The effect of short-term estradiol therapy on clotting and inflammatory markers in older men receiving hormonal suppression therapy for prostate cancer

Abstract

The cardio protective effect of estrogen in women has come under scrutiny as recent evidence from long-term trials has demonstrated negative findings. In contrast, the effect of endogenous sex hormones, specifically estrogen, on cardiovascular disease, inflammation and clotting parameters in men has not been well-studied. Men receiving androgen deprivation therapy for prostate cancer provide a unique model to study the effect of estrogen alone on inflammation and clotting factors. In a short-term randomized controlled trial of 17-β estradiol (E₂) versus placebo, we measured sex hormones, markers of inflammation including homocysteine (HC), C-reactive protein (CRP), interleukin-6 (IL-6) and coagulation factors including fibrinogen, plasminogen activator-inhibitor-1 (PAI-1) and anti-thrombin-III (AT-III) in 27 older men without bone metastases receiving androgen deprivation therapy or neoadjuvant treatment for prostate cancer. After 9 weeks of E₂ treatment, there was no difference in inflammation or clotting parameters between groups, but after 9 weeks of treatment AT-III increased in the E₂ treated group and decreased in the placebo group. CRP, homocysteine and IL-6 did not show any significant differences. We also evaluated the above parameters in 12 men 3 weeks after acute steroid withdrawal with androgen deprivation therapy and found no significant changes.

We found an increase in AT-III in men receiving E₂ which may be related to gonadal steroid withdrawal, but no significant differences in other inflammatory or clotting factor parameters. While the current report is very preliminary in a small group of subjects, further studies are needed to determine the long-term effects of E₂ in this population of hypogonadal men.

Keywords:

• Inflammation
• estrogen
• male
• atherosclerosis

The universally high coronary heart disease (CHD) risk in men and the increased CHD risk in women experiencing early menopause led to the hypothesis that endogenous estrogen and exogenous hormone replacement therapy might be cardioprotective in women, in part due to the lipid lowering effect as well as effects on the vascular endothelium. This was initially supported by observational studies with postmenopausal hormone replacement. However, recent evidence from large randomized controlled trials in postmenopausal women has shown that hormone replacement therapy with or without progesterone causes an increased risk of cardiovascular endpoints [1],[2]. This phenomenon may be related to the role of inflammation and coagulation factors. Studies evaluating this issue in women have demonstrated variable results. Conclusive long-term data are not yet available.

As men make estrogen (E₂) throughout their lives from the aromatization of testosterone (T) primarily in peripheral tissues [3] and levels are higher than in postmenopausal women, it is unclear whether E₂ plays any cardio protective role in men. High dose E₂ was found to increase death rate in men in the Coronary Drug Project [4]. Similar increases in cardiovascular and thrombotic events have been reported when high dose E₂ (diethylstilbestrol) was used to treat prostate cancer [5]. Despite these results, there is a renewed interest in the effect of E₂ on CHD, in part because of reports that E₂ may have beneficial effects on vascular tone. Recent data from Framingham on 2084 men without cardiovascular disease at baseline demonstrated that higher baseline E₂ levels were associated with lower risk of cardiovascular events in older men supporting a positive effect of endogenous estrogen [6]. We have previously reported that oral E₂ in healthy elderly men reduces homocysteine, fibrinogen, PAI-1 concentration and favourably influences lipoprotein levels without increasing markers of thrombotic risk [7]. However, the T levels in these men decreased significantly when they are treated with E₂, due to feedback at the level of the hypothalamus and pituitary gland.

Thus, men receiving luteinizing hormone releasing hormone agonist (LHRH-A) therapy used in the treatment of prostate cancer is a unique model in which to study the role of E₂ on inflammation and clotting parameters. This treatment leads to hypogonadism with low circulating levels of both T and E₂. In the present study we have used this model to add back a physiological dose of E₂ to these older men in order to determine the early effect of E₂ alone on markers of coagulation and inflammation.

Subjects and methods

Study population

Community-living men receiving treatment with LHRH-A therapy, either for failure of primary treatment for localized prostate cancer or as neoadjuvant treatment prior to radiation or seed implantation took part in a study to evaluate the effect of oral estradiol on markers of bone metabolism [8]. Subjects with bone metastases, chronic medical conditions, significant coronary disease or thromboembolic disorders, diseases of bone metabolism or on medications known to cause or treat osteoporosis (other than calcium or Vitamin D supplementation) were excluded.

Twenty-seven subjects were randomized to a 9-week, double-blind placebo-controlled trial of 1 mg/d of micronized 17-β E2 versus placebo (PL). All subjects were receiving depot LHRH-A therapy. Fourteen men were on established LHRH-A therapy (EST group) with clinical Stage T4 disease, and 13 men were receiving neoadjuvant treatment (NEO group) prior to external beam radiation or seed implantation, clinical Stage T2-3 disease. The EST and NEO groups were randomized independently, so that approximately equal numbers of men were in the E2 and placebo arm. The NEO group had measurements prior to initiation of therapy of LHRH-A treatment (screening visit), and 3 weeks after this initial injection, which is designated as their baseline (BL) visit (Figure 1). Twenty-five men completed the study. Sex hormones, markers of inflammation and coagulation factors were measured at screening, baseline and after 9 weeks on study protocol. Each man was instructed to bring remaining tablets to the end of treatment visit for compliance check, as estimated by pill count. Subjects were asked to complete a side effects questionnaire at all visits. All subjects gave informed consent according to the guidelines of the Institutional Review Board of the University of Connecticut Health Center (UCHC).

Figure 1. The figure represents the time-line for the study. Men in the NEO group were screened and then randomized (E₂ versus PL) 3 weeks after their initial LHRH-A injection (baseline), and men in the EST group were randomized 3 weeks after initial screening.

Hormone measurements

Sex-hormone measurements included E₂, E₁, Total T and SHBG. Assays for total T, E₂, E₁ SHBG, FSH and LH were measured by Endocrine Sciences (Calabasas Hills, CA). T, E₂ and E₁ were measured by radioimmunoassay assay (RIA) and SHBG by a competitive binding assay.

Clotting factors

Clotting factors included fibrinogen, measured by a clot-based technique, PAI-1 by ELISA and AT-III by chromogenic assay. Inflammatory markers included homocysteine measured by gas chromatography/mass spectrometry, highly sensitive C-reactive protein (hsCRP) measured by nephelometry (BN-2 Instrument, Dade Behring, Inc., Tarrytown, NY) and IL-6 measured by ELISA.

Data analysis

A series of repeated measures analysis of variance (ANOVA) models were conducted to examine changes in clotting and inflammatory factors over time, by group and time by group interactions. All of the subjects were assessed at baseline (BL) and after 9 weeks of intervention (post-treatment). Several measures were skewed and therefore required transformation for statistical analysis: these included PAI-I (square root transformation), Antithrombin III (natural log transformation), CRP (natural log transformation) and HC (natural log transformation).

Results

Subject characteristics

Baseline characteristics of the 25 men in the E₂ and PL groups are shown in Table I. There were no differences in age, BMI or PSA or hormone levels at baseline between the E₂ and PL groups. The mean duration of LHRH-A therapy in the EST group was 31 months with a range of 1-99 months. There was no difference in mean duration of therapy in the E₂ versus PL group at BL. After 9 weeks of E₂ treatment, mean estradiol levels rose significantly (p < 0.001) to 64 ng/ml in the treated group, with no significant changes in the other hormones in either group (data not shown).

Table I.  Baseline characteristics of study subjects.

Characteristic	E2 group	Placebo
N	12	13
Age (years)	71 ± 2	71 ± 2
BMI (kg/m^2)	27.8 ± 1.2	27.7 ± 1.3
PSA (ng/ml)	7.4 ± 2.8	7.1 ± 2.8
Nl: (X–4.0)
Testosterone (nmol/L)	4.7 ± 2.7	6.8 ± 2.3
(Nl: 10.4–34.7)
Estradiol (pmol/L)	21.1 ± 3.1	27.0 ± 6.3
(Nl: 37–220)
SHBG (nmol/L)	300 + 51	229 + 51
(Nl: 9–111)

Values are means ± standard error (SE).

Clotting and inflammatory markers

Table II shows the results of hormone levels, clotting and inflammatory markers in the E₂ and placebo groups at baseline and 9 weeks post-treatment. There were no differences in any of the parameters by group but fibrinogen did change significantly over time (p = 0.047). AT-III had a significant time by treatment interaction (p = 0.023), which when further analysed showed a significant increase in the PL group (p = 0.014, data not shown). Percentage changes in all the factors were also analysed, and we found no significant differences between the groups or in changes from baseline. When we looked at men in the NEO group receiving placebo for 9 weeks (n = 5), we found a trend towards an increase in AT-III of 13% (p = 0.06) and an increase in HC levels of 11% (p = 0.08)(data not shown).

