Correlations of Androgen Deficiency with Clinical Symptoms in Taiwanese Males

ORIGINAL ARTICLE Correlations of androgen deficiency with clinical symptoms in Taiwanese males

YC Lin1, TIS Hwang1,2,3, HS Chiang2, CR Yang4,HCWu5,TLWu6 and SP Huang7

1Division of Urology, Department of Surgery, Shin-Kong WHS Memorial Hospital, Taipei, Taiwan; 2Department of Urology, School of Medicine, Fu-Jen Catholic University, Taipei, Taiwan; 3Department of Urology, Taipei Medical University, Taipei, Taiwan; 4Division of Urology, Taichung Veterans General Hospital, Taichung, Taiwan; 5Department of Urology, China Medical University Hospital, Shenyang, China; 6Division of Urology, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan and 7Division of Urology, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan

The purpose of this study was to elucidate correlations between different biochemical measurements of androgen deficiency and clinical symptoms in male residents of Taiwan. An investigation of the serum biochemical markers for androgen deficiency in 650 males, including total testosterone, calculated free testosterone, and bioavailable testosterone, was conducted. Measure- ments of clinical symptoms were obtained using a questionnaire of the androgen deficiency in the aging male (ADAM) by St Louis University (SLQ). Correlations among the biochemical markers, correlations of the biochemical markers and age, and relationships between the biochemical markers and the SLQ were evaluated. The sensitivity and specificity of the SLQ were determined. Bioavailable and calculated free testosterone correlated better with age than did total testosterone. Eighty percent of the men had a positive SLQ, and 20% had a negative SLQ. The percentage of positive SLQ results increased with age. No statistically significant difference was noted between the biochemical markers of bioavailable and calculated free testosterone levels and the SLQ status except for men aged over 70 years. The SLQ in this study showed an acceptable sensitivity of about 80%, but the specificity was poor (about 20%). In conclusion, bioavailable testosterone and calculated free testosterone were more-closely correlated with age and may be better biochemical markers for androgen deficiency. SLQ might not be a suitable single measurement for androgen deficiency and should be used together with biochemical markers. International Journal of Impotence Research (2006) 18, 343–347. doi:10.1038/sj.ijir.3901417; published online 17 November 2005

Keywords: androgen deficiency; testosterone; bioavailable testosterone; calculated free testosterone

