Peripheral autonomic dysregulation in response to standing: a new form associated with cardiovascular disease, mobility reduction, and cognitive alterations

Abstract

Background and Aims

Cardiovascular responses to standing are essential for maintaining cerebrovascular blood flow. While orthostatic blood pressure (BP) responses have been studied extensively, the role of cardiac output (CO) remains unclear. This study aimed to investigate patterns of orthostatic CO responses, their physiological determinants, and clinical relevance.

Methods

Real-time haemodynamic responses to active standing were recorded in two prospective cohorts: 3074 young (Melbourne) and 3025 older (Ireland) adults. In the young, five subgroups assessed phenotype reproducibility (same-day and two-week retests), responses to vasodilation (sublingual glyceryl trinitrate, GTN) or sympathetic activation (SA; handgrip with phenylephrine), and associations with urinary electrolyte excretion. In older adults, CO patterns were related to clinical characteristics and outcomes.

Results

A reproducible dichotomy in the initial orthostatic CO response was identified: either a transient rise (CO_R: 66% young, 78% older) or fall (CO_F: 34% young, 22% older). CO_F subjects had less heart rate increase and greater transient reductions in stroke volume and BP upon standing, despite an increased sympathetic response. GTN exaggerated CO_F, whereas SA mitigated it. In older adults, CO_F was more prevalent with cardiovascular disease, higher body mass index, smoking, diabetes, or antihypertensive use. After adjustment, CO_F remained associated with slower gait, frailty, and diminished cognitive function (all P < .001). Among subjects on negatively chronotropic drugs (n = 487), falls were more frequent in CO_F (odds ratio 1.82, 95% confidence interval 1.18–2.80).

Conclusions

These findings reveal a novel variant of the autonomic cardiovascular response to standing, with important clinical implications for functional decline in older adults.

Structured Graphical Abstract

Graphical Abstract.

Physiological characteristics and clinical associations of a new form of peripheral autonomic dysregulation.

See the editorial comment for this article ‘Falling for cardiac output: an enticing new orthostatic phenotype’, by A.H. Petriceks et al., https://doi.org/10.1093/eurheartj/ehag065.

Introduction

The complex cardiovascular responses to active standing provide insights into fundamental homeostatic autonomic control mechanisms.¹ Upon standing, these responses² counteract the gravitational redistribution of blood below the diaphragm and stabilize blood pressure (BP). The normal pattern is a brief fall in BP that reaches a nadir 7–10 s after active standing. Failure to maintain BP on continued standing (orthostatic hypotension, OH) is associated with an increased risk of falls that are the predominant cause of injury among older individuals.³ Distinct from OH, another early and transient abnormality in BP regulation is initial orthostatic hypotension (IOH), defined as a ≥40 mmHg drop in systolic BP (SBP) and/or ≥20 mmHg drop in diastolic BP (DBP) within the first 15 s of active standing. Interestingly, previous analyses showed that these large initial BP drops do not necessarily correlate with a heightened risk of falls.³ Therefore, it has been proposed that other postural cardiovascular changes might be important.³

The initial fall in BP on standing is typically accompanied by transient increases in heart rate (HR) and falls in stroke volume (SV) and, perhaps unexpectedly, systemic vascular resistance (SVR).¹ Active muscle contraction is believed to be a key element in these responses, as they are not observed with tilt-table testing.¹ Muscle activity engages central and cardiopulmonary autonomic reflexes, along with local metabolic mechanisms, contributing to the rise in HR and drop in SVR.¹ The reduction in SV is primarily explained by the gravitational translocation of approximately 500–1000 mL of blood into the abdominal capacitance veins, which normally hold around 20% of the total blood volume.^4,5 The fall in BP per se activates the arterial baroreflexes, triggering sympathetic outflow that restores SVR via arteriolar constriction and augments venous return by constricting the capacitance veins, thereby supporting SV recovery.^6–9

The changes in cardiac output (CO) upon standing merit attention as they are critical in determining cerebral blood flow.^10,11 Somewhat unexpectedly, CO is usually reported to increase by about 24% within 7 s of active standing. This is the result of an early rise in HR that counterbalances the fall in venous return and SV.¹ However, our serendipitous observations suggested that an initial significant decrease in CO upon standing is not uncommon. Such variation in CO responses to standing may indicate fundamental differences in autonomic and neural cardiovascular regulation, which could hold significant clinical implications, but remain largely unexplored. This study aimed to investigate the variability in initial CO responses to standing in both young and elderly populations. By using real-time haemodynamic recordings and autonomic analyses, we sought to characterize the distinct physiological profiles and explore their potential associations with important clinical endpoints and treatments relevant to the health of older adults.

Methods

Study design and participants

Two independent prospective cohorts were used in this study: A younger adult cohort from the University of Melbourne and an older adult cohort from The Irish Longitudinal Study on Ageing (TILDA). A schematic overview of the study design is provided in Supplementary data online, Figure S1. Ethical approval was obtained from the Human Research Ethics Committee of the University of Melbourne (Melbourne cohort) and the Faculty of Health Sciences Research Ethics Committee at Trinity College Dublin (TILDA cohort), and all participants provided written informed consent before enrolment. Details of participant exclusions are provided in the Supplementary Methods, with a flow diagram illustrating the inclusion process (Supplementary data online, Figure S2).

The Melbourne cohort included a total of 3074 medical students (mean age 22 years; 50% female; racial composition: 59% non-Asian and 41% Asian) recruited annually between 2011 and 2019 at the University of Melbourne. Participants were eligible if they had no history of cardiovascular disease, neurological disorders, or other chronic conditions affecting cardiovascular function. Standardized measurements of body weight and height were taken for each participant, from which body mass index (BMI) was calculated. Participants self-nominated their racial backgrounds, which were subsequently categorized into two groups: Asian and non-Asian, for the purposes of statistical analyses.

TILDA is a large-scale, nationally representative, prospective study focused on the ageing process in Irish adults aged 50 years and over.¹² All participants are of Caucasian descent. Data collected through TILDA include a range of health, lifestyle, and socio-economic measures, alongside clinical and biological assessments. The detailed design and methods of the study have been published elsewhere.¹³ The present study included 3025 participants from wave 1 (2009–13) and wave 3 (2014–15) who underwent a standardized active-stand test with continuous beat-to-beat cardiovascular monitoring.

Orthostatic challenge and group classification

All participants underwent a standardized active standing test with continuous beat-to-beat haemodynamic monitoring using the Finometer Midi system (Finapres Medical Systems, Amsterdam, The Netherlands). After ≥ 5 min of supine rest, participants stood independently and remained upright for 2 min. Continuous cardiovascular signals were acquired throughout the test, including SBP, DBP, HR, SV, CO, SVR, and the maximal steepness of the arterial pulse wave upstroke (dp/dt).² These measures were derived from pulse contour analysis of the finger arterial pressure waveform. SV, a key determinant of CO, was estimated using the Modelflow algorithm implemented in the Finometer system (further details, including considerations of arterial stiffness, are provided in the Supplementary Methods). This method reconstructs aortic flow from the finger arterial pressure waveform using a three-element Windkessel model of the arterial tree. SV was then computed as the numerical integration of simulated flow during systole.¹⁴ Its accuracy has been validated against invasive thermodilution (for point measures) or ultrasound (for beat-to-beat measures) methods, where it tracks closely during orthostatic testing,¹⁵ postural change,¹⁶ and other physiological and clinical challenges.^17–20

Participants were classified based on their CO response to standing. Those exhibiting a rise in CO relative to supine values were categorized as CO rise (CO_R), while those showing a predominant fall were categorized as CO fall (CO_F). Ambiguous or biphasic CO responses were classified according to criteria detailed in the Supplementary Methods (Supplementary data online, Figure S3). In the TILDA cohort, participants were classified as CO_F if they exhibited a predominant CO fall upon standing during either wave 1 or wave 3, to identify those with an underlying predisposition to this phenotype. Those who showed a CO rise in both waves were classified as CO_R. Where available in TILDA, clinical and outcome variables were averaged over the two waves, including for individuals classified as CO_F based on a single observation.

Haemodynamic and autonomic analysis

The protocol for obtaining supine, on standing, and standing values, as well as the changes upon standing, is detailed in the Supplementary Methods and illustrated in Supplementary data online, Figure S4. Resting supine and standing values were averages over 30 s periods, while values upon standing represent point estimates of the maximum deviation (fall or rise) rather than the average during this period.

Sympathetic vascular tone and baroreflex sensitivity (BRS) were continuously assessed throughout the active standing protocol (see Supplementary Methods for details). Sympathetic tone was assessed via spectral analysis of SBP variability using time-frequency analysis. The low frequency (LF) power of SBP variability (SBP_LF), which is highly correlated with muscle sympathetic activity, was used to describe the sympathetic vasomotor activity.²¹ Baroreflex sensitivity was computed continuously using the time-course BRS (xBRS) method.^22,23

Experimental subgroups and interventions

Five independent experimental subgroups were derived from the Melbourne cohort to facilitate more detailed mechanistic investigations (see Supplementary data online, Figure S5). All groups were representative of the entire cohort. Among them, four subgroups underwent a second orthostatic challenge, with their first challenge acting as a control. The fifth subgroup provided additional experimental data.