Table II.  Changes in clotting factors and inflammatory markers at baseline and after 9 weeks of E₂ treatment.

		Baseline		Post-Rx
Marker	Normal values	E2	Placebo	E2	Placebo	P value
Fibrinogen (mg/dl)	150–400	345 ± 68	348 ± 104	354 ± 66	400 ± 115	0.047^a
PAI-1 (IU/ml)	0.0–22.0	11.3 ± 12.2	6.7 ± 9.6	12.1 ± 10.4	11.1 ± 10.4	NS
AT-III	84–124%	98.7 ± 12.3	101.7 ± 17.9	96.5 ± 12.5	107.7 ± 16.9	0.023^b
hsCRP (mg/L)	0.068–8.2	2.1 ± 1.9	2.9 ± 3.6	1.7 ± 1.9	3.9 ± 4.5	NS
HC (umol/L)	4.0–12.0	10.1 ± 2.6	10.8 ± 3.1	10.9 ± 6.2	10.9 ± 2.7	NS
IL-6 (pg/ml)	0.378–10.1	4.3 ± 1.5	4.2 ± 2.1	4.4 ± 1.6	4.3 ± 1.9	NS

Values are mean ± standard deviation.

TPA, total plasminogen activator; PAI-1, plasminogen activator inhibitor-1; AT-III, antithrombin III; CRP, C reactive protein; HC, homcysteine; IL-6, interleukin-6.

*p ≤ 0.05 for significant differenece between groups post-treatment by repeated measures anova; ^aby time, ^btime by group.

Table III shows the results of inflammatory markers in 12 men in the NEO group 3 weeks after initial LHRH-A injection when both T and E₂ levels would be at nadir. There was no significant change from baseline in any of the parameters, however, there was a trend in the fibrinogen which showed a 22% increase from baseline (p = 0.08).

Table III.  Changes in inflammatory markers in 12 subjects 3 weeks after initial LHRH-A injection.

Marker	Normal values	Baseline	3 weeks post LHRH-A	Mean % change from BL	P value (for % change from BL)
hsCRP (mg/L)	0.068–8.2	2.0 ± 1.8	1.8 ± 1.5	−2.7	NS
HC (umol/L)	4.0–12.0	11.2 ± 2.6	11.0 ± 2.5	−0.6	NS
IL-6 (pg/ml)	0.378–10.1	3.7 ± 1.1	4.0 ± 1.6	7.8	NS

Values are mean ± standard deviation.

Discussion

Inflammation is recognized as an important mediator of all phases of atherosclerosis, from the initial recruitment of circulating leucocytes to the endothelial wall to final rupture of the unstable plaque. Historically, high dose E₂ use in the form of diethylstilbesterol (DES) in men with prostate cancer led to an increased risk of cardiac and thromboembolic events [5] and thus went out of favour. In the Coronary Drug Project in the 1970s, high-dose E₂ was also evaluated as secondary prevention in eugonadal men with CAD, but led to an increase in thrombotic events [4]. Recent studies in men have used different formulations and lower doses of E₂ that led to serum levels in the physiologic male range. In the present study we used micronized estradiol, as other formulations of estrogen contain many different estrogen compounds that cannot be accurately measured.

In this double-blinded, randomized, controlled short-term study in hypogonadal men receiving LHRH-A, we showed that the addition of E₂ in hypogonadal men is not associated with significant changes in clotting or inflammatory markers. An earlier study done by our group in eugonadal elderly men before and after 9 weeks of treatment with three different doses (0.5, 1.0 and 2.0 mg/day) of micronized 17β-E₂, showed that oral E₂ reduced homocysteine, fibrinogen, PAI-1 and favourably influenced VLDL, LDL and HDL subclasses. There were no changes in thrombotic markers including thrombin-antithrombin-III complex, prothrombin fragment 1.2, D-dimer, antithrombin activity, protein-C and S and von Willebrand factor antigen or LP(a). However, in that study, subjects were not randomized: they served as their own controls and the analysis included all three E₂ doses. In addition, men on the highest E₂ dose had supraphysiologic levels of estradiol measured [7].