Introduction cognitive changes, and mood instability. Most importantly, sexual symptoms, such as erectile Testosterone (T) is the key hormone for maintaining dysfunction (ED) and decreased libido, are asso- male characteristics. Deficiency of this hormone in ciated with testosterone deficiency.1–3 Several bio- middle-aged and elderly men, or so-called aging chemical markers including free testosterone (FT), males, may cause physiological changes such as bioavailable testosterone (BT), and sex hormone- decreased body hair, skin changes, and decreased binding globulin (SHBG) have been developed in lean body mass and muscle strength. It can also the past few years to detect androgen deficiency. cause psychological symptoms, such as depression, Validation of these markers has been reported, but controversies with their efficacy still exist.4,5 Moreover, noninvasive screening questionnaires have been developed to detect populations with Correspondence: Dr TIS Hwang, Division of Urology, possible androgen deficiencies. Among the ques- Department of Surgery, Shin-Kong WHS Memorial Hospi- tionnaires, the questionnaire of androgen deficiency tal, No. 95 Wen Chang Rd., Shih Lin District, Taipei, of the aging male (ADAM) developed at St Louis Taiwan 111, ROC. 6 E-mail: [email protected] University (SLQ) is the commonly used one. The Received 8 September 2005; revised 28 September 2005; validation of this questionnaire was reported, and accepted 29 September 2005; published online 17 results have been thought to be good enough such November 2005 that it can be used as a screening tool for men with Correlations of androgen deficiency with clinical symptoms YC Lin et al 344 androgen deficiency.6 However, the correlations (Diamond Diagnostics, Holliston, MA, USA). Calcu- between clinical symptoms and biochemical test lated free testosterone (cFT) and bioavailable testost- results were weak, and doubts remained about its erone (BT) were determined using the formula of efficacy in Western countries.7,8 the International Society for Study of the Aging To clarify these doubts and to investigate the Male (ISSAM), which is available on the web ADAM in the Taiwanese population, a cross- (www.issam.ch). sectional study across the entire island of Taiwan was conducted. A survey for androgen deficiency and the correlations of a clinical screening question- Questionnaires naire with various biochemical markers was under- Anthropometric information including body height, taken. body weight, and body mass index (BMI) was obtained. Data on health behaviors and the subject’s current medical condition were also collected. Materials and methods Clinical symptoms of androgen deficiency in the aging male (ADAM) were assessed using the SLQ of Subjects ADAM.6 A total of 749 males aged between 40 and 80 years participated in this study in August 2004 from five medical centers distributed from north to south in Statistical analysis Taiwan. Equal numbers of rural and urban subjects Data were processed with SPSS 12.0 (SPSS, Chica- were included. Subjects were all volunteers and go, IL, USA), and one-way ANOVA was used for informed consents were obtained. Men who cur- correlation of the variables. Correlations of the rently have malignant diseases or liver cirrhosis, biochemical markers and age were evaluated with or who were using hormones, antiandrogen or a linear regression test. The sensitivity (the propor- antifungal agents, or steroidal agents were excluded tion of true positives) and specificity (the proportion from this study. Six hundred and fifty men were of true negatives) were calculated using biochemical eligible for enrollment in this study. Although the measurements as standards. activity was open to the general population, all of the study subjects were Taiwanese males. Table 1 Mean biochemical markers of androgen stratified into 10-year ranges

Hormonal measurements Age (years) Number TT (nmol/l) cFT (nmol/l) BT (nmol/l) Blood samples were collected and the following measurements were made. Serum albumin, follicle- 40–49 82 16.45 0.32 7.71 stimulating hormone (FSH), luteinizing hormone 50–59 236 15.50 0.29 6.95 (LH), prolactin, total testosterone (TT), and sex 60–69 213 16.23 0.26 6.19 hormone-binding globulin (SHBG) were checked. 70–80 119 15.67 0.22 5.19 Serum albumin was analyzed with a Beckman CX-7 Mean 650 15.89 0.27 6.47 chemistry analyzer (Global Medical Instrumenta- tion, Ramsey, MN, USA). LH, FSH, TT, and SHBG TT: total testosterone; cFT: calculated free testosterone; BT: were determined with a DPC Immulite analyzer bioavailable testosterone.

50.00 r=0.011 0.700 r=-0.330 15.000 r=-0.372 p=0.782 p<0.0001 p<0.0001 0.600 40.00 12.000 0.500

30.00 0.400 9.000 BT cFT TT 20.00 0.300 6.000 0.200 10.00 3.000 0.100

0.00 0.000 0.000 40 50 60 70 80 40 50 60 70 80 40 50 60 70 80 Age Age Age Figure 1 Correlations between the biochemical markers, total testosteron (TT), calculated free testosterone (cFT) and bioavailable testosterone (BT) with age. cFT and BT were correlated with age.

International Journal of Impotence Research Correlations of androgen deficiency with clinical symptoms YC Lin et al 345 50.00 r=0.62 50.00 r=0.63 15.000 r=0.98 p<0.0001 p<0.0001 p<0.0001 40.00 40.00 12.000

30.00 30.00 9.000 TT TT BT 20.00 20.00 6.000

10.00 10.00 3.000

0.00 0.00 0.000

0.000 3.000 6.000 9.000 12.000 15.000 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 BT cFT cFT Figure 2 Correlations among the three biochemical markers, cFT, and BT were well correlated with TT.

Table 2 Distributions of men with androgen deficiency in Table 3 Distribution of ADAM questionnaire results stratified by different age groups age

Age (years) TTo11 nmol/l cFTo0.23 nmol/l BioTo5 nmol/l Age (years) SLQ positive SLQ negative

Number % Number % Number % Number % Number %

40–49 14 11 15 7 11 7 40–49 60 73 22 27 82 50–59 47 39 56 27 36 23 50–59 178 75 58 25 236 60–69 37 30 75 36 52 33 60–69 175 82 38 18 213 70–80 24 20 70 33 60 38 70–80 108 91 11 9 119

Total 122 19 216 33 159 24 521 80 129 20 650

SLQ: St Louis University questionnaire of androgen deficiency in aging male. Results Table 4 Correlations between biochemical markers of androgen A total of 650 men were included in this study and deficiency and SLQ in 10-year ranges were stratified into four categories according to age. The hormonal status of the enrolled subjects is Age (years) SLQ positive SLQ negative presented in Table 1. A declining trend was noted mean7s.d. mean7s.d. P-value for the biochemical markers (TT, cFT, and BT) with age. However, TT did not correlate with age, but cFT 40–49 n ¼ 60 n ¼ 22 and BT correlated well with age (Figure 1). The TT (nmol/l) 16.5875.76 16.4375.97 0.962 relationship between age and cFT was À0.330 cFT (nmol/l) 0.3370.10 0.3070.09 0.571 (Po0.0001), between age and BT was À0.372 BT (nmol/l) 7.8672.41 7.2972.15 0.430 (P 0.0001), and between age and TT was 0.011 o 50–59 n ¼ 178 n ¼ 58 (P ¼ 0.782). Correlations among TT, cFT, and BT are TT (nmol/l) 15.7474.83 14.5875.00 0.197 shown in Figure 2. cFT and BT correlated well with cFT (nmol/l) 0.2970.09 0.2870.08 0.431 TT (cFT vs TT, r ¼ 0.63, Po0.0001; BT vs TT BT (nmol/l) 7.0372.12 6.7171.98 0.302 r ¼ 0.62, Po0.0001). Moreover, cFT and BT corre- 60–69 n ¼ 175 n ¼ 38 lated with each other best (r ¼ 0.98, Po0.0001). TT (nmol/l) 16.4376.10 15.8374.74 0.572 Androgen deficiency could be detected using the cFT (nmol/l) 0.2670.07 0.2870.07 0.146 hormonal study, and 122 (19%) subjects had low TT BT (nmol/l) 6.1071.72 6.6371.54 0.085 levels (using a cutoff value of o11 nmol/l), 209 (32%) had low cFT levels (using a cutoff value of 70–80 n ¼ 108 n ¼ 11 TT (nmol/l) 15.9775.53 13.6178.02 0.207 o0.23 nmol/l), and 159 (24%) had low BT levels 7 7 1 cFT (nmol/l) 0.22 0.07 0.18 0.07 0.033 (using a cutoff value of o5 nmol/l). The distribu- BT (nmol/l) 5.3071.60 4.1271.46 0.023 tion of males with androgen deficiency was stratified into four categories by age (Table 2). The bold values mean statistically significant results. There were 521 (80%) subjects with positive clinical symptoms of ADAM as surveyed by the age and a negative rate decreasing with age were SLQ and 129 (20%) with negative SLQ results. noted (Table 3). The relationship between the SLQ Trends of an increasing positive rate with increased and androgen deficiency is shown in Table 4. The

International Journal of Impotence Research Correlations of androgen deficiency with clinical symptoms YC Lin et al 346 SLQ could only be used to screen men with was neither a statistically significant correlation androgen deficiency aged over 70 years. No statis- between the biochemical parameters and SLQ tically significant difference was elucidated for the except for men aged over 70 years. The SLQ may biochemical markers between men with positive have its role in screening patients with symptoms of and negative SLQ results. androgen deficiency, but it should not be the only The sensitivity of the SLQ in this study was tool used to survey men with these symptoms. 75.4% for TT, 81.5% for cFT, and 81.8% for BT, but For questions 1 and 7 of the SLQ surveying sexual the specificity was 18.6% for TT, 21.1% for cFT, and dysfunction, we found that 55% of middle-aged to 20.2% for BT. According to SLQ questions 1 and 7, elderly men in Taiwan have decreased libido, and 55% of the men in this study had decreased libido 61% have ED. Trends of increasing prevalences of and 61% had ED. We also found a trend of an decreased libido and ED with age were also demon- increasing prevalence of these two symptoms with strated. When comparing the status of ED and increasing age. decreased libido, the biochemical markers did not differentiate men with or without symptoms except for men older than 70 years of age. The biochemical study results showed no correlations with androgen Discussion deficiency. These results support the fact that the etiology of ED and decreased libido are multi- This cross-sectional study enrolled 650 males aged factorial and possibly not influenced by a single between 40 and 80 years from different regions of factor. Androgen deficiency as the main factor Taiwan. As such, these data are representative of the causing ED was only found in a small fraction of current hormonal status of men living in Taiwan. patients.14–16 The prevalence of androgen deficiency was 9–32% Sensitivity of the SLQ has previously been based on the biochemical measurements evaluated validated by Morley et al.6 to be 88%, with a in this study. This result is comparable with specificity of 60%. The sensitivity of the SLQ in this previous studies in a group of men aged below 70 study was around 80% for the different biochemical years, but is low for men aged over 70 years.9–11 markers, which is similar to the report by Morley Controversies still exist over the determination of et al.;18 however, the specificity was only around ideal biochemical measurements for people with 20%. Furthermore, there were more enrolled sub- androgen deficiency. Serum TT is a less-expensive jects in this study and they were community-based, and simpler biochemical measurement for aging while in the previous study, subjects were confined males with androgen deficiency. However, a poor to physicians. Our results may be more comprehen- correlation between TT and age has been reported,12 sive and may suggest that the questionnaire should similar to that demonstrated in this study. cFT and not be the sole tool used in screening for androgen BT are better parameters, but the procedure is time deficiency. consuming and the costs are relatively higher.5,12 The prevalence of androgen deficiency being However, Christ-Crain et al.13 reported no good similar in this study to studies conducted in correlations between BT and TT. We used all three of Western countries might imply that no differences these parameters for the screening test to verify their exist between races.6,17–19 However, the lack of a correlations. In this study, the TT level did not correlation between androgen deficiency and the correlate with age, but cFT and BT correlated better SLQ might be a racial factor, and further large with age. However, TT showed good correlations surveys among different races should be carried out with cFT and BT. Accordingly, cFT and BT corre- to elucidate this observation. lated well with TT, and a correlation with age was also noted. In checking men with possible androgen deficiency, cFT and BT are better and more-reason- able biochemical markers, as was reported by Morley et al.6 and Tsujimura et al.12 Conclusions The ADAM questionnaire of St Louis University had been used as a noninvasive screening tool to BT and cFT are better parameters than TT for detect possible androgen deficiency in aging males. detecting androgen deficiency. BT and cFT correlate Morley et al.6 validated this questionnaire with BT, well with TT, but the correlation is best between cFT and they concluded that the SLQ showed high and BT. No correlations between the SLQ and sensitivity, acceptable specificity, and good intrara- biochemical measurements were found except in ter reliability. Since then, the questionnaire had men aged over 70 years of age. Biochemical testing been used to define symptoms of ADAM. However, cannot adequately reflect ED and decreased libido. Christ-Crain et al.13 reported that testosterone did The SLQ may not be a convincing single screening not have a statistically significant correlation with test for men with androgen deficiency, and bio- the SLQ, although the study population they chemical measurements should be done simulta- evaluated was small. In the present study, there neously.

