Cross-sectional analysis between prostate volumes and serum sex hormones in Japanese men attending a urological clinic

Abstract

Aim: To evaluate epidemiologically the association between measured prostate volume and sex hormones.

Methods: Between December 2012 and September 2014, 226 patients attending the urological clinic were assessed for the relationship between prostate volume (PV) and, serum sex hormones, physical size, personal habits, etc. Prostate volume was measured by using transabdominal ultrasonography. Statistically, the Pearson correlation coefficients test was used.

Results: Total cases, the cases of PV ≤ 25 ml, and the cases of PV > 25 ml were evaluated respectively. Total cases and the cases of PV > 25 ml showed a positive significant correlation with testosterone (T), but the cases of PV ≤ 25 ml showed no such correlation. The cases of PV > 25 ml had a positive significant correlation with estradiol (E2), but total cases and cases of PV ≤ 25 ml did not. Dehydroepiandrosterone sulfate (DHEAS) showed no correlation with any case of PV, however it decreased significantly with age and had a correlation with alopecia. The E2/T ratio had no correlation with any case of PV, but on the other hand, the T/DHEAS and E2/DHEAS ratios had significant positive correlation with PV > 25 ml.

Conclusions: Serum T and E2 had significant positive correlation with measured PV especially in larger prostates. This result seems to correspond with the conventional theory that T and E2 have an etiological effect on benign prostatic hyperplasia. DHEAS did not show direct correlation with PV, however it appeared to suppress the role of T and E2 on benign prostatic hyperplasia growth. DHEAS might be a key to understanding the etiology of benign prostatic hyperplasia with aging.

• Benign prostatic hyperplasia
• DHEAS
• estradiol
• prostate volume
• testosterone

Introduction

Benign prostatic hyperplasia (BPH) is one of the most common diseases treated as a cause of male lower urinary tract symptoms (LUTS) in urological clinics. Pathologically, the prevalence of BPH is only 5% at the fourth decade, however, it is 50% at the fifth decade, and 90% at the eighth decade [1,2]. In recent years new treatments of BPH have been developed – involving new medicines and new operating techniques. Also, the molecular and the endocrinological mechanisms of BPH have been the subjects of many researches. However, the epidemiology of BPH is undoubtedly multifactorial, and therefore still controversial [2–7]. Concerning sex hormones, testosterone (T) is supposed to be the most important steroidal hormone which affects the function and development of the prostate. T is converted to dihydrotestosterone (DHT) by 5α-reductase, then, transported to target tissues, binding to the androgen receptor with a consequent modulation of genes [2]. Estradiol (E2) is the main biological active estrogen produced in testis or converted from androgens in males [8] and has a fundamental role in the regulation of body fat and sexual function [9]. E2 is also implicated as one of the causes of BPH. E2 proliferates BPH stromal cells in culture, moreover, estrogen receptor and aromatase activity, are detected in BPH stroma [10]. Dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS) which decrease with age in both male and female [11], are the most prevalent adrenal androgens and are thought to be converted to DHT in prostate tissue [12], which affects the pathogenesis of BPH. Whereas there are a lot of arguments over the epidemiology of BPH, only a few studies have been published concerning the relationship between prostate volume (PV) examined by ultrasonography and serum sex hormones epidemiologically. Roberts et al. carried out a population–based study in the USA [6], Schatzl et al. evaluated the correlation between endocrine parameters and clinical parameters such as PV in elderly men with LUTS [13], and Liu et al. completed a study using a free health screening program from Taiwan [14]. However these studies, except for minor correlations, did not clearly show a relationship between PV and sex hormones. In this study, we tried to find the correlation between measured PV and sex hormones in the patients attending our clinics in Japan.

Methods

Subjects

The study concerns 226 ambulatory patients aged 48–98 years (mean: 71.7, SD: 8.76), who attended the urological clinic located in rural areas in Kyoto prefecture from December 2012 to September 2014. Patients who had prostate cancer or other critical diseases, or were prescribed 5α–reductase inhibitor were excluded. The patients studied complained of LUTS (206), or other urological diseases such as ureteral stone and microhematuria (20), and had their kidney, bladder, and prostate checked by ultrasonography. At the same time, blood examinations were performed for prostate-specific antigen (PSA), renal function, and sex hormones. Permission for these examinations was received from the patients. Furthermore, patients' general status, tobacco and alcohol use, and alopecia were also checked.

Measurement of sex hormones

Blood sampling was performed during clinic hours between 9:00 am and 12:30 pm. Serum T, E2, and DHEAS, were measured. These analyses were conducted in the KYOTO BISEIBUTSU KENKYUSYO (Kyoto Microbiological Institute). T, E2, and DHEAS were measured by enzyme immunoassay using kits of COBAS (Roche Diagnostics, Tokyo, Japan), ADVIA Centaur (Siemens Healthcare Diagnostics, Tokyo, Japan), and Access (BECKMAN COULTER, Tokyo, Japan), respectively.