In the current study a significant finding was an increase in AT-III that occurred in the placebo group over time. The placebo group included both men on long standing LHRH-A therapy (EST group) and men new to this therapy (NEO group) in equal distribution. When we analysed the men in the NEO group receiving PL separately, we did find a trend towards an increase in this parameter (p = 0.06). Therefore, it is likely that the men in the NEO group, who were acutely experiencing decline of gonadal steroids (T and E), influenced the PL group results. There was also a trend for an increase in fibrinogen in the PL group, and the explanation for this may be similar to that for the increase in AT-III. In healthy postmenopausal women receiving continuous combined HRT with estradiol and norethisterone acetate, AT-III levels significantly decreased after 3 months of treatment [9] which could theoretically increase the risk of thromboembolic events. Other clinical studies have demonstrated a decrease of fibrinogen with HRT [10]. Thus, the increases in anti-thrombin III and fibrinogen in the PL group in our study likely reflects the severe decline in levels of both T and E₂ in these men.

High HC levels have also been associated with cardiovascular disease [11]. HC levels tend to rise in postmenopausal women [12] and hormone replacement therapy decreases levels in a number of studies [13-16]. Few data about the effects of E₂ on clotting and inflammatory markers are available in men. HC has been evaluated in male to female transsexuals treated with oral and transdermal 17β E₂, and both regimens led to reduced plasma homocysteine levels [17]; however, levels increased after androgen administration in female to male transsexuals [18]. This is supported by data from our previous study of older men, in which 0.5 to 2.0 mg of oral 17β E₂ also reduced homocysteine levels [6]. When we analysed the NEO men in the PL group alone (N = 6, data not shown), we found a significant decrease in homocysteine levels from baseline (pre-LHRH-A) to 12 weeks post-LHRH agonist (p = 0.05) which may imply that the decline of T levels in our subjects was most influential on this effect.

CRP is an inflammatory marker that when elevated, indicates increased cardiovascular risk in both men and women [19]. We did not determine any change in the present study. Ng et al. studied the levels of CRP in two randomized trials after 3 months of androgen treatment with dihydrotestosterone (DHT), a non-aromatizable androgen, or recombinant human chorionic gonadotropin (rHCG) in healthy men >60 yrs with partial androgen deficiency (T < 15 nmol/l). Androgen levels were consistently maintained at eugonadal levels throughout the trials, with E₂ levels markedly increased in the rHCG group only. There was no significant change in CRP in either group after 3 months, and an increase in serum E₂ in men treated with rHCG also had no significant effect on CRP [20]. Data from trials in women supports an increase in CRP due to HRT [2],[21],[22].

In a previous study we have shown that E₂ therapy in healthy elderly men reduced PAI-1 levels [6], but in the current study there were no changes. In postmenopausal women HRT has been shown to increase plasma fibrinolytic activity by reducing PAI-1 levels [23-25]. More studies are needed to clarify this further.

The pro-inflammatory cytokine IL-6 may have a key role in the etiology of CHD. Data in men is limited. In a study of male to female transsexuals [26] with a mean age of 34 years taking very high doses of E₂ (1.5-5.0 mg/d premarin for 6 months), 25 men demonstrated reduced levels of IL-6 compared with 13 controls (p < 0.05). In a small study of 3 different doses of micronized 17B-estradiol for 12 weeks in postmenopausal women, IL-6 decreased by 35% at 12 weeks compared with placebo (p < 0.001) [22]. These reductions in IL-6 have been seen in some HRT trials but not in others [27],[28].

This study is limited by the small number of subjects studied and the short time period of treatment with 1 mg/d of micronized E₂. It is unclear whether the changes noted in these markers would persist with long-term treatment with estrogen. The mixing of the groups (EST and NEO), although evenly distributed, may have had an effect on response as well. These subgroups were subject to statistical analysis; however, they are too small to draw conclusions.

Conclusions

We found no significant differences in inflammatory and clotting factor parameters after short-term treatment with 1 mg/day micronized E₂ in hypogonadal men receiving androgen deprivation therapy. Future long-term studies are warranted to detect whether these observations hold true and if gender differences of clotting and inflammatory markers in response to E₂ treatment exist.

Acknowledgement

Supported by Claude Pepper Older Americans Independence Center Grant award # 5P60-AG13631, a Clinical Assistant Physician award under the General Clinical Research Center NIH Grant # MO1 RR06192 and a Donahue Foundation Award # DF99-072.