Exogenous estradiol increases cardiovagal baroreflex sensitivity during a hypertensive stimulus in premenopausal young women

Abstract

Women have a greater reduction in cardiovagal baroreflex sensitivity (CVBRS) as they age which contributes to an elevated cardiovascular disease risk. One potential contributor to the attenuated CVBRS may be reductions in estrogen. Therefore, we tested the hypothesis that isolated estradiol (E₂) and combined E₂ and progesterone (P₄) would increase CVBRS to the same extent. To examine the acute effects of E₂ and P₄ on CVBRS in healthy pre-menopausal women, we used a gonadotropin releasing hormone antagonist to suppresses endogenous sex hormones. We tested 29 young adult women over 10–12 days of hormone manipulation. After 4 days of hormone suppression participants were given either 0.1–0.2 mg transdermal estradiol (E₂) or 200 mg oral micronized progesterone (P₄) per day. Following 3–4 days of isolated hormones, combined hormones (200mg of progesterone and 0.1–0.2mg estradiol) were administered. CVBRS was assessed during a modified Oxford protocol approximately 4 days following each change in hormones. Overall CVBRS was determined as the slope of the response between the R-R interval and systolic blood pressure. CVBRS slope during the hypotensive (sodium nitroprusside) and hypertensive (phenylephrine) phases of the modified Oxford were also assessed. There was no change in the overall R-R Interval or hypotensive CVBRS during any of the hormone trials. Interestingly, the E₂ group’s CVBRS increased in response to the hypertensive stimulus while the P₄ group had no change. These data suggest that estradiol alone augments CVBRS to a hypertensive stimulus, but progesterone alone and combined E₂ and P₄ do not change CVBRS.

Graphical Abstract

NEW & NOTEWORTHY

Attenuated cardiovagal baroreflex sensitivity is associated with cardiovascular disease. Female sex hormones have a significant impact on the development of cardiovascular disease across the lifespan. Using an experimental model of hormone suppression and add-back, estradiol add-back alone increased cardiovagal baroreflex sensitivity to a hypertensive stimulus unlike the other hormone manipulation conditions. These data suggest that estradiol alone plays a significant role in cardiovagal baroreflex sensitivity and reductions may contribute to potential cardiovascular disease development.

INTRODUCTION

Cardiovagal baroreflex sensitivity (CVBRS) is a highly regulated physiological process that captures changes in heart rate or R-R interval to an acute change in blood pressure. The importance of studying CVBRS is further reinforced when we consider that it provides insight into parasympathetic activity of the autonomic nervous system. Dysfunction in this reflex or an attenuated CVBRS is associated with high blood pressure (1) and increased blood pressure variability (2) and, as such, is also predictive of increased cardiovascular disease risk (1, 3–8). Women exhibit greater reductions in CVBRS as they age, specifically after the menopausal transition, compared to men, contributing to higher cardiovascular disease risk (9–13). One potential contributor to the attenuation in CVBRS may be female sex hormones, such as estrogens and progesterone.

The impact of changes in female sex hormones on baroreflex sensitivity, both across the menstrual cycle and with hormonal contraception use, have had varied results (14–20). Furthermore, the equivocal nature of CVBRS and female sex hormone fluctuations have often been further confounded in studies that have utilized the modified Oxford technique and combined the hypo- (sodium nitroprusside) and hypertensive (phenylephrine) stimuli. These data analysis methods have potentially masked interactions between sex hormones and baroreflex sensitivity (21).

Traditionally, a binning method has been utilized in which data from both the falling (hypotensive) and the rising (hypertensive) pressure phases of the modified Oxford are binned into 2 mmHg pressure bins. However, combining all data may mask differences in CVBRS to hypo- versus hypertensive stimuli. One way to separate these responses is to perform a piecewise linear regression, first proposed by Studinger et al. (21) and also subsequently performed by Wenner et al. (22) with a similar hormone suppression/ add-back model to what we have employed. This model divides CVBRS during the hypotensive stimuli (pressure decrease) and the hypertensive stimuli (pressure increase). Studinger et al. found that the falling (decrease) and rising (increase) pressure phases in all of their participants demonstrated hysteresis or a ‘delayed’ response that was indicative of changes in the interaction of set point and gain during each pressure change (21). Therefore, we have chosen to use this method and isolate the hypo- and hypertensive stimuli as it may allow for an understanding of how fluctuations in female sex hormones across the menstrual cycle and the lifespan may govern hypo- and/or hypertensive events that occur due to environmental factors (i.e., stress, sleep deprivation, etc.,).

Those that have separated these responses across various hormone phases have reported differential results. For example, Hiroshoren and colleagues reported that CVBRS was elevated during a hypertensive stimuli in the mid and late luteal phases (16). However, Tanaka and colleagues found no difference in CVBRS in response to a hypertensive stimulus during the midluteal phases yet reported an increased CVBRS during a hypotensive stimulus during the early follicular phase (17). Furthermore, they also reported that during the preovulation phase, when estrogen levels are rising but progesterone (P₄) levels remain low, CVBRS was increased during a hypertensive stimulus (17). Taken together, there is no consensus as to the effect female sex hormones have on CVBRS.

Investigations that have been conducted in humans have generally utilized the endogenous hormone fluctuations and measured the associated changes in baroreflex sensitivity. Unfortunately, this method has the influence of multiple female sex hormones at one time. Therefore, to better understand the independent and combined effects of female sex hormones on cardiovascular regulation, our group has previously used a tightly controlled model of hormone suppression and add-back and have subsequently implemented this model in this research (23, 24). This model allows us to strictly regulate the female sex hormones at each time point throughout our investigations giving us a clear indication of the impact each has on CVBRS.

Thus, the purpose of this study was two-fold, first, to investigate whether estradiol (E₂) and/or P₄ independently changes CVBRS. Second, to determine if E₂ and P₄ independently and in combination change CVBRS during the hypotensive and hypertensive stimulus of the modified Oxford. We hypothesized that by using a tightly controlled model of hormone suppression and add-back, E₂ alone would increase overall R-R interval CVBRS, whereas P₄ would not increase CVBRS and would counteract the effects of E₂ alone when administered first. Furthermore, we hypothesized that combined administration of E₂ and P₄ would result in a similar CVBRS as E₂ alone when E₂ was administered first. Finally, we hypothesized that there would be E₂ and P₄ dependent changes in CVBRS during the isolated add back and combination hormone administration when we separated the hypotensive and hypertensive stimulus.

MATERIALS AND METHODS

This study was approved by the Institutional Review Board at the University of Oregon (Protocol #: 03152011.051). Written informed consent was obtained from all participants and the study conformed to the principles of the Declaration of Helsinki, except for registration in a database.

Participants

Twenty-nine healthy, nonsmoking, recreationally active females (activity limited to five hours per week of light or moderate exercise), between the ages of 18 and 31 participated in this study (See Table 1 for demographic information). A subset of these subjects (n=20) also participated in an endothelial function protocol, and results are published elsewhere (24). Subjects were excluded if they were taking any prescription medications, except for combined hormonal contraceptives (n=20; see Table 1 for detailed list of hormonal contraceptives used prior to hormone suppression). Those who were using hormonal contraceptives were asked to discontinue use throughout the duration of the study. Participants were required to have a 30-day washout from any hormonal contraceptives they were previously taking. Washout was confirmed by the start of regular menstruation. Prior to participating in the study, participants were screened by researchers and our collaborating physician specializing in gynecological medicine. Women were excluded if they had a history of cardiovascular disease, hypertension, hypercholesterolemia, diabetes, medical allergies, clotting disorders, endocrine and/or menstrual disorders, and or had recently undergone any surgical procedures.

Table 1.

Participant Demographics

Group, N	Estradiol-first 13	95% CI	Progesterone-first 15	95% CI
Age, yr	22 ± 3	[20, 23]	21 ± 3	[20, 23]
Weight, kg	61.4 ± 9.2	[56.0, 66.7]	59.7 ± 7.3	[55.6, 63.7]
Height, cm	165.3 ± 7.3	[161.1, 169.5]	162.1 ± 6.8	[158.3, 165.9]
BMI, kg/m2	22.4 ± 2.9	[20.7, 24.1]	22.7 ± 2.8	[21.1, 24.2]
Hormonal manipulation, N	8		12
Aviane	1		1
Desogen			2
Jolessa			1
Junel	1
Kariva			1
Levora	1
Nuva Ring	1		4
Ortho-Evra	1
Orthotricylcin Low	1
Reclipsen			1
Solia			1
Yasmin	1
Yaz	1		1

Data are reported as mean ± standard deviation, with 95% confidence intervals in the adjacent column. BMI, body mass index. All hormonal manipulation was stopped 30 days prior to the first experimental visit.

Participants were required to abstain from exercise, alcohol, vitamins, and over-the-counter medications for at least 24 hours and caffeine for 12 hours prior to each testing visit. Participants also abstained from food and drink (with the exception of water) for 12 hours prior to each morning testing visit. For the afternoon visits, participants were asked to eat a light, low-fat meal no less than two hours prior to the testing start time to prevent hypoglycemia and were asked to replicate their meals prior to each testing visit.

Experimental Design

We used a hormone suppression model to independently determine the effect of E₂ and P₄ on CVBRS. Endogenous hormone suppression began within three days of menstruation (day 1; see Figure 1) and continued for the entirety of the experiment (10–12 days). Female sex hormones were suppressed via daily subcutaneous injection of a gonadotropin-releasing hormone antagonist (GnRHa). This GnRHa suppression model fully suppresses endogenous estrogens and progesterone within 36–48 hours of the first injection (25).

