A group of 52 community-dwelling women of advanced age (mean age 79.5 ± 2.8; range 76-87 years), consecutive patients of an outpatient geriatric clinic were recruited for the study. The inclusion criteria defined were the following: female sex, age over 75 years, stable clinical status, no history of hormonal treatment.

The exclusion criteria were: mental or functional inability to participate in the study, a history of panhysterectomy, diagnosis of hypo- and hyperthyroidism, adrenal or pituitary disorders, hemodialysis therapy, hormone therapy (including use of anabolic-androgenic steroids) and lack of patient’s consent. To obtain an adequate echocardiographic measurement and eliminate factors potentially affecting cardiac parameters, patients with atrial fibrillation or severe arrhythmias were excluded from the study.

Laboratory tests were performed on fasting serum samples (within one hour after blood collection). Total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), and glucose (GLU) were measured using the DXC700 analyser, Beckman Coulter, Brea, USA. All the tests and analysers applied in the study were certified for routine in-vitro diagnostic.

Body mass index (BMI) and waist to height ratio (WHtR) were calculated using the collected anthropometric data (body mass and height were measured using RADWAG personal weight scales (WPT60 150OW; Radwag Balances and Scales, Radom, Poland) and waist circumference was assessed using the SECA measuring tape (SECA Deutschland, Hamburg, Germany).

A transthoracic echocardiogram was performed and serum hormones levels were determined in all the subjects who met the criteria of the study. Elevated blood pressure (BP) was defined according to the European guidelines for arterial hypertension: ≥ 140 mmHg for systolic blood pressure (SBP) and/or ≥ 90 mmHg for diastolic blood pressure (DBP) (3).

The study protocol was approved by the Bioethics Committee of the Medical University of Lodz in accordance with the Declaration of Helsinki (RNN/345/08/KB) and all the patients gave their written consent for participation in this study.

Serum levels of estradiol, testosterone, follicle-stimulating hormone(FSH), luteinising hormone (LH), dehydroepiandrosterone sulphate (DHEAS), and cortisol were determined in blood samples taken in the morning from the cubital vein from overnight fasting patients. Hormone concentrations were determined using the technology based on competitive or sandwich chemiluminescence immunoassays (CLIA). Measurements were made with sets from Saluggia (Italy or Stillwater, MN, USA) on the LIAISON XL analyzer from DiaSorin Inc.). The testosterone to estradiol ratio (T/E) was calculated by dividing the respective total serum levels after converting both to pmol/L.

Transthoracic echocardiography was performed in all the patients by the same experienced cardiologist with the use of the Vivid S70 ultrasound system (GE Medical Systems, 2018). The parameters of cardiac structure and function were assessed according to the current recommendations (24, 25). The following internal cardiac dimensions were obtained with the use of two-dimensional linear measurements: interventricular septal diastolic dimension (IVSd), left ventricular diastolic dimension (LVDd), posterior wall diastolic dimension (PWd), and left atrial (LA) antero-posterior diameter. Relative wall thickness (RWT) was calculated according to the following formula: 2 × PWd divided by LVDd. Left ventricular mass was automatically calculated using the Devereux formula and indexed to body surface area (LV mass index – LVMI). Left ventricular hypertrophy (LVH) was defined according to the criteria as LVMI > 95 g/m² for women (24). Left ventricular end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), LV ejection fraction (LVEF) were determined using the modified biplane Simpson’s method from the apical four- and two-chamber views. LVEDV and LVESV were indexed to body surface area - LVEDV index (LVEDVI) and LVESV index (LVESVI). Left atrial volume was measured using the biplane area-length method and divided by body surface area (LA volume index – LAVI). The peak early (E), and late-atrial (A) diastolic velocities were obtained by pulsed-wave Doppler from the mitral inflow profile. The E/A ratio was calculated. Pulsed tissue Doppler imaging (TDI) was used to measure the LV peak systolic velocity (S’) and early diastolic velocity (E’) from the median (septal) and lateral mitral annular velocities. Then, the mean values of these parameters were calculated. Additionally, the average E/E’ ratio from septal and lateral measurements was calculated. Systolic and diastolic LV function was interpreted in accordance with the current guidelines (2, 24, 25).

The total sample size required to determine whether a correlation coefficient differs from zero (α (two-tailed) =0.05; β=0.20; r=0.4) is 47. The normality of distribution was verified using the Shapiro-Wilk test. Not normally distributed continuous variables were presented by median and the first (25%) to the third (75%) quartile. Spearman correlations between hormones level and numerical data were calculated. The quantitative variables (between the hypertrophy groups), were compared using the Mann-Whitney U-test. The occurrence of differences between the groups was assessed using a one-way ANOVA, the chi-square test, or Fischer’s Exact test.

After log transformation of the age and testosterone level, multiple linear regression was performed to assess what independent variables determine IVSd, E’, E/E’ ratio and LVMI.

Stepwise logistic regression (odds ratios and corresponding 95% confidence intervals (95%CI) was used to assess which independent variables predicted the presence of hypertrophy. When building the model, the following variables were taken into account: age, testosterone level and hypertension.

Statistical significance was set at p ≤ 0.05. The analyses were performed using Statistica 13.1 (StatSoft Polska, Cracow, Poland)

Numerous studies have revealed the beneficial effect of estrogens or estrogen receptors stimulation on proper cardiovascular function in women. Earlier reports are also confirmed by more recent observations, where the decline of estrogens level and a higher testosterone/estradiol ratio were recognized as the main cause of increased cardiovascular risk in post-menopausal women (30, 31). As an example, in a study conducted by Zhao et al. among 2,834 women at a mean age of 64.9 ± 8 years, a higher endogenous estradiol level was associated with a lower risk of CAD and HF incidence during 12 years of follow-up (31). In relation to the heart muscle, estrogens via receptors ERα, ERβ and GPER have an inhibitory effect on cardiac hypertrophy, inflammation, fibrosis, and oxidative stress, which has been confirmed in animal and humans models (32–35). However, as our results show, these relationships are not obvious in very old women in whom age- and disease-related cardiac remodeling may overlap, especially that associated with HT. In our group of females at mean age of 79.5 years, a positive correlation between serum estradiol level and posterior wall thickness was observed, with no association found with other echocardiographic parameters. Moreover, estradiol levels were slightly higher in the subgroup with LVH (p=0.051); however, they did not correlate directly with any of the echocardiographic indicators of potential impairment of LV diastolic function (E, E/A ratio, E’, E/E’ ratio or LA enlargement). The relationships found to some extent differ from most of previous reports. Nevertheless, what has been recently emphasized, the interaction of estradiol with various receptors can result in different effects depending on multiple factors (36). In our study, about 80% of the women with LVH were hypertensive, which is an important fact in the context of the study by Wittnich et al. (37). In this animal study, a weaker protective effect of estradiol on LVH and LV function was observed in the case of HT, which resulted from a 35% lower content of ERβ in myocardial cells. However, these relationships require more detailed analyses, and the impact of estradiol on the heart should be definitely considered in the context of other factors among which testosterone seems to play a significant role.

As it was demonstrated by several studies, endogenous testosterone has a favorable impact on cardiac and vascular function in men, however, in the case of women, the data are inconclusive. Physiologically, in females testosterone production decreases steadily from middle age, and the serum levels are significantly lower than in men, including individuals of advanced age. The results of the reports on the influence of endogenous testosterone on the cardiovascular system in post-menopausal women vary considerably, showing a whole range of effects, from protective to harmful ones (38–40). A recently published epidemiological study has highlighted the direct correlation between concentration of serum total testosterone and all-cause mortality in 93,314 post-menopausal women followed-up for approximately nine years, as opposed to a group of 154,965 men in whom that risk decreased with higher testosterone levels (41). Additionally, in another study including a similar group of women, higher testosterone levels and testosterone/estradiol ratio significantly enhanced the incidence of CAD and HF events over 12 years (31). In the present study, we found a stronger relationship between testosterone and echocardiographic parameters specific for cardiac remodeling in comparison to other hormones. The association was mostly expressed in cardiac hypertrophy, where a positive correlation with IVSd and LVMI was observed. Moreover, the LVH patients were characterized by a higher serum testosterone level, and testosterone/estradiol ratio.

Pressure overload is an essential cause of pathological LVH, and adequate treatment aimed at blood pressure lowering can prevent or even regress it, which decreases the cardiovascular risk (3, 5, 6, 42, 43). We did not show the association between LVH and SBP and DBP. This may be due to received anti-hypertensive treatment. However, as Muiesan et al. concluded, LVH is more frequent in women than in men, regardless of anti-hypertensive treatment and adequate blood pressure control, which indicates the important role of other factors in the pathophysiology of LVH in females (44).

Similarly, the relationship between the indicators of excessive body weight and LVH, particularly expressed in women, is confirmed in the literature (45, 46). It was also observed in our study in the form of higher values of WHtR (p=0.009), as well as a trend towards higher values of BMI (p= 0.058) in the LVH subgroup.

Among the factors analyzed, only the presence of HT and serum testosterone level were found to be independent factors associated with LVH. This may confirm the recently published results of other authors who report a positive correlation between testosterone level and LVMI in a large group of post-menopausal women (47). It is worth noting that the processes dependent on HT and the action of testosterone may overlap, and testosterone itself is also involved in the etiology of HT (48).

Subramanya et al. observed a lower LVEDV and a higher mass/volume ratio assessed in cardiac magnetic resonance imaging, with a higher testosterone level, which indicates concentric LV remodeling resulting from pathological processes (47). In contrast to our study, the participants were younger (mean age 65 ± 9 years), hypertensive in 48% (vs. 67% in our group), had a slightly lower SBD and DBP, a comparable BMI, and higher total testosterone levels (median 0.9 nmol/l). Although the patients’ characteristics and methods differ to some extent, the direction of the observed changes is consistent.

Contrary to men, in females testosterone may negatively influence myocardial relaxation, as was reported by Olszanecka et al. (49). In this study, the free testosterone level was inversely associated with E/A ratio (β= -0.19, p= 0.05), however, it was only one parameter of LV diastolic function obtained. In the previously cited study, similar relationships have been shown, however, only the indirect assessment of cardiac diastolic function was used (LV mass/volume ratio) (47). Our analysis was carried out in compliance with the current guidelines on cardiac imaging using the recommended cardiac measurements, which enabled us to demonstrate a significant negative relationship between testosterone level and E’ and a positive one with E/E’ ratio. The directions of changes in both parameters indicate worsening LV relaxation, which may contribute to diastolic dysfunction and HFpEF. Interestingly, these correlations appeared within the ranges of generally normal values of the above parameters.

In a large prospective cohort study, lower levels of endogenous testosterone and DHEAS in men and DHEAS in postmenopausal women were associated with the development of heart failure (50). After a median of 9.1 years follow-up, higher dehydroepiandrosterone and estradiol levels were associated with increased LVM in older men (47). Likewise, glucocorticoids may play an important role in cardiac remodeling and progression to heart failure (51, 52). The data on potential role of gonadotropins in an aging heart are especially poor. Some recent studies confirm the possibility of direct actions of gonadotropins beyond the reproductive system (53). Elevated serum FSH has been suggested to increase the risk of atherosclerosis and cardiovascular diseases (54, 55). In one available study in young adults, absolute value of circumferential strain assessed on left ventricular short-axis images at the papillary muscle level (mid-ventricular) was positively correlated to FSH (56).

In the present study, we were not able to show any relationship of DHEAS, cortisol, LH and FSH to echocardiographic parameters. This may be due to the fact that previously investigated populations were much younger in comparison to our older women. It cannot be excluded that in the more advanced age the influence of these hormones on the heart may be alleviated (as for DHEAS or cortisol) or clinically negligible. However, the role of hormone receptors in the heart muscle of older subjects needs further detailed studies.

The strength of our research is unique age group of the women examined. Moreover, the study group is quite homogeneous in terms of several characteristics, such as sex, advanced age, and clinical features. A number of cardiac parameters were precisely obtained and analyzed. Thus, it was possible for us to observe subtle relationships that are mostly not found by other researchers.

The limitations of our study include the small sample size. Other shortcomings include a single measurement of hormones level in relation to echocardiographic parameters and cross-sectional character of the study. The strength of the associations found was moderate and may be affected by other factors than those presented in our study. Further studies are required to expand the knowledge on sex hormones affecting the heart muscle in women at an advanced age.