International Journal of Impotence Research Correlations of androgen deficiency with clinical symptoms YC Lin et al 347 Acknowledgments 9 Harman SM, Metter EJ, Tobin JD, Pearson J, Blackman MR. Longitudinal study of aging. Longitudinal effects of aging on serum total and free testosterone levels in healthy men. J Clin This study was in part financially supported by the Endocrinol Metab 2001; 86: 724–731. Taiwanese Association of Andrology. 10 Feldman HA, Longcope C, Derby CA. Age trends in the levels of serum testosterone and other hormone in middle age men: longitudinal results of the Massachusetts Male Aging Study. J Clin Endocrinol Metab 2002; 87: 589–596. 11 Morley JE, Kaiser FE, Perry III HM. Longitudinal changes in References testosterone, luteinizing hormone and follicle stimulating hormone in healthy older men. Metabolism 1997; 46: 1 Morales A, Lunenfeld B. Investigation, treatment and mon- 410–415. itoring of late-onset hypogonadism in males. Official recom- 12 Tsujimura A, Matsumiya K, Matsuoka Y, Takahashi T, Koga M, mendations of ISSAM. International Society for the Study of Iwasa A et al. Bioavailable testosterone with age and erectile the Aging Male. Aging Male 2002; 5: 151–158. dysfunction. J Urol 2003; 170: 2345–2347. 2 Davidson JM, Chen JJ, Crapo L, Gray GD, Greenleaf WJ, Catania 13 Christ-Crain M, Mueller B, Gasser TC, Kraenzlin M, Trummler JA. Hormonal change and sexual function in aging men. J Clin M, Huber P et al. Is there a clinical relevance of partial Endocrinol Metab 1983; 57: 71–77. deficiency of the aging male. J Urol 2004; 172: 624–627. 3 Comhaire FH. Andropause: hormone replacement in the 14 Morales A, Heaton JP. Hormonal erectile dysfunction. ageing male. Eur Urol 2000; 38: 655–662. Evaluation and management. Urol Clin North Am 2001; 28: 4 Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation 279–288. of simple methods for the estimation of free testosterone in 15 Foresta C, Caretta N, Rossato M, Garolla A, Ferlin A. serum. J Clin Endocrinol Metab 1999; 84: 3666–3671. Role of androgens in erectile function. J Urol 2004; 171: 5 Morley JE, Patrick P, Perry III HM. Evaluation of assays 2358–2362. available to measure free testosterone. Metabolism 2002; 51: 16 Morales A, Heaton JP. Hypogonadism and erectile dysfunc- 554–559. tion: pathophysiological observations and therapeutic out- 6 Morley JE, Charlton E, Patrick P, Kaiser FE, Cadeau P, comes. BJU Int 2003; 92: 896–899. McCready D et al. Validation of a screening questionnaire for 17 Tenover JS. Effects of testosterone supplementation in the androgen deficiency in aging male. Metabolism 2000; 49: aging male. J Clin Endocrinol Metab 2001; 86: 1092–1097. 1239–1242. 18 Morley JE, Perry III HM, Kaiser FE, Kraenzle D, Jensen J, 7 Barrett-Connor E, Von Muhlen DG, Kritz-Silverstein D. Houston K et al. Effects of testosterone replacement therapy in Bioavailable testosterone and depressed mood in older men: old hypogonadal males: a preliminary study. J Am Geriatr Soc the Rancho Bernardo Study. J Clin Endocrinol Metab 1999; 84: 1993; 41: 149–155. 573–577. 19 Wang C, Swedloff RS, Iranmanesh A, Dobs A, Snyder PJ, 8 Synder PJ, Peachey H, Hannoush P, Berlin JA, Loh L, Holmes Cunningham G et al. Transdermal testosterone gel improve JH et al. Effect of testosterone treatment on bone mineral sexual function, mood, muscle strength, body composition density in men over 65 years of age. J Clin Endocrinol Metab parameters in hypogonadal men. J Clin Endocrinol Metab 1999; 84: 1966–1972. 2000; 85: 2839–2848.

International Journal of Impotence Research