In addition to critical development and reproductive

Transcription

1 Journal of Andrology, Vol. 33, No. 6, November/December 2012 Copyright E American Society of Andrology Seasonal Fluctuations in Testosterone-Estrogen Ratio in Men From the Southwest United States DANIEL J. MOSKOVIC, MICHAEL L. EISENBERG, AND LARRY I. LIPSHULTZ From the Scott Department of Urology, Baylor College of Medicine, Houston, Texas. ABSTRACT: Although controversial, seasonal variations in testosterone have been observed in several populations of men throughout the world. This finding might have an impact on screening and treatment of hypogonadism. We examined the circannual patterns of sex hormones in the Southwest United States. A prospectively assembled database of almost patients in a men s health practice was used to collect data on testosterone, estradiol, sex hormone binding globulin (SHBG), follicle-stimulating hormone (FSH), luteinizing hormone (LH), and dehydroepiandrosterone-sulfate (DHEA-S). Patient age, address, and date of visit were recorded. Of note, testosterone-estrogen ratio (T/E ratio) and free testosterone were calculated values. The data were grouped by month and by season (3-month intervals beginning with June, July, and August as summer). Analysis of variance was used to compare hormone levels between seasonal and monthly data sets, with P,.05 regarded as statistical significance. Statistically significant differences in estradiol (P 5.02), T/E ratio (P,.01), FSH (P 5.02), and SHBG (P,.01) were observed between seasons. Peak-to-trough variations were as follows: 6% for estradiol, 16.5% for T/E ratio, 11.0% for FSH, and 11.6% for SHBG. The T/E ratio peaked in the spring and was at its nadir in the fall. No differences in testosterone (P 5.21), LH (P 5.25), free testosterone (P 5.08), and DHEA-S (P 5.11) were observed. Statistically significant evidence of variation in estradiol and T/E ratio were identified in the men included in this study. Although this is consistent with seasonal body habitus changes, physical activity levels, and hypothesized hormonal patterns, the variability reported in the literature makes further trials covering a broader geographic region important to confirm the findings. Key words: Androgens, sex hormones, seasonal variation. J Androl 2012;33: In addition to critical development and reproductive function, testosterone has been identified as a significant contributor to the health of the aging male (Matsumoto, 2002). Specifically, deficiencies in testosterone have been found to be associated with feminized body habitus, sexual dysfunction, insulin resistance, and cardiovascular morbidities (Selvin et al, 2007; Meier et al, 2008; Nettleship et al, 2009). Given these important observations, testosterone screening is increasingly more common in the primary care setting (Tomlinson, 2007). It is therefore critical for practitioners to understand how to interpret androgen tests and identify sources of variation (Sadovsky et al, 2007). There have been several studies that have observed seasonal variations in testosterone (Andersson et al, 2003; Visscher and Seidell, 2004; Tancredi et al, 2005; Brambilla et al, 2007; Ruhayel et al, 2007). This finding, if confirmed, could have considerable impact on the management of male hypogonadism, given that treatment goals and dosage could vary seasonally. It is clear that testosterone levels vary within a 24-hour period, Correspondence to: Dr Larry I. Lipshultz, Scott Department of Urology, Baylor College of Medicine, 6624 Fannin St, Suite 1700, Houston, TX ( larryl@bcm.tmc.edu). Received for publication January 2, 2012; accepted for publication June 7, DOI: /jandrol conforming to a diurnal pattern with a maximum value in the morning (Crawford et al, 2007). It has been suggested that circadian rhythm and the sleep-wake cycle may be responsible for these daily changes (Luboshitzky et al, 1999). It is conceivable that variations in diurnal patterns may lead to seasonal variability of testosterone, where changes in sunlight exposure and activity levels may influence circadian rhythm. A study of Norwegian men by Svartberg et al (2003) demonstrated remarkable variation in testosterone levels, where differences as high as 31% were observed between peak and trough months; the highest levels were observed in the mid to late fall, and the lowest levels were observed in the summer. The authors hypothesized that such variations may be due to daylight exposure or ambient temperature. Surprisingly, a similar analysis of San Diego men by Svartberg and Barrett-Connor (2004) did not demonstrate circannual variation in testosterone levels. Similarly, another USbased study by Brambilla et al (2007) concluded that men in the Boston area do not experience seasonal variations in testosterone. Small samples sizes, however, make interpretation of the latter studies difficult. Given the observed variability of testosterone in some patient populations described above, we conducted an analysis of seasonal patterns of hormones in men living in the Southwest United States. Specifically, we 1298

