| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Andrology, ANZAC Research Institute and Concord Hospital (P.Y.L., S.M.W., M.J., D.J.H.), and Woolcock Institute of Medical Research and Royal Prince Alfred Hospital (B.Y., D.G.J., R.R.G.), University of Sydney, Sydney, 2139 Australia
Address all correspondence and requests for reprints to: Prof. David J. Handelsman, ANZAC Research Institute, Sydney NSW 2139 Australia. E-mail: djh{at}med.usyd.edu.au.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
1 h), increased the duration of hypoxemia (5 min/night), and disrupted breathing during sleep (total and non-rapid eye movement respiratory disturbance indices both increased by approximately seven events per hour) (all P < 0.05). Despite expected effects on body composition (increase in total and lean mass, reduction in fat mass, P < 0.05, bioimpedance method), upper airway dimensions did not change (acoustic reflectometry). Driving ability (computer simulation), physical activity (accelerometry, Physical Activity Scale in the Elderly), quality of life (SF36, Functional Outcomes of Sleep Questionnaire), mood (Profile of Mood States Questionnaire), sleepiness (Epworth, Stanford scales), and insulin resistance (homeostasis model) also were not changed by treatment. Short-term administration of high-dose testosterone shortens sleep and worsens sleep apnea in older men but did not alter physical, mental, or metabolic function. These changes did not appear to be due to upper airway narrowing. Further study of longer-term lower-dose androgen therapy on sleep and breathing is needed to evaluate its safety in older men. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Androgen treatment has long been believed to precipitate OSA in some men (3). However, few studies have systematically examined the effect of androgen treatment on breathing or sleep (4, 5, 6, 7, 8, 9). There are no randomized placebo-controlled studies that have examined the effect of androgen treatment on both sleep and breathing, although one has reported no reduction in sleep (7) and the other no worsening of breathing (9). Furthermore, no study has shown that androgen treatment shortens sleep or impairs physical or mental function. Although inconsistent effects on sleep architecture are reported, the paucity of well-designed studies using gold standard endpoints (such as overnight polysomnography) means that the safety of androgen therapy during sleep remains uncertain.
Concurrently there is increased interest in the use of androgen supplementation in older men to improve body composition, muscle and bone strength, and physical functioning (10, 11, 12, 13, 14, 15, 16, 17, 18, 19). These studies have consistently recruited men over 60 yr of age with baseline plasma testosterone concentrations 
430 ng/dl (15 nmol/liter, which includes the lower end of the young eugonadal male reference range) because men with lower baseline testosterone concentrations are more likely to be androgen responsive (14, 20). Most of these studies have treated men for less than 6 months, and there are no long-term studies of high-dose testosterone treatment. Because of the paucity of data, this study was designed to examine the safety of short-term, high-dose testosterone treatment on sleep, breathing, and mental and physical function in older men, and, therefore, the duration of exposure was minimized to reduce the risk to participants.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study was designed as a randomized sequence, double blind, placebo-controlled cross-over study of 17 eligible and evaluable men (Fig. 1
). Subjects were randomly assigned to receive either testosterone esters (250 mg/ml, Sustanon 250, Organon Australia, Sydney, Australia) or arachis oil volume-matched placebo (Department of Pharmacy, Concord Hospital) during the first period before crossing over to the other treatment. Each treatment period consisted of three im injections spaced 1 wk apart. The first injection was 2 ml (equal to 500 mg testosterone esters in the treatment group), and the subsequent two injections were 1 ml (250 mg testosterone ester) each. There was a washout period of 8 wk between the two treatment periods, which would allow complete clearance of administered testosterone. All injections were administered by staff of the Department of Andrology. Subjects were advised to report any injection-related (stinging, bruising, or pain) adverse event. Adverse events or intercurrent illness and their likely relationship with drug administration were recorded prospectively. Anthropometry, sleep studies, driving, and physical and mental function were assessed at baseline and after each study period (2–4 d after the last im injection to capture the maximal circulating testosterone concentrations). More easily obtained variables [blood, SF36, International Prostate Symptom Score (IPSS), Physical Activity Scale in the Elderly (PASE) ] in subjects were obtained at 1- to 2-wk intervals (Fig. 1). All study procedures were approved by the CSAHS Ethics Committee within National Health and Medical Research Council (Australia) guidelines. The trial was conducted under the TGA (Australia) clinical trial notification scheme.
|
Healthy, ambulatory and community-dwelling men older than 60 yr of age were recruited if they had a plasma testosterone 430 ng/dl (15 nmol/liter) on two separate occasions. They were excluded because of: 1) presence of severe snoring or sleepiness requiring investigation and treatment; 2) contraindication to testosterone use including breast cancer, prostate cancer, or disease requiring further treatment; 3) unstable, uncontrolled or severe chronic medical disease; 4) medical conditions that might interfere with evaluation of sleep or physical activity (including physical incapacity or dementia); or 5) medication use known to interact with sex steroid action or alter sleep
Study procedures
Volunteers were recruited through public advertisement. Respondents were provided with an explanation of the study and a written information sheet and were required to sign a consent form. Standardized medical history, physical examination, and blood samples were obtained at entry. Fasting blood samples were taken between 0830 and 1030 h before injection of study drug, and participants were asked not to vary their diet or exercise patterns. Eligible subjects who satisfied all entry criteria were randomly assigned a study number that corresponded with individually numbered drug supplies. The Department of Pharmacy (Concord Hospital) held the randomization list (obtained by random-number generator) and dispensed all study medications. At the end of the study, subjects nominated the period during which active treatment was administered to assess adequacy of blinding.
Assays and insulin sensitivity
Serum samples were stored frozen for measurement, within a single assay, using commercially available DELFIA immunoassays (Perkin-Elmer Life Sciences, Rowville, Australia) for LH, FSH, total testosterone, SHBG, insulin, and prostate-specific antigen (PSA). Leptin was measured by double-antibody RIA (Linco Research Inc., St Louis, Mo) assay. Blood glucose was determined by the hexokinase method (Hitachi 917 autoanalyzer, Roche Diagnostics, Castle Hill, Australia) from blood collected in sodium fluoride (Vacutainer; Becton Dickinson, Rutherford, NJ) during the previous 2 h. Within assay coefficients of variation were less than 10% for all assays, and detection limits for gonadotropin assays were 0.1–0.2 IU/liter. Insulin resistance and production were calculated according to the homeostasis model (21).
Anthropometry and upper airway caliber
Anthropometry and upper airway caliber were assessed in all subjects on three occasions (baseline, end of phase 1, and end of phase 2) by a single observer. Anthropometric (22) and bioimpedance measurements were taken while fasting before the first and last injection of each study phase immediately after micturition as previously reported (19). Lean mass was estimated from bioimpedance (SEAC model BIM 3.0 bioimpedance meter, Inderlec, Brisbane, Australia) according to Lukaski and Bolunchuk’s formula and fat mass, obtained by subtraction from body weight (23).
Upper airway caliber was assessed by acoustic reflectometry (24) (Eccovision pharyngometer; E. Benson Hood Laboratories, Pembroke, MA). Airway dimension at the pharynx, glottis, and oropharyngeal junction were measured with the subject awake and in four positions (supine, erect, and in the left and right lateral) while breathing quietly at tidal volume.
Polysomnography
Sleep and breathing were assessed by full overnight polysomnography (at RPAH Sleep Disorders Centre) after one night of laboratory acclimatization in all subjects on three occasions (baseline, end of phase 1, and end of phase 2) and as previously described (25). Chest wall and abdominal motion were recorded using Respitrace (Ambulatory Monitoring Inc., Ardsley, NY), airflow was measured by nasal pressure (pressure transducer; Validyne, Northridge, CA) and arterial oxyhemoglobin saturation with a Biox 3700e (Ohmeda, Louisville, CO). Subjects were asked to maintain their usual sleeping pattern (and not to nap during the day) before sleep studies. Data were recorded continuously on a computerized polygraph (Compumedics E series, Melbourne, Australia) and scored by a single blinded technologist using American Academy of Sleep Medicine consensus methods (26) to calculate respiratory disturbance indices and sleep stage. In brief, apneas occurred when airflow ceased, whereas hypopneas occurred when airflow or thoracoabdominal movement fell by at least 50%. Apneas and hypopneas had to be of at least 10 sec duration and were classified as obstructive if there was continued diaphragm electromyographic activity or thoracoabdominal wall movement otherwise they were central. The total respiratory disturbance index (RDI) was calculated as the number of apneas and hypopneas that occurred per hour of sleep and were subdivided into rapid eye movement (REM) RDI and non-REM (NREM) RDI according to whether events occurred in REM or NREM sleep, respectively. Sleep stages [sleep stage 1 (S1), 2 (S2), 3 (S3), 4 (S4), NREM, and REM sleep] were defined by standard electroencephalographic criteria (26). Sleep efficiency (the proportion of time slept while in bed), sleep-related hypoxemia (oxygen saturation <90%, during sleep), and nadir oxygen saturation during sleep were also determined.
Driving performance task
Driving ability was assessed by computer simulation (27) in a darkened room (AusEd, Diagnose IT, Sydney, Australia) at baseline and after each study phase. In brief, all subjects performed a 30-min driving performance task after a 20-min practice session in a darkened room at 2000 h in the evening. The object of the test was to steer an image of a car hood along a predefined monotonous, night-driving course as accurately as possible within a velocity range indicated on a screen speedometer. To test vigilance and reaction time, subjects had to brake when a visual image of a truck randomly appeared. The driving simulator recorded tracking error (ability to steer accurately), velocity deviation (time outside mandatory speed range), reaction time (average time to respond to visual stimulus), and number of crashes.
Physical activity and questionnaires
Physical activity was determined objectively by accelerometry (TriTrac R3D ergometer; ZMD Reining Inc., Madison, WI) (28) using the manufacturer’s proprietary software (Tritrac R3D software, version 6.05) and by PASE questionnaire (29).
General [SF36 (30)] and sleep-specific [Functional Outcomes of Sleep Questionnaire (FOSQ) (31)[ quality of life as well as sleepiness [Stanford and Epworth sleepiness scale (32)] were assessed before and after treatment in each phase of the study. Mood was assessed by the Profile of Mood States Questionnaire (POMS) (33) and lower urinary tract symptoms associated with prostate growth were assessed by the IPSS (34).
Statistical analysis
The randomization code was not broken until all data were collected and the resulting database cleaned and locked. Standard techniques to transform nonnormal data (Shaprio-Wilk test) and examine for carryover and period effects were used (35). Response to testosterone treatment was analyzed by paired t test (35). For data with repeated measures, treatment and placebo data were pooled separately by averaging and then analyzed analogously. These results were confirmed by a two-way ANOVA model with main effects for order (placebo or testosterone treatment first) and time, with repeated measures on time. In this model, treatment effect was determined by the interaction between order and time and period effects by the time variable. The results of this method are not presented unless they differ significantly from the previous. Pearson’s correlations were determined to assess linear associations between baseline hormonal data and change in breathing or sleep. All data are expressed as mean ± SEM. Data were considered significantly different at P 0.05 (two sided). Analyses were performed using SPSS 10.0 (SPSS Inc., Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Seventeen eligible men aged 67.5 ± 0.8 (SEM) yr were recruited with nine receiving testosterone first. One subject was discontinued at the end of the second period (while on placebo) because of recurrence of long-standing cervical neck pain. There were no other serious adverse events, and no adverse event was considered to be drug related except for two men who had transient nipple tenderness while receiving testosterone and injection-related pain that occurred in three of 102 injections.
Sixteen evaluable men completed the study. There were no missing appointments and minimal missing data (<1%) arising from equipment failure. Period and carryover effects were not detected. Subjects were unable to correctly predict their treatment assignments at the end of the study (only 65% guessed correctly, P = 0.33) suggesting adequate blinding.
Hormonal and anthropometric effects
Testosterone treatment significantly suppressed LH and FSH to the detection limit of the assay and markedly increased serum total testosterone and estradiol by 2- to 3-fold (Fig. 2). Leptin fell, but there were no significant changes in SHBG, insulin, glucose, or PSA (not all data shown). IPSS (lower urinary tract symptoms) was not changed by treatment (not shown).
|
2 kg) and lean mass (3 kg) and reduced fat mass (1 kg) (Fig. 3). There was no significant change in either abdominal or neck circumferences or in upper airway caliber (data not shown).
|
Total time slept was shorter on testosterone treatment (1 h) with a reduction in both NREM (particularly stage 2) and REM sleep as well as sleep efficiency (Fig. 4). The proportion of total time slept in NREM and REM sleep did not change (data not shown). The total RDI (7 events/h) was increased by testosterone, and this was due predominantly to events occurring in NREM sleep (7 events/h, Fig. 5). Absolute (5 min) and relative (as a proportion of total time slept, by 2%) hypoxemia was prolonged, but nadir oxygen saturation was not altered by testosterone treatment (Fig. 5).
|
|
Physical and mental function and quality of life
Driving ability (ability to steer and maintain constant speed, crashes, and divided attention) was not altered by testosterone treatment. Physical activity (accelerometry and PASE), mood (POMS), sleepiness (Epworth and Stanford), and quality of life (FOSQ and SF-36) were not altered by testosterone treatment. Metabolic function (homeostasis model assessment-derived insulin resistance and production) was unchanged. (Fig. 6, not all data shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Despite worsened breathing, lengthened hypoxemia, and shortened sleep, no alteration in physical or mental function (driving ability, self-reported and objectively recorded physical activity, mood, sleepiness, or quality of life) or metabolic function (insulin resistance) was detected, although changes may have occurred with prolonged therapy. Insulin resistance is associated with OSA (39, 40, 41) independent of visceral obesity (42, 43), and it may be that there was no net effect because disordered breathing worsens and reduced adiposity improves insulin action. It is possible that the short duration of high-dose androgen therapy chosen (to minimize any potential adverse effects) was not long enough to cause detectable physical, mental or metabolic changes, particularly because the increase in sleep-disordered breathing and duration of hypoxemia were both moderate in magnitude. However, prolonged therapy, which would be required to maintain body compositional and other physical performance improvements in older men, may have resulted in chronic low-level hypoxemia leading to impaired cardiovascular, neurological, and physical function. In addition, sleep deprivation, particularly when sustained, will impair physical and mental function in older individuals, although the magnitude of the effect is reduced, compared with the young (44).
Alternatively, napping during the day (against specific instructions) may have compensated for the decrease in sleep at night, but there is no evidence for this. It is also possible that the methods used were not sensitive enough to detect subtle changes in physical, mental, or metabolic function. However, our methods to measure driving ability, physical activity, and insulin resistance are validated (21, 27, 28, 29, 45) and widely used.
Testosterone may worsen breathing by a number of mechanisms because upper airway patency is determined by many structural and neuromuscular factors that control pharyngeal airway size and collapsibility. Obesity, as reflected by body mass index (BMI) and neck circumference (46), is associated with OSA. However, this association is explained by upper airway narrowing (24) probably via pharyngeal fat deposition (47, 48). Although testosterone treatment resulted in an increase in BMI, adiposity (and presumably pharyngeal fat) was reduced (as measured by bioimpedance), consistent with the observed reduction in serum leptin concentration. It is also possible that these rapid change in body composition reflect changes in hydration because of high doses of testosterone used in a short period of time and that the disturbed breathing arose from upper airway edema. However, fluid shifts do not occur in young men treated with comparable doses of testosterone esters (49). Furthermore, direct measurement of upper airway caliber (by acoustic reflectometry) at the narrowest (the oropharyngeal junction) and widest (maximal pharyngeal area) locations, as well as neck or abdominal circumferences, did not change with treatment. Because these measurements were performed with the subjects awake, we cannot exclude the possibility that upper airway narrowing may have occurred specifically during sleep. Although a direct anabolic effect on upper airway soft tissue growth could result in a physical reduction in upper airway dimension and this is thought to be how androgens induce OSA in women (50, 51, 52), the duration of therapy in this study was too short for this to have occurred.
Testosterone may worsen breathing by neuromuscular mechanisms. Testosterone treatment increases upper airway collapsibility (53), ventilation (54), and hypoxic (54) and hypercapnic ventilatory responses (55), leading to a reduced apneic threshold. These effects on breathing may occur particularly during NREM sleep (55), which is consistent with our findings.
Although androgen deficiency, particularly in older men, is associated with insomnia and other sleep disturbances, a marked reduction in time slept with testosterone administration was actually observed. It may be that high and low circulating testosterone concentrations are both associated with sleep disturbances. Testosterone may directly alter sleep through central nervous system effects including altered serotingergic neurotransmission (56). Testosterone increases nocturnal metabolic rate (54), potentially impairing sleep quality (57).
An important caveat of this study was that, as expected, treatment resulted in high supraphysiological serum testosterone concentrations and sleep and function were measured at the time of expected peak concentrations. Such high doses have been used therapeutically in pilot studies in older men (58), but there are no long-term studies and on the basis of our findings, long-term, high-dose androgen use in older men cannot be recommended, although the safety of long-term, low-dose androgen use remains to be clarified. Hence, our findings are applicable mostly to short-term, high-dose testosterone ester administration and may not be applicable to lower-dose physiological replacement, particularly with preparations (such as transdermal or subdermal testosterone therapy), which do not result in supraphysiological peak serum concentrations (9). Nevertheless, im testosterone ester administration is a widely used and studied form of therapy among older men (10, 11, 12, 13, 17).
We also observed the phenomenon of "regression to the mean" (59) in baseline testosterone concentrations, whereby despite screening requirements of less than 430 ng/dl (15 nmol/liter), the subsequent pretreatment baseline testosterone concentration was modestly higher (16, 19). Hence, many of our men had a baseline testosterone within the young normal range. We also showed that a lower baseline testosterone or a higher baseline total respiratory disturbance index was not linearly associated with greater worsening of any breathing or sleep parameter. This suggests that excluding subjects for androgen therapy on the basis of preexisting OSA may not be sufficient, and conversely, hypogonadal men with coexistent OSA should not necessarily be denied androgen therapy. Nevertheless, monitoring of sleep and breathing would be prudent in hypogonadal men with symptoms (e.g. snoring, daytime sleepiness, morning headache) of or risk factors (e.g. obesity) for OSA commencing androgen replacement, particularly replacement with im testosterone esters (3, 4, 5, 6, 7, 53). This is particularly prudent because this study specifically included men without symptoms of severe OSA and may, therefore, have excluded men who may have been more adversely affected by androgen therapy. However, at baseline many of the men had disturbed breathing, defined polysomnographically, consistent with men of this age. The clinical significance of these laboratory-based measurements is uncertain in the absence of symptoms of OSA. However, no man was diagnosed with severe OSA requiring treatment, although four men were offered further evaluation and two men subsequently received a trial of continuous positive airway pressure.
Short-term prostate safety was suggested by the lack of change in PSA and IPSS, consistent with other longer-term but lower-dose studies (14, 16, 19). Uniquely, this study shows that high-dose treatment, sufficient to increase serum testosterone and estradiol by 2- to 3-fold will not acutely alter PSA or induce lower urinary tract symptoms in older men if ultrashort duration therapy is used.
Nevertheless, this study raises concerns about the safety of long-term androgen therapy in older men with partial age-related androgen deficiency. Consideration for breathing and sleep safety must now be placed alongside concerns regarding prostate disease (60) and polycythemia (61). Given the recently recognized adverse effects of hormone replacement therapy in older women (62), this is a salutatory reminder that high-quality studies examining both efficacy and safety are required before widespread adoption of androgen replacement in older men can be recommended.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: BMI, Body mass index; FOSQ, Functional Outcomes of Sleep Questionnaire; IPSS, International Prostate Symptom Score; NREM, non-rapid eye movement; OSA, obstructive sleep apnea; PASE, Physical Activity Scale in the Elderly; POMS, Profile of Mood States Questionnaire; PSA, prostate-specific antigen; RDI, respiratory disturbance index; REM, rapid eye movement.
Received February 12, 2003.
Accepted April 24, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Wald, M. Miner, and A. D. Seftel State of the Art Reviews: Male Menopause: Fact or Fiction? American Journal of Lifestyle Medicine, April 1, 2008; 2(2): 132 - 141. [Abstract] [PDF]
|
|
|
|
 |
|
 D. Ahuja, J. H. Mateika, M. P. Diamond, and M. Safwan Badr Ventilatory sensitivity to carbon dioxide before and after episodic hypoxia in women treated with testosterone J Appl Physiol, May 1, 2007; 102(5): 1832 - 1838. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Bhasin, G. R. Cunningham, F. J. Hayes, A. M. Matsumoto, P. J. Snyder, R. S. Swerdloff, and V. M. Montori Testosterone Therapy in Adult Men with Androgen Deficiency Syndromes: An Endocrine Society Clinical Practice Guideline J. Clin. Endocrinol. Metab., June 1, 2006; 91(6): 1995 - 2010. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. L. Ding, Y. Song, V. S. Malik, and S. Liu Sex Differences of Endogenous Sex Hormones and Risk of Type 2 Diabetes: A Systematic Review and Meta-analysis JAMA, March 15, 2006; 295(11): 1288 - 1299. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Kaufman and A. Vermeulen The Decline of Androgen Levels in Elderly Men and Its Clinical and Therapeutic Implications Endocr. Rev., October 1, 2005; 26(6): 833 - 876. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Morton Frequent Occurrence of Hypogonadotropic Hypogonadism in Type 2 Diabetes J. Clin. Endocrinol. Metab., March 1, 2005; 90(3): 1903 - 1903. [Full Text] [PDF]
|
|
|
|
 |
|
 R. F Fregosi Castrating the respiratory controller J. Physiol., December 1, 2004; 561(2): 353 - 353. [Full Text] [PDF]
|
|
|
|
|
|
P. Y. Liu, R. S. Swerdloff, and J. D. Veldhuis The Rationale, Efficacy and Safety of Androgen Therapy in Older Men: Future Research and Current Practice Recommendations J. Clin. Endocrinol. Metab., October 1, 2004; 89(10): 4789 - 4796. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