Prostate measurement

Prostates were scanned by trans-abdominal ultrasonography by the same operator and the size was calculated from the longitude, the transverse, and the height of prostates following the method of Stamey [15,16].

Statistical method

The correlation between PV and sex hormones was analyzed by the Pearson correlation coefficients test. The correlation between other factors such as age, height, weight, BMI, smoking (divided into 0: never, 1: past, 2: current), alcohol drinking (divided into 0: never, 1: past, 2: current) and alopecia (divided into 0: no alopecia, 1: mild, 2: severe) were also considered. A p value of <0.05 was considered statistically significant on the two–sided test. All statistical analyses were carried out with the IBM SPSS Statistics version 22 (IBM Corp., Armonk, NY).

Results

The total number of patients was 226, and they were divided into two groups (PV > 25 ml:125 patients, PV ≤ 25 ml:101 patients) which were evaluated respectively for the correlation between PV, sex hormones and other variables. We had already examined the PV of 436 Japanese males aged from their 40s to their 80s using trans-rectal ultrasonography. 373 of them had normal prostates, the mean PV of which was 19.7 ml [17]. Evaluating 286 men, Masumori et al. also reported the mean PV of Japanese men in their 40s and 50s was 17.4–19.4 ml [18]. Therefore, we presumed that a PV of 25 ml showed an enlarged prostate. Patients’ characteristics are showed in the Table 1.

Table 1. Characteristics of patients.

Variable	All patients	PV ≤ 25	PV > 25
	(N = 226)	(N = 101)	(N = 125)
Testosterone (ng/ml)	4.86 ± 1.81	4.91 ± 1.84	4.82 ± 1.79
E2 (pg/ml)	29.5 ± 11.4	29.5 ± 12.4	29.6 ± 10.5
DHEAS (μg/dl)	112.0 ± 61.6	117.4 ± 65.1	107.7 ± 58.4
Height (cm)	165.4 ± 6.3	164.4 ± 6.6	166.2 ± 6.0
Body weight (kg)	62.8 ± 9.4	61.5 ± 10.1	63.8 ± 8.6
BMI	22.9 ± 2.9	22.7 ± 3.3	23.1 ± 2.6
Smoking (N)
0	89	44	45
1	95	38	57
2	40	19	21
unknown			2
Alcohol (N)
0	69	24	45
1	15	8	7
2	140	69	71
unknown			2
Alopecia (N)
0	107	51	56
1	67	30	37
2	47	20	27
unknown			5
Age	71.7 ± 8.8	71.7 ± 9.3	71.7 ± 8.36

Mean ± SD.

PV and age

Overall PV (mean: 33.0, SD: 20.4) and age showed no correlation (p = 0.290) (Figure 1). PV ≤ 25 (p = 0.767) and PV > 25 (p = 0.169) also had no correlation with age (Table 3). However, in patients under 70 years old, there was a positive significant correlation between PV and age (Figure 1).

Figure 1. Relationship between PV and age. PV showed no correlation with age (r = 0.071, p = 0.290, N = 226). However, in patients under 70 years old, there was a significant positive correlation between PV and age (r = 0.284, p = 0.008, N = 85).

Table 3. Pearson correlations between PV (all, ≤25, >25) and the variables.

	All patients	PV ≤ 25	PV > 25
Testosterone	0.149*	0.053	0.287†
	(0.163*)	(0.068)	(0.303†)
E2	0.112	0.091	0.193*
	(0.135*)	(0.117)	(0.213*)
DHEAS	−0.071	0.103	−0.058
	(−0.049)	(0.120)	(−0.001)
Height	0.037	0.136	−0.117
	(0.071)	(0.172)	(−0.081)
Body weight	0.046	0.293†	−0.117
	(0.081)	(0.330†)	(−0.084)
BMI	0.028	0.258†	−0.074
	(0.045)	(0.269†)	(−0.056)
Smoking	0.068	0.049	0.072
	(0.067)	(0.048)	(0.082)
Alcohol	−0.169*	−0.237*	−0.111
	(−0.172*)	(−0.246*)	(−0.102)
Alopecia	0.071	−0.098	0.094
	(0.049)	(−128)	(0.061)
Age	0.071	0.030	0.124

*p < 0.05; †p < 0.01; ( ) adjusted by age.

Testosterone

PV was significantly correlated with T in 226 patients (Table 2, Figure 2a). But, PV ≤ 25 had no correlation with T (p = 0.596), on the other hand PV > 25 had significant correlation with T. PV and PV > 25 also had significant correlation with T after age adjustment, as shown in Table 3 by the figures in brackets.

