Sex differences in integrated neurocardiovascular control of blood pressure following acute intermittent hypercapnic hypoxia

Abstract

Repetitive hypoxic apneas, similar to those observed in sleep apnea, result in resetting of the sympathetic baroreflex to higher blood pressures (BP). This baroreflex resetting is associated with hypertension in preclinical models of sleep apnea (intermittent hypoxia, IH); however, the majority of understanding comes from males. There are data to suggest that female rats exposed to IH do not develop high BP. Clinical data further support sex differences in the development of hypertension in sleep apnea, but mechanistic data are lacking. Here we examined sex-related differences in the effect of IH on sympathetic control of BP in humans. We hypothesized that after acute IH we would observe a rise in muscle sympathetic nerve activity (MSNA) and arterial BP in young men (n = 30) that would be absent in young women (n = 19). BP and MSNA were measured during normoxic rest before and after 30 min of IH. Baroreflex sensitivity (modified Oxford) was evaluated before and after IH. A rise in mean BP following IH was observed in men (+2.0 ± 0.7 mmHg, P = 0.03), whereas no change was observed in women (−2.7 ± 1.2 mmHg, P = 0.11). The elevation in MSNA following IH was not different between groups (4.7 ± 1.1 vs. 3.8 ± 1.2 bursts/min, P = 0.65). Sympathetic baroreflex sensitivity did not change after IH in either group (P > 0.05). Our results support sex-related differences in the effect of IH on neurovascular control of BP and show that any BP-raising effects of IH are absent in young women. These data enhance our understanding of sex-specific mechanisms that may contribute to BP changes in sleep apnea.

INTRODUCTION

Over 35% of the United States population will develop sleep apnea during their lifetime (87). Individuals with sleep apnea are at increased risk of developing hypertension and cardiovascular disease (57). A hallmark of sleep apnea is persistent activation of the sympathetic nervous system or “sympathoexcitation” (72). Sympathoexcitation occurs in prevalent conditions like obesity, hypertension, and heart failure and carries an increased mortality risk (1, 3, 37). In the context of sleep apnea, repeated exposure to low oxygen (intermittent hypoxia, IH) and subsequent arterial oxygen desaturation is the primary physiological stimulus by which increases in muscle sympathetic nerve activity (MSNA) and blood pressure occur (40, 60, 71, 72). Along these lines, acute IH (an experimental model that mimics these arterial oxygen desaturations) has been shown to reset the sympathetic baroreflex upward to higher levels of MSNA and rightward to higher levels of blood pressure (50, 78). This resetting of the baroreflex is associated with the development of hypertension in preclinical models of sleep apnea (chronic intermittent hypoxia, CIH) (86).

Although controversy exists (47, 49, 52, 57, 65, 88), there have been reports of a lower incidence of hypertension in women with sleep apnea compared with men (7, 22, 48). Similarly, some (23), but not all (73, 74), data show that female rats exposed to CIH do not develop high blood pressure. Together these findings suggest that mechanisms contributing to the development of high blood pressure following IH exposure differ between men and women; however, sex differences in the neurovascular response following IH in humans have not been assessed. To begin to address this gap in knowledge, we examined MSNA, blood pressure, and the baroreflex before and after exposure to acute IH in healthy young men and women. We hypothesized that we would observe an increase in MSNA and blood pressure in young men after acute IH that would be absent in young women.

METHODS

Participants.

All experiments and procedures were approved by the Institutional Review Boards at the Mayo Clinic (16-004563) and the University of Missouri (2007973), were in accordance with institutional guidelines, and conformed to the Declaration of Helsinki. Experiments were conducted prospectively by each of the two centers, and data were merged to assess the hypothesis presented here. Data unrelated to present hypotheses were published previously from a subset of participants (53).

All participants (n = 49) were young (<45 yr of age), healthy, nonobese [body mass index (BMI) <30 kg/m²], and nonsmokers. Individuals were excluded if they carried a chronic disease diagnosis or were taking medications known to affect autonomic, respiratory, or cardiovascular control. Women were premenopausal and were studied in the early follicular phase of the menstrual cycle (days 1–7) or the placebo phase of oral contraceptive use (self-report, n = 9). We observed no discernible differences in main outcome variables between naturally cycling women and those taking oral hormonal contraceptives, and thus data were pooled; however, we report individual data points within the figures for transparency. A urine pregnancy test was performed on the morning of the study, and a negative test was needed to proceed. Participants were asked to refrain from alcohol, caffeine, nonsteroidal anti-inflammatory drugs (NSAIDS), and strenuous exercise for 24 h before study participation, following recently published guidelines (21, 42).

Written informed consent was obtained from all participants on a screen visit. After the screen visit, participants were instructed to sleep for a standard 8–9 h with oxygen saturation and heart rate continuously monitored and recorded (Wrist Ox, model 3150; Nonin Medical, Plymouth, MN). The device was returned to the laboratory and provided data for an automated scoring algorithm that calculated an oxygen desaturation event index [ODI; 4% drop in arterial O₂ saturation by pulse oximetry ($\text{S}{\text{p}}_{{\text{O}}_2}$) for a minimum duration of 10 s] adjusted for artifacts. Participants with ODI > 10 events/h were excluded to avoid potential effects of undiagnosed sleep apnea.

Instrumentation.

On the study day, participants were admitted to the laboratory in the morning after an overnight fast. Participants rested supine during instrumentation, which included electrocardiography (lead II) and finger pulse oximetry. An intravenous catheter was placed in a forearm vein for systemic intravenous infusions. In n = 24 [18 male (M), 6 female (F)], a 20-gauge, 5-cm catheter was placed in the brachial artery under aseptic conditions after local anesthesia for beat-by-beat arterial blood pressure measurement. In n = 25 (12 M, 13 F), arterial blood pressure was assessed noninvasively by finger photoplethysmography calibrated to automated upper arm sphygmomanometry measurements taken several times throughout the duration of the protocol. Stroke volume was estimated from the blood pressure waveform with the Modelflow method (LabChart; ADinstruments, Sydney, Australia), which incorporates age and sex. Cardiac output was calculated as the product of stroke volume and heart rate. Total peripheral resistance was calculated as cardiac output divided by mean arterial blood pressure. Given that blood pressure was monitored by two different methods, it is important to note that 1) beat-to-beat changes in noninvasive measurements closely follow changes in intra-arterial pressure (54, 58) and 2) invasive and noninvasive baroreflex sensitivity measurements are highly correlated [r = 0.91–0.98 (25, 54, 59)] and have been shown to provide equivalent information (59). Furthermore, the Modelflow method was developed for application of both noninvasive pressure and intra-arterial pressure (4, 5, 26, 82). However, because pulse pressure tracings from the finger may yield an overestimate of beat-to-beat cardiac output compared with intra-arterial pulse pressure tracings (2), we also report changes relative to baseline. This analysis approach has been shown to be accurate, reliable, and in good agreement with Doppler echocardiography (2, 34).

Muscle sympathetic nerve activity.

Measurement of MSNA was attempted in all study participants with the technique of microneurography, and stable MSNA recordings were available from a subset (n = 31; 21 M, 10 F). Multiunit MSNA was recorded with a tungsten microelectrode. The microelectrode was placed percutaneously into the peroneal nerve, posterior to the fibular head, under direct two-dimensional (2D) ultrasound guidance (10). A muscle sympathetic fascicle was identified when taps on the muscle belly or passive muscle stretch evoked mechanoreceptive impulses and no afferent neural response was evoked by skin stimuli. A reference electrode was positioned subcutaneously ∼4 cm from the recording electrode. The recorded signal was amplified, band-pass filtered (700–2,000 Hz), rectified, and integrated (time constant 0.1 s). Sympathetic neurograms were analyzed with a semiautomated program (Ensemble-C; Elucimed LTD). MSNA was expressed as burst frequency (bursts/min), burst incidence (bursts/100 heartbeats), and burst area per minute.

Acute intermittent hypoxia.

Acute IH was achieved for 30 min with methods published previously (53) (see Fig. 1). Briefly, participants were instrumented with a mask connected to a nonrebreathing valve. Breath-by-breath tidal volume, respiratory rate, and inspired/expired gasses were monitored continuously. Individuals alternated between a hypercapnic hypoxic (3% carbon dioxide, 5% oxygen) gas cylinder and room air (21% oxygen) to target 15 hypoxic events over 30 min while preventing hypocapnia typical of hypoxia-induced hyperventilation [men end-tidal CO₂ (ETCO₂): baseline 43.7 ± 0.7 mmHg vs. IH average 43.7 ± 0.7 mmHg; women ETCO₂: baseline 41.5 ± 0.6 mmHg vs. IH average 43.1 ± 0.8 mmHg]. A 50-L meteorological balloon served as a volume reservoir. This technique results in arterial oxygen desaturation events similar to those observed in sleep apnea (5–10% desaturations; 15–30 events/h; see Fig. 2), and duration was chosen based on prior work showing that a consistent rise in MSNA is achieved within this time frame (28, 29, 41, 75, 85). A hypoxic ventilatory response (HVR) test was conducted immediately before and after acute IH. Tests took ∼15 min to complete, at which time hypoxia was achieved with 2–6 breaths of 5% oxygen, 3% carbon dioxide followed by 2 min of room air (43). This was repeated up to five times per test. Chemosensitivity was assessed as the slope of the regression line for minute ventilation vs. oxygen saturation ($\text{S}{\text{p}}_{{\text{O}}_2}$) (hypoxic ventilatory response).

Fig. 1.