Figure 1.

Experimental timeline. Day 1 was the start of the experimental timeline with administration of hormone suppression (GnRHa) alone for 4 days prior to testing. Day 4 was the first experimental testing day followed by assignment to either estradiol (E₂) first or progesterone (P₄) first. Day 7–8 was testing visit 2 followed by the combined hormone administration (E₂ + P₄ or P₄ + E₂), finally finishing with experimental visit 3. Created in BioRender.

On day 4 of hormone suppression, participants underwent the first CVBRS experimental testing and were randomized to first receive either E₂ “add-back” (0.1–0.2 mg transdermal estradiol patch) or P₄ “add-back” (oral administration of 200 mg of progesterone daily). Participants returned to the laboratory on day 7 or 8 for testing and then started the combined E₂ and P₄ administration (either E₂ + P₄ or P₄ + E₂ dependent on which sex hormone was administered first). Participants returned to the laboratory for final testing between days 10–12.

Hormone Manipulation

GnRHa was administered daily through subcutaneous injections by participants. Injections were 250 μg/0.5 ml of the GnRHa, ganirelix acetate (Organon International, Roseland, NJ). Participants were trained by research personnel in how to administer the subcutaneous injections independently prior to leaving the laboratory. E₂ transdermal patches (31mm²; Estradiol; Mylan Pharmaceuticals Inc, Morgantown, WV) were required to be worn for a week and contained 3.88 mg of estradiol USP. Each patch delivered either 0.1 mg (single-estradiol group; N = 9) or 0.2 mg (double-estradiol group; N = 5). P₄ (Prometrium; Solvay Pharmaceuticals, Marietta, GA) was administered orally at a dose of 200mg and was a bio-identical equivalent to endogenous progesterone. Combination hormone administration included both modes of delivery for E₂ and P₄ (i.e., transdermal patch + oral progesterone). The order of administration was dependent on which hormone was administered first.

Measurements

Laboratory temperature was maintained between 21 °C – 23 °C for all experimental visits. Time of day for all visits were kept consistent for each participant. Baseline measures were completed following 20 minutes of supine rest. Blood samples were collected and centrifuged at 1300 g for 15 minutes at 4 °C, separated and stored at −80 °C within 30 minutes. Plasma and serum were later analyzed for circulating E₂ and P₄ concentrations offsite by the Oregon Clinical and Translation Research Institute (OCTRI) Core Laboratory.

Heart rate and rhythm were measured continuously throughout the protocol (CardioCap; Datex-Ohemda, Louisville, CO) via five-lead electrocardiogram. A beat-by-beat blood pressure cuff was placed on the middle finger, and pressure was continuously measured via the Peñaz method (Nexfin; BMEye, Amsterdam, The Netherlands) (26, 27). The electrocardiogram and blood pressure tracings were digitized and recorded using a signal-processing software system (Windaq; Dataq Instruments, Akron, OH).

CVBRS was assessed using the modified Oxford technique as previously described (22, 28, 29). Briefly, the modified Oxford technique is as follows: a 100 μg sodium nitroprusside (Nitropress; Hospira Inc, Lake Forest, IL) bolus is infused via IV catheter, followed 60 seconds later by a 150 μg phenylephrine HCL (Phenylephrine; Sandoz Inc, Princeton, NJ) bolus (30). Heart rate and blood pressure are recorded for two additional minutes following the administration of phenylephrine, for a total of three minutes. Participants rested for at least 20 minutes between the first and second modified oxford technique.

Data and Statistical Analysis

These data were collected and stored off-line for later analysis (Windaq; Dataq Instruments, Akron, OH). R-R Intervals and systolic blood pressure were synchronized over the cardiac cycle and analyzed using the traditional binning method (30) and the piecewise linear regression method (31).

Traditional Binning Method:

To obtain an overall measure of CVBRS, R-R intervals were pooled over 2 mmHg bins, and bins were weighted based on the number of cardiac cycles contained within a given pressure bin. The linear portion of the curve was identified via a combination of visual inspection and by comparing the weighted slopes. To ensure accurate identification, pressure bins during the plateau response were removed. Agreement by at least two separate researchers (JCM, VEB), who were blinded to the phase of hormone manipulation, confirmed identification of the linear portion of the curve.

