnosing testosterone deficiency in an older population is certainly difficult, given the lack of definitive criteria and nonspecific symptoms. The EMAS provides the best evidenced-based criteria to aid in the diagnosis.

Based on several studies, including the EMAS study, Lunenfeld et al. [12] have provided recommendations to aid in the diagnosis and treatment of late-onset hypogonadism in the aging male. Diagnosis requires assessing for signs and symptoms of late-onset hypogonadism including low libido, poor morning erections, depression, poor concentration, sexual dysfunction, impotence and fatigue [12]. Screening questionnaires may help evaluate for presence of symptoms before initiating treatment as well as for monitoring response to treatment, although their specificity is low. The authors recommend obtaining morning testosterone levels with values above 15 nmol/L generally rules out hypogonadism. If the testosterone levels are equivocal, free testosterone levels below 225 pmol/L and the presence of one or more symptoms of late-onset hypogonadism can support treatment initiation.

Testosterone threshold in young men

While many studies have examined the association between testosterone thresholds and symptoms, most have been done in older men and few have explored levels in young men. Recent work from our group identified a total testosterone threshold for younger men (age <40 years) of 400 ng/dL, with psychological and physical symptoms being strongly associated with low testosterone levels [13]. This is in contrast to the previous studies, which identified associations between sexual symptoms and low testosterone in older men [9,14]. This suggests that sexual symptoms, while they may be better predictors of low testosterone in older men, are not necessarily good predictors in young men.

Emerging evidence for use of symptom-specific T cutoffs

Despite historical studies, a testosterone threshold at which symptoms of hypogonadism and adverse health outcomes occur still has not been clearly identified. Recent studies have continued to correlate testosterone levels with symptoms of hypogonadism (Table 1).

Table 1. Testosterone thresholds as suggested by various studies.

Study	Testosterone threshold
Isidori et al. [14]	346 ng/dL
Kelleher et al. [18]	256 ng/dL
EMAS [9]	317 ng/dL
Finkelstein et al. [6]	200 ng/dL
Lackner et al. [19]	None identified

A cross-sectional cohort study of 434 men age >50 years was performed to examine the correlation between psychosomatic complaints and testosterone levels [15]. This study was unable to find a clear-cut threshold for late onset hypogonadal symptoms but reported that the prevalence of psychosomatic and metabolic risk factors gradually accumulated as testosterone levels declined. In addition, it was found that some symptoms occurred more frequently at higher levels of testosterone than other symptoms. For example, loss of libido and vigor tended to occur at testosterone concentrations of 15 nmol/L (432 ng/dL), while depression and diabetes mellitus were more prevalent in men with testosterone concentrations below 10 nmol/L (288 ng/dL). In addition, levels below 230 ng/dL seemed more likely to be associated with erectile dysfunction. This suggests that symptom-specific thresholds may be a better approach to patients rather than a uniform threshold.

Several other studies have led to similar conclusions regarding symptom-specific thresholds. In their randomized double-blinded study of 61 eugonadal men, Bhasin et al. also concluded that that different androgen-dependent processes have different testosterone dose-response relationships [16]. Seftel et al. randomized 406 hypogonadal men (testosterone ≤ 300 ng/dL) to receive 50 mg/d testosterone gel, 100 mg/d testosterone gel, testosterone patch or placebo, and sexual function was assessed at 30 and 90 d [17]. Their threshold levels appeared to be approximately 400 ng/dL for nighttime erections, 500 ng/dL for sexual intercourse and 600 ng/dL for sexual desire.

Perhaps the two best studies examining the testosterone threshold at which male hypogonadism occurs are EMAS [9] and another study by Finkelstein et al. [6]. The EMAS, as previously described, further supported the concept of a symptom-specific testosterone threshold. This study included a random sampling of 3369 community volunteers between the ages 40 and 79 years [9]. This study reported nine symptoms that were related to testosterone level, including three sexual symptoms, three physical symptoms and three psychological symptoms. Various thresholds were identified for specific symptoms. For example, thresholds for poor morning erection, erectile dysfunction and decreased level of vigorous activity were 11 nmol/L (317 ng/dL), 8.5 nmol/L (245 ng/dL) and 13 nmol/L (375 ng/dL), respectively. As previously described, this study also analyzed clusters of symptoms and found that three sexual symptoms were consistently associated with total testosterone levels <11 nmol/dL in aging men. The use of a free testosterone threshold was explored and was found not to strengthen the association between symptoms and levels when total testosterone was below 8 nmol/L. However, the use of free testosterone may be more useful in those with a borderline total testosterone level (8 nmol/L to 11 nmol/L) and multiple symptoms. While this study is limited by the cross-sectional design and the fact that the cohort does not represent those seeking treatment, it does support the notion that different testosterone thresholds exist for different clinical symptoms.

In the study by Finkelstein et al., 198 healthy men (20–50 years of age) had their endogenous testosterone and estradiol production suppressed by injection of goserelin acetate [6]. They were then randomly assigned to receive a placebo gel or testosterone supplementation at four different doses for 16 weeks. They reported that serum testosterone in men who received placebo, 1.25 g, 2.5 g, 5 g or 10 g of testosterone gel daily, corresponded to mean testosterone levels of 44 ± 13, 191 ± 78, 337 ± 173, 470 ± 201 and 805 ± 355 ng/dL, respectively. In this study, men who received placebo, the 1.25 g and 2.5 g doses of testosterone had increases in percentage body fat and subcutaneous fat area, while only the placebo and 1.25 g doses resulted in decreases in lean body mass, thigh-muscle area and erectile function. Leg press strength fell only with placebo administration. It appeared that sexual desire was dose-dependent, and differences were seen in all groups except those who received 2.5 and 5 g doses. The authors also explored the role of using an aromatase inhibitor in order to tease out the differential effects of testosterone and estrogen on symptoms. When aromatization was intact, fat accumulation began around testosterone levels of 350–300 ng/dL. Interestingly, the various measures of strength were not diminished until levels of <200 ng/dL. Sexual function showed a dose-dependent decline. With the use of aromatase inhibitors, the authors were able to attribute fat changes to estradiol, while changes in libido could be attributed to both testosterone and estrogens.