Study timeline. Acute intermittent hypoxia (IH) was achieved for 30 min. A hypoxic ventilatory response (HVR) test was conducted immediately before and after acute IH. HVR tests took ∼15 min to complete. Before and after the HVR test and once oxygen saturation returned to baseline, individuals completed a 5-min quiet resting period. After the 5 min, an intravenous bolus of sodium nitroprusside was administered, followed by an intravenous bolus of phenylephrine [modified Oxford technique (ModOx)].

Fig. 2.

Acute intermittent hypoxia (IH): representative data during the IH protocol from a female participant. Individuals alternated between a hypercapnic-hypoxic (3% carbon dioxide, 5% oxygen) gas cylinder and room air (21% oxygen) to target 15 hypoxic events over 30 min while preventing hypocapnia typical of hypoxia-induced hyperventilation. This technique results in arterial oxygen desaturation events similar to those observed in sleep apnea (5–10% desaturations; 15–30 events/h).

Modified Oxford technique.

Before and after the hypoxic ventilatory response test (described above; Fig. 1) and once oxygen saturation ($\text{S}{\text{p}}_{{\text{O}}_2}$) returned to baseline, individuals completed a 5-min quiet resting period. The timing of measurements was consistent with prior studies examining the sympathetic and cardiovascular effects following acute IH (28, 29, 41, 51, 75). After the 5 min, an intravenous bolus of sodium nitroprusside (100 µg) was administered, followed by an intravenous bolus of phenylephrine (150 µg). This approach is considered the gold standard to examine sensitivity of the arterial baroreflex (70). Data were analyzed during the modified Oxford test for measures of cardiac and sympathetic baroreflex sensitivity (Ensemble-R; Elucimed LTD).

Cardiac baroreflex sensitivity.

Sequences used for systolic blood pressure and R-R interval signals were required to rise or fall monotonically in the same direction for at least 3 consecutive beats (55, 79). Values belonging to the identified sequences were formed into xy-pairs, and a regression curve was fitted (R value ≥ 0.8 acceptance level), with the slope of the curve equaling cardiac baroreflex sensitivity (ms/mmHg). Analysis was performed separately for ascending (up-up) and descending (down-down) sequences, and results were pooled. Those individuals in whom no sequences could be identified were excluded from the analysis.

Sympathetic baroreflex sensitivity.

Diastolic blood pressures were assigned 3-mmHg bins, and for each bin the corresponding MSNA burst incidence (bursts/100 heartbeats) was determined (32). Sympathetic baroreflex sensitivity was quantified by plotting MSNA burst incidence (bursts/100 heartbeats) and total activity [arbitrary units (AU)/100 heartbeats] against mean diastolic pressure for each millimeter of mercury bin. Each data point was weighted according to the number of times the particular value occurred, based on the technique described by Kienbaum et al. (32). The value of the slope determined via linear regression analysis provided the sympathetic baroreflex sensitivity for each participant. The midpoint (T₅₀) for each individual’s curve was calculated, representing the diastolic pressure (mmHg) at which there was a 50% likelihood of a burst occurring. Stable MSNA data were available during the modified Oxford test from a subset of individuals (16 M, 9 F).

Data analysis.

Hemodynamic and respiratory variables were recorded at 1,000 Hz and MSNA signals were recorded at 10,000 Hz with a computer data acquisition system (PowerLab; ADInstruments). All data were stored for off-line analysis. Normoxic resting data were analyzed over a 5-min baseline period before and after acute IH, and change values were calculated (Δ = Post-IH – Pre-IH). Statistical analysis was completed with SigmaPlot 14.0 (Systat Software, Inc). The effect of IH (Pre-IH, Post-IH) on main outcome variables was assessed within groups with a one-way repeated-measures analysis of variance (ANOVA). Normality was assessed with the Shapiro–Wilk test and equal variance by the Brown–Forsythe test. Pairwise comparisons were done with the Holm–Sidak method. One-way ANOVAs were used to assess potential group differences in change (Δ, Post-IH – Pre-IH) variables and pairwise comparisons made with the Holm–Sidak method. Nonparametric tests were conducted with the Friedman repeated-measures analysis of variance on ranks or Kruskal–Wallis one-way analysis of variance on ranks (Δ), as appropriate. To examine potential associations between changes in blood pressure and key regulatory variables following IH (e.g., cardiac output, total peripheral resistance), regression analysis was performed and Pearson correlation coefficients were calculated. Statistical significance was set a priori at P < 0.05. Data are reported as means ± standard error of the mean (SE).

RESULTS

Participant characteristics are reported in Table 1, and representative data are shown in Fig. 3. Acute IH was achieved via repeated reductions in inspired oxygen that resulted in significant falls in arterial oxygen saturation (men $\text{S}{\text{p}}_{{\text{O}}_2}$: baseline 98 ± 1% vs. average nadir 91 ± 1%; women $\text{S}{\text{p}}_{{\text{O}}_2}$: baseline 98 ± 1% vs. average nadir 91 ± 1%; main effect of IH, P < 0.01).

Table 1.

Subject demographics

	Men	Women	P
Count	30	19
Age, yr	30 ± 1	26 ± 1	<0.01
Height, cm	181 ± 1	163 ± 2	<0.01
Weight, kg	84.3 ± 2.4	65.0 ± 2.7	<0.01
Body mass index, kg/m2	25.7 ± 0.5	24.2 ± 0.7	0.07
Systolic blood pressure, mmHg	117 ± 2	111 ± 2	0.02
Diastolic blood pressure, mmHg	72 ± 2	69 ± 1	0.15

Data are reported as means ± SE from n = 49 participants. Blood pressures from upper arm automatic sphygmomanometry were determined on the screening visit. Variables compared between sexes via 1-way ANOVA.

Fig. 3.

Steady-state normoxia before (left) and after (right) acute intermittent hypoxia (IH): representative data from a male (top) and a female (bottom) participant. BP, blood pressure (mmHg); ECG, electrocardiogram; MSNA_Int, integrated muscle sympathetic nerve activity.

Cardiovascular control.

IH resulted in an increase in heart rate in both men (P < 0.01) and women (P = 0.03). The change (Δ) in heart rate following IH did not differ between the sexes (P = 0.47; Table 2). There was no effect of IH on systolic blood pressure in men (Δ = +2.6 ± 1.5 mmHg, P = 0.24) or women (Δ = −1.6 ± 1.9 mmHg, P = 0.82), and the change (Δ) did not differ between the sexes (P = 0.08). There was an increase in diastolic and mean blood pressure in the men (P = 0.03 and P = 0.03, respectively) after IH, whereas a decrease in diastolic (P = 0.02) and no change in mean (P = 0.11) blood pressure were observed in the women (Table 2, Fig. 4, A and B).

Table 2.

Neurocardiovascular control variables before and after acute intermittent hypoxia

	Men			Women			Δ
	Pre	Post	P	Pre	Post	P	Men	Women	P
Mean blood pressure, mmHg	94 ± 1	96 ± 1*	0.03	91 ± 2	88 ± 3	0.11	2.0 ± 0.7	−2.7 ± 1.2†	<0.01
Heart rate, beats/min	57 ± 1	59 ± 2*	0.01	63 ± 2	66 ± 2*	0.03	1.9 ± 0.7	2.8 ± 1.2	0.47
Stroke volume, mL/beat	97 ± 3	96 ± 3	0.72	80 ± 3	83 ± 4	0.11	−0.8 ± 2.6	3.2 ± 2.0	0.32
Cardiac output, L/min	5.5 ± 0.2	5.6 ± 0.2	0.68	4.8 ± 0.2	5.2 ± 0.2*	0.01	0.1 ± 0.2	0.4 ± 0.2	0.13
Total peripheral resistance, mmHg·s·mL−1	1.13 ± 0.05	1.10 ± 0.04	0.72	1.23 ± 0.07	1.10 ± 0.07*	0.04	−0.02 ± 0.04	−0.13 ± 0.04†	0.04
Burst incidence, bursts/100 heartbeats	34 ± 2	41 ± 3*	<0.01	30 ± 4	35 ± 5	0.056	7.2 ± 1.8	5.5 ± 2.2	0.59
Burst area, AU/min	4,372 ± 273	4,996 ± 327*	0.01	3,785 ± 418	4,327 ± 470*	0.02	625 ± 231	542 ± 154	0.29
Cardiac baroreflex sensitivity, ms/mmHg	14.3 ± 1.4	13.0 ± 0.9*	0.04	14.6 ± 1.2	12.1 ± 1.1	0.08	−1.3 ± 1.0	−2.6 ± 1.4	0.47

Results are reported as means ± SE from men (n = 30) and women (n = 19) unless otherwise noted [muscle sympathetic nerve activity (MSNA): men n = 21, women n = 10; cardiac baroreflex sensitivity: men n = 29, women n = 18]. Data were analyzed during 5 min of quiet normoxic rest. The effect of intermittent hypoxia (IH) (Pre-IH, Post-IH) was assessed with a 1-way repeated-measures ANOVA. Multiple comparisons via Holm–Sidak method. Change (Δ = Post-IH – Pre-IH) variables compared between sexes via 1-way ANOVA. AU, arbitrary units.

^*P < 0.05 vs. Pre-IH.

^†P < 0.05 vs. Δ men.

Fig. 4.

Sex differences in sympathetic support of blood pressure after acute intermittent hypoxia (IH). Data are reported as individual data points (symbols) and group means (bars). The effect of IH (Pre-IH, Post-IH) within sex was assessed with a 1-way repeated-measures ANOVA. Multiple comparisons via Holm–Sidak method. Change (Δ = Post-IH – Pre-IH) variables compared via 1-way ANOVA. Data are reported for diastolic blood pressure [n = 30 men (M) and 19 women (W); A and B], burst frequency (n = 21 M and 10 W; C and D), and sympathetic baroreflex sensitivity (n = 16 M and 9 W; E and F). ○, Men; □, naturally cycling women; Δ, women taking oral hormonal contraceptives.

There was no change in total peripheral resistance or cardiac output after IH in men (P = 0.72 and P = 0.68, respectively). Women exhibited a fall in total peripheral resistance (P = 0.04) and an increase in cardiac output (P = 0.01) after acute IH. There was no effect of IH on stroke volume in men or women (P = 0.72 and P = 0.11, respectively), and the change (Δ) did not differ between the sexes (P = 0.32) (see Table 2). Of note, conclusions were maintained when data were assessed relative (%) to baseline. Specifically, the analysis supported no change in total peripheral resistance (P = 0.72) or cardiac output (P = 0.36) after IH in men, whereas women exhibited a 10 ± 3% fall in total peripheral resistance (P < 0.01) and a 10 ± 3% increase in cardiac output (P < 0.01). There was no effect of IH on stroke volume in men (P = 0.72) or women (P = 0.10) when calculated relative to baseline, and the change did not differ between sexes (P = 0.28).

Sympathetic nervous system activity.

MSNA data were available from a subset of individuals (men, n = 21; women, n = 10). Compared with baseline, MSNA burst frequency and burst area were increased after IH in both men and women (P < 0.05; Table 2). The increase in MSNA with IH (Δ) did not differ between men and women (P > 0.05; Fig. 4, C and D). There was no change in sympathetic baroreflex sensitivity after IH in either group when assessed as burst incidence (P > 0.05; Table 2, Fig. 4, E and F) or total activity (men: −7.1 ± 0.7 to −6.2 ± 0.5 AU·100 heartbeats⁻¹·mmHg⁻¹, P = 0.10; women: −4.9 ± 0.6 to −6.0 ± 0.8 AU·100 heartbeats⁻¹·mmHg⁻¹, P = 0.51). There was, however, an increase in the T₅₀ in men (71 ± 1 to 75 ± 2 mmHg, P < 0.01) that was not observed in women (69 ± 4 to 70 ± 4 mmHg, P = 0.61). Men also exhibited a fall in cardiac baroreflex sensitivity after IH (P = 0.04). In contrast, there was no effect of IH on cardiac baroreflex sensitivity in women (P = 0.08; Table 2).

Respiration.

The hypoxic ventilatory response was increased after acute IH in both men (P = 0.03) and women (P = 0.02). The change (Δ) in the hypoxic ventilatory response following IH did not differ between the sexes (P = 0.28; Table 3). After return to normoxic breathing, there was no effect of acute IH on tidal volume in either group (men: P = 0.39, women: P = 0.52). There was an increase in breathing frequency and a trend for greater minute ventilation in the men (P = 0.04 and P = 0.07, respectively) after IH, whereas no change was observed in the women (P = 0.69 and P = 0.18, respectively) (Table 3).

Table 3.

Breathing variables before and after acute intermittent hypoxia

	Men			Women			Δ
	Pre	Post	P	Pre	Post	P	Men	Women	P
Tidal volume, mL/breath	581 ± 46	611 ± 44	0.39	443 ± 55	461 ± 69	0.52	30 ± 35	18 ± 27	0.81
Breathing frequency, breaths/min	11.8 ± 0.5	12.6 ± 0.4*	0.04	13.5 ± 0.7	13.8 ± 0.7	0.69	0.9 ± 0.4	0.2 ± 0.6	0.38
Minute ventilation, L/min	6.5 ± 0.4	7.4 ± 0.4	0.07	6.0 ± 0.5	6.3 ± 0.4	0.18	1.0 ± 0.4	0.4 ± 0.3	0.25
Exhaled carbon dioxide, mmHg	43.7 ± 0.7	41.8 ± 0.7*	<0.01	41.5 ± 0.6	40.5 ± 0.8*	0.02	−1.9 ± 0.5	−1.1 ± 0.4	0.38
Hypoxic ventilatory response, L·min−1·%−1	−0.65 ± 0.06	−0.75 ± 0.05*	0.03	−0.59 ± 0.15	−0.83 ± 0.16*	0.02	−0.10 ± 0.04	−0.24 ± 0.10	0.28

Results are reported as means ± SE from men (n = 30) and women (n = 19) unless otherwise noted (hypoxic ventilatory response: men n = 29, women n = 18). Data were analyzed during 5 min of quiet normoxic rest. The effect of intermittent hypoxia (IH) (Pre-IH, Post-IH) was assessed with a 1-way repeated-measures ANOVA. Multiple comparisons via Holm–Sidak method. Change (Δ = Post-IH – Pre-IH) variables compared between the sexes via 1-way ANOVA.

^*P < 0.05 vs. Pre-IH.

Correlation analysis.

Post hoc analyses examining associations between IH-mediated changes in blood pressure (Δ) and key regulatory variables uncovered significant correlations between the change in diastolic blood pressure and total peripheral resistance in both men (R = 0.680, P < 0.01) and women (R = 0.606, P < 0.01) (Fig. 5, A and B). An association between the change in diastolic blood pressure and cardiac output (R = −0.486, P < 0.01) was present in men but not in women (R = −0.258, P = 0.29) (Fig. 5, C and D). Men also exhibited an association between the change in systolic blood pressure and cardiac output (R = 0.508, P < 0.01), which was not observed in women (R = 0.126, P = 0.61) (Fig. 5, E and F). Associations between systolic blood pressure and stroke volume (men: R = 0.409, P = 0.03; women: R = 0.028, P = 0.91) and heart rate (men: R = 0.443, P = 0.01; women: R = 0.144, P = 0.56) were observed in men but not in women.

Fig. 5.

Relationships between changes in blood pressure and key hemodynamic variables following acute intermittent hypoxia (IH). Regression analysis was performed, and Pearson correlation coefficients were calculated. There were significant correlations between the change in diastolic blood pressure and total peripheral resistance in both men (A) and women (B). An association between the change in diastolic blood pressure and cardiac output was present in men (C) but not in women (D). Men also exhibited an association between the change in systolic blood pressure and cardiac output that was not observed in women (E and F). ●, Men; □, naturally cycling women; Δ, women taking oral hormonal contraceptives.

When associations between IH-mediated changes in blood pressure were compared with relative (%) changes in key regulatory variables, conclusions were maintained. Specifically, we continued to observe correlations between the change in diastolic blood pressure and total peripheral resistance in both men (R = 0.701, P < 0.01) and women (R = 0.640, P < 0.01). An association between the change in diastolic blood pressure and cardiac output was present in men (R = −0.453, P = 0.01) but not women (R = −0.291, P = 0.23). Men continued to exhibit an association between the change in systolic blood pressure and cardiac output (R = 0.557, P < 0.01), which was not observed in women (R = 0.037, P = 0.88). Similar conclusions were also made when examining relative changes in stroke volume compared with systolic blood pressure (men: R = 0.365, P = 0.047; women: R = 0.075, P = 0.76).

DISCUSSION

Our data demonstrate sex differences in sympathetic control of blood pressure in humans after acute IH exposure. Major findings include the following: 1) increases in blood pressure following acute IH observed in young men are not present in young women despite similar increases in MSNA and 2) any effect of IH on persistent increases in MSNA and blood pressure is independent of a change in sympathetic baroreflex sensitivity. Group differences in blood pressure after acute IH may instead be attributed to sex-specific changes in cardiac output and total peripheral resistance.

Repeated, intermittent exposure to hypoxia in humans results in consistent increases in MSNA that persist beyond the duration of the hypoxic exposure (11, 12, 28, 41, 51, 75, 76, 85). A portion of the increase in MSNA following acute IH in humans is chemoreflex mediated (53, 63). Hypoxia-mediated increases in chemoafferent activity result in activation of neurons in the nucleus tractus solitarii and downstream rostroventrolateral medulla to elicit increases in efferent sympathetic activity (17, 60). Consistent with this, we observed an increase in both MSNA (Fig. 4D, Table 2) and the hypoxic ventilatory response (a measure of peripheral chemoreceptor sensitivity; Table 3) after IH that did not differ between men and women. Our results agree with previous work showing that the peak MSNA response to hypoxia is similar in men and women (27, 46), as well as data that show a significant positive correlation between MSNA and the hypoxic ventilatory response after acute IH exposure (45). With this, it is important to acknowledge we did not assess the MSNA response during the hypoxic ventilatory response test. Work from others shows that the individual ventilatory response to peripheral chemoreceptor stimulation does not predict sympathetic sensitivities (30, 61); therefore, we cannot be certain that the rise in MSNA we observe after IH in both men and women is due to similar chemoafferent activity.

Despite similarities in the MSNA response, we present exciting data supporting an increase in blood pressure following acute IH in young men that is not observed in young women (Fig. 4B). These results extend preclinical work showing that female rats exposed to CIH do not develop high blood pressure (23) and are consistent with a lower incidence of hypertension in women with sleep apnea compared with men (7, 22, 48). Importantly, although the magnitude of the change in mean arterial blood pressure following IH is relatively small (men: +2.0 ± 0.7 mmHg; women: −2.7 ± 1.2 mmHg), we observed a large effect size (Hedges’ g: 1.03), supporting statistical significance. Furthermore, even a 2-mmHg difference in diastolic blood pressure can have significant clinical implications (83). For example, for each 2-mmHg reduction in diastolic blood pressure the risk of stroke can be reduced 11.5% (36, 62) and the risk of coronary heart disease reduced by 6% (9). Given sex disparities in cardiovascular disease risk in patients with sleep apnea (7, 22, 48), even relatively small differences in the effect of IH on blood pressure observed in the present study could have large implications clinically.

Monahan and colleagues (50) have shown that repetitive hypoxic apneas do not impair sympathetic baroreflex sensitivity, and our data agree with their finding. They also demonstrated that, rather than a change in sensitivity, the baroreflex stimulus-response curve was reset rightward (to higher levels of blood pressure) and upward (to higher levels of MSNA) in the seven men and seven women studied (50). This observation was similar to baroreflex resetting seen during continuous hypoxia (18, 19, 64) and is consistent with our results in men (Fig. 4, Table 2). Baroreflex resetting in this context may be analogous to observations made during exercise (66). With this, sex-related differences in baroreflex resetting during exercise have not been observed (33); however, the authors found that women were better at protecting against increases in blood pressure during exercise compared with men (33). Unfortunately, they were not able to discern whether differences were due to variations in afferent signaling, autonomic efferent activity, or neural-vascular transduction (33). On the basis of similarities in the increase in hypoxic ventilatory response and MSNA in men and women following acute IH, we speculate that the lack of change in blood pressure after IH in young women is the result of factors that may alter the relationship between MSNA and downstream vasoconstriction.

An increase in MSNA initiates release of neurotransmitters (e.g., norepinephrine) from sympathetic nerve terminals that bind onto α-adrenergic receptors and cause vasoconstriction. With this, the vasoconstrictor effects of MSNA following IH appear blunted in young women relative to those in young men. There are data to suggest that compared with men young women exhibit greater vasodilation (8) and blunted vasoconstriction (56) during acute hypoxia, although controversy exists (46). Blunted vasoconstriction during systemic hypoxia could be due to decreased α-adrenergic sensitivity to norepinephrine. However, previous work in humans has shown sympathetic-mediated vasoconstriction and vascular responsiveness to sympathetic activation (i.e., neurovascular transduction) to be preserved or even exaggerated with exposure to hypoxia (13, 75, 77, 84)—importantly, the results come primarily from men. On the other hand, there are data to suggest that norepinephrine causes less vasoconstriction in women compared with men under normoxic conditions (35) and women have lower neurovascular transduction (defined as the transfer of MSNA into vascular tone) compared with men (6), although this is not a universal finding (24, 69, 80). Unfortunately, we do not have measures of vascular responsiveness to α-adrenergic receptor activation from the present cohort; nor is it clear whether there are sex differences in norepinephrine release after acute IH exposure (39). With this, future studies are warranted.

We observed a significant fall in total peripheral resistance in women following IH that did not occur in the men. Our data in men agree with previous work (15), and we now extend these findings to highlight novel observations in women. In women, we speculate that a decrease in α-adrenergic-mediated vasoconstriction or possibly an increase in β-adrenergic receptor-mediated vasodilation following IH may contribute. Prior work has shown that associations between sympathetic activity and blood pressure are absent in young women when β-adrenergic receptors are active (6, 20). Because anywhere from 10% to 50% of hypoxic vasodilation is attributed to the β-adrenergic receptors (81), we speculate that attenuated vasoconstriction following IH-mediated increases in MSNA in women occurs via catecholamines preferentially binding to β-adrenergic receptors in the skeletal muscle vasculature. Estrogen is an established regulator of β-adrenergic vascular receptors and increases β-adrenergic receptor expression (68). This notion strengthens the likelihood that hormone levels (e.g., estradiol) may play a role in the present findings. It is important to note that not all clinical data support sex differences in the incidence of hypertension in sleep apnea (47, 49, 52, 57, 65, 88). Much of this controversy may be due to the inclusion of both pre- and postmenopausal female patients. Along these lines, preclinical data demonstrate that the protective effects of female sex in the setting of IH are due, at least in part, to circulating sex hormones (e.g., estradiol) (38, 67). Given that estrogen receptor expression may differ before to after menopause (16), it is reasonable to speculate that any sex differences we observed in the effect of IH on blood pressure may be altered after menopause.

In addition to sex-specific differences in sympathetic neurovascular transduction after acute IH, divergent blood pressure responses between men and women may be the result of differences in the cardiac response to IH. Along these lines, there is a tendency for the systolic blood pressure response to IH to be greater in men compared with women, and post hoc analysis uncovered a significant correlation between cardiac output and systolic blood pressure in men but not women. We speculate that at least a portion of the overall effect of IH on blood pressure in men may be related to greater cardiac output via a greater inotropic response, an idea with support in the literature (44). Changes in cardiac control of blood pressure following acute IH are further supported by a fall in cardiac baroreflex sensitivity in men (Table 2). Similarly, Tremblay and colleagues (78) have previously reported a blunted response to carotid baroreflex loading after IH.

Experimental considerations.

Strengths of the present investigation include a large sample of healthy young men and women and gold standard methodology including direct recordings of sympathetic activity via microneurography and the modified Oxford approach to assess baroreflex sensitivity. However, it is important to acknowledge that the modified Oxford technique does not allow full characterization of the baroreflex stimulus-response curve (18). Although we cannot determine with certainty whether the set point of the stimulus-response curve was altered after acute IH, differences in baseline blood pressure and T₅₀ suggest this is the case.

An additional strength of the investigation was the addition of carbon dioxide during the IH protocol. This experimental control was applied to prevent hypocapnia during IH, which can occur with acute hyperventilation (see Fig. 2). With this, we observed a significant increase in the hypoxic ventilatory response following IH in the setting of relatively maintained carbon dioxide (baseline ETCO₂ before the HVR test: men pre-IH HVR 42.9 ± 0.7 mmHg vs. post-IH HVR 42.7 ± 0.8 mmHg; women: pre-IH HVR 41.2 ± 0.8 mmHg vs. post-IH HVR 41.5 ± 0.9 mmHg). Although we were able to maintain carbon dioxide levels during the IH and HVR protocols, we observed an increase in breathing frequency after return to normoxia. This increase in ventilation after IH led to relative hypocapnia during steady-state measures (Table 3), which has the potential to influence central sympathetic outflow and cardiovascular responses following IH. Importantly, hypocapnia during post-IH measurements was observed in both sexes, and such changes would attenuate chemoreceptor-mediated increases in MSNA, leading us to underestimate (rather than overestimate) the sympathoexcitatory effect of IH on MSNA following the exposure.

Finally, data were conducted prospectively by two centers, and the collected data were merged to assess the hypothesis presented here. Because this was a secondary analysis of prospectively collected data, we therefore present an unbalanced sample size between men and women (i.e., substantially greater number of men than women). The men were also relatively older (30 ± 1 yr) than the women (26 ± 1 yr) studied. Importantly, when data were examined from a subset of participants ≤30 yr of age (men, n = 19; women, n = 17), conclusions were maintained. Specifically, a rise in mean blood pressure following IH was greater in men compared with women (+2.1 ± 1.1 vs. −3.1 ± 1.3 mmHg, P < 0.01). Furthermore, the elevation in MSNA following IH was not different between groups (5.2 ± 1.2 vs. 3.7 ± 1.5 bursts/min, P = 0.48). Given the robust sample size and reproducibility of findings in a cohort of individuals specifically matched for age, these are minor limitations.

Perspectives and Significance

These data are some of the first to highlight sex differences in the blood pressure response following acute IH in humans; our data clearly show that men and women respond differently and highlight the need for consideration of sex as a biological variable in future research in this area. The presence of sex differences in the neuro-cardiovascular response to acute IH is clinically important given that females are relatively underrepresented in preclinical models of IH, as well as clinical and epidemiological studies in humans (87). As increased daytime blood pressure can be used as a clinical rationale to evaluate patients for sleep apnea (14) and because our data demonstrate that young women do not exhibit an increase in blood pressure after acute IH, this could contribute to underdiagnoses. This notion is supported by data showing that women are typically underdiagnosed for sleep apnea compared with men and as a result are more likely to have poor cardiovascular outcomes (31). Our data thus highlight the need to continue to develop a better understanding of the pathogenesis of conditions related to hypoxemia, such as sleep apnea, and how it may differ by sex.

Summary.

The present study fills a key gap in knowledge regarding sex differences in the effect of IH on sympathetic nervous system activity and resulting changes in blood pressure in humans. Despite similarities in the MSNA response, we present exciting data supporting an increase in blood pressure following acute IH in young men that is not observed in young women. Together these preliminary findings are suggestive of interesting sex-related differences in the neurovascular response to acute IH that merit further investigation in follow-up studies including those in older adults and patients with sleep apnea.

GRANTS

This work was supported by National Institutes of Health Grant HL-130339 (to J. K. Limberg), the Mayo Clinic Center for Biomedical Discovery (J. K. Limberg), and the National Institutes of Health Grant-U54 AG044170 (to S. E. Baker).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.M.M. and J.K.L. conceived and designed research; D.W.J., E.P.O., S.E.B., Z.M.S., C.L.I., J.L.H., C.M.M., and J.K.L. performed experiments; D.W.J., E.P.O., Z.M.S., C.L.I., J.L.H., and J.K.L. analyzed data; D.W.J., E.P.O., S.E.B., Z.M.S., C.L.I., J.L.H., C.M.M., and J.K.L. interpreted results of experiments; D.W.J., C.L.I., and J.K.L. prepared figures; D.W.J. and J.K.L. drafted manuscript; D.W.J., E.P.O., S.E.B., Z.M.S., C.L.I., and J.K.L. edited and revised manuscript; D.W.J., E.P.O., S.E.B., Z.M.S., C.L.I., J.L.H., C.M.M., and J.K.L. approved final version of manuscript.