2 Moskovic et al N Seasonal Fluctuations in T ÙE Ratio 1299 aggregated data on a variety of sex hormones in more than men, making this the largest analysis conducted of circannual testosterone patterns. Materials and Methods Patient Population A prospectively collected database was reviewed to assess serum hormone levels recorded in men being evaluated in a large men s health practice in the Southwest United States. Because of changes in hormone assay techniques over time, the analysis was restricted to only include samples analyzed with a similar or the same assay as the system currently in place in the practice (September 1997 to July 2010). Data were evaluated from the first patient visit only to ensure exclusion of men whose hormone levels might reflect treatment. Serum testosterone, estradiol, sex hormone binding globulin (SHBG), follicle-stimulating hormone (FSH), luteinizing hormone (LH), and dehydroepiandrosterone-sulfate (DHEA-S) were collected. Free testosterone and testosterone-estrogen ratio (T/ E ratio) were calculated values. Patient age, address, and date of visit were recorded. Patient serum was collected any time between 0900 and 1900 hours. Hormone data were analyzed to identify seasonal and monthly variations in values and ratios. For the seasonal analysis, the visit date was categorized by month of evaluation where visits in June, July, and August were grouped as summer; September, October, and November as fall; December, January, and February as winter; and March, April, and May as spring. Importantly, the results of monthly vs seasonal segmentation were internally consistent, and we elected to present the seasonal analysis herein. Statistical Analysis Descriptive statistics (mean, median, and SD) were calculated for all hormone levels. Analysis of variance was used to compare hormone levels between seasonal and monthly data sets (to identify internal consistency). Values with P,.05 were regarded as representing significant differences between chronologic cohorts; values of P,.01 are reported as such. Additionally, a study by Crawford et al (2007) suggested that diurnal variation in hormone levels diminishes after age 60 years. To account for physiologic differences in hormonal regulation by age, we dichotomized the data to compare men age,60 years to men $60 years. All statistical analyses were performed using Stata 10 (Stata Inc, College Station, Texas). Results Review of the database identified patients to be evaluated, with a mean age of years (mean 6 SD). The data were dichotomized by age, with 9669 patients younger than 60 years (mean age, years) and 1954 patients 60 years or older (mean age, years). By far, most patients (.90%) in the database reported permanent residence in Texas, Louisiana, Oklahoma, Arizona, and Southern California. Table 1 presents the comparative statistical analysis for the entire patient population and the dichotomized patient cohorts regarding seasonal variation in the total population. Statistically significant differences in seasonal peak-to-trough values were observed in estradiol (P 5.02), T/E ratio (P,.01), FSH (P 5.02), and SHBG (P,.01); there were no statistically significant changes in testosterone (P 5.21), LH (P 5.25), free testosterone (P 5.08), and DHEA-S (P 5.11). The relative and absolute differences between peak and trough values for each season are presented in Table 2. The Figure presents a graphical representation of total and free testosterone, LH, estradiol, and T/E variation by season. Dichotomizing by age did not substantially alter the above observations (Table 1). A 10% change was observed in free testosterone between peak (summer) and trough (fall) values, although this did not achieve statistical significance. In general, the younger patient cohort experienced greater statistical variation in sex hormones than did the older men, although the younger cohort was also larger (n vs 1954). Variation in T/E ratio was substantial in both cohorts but achieved statistical significance only in the younger group of patients (P,.01 vs P 5.06). SHBG was significantly variable from peak to trough in both age cohorts, with an approximately 15% variation between spring and summer. Older men experienced a nearly 30% decline in DHEA-S between winter and spring (P 5.03). Discussion This assessment of seasonal variation of sex hormones in men from the Southwest United States revealed that there are significant circannual differences in the T/E ratio in men. However, the magnitude of these differences was only significant in the younger cohort of patients. Seasonal variation of SHBG was significant in both age cohorts, with variations of up to 15% between peak and trough values. In contrast, testosterone and LH were similar throughout the year in both age cohorts. Additionally, DHEA-S and free testosterone are statistically similar throughout the year. The absence of circannual fluctuations in testosterone must be interpreted in the context of other studies assessing seasonal variations in androgens. Perhaps the most striking variation in androgen levels was reported by Svartberg et al (2003). Seasonal variation was assessed by comparing monthly mean serum values of 1548 men from Tromso, Norway, who underwent two

3 1300 Journal of Andrology N November ÙDecember 2012 Table 1. Seasonal variation of sex hormones for the general database and the dichotomized patient cohorts a Seasons Summer Fall Winter Spring P Testosterone, nmol/l Estradiol, pmol/l T/E ratio (T & E in nmol/l) FSH, IU/L LH, IU/L Free testosterone, pmol/l DHEA-S, mmol/l SHBG, nmol/l ,.01 Age,60 y (n ) Testosterone, nmol/l Estradiol, pmol/l T/E ratio (T & E in nmol/l) ,.01 FSH, IU/L LH, IU/L Free testosterone, pmol/l DHEA-S, mmol/l SHBG, nmol/l ,.01 Age $60 y (n ) Testosterone, nmol/l Estradiol, pmol/l T/E ratio (T & E in nmol/l) FSH, IU/L LH, IU/L Free testosterone, pmol/l DHEA-S, mmol/l SHBG, nmol/l Abbreviations: DHEA-S, dehydroepiandrosterone-sulfate; FSH, follicle-stimulating hormone; IGF-1, insulin growth factor 1; LH, luteinizing hormone; SHBG, sex hormone binding globulin; T/E ratio, testosterone-estrogen ratio. a Data presented include mean 6 SD. blood draws somewhere between 4 and 12 weeks apart. There was an 8-hour window within the day where blood drawing was acceptable (0800 to 1600 hours). Total testosterone, free testosterone, and LH showed substantial seasonal variability, with peak values increased by 19% and 31% for total and free testosterone, respectively (P,.001). Although peak values for testosterone were observed in the fall, trough values were seen in the summer. This relatively large study raised concern for wide variation in sex hormone levels between seasons and how this difference could have an impact on patient evaluation. Given the large window permitted for acceptable blood samples and variations in age and body habitus, it is conceivable that the results might be confounded by a variety of other factors. However, the authors state that adjustments to their data on the basis of these parameters did not impact the results. Therefore, in this cohort of men, there was evidence of seasonal fluctuation in testosterone that was substantial enough to be clinically important. A study of men reported by Andersson et al (2003) reported similar variability in circulating sex hormones, with a somewhat similar pattern of peak and trough values compared with the previously described study. Peak values of total testosterone were approximately 20% greater than nadir values. LH changes exhibited a similar trend. In this study, testosterone levels peaked in the late summer and hit the nadir in the late winter and spring. This small study included only 27 men who underwent monthly blood collection for approximately 1 year, and variation was based on month-to-month changes in hormones. Despite the small sample size, this study helped provide confirmation of seasonal variation in sex hormones. It is important to note that both Norway and Denmark show larger variations in temperatures and hours of sunlight between seasons than does the Southwestern United States, the location of the current report. Ruhayel et al (2007) reported on a series of men from the Arctic circle to assess how extremes of temperature and light-dark cycles might have an impact on hormonal fluctuations. A total of 205 Norwegian men were enrolled and underwent strict sample collection to minimize diurnal variations in circulating androgens.

4 Moskovic et al N Seasonal Fluctuations in T ÙE Ratio 1301 Table 2. Peak and trough seasonal variations in sex hormones a Seasonal Variation Peak Value Trough Value Relative Difference, % Absolute Difference Testosterone, nmol/l Estradiol, pmol/l T/E ratio (T & E in nmol/l) FSH, IU/L LH, IU/L Free testosterone, pmol/l DHEA-S, mmol/l SHBG, nmol/l Age,60 y (n ) Testosterone, nmol/l Estradiol, pmol/l T/E ratio (T & E in nmol/l) FSH, IU/L LH, IU/L Free testosterone, pmol/l DHEA-S, mmol/l SHBG, nmol/l Age $60 y (n ) Testosterone, nmol/l Estradiol, pmol/l T/E ratio (T & E in nmol/l) FSH, IU/L LH, IU/L Free testosterone, pmol/l DHEA-S, mmol/l SHBG, nmol/l Abbreviations: DHEA-S, dehydroepiandrosterone-sulfate; FSH, follicle-stimulating hormone; LH, luteinizing hormone; SHBG, sex hormone binding globulin; T/E ratio, testosterone-estrogen ratio. a Relative difference 5 (peak 2 trough) / peak. Absolute difference 5 (peak 2 trough). Approximately half of the enrollees lived north (Tromso) of the Arctic Circle, whereas the other half lived to the south (Oslo). Four samples were collected from each participant to attempt to collect data within the varying phases of the local light-dark cycle (early and late samples in the winter and summer). Seasonal variation was defined as a significant change in the total (early and late) winter levels compared with the total summer levels. Both regions exhibited a decline in LH levels in the early winter relative to the early summer, whereas only the northern region demonstrated seasonal variation. Interestingly, despite this finding, it was the southern region that demonstrated variation in total testosterone levels. Closer inspection of the data reveals that both samples experienced a decline in LH in the early summer, although participants from Oslo experienced a decline in testosterone paralleling this observation whereas men from Tromso experienced a borderline significant rise in testosterone (P 5.06). Free testosterone levels mirrored those of total testosterone. Estradiol levels varied significantly within both populations, with trends comparable to free and total testosterone levels. Of note, FSH did not vary by season. One suggested mechanism for the seasonal variation of sex hormones is their relationship to melatonin and/ or sleep patterns (Martikainen et al, 1985; Wehr, 1998; Svartberg and Barrett-Connor, 2004). Given the central regulation of melatonin and sleep, it has been proposed that these factors may influence the hypothalamicpituitary-gonadal (HPG) axis. Ruhayel and colleagues (2007) further postulated that there might be a relationship between circannual variations in melatonin and reproductive hormones. In fact, the authors observed a relationship between 6-sulfatoxymelatonin, a melatonin metabolite, and testosterone in Oslo and LH in Tromso. Although the results of this study are difficult to explain, their findings may represent an inadequately powered study. The authors did, however, demonstrate variability in melatonin levels in both locations where no consistent variations in any hormones were observed, perhaps weakening the argument of a relationship between melatonin secretion, circadian rhythm, and sex hormones. There has only been one US-based study, conducted by Brambilla et al (2007), examining seasonal fluctuations in androgens. In this Boston-area study, 121 men

5 1302 Journal of Andrology N November ÙDecember 2012 Figure. Representation of seasonal variation of sex hormones: (a) total testosterone, (b) luteinizing hormone (LH), (c) free testosterone, (d) estradiol, and (e) testosterone-estrogen (T:E) ratio. Error bars represent interquartile ranges for the values. were followed with morning blood draws at times 0, 3, and 6 months to obtain a composite for circannual sex hormone levels. The time of enrollment was random, so the investigators were able to capture sample representation in all 12 months. The results of their study did not demonstrate either empiric or statistical fluctuations in hormone levels. In fact, peak levels were within 4% of the mean level for all hormones examined. In aggregate, there is great heterogeneity in the existing literature regarding the impact that seasons may have on a man s circulating sex hormone levels. The current report, which is the largest to date, does not identify seasonal variation in androgens, although there is a statistically significant change in T/E ratio driven by slight fluctuations in estrogen levels. Initiation of testosterone replacement is typically based on both clinical and laboratory assessment of the patient (Sadovsky et al, 2007). It follows that if there is seasonal variability of circulating androgen and/or sex hormone levels, this may have an impact on therapeutic decisions. Sex hormones are known to vary throughout the day, thus exhibiting diurnal variation, a factor that is important for patients with borderline hypogonadism (Crawford et al, 2007). If there is seasonal variation in circulating androgen levels, it could have an impact on criteria for hormone repletion and might even allow for

6 Moskovic et al N Seasonal Fluctuations in T ÙE Ratio 1303 alternative treatments that would manipulate circadian rhythms to better manage androgen status. It is conceivable that seasonal changes may be due to mediating factors that vary during the year. For example, as body mass index changes during the year, it is possible that changes in body habitus may lead to changes in serum hormone levels due to aromatization rather than differences due to length of day or ambient temperature (Visscher and Seidell, 2004). There is clear evidence that exercise increases considerably in the summer months, potentially explaining the higher T/E ratio in the spring and summer seasons (Dannenberg et al, 1989; Matthews et al, 2001). Further support for this hypothesis can be derived from the dichotomized analysis presented herein; changes in the T/E ratio were more exaggerated (ie, statistically significant) in younger patients, who are more likely to be physically active. Additionally, there is evidence to suggest that the HPG axis is more sensitive in younger men, perhaps hinting at the potential for a central cause, wholly or partially, for these observations (Tenover and Bremner, 1991). Curiously, there was no decrease in testosterone associated with slight increases in estradiol. This observation would be expected, given the negative feedback on the HPG axis exerted by estradiol. Perhaps this suggests that the fluctuations are not clinically relevant and represent a statistical finding, given the large sample size. Indeed, other factors such as age, health status, and ethnicity have been implicated as predisposing patients to differences in sex hormone levels (Harman et al, 2001; Heald et al, 2003; Travison et al, 2007). Unfortunately, seasonal studies require large amounts of data to draw definitive conclusions, and it is difficult to control for all of these variables in such studies. Additionally, any of these factors may confound one another, suggesting that regional studies must be performed to elucidate the relationship between seasons and sex hormone variations. As the single largest study of circannual variation in sex hormones, the current study suggests that there are slight circannual differences in systemic sex hormone levels, particularly in younger men. The large number of patients included in the study validates these findings, as does the pattern between interrelated hormones (eg, LH and testosterone), although the finding may be statistically an anomaly without clinical significance. Importantly, our study has several important limitations that must be discussed. The aforementioned studies relied on longitudinal data on the sample patient cohort, whereas our data were obtained from single observations in more than patients. The Southwest United States also experiences different temperature and day-night patterns relative to some of the other locations discussed. And because of the nature of the database, there was no standard protocol for the timing of blood collection within the database. Additionally, parameters such as body mass index, patient medications, smoking status, glucose tolerance, total cholesterol, high-density lipoprotein cholesterol, and triglycerides were not available in this database. All of these factors may confound findings related to testosterone levels. However, given the very large sample size, it is likely that these parameters are similar between the study cohorts and that the distribution of blood collection times occurred similarly between the seasonal cohorts. Although these data are critically important in smaller studies to normalize for confounders that have an impact on testosterone levels, perhaps their influence is mitigated in this study on the basis of sample size. Conclusion This study represents the largest to date to explore the relationship in men between serum hormone levels and seasons. Estradiol, FSH, and SHBG were found to vary from peak to trough over the calendar year. Most interesting was the fluctuation of the T/E ratio, perhaps a statistically significant finding related to the large sample size included in this study. Alternatively, this finding may represent seasonal lifestyle or hormonal changes that have been postulated, although never directly linked to sex hormone variation in men. Given other publications on seasonal hormone variations, it is conceivable that circannual variation in sex hormones may be region specific and based on local climate patterns. References Andersson AM, Carlsen E, Petersen JH, Skakkebaek NE. Variation in levels of serum inhibin B, testosterone, estradiol, luteinizing hormone, follicle-stimulating hormone, and sex hormone-binding globulin in monthly samples from healthy men during a 17-month period: possible effects of seasons. J Clin Endocrinol Metab. 2003;88: Brambilla DJ, O Donnell AB, Matsumoto AM, McKinlay JB. Lack of seasonal variation in serum sex hormone levels in middle-aged to older men in the Boston area. J Clin Endocrinol Metab. 2007;92: Crawford ED, Barqawi AB, O Donnell C, Morgentaler A. The association of time of day and serum testosterone concentration in a large screening population. BJU Int. 2007;100: Dannenberg AL, Keller JB, Wilson PW, Castelli WP. Leisure time physical activity in the Framingham Offspring Study: description, seasonal variation, and risk factor correlates. Am J Epidemiol. 1989;129: Harman SM, Metter EJ, Tobin JD, Pearson J, Blackman MR. 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: Heald AH, Ivison F, Anderson SG, Cruickshank K, Laing I, Gibson JM. Significant ethnic variation in total and free testosterone concentration. Clin Endocrinol (Oxf). 2003;58:

7 1304 Journal of Andrology N November ÙDecember 2012 Luboshitzky R, Herer P, Levi M, Shen-Orr Z, Lavie P. Relationship between rapid eye movement sleep and testosterone secretion in normal men. J Androl. 1999;20: Martikainen H, Tapanainen J, Vakkuri O, Leppaluoto J, Huhtaniemi I. Circannual concentrations of melatonin, gonadotrophins, prolactin and gonadal steroids in males in a geographical area with a large annual variation in daylight. Acta Endocrinol (Copenh). 1985;109: Matsumoto AM. Andropause: clinical implications of the decline in serum testosterone levels with aging in men. J Gerontol A Biol Sci Med Sci. 2002;57:M76 M99. Matthews CE, Freedson PS, Hebert JR, Stanek EJ III, Merriam PA, Rosal MC, Ebbeling CB, Ockene IS. Seasonal variation in household, occupational, and leisure time physical activity: longitudinal analyses from the seasonal variation of blood cholesterol study. Am J Epidemiol. 2001;153: Meier C, Nguyen TV, Handelsman DJ, Schindler C, Kushnir MM, Rockwood AL, Meikle AW, Center JR, Eisman JA, Seibel MJ. Endogenous sex hormones and incident fracture risk in older men: the Dubbo Osteoporosis Epidemiology Study. Arch Intern Med. 2008;168: Nettleship JE, Jones RD, Channer KS, Jones TH. Testosterone and coronary artery disease. Front Horm Res. 2009;37: Ruhayel Y, Malm G, Haugen TB, Henrichsen T, Bjorsvik C, Grotmol T, Saether T, Malm J, Figenschau Y, Rylander L, Levine RJ, Giwercman A. Seasonal variation in serum concentrations of reproductive hormones and urinary excretion of 6-sulfatoxymelatonin in men living north and south of the Arctic Circle: a longitudinal study. Clin Endocrinol (Oxf). 2007;67: Sadovsky R, Dhindsa S, Margo K. Testosterone deficiency: which patients should you screen and treat? J Fam Pract. 2007;56:S1 S20; quiz S21 S23. Selvin E, Feinleib M, Zhang L, Rohrmann S, Rifai N, Nelson WG, Dobs A, Basaria S, Golden SH, Platz EA. Androgens and diabetes in men: results from the Third National Health and Nutrition Examination Survey (NHANES III). Diabetes Care. 2007;30: Svartberg J, Barrett-Connor E. Could seasonal variation in testosterone levels in men be related to sleep? Aging Male. 2004;7: Svartberg J, Jorde R, Sundsfjord J, Bonaa KH, Barrett-Connor E. Seasonal variation of testosterone and waist to hip ratio in men: the Tromso study. J Clin Endocrinol Metab. 2003;88: Tancredi A, Reginster JY, Luyckx F, Legros JJ. No major month to month variation in free testosterone levels in aging males: minor impact on the biological diagnosis of andropause. Psychoneuroendocrinology. 2005;30: Tenover JS, Bremner WJ. The effects of normal aging on the response of the pituitary-gonadal axis to chronic clomiphene administration in men. J Androl. 1991;12: Tomlinson JM. Testosterone and the primary care physician. Int J Clin Pract. 2007;61: Travison TG, Araujo AB, Kupelian V, O Donnell AB, McKinlay JB. The relative contributions of aging, health, and lifestyle factors to serum testosterone decline in men. J Clin Endocrinol Metab. 2007;92: Visscher TL, Seidell JC. Time trends ( ) and seasonal variation in body mass index and waist circumference in the Netherlands. Int J Obes Relat Metab Disord. 2004;28: Wehr TA. Effect of seasonal changes in daylength on human neuroendocrine function. Horm Res. 1998;49: