Combined androgen excess polycystic ovary syndrome and obesity are associated with elevated beat-to-beat blood pressure variability in adult women

Abstract

Androgen excess polycystic ovary syndrome (AE-PCOS) is associated with elevated risk for hypertension and cardiovascular disease. Greater beat-to-beat blood pressure variability (BPV) has emerged as a novel marker of cardiovascular risk. Little is known about beat-to-beat BPV in AE-PCOS, but prior studies suggest blunted baroreflex sensitivity (BRS), a major contributor to beat-to-beat BP regulation. Both obesity and androgen excess are implicated as driving factors of cardiovascular risk in AE-PCOS. We hypothesized that 1) beat-to-beat BPV would be greater in AE-PCOS compared to both women with overweight/obesity (OW/OB) and lean controls; 2) beat-to-beat BPV would be augmented in OW/OB compared to lean controls. We measured beat-to-beat BP (finger photoplethysmography) during ~5–10 minutes of supine rest in 18 lean controls, 14 OW/OB controls, and 14 women with AE-PCOS. Serum testosterone was measured and spontaneous cardiovagal BRS was assessed via sequence technique. In AE-PCOS, resting BP was greater than lean controls (P<0.05) and serum testosterone greater than both control groups (P<0.001). Diastolic BP standard deviation was greater in AE-PCOS (3.9 ± 1.0 mmHg) compared to both control groups (lean: 3.0 ± 1.5; OW/OB: 2.9 ± 1.2 mmHg; P=0.007). Cardiovagal BRS was not different between groups (P=0.196). Neither cardiovagal BRS nor serum testosterone were related to beat-to-beat BPV in AE-PCOS. Collectively, these data suggest greater beat-to-beat BPV in AE-PCOS which cannot be accounted for by obesity alone and is not associated with cardiovagal BRS or circulating androgens.

NEW AND NOTEWORTHY

This is the first study to comprehensively examine beat-to-beat blood pressure variability (BPV) in women with androgen excess polycystic ovary syndrome (AE-PCOS). Our results suggest early alteration(s) in beat-to-beat BPV in women with AE-PCOS before the manifestation of overt cardiovascular disease, which cannot be fully accounted for by co-incident obesity alone and is not associated with cardiovagal baroreflex sensitivity or androgen excess.

INTRODUCTION

Androgen excess polycystic ovary syndrome (AE-PCOS) is the most common endocrine disorder in young women, impacting over 10% of reproductive age women and representing an annual economic burden of $8 billion per year in the United States.^1,2 AE-PCOS is an independent risk factor for cardiovascular disease (CVD) but also often coincides with other metabolic and CVD risk factors such as obesity, and hypertension. In this regard, recent meta-analyses demonstrate a ~1.7-fold increased risk of hypertension in AE-PCOS.^3,4 Notably, women with AE-PCOS already start to exhibit higher resting blood pressure (BP) in their second decade of life compared to women without AE-PCOS, indicating early BP dysregulation.^5,6 Although the heightened incidence of hypertension/CVD in AE-PCOS is well-documented, the physiological mechanism(s) underlying this risk remains incompletely understood.

In addition to the measurement of resting mean BP, there is a growing interest in the assessment of resting beat-to-beat BP variability (BPV) as a novel marker of enhanced cardiovascular risk. This is because prior studies suggest that augmented beat-to-beat BPV provides prognostic information regarding risk of cardiovascular events and target organ damage that is independent of mean BP, per se.^7–9 To our knowledge, only two previous studies by the same group have reported beat-to-beat BPV in women with AE-PCOS. In these investigations, systolic BPV was analyzed using spectral analysis but only one study included control participants, suggesting no difference in beat-to-beat BPV in AE-PCOS compared to controls.^10,11 However, these studies did not account for the impact of obesity independent of PCOS and did not report androgen levels. These are key questions because obesity and androgen exposure have important discrete effects on BP. Accumulating evidence suggests that androgens, particularly through their influence on the autonomic nervous system, play an important role in mediating PCOS-induced cardiovascular dysfunction in AE-PCOS.¹² In addition, free testosterone index (an index of the biological activity of testosterone) is associated with increased target organ damage and 24-hour ambulatory systolic and diastolic BPV in postmenopausal women with hypertension but without PCOS.¹³ Taken together, these data suggest that androgen excess may be a mechanism of augmented beat-to-beat BPV in women with PCOS.

The regulation of BP on a beat-to-beat basis is governed primarily by the arterial baroreflex, and it is well known that impairments in baroreflex function lead to increases in beat-to-beat BPV.¹⁴ Prior studies suggest that women with AE-PCOS exhibit blunted cardiovagal baroreflex sensitivity (BRS).^11,15 Moreover, a recent study from our group¹⁶ demonstrated that, despite no apparent differences in baseline integrated baroreflex gain, testosterone-suppression improved baroreflex gain in women with AE-PCOS but not in healthy control women indicating that testosterone was blunting baroreflex gain in women with PCOS. However, one of these prior studies observed that cardiovagal BRS was blunted only in women with obesity and AE-PCOS, whereas non-obese AE-PCOS exhibited no difference in cardiovagal BRS compared to controls.¹¹ Therefore, in addition to androgen excess, obesity may also contribute to augmented beat-to-beat BPV in women with AE-PCOS.

Accordingly, in the present investigation we examined resting beat-to-beat BPV in women with AE-PCOS compared to control women. In order to account for plausible effects of body mass on BPV, we combined data from two previous investigations^16,17 performed in our laboratory and separated them into overweight/obese (OW/OB) and lean control groups. We also included BPV data from seven participants from a collaborator’s laboratory (C. Usselman) to increase sample size. We tested the hypotheses that: 1) beat-to-beat BPV would be greater in women with AE-PCOS compared to both OW/OB and lean controls; 2) beat-to-beat BPV would be augmented in OW/OB controls compared to lean controls. Finally, to begin to understand mechanism(s), we quantified circulating androgen concentration and cardiovagal BRS and performed correlational analyses with our BPV measures.

METHODS

Ethical Approval

Participants provided written informed consent after explanation of the study procedures and experimental measures, which conformed to the standard set by the Declaration of Helsinki and were approved by the Human Investigation Committee at the Yale School of Medicine (HIC Protocol No. 0801003437, 2000020950), and the Faculty of Medicine Institutional Review Board at McGill University (IRB Study No. A05-M14–18A). These studies were registered with ClinicalTrials.gov (NCT00757185, NCT04327934).

Participants

A total of 46 young women were recruited, 18 lean controls, 14 OW/OB controls, and 14 women with OW/OB and AE-PCOS through local advertisement(s) in the greater New Haven area, Yale School of Medicine online listings for clinical research studies, or by contacting women who had previously consented to being contacted for future studies through the Yale Center for Clinical Investigation and/or the Yale New Haven Hospital’s MyChart system. Seven participants were recruited through posters distributed in the greater Montreal area and on McGill campus. Participants were excluded if they had a previous history of smoking, diabetes, clinically diagnosed hypertension, renal disease, other endocrine or reproductive disorders, were taking insulin-sensitizing/cardiovascular medications, or were taking hormonal contraception for >4 weeks prior to participating in the study. Control participants reported regular menstrual cycles without known gynecological abnormalities and were tested in the early follicular phase of the menstrual cycle. All women included in the study were screened for PCOS with a clinical exam including a transvaginal ultrasound scan as previously described.¹⁸ For the diagnosis of AE-PCOS, in addition to serum androgen excess, at least one of the two following criteria were required: oligo/anovulation (intermenstrual period ≥45 days or ≤8 total menses per year) and/or morphological appearance of polycystic ovary syndrome (≥12 small follicles in the range of 3–9 mm average diameter in one or both ovaries) in accordance with guidelines.¹⁹ Control participants were not included if the total testosterone was greater than 45 ng/dL in accordance with our prior studies.¹⁶ Data from some of these participants (n=40) have been reported in previously published studies from our group.^16,17,20 However, the current hypotheses and analyses are unique and were not addressed by that prior work.

Experimental Measures

Beat-to-beat BP was measured using a commercially available finger photoplethysmography device (Finometer, Finapres Medical Systems, Amsterdam, Netherlands) and supine absolute brachial BP was measured in triplicate with a manual sphygmomanometer. Respiratory excursions were monitored using a strain-gauge pneumobelt around the abdomen or a respiratory capnography device connected to a nasal canula to ensure consistent respiration throughout the resting period. Beat-to-beat BP data were sampled at 1,000 Hz using commercial acquisition software (LabChart; ADInstruments, Sydney, Australia) and stored offline for later analysis.

For both studies, aliquots of whole blood were transferred into a vacutainer without anticoagulant for assessing serum estradiol, progesterone, and total testosterone. The tubes were immediately centrifuged at 4°C, and the serum extracted and analyzed for reproductive hormones. All centrifuged samples were frozen immediately and stored at −80°C until analysis.

Estradiol and progesterone were measured using enzyme-linked immunosorbent assay (ELISA) methods (Alpco Diagnostics, Salem, NH) as described in detail in the previous studies of Usselman et al.¹⁷ and Stone et al.¹⁶ Total testosterone was measured via ELISA or mass spectrometry (ABSCIEX 5500 QTRAP, SCIEX, Inc. Framingham, MA) coupled with high-performance liquid chromatography (Ultimate 3000 Nanocap, Thermo Fisher Scientific, Waltham, MA) at the University of Mississippi Medical Center, Department of Pharmacology and Toxicology (Jackson, MA).¹⁶ Quantitation and statistical analyses were facilitated with MultiQuant v 2.0 (SCIEX, Framingham MA). Testosterone and heavy-labeled internal standard d3-testosterone were purchased from Cerilliant, Corp. (Roundrock, TX). Due to difficulty (poor clotting) with one blood sample, we had to exclude data for one of our AE-PCOS participants. We estimated free testosterone based on the methods described by Vermeulen et al.²¹

Experimental Protocol

Prior to the morning of the experiment, participants attended the lab for a familiarization visit to orient the participants to all study protocols and most aspects of instrumentation. Experiments were performed in a dimly lit, temperature-controlled room in the morning following an overnight fast. Participants were instructed to avoid alcohol consumption and strenuous activity/exercise for at least 24 hours prior to the study. Hydration on the morning of the study was confirmed by measuring urine specific gravity (≤1.020) from samples collected upon arrival to the lab. Then, participant height and weight were measured. After anthropometric assessment participants were asked to lay in the supine position for ~30 minutes for instrumentation on a bed in the environmental chamber. An antecubital venous catheter was then inserted and a resting blood sample was taken. After which, beat-to-beat data were collected over a quiet resting baseline period of 4–13 minutes (average 6.01 ± 1.48 minutes).

Data Analysis

Absolute mean arterial pressure (MAP) was calculated from the average of the 3 supine brachial blood pressures using the equation one-third systolic BP (SBP) plus two-thirds diastolic BP (DBP). Beat-to-beat MAP was calculated as the average of the arterial pressure waveform from diastole to diastole. Modelflow^® was used to estimate stroke volume (SV) on a beat-to-beat basis from Beatscope (Finapres Medical Systems, Amsterdam, Netherlands) and cardiac output (CO) subsequently calculated as SV multiplied by heart rate (HR). Total peripheral resistance (TPR) was calculated as MAP divided by CO. Pulse interval was calculated as the time from diastole to diastole. For all beat-to-beat variable analyses, three beats of data following Physiocals (average 3.6 ± 2.2 per recording) were removed to account for the smoothing algorithm applied in this software.

Beat-to-Beat BP, CO, and TPR variability

Standard measures of variability were calculated for all variables (i.e., SBP, DBP, MAP, CO, TPR) as previously described by our group.²² Briefly, standard deviation (SD) and range, were derived using SPSS statistical software (IBM; Armonk, NY, USA). Coefficient of variation was calculated as SD divided by mean and then expressed as a percentage (i.e., multiplied by 100). Average real variability (ARV) was calculated according to the following standard formula:

$$ARV=\frac{1}{N-1}{\sum }_{k=1}^{N-1}|X_{K+1}-X_K|$$

In this equation, N represents the absolute number of measurements in the recording (total number of heartbeats), k denotes the chronological order of those measurements, and X is the variable of interest. The ARV parameter was originally developed for quantifying BPV in 24-hour ambulatory recordings,²³ and has become increasingly adopted in studies on beat-to-beat BPV.^8,22,24,25

Cardiovagal Baroreflex Sensitivity

Spontaneous cardiovagal BRS was quantified using the sequence method (CardioSeries v2.7; Sao Paulo, Brazil). For this method, sequences of three or more consecutive cardiac cycles where SBP and pulse interval change in the same direction are identified, and the slope is calculated for each individual sequence and averaged to provide an overall gain. Sequences were also separated into up sequences where pulse interval and SBP are both increasing, and down sequences, where SBP and pulse interval are both decreasing. For this analysis, BRS sequences were identified using a zero-beat delay, R² of 0.85 for inclusion, and thresholds of SBP and pulse interval change(s) of 1 mmHg and 5 msec, respectively.²⁶ Previous studies have demonstrated good agreement between cardiovagal BRS gain calculated using sequence technique with either pulse interval or R-R interval.^27,28

Statistical Analyses

Normality was assessed using the Shapiro-Wilk test, and non-parametric tests were used when appropriate. Between group comparisons of participant characteristics and beat-to-beat variability parameters were made using ordinary one-way analysis of variance (ANOVA) with post-hoc comparisons made using Tukey’s post hoc when appropriate. The relationship between cardiovagal BRS or androgens (i.e., total or free testosterone) and beat-to-beat BPV parameters was assessed by Pearson correlation coefficients. Statistical significance was set a priori at an α<0.05. All data are presented as mean ± SD for consistency.

RESULTS

Participant characteristics are presented in Table 1. There were no differences in age between groups. However, as expected, lean controls had a lower body mass and body mass index (BMI) than OW/OB or AE-PCOS individuals. Absolute SBP and DBP were significantly greater in AE-PCOS compared to lean controls. Estradiol and progesterone were not different between groups whereas total testosterone and free testosterone were greater in AE-PCOS compared to both OW/OB and lean controls.

Table 1.

Participant Characteristics

	Lean CON (n=18)	OW/OB CON (n=14)	AE-PCOS (n=14)	P value
Age (years)	26 ± 7	27 ± 5	27 ± 6	0.728
Height (cm)	168 ± 9	162 ± 6*	162 ± 6*	0.036
Body weight (kg)	60 ± 7	92 ± 14*	99 ± 15*	<0.001
BMI (kg/m2)	21 ± 2	35 ± 6*	38 ± 6*	<0.001
Systolic BP (mmHg)	114 ± 13	120 ± 12	125 ± 9*	0.030
Diastolic BP (mmHg)	72 ± 9	78 ± 9	82 ± 8*	0.007
Estradiol (pg/mL)	63 ± 42	56 ± 35	92 ± 45	0.099
Progesterone (ng/mL)	1.0 ± 2.2	1.0 ± 1.3	1.0 ± 1.7	0.994
Total Testosterone (ng/dL)	9 ± 11	11 ± 11	87 ± 67*#	<0.001
Free Testosterone (ng/dL)	1.5 ± 0.9	4.5 ± 3.6	18.2 ± 15.1*#	<0.001

Data are displayed as mean ± SD. Abbreviations: CON, control; OW/OB, overweight/obese; AE-PCOS, androgen-excess polycystic ovary syndrome; BMI, body mass index; BP, blood pressure.

^*P<0.05 vs lean CON;

^#P<0.05 vs OW/OB CON.

Beat-to-Beat BP, CO, and TPR Variability

Figure 1 depicts SD and ARV measurements for SBP, DBP, and MAP in the three groups. The DBP SD was greater in AE-PCOS compared to both lean and OW/OB controls. However, there were no significant differences in SBP or MAP SDs between groups. The SBP, DBP, and MAP ARV were all significantly greater in AE-PCOS compared to lean controls. In addition, SBP ARV was greater in OW/OB than lean controls. There were no significant differences in the range (data not shown, all P>0.05) or coefficient of variation for SBP (lean controls: 3.7 ± 1.5 %; OW/OB controls: 3.1 ± 0.8 %; AE-PCOS: 3.4 ± 0.6 %, P=0.377), DBP (lean controls: 4.2 ± 2.3 %; OW/OB controls: 3.7 ± 1.4 %; AE-PCOS: 4.8 ± 1.5 %, P=0.278), or MAP (lean controls: 3.7 ± 1.6 %; OW/OB controls: 3.2 ± 1.3 %; AE-PCOS: 3.7 ± 1.1 %, P=0.577) across groups.

Figure 1.

Group means and individual data for traditional (standard deviation, top panels) or novel (average real variability, ARV; bottom panels) parameters of variability for systolic blood pressure (SBP; left), diastolic blood pressure (DBP; middle), and mean arterial pressure (MAP; right) in lean control (CON) women (open bars, n=18) or women with overweight/obesity (OW/OB; light blue bars, n=14) and women with androgen-excess polycystic ovary syndrome (AE-PCOS; dark blue bars, n=14). Comparisons between groups were made using one-way ANOVA or the Kruskal-Wallis test where appropriate, with significance set a priori as P<0.05. Data are presented as mean ± SD.

Figure 2 depicts CO and TPR SD and ARV in the three groups. Women with AE-PCOS and OW/OB controls both exhibited greater CO SD than lean controls (P=0.002). Only women with AE-PCOS exhibited greater CO ARV compared to lean controls (P=0.002). In contrast, lean controls exhibited greater TPR SD (P=0.004) but there were no differences in TPR ARV between groups (P=0.178). Regarding range and coefficient of variation, both women with AE-PCOS and OW/OB controls exhibited greater CO range than lean controls (lean controls: 1.82 ± 0.75 L/min; OW/OB controls: 3.10 ± 0.89 L/min; AE-PCOS: 3.56 ± 2.05 L/min, P=0.001), whereas lean controls exhibited a greater TPR range than OW/OB controls (lean controls: 10.87 ± 5.54 mmHg/L/min; OW/OB controls: 6.83 ± 2.09 mmHg/L/min; AE-PCOS: 8.59 ± 2.70 mmHg/L/min, P=0.022). However, there were no group differences in CO or TPR coefficient of variation(s) across groups (data not shown, all P>0.05).

Figure 2.

Group means and individual data for traditional (standard deviation, top panels) or novel (average real variability, ARV; bottom panels) parameters of variability for cardiac output (CO; left), and total peripheral resistance (TPR; right) in lean control (CON) women (open bars, n=18) or women with overweight/obesity (OW/OB; light blue bars, n=14) and women with androgen-excess polycystic ovary syndrome (AE-PCOS; dark blue bars, n=14). Comparisons between groups were made using one-way ANOVA or the Kruskal-Wallis test where appropriate, with significance set a priori as P<0.05; only significant P-values are displayed. Data is presented as mean ± SD.

Cardiovagal BRS and Relationship to BPV

Figure 3 depicts overall cardiovagal BRS in the three groups and the correlation between overall cardiovagal BRS and DBP SD or DBP ARV. There were no group differences in overall cardiovagal BRS gain (P=0.196, Figure 3), or cardiovagal BRS up gain (lean controls: 26.1 ± 13.6 ms/mmHg; OW/OB controls: 20.0 ± 9.7 ms/mmHg; AE-PCOS: 24.1 ± 15.3 ms/mmHg, P=0.392) and down gain (lean controls: 28.2 ± 10.7 ms/mmHg; OW/OB controls: 22.8 ± 11.1 ms/mmHg; AE-PCOS: 23.1 ± 9.9 ms/mmHg, P=0.173). Cardiovagal BRS was not related to DBP SD or DBP ARV in women with AE-PCOS. However, cardiovagal BRS was positively related to DBP ARV in OW/OB controls and tended (P=0.053) to be positively related in lean controls (Figure 3). There was no relationship between overall, up, or down cardiovagal BRS gain and any other BP variability metric in women with AE-PCOS (data not shown, all P>0.05).

Figure 3.

Group means and individual data for cardiovagal baroreflex sensitivity (BRS; top) overall gain in lean control (CON) women (open bars, n=18) or women with overweight/obesity (OW/OB; light blue bars, n=14) and women with androgen-excess polycystic ovary syndrome (AE-PCOS; dark blue bars, n=14) and the association(s) between cardiovagal BRS and markers of augmented beat-to-beat diastolic blood pressure (DBP) variability, particularly standard deviation (SD; bottom left) and average real variability (ARV; bottom right). Cardiovagal BRS was related to DBP ARV only in CON women with OW/OB. Comparisons between groups were made using one-way ANOVA or the Kruskal-Wallis test where appropriate, with significance set a priori as P<0.05. The linear relationship between cardiovagal BRS and DBP variability parameters was assessed using Pearson’s correlation coefficients.

Association between Androgen Excess and BPV in AE-PCOS

Neither total testosterone (r=−0.144, P=0.640) nor free testosterone (r=0.457, P=0.184) were related to DBP SD in the women with AE-PCOS. Likewise, neither total testosterone (r=0.125, P=0.684) nor free testosterone (r=0.106, P=0.770) were related to DBP ARV in AE-PCOS. There was no relationship between total or free testosterone and any other metric of BPV in women with AE-PCOS (data not shown).

DISCUSSION

The major novel findings of this investigation are five-fold: 1) women with AE-PCOS exhibited greater beat-to-beat diastolic BPV as assessed by traditional measures (i.e., SD) compared to both lean and OW/OB controls; 2) Control women with OW/OB also exhibited greater beat-to-beat BPV than lean control women, particularly, as quantified by ARV, a newer metric of BPV; 3) the augmented BPV in AE-PCOS and OW/OB controls compared to lean controls was accompanied by greater CO variability; 4) there were no differences in cardiovagal BRS across groups, however, BRS was positively related to beat-to-beat ARV in OW/OB controls, whereas this relationship was not present in AE-PCOS; 5) there was no association between circulating androgens and any measure of beat-to-beat BPV in women with AE-PCOS. Altogether, these data indicate greater beat-to-beat BPV in AE-PCOS, that cannot entirely be accounted for by BMI, and that coincided with greater CO variability but was not associated with cardiovagal BRS or circulating androgens.

Our study is the first comprehensive examination of beat-to-beat BPV in young women with AE-PCOS. This is important because short-term BPV is associated with the future development, progression, and overall severity of cardiovascular and renal diseases,²⁹ which are both more common in women with AE-PCOS.^30,31 In addition, beat-to-beat BPV provides prognostic information regarding target organ damage that extends beyond traditional 24-hour ambulatory or home-based BP monitoring.^7–9 Thus, these early alterations in beat-to-beat BP dynamics may represent enhanced cardiovascular risk in women with AE-PCOS that manifests prior to impending CVD or renal dysfunction. Importantly, DBP SD was greater in AE-PCOS compared to both lean and OW/OB control groups, indicating greater BPV in the women with AE-PCOS independent of body mass (i.e., attributable to AE-PCOS not obesity per se). In line with the clinical significance of beat-to-beat DBP SD, previous studies have demonstrated that this metric was associated with indices of kidney damage (urinary albumin-creatinine ratio).³² These findings are also congruent to the increased renal injury observed in rodent models of AE-PCOS.³³ Further, beat-to-beat DBP SD has been shown to provide prognostic information following ischemic stroke which was not captured by other variability outcomes (e.g., SBP SD) or absolute BP values.⁷

We also observed that SBP ARV was greater in both AE-PCOS and OW/OB compared to lean controls, but no difference between the AE-PCOS and OW/OB groups. Notably, beat-to-beat SBP ARV has been reported to be associated with end organ damage (i.e., echocardiography-derived left ventricular mass index).⁸ The concept that obesity itself may play a role in augmenting beat-to-beat BPV is expected because obesity is also an independent risk factor for hypertension/CVD with estimates suggesting that obesity can explain 60–75% of the risk for primary hypertension.^34,35 Within the context of AE-PCOS, there is also precedent that obesity drives 24-hour ambulatory BP abnormalities (e.g., non-dipping pattern).⁶ Taken together with our observation of greater DBP SD in women with AE-PCOS compared to OW/OB controls, these data suggest early BP dysregulation in AE-PCOS that may also be independent of obesity.

In our study, the greater beat-to-beat BPV was accompanied by greater CO variability, which is in line with prior studies of BP regulation in young women. Indeed, seminal studies by Joyner and colleagues^36,37 as well as prior studies from our group³⁸ generally indicate that young women regulate their BP more with CO than TPR under resting conditions, compared to men who regulate mainly using TPR. Briefly, young women display a reduced overall autonomic support of BP (i.e., ganglionic blockade reduces BP less in young women) and the contribution of CO to BP is relatively greater than TPR in young women compared to young men or older women. These effects are thought to be, in part, mediated by peripheral β-adrenergic vasodilation (which may oppose vasoconstrictor pathways), and enhanced cardiovagal support of BP.^36,37,39,40 In addition, in young women CO contributes more than TPR to carotid-baroreflex mediated BP responses such that during transient simulated hypertension, CO accounted for the entirety of the baroreflex-mediated depressor response in young women.³⁸ Nevertheless, we did not observe an association between cardiovagal BRS and beat-to-beat BPV in our women with AE-PCOS. However, we observed that cardiovagal BRS was positively associated with BPV (specifically, ARV) in the OW/OB controls with a strong tendency in lean controls. This latter finding would seem to imply that spontaneous cardiovagal activity drives some degree of natural beat-to-beat ARV in healthy young women which is lost or weakened in AE-PCOS.

We also did not observe a relationship between circulating androgens and beat-to-beat BPV in AE-PCOS. This is notable given our recent study demonstrating that short-term testosterone suppression improved baroreflex function in AE-PCOS but not control participants.¹⁶ However, it is important to point out that the baroreflex gain in our previous study was derived from a forearm vascular resistance response to low-level lower body negative pressure, a protocol which primarily captures the peripheral (i.e., sympathetic) vascular component of the baroreflex rather than the cardiovagal component. Furthermore, in that study, differences in integrated baroreflex gain between AE-PCOS and controls were most evident following short-term gonadotropin releasing hormone suppression, whereas baseline baroreflex function did not appear different. Other studies have also observed that circulating androgen concentrations are related to elevated resting sympathetic outflow in lean women with AE-PCOS,^41,42 but not in women with OW/OB and AE-PCOS.^43,44 Taken together, these data suggest that androgens play a greater role in modulating the peripheral sympathetic component than the cardiovagal component, and that this relationship could be modified by obesity in AE-PCOS. Nevertheless, given the relatively small spread of testosterone data in our cohort, additional studies are warranted to better understand the role of androgens in beat-to-beat BPV, and the extent to which obesity modifies the contribution of androgens to BP regulation in women with AE-PCOS.

In summary, this is the first study to comprehensively study beat-to-beat BPV in AE-PCOS, and we observed greater beat-to-beat BPV in these women which could be partially, but not entirely, explained by the presence of obesity. Our findings highlight an early BP phenotype which may predispose women with AE-PCOS to future cardiovascular and/or renal disease.