Piecewise Linear Regression:

To differentiate the sensitivity of the cardiovagal baroreflex in response to a hypotensive stimulus (decreasing pressure, sodium nitroprusside) and hypertensive stimulus (increasing pressure, phenylephrine) we analyzed the R-R intervals and corresponding systolic blood pressure for each cardiac cycle during the hypo- and hypertensive stimuli, respectively. The hypotensive stimulus was defined as the onset of a decrease in pressure after sodium nitroprusside administration and continued until it reached a nadir. The hypertensive stimulus was defined as the onset of increasing pressure following phenylephrine administration and continued until the peak pressure was reached. To avoid subjectivity, the linear portion for each of these curves was determined using a piecewise linear regression analysis method, requiring a minimum of five data points to determine the presence of a threshold and/or saturation (31). The presence of a threshold and/or saturation was determined by repeatedly regressing the unweighted data, starting with the first five data points. The regression that returned the smallest error sum of squares, and a significance value of P > 0.05, was identified as the threshold/saturation. If these values were reached using only the first five data points, this was indicative of no threshold/saturation. After the linear portion of the curve was identified, outliers were detected by assessing Cook’s Distance values (32), and removed accordingly.

Data are reported as mean with 95% confidence intervals and/or standard deviations. Alpha was set at 0.05 for all statistical inferences including familywise error rates. All data inferences were drawn from two-way repeated measures mixed-effects models (for measurements across time and between groups) with pre-planned comparisons via contrasts. Statistical differences from our mixed effects model were evaluated using a Tukey post-hoc analysis (JMP 18.1, SAS Institute). We also assessed our data for outliers. Based on our analysis, it was determined that any calculated outliers did not impact the statistical data interpretation and are physiologically relevant values. Cohen’s d effect sizes were also calculated for blood pressure findings that were below our pre-determined alpha value. Simple linear regressions were used to analyze the relationship between change (combination hormone trial – hormone suppression trial in each group) in E₂:P₄ ratios (ΔE₂:P₄) and the associated change in CVBRS measures (i.e., R-R interval overall slope, hypotensive slope, and hypertensive slope) Other correlations between E2 or P4 and the slopes of the overall CVBRS R-R interval, hypotensive, and hypertensive slopes were also analyzed (GraphPad Prism 10.4.2, GraphPad).

RESULTS

Participant Characteristics

Participant demographics were well matched between groups and are shown in Table 1. E₂ and P₄ levels measured following four days of hormone suppression were lower compared to hormone levels measured during each isolated add-back (E₂, P < 0.01; and P₄, P <0.01) and the combined E₂ + P₄ or P₄ + E₂ conditions (P <0.01; see Figure 2). Degrees of freedom (DF) and F ratio for estrogen and progesterone group x time interaction and trial were as follows. Estrogen: DF = 2 and F ratio = 20.2 (interaction), DF = 2 and F ratio = 32.3 (trial). Progesterone: DF = 2 and F ratio = 11.4 (interaction), DF = 2 and F ratio = 52.0 (trial).

Figure 2.

Estradiol and progesterone concentrations for each condition. Estradiol and progesterone were successfully reduced following hormone suppression (GnRHa). Add back of estradiol (E₂) and progesterone (P₄) increased concentrations of each hormone, and the combined add-back (E₂ + P₄ and P₄ + E₂) increased both hormones compared to the hormone suppression condition. * P < 0.01 vs. hormone suppression within group; †P < 0.01 vs. hormone suppression between groups; ‡ P < 0.01 between groups at that time point; # P <0.01 vs. E₂

Hemodynamics

Systolic blood pressure was 2–3 mmHg lower during the combined hormone administration than with hormone suppression or isolated add-back (E₂ or P₄; P < 0.05). Diastolic blood pressure and mean arterial pressure were higher during the hormone suppression and isolated add-back phases than in the combined hormone administration phase (P < 0.01). Heart rate did not change during any of the hormone manipulation trials. Effect sizes calculated as Cohen’s d were calculated for comparisons. The E₂ first group blood pressure comparisons had a medium effect (Cohen’s d: 0.60 – 0.72), whereas the P₄ first group had a small effect (Cohen’s d: 0.18 – 0.36). Thus, indicating that the differences in blood pressure had only a small or moderate difference. Data are reported in Table 2. DF and F ratio for heart rate, systolic blood pressure, diastolic blood pressure, and mean arterial pressure group x time interaction and trial were as follows. Heart rate: DF: 2 and F ratio = 0.03 (interaction). Systolic blood pressure: DF = 2 and F ratio = 1.86 (interaction), DF = 2 and the F ratio = 3.8 (trial). Diastolic blood pressure: DF = 2 and F ratio = 0.91 (interaction), DF = 2 and F ratio = 7.1 (trial). Mean arterial pressure: DF = 2 and the F ratio = 0.25 (interaction), DF = 2 and the F ratio = 8.1 (trial).

Table 2.

Hemodynamics

	Estradiol-First Group
	Hormone Suppression	95% CI	Cohen’s d	E2	95% CI	Cohen’s d	E2 + P4	95% CI	Cohen’s d
Heart Rate, bpm	60 ± 12	[55, 65]	--	59 ± 10	[54, 64]	--	61 ± 8	[55, 66]	--
Systolic Blood Pressure, mmHg	111 ± 6	[107, 114]	--	108 ± 6	[104, 111]	--	107 ± 5 *	[105, 111]	0.72
Diastolic Blood Pressure, mmHg	70 ± 5 †	[66, 73]	0.72	70 ± 7 †	[66, 73]	0.61	66 ± 6	[63, 69]	--
Mean Arterial Pressure, mmHg	83 ± 5 †	[80, 86]	0.60	82 ± 7 †	[79, 85]	0.34	80 ± 5	[77, 83]	--
	Progesterone-First Group
	Hormone Suppression	95% CI	Cohen’s d	P4	95% CI	Cohen’s d	P4 + E2	95% CI	Cohen’s d
Heart Rate, bpm	61 ± 10	[56, 66]	--	59 ± 7	[54, 64]	--	61 ± 8	[56, 66]	--
Systolic Blood Pressure, mmHg	109 ± 6	[106, 113]	--	110 ± 6	[107, 113]	--	107 ± 7 *	[104, 111]	0.31
Diastolic Blood Pressure, mmHg	70 ± 6 †	[67, 73]	0.36	69 ± 6 †	[66, 72]	0.18	68 ± 5	[64, 71]	--
Mean Arterial Pressure, mmHg	83 ± 6 †	[80, 86]	0.36	83 ± 6 †	[80, 86]	0.36	81 ± 5	[78, 84]	--

Data are reported as mean ± standard deviation. The adjacent column is the 95% confidence interval (CI);

^*P < 0.05 vs. Hormone Suppression

^†P < 0.01 vs. combined hormone, either P₄ + E₂ or E₂ + P₄; Effect sizes calculated as Cohen’s d are shown for comparisons.

Traditional Binning Method: R-R Interval CVBRS

Overall R-R interval CVBRS, as analyzed by the traditional binning method, is shown in Figure 3. There were no differences in CVBRS slope across any of the hormone manipulation conditions (P = 0.3). DF and F ratio for the overall R-R interval CVBRS group x time interaction and trial were as follows: DF: 2 and F ratio = 1.1 (interaction) DF = 2 and the F ratio = 1.4 (trial).

Figure 3.

Overall binning method of analysis for CVBRS calculating the R-R Interval slope. There were no differences between any phase of hormone manipulation (P = 0.3).

Piecewise Linear Regression Analysis Method

Similarly, there were no differences in CVBRS during hormone suppression, isolated hormone administration, or combined hormone administration for either E₂ or P₄ first groups following the hypotensive stimulus (P ≥ 0.3). In response to a hypertensive stimulus, CVBRS was increased with E₂ administration alone vs. hormone suppression (P < 0.01), but the combined hormone administration had no increase (P = 0.1). Separated hypotensive and hypertensive responses are shown in Figure 4. DF and F ratio for the hypotensive and hypertensive slopes group x time interaction and trial were as follows; Hypotensive slope: DF = 2 and F ratio = 0.3 (interaction) DF = 2 and the F ratio = 0.3 (trial). Hypertensive slope: DF = 2 and F ratio = 0.7 (interaction) DF = 2 and the F ratio = 4.1 (trial).

Figure 4.

Separating the hypotensive (sodium nitroprusside) and hypertensive (phenylephrine) stimuli used during the modified Oxford protocol using Piecewise Linear Regression Analysis. In the E₂ first group there was an increase in CVBRS hypertensive slope following E₂ add-back which was maintained with combined E₂ + P₄. *P = 0.02 vs. Hormone Suppression in the E₂ group.

The correlations between ΔE₂:P₄ ratio and ΔCVBRS for overall, hypotensive, and hypertensive slope are shown in Figure 5. There was no correlation (P > 0.2 for all correlations) between the ΔE₂:P₄ ratio and ΔCVBRS for overall R-R interval, hypotensive, or hypertensive slope in either group (E₂ first or P₄ first). Correlations between E₂ and P₄ for overall CVBRS R-R interval, hypotensive, and hypertensive slope can be found in the Supplemental data (Figures S1–S3). There was no relationship (r² ≤ 0.22) between different phases of hormone manipulation (GnRH, E₂, P₄, E₂ + P₄ or P₄ + E₂) and CVBRS slopes.

Figure 5.

Correlations between the change (Combination hormone trial – hormone suppression trial; Δ) in E₂:P₄ ratio and overall R-R Interval slop, hypotensive slope, and hypertensive slope. There was no correlation between the Δ E₂:P₄ ratio and any measure of CVBRS in either group (E₂ first or P₄ first).