Figure 2. Correlation between PV and sex hormones. (a) PV and testosterone (r = 0.149, p = 0.025), (b) PV and E2 (r = 0.112, p = 0.094), (c) PV and DHEAS (r = −0.071, p = 0.290). PV significantly correlated with testosterone. N = 226.

Table 2. Pearson correlations between variables of all patients.

	PV	T	E2	DHEAS	Smoking	Alcohol	Alopecia
PV	–	0.149*	0.112	−0.071	0.068	−0.169*	0.071
		(0.163*)	(0.135*)	(−0.049)	(0.067)	(−0.172*)	(0.049)
Testosterone	0.149*	–	0.343†	0.003	0.047	−0.060	0.054
	(0.163*)		(0.307†)	(0.035)	(0.038)	(−0.049)	(0.100)
E2	0.112	0.343†	–	0.044	0.128	−0.077	0.035
	(0.135*)	(0.307†)		(0.081)	(0.125)	(−0.069)	(0.042)
DHEAS	−0.071	0.003	0.044	–	−0.112	0.026	−0.320†
	(−0.049)	(0.035)	(0.081)		(−0.115)	(−0.020)	(−0.186†)
Height	0.037	−0.102	−0.013	0.254†	−0.027	−0.030	−0.190†
	(0.071)	(−0.090)	(−0.029)	(0.114)	(−0.020)	(−0.063)	(−0.060)
Body weight	0.046	−0.250†	0.088	0.194†	−0.004	0.017	−0.106
	(0.081)	(−0.240†)	(0.119)	(0.012)	(0.006)	(−019)	(0.063)
BMI	0.028	−0.245†	0.103	0.074	0.005	0.034	−0.013
	(0.045)	(−0.231†)	(0.140*)	(−0.054)	(0.010)	(0.011)	(0.100)
Smoking	0.068	0.047	0.128	−0.112	–	−0.041	0.023
	(0.067)	(0.038)	(0.125)	(−0.115)		(−0.049)	(0.008)
Alcohol	−0.169*	−0.060	−0.077	0.026	−0.041	–	−0.022
	(−0.172*)	(−0.049)	(−0.069)	(−0.020)	(−0.049)		(0.001)
Alopecia	0.071	0.054	0.035	−0.320†	0.023	−0.022	–
	(0.049)	(0.020)	(0.042)	(−0.186†)	(0.008)	(0.001)
Age	0.071	0.082	−0.021	−0.409†	0.037	−0.057	0.390†

*p < 0.05; †p < 0.01; ( ) adjusted by age N = 226.

E2

PV was not significantly correlated with E2 (p = 0.094) (Figure 2b), but after adjustment according to age significant correlation was observed (Table 2). PV > 25 was significantly correlated with E2, but PV ≤ 25 had no correlation with E2 (p = 0.363). PV > 25 was also significantly correlated with E2 after age adjustment (Table 3).

DHEAS

PV (p =0.290), PV ≤ 25 (p = 0.306) and PV > 25 (p = 0.522) had no significant correlation with DHEAS (Tables 2 and 3, Figure 2c). But DHEAS was significantly correlated with alopecia (Table 2).

Sex hormone ratios

E2/T ratio had no correlation with PV (p = 0.682), PV ≤ 25 (p = 0.858), and PV > 25 (p = 0.455). T/DHEAS and E2/DHEAS ratio had a significant positive correlation with PV > 25. T/DHEAS ratio still had a significant positive correlation with PV > 25 after age adjustment (Table 4).

Table 4. Pearson correlations between PV and sex hormone ratios.

	All patients	PV ≤ 25	PV > 25
E2/T	−0.027	−0.081	−0.067
	(−0.028)	(−0.015)	(−0.079)
E2/DHEAS	0.046	0.104	0.186*
	(0.034)	(0.101)	(0.155)
T/DHEAS	0.096	0.096	0.226*
	(0.082)	(0.092)	(0.201*)

*p < 0.05; ( ) adjusted by age.

Relationship between sex hormones

T and E2 had significant correlation, but other hormones had no correlation (Figure 3).

Figure 3. Correlation between sex hormones. (a) E2 and testosterone (r = 0.343, p = 0.000), (b) DHEAS and E2 (r = 0.044, p = 0.508), (c) DHEAS and testosterone (r = 0.003, p = 0.968). E2 significantly correlated with testosterone. N = 226.

Sex hormones and age

DHEAS had a strong negative correlation with age. But other sex hormones had no such correlation (Figure 4).

Figure 4. Relationship between sex hormones and age. (a) testosterone (r = 0.082, p = 0.218), (b) E2 (r = −0.021, p = 0.749), (c) DHEAS (r = −0.409, p = 0.000). DHEAS significantly correlated with age. N = 226.

Discussion

Although we have been treating significant numbers of BPH patients for many years at our urological clinics, the exact etiology of BPH has not been fully understood. Historically despite many studies carried out, the trigger for BPH formation is still not yet clear. Sugimura in 1996 summarized the hypotheses of BPH pathogenesis from previous studies [19,20]. He divided them into four theories as follows: (1) the DHT hypothesis; (2) the embryonic reawakening theory; (3) the stem cell theory; and (4) the effect of estrogen. In these theories, androgens, including T and DHT, and estrogen, have been the principal subjects. T is converted to DHT by 5α-reductase, and then binds to an androgen receptor. According to Carson et al, the binding of DHT to the androgen receptor complex, results in a cascade of events necessary for the formation of the signaling factors that regulate cellular growth. Unlike the decrease of circulating T with age [21], DHT does not appreciably decease. Thus, the DHT level remains constant throughout the time when clinical BPH is most likely to appear [2]. Marberger et al., when evaluating the treatment of BPH patients by dutasteride, also suggested that the high levels of 5α-reductase and DHT in the prostate, allow the development and progression of BPH even at low circulating T levels, when considering the effectiveness of dutasteride to BPH patients at all serum T levels [22]. On the other hand, the studies of T therapy for one year in hypogonadal men did not show any apparent increase of PV [23,24]. A similar study for a more than five-year term showed a significant PV increase, however the authors could not ascertain whether the increase was significantly above that of the normal aging population [25]. Considering such findings, T or DHT do not seem to be enough to progress BPH, even though they are effective in prostate growth.

In 1976, Walsh and Wilson reported that they grew BPH in castrated dogs treated with E2 and androstanediol, hormones which are normally produced in testis [26]. Krieg and coworkers examined the effect of aging on sex hormones in the human prostate, and showed that the stromal DHT and T level of the normal prostate and BPH, had no correlation with age. At the same time, the stromal E2 level increased with age in BPH, and was also higher than the epithelial level. This result suggested that the increasing estrogen/androgen ratio in stroma due to aging had an effect on BPH growth [27]. Epidemiologically, the correlation between estrogen and BPH has also been suggested. For example, Partin et al. reported the positive correlation of BPH volume with E2 and free T, when inspecting prostates removed by operation, which had low volume prostatic cancer [5]. Suzuki et al. examined the correlation between total T, free T, and E2, with prostate size estimated by digital examination. They indicated the positive correlation between prostatic size and E2 level [7]. Roberts et al., reported a dose–response relationship of E2 levels with prostate size, estimated by trans-rectal ultrasound among men with high bioavailable T [6]. Some researches have indicated no relationship between E2 and BPH [3], nevertheless, it seems consistent that the synergetic effect of E2 and androgen plays an important role on BPH pathogenesis.

A significant number of epidemiological studies did not show a significant relationship of T with prostatic size [6,7,13,14], in our study however, T and E2 had significant positive correlation with PV, especially in large prostates. This result seems to indicate the synergetic effect of E2 and androgen on BPH pathogenesis, even if there is an association between T and E2. Furthermore, although the estrogen/androgen ratio has been generally considered one of the important factors of BPH growth, we did not find a correlation between E2/T ratio in serum and PV, as Schatzl et al. have already reported [13]. However, we did find that the T/DHEAS and E2/DHEAS ratios had a positive correlation with larger prostates. In other studies, the relationship of BPH, and DHEA or DHEAS, is controversial [3–5,13]. However, the results of our study, do not only show the effect of T and E2 on BPH growth, but also may imply some suppressive role of DHEAS on T and E2.

This study failed to demonstrate serum T decreasing with age. This is in keeping with other Japanese and European studies, which also demonstrated no such correlation [13,28]. However, our study showed significant decrease of DHEAS with aging. DHEAS is a hormone which dramatically decreases with age and is thought to have an anti-aging function [11]. Actually, our study also showed a negative correlation of DHEAS with alopecia. Several studies demonstrated the protective role of DHEA, for example, protecting the mammary gland of rodents from the development of cancer, and also the inhibitive effect of DHEA was shown in the study of prostate adenocarcinoma in rats [29]. From these data, therefore, DHEAS or DHEA may well have a negative effect on BPH growth.

In conclusion, our study showed that PV had a positive correlation with T and E2, especially in larger prostates. This result supports the conventional theory of the synergetic effect of androgen and estrogen on BPH synthesis, by evaluating measured PV and sex hormones. DHEAS had no significant correlation with PV, but may have a suppressive effect on T and E2 in BPH growth considering the relationship between PV and T/DHEAS, E2/DHEAS ratios. DHEAS decreasing with age might be one of the important factors in BPH growth with aging.

Declaration of interest

Authors do not have conflicts of interest.