| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RAPID COMMUNICATION |
New England Research Institutes (T.G.T., A.B.A., A.B.O., J.B.M.), Watertown, Massachusetts 02472; and St. Louis University Division Of Geriatric Medicine and St. Louis Veterans Affairs Medical Center (J.E.M.), St. Louis, Missouri 63104
Address all correspondence to: Thomas G. Travison, Ph.D., New England Research Institutes, 9 Galen Street, Watertown, Massachusetts 02472. E-mail: ttravison{at}neriscience.com. Address requests for reprints to: John B. McKinlay, Ph.D., New England Research Institutes, 9 Galen Street, Watertown, Massachusetts 02472. E-mail: jmckinlay{at}neriscience.com.
Abstract
Context: Although it is known that serum testosterone (T) concentrations are related to libido, the strength of that relationship in community-dwelling men has not yet been determined.
Objective: Our objective was to assess the strength and significance of the association between aging men’s self-reports of libido and serum T concentrations.
Design: Our study was a community-based evaluation of men’s health and aging, including three data collection waves: baseline (T1, 1987–1989) and follow-ups (T2, 1995–1997; T3, 2002–2004). Libido was measured on a 14-point scale assessing self-reported frequency of desire and thoughts/fantasies; low libido was defined as a score of less than 7 of 14.
Setting: We conducted an epidemiological study in greater Boston, Massachusetts.
Participants: There were 1632 men aged 40–70 yr at baseline, with follow-up on 922 (56%) at 9 yr (T2) and 623 (38%) at 15 yr (T3).
Main Outcome Measures: We assessed total and calculated bioavailable T .
Results: Three hundred eighteen (19%) subjects reported low libido at baseline. Libido and T displayed a significant association. However, the difference in mean T levels between those subjects with low libido and those without was small; analyses indicated a 3.4 ng/dl (0.12 nmol/liter) increase in total T per unit increase in libido. Subjects reporting low libido exhibited an increased but modest probability of exhibiting low T. Dividing T concentrations by the number of androgen receptor gene cytosine, adenine, guanine repeats did not enhance associations.
Conclusions: Libido and T concentrations are strongly related at the population level. However, the value of individual patient reports of reduced libido as indicators of low T levels is open to question.
REDUCED LIBIDO IS widely considered the most prominent symptomatic reflection of low testosterone (T) levels in men (1, 2), and the correction of low T via supplementation therapy is regarded as a promising treatment for the symptoms of androgen deficiency (AD), including reduced libido. It is unclear, however, how often an individual patient report of low libido indicates truly low T concentrations.
T supplementation has been associated with increases in sexual functioning, mood, and strength (3, 4, 5), but there is yet no consensus regarding its long-term safety and efficacy (6). Nor is the association between naturally occurring T levels and libido completely understood. Although decreased libido is a concern often expressed by aging patients (7), it is difficult to measure comprehensively, being multifactorially determined and associated with both psychosocial and organic factors. Normal and typical T concentrations are also the subjects of substantial debate (8).
The male aging process is itself often accompanied by a decline in sexual functioning and a variety of symptoms reminiscent of those of AD in younger men (9, 10). To what degree these symptoms can be attributed to incipient AD remains unresolved. The specific range of T values that is associated with symptoms may differ by patient (11), and response to supplementation varies with dose, age, and endogenous T levels (12, 13, 14). At the same time, although libido is a leading component of questionnaires used to screen for AD, these have generally displayed only a modest ability to distinguish biochemically hypogonadal from eugonadal men (15, 16, 17).
Both libido and T may be influenced by androgen receptor transactivation, itself associated with the androgen receptor gene polymorphic cytosine, adenine, guanine repeat length (CAG RL) (18). Previous analyses (19) have indicated an association between CAG RL and longitudinal decline in T concentrations and that CAG RL may mediate an association between T and depressive symptoms (20), the latter likely associated with libido although not always with T itself (21). It may be that T concentrations corrected for CAG RL are more strongly associated with libido than is T alone. Likewise, it is possible that libido is most strongly related not to T but to a related factor such as LH, which may be elevated in men who exhibit symptoms of hypogonadism even in the presence of normal T concentrations.
We analyzed data on 1632 men enrolled in the Massachusetts Male Aging Study (MMAS), a longitudinal study of men’s health and aging, to address three specific queries: first, whether self-reported libido is associated with mean serum T or LH levels; second, whether low libido is an independent indicator of low serum T; and third, whether the ratio of T to CAG RL is more strongly associated with libido than is T itself. Measures of total T (TT), calculated bioavailable T (BT), and LH were evaluated with respect to frequency of sexual thoughts and desire as well as age and other relevant covariates.
Subjects and Methods
Study sample and design
The MMAS is an observational cohort study of aging men. The design has been described previously (22). Men of age 40–70 yr were randomly sampled from communities in and around Boston, Massachusetts; 1709 completed a baseline (T1) visit between 1987 and 1989. Follow-up visits were conducted after approximately 9 yr (T2, 1995–1997) and 15 yr (T3, 2002–2004). All procedures were approved by the Institutional Review Board of the New England Research Institutes.
The analysis reported here employs data on 1632 men who had TT, BT, and libido measurements at baseline; 922 (56%) and 623 (38%) of these men had T and libido data at T2 and T3, respectively. MMAS subjects were typically married (75%), had some education beyond high school (71%), and were employed (78%); 83% stated that they had one or more current sexual partners, compared with 81% at T2 and 78% at T3. The sample was racially/ethnically homogeneous (with 95% of subjects self-identifying as Caucasian), reflecting the demographic profile of residents of Massachusetts at baseline.
Data acquisition
A trained technician/interviewer visited each subject’s home. Written informed consent was obtained at the beginning of each visit. Medical history, self-assessed health and comorbidities (heart disease, hypertension, and diabetes), cigarette smoking, and typical daily alcohol consumption (23) were ascertained by prompted self-report. Depressive symptoms were measured using the Centers for Epidemiologic Studies–Depression (CES-D) scale (24). Physical activity and energy expenditure were derived from subjects’ 7-d recall of their activities (25, 26). CAG RL data were obtained (19) at T2 only (n = 922).
Serum hormone data were obtained by standard methods as previously described (8, 22). All assays were performed at the Endocrine Laboratory, University of Massachusetts Medical School (Worcester, MA), under the supervision of Christopher Longcope, M.D. Serum TT measurements were obtained using RIA kits from Diagnostic Products Corp. (Los Angeles, CA). BT was determined via calculation (27).
Previous analyses (28) indicate that aspects of experimental protocol can induce bias in measuring T concentrations. The design of the MMAS took account of these issues (8, 22). Most importantly, nonfasting blood samples were obtained within 4 h of each subject’s waking to minimize the influence of diurnal variation in hormone levels, and, to smooth episodic hormone secretion, two blood samples were obtained 30 min apart and pooled in equal aliquots at the time of assay.
Construction of libido score
At the end of each in-home visit, the subject completed a self-administered questionnaire, which he then gave to the interviewer in a sealed envelope. Two items were used to measure libido. The first asked: How frequently do you feel sexual desire? This feeling may include wanting to have a sexual experience (masturbation or intercourse), planning to have sex, feeling frustrated due to lack of sex, etc. The second asked: How frequently do you have sexual thoughts, fantasies, or erotic dreams? Responses were chosen from integer ordinal scales ranging from 0 (not at all) to 7 (more than once per day.) For this analysis, the two responses were added together to form a crude score with range 0–14. Presentation of low libido was explicitly defined as exhibiting a score of less than 7. Subjects who experienced neither sexual desire nor thoughts/fantasies at least once per week would meet this criterion, as would subjects who experienced thoughts once per week but desire only once per month (or vice versa).
Statistical analysis
Partial correlation statistics were employed to assess cross-sectional associations controlling for age. Mixed-effects linear regression analyses using data from all three study waves were employed to estimate mean hormone (TT, BT, and LH) levels as a function of concurrent libido status in controlling for study wave, subject age, and baseline covariates. To accommodate the effect of substantial skew in the distribution of LH, its natural logarithm was also considered. Parallel analyses of TT/CAG and BT/CAG were conducted.
Sample associations were considered statistically significant if the null hypothesis of no population association could be rejected at the 0.05 level via Wald or likelihood ratio tests. The predictive utility of low libido was assessed via point and interval estimation of the positive predictive value (PPV) of libido reports as indicators of serum TT, TT/CAG, BT, and BT/CAG values below various threshold levels at T2 (we highlight T2 because CAG RL data were obtained at that time). Odds ratios for predicting T levels below thresholds were estimated using multiple logistic regression.
Results
Baseline summaries are given in Table 1
. Mean ± SD TT and BT at T1 were 516.9 ± 175.9 ng/dl (17.92 ± 6.1 nmol/liter) and 249.0 ± 97.5 ng/dl (8.63 ± 3.4 nmol/liter), respectively. The mean ± SD baseline libido score was 9.2 ± 3.1. Three hundred fifty-one (22%) subjects reported experiencing both sexual desire and thoughts/fantasies more often than once per week but less frequently than once per day; these subjects had the mode libido score of 10. Ten was also the most common score at T2 (20%) and T3 (12%). Three hundred eighteen (19%) subjects exhibited low libido (a score of less than 7) at T1, whereas 23% (216 of 922) and 28% (173 of 623) reported low libido at T2 and T3, respectively.
|
T and libido were significantly associated. At T2, the partial correlation between libido scores and TT concentrations was 0.13 (P < 0.001) and that between libido and BT was 0.14 (P < 0.001); the corresponding quantities for TT/CAG and BT/CAG were 0.11 (P < 0.001) and 0.12 (P < 0.001), respectively. Thirteen of 26 (50%) subjects with TT less than 200 ng/dl, and 49 of 133 (37%) with TT less than 300 ng/dl reported low libido. Similarly, 24 of 46 (52%) and 38 of 89 (43%) subjects with BT less than 100 ng/dl and BT less than 120 ng/dl, respectively, reported low libido.
In keeping with these results, the overall distributions of hormone outcomes varied with libido status. However, there was substantial overlap in hormone concentrations over the range of scores. Figure 1 displays exploratory results for BT by raw libido score and rough quartiles; results for TT, BT/CAG, and TT/CAG (data not shown) were similar. The subjects with the lowest 23% of scores, i.e. those who displayed low libido, exhibited T concentrations quite close to those in the next lowest grouping. On the other hand, those subjects in the highest 20% (a score of at least 12) exhibited mean BT about 58 ng/dl (2.0 nmol/liter) higher than did those subjects who reported low libido.
|
. In multivariate models, libido remained significantly associated with T concentrations, but differences in TT and BT concentrations attributable to libido were reduced, to approximately 34 ng/dl (1.2 nmol/liter) TT and 12 ng/dl (0.42 nmol/liter) BT per 10-point difference in libido. (Note that such a 10-point difference is substantial in practical terms, corresponding, for instance, to the difference between subjects who report no desire or thoughts/fantasies at all vs. those who experience each weekly or to the difference between those subjects who experience desire and thoughts/fantasies two or three times per month and those who experience both more than once per day.) Competing factors such as diabetes and smoking status tend to account for differences as great as those associated with libido (Table 2). As is consistent with previous analyses of MMAS data, current cigarette smoking was associated with substantial increases in TT and BT. Covariates not depicted in Table 2 exhibited no significant influence in multivariate models and have been omitted.
|
The association between LH and libido was not significant in multivariate models. Dividing TT and BT by CAG RL did not appear to affect the relationship between libido and T concentrations across the spectrum of potential T values. CAG RL appeared to add little information to the analyses; at T2, for instance, the correlation between BT and BT/CAG was 0.9, and CAG RL displayed no association with libido scores.
Predictive value of independent reports of low libido
Subjects who reported low libido were more likely to exhibit low T levels than were subjects who did not. For instance, at T2, 23% of subjects who reported low libido had TT concentrations less than 300 ng/dl (10.4 nmol/liter), whereas the corresponding quantity among subjects who did not report low libido was 12% (age-adjusted odds ratio, 2.0; P = 0.002).
Table 3 presents an examination of self-report of low libido as an indicator of hormone concentrations below various TT and BT thresholds at T2. The proportions of subjects correctly classified by the low libido indicator are fairly high at all reasonable thresholds. However, the PPV of a low libido report in indicating low serum T was small. Among the subjects who reported low libido, only about 23% [95% confidence interval (CI) = 17.3–28.9] had serum TT concentrations less than 300 ng/dl (10.4 nmol/liter); the corresponding proportion for all subjects, regardless of libido status, was 14.4%. Results for BT were similar (Table 3).
|
Using the MMAS cohort data, we examined mean T (TT and BT) levels as a function of men’s libido in the context of age and other covariates. The relationship between libido and T levels was statistically significant in both unadjusted and adjusted models.
This result might lead one to interpret an individual patient’s report of reduced libido as a likely manifestation of T deficiency. Additional analyses suggest, however, that a simple individual report of low libido has a low predictive value in indicating T levels below commonly applied TT thresholds; at T2, the PPVs for predicting values less than 200 and 300 ng/dl were 6 and 23%, respectively. Likewise, the magnitude of regression effects indicated that large differences in libido were associated with only modest differences in T concentrations after other covariates were taken into consideration.
It should be noted, however, that the performance properties of a report of low libido as a crude test for low T levels are subject to the reality that the MMAS cohort is made up of community-dwelling (i.e. nonpatient), relatively healthy men. It may be that individual reports of decreased libido would exhibit greater utility as indicators of reduced T levels in a population of men at greater risk of AD.
Some authors have suggested that a BT may be a better measure of tissue available T than is TT (29) and thus potentially more strongly associated with clinically manifest symptoms such as reduced libido. Our results suggest that any such difference is inconsequential.
Certain limitations of this investigation should be noted. We view the composite libido score as a reasonable proxy for an individual patient’s report of his frequency of experiencing sexual desire, but it is a simple measure bound by the constraints of self-administered questionnaire in the context of a larger epidemiological survey. The two libido questions are limited to frequency of sexual feelings, making no effort to evaluate their duration or intensity. Given these considerations, the strength of the association between libido and T as observed here should not be undersold, with the caveat that the evidence at hand suggests that low libido exhibits modest PPV in predicting low T.
Recent studies have suggested that the effects of T supplementation on sexual function and desire may be modest and diminish over time (30). Our analyses indicate that, similarly, the role of endogenous T in modulating libido is complex and that an individual’s report of low libido should not necessarily be interpreted as evidence of low serum T.
Acknowledgments
We acknowledge the many contributions of Dr. Christopher Longcope, who passed away in 2004. For nearly 20 yr, he was an indispensable colleague on the MMAS. His scientific expertise and collegiality are missed. We also thank Taylor Aichel for her assistance in preparing this manuscript.
Footnotes
This work was supported by the National Institutes of Health (National Institute of Diabetes and Digestive and Kidney Diseases: DK44995, DK51345; National Institute on Aging: AG04673; and National Institute of Child Health and Human Development: HD39206).
J.E.M. has equity interests in Mattem Vet Pharmaceuticals, has consulted and lectured for Solvay Pharmaceuticals, and has received grant support from PAR Pharmaceuticals. All other authors have nothing to declare.
First Published Online May 2, 2006
Abbreviations: AD, Androgen deficiency; BT, bioavailable T; CAG RL, cytosine, adenine, guanine repeat length; CI, confidence interval; MMAS, Massachusetts Male Aging Study; PPV, positive predictive value; T, testosterone; TT, total T.
Received November 17, 2005.
Accepted April 21, 2006.
References
This article has been cited by other articles:
 |
|
 |
|
  A. B. Araujo, G. R. Esche, V. Kupelian, A. B. O'Donnell, T. G. Travison, R. E. Williams, R. V. Clark, and J. B. McKinlay Prevalence of Symptomatic Androgen Deficiency in Men J. Clin. Endocrinol. Metab., November 1, 2007; 92(11): 4241 - 4247. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