DISCUSSION

The purpose of this study was to differentiate the independent and combined effects of E₂ and P₄ on CVBRS. Furthermore, we also sought to determine whether there were independent and combined hormone administration effects during the hypo- and hypertensive stimulus of the modified Oxford. Contrary to our hypotheses, we found no differences in CVBRS when we analyzed the overall R-R interval CVBRS using the traditional binning method. Interestingly, when we analyzed CVBRS using a piecewise linear regression analysis, E₂ alone increased CVBRS during a hypertensive stimulus but the combination hormone administration yielded no increase compared to hormone suppression. Furthermore, there was no change in CVBRS with P₄ alone or when combined with E₂, nor were there any differences in CVBRS between either group or trial for the hypotensive stimulus. These data indicate that E₂ alone contributes to an increased CVBRS, and that hormone order has no effect on the response. Interestingly, we also report the direct impact that hormone manipulation has on blood pressure, with the combination of hormones reducing systolic, diastolic, and mean arterial pressure compared to the other hormone conditions.

Independent effects of E₂ and P₄ on CVBRS

E₂ alone increased CVBRS to the hypertensive stimuli while P₄ alone did not change CVBRS and antagonized the effects of isolated E₂. Although when both the hypertensive and hypotensive stimuli were binned together with the traditional method, as reported in the R-R interval CVBRS slope, there was no change with hormone suppression, isolated add-back, or combined administration.

Previous work by members of our group tested CVBRS in premenopausal women who were taking oral contraceptives (33), their results differed from that of the naturally cycling women (34). During the low hormone phase (which they liken to the early follicular phase) CVBRS was increased compared to the high hormone phase in which CVBRS was decreased (33). Parallels between overall R-R Interval CVBRS during their high hormone phase and our combined hormone administration trial would lead to a hypothesis that CVBRS would be reduced when we administered both hormones together. This was not what occurred and, in fact, CVRBS was not different across all the trials.

Moreover, between our work and that of Minson and colleagues in women who were taking oral contraceptives, we should note that the concentrations of ethinyl estrogen and progestins were not the same (33). Minson et al. tested women who were taking a monophasic oral contraceptive where the ethinyl estrogen concentration was 30–35 µg combined with a low-dose progestin (33). In our work we administered isolated E₂ and P₄ add-back of roughly 3.88mg in transdermal patches and 200 mg oral tablet respectively. It may be that had the concentrations of E₂ been higher it would have yielded results similar to their data (33), but this may be unlikely given the exogenous hormone suppression that we administered concomitantly. Thus, while the isolated add-back in our work does provide insight into the potential effect of exogenous hormones on aspects of parasympathetic function; our work better describes how each hormone individually and together may impact CVBRS and the associated aspects of autonomic function.

The interesting aspect of our data is the increase in CVBRS following estrogen administration during the hypertensive stimulus. Mascone and colleagues reported that in women who were naturally cycling and in women who were taking oral contraceptives those who had higher concentrations of estrogen had a reduced CVBRS during the low hormone phase (20). We did not see those same results. It is well documented that E₂ is a cardio-protective hormone (35–37). Thus, we speculate that the reason for the increase in CVBRS during a hypertensive stimulus may be due to the downstream effects of estrogen (such as increasing nitric oxide bioavailability which may have a direct impact on myocardiocyte health and function (38, 39)) that allow for rapid transient adjustments to a hypertensive stimulus ( environmental effects of stress such as a lack of sleep or daily stressful events).

We had hypothesized that P₄ would not increase CVBRS and would counteract the increase seen from E₂ administration. This hypothesis came from work that showed P₄ has a role in activating the renin-angiotensin-aldosterone system (40) increasing both peripheral vasoconstriction and arterial pressure (41, 42). As this influences a separate aspect of autonomic function, we had supposed there would be a counteracting effect of P₄ on CVBRS. Especially given the previous work indicating that P₄ and E₂ have counteracting effects on autonomic function. Surprisingly we did not see any effect of P₄. Notably, isolated increases in P₄ concentration alone are not seen endogenously and only exogenously would P₄ be administered in isolation such as oral contraceptives that are progestin only. Therefore, these results may be beneficial as a number of hormonal manipulation methods for women use only progestins. In summary, our data suggest that some exogenous ovarian sex concentrations may have a limited impact on the CVBRS and cardiovascular disease risk.

Effects of Combined E₂ and P₄ on CVBRS

The last phase of our experimental design included the combined administration of both E₂ and P₄. Interestingly, CVBRS did not change thereby demonstrating the resilience of CVBRS with changing circulating hormone concentrations. These data could also demonstrate a potential maladaptation in physiological responsiveness, given that adjusting to various exogenous circulating hormone concentrations would be beneficial to regulating CVBRS. In mice, Goldman and colleagues reported that there is a fluctuation in cardiovagal spontaneous baroreflex sensitivity across the estrous cycle (43). Whereas in previous work by members of our group, CVBRS was not different between two menstrual cycle phases (between the follicular and luteal phases of the menstrual cycle, i.e., where there are large adjustments in endogenous circulating E2 concentration) (34). Our data do mirror that of Minson and colleagues (35), with the exception that we were not measuring CVBRS during natural hormone fluctuations but instead, exogenous hormone administrations.

We administered bio-identical P₄, which is quite different from the wide array of progestins used in oral contraceptives. It has been demonstrated in previous studies by our group that the exogenous form of P₄ used in oral contraceptives (i.e., progestins) can greatly impact vascular regulation by countering estrogens’ benefits (24, 44, 45). Thus, it is interesting that there was no change in CVBRS following combination P₄ administration during both the overall CVBRS and the separated hypo- and hypertensive stimuli.

It could be that in premenopausal healthy women, P₄ does not influence parasympathetic activity to the extent that it may later in life. Another explanation is that P₄ does not have an impact on acute transient stimuli such as those agents used in the modified Oxford technique but instead may influence cardiovascular and autonomic responses over time given increased exposure to hypertensive stimuli (e.g., stress, sleep deprivation, etc.,) or hypotensive responses (e.g., autonomic dysfunction, heat events).

Ideally, exogenous hormone manipulation (i.e., hormonal contraception, hormone replacement therapy) would have no adverse effect on cardiovascular or autonomic function, but at this time the results are equivocal, between previous work and our data. Previous work from members of our group have reported that the placebo phase of a combined hormone oral contraceptive increased CVBRS compared to the high hormone or “active” phase when ethinyl estrogen-progestin are taken (33, 46). Mascone and colleagues reported similar results in that the low hormone phase of women taking hormonal manipulation had an increased CVBRS (20).

Furthermore, although we have demonstrated that E₂ increased the hypertensive response in CVBRS, when the combined hormones were administered, namely the P₄ + E₂, the hypertensive slope was not different compared to the P₄ only condition. One possibility for this discrepancy is that the combination of estrogen and progesterone have a counter-balance effect where both acting together produce a zero-sum effect on autonomic function, similar to that proposed by Fu and colleagues (47). As we stated in the introduction, understanding the responses to the pharmacological hypotensive and hypertensive stimuli could be indicative of hypotensive and/or hypertensive events that may occur across the lifespan. These events may occur at the same time as hormone fluctuations or because of a decline in hormones (i.e., post-menopausal hypertension, (48, 49); hot flash induced hypotension, (50)). If we can understand the responses due to isolated estrogen and progesterone, we may be able to determine a cause and effect relationship between environmental events and endogenous hormones with greater certainty. That said, an initial rise in P₄ prior to a rise in E₂ is not a physiological pattern within the normal menstrual cycle and therefore should be considered during interpretation.

Additionally, we analyzed the relationship between ΔE2:P4 ratio and Δ in slope of the CVBRS in the overall, hypo-, and hypertensive conditions and there was no relationship. This may indicate that the ratio of hormones, especially E2:P4 which gives an indication of estrogen dominance (51), is not the best way to assess the relationship between hormone fluctuation and cardiovascular function or health.

Impact of hormone manipulation on blood pressure.

Interestingly, blood pressure changed during the various phases of hormone administration during our investigation. P₄ administration alone had no effect on systolic, diastolic, and mean arterial pressure. In contrast, E₂ alone reduced systolic pressure by 3 mmHg but did not change diastolic pressure and reduced mean arterial pressure by only 1 mmHg. The largest reduction of blood pressure was in the E₂-first combined hormone condition. This group had a systolic and diastolic pressure reduction of 4 mmHg and a reduction in mean arterial pressure of 3 mmHg. This is not surprising as E₂ has a positive effect on blood pressure regulation, especially prior to menopause (52).

Furthermore, these are not insignificant reductions, especially in young, healthy women, as the well-established clinically meaningful reductions in blood pressure are 5–10 mmHg for systolic and 3–4 mmHg for diastolic (53). Furthermore, these blood pressure changes also align with the reductions reported from aerobic exercise in normotensive individuals of 2–4 mmHg (54, 55), underscoring the impact that hormone manipulation has on blood pressure. As a caveat to these findings, it is important to note that our participants started out with normotensive, “healthy” blood pressure. As such, while these are substantial reductions in pressure may not be clinically meaningful and only representative of an intact mechanisms of blood pressure regulation.

From our data we established that isolated E₂ administration has a significant, clinically relevant impact on reducing blood pressure in young healthy pre-menopausal participants. Hirshoren and colleagues reported that the fluctuations in endogenous ovarian hormones across the menstrual cycle do not change blood pressure regulation (56). Whereas Shufelt et al. recently reported that combined oral contraceptives cause hypertension with increases of 7–8 mmHg in systolic blood pressure (57). Interestingly, they also report that with progestin only hormonal contraception there is no change in blood pressure (57) fitting with what we report, of no significant change in blood pressure in the P₄ first group.

It is also important to consider the role of luteinizing and follicle stimulating hormones on blood pressure given their role in estrogen regulation and feedback (58, 59). Yu and colleagues report that high levels of follicle stimulating hormone led to the increased production of renin and in turn an increase in blood pressure (60). Furthermore, an imbalance in the ratio of luteinizing hormone to follicle stimulating hormone is a primary marker in polycystic ovary syndrome (61, 62), a syndrome accompanied by hypertension and an increased risk of cardiovascular diseases (63–66). Therefore, more work is needed to understand the role of these hormones in cardiovascular regulation, as the primary focus for most physiologists has often been the effect of estrogen.

Perspectives and Considerations

Perspective on quantifying CVBRS R-R interval v. separated hypo- and hypertensive stimulus responses.

Our results demonstrate that R-R interval CVBRS from a modified oxford is not different between variations of hormone manipulation (i.e., E₂ first or P₄ first add-back or combined administration). When we separate the modified oxford response between hypo- and hypertensive stimuli (e.g., sodium nitroprusside vs. phenylephrine) our results indicate that E₂ alone increased CVBRS during the hypertensive phase, suggesting that E₂ alone increases CVBRS in response to rising pressure. These results would have been masked if we had only used the traditional binning method for analysis.

Early work led by co-authors of this manuscript demonstrated that changes in sex hormone concentrations in women taking oral contraceptives directly affect baroreflex sensitivity using the modified Oxford technique (33). They reported that CVBRS was greater in a low hormone/placebo phase than in the high hormone phase (33). Furthermore, E₂ administration alone increased CVBRS during the hypertensive stimulus. Therefore, we would suggest that especially when trying to interrogate the mechanisms of change in baroreflex sensitivity with hormone manipulation, it is important to also analyze the separate hypo- and hypertensive responses to avoid masking any important physiology.

We also ran various correlation analyses to determine whether there was a relationship between E₂ and P₄ concentrations and CVBRS slopes (i.e., R-R interval, hypotensive, and hypertensive). Based on our analyses there were no relationships between these variables indicating that hormone concentration levels do not impact CVBRS gain.

Experimental considerations

A few experimental considerations are important when interpreting these data. First, the method of administration for E₂ and P₄ were different, and E₂ was administered via transdermal patch in two different concentrations. We initially started this investigation by administering 0.1 mg per day via transdermal patches but following an initial analysis of participant E₂ levels, we noted we were not achieving our desired blood hormone levels, especially in the E₂-first group. Therefore, we chose to increase the transdermal patch dose of E₂ to 0.2 mg per day to ensure E₂ concentrations would reach the desired level. Additionally, we also instructed our participants to change their transdermal patch every three days, instead of the manufacturer’s recommendation of every seven days. In our final analysis we noted no difference in CVBRS between the differing concentrations of E₂ and therefore have collapsed both groups of E₂ first (i.e., 0.1 and 0.2) into one E₂ first group.

Second, as we showed in Figure 2 (bottom panel), progesterone is reported to be at a concentration of 2 ng ml⁻¹ during hormone suppression. Given that the reduction of progesterone in these women was not at zero during hormone suppression, there may be an influence of the low circulating progesterone concentration on CVBRS, primarily in the E₂ first group, but also during hormone suppression.

Third, as this was an initial investigation into the effect of hormone manipulation on CVBRS we chose to study young healthy premenopausal women. Therefore, these data do not extend to populations with hormonal imbalance, gynecological, metabolic, or cardiovascular disease, nor to peri- or post-menopausal women taking hormone replacement therapy.

Conclusion

We tested the hypotheses that hormone manipulation would alter the overall R-R interval CVBRS and the separate responses to hypo- and hypertensive stimuli in the modified Oxford technique of CVBRS using a well-controlled hormone suppression and add-back model. Contrary to our hypothesis, there were no changes across any of our hormone manipulation conditions in overall R-R interval CVBRS. There were also no changes in CVBRS with P₄ administration alone or when we separated the hypo- and hypertensive stimuli. E₂ administration alone was the only condition to increase CVBRS during the hypertensive stimulus. These results help us to understand the independent and combined effects that P₄ and E₂ have on CVBRS. Furthermore, it also helps to establish the impact of exogenous hormone manipulation on CVBRS, providing direction for future investigations into the effect of exogenous hormones on female hormone imbalances and subsequent cardiovascular disease risk.

Supplementary Material

Supplemental figures S1– S3: DOI: https://doi.org/10.6084/m9.figshare.29190791.v1

DATA AVAILABILITY

Source data for this study are not publicly available due to privacy or ethical restrictions. The source data are available to verified researchers upon request by contacting the corresponding author.