Testosterone thresholds may vary not only with symptoms but also among individuals. A recent study recorded serum testosterone levels in men undergoing testosterone repeated pellet implantation (200 mg) upon return of hypogonadal symptoms [18]. Data collected over five years of reimplantation cycles revealed that when symptoms returned, the blood testosterone levels were highly reproducible within individuals. However, there was significant variation between individuals in the testosterone level at which the symptoms returned (9.2 nmol/L ± 4, p < 0.001). The unique feature of this study was that the cohort of men used depot testosterone that resulted in more stable blood levels that declined gradually, making it easier to identify levels at which symptoms began. This study concluded that hypogonadal men have an individual threshold at which they are symptomatic, but this threshold varies among men. The threshold on an average is at the lower limit of normal for young men (9.2 nmol/L or 265 ng/dL).

In contrast to these previous studies, Lackner et al. were unable to identify symptom-specific testosterone thresholds [19]. While they reported associations between typical symptoms of androgen deficiency and testosterone levels, using multiple logistic regression, they found age to be a significant confounder. Symptoms such as decreased libido and erectile dysfunction were significantly associated with age and not testosterone level.

Taken together, these studies illustrate that there are different serum testosterone thresholds for individual clinical symptoms and individual patients. The reason for variation remains unclear. The changing responses may reflect differing levels of activity, variations in testosterone metabolism or polymorphisms of the androgen receptor [16,20].

Current guidelines

Although there is a lack of consensus regarding testosterone treatment thresholds, many academic societies have published recommendations. The American Urological Association’s most recent guidelines regarding testosterone threshold suggest that the decision to treat should take into account both the presence of symptoms and the serum testosterone measurements; at least until testosterone assays have been optimized [21]. The Endocrine Society recently updated their guidelines stating that while a true serum testosterone threshold is unknown, for most symptoms, the average testosterone threshold corresponded to the lower limit of the normal range for young men, i.e. approximately 300 ng/dL (10.4 nmol/L), with a greater likelihood of having symptoms below this threshold than above it [1]. However, this reference range may vary among laboratories and therefore they recommend using the lower limit of normal as established by the local laboratory. International Society of Andrology, International Society for the Study of Aging Male, European Association of Urology, European Academy of Andrology and American Society of Andrology also recently published guidelines. These recommended against supplementation in men with total testosterone level above 12 nmol/L (350 ng/dL) and for supplementation in men with testosterone below 8 nmol/L (230 ng/dL). However in men between 8 and 12 nmol/L, they recommended repeating the measurement of total testosterone and calculating the free testosterone using sex hormone-binding globulin (SHBG) or equilibrium dialysis [22]. SHBG measurements may in fact be particularly important in the aging male since SHBG levels increase with age, which causes a decrease in bioavailable testosterone [23]. Therefore, if testosterone levels are normal but symptoms of hypogonadism are present, elevated SHBG levels can support the diagnosis. In addition, reliable free testosterone assays are not always readily available, making calculation of bioavailable testosterone using total testosterone and SHBG measurements a reasonable surrogate. While no threshold for free testosterone has been established, expert opinion suggests that a free testosterone level below 65 pg/mL can provide supportive evidence for treatment [22,24].

Conclusion

It is unclear what the testosterone threshold should be, regardless of age. Although historical studies have attempted to identify the levels at which symptoms occur, none have been successful. More recent evidence seems to suggest that the idea of one testosterone threshold may not be clinically useful, given that thresholds vary for symptoms, among individuals and among age groups. In fact, the word “threshold” may be inappropriate given that adverse events tend to occur on a continuum rather than at a strict threshold. Therefore, perhaps a more useful approach may be to treat symptomatic hypogonadal men with testosterone. If the symptoms resolve with testosterone treatment, then therapy can continue. Symptom resolution occurs at various times depending on the type of symptom and the organ system involved [12,25]. For example, improvement in libido may occur between 3 and 6 weeks, while erectile function may take up to six months for noticeable effects [12]. Testosterone therapy should be stopped if symptoms do not improve or are not alleviated. Hence, it is important to clearly document the symptoms for which testosterone therapy is initiated. Strategies such as manipulating SHBG levels to alter the levels of free testosterone and addition of supplements such as DHEA, nandrolone and sermorelin are experimental with no controlled studies to justify their routine use. If symptoms fail to resolve within a reasonable timeframe, then testosterone should be stopped (given the uncertainty of long-term risks) and other etiologies should be explored. For example, chronic illnesses such as type 2 diabetes mellitus, thyroid dysfunction, HIV, metabolic syndrome, depression, chronic obstructive lung disease and drug side-effects may be contributing to symptoms [12].

While treatment of symptomatic hypogonadism is important, physicians should also be wary of over treatment. This has become more of concern, especially given new studies suggesting testosterone replacement therapy may be associated with cardiovascular risk [26–29]. While there are no randomized trials to definitively address the association between testosterone and cardiovascular risk, physicians should be cautious when treating men older than 65 years, particularly those with a cardiac history. In addition, there is the concern that testosterone therapy could stimulate prostate cancer cells, although the evidence is mixed.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

R. R. is a NIH K12 Scholar supported by a Male Reproductive Health Research Career (MHRH) Development Physician-Scientist Award (HD073917-01) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD).