Sexual Function/Infertility Validity of Midday Total Testosterone Levels in Older Men with Erectile Dysfunction R. Charles Welliver, Jr.,* Herbert J. Wiser, Robert E. Brannigan, Kendall Feia, Manoj Monga and Tobias S. K ohler From the Division of Urology, Southern Illinois University School of Medicine (RCW, TSK) and Springfield Clinic (HJW), Springfield and Department of Urology, Northwestern University Feinberg School of Medicine (REB), Chicago, Illinois, Department of Urology, University of Minnesota School of Medicine and Minneapolis Veterans Affairs Healthcare System, Minneapolis (KF), Minnesota, and Glickman Urological and Kidney Institute, Cleveland Clinic Foundation (MM), Cleveland, Ohio Purpose: Based on studies showing the circadian rhythmicity of testosterone the optimal time of day to draw total testosterone in men has classically been reported as between 8 and 11 a.m. However, further studies demonstrated that the testosterone circadian rhythmicity becomes blunted with age. Materials and Methods: We retrospectively reviewed the charts of 2,569 men who presented with erectile dysfunction for total testosterone and draw times. We compared the men by age group, including less than 40 years and 5-year groupings after age 40 years. Total testosterone was analyzed for variability during the most common draw time hours (7 a.m. to 2 p.m.). Results: Mean total testosterone at 7 to 9 a.m. and 9 a.m. to 2 p.m. clinically and statistically differed only in men younger than 40 vs 40 to 44 years old (mean difference 207 ng/dl, 95% CI 98e315, p ¼ 0.0004 vs 149 ng/dl, 95% CI 36e262, p ¼ 0.01). No other group showed a clinically and statistically significant difference between those periods. Conclusions: Total testosterone in men with erectile dysfunction who are younger than 45 years should be drawn as close to 7 a.m. as possible because a statistically and clinically relevant decrease in testosterone will occur during the course of the day. Men older than 45 years with erectile dysfunction can have total testosterone drawn at any time before 2 p.m. without misleading results. Key Words: testis, erectile dysfunction, testosterone, circadian rhythm, age groups THE latest practice guidelines of the Endocrine Society recommend measuring morning TT in men presenting with symptoms of androgen deficiency. 1 The optimal time to draw TT in men has been classically reported as morning 1 to obtain the peak value or specifically between 8 and 11 a.m. 2 Based on 24-hour sequential blood sampling in small subsets of healthy men diurnal variation of TT was clearly demonstrated but this effect is attenuated as men age. 3e7 A large study in men with a mean age of 60 years participating in prostate cancer screening demonstrated that TT remains stable throughout the morning and early afternoon, decreasing only after 2 p.m. 8 No difference was noted in diurnal variation before 2 p.m. In contrast, other researchers reported complete loss of diurnal variation in TT in men with androgen deficiency. 9 Other studies clearly showed the blunting of diurnal variation in TT in older Abbreviations and Acronyms ADT ¼ androgen deprivation therapy ED ¼ erectile dysfunction LH ¼ luteinizing hormone TRT ¼ testosterone replacement therapy TT ¼ total testosterone VA ¼ Veterans Affairs Accepted for publication January 14, 2014. Study received Minneapolis VA Hospital institutional review board approval. * Correspondence: Division of Urology, Department of Surgery, Southern Illinois University School of Medicine, 301 North Eighth St., P.O. Box 19649, Springfield, Illinois 62794 (e-mail: rwelliver@ siumed.edu). Financial interest and/or other relationship with Fertility and Sterility. Financial interest and/or other relationship with Auxilium, Allergan, AMS, Coloplast, AbbVie and Hollister. For another article on a related topic see page 267. 0022-5347/14/1921-0165/0 THE JOURNAL OF UROLOGY 2014 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH,INC. http://dx.doi.org/10.1016/j.juro.2014.01.085 Vol. 192, 165-169, July 2014 Printed in U.S.A. www.jurology.com j 165
166 VALIDITY OF MIDDAY TOTAL TESTOSTERONE IN OLDER MEN WITH ERECTILE DYSFUNCTION This is not surprising since TT decreases with age in men. This is thought to occur at a rate of approximately 1% per year beginning around age 40 years. 12 However, the conclusions drawn from screening the general population for low testosterone should be used carefully. Recent Endocrine Society guidelines recommend against screening the general population for androgen deficiency and only recommend screening men with the characteristic symptoms of androgen deficiency. 1 Age dependent effects of diurnal variation remain unclear, particularly in younger men. Additionally, knowing the diurnal variation in TT in men with a classic symptom of androgen deficiency such as ED is likely to be more helpful for determining who and when to screen. Because hypogonadal men typically notice an energy nadir later in the day, perhaps investigating TT levels at that time of maximum bother may provide more insight. Previous literature necessitates asking the question of what time of day testosterone levels can reliably be drawn when considering patient age. 13 We assessed the effect of age on the diurnal variance of TT and reassessed draw time recommendations by determining peak TT values. More clearly establishing acceptable draw times may lead to more patient flexibility in obtaining laboratory values and decrease redundant laboratory testing. men. 5e7,10,11 Separate analysis was done to compare mean TT for the draw times 7 a.m. to 2 p.m. and after 2 p.m. for all patients older than 45 years. Data were analyzed with ExcelÒ and GraphPadÒ statistical calculators using the 2-tailed test with continuity corrections. All results at p <0.05 were considered significant. Checking LH became standard at Minneapolis VA in 2000. A subset of men (653 of 2,794 with available TT) also had a LH level drawn as part of the laboratory assessment before the visit. LH and TT in these men were evaluated to determine what percent in the overall study group of 2,569 with TT 50 to 1,000 ng/dl were likely to have hormonal manipulation. Men with undetectable LH but TT between 50 and 1,000 ng/dl were considered to have a confounding additional factor of ADT or TRT. RESULTS Mean age of these patients was 63 years (range 26 to 89). The 65 to 69-year age group had the largest number of patients (532). The extremes of age (26 to 39 years in 46 patients and 80 to 84 years in 43) were least represented. Graphic analysis with smoother curves of scatterplot data of TT by draw times of different age groups showed a large decrease in TT during the day in the youngest 2 groups (less than 40 and 40 to 44 years) (see figure). All other age groups did not MATERIALS AND METHODS This study was approved by the institutional review board at Minneapolis VA Hospital. The computerized charts of all men who presented to the Minneapolis VA ED clinic from 1986 to 2004 with a primary complaint of ED were collected as previously described. 14 These records were then retrospectively reviewed. A total of 3,547 men were evaluated at the ED clinic during this 15-year period. Of these men 2,794 between ages of 26 and 84 years had TT evaluated at least once, including 114 excluded from primary study analysis due to draw time outside a 7 a.m to 2 p.m. range. The 111 men with TT outside the laboratory reference range of greater than 1,000 or less than 50 ng/dl were excluded from study. The remaining 2,569 men were stratified into groups by age, including a 26 to 39-year age group and groups at 5-year age intervals thereafter with 46 or more men per group. The groups were then differentiated into hourly draw times spanning 7 a.m. to 2 p.m. No patient had repeat testosterone measurements included in the data set. These data subgroups were analyzed by scatterplot techniques. Based on the scatterplots and the need for adequate numbers for statistical purposes each age group was separated into 2 groups, including men with a draw time of 7 to 9 a.m. and those with a draw time of 9 a.m. to 2 p.m. The mean TT in each draw time subgroup was compared by each age group using the Student t-test. TT vs draw time in men of different age groups
VALIDITY OF MIDDAY TOTAL TESTOSTERONE IN OLDER MEN WITH ERECTILE DYSFUNCTION 167 demonstrate this obvious decrease in TT. In men younger than 45 years the highest TT was predictably seen at the earliest points in the day and the rate of decrease throughout the day was most notable in the morning. Comparing TT measured between 7 and 9 a.m. vs 9 a.m. to 2 p.m. in the different age groups yielded a statistically significant difference in only 3 groups, that is less than 40, 40 to 44 and 70 to 74 years (p ¼ 0.0004, 0.01 and 0.038, respectively, see table). The difference in mean TT was 207, 149 and 34 ng/dl, respectively. To analyze the TT decrease after 2 p.m. in men older than 45 years TT values in samples drawn after 2 p.m. in 100 patients were also compared to TT values at all draw times before 2 p.m. in 2,462. A statistical difference was seen with mean TT for draw times before 2 p.m. greater than values after 2 p.m. (428 169 vs 387 168 ng/dl, mean difference 41, 95% CI 7, 75, p ¼ 0.017). Of the 653 men with known LH values only 15 had LH below reference values, implying hormonal manipulation. Six of these 15 men had TT outside the 50 to 1,000 ng/dl range, including 3 with values less than 50 and 3 with values greater than 1,000 ng/dl. By restricting TT to between 50 and 1,000 ng/dl we could then identify 6 of the 15 men (40%) who were Early morning vs late morning/midday TT in men of different age groups Age (No. pts) Time Mean SD TT (ng/dl) Mean ng/dl TT Difference (95% CI) p Value 26e39: 207 (98e315) 0.0004 17 7e9 a.m. 647 233 29 9 a.m.e2 p.m. 440 134 40e44: 149 (36e262) 0.0103 16 7e9 a.m. 660 242 51 9 a.m.e2 p.m. 511 181 45e49: 3 ( 64e58) 0.9231 43 7e9 a.m. 474 184 113 9 a.m.e2 p.m. 477 169 50e54: 20 ( 30e70) 0.4318 74 7e9 a.m. 452 164 205 9 a.m.e2 p.m. 432 195 55e59: 28 ( 15e71) 0.1986 83 7e9 a.m. 445 186 216 9 a.m.e2 p.m. 417 161 60e64: 5 ( 32e42) 0.7879 123 7e9 a.m. 443 174 298 9 a.m.e2 p.m. 438 173 65e69: 7 ( 26e40) 0.6763 150 7e9 a.m. 441 171 382 9 a.m.e2 p.m. 434 175 70e74: 34 (2e66) 0.0379 146 7e9 a.m. 452 165 339 9 a.m.e2 p.m. 418 165 75e79: 2 ( 53e57) 0.9425 65 7e9 a.m. 358 90 168 9 a.m.e2 p.m. 356 216 80e84: 3 ( 75e69) 0.9333 14 7e9 a.m. 361 99 29 9 a.m.e2 p.m. 364 114 likely undergoing hormonal manipulation. Only 9 men in the group of 653 had normal range TT and undetectable LH. These patients with potentially unrecognized hormonal manipulation comprised only 1.4% of the study group. DISCUSSION This study further solidifies the circadian rhythmicity of testosterone and the decrease in the effect of this rhythm with increasing age. 3e6,10,15 However, to our knowledge our study differs since we verified this in men with a classic symptom of hypogonadism and quantified the severity of testosterone blunting in relation to patient age. The magnitude of change when comparing the 7 to 9 a.m. interval draw against draws at all other times was clinically and statistically significant in men younger than 40 and 40 to 44 years old (207 and 149 ng/dl, respectively). Although the 70 to 74-year age group showed a statistically significant change (p ¼ 0.038), the clinical significance of a 34 ng/dl difference is questionable. Although statistical analysis was only done to compare the 7 to 9 a.m. and 9 a.m. to 2 p.m. draw times, scatterplots indicated that testosterone decreased rapidly after 7 a.m. in men younger than 45 years. This corresponds to previous sequential blood draw studies showing peak testosterone at approximately 6 to 7 a.m. 7 Values of TT drawn after 2 p.m. were also compared with values at all draw times before 2 p.m. in patients older than 45 years. Our results were similar to those of Crawford et al 8 in that a statistically significant decrease in TT (41 ng/dl, p ¼ 0.017) was noted after 2 p.m. However, definitive conclusions from this point should be tempered due to the limited number of men (100) who had blood drawn after 2 p.m. and the questionable clinical significance of a 41 ng/dl change in TT. Most experts recommend routine TT measurement as a diagnostic tool in men who present with hypogonadal symptoms. Low or low normal TT should be confirmed by repeat TT measurement and in some patients by the free or bioavailable testosterone level. 1 Furthermore, a true diagnosis of hypogonadism requires at least some adverse effect such as ED, low libido, depression, lethargy, osteoporosis, regression of secondary sexual characteristics, or loss of muscle mass or strength. 16 The clinical symptom present in all patients differentiates this study from the prostate cancer screening subgroup in another large study. 8 The cause of a decrease in testosterone production with age is multifactorial and poorly understood. Potential causes include a decrease in the number and alteration of Leydig cells, 17 and age
168 VALIDITY OF MIDDAY TOTAL TESTOSTERONE IN OLDER MEN WITH ERECTILE DYSFUNCTION related alterations in the hypothalamic-pituitarygonadal axis. 18 Notable correlations have also been made between decreased sleep quantity with advancing age and decreasing testosterone. 19 A potential confounding factor of this study is the potentially unrecognized use of ADT and TRT in these patients. We mitigated the influence of this possible confounder by analyzing the effect of our exclusion criteria, namely in excluding TT values outside the laboratory reference range. When available, LH values were included in analysis to enable us to draw conclusions on confounding risk in our overall cohort. Generally men who are on ADT or TRT would have undetectable LH due to central repression or feedback, respectively, from these treatments. Checking LH became a standard practice at Minneapolis VA Hospital in 2000 and 94% of men in whom TT was measured between 2000 and 2002 also had LH assessed. This subgroup of 653 men represented 23% of the entire study cohort. After applying our exclusion criteria an acceptably small cohort of only 1.4% of the subgroup would have been on unrecognized hormonal manipulation. Including these men in the data set did not significantly change our results. As we reported previously, 23% of the entire cohort could have been considered hypogonadal using a cutoff point of 300 ng/dl. 20 An ideal study to determine the optimum time to draw TT levels in men stratified by age would be well powered and include multiple repeat bioavailable testosterone samples from the same man on the same day. An ideal study would also record and account for potential confounders of TT such as past medical problems, lifestyle variations, waking time and medication history. Additionally, the study would record any of the well recognized sequelae of hypogonadism (low libido, poor energy, sarcopenia, etc) and stratify patients by these symptoms. Given that patients in our study had TT drawn only once and potential confounders were not recorded, we relied on the law of large numbers to have aggregate mean values of many men (with an averaged confounding risk) for a given period approximate the TT of a future single patient for the same period. Another group drew conclusions based on a similar premise. 8 Although our methodology cannot account for several confounders, our data appropriately corroborate 2 previously proposed principles. 1) The data show a smooth decrease in TT with age. 2) Obvious diurnal variation exists in men younger than 45 years and this effect is blunted in older men. CONCLUSIONS Our study of a single blood draw in a large number of men older than 45 years with ED revealed no difference in mean TT at any time between 7 a.m. and 2 p.m. In men younger than 45 years with ED a draw time after 9 a.m. resulted in significantly lower mean TT. Our data provide unique insight into a large cohort of men with ED, a key symptom of hypogonadism. REFERENCES 1. Bhasin S, Cunningham GR, Hayes FJ et al: Testosterone therapy in adult men with androgen deficiency syndromes: an endocrine society clinical practice guideline. J Clin Endocrinol Metab 2010; 95: 2536. 2. Wespes E and Schulman CC: Male andropause: myth, reality, and treatment. Int J Impot Res, suppl., 2002; 14: S93. 3. Marrama P, Carani C, Baraghini GF et al: Circadian rhythm of testosterone and prolactin in the ageing. Maturitas 1982; 4: 131. 4. Plymate SR, Tenover JS and Bremner WJ: Circadian variation in testosterone, sex hormone-binding globulin, and calculated nonsex hormone-binding globulin bound testosterone in healthy young and elderly men. J Androl 1989; 10: 366. 5. Bremner WJ, Vitiello MV and Prinz PN: Loss of circadian rhythmicity in blood testosterone levels with aging in normal men. J Clin Endocrinol Metab 1983; 56: 1278. 6. Tenover JS, Matsumoto AM, Clifton DK et al: Age-related alterations in the circadian rhythms of pulsatile luteinizing hormone and testosterone secretion in healthy men. J Gerontol 1988; 43: M163. 7. Diver MJ, Imtiaz KE, Ahmad AM et al: Diurnal rhythms of serum total, free and bioavailable testosterone and of SHBG in middle-aged men compared with those in young men. Clin Endocrinol (Oxf) 2003; 58: 710. 8. Crawford ED, Barqawi AB, O Donnell C et al: The association of time of day and serum testosterone concentration in a large screening population. BJU Int 2007; 100: 509. 9. Gupta SK, Lindemulder EA and Sathyan G: Modeling of circadian testosterone in healthy men and hypogonadal men. J Clin Pharmacol 2000; 40: 731. 10. Montanini V, Simoni M, Chiossi G et al: Age-related changes in plasma dehydroepiandrosterone sulphate, cortisol, testosterone and free testosterone circadian rhythms in adult men. Horm Res 1988; 29: 1. 11. Boyce MJ, Baisley KJ, Clark EV et al: Are published normal ranges of serum testosterone too high? Results of a cross-sectional survey of serum testosterone and luteinizing hormone in healthy men. BJU Int 2004; 94: 881. 12. Harman SM, Metter EJ, Tobin JD et al: Longitudinal effects of aging on serum total and free testosterone levels in healthy men. Baltimore Longitudinal Study of Aging. J Clin Endocrinol Metab 2001; 86: 724. 13. Morgentaler A: Commentary: guideline for male testosterone therapy: a clinician s perspective. J Clin Endocrinol Metab 2007; 92: 416. 14. Bodie J, Lewis J, Schow D et al: Laboratory evaluations of erectile dysfunction: an evidence based approach. J Urol 2003; 169: 2262.
VALIDITY OF MIDDAY TOTAL TESTOSTERONE IN OLDER MEN WITH ERECTILE DYSFUNCTION 169 15. El-Sakka AI and Hassoba HM: Age related testosterone depletion in patients with erectile dysfunction. J Urol 2006; 176: 2589. 16. Nieschlag E, Swerdloff R, Behre HM et al: Investigation, treatment and monitoring of lateonset hypogonadism in males: ISA, ISSAM, and EAU recommendations. Eur Urol 2005; 48: 1. 17. Neaves WB, Johnson L, Porter JC et al: Leydig cell numbers, daily sperm production, and serum gonadotropin levels in aging men. J Clin Endocrinol Metab 1984; 59: 756. 18. Deslypere JP and Vermeulen A: Aging and tissue androgens. J Clin Endocrinol Metab 1981; 53: 430. 19. Penev PD: Association between sleep and morning testosterone levels in older men. Sleep 2007; 30: 427. 20. K ohler TS, Kim J, Feia K et al: Prevalence of androgen deficiency in men with erectile dysfunction. Urology 2008; 71: 693.