Anticipatory central command on standing decreases cerebral blood velocity causing hypocapnia in hyperpneic postural tachycardia syndrome

Abstract

Fifty percent of patients with postural tachycardia syndrome (POTS) are hypocapnic during orthostasis related to initial orthostatic hypotension (iOH). We determined whether iOH drives hypocapnia in POTS by low BP or decreased cerebral blood velocity (CBv). We studied three groups; healthy volunteers (n = 32, 18 ± 3 yr) were compared with POTS, grouped by presence [POTS-low end-tidal CO₂ (↓ETCO₂), n = 26, 19 ± 2 yr] or absence [POTS-normal upright end-tidal carbon dioxide (nlCO₂), n = 28, 19 ± 3 yr] of standing hypocapnia defined by end-tidal CO₂ (ETCO₂) ≤ 30 mmHg at steady-state, measuring middle cerebral artery CBv, heart rate (HR), and beat-to-beat blood pressure (BP). After 30 min supine, subjects stood for 5 min. Quantities were measured prestanding, at minimum CBv, minimum BP, peak HR, CBv recovery, BP recovery, minimum HR, steady-state, and 5 min. Baroreflex gain was estimated by α index. iOH occurred with similar frequency and minimum BP in POTS-↓ETCO₂ and POTS-nlCO₂. Minimum CBv was reduced significantly (P < 0.05) in POTS-↓ETCO₂ (48 ± 3 cm/s) preceding hypocapnia compared with POTS-nlCO₂ (61 ± 3 cm/s) or Control (60 ± 2 cm/s). The anticipatory increased BP was significantly larger (P < 0.05) in POTS (8 ± 1 mmHg vs. 2 ± 1) and began ∼8 s prestanding. HR increased in all subjects, CBv increased significantly (P < 0.05) in both POTS-nlCO₂ (76 ± 2 to 85 ± 2 cm/s) and Control (75 ± 2 to 80 ± 2 cm/s) consistent with central command. CBv decreased in POTS-↓ETCO₂ (76 ± 3 to 64 ± 3 cm/s) correlating with decreased baroreflex gain. Cerebral conductance [meanCBv/mean arterial blood pressure (MAP)] was reduced in POTS-↓ETCO₂ throughout. Data support the hypothesis that excessively reduced CBv during iOH may intermittently reduce carotid body blood flow, sensitizing that organ and producing postural hyperventilation in POTS-↓ETCO₂. Excessive fall in CBv occurs in part during prestanding central command and is a facet of defective parasympathetic regulation in POTS.

NEW & NOTEWORTHY Dyspnea is frequent in postural tachycardia syndrome (POTS) and is associated with upright hyperpnea and hypocapnia that drives sinus tachycardia. It is initiated by an exaggerated reduction in cerebral conductance and decreased cerebral blood flow (CBF) that precedes the act of standing. This is a form of autonomically mediated “central command.” Cerebral blood flow is further reduced by initial orthostatic hypotension common in POTS. Hypocapnia is maintained during the standing response and might account for persistent postural tachycardia.

INTRODUCTION

The purpose of this study was to determine whether initial orthostatic hypotension (iOH) drives hypocapnia in postural tachycardia syndrome (POTS) by low blood pressure (BP) or decreased cerebral blood flow (CBF). Postural hypocapnia can cause postural tachycardia because hypocapnia produces sinus tachycardia (1–5). Prior observations suggested that reduction of middle cerebral artery (MCA) blood velocity (CBv) preceded hypocapnia (2). Blood flow to the MCA and carotid body originate from the common carotid artery, may change in parallel, and if intermittently reduced, effect intermittent ischemic hypoxia of the carotid body sensitizing and activating the peripheral chemoreflex (3, 4). Hypocapnia ensues, increasing heart rate (HR) and blood pressure (BP) (5). The hypothesis of increased carotid body sensitization received support from our work indicating enhancement of the hypoxic ventilatory response in patients with postural tachycardia syndrome (POTS) (6). Hypocapnia effectively blunted central chemoreflex activity supine and upright (7).

In patients with hypocapnic POTS [end-tidal CO2 (ETCO₂) <30 mmHg], reduced CBv is driven by the initial orthostatic hypotension (iOH) (8) that occurs more frequently compared with healthy volunteers during active standing (9). However, these studies did not measure CBv to better understand the physiological production of hypocapnia.

iOH during standing results from the shift of an estimated 500–700 mL from central stores to predominantly venous structures below the diaphragm (10). A brief increase in BP is driven by the mechanical compression of leg and splanchnic venous reserves, BP then falls due to gravitational central hypovolemia producing reflex increases in HR with a local maximum near the BP minimum (nadir) ∼10 s poststand (9). BP is rapidly restored with an initial overshoot in ∼10 s by sympathetic vasoconstriction and splanchnic venoconstriction (11). In contrast, vagal withdrawal occurs within a heartbeat producing rapid reflexively increased HR often prolonged in POTS (12).

iOH is defined as a transient fall of BP on rapid standing exceeding 40 mmHg systolic or 20 mmHg diastolic, although some transient decrease in BP is almost always observed (9). iOH is distinguished from sustained orthostatic hypotension (OH) in which BP progressively falls without recovery after standing (13). iOH findings are reproducible (14). In previous studies, we have shown dynamic and static autoregulation of CBF are less effective in patients with POTS compared with control subjects during orthostatic challenge (15).

We propose that in subjects with hypocapnic POTS, the high frequency of iOH on standing reduces CBv that sensitizes and activates the carotid body by intermittent ischemic hypoxia. We further hypothesize that an excessive fall in CBv may occur in part from a reduction of CBv during central command, i.e., feedforward signals descending from higher centers affecting cardiovascular and respiratory control, in anticipation of standing and is a consequence of impaired autonomic regulation in POTS (16, 17).

METHODS

We recruited healthy control subjects and patients with POTS, a form of chronic orthostatic intolerance (OI) defined as “difficulty tolerating the upright posture because of symptoms that abate once supine” (18). Symptoms of OI include lightheadedness, visual difficulties, impending loss of consciousness, headache, hypotension, tachycardia or bradycardia, pallor, and diaphoresis (19). A diagnosis of POTS requires >3 mo of OI plus persistent excessive upright tachycardia in the absence of sustained hypotension (20). Excessive upright tachycardia is defined in adults by an increase of HR exceeding 30 beats/min (bpm) or to a HR > 120 beats/min during a 10-min orthostatic test (21). An increase >40 beats/min is required for the diagnosis in children less than 19 yr old (22).

Subjects

We prospectively recruited 54 consecutive patients with POTS, ranging in age from 14 to 23 yr over a 2-yr period at the Center for Hypotension at New York Medical College. Approximately half (26/54) gave a history of upright dyspnea. We also recruited 32 healthy nonfainting control subjects, ranging in age from 14 to 22 yr. Patients with POTS were referred for suspected POTS with OI lasting longer than 6 mo. POTS diagnoses were confirmed by a 70° upright tilt test performed at a prior visit. Subjects with POTS were either medication naïve or weaned off all medications for at least 2 wk before being studied.

All subjects provided a medical history and a physical examination with electrocardiography was performed. Cardiovascular disease, systemic disease, and transient infections were always excluded. Patients on long-term bed rest were excluded.

Control subjects were recruited from age-, BMI-, and sex-matched healthy volunteers in whom a 70° upright tilt for 10 min was normal [Control, 18 ± 3 yr, BMI 21 ± 4, n = 32, 3 males; POTS-low end-tidal carbon dioxide (↓ETCO₂), 19 ± 2 yr, BMI 20 ± 3, n = 26, 2 males; POTS-normal upright end-tidal carbon dioxide (nlCO₂), 19 ± 3 yr, BMI 21 ± 3, n = 28, 2 males; N.S.]. Control subjects reported no illness or orthostatic intolerance and had never fainted. Approximately 50% reported mild dizziness upon rapid standing from the supine position. Exclusion criteria for participation included infectious or systemic illness, competitive athletic training, recent long-term bed rest, hypertension, or pregnancy within the last year. Prior medication, if any, was discontinued for at least 2 wk before participation in this study, except for the use of oral contraceptives. Each female subject was studied during the mid-luteal phase of their menstrual cycle.

All subjects refrained from caffeine, nicotine, or xanthine-containing products for at least 72 h before testing. This study was approved by the Institutional Review Board of New York Medical College. Subjects 18 yr or older signed an informed consent; those younger than 18 yr assented to participate, and their legal guardian signed an informed consent.

Instrumentation.

All subjects were instrumented while lying supine by the research technician. An electrocardiograph measured HR from the interbeat cardiac electrical intervals and beat-to-beat BP were measured using finger photoplethysmography (Finometer; FMS, Amsterdam) placed on the right index finger and corrected for tilt angle, using ModelFlow software (23). Finometer BP readings were automatically calibrated to brachial artery pressure and oscillometric BP measurements. Nasal cannula capnography combined with a pulse oximeter (Smith Medical PM, Waukesha, WI) measured end-tidal CO₂ (ETCO₂), and oxygen saturation. All subjects were instructed to breathe only through their nostrils; when this did not occur, it was immediately evident on the capnograph signal and subjects were verbally corrected. Respiratory inductive plethysmography (Respitrace, NIMS Scientific, Miami Beach) was only intermittently effective because of rapid standing dislodged transducers. Nevertheless, supine ETCO₂ and respiratory intervals, when observed, were perfectly synchronized.

Transcranial Doppler (TCD) (Neurovision; Multigon, Yonkers, NY) measured CBv of the left middle cerebral artery (MCA) using a 2-MHz probe fixed to the subject’s head by a headband. TCD probes are vulnerable to movement from altered MCA insonation focus on standing. Strict successful efforts were made to optimize and stabilize probes at rest and during orthostasis.

Samples of all signals were acquired at 200 samples/s, multiplexed, and A/D converted using USB data acquisition devices DI-706 and DI-720 (DATAQ, Akron Ohio) and analyzed using custom, proprietary, copyrighted software written by one of the authors (J.M.S.).

In previous studies, healthy volunteers had supine ETCO₂ values that ranged between 38 and 43 mmHg supine, which decreased with orthostasis but was always ≥33 mmHg. We therefore defined upright hypocapnia during steady-state standing as ETCO₂ ≤ 30 mmHg (5). Patients with hypocapnic POTS were designated POTS-↓CO₂ and patients with non-hypocapnic POTS were designated POTS-nlCO₂. Patients with POTS-↓CO₂ uniformly exhibited dyspnea on hyperpnea.

Protocol

All subjects fasted for >3 h before testing. Subjects arrived at our center at 9:00 AM and were prepared for study while supine on a tilt table. Following instrumentation, subjects remained awake and supine for 30 min to acclimate. Baseline HR, BP, ETCO₂, and CBv data were collected for 10 min in the supine position to use for HR variability (HRV) and BP variability (BPV) calculations. Subjects stood up over <5 s and remained standing for 5 min while HR, BP, ETCO₂, and CBv were continuously monitored. We explained the standing procedure to all study participants who were told that the research technician would approach the tilt table 5–10 s before standing. The act of approach provoked an anticipatory response comprising an increase in HR, and BP in all subjects, an increase in CBv in control subjects and subjects with POTS-nlCO₂, but a decrease in CBv in subjects with POTS-↓CO₂. Five minutes of standing was chosen because we have not observed vasovagal syncope within this time period, and 5 min sufficed to establish steady-state HR. Some patients with POTS fulfilled criteria for iOH during the standing procedure with brief OI symptoms. After 5 min, subjects returned to supine for recovery during which baseline HR, BP, ETCO₂, and CBv were restored.

Event data analysis.

We compared data for control and both POTS groups at baseline, and at events depicted in Fig. 1 which shows BP, HR, and CBv from a representative healthy individual.

Figure 1.

Representative phasic cerebral blood flow velocity (CBFv, A), blood pressure (BP, B), and time averaged heart rate (HR, C), and for a Control subject who had initial orthostatic hypotension (iOH). Marked events are: Baseline; Stand = Standing initiated; Min CBv = the nadir or lowest CBv achieved after standing; Min BP = the nadir or lowest blood pressure achieved after standing; Peak HR = the maximum HR achieved after standing; CBv Recovery = the first local maximum of CBv after standing; BP Recovery = the first local maximum of BP after standing; Min HR (HR Recovery) = the lowest HR after standing; Steady-State HR = HR increases after HR Recovery until it achieves a steady slope.

In time sequence these events are:

Baseline = data collected supine before the anticipatory phase

Stand = standing initiated at end of anticipatory phase

Min CBv = the nadir or lowest mean CBv after standing

Min BP = the nadir or lowest BP achieved after standing

Peak HR = the local maximum HR achieved after standing

CBv Recovery = the first local maximum of mean CBv after standing

BP Recovery = the first local maximum of BP after standing

Min HR (HR Recovery) = the lowest HR after standing

Steady-State HR = HR increases following recovery, at steady-state slope turns downward computed by the running slope differences/t test technique of Zuo et al. (24).

Averaged HR, BP, CBv, ETCO₂, and time from standing ± standard deviation were tabulated for each event using a moving average window with a 3 s aperture in Table 1 for POTS-↓CO₂, POTS-nlCO₂, and control. The start of standing was referenced as time 0. Anticipatory changes in HR, BP, and CBv were measured during the 10 s preceding standing, then averaged, and presented as time 0 (“Standing”). Graphical data in Fig. 2 compare arterial blood pressure, HR, CBv, and ETCO₂ at events for representative subjects from each group. Graphical data in Fig. 3 compares mean arterial blood pressure (MAP), mean CBv, and mean cerebral conductance at events for each group averaged over subjects. Cerebral conductance was calculated from the ratio of mean CBv to MAP for each subject at every event.

Table 1.

Time of event, systolic blood pressure, mean cerebral blood velocity, and heart rate

	POTS-↓CO2 < 30 Torr (n = 26)					POTS-nlCO2 ≥ 30 Torr (n = 28)					Control (n = 34)
Event	Time	SBP	mCBv	ETCO2	HR	Time	SBP	mCBv	ETCO2	HR	Time	SBP	mCBv	ETCO2	HR
Baseline		116 ± 3	76 ± 3	36 ± 1*#	83 ± 4*		115 ± 4	76 ± 2	41 ± 1	75 ± 3		115 ± 4	77 ± 2	40 ± 2	72 ± 2
Stand	0	124 ± 3*	64 ± 3*#	35 ± 2	94 ± 4*#	0	123 ± 3*	85 ± 2	39 ± 2	80 ± 3	0	116 ± 3	80 ± 2	39 ± 2	82 ± 4
CBFv nadir	9 ± 1	83 ± 4	48 ± 3*#	27 ± 2*#	116 ± 3*	10 ± 1	86 ± 5	61 ± 3	37 ± 1	106 ± 5	9 ± 3	84 ± 7	60 ± 2	36 ± 2	105 ± 4
BP nadir	12 ± 1	73 ± 2*	52 ± 3*#	30 ± 2*#	117 ± 3*	12 ± 1	73 ± 3*	64 ± 3	37 ± 1	112 ± 7	10 ± 1	82 ± 3	61 ± 3	37 ± 2	106 ± 4
HR peak	14 ± 2	123 ± 7	70 ± 5	28 ± 1*#	131 ± 3*	19 ± 2	118 ± 6	82 ± 4	36 ± 1	127 ± 3*	15 ± 1	111 ± 5	78 ± 3	35 ± 2	116 ± 4
CBv recovery	18 ± 1	88 ± 5*	72 ± 4*#	27 ± 2*#	127 ± 4*#	19 ± 1	95 ± 5*	88 ± 4	36 ± 1	113 ± 5	16 ± 1	121 ± 7	88 ± 4	35 ± 2	100 ± 5
BP recovery	21 ± 2	150 ± 6*	62 ± 5*	28 ± 2*#	110 ± 3*	27 ± 3	148 ± 6	74 ± 4	37 ± 1	102 ± 4	21 ± 1	139 ± 7	76 ± 4	36 ± 2	91 ± 6
HR recovery	42 ± 8*	140 ± 6*#	63 ± 4*#	28 ± 2*#	105 ± 4*#	39 ± 6*	134 ± 7	71 ± 4	37 ± 1	87 ± 4	25 ± 1	122 ± 6	72 ± 4	37 ± 2	78 ± 3
Steady state	85 ± 12*	135 ± 5*#	55 ± 4*#	26 ± 2*#	115 ± 5*	101 ± 13*	120 ± 4	68 ± 4	37 ± 1	105 ± 4*	55 ± 12	122 ± 4	72 ± 4	37 ± 2	90 ± 3
End stand	300 ± 12	136 ± 6*#	56 ± 4*#	26 ± 2*#	116 ± 5*	300 ± 6	121 ± 3	67 ± 5	35 ± 1	107 ± 4*	300 ± 10	120 ± 3	71 ± 4	37 ± 2	89 ± 4

All values = means ± SD. Comparisons of study groups on baseline spectral analysis of heart rate and blood pressure variability were conducted using independent t tests for all postural tachycardia syndrome (POTS) vs. controls and for each POTS subgroup vs. controls. Similarly, independent t tests were used for comparing each POTS subgroup. For these univariable tests, there was no adjustment for multiple comparisons. P values are unadjusted for multiple comparisons. BP, blood pressure; CBv, cerebral blood velocity; CBFv, cerebral blood flow velocity; ETCO₂, end-tidal CO₂; HR, heart rate; mCBv, mean cerebral blood velocity; POTS-↓CO₂, postural tachycardia syndrome with low carbon dioxide; POTS-nlCO₂, postural tachycardia syndrome with normal upright carbon dioxide; SBP, systolic blood pressure.

^*P < 0.05 compared with controls; #P < 0.05 compared with baseline.

Figure 2.

Representative tracings in blood pressure (BP, A), heart rate (HR, B), cerebral blood velocity (CBv, C), and end-tidal CO₂ (ETCO₂, D) from (top to bottom) postural tachycardia syndrome with low ETCO₂ (POTS-↓ETCO₂), postural tachycardia syndrome with normal upright ETCO₂ (POTS-nlETCO₂), and healthy control subjects. Anticipatory increases in BP and HR were present in all. CBv increased in POTS-nlETCO₂ but decreased in POTS-↓ETCO₂.

Figure 3.

Mean arterial pressure (MAP, A), mean cerebral blood velocity (mCBv, B), and cerebral conductance (C) for postural tachycardia syndrome with low end-tidal carbon dioxide (POTS-↓ETCO₂) (in gray), postural tachycardia syndrome with normal upright end-tidal carbon dioxide (POTS-nlETCO₂) (in red), and Control (in black). Reduced conductance maintains decreased mCBv in POTS-↓ETCO₂. While the middle panel shows no significant differences for a group-by-time interaction, however, main effects for group and time were significant (P < 0.05). The bottom panel shows a significant group-by-time interaction for conductance (P < 0.05). Specifically, the POTS ↓CO₂ had significantly a lower mean conductance value at the blood pressure (BP)_min time point (adjusted P < 0.01).

HR variability, BP variability, and cardiovagal baroreflex gain.

We previously showed reduced HRV and gain at the operating point of the cardiovagal baroreflex (25). We repeated these calculations in POTS-↓CO₂, POTS-nlCO₂, and controls using custom software to collect data for at least 10 min (the “baseline period”) based on the spontaneously fluctuating R-R interval and systolic BP signals. For power spectra analysis, we used an autoregressive model based on the methods by Montano et al. (26) and Pagani et al. (27). Signals recorded as discrete time series were converted to an impulse train with equal intervals (28) and we computed the autoregressive power spectra (29). The Yule–Walker equation, modified by the Levinson-Durbin algorithm and succeeded by Anderson’s test, used Akaike’s final prediction error minimization to find the model order. The resultant interval spectrum was transformed to the spectrum of counts via division by the mean signal sequence (28). The defined frequency band total power (0.05–0.40 Hz) was used to calculate individual spectral powers for the low-frequency (LF, 0.05–0.15 Hz), and high-frequency (HF, 0.15–0.40 Hz) power bands. We used the ratio of LF RR interval power to LF BP power [the α index] to describe changes in the cardiovagal baroreflex sensitivity at the operating point (25). HF power corresponds to the respiratory sinus arrhythmia and is vagal dependent but only partly baroreflex dependent (30). LF BP power corresponds to arterial vasomotion and is sometimes used as an index of sympathetic activity (31). LF RR interval power depends on LF BP power transduced by the cardiovagal baroreflex (32). These results are presented in Table 2.

Table 2.

Baseline spectral analysis of heart rate and blood pressure variability

	All Controls	All POTS	Pots ↓ETCO2	POTS nlETCO2
Total HRV	2,838 ± 470	1,066 ± 119*	835 ± 118*	1,208 ± 125*#
LF HRV	754 ± 143	379 ± 65*	247 ± 58*	479 ± 96*#
HF HRV	1,444 ± 297	389 ± 51*	339 ± 106*	428 ± 62*
Total BPV	10.3 ± 2.0	11.8 ± 1.3	9.4 ± 1.8	13.1 ± 1.7
LF BPV	2.5 ± 0.3	3.0 ± 0.3	2.9 ± 0.4	3.0 ± 0.3
HF BPV	1.5 ± 0.2	1.3 ± 0.1	1.5 ± 0.2	1.3 ± 0.2
Alpha index	20.5 ± 2.1	15.4 ± 2.0	10.9 ± 2.0*	17.3 ± 2.1#

Pearson’s correlation was used to assess the linear associations for the spectral analyses. Comparison of study groups on baseline spectral analysis of heart rate and blood pressure variability was conducted using independent t tests for all postural tachycardia syndrome (POTS) vs. controls and for each POTS subgroup vs. controls. Similarly, independent t tests were used for comparing each POTS subgroup. For these univariable tests, there was no adjustment for multiple comparisons. P values are unadjusted for multiple comparisons. HF BPV, high-frequency blood pressure variability; HF HRV, high-frequency heart rate variability; HRV, heart rate variability; LF BPV, low-frequency blood pressure variability; LF HRV, low-frequency heart rate variability; POTS-↓ETCO₂, postural tachycardia syndrome with low end-tidal carbon dioxide; POTS-nlETCO₂, postural tachycardia syndrome with normal upright end-tidal carbon dioxide.

^*P < 0.05 compared with all controls; #P < 0.05 ↓ETCO₂ compared with POTS nlETCO₂.

Statistical methods.

For Tables 1 and 2, comparisons of study groups on baseline spectral analysis of heart rate and blood pressure variability were conducted using independent t tests for all POTS versus controls and for each POTS subgroup versus controls. Similarly, independent t tests were used for comparing each POTS subgroup. For these univariable tests, there was no adjustment for multiple comparisons. For Fig. 3, comparison of POTS groups and controls over the time-ordered events was analyzed as repeated-measures ANOVA (rmANOVA), with time-ordered events considered as a fixed effect. Three dependent variables were considered: MAP, mean CBv, and conductance at each event. Each dependent variable was assessed in a separate model. For these models, we assumed a covariance structure of compound symmetry. To determine whether the pattern of responses over the periods differed between patients with POTS-↓CO₂, POTS-nlCO₂ and controls, we evaluated a group-by-period interaction term. P values for the interaction term were adjusted using the Greenhouse–Geisser correction. If the interaction term was significant, post hoc comparisons of the study groups used Scheffe’s test to control for multiple comparisons. Significance was established at P < 0.05. To reduce inter-rater variability, all data were collected by the research technician and were analyzed by one investigator who was blinded to subject group membership. Results were calculated using SPSS version 15.

RESULTS

Figure 2 shows representative tracings of (left to right) arterial pressure, HR, CBv, and ETCO₂ from among (top to bottom) POTS-↓ETCO₂ (n = 26), POTS-nlETCO₂ (n = 28), and healthy control (n = 34) subjects. HR also increased in all during anticipatory period. CBv increased in POTS-nlETCO₂ and control during the anticipatory period as expected (33, 34) but decreased in POTS-↓ETCO₂ despite increasing BP. The overall decrease of CBv from Stand to nadir of CBv was consequently largest in the patient with POTS-↓ETCO₂.

Timing of Events Is Prolonged in POTS Compared with Control

Compilation of the data for all subjects is shown in Table 1, and all showed anticipatory increases in BP preceding standing that were smallest in control. Table 1 shows event times, systolic blood pressure (SBP) ± SD, mean CBv ± SD, ETCO₂ ± SD, and HR ± SD as dependent variables using the ordinal event scale in POTS-↓ETCO₂, POTS-nlETCO₂, and healthy control subjects. The time to achieve HR Recovery and Steady-State HR was significantly prolonged for POTS following stand compared with control subjects (P < 0.05).

The BP Nadir Was Lower in All POTS but Afterward Higher in POTS-↓ETCO₂. HR Was Increased for All Events for POTS-↓ETCO₂

Systolic blood pressure at the Stand was higher (P < 0.05), but at BP Nadir and CBv Recovery was lower in all POTS compared with control (P < 0.05), whereas SBP was thereafter only increased for POTS-↓ETCO₂ compared with POTS-nlETCO₂ and control (P < 0.05). HR was increased at all events for POTS-↓ETCO₂ compared with control (P < 0.05). HR was increased above control for POTS-nlETCO₂ at steady state and at the 5-min event (end stand) (P < 0.05).

ETCO₂ Was Decreased in POTS-↓ETCO₂ at Baseline but Decreased CBv Beginning at Stand

Mean CBv was decreased beginning at Stand for POTS-↓ETCO₂ compared with either POTS-nlETCO₂ or control (P < 0.05); reduced CBv was initiated preceding the onset of hypocapnia. ETCO₂ was decreased in POTS-↓ETCO₂ by design but was also significantly reduced at baseline without difference in baseline CBv (P < 0.05).

Mean Arterial Pressure Was Lower than Control in POTS-↓ETCO₂ and POTS-nlETCO₂ at the BP Nadir but Elevated in POTS-nlETCO₂ Thereafter. Mean CBv and Conductance Were Reduced following Stand in POTS-↓ETCO₂

Figure 3 demonstrates the dominant role of reduced cerebral blood velocity conductance in maintaining decreased mean CBv (mCBv) in POTS-↓ETCO₂. Orthostatic hypertension accompanied upright hypocapnia. As shown in Fig. 3A, no significant differences emerged for main effects of the study group and time or for a group-by-time interaction. Figure 3B shows no significant differences for a group-by-time interaction, however, main effects for group and time were significant (P < 0.05). Of interest, the POTS ↓CO₂ group had a lower mean CBv value over all time points compared with the other two study groups (adjusted P < 0.05). Finally, Fig. 3C shows a significant group-by-time interaction for conductance (P < 0.05). Specifically, the POTS ↓CO₂ had significantly a lower mean conductance value at the BP_min time point (adjusted P < 0.01).

Baseline Baroreflex Gain as Measured by the α Index Was Reduced for POTS-↓ETCO₂ Compared with POTS-nlETCO₂ and Control

Measures of HR and BP variability were used to investigate autonomic and baroreflex-related parasympathetic modulation in the frequency domain. Table 2 shows measures of HRV and BPV. The ratio of HR to BP power (the α index of baroreflex gain) and HF-HRV (measures respiratory sinus arrhythmia) reflect different aspects of parasympathetic activity. Total HRV and HF-HRV—are low for all POTS compared with control (P < 0.5) signifying reduced vagal modulation of HR. HF-HRV represents the respiratory sinus arrhythmia driven largely by vagal afferents arising from pulmonary stretch receptors and arterial baroreflex, plus intrinsic rhythmicity of brain respiratory centers (35). The α index, representing baseline cardiovagal baroreflex gain at the operating point of the sigmoidal baroreflex curve, was significantly decreased (P < 0.05) only in POTS-↓ETCO₂. A correlational analysis of α and mean CBv resulted in a significant Pearson correlation coefficient of 0.61 (P < 0.001).

DISCUSSION

This study determined whether iOH drives hypocapnia in POTS by low BP or decreased CBv. We had hypothesized that hypocapnic POTS is driven by the high frequency of iOH in POTS on standing producing lower BP. Our data disprove this hypothesis as there is no direct dependence of hypocapnia on the degree of initial hypotension in POTS. However, reduced minimum CBv during early standing was highly related to hypocapnia in POTS, potentially by sensitizing and activating the carotid body via intermittent ischemic hypoxia. We had also hypothesized that an excessive fall in CBv may occur in part from a reduction of CBv during central command in anticipation of standing that is autonomically driven. Our data are consistent with this hypothesis and demonstrate a significant correlation with impaired parasympathetic regulation.

The major findings of this study are that postural hypocapnia in POTS is preceded by a reduction in cerebral conductance and CBv in the middle cerebral artery and that reduced CBv, and conductance begin before the physical act of standing, likely initiated via central command, i.e., descending signals from higher brain centers capable of influencing cardiovascular and respiratory responses, and are maintained throughout the initial standing response. Since iOH occurs with similar frequency in POTS-↓ETCO₂ and POTS-nlETCO₂, it is unlikely that hypocapnic POTS is driven directly by the Standing reduction in BP. The resulting reduced CBv and cerebral conductance are subsequently sustained by hypocapnia.

Evidence suggests that reduced blood flow in the middle cerebral artery reflects reduced blood flow in the internal carotid artery and also reduced blood flow to the carotid body (36, 37). Reduced blood flow to the carotid body decreases oxygen delivery and results in ischemic or stagnant hypoxia, often seen in heart failure (38). Ischemic hypoxia would occur intermittently upon standing in POTS-↓ETCO₂. Intermittent hypoxia is a potent sensitizer of the carotid body chemoreflex (4, 39) and we therefore speculate that this results in hyperventilation and hypocapnia. Hypocapnia also interferes with the central CO₂-sensitive chemoreflex (7, 40). Our determination that baseline ETCO₂ is reduced in POTS-↓ETCO₂ despite no difference in baseline CBv or conductance supports the hypothesis of carotid body sensitization.

Thus, based on our data, we hypothesize that an aberrant response to central command results in the development of hypocapnic POTS. The response of control subjects presumably occurs via a “normal” anticipatory central command response to standing (as a form of exercise), which includes an increase in BP, HR, and CBv (41). A similar response is seen in POTS-nlETCO₂. Changes of CBv when BP is increased at the Standing event represents a complex interaction among opposing effects: direct effects of central command should increase CBv as occurs during the initiation of exercise (42), myogenic autoregulation of CBF should constrict cerebral vasculature in response to increased BP to maintain CBF relatively constant (43), and arterial baroreflexes should enhance parasympathetic activity and withdraw sympathetic activity at the level of the cerebral and pial arteries (44). Sympathetic fibers regulate cerebral blood flow only under conditions of adrenergic activation/deactivation (45) and otherwise have controversial effects (46). Muscarinic parasympathetic fibers have also been shown to be relatively ineffectual in producing cerebral artery vasodilation in humans (47). But parasympathetic innervation of cerebral arteries is particularly dense and relies upon cotransmitters in addition to or in preference to acetylcholine (48). Such “noncholinergic parasympathetic nerves” produce diverse vasoactive substances including nitrergic nitric oxide (NO), vasoactive intestinal polypeptide and pituitary adenylate cyclase pathway-activating peptide, with NO functioning as a common vasodilator pathway (49, 50). Thus, increased BP via central command should activate nitrergic nerves producing cerebral vasodilation and enhanced cerebral blood flow—as observed in Control and in POTS-nlCO₂, but not in POTS-↓CO₂.

Past work suggested that a subset of POTS has a defect in nitrergic parasympathetic activity (51) and blunted cerebral response to NO (52). However, those patients were not sorted by eucapneic or hypocapnic orthostatic responses. In the present study, we have shown that reduced cerebral conductance and postural hypocapnia correlates with decreased α index indicative of parasympathetic baroreflex deficiency in POTS (25). Thus, patients with POTS-↓ETCO₂ fail to dilate cerebral vessels, cerebral conductance is reduced, and cerebral blood flow falls.

Therefore, postural hypocapnia does not directly depend on the magnitude of iOH, but rather on the interactions of central command with defective arterial baroreflex-mediated cerebral vasodilation. Anticipatory effects of initial standing on reducing cerebral blood flow are key. We surmise that upon standing in POTS-nlCO₂ and control subjects, cerebral blood flow is insufficiently reduced to sensitize carotid body chemoreflexes because nitrergic parasympathetic cerebral innervation is intact.

Limitations

Subjects for this study were recruited as referred to our clinic for evaluation of suspected POTS. There may be recruitment bias that is evident in the age range studied. Therefore, we have no specific insight into how this subgroup of patients with POTS fits into the larger population of patients affected by POTS. Our data are from subjects of a restricted age range and thus conclusions of this study cannot be generalized to different ages.

Transcranial Doppler measures blood velocity through specific large cerebral arteries. The middle cerebral artery was used because it derives its blood flow almost entirely from the carotid artery as does the carotid body. Evidence supports the hypothesis that MCA flow reflects carotid body flow, and that reduced MCA may imply stagnant ischemia of the carotid body. Direct noninvasive measurement of carotid body blood flow is not feasible. We also did not measure transcranial Doppler in both hemispheres because previous work showed that middle cerebral artery CBv was not different between the hemispheres during orthostatic stress (53). Transcranial Doppler only measures cerebral blood velocity rather than flow. However, changes in CBv are thought to accurately reflect changes in cerebral blood flow during orthostasis (54).

Carotid body sensitization was not directly assessed via measurement of the hypoxic ventilatory response. This is a limitation to the explanation of results. At this time, we lacked appropriate instrumentation to perform such testing. Going forward, hypoxic and hyperoxic responses will be tested as well as hypercapnic + hyperoxic responses to better test carotid body sensitization hypothesis and the effects on central chemoreflexes.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

Funding for this project was provided by National Heart Lung and Blood Institute (NHLBI) under Grants RO1 HL134674 and 1R56 HL162752.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

J.M.S. and M.S.M. conceived and designed research; performed experiments; analyzed data; interpreted results of experiments; prepared figures; drafted manuscript; edited and revised manuscript; approved final version of manuscript.