Hypogonadal symptoms are associated with different serum testosterone thresholds in middle-aged and elderly men

Abstract

OBJECTIVES

To determine the association between hypogonadal symptoms and total serum testosterone levels in middle-aged and elderly men (>40 years of age) and to identify whether there exists a clear-cut discriminatory threshold of total testosterone below which probability of hypogonadal symptoms increases.

METHODS

We retrospectively reviewed the charts of 360 men who presented to an outpatient men’s health clinic with chief complaint of low testosterone. Sexual, psychological and physical symptoms were evaluated using the Androgen deficiency in Aging Male (ADAM) questionnaire. Serum levels of total testosterone were collected on the same day that men completed their ADAM questionnaires. We performed univariate (t-test, chi-square, binary logistic regression) and multivariate analyses (binary logistic regression) to evaluate the total testosterone threshold and the symptoms that predicted a low testosterone level.

RESULTS

A cluster of symptoms: one sexual (decreased libido), one psychological (decreased energy), and three physical (decreased strength/endurance, decreased ability to play sports, and falling asleep after dinner) were most associated with total serum testosterone levels of ≤300ng/dL. The threshold testosterone serum levels that were associated with increased prevalence of these hypogonadal symptoms ranged from 320 to 375ng/dL. On multivariable analysis, age, but not symptoms on the ADAM questionnaire, predicted a total testosterone of less than 300ng/dL.

CONCLUSIONS

A distinct constellation of hypogonadal symptoms exists at various serum testosterone levels. Consequently, identification of the thresholds for specific symptom management will be critical in establishing patient-centered treatment algorithms.

INTRODUCTION

Hypogonadism is a clinical disorder consisting of a myriad of symptoms in the presence of low serum levels of total testosterone – traditionally defined as a solitary threshold value of <300 ng/dL¹. The long list of nonspecific symptoms associated with hypogonadism makes diagnosis particularly difficult. Late-onset hypogonadism exhibits particular challenges since numerous comorbidities and the natural course of aging masks the true nature of the clinical condition. Given that the total testosterone level of 300ng/dL was based on a panel consensus², the uniform application of a single serum threshold to define hypogonadism is not appropriate given the propensity of men to exhibit variable symptomology at different serum testosterone levels³. In attempts to characterize the clinical symptoms associated with low testosterone levels, studies have been conducted in a general population of elderly men⁴ and in patients receiving testosterone therapy⁵. In the absence of direct evidence of the relationship between serum testosterone levels and hypogonadal symptoms, we conducted a study in middle-aged and elderly men who presented with a chief complaint of either “low testosterone” or symptoms of hypogonadism in order to characterize the specific symptom profiles associated with low serum testosterone levels.

PATIENTS AND METHODS

We evaluated a total of 360 men (40–90 years of age) seen consecutively between May 2013 and March 2014 who presented with symptoms of hypogonadism and who had never received testosterone supplementation therapy (TST). All men were surveyed with the Androgen Deficiency in the Aging Male (ADAM) questionnaire on the same day that their testosterone levels (AM blood samples, radioimmunoassay) were measured. Men using testosterone or other androgenic anabolic steroids (AAS) either at the time of the survey or in the 6 weeks prior to the visit were excluded. Men with significant comorbidities such as cancer and end-organ failure were excluded as well. A locally weighted, linear regression model was used to identify threshold levels of testosterone below, which the probability of symptoms increased above the background prevalence of the overall study population. Univariate and multivariable analyses were performed to evaluate factors that would predict a total testosterone ≤300 ng/dL, as well as their relationship to age. Data was analyzed using Microsoft Excel (Microsoft, Redmond, WA) and Minitab16 (Minitab Inc.). All values were reported as mean ± SD and t-tests were used to evaluate differences in means between groups. A p-value ≤0.05 was considered statistically significant.

RESULTS

Of the 360 men, 160 had T ≤300ng/dL, and the remaining 200 men had T >300ng/dL. The mean age of men was 57.1 ± 11.4y; mean total testosterone was 337.8 ± 147.2 ng/dL. Of the 10 hypogonadal symptoms that are part of the ADAM questionnaire, the 5 symptoms differed significantly (p<0.05) in prevalence between men with serum testosterone levels of ≤300ng/dL and men with T >300ng/dL (Table 1). These included one sexual symptom (decreased libido), one psychological symptom (lack of energy), and three physical symptoms (decreased strength and endurance, decreased ability to play sports, and falling asleep after dinner).

Table 1.

Identification and Prevalence of ADAM Symptoms Related to Levels of Total Testosterone

Category	Question	N	Prevalence in men with T < 300 ng/dL	N	Prevalence in men with T >300 ng/dL	p-value
Sexual	Do you have decreased libido?	143	66%	195	52%	0.0107
	Are your erections less strong?	151	78%	200	75%	0.5224
Psychological	Have you noticed a decreased “enjoyment in life”?	144	35%	200	28%	0.1938
	Are you sad and/or grumpy?	144	23%	198	21%	0.6002
	Do you have a lack of energy?	149	50%	199	39%	0.0397
Physical	Do you have a decrease in strength and or endurance?	148	56%	197	39%	0.0021
	Has there been a recent deterioration in your work performance?	156	20%	200	13%	0.085
	Are you falling asleep after dinner?	156	42%	200	29%	0.0143
	Have you noticed a recent deterioration in your ability to play sports?	143	41%	200	29%	0.0153
	Have you lost height?	152	26%	199	20%	0.246

The probability of middle-aged or elderly men expressing symptoms was inversely related to serum testosterone levels (Figure 1). A range of serum testosterone levels (320–375ng/dL), rather than a solitary testosterone level, was found to differentiate between the five, most significantly different symptoms. Specifically, serum testosterone thresholds of 320ng/dL for decreased ability to play sports, 340ng/dL for decreased strength and endurance, 350ng/dL for lack of energy, 360ng/dL for increasingly falling asleep after dinner and 375ng/dl were identified for decreased libido. On a univariate analysis, age, energy, strength/endurance, work performance, falling asleep after dinner, and ability to play sports was associated with a T ≤300ng/dL. Importantly, age was strongly associated with a low testosterone level (Table 2). On multivariable analysis, only age predicted a T ≤300ng/dL (Table 3). Sexual symptoms (decreased libido and poor erectile function) commonly thought to be associated with low testosterone did not predict men with T ≤300ng/dL.

Figure 1.

Prevalence of Hypogonadal Symptoms on the Basis of Levels of Total Testosterone

Table 2.

Univariate binary logistic regression analysis comparing serum testosterone (≤300 and >300ng/dl) to age and ADAM questionnaire responses

	Odds Ratio	95% Confidence Interval	P
Age	0.96	0.94–0.97	<0.001
Do you have decreased libido?	1.14	0.91–2.16	0.121
Do you have a lack of energy?	1.85	1.22–2.82	0.004
Do you have a decrease in strength or endurance?	2.09	1.37–3.20	0.001
Have you noticed a decreased “enjoyment in life”?	1.28	0.83–2.00	0.266
Are you sad and/or grumpy?	1.44	0.91–2.27	0.115
Are your erections less strong?	0.68	0.43–1.07	0.098
Has there been a recent deterioration in your work performance?	2.10	1.25–3.52	0.005
Are you falling asleep after dinner?	1.83	1.20–2.80	0.005
Have you noticed a recent deterioration in your ability to play sports?	1.76	1.14–2.72	0.011
Have you lost height?	0.97	0.56–1.66	0.898

Table 3.

Multivariable binary logistic regression analysis comparing serum testosterone (≤300 and >300ng/dl) to age and ADAM questionnaire responses

	Odds Ratio	95% Confidence Interval	P
Age	0.95	0.93–0.98	<0.001
Do you have a lack of energy?	1.22	0.65–2.30	0.540
Do you have a decrease in strength or endurance?	1.70	0.90–3.24	0.104
Has there been a recent deterioration in your work performance?	1.12	0.56–2.26	0.753
Are you falling asleep after dinner?	1.46	0.85–2.50	0.172
Have you noticed a recent deterioration in your ability to play sports?	0.90	0.47–1.70	0.736

COMMENT

A clinical threshold of 300 ng/dL is often cited in the literature as the biochemical definition of hypogonadism² but it is important to remember that the value was based on panel consensus. When evaluating prevalence of symptoms at the often-used cutoff of 300ng/dL to diagnose biochemical hypogonadism, we found a significant difference between men above and below 300ng/dL of serum testosterone. In our clinical experience, many patients with serum testosterone levels between 300 and 400ng/dL still report hypogonadal symptoms. Due to this discrepancy between standard practice and clinical experience, we chose to assess whether or not any hypogonadal symptoms predicted serum testosterone levels above 300ng/dL in our population. Our findings support our clinical experience that many men with serum testosterone levels between 300 and 400ng/dL can still experience hypogonadal symptoms.

Our finding that multiple, symptom-specific testosterone thresholds exist, supports earlier studies showing that different functional testosterone levels exist for various hypogonadal symptoms^6,7. A large general population of men between ages of 40 and 79 was evaluated in the EMAS (European Male Aging Study)⁴. The EMAS identified erectile dysfunction as the most prevalent symptom in men with T <300 ng/dL. We found that poor erectile function was the most common symptom in our study. Interestingly however, our study showed that the prevalence of erectile dysfunction was similar amongst men with testosterone levels ≤300ng/dL and >300 ng/dL. Therefore, we suggest that although the symptom of erectile dysfunction may be a good indicator of low testosterone in community-dwelling men, it is a poor indicator for identifying low testosterone in men >40 years of age presenting to a men’s health clinic. This finding suggests that appropriate age and population-based questions need to be identified prior to treating hypogonadism.

Previous studies analyzing the association between testosterone threshold and symptoms have varied findings. Wu and colleagues found that only sexual symptoms are associated with testosterone levels⁴. A contrast to these data, Lackner and colleagues found that hypogonadal symptoms correlate with age rather than T levels, and that on univariate analysis there was only a correlation between psychological symptoms and T levels. In this study, no symptoms were predictive of low testosterone on multivariable analysis. Neither of these studies utilized the often used ADAM questionnaire. In our study, on multivariable analysis no question on the ADAM questionnaire predicted testosterone ≤300ng/dl. Importantly, increasing age was the only predictive factor for low T levels. These findings are not surprising given other clinical data. Although testosterone has been shown to be important in erectile function,^2,4 age related factors affecting vasculature are likely more important in age related erectile dysfunction⁸. The interesting finding that a positive mood predicted an increasing age has been shown in multiple studies outside the field of Urology⁹. Despite the associations that have been shown in other studies between hypogonadal symptoms and T levels,^4,6,7 our data suggests that the strength of these relationships at a cut-off of 300ng/dl may non-diagnostic. There exist a proportion of men with low-‘normal’ serum testosterone levels above 300ng/dl that experience hypogonadal symptom suggesting that the commonly used cut-off of 300ng/dl for biochemical hypogonadism may not be appropriate.

Our study has both strengths and limitations. We minimized heterogeneity by surveying men as they were seen consecutively and by excluding men who had received AAS and TST in the previous 6 weeks. Men <40 years of age were excluded for primarily two reasons: 1) The ADAM questionnaire was originally validated in elderly men and is used to screen for adult-onset hypogonadism¹⁰ and 2) Given that TST is prescribed most often for elderly men¹¹, an age cut-off of 40 years provides results that are more generalizable to outpatient settings. We also recorded a single AM serum testosterone level for patients at the same visit that they filled out the ADAM questionnaire. While this provided a good record of the relationship between serum testosterone levels and hypogonadal symptoms, multiple evaluations of T levels and symptoms could have yielded a more robust comparison.

In summary, among men >40 years of age visiting a men’s health clinic, symptoms of decreased libido, decreased energy, decreased strength and endurance, decreased ability to play sports, and falling asleep after dinner were conditions most frequently associated with low total serum testosterone levels. The threshold level of serum testosterone for these hypogonadal symptoms ranged from 320 – 375 ng/dL. Consequently, we propose the existence of unique thresholds at which hypogonadal symptoms become increasingly prevalent. Rather than using solitary, predefined levels of serum testosterone (i.e. <300ng/dL) as cut-points for treatment, we recommend using a variable, yet specific scale based on symptomatology.