Reproducibility studies

To evaluate the repeatability of the CO phenotype, only participants from the untreated control arm who completed repeated assessments were studied. One untreated subgroup (n = 330) repeated the experiment on the same day, while another separate untreated subgroup (n = 328) completed a second experiment 2 weeks after their first test.

Acute vasodilation and sympathetic augmentation

In two subgroups, interventions were applied before the second orthostatic challenge. In one, acute vasodilation was induced by sublingual administration of glyceryl trinitrate (GTN, 300 µg) in 390 subjects, 15 min before the second orthostatic challenge. In a separate group of 519 subjects, sympathetic augmentation was achieved through sustained handgrip exertion (30% of maximum force for 30 s) in combination with phenylephrine administration (average dose 18 mg) before the second orthostatic challenge.

Urinary electrolyte excretion

Considering the potential influence of sodium intake on extracellular fluid volume and therefore on CO, 24-hour urine collections were obtained from 398 participants for assessment of urinary electrolyte excretion. Sodium and potassium concentrations were measured, and daily excretion rates (Na/d, K/d) as well as ratios to creatinine excretion (Na/Creat, K/Creat) were calculated.

Clinical measures

The detailed methods for the clinical measures and characteristics in the TILDA study have been published previously.^12,13 For this study, demographic data were obtained from waves 1 and 3, including age (from the last wave attended), sex, BMI, highest education level (from the last wave attended), smoking status, diabetes, hypertension, cardiovascular disease [CVD; defined as any of angina, heart attack, heart failure, irregular heart rhythm, heart murmur, stroke and/or transient ischaemic attack (TIA)], and reported use of antidepressant, antihypertensive or benzodiazepine medications. Clinical outcomes comprised reported falls (defined as a fall within the past year or since the previous wave), falls and/or blackouts (defined as a report of either a recent fall or a fainting/blackout episode), standing time, Timed Up and Go (TUG), normal gait velocity, maximum gait velocity (from wave 3 only), grip strength, Mini-Mental State Examination (MMSE)²⁴ score, Montreal Cognitive Assessment (MoCA)²⁵ and Fried frailty index. No primary outcome was pre-specified as the study was exploratory.

Statistical analysis

A full description of the statistical methods is provided in the Supplementary Methods. Phenotypes with skewed distributions were natural log-transformed before parametric analyses. Continuous variables are reported as mean with standard deviations (SD) or median with interquartile range (IQR) where appropriate. Two-tailed tests were used throughout. In the Melbourne cohort, differences in continuous phenotypes between CO_R and CO_F groups were analysed using linear regression models adjusting for age, sex, BMI, height, year of recording, self-reported race (Asian vs non-Asian), and CO group. For each standing-phase outcome, the corresponding supine measurement was added as a covariate. The primary model coefficients of interest were the adjusted differences between CO_R and CO_F groups, presented with 95% confidence intervals (CI). In the TILDA cohort, associations between CO phenotype and clinical outcomes were examined using logistic or linear regression models. Models were adjusted for age, sex, educational attainment, diabetes, BMI, smoking, medication use, and the presence of CVD (with CVD excluded when evaluated as an outcome). Medication use was included as dummy variables (yes/no) for antihypertensives, antidepressants, and benzodiazepines. Negative chronotropic drugs (β-blockers and non-dihydropyridine calcium channel blockers), being a subset of antihypertensives, were not included separately to avoid collinearity; instead, stratified analyses by negative chronotropic drug use were performed within the antihypertensive group. Covariates in both cohorts were selected a priori based on their established physiological or clinical relevance to CO and BP, consistent with prior analyses in each cohort. Statistical tests were performed using SPSS Statistics 30 (SPSS Inc, Chicago, IL, USA) in Melbourne and STATA 15.1 (StataCorp LLC, College Station, TX, USA) in the TILDA cohort. Details of model specification, diagnostic checks, and sensitivity analyses—including stratification by sex and antihypertensive medication use—are provided in the Supplementary Methods.

Results

Discovery of a novel cardiovascular phenotype

In 3074 healthy young adults who underwent non-invasive haemodynamic assessment during an active stand test, we observed the expected average response to standing: within the first 30 s of standing, SBP, SV, and SVR decreased, coupled with transient increases in HR and CO (Figure 1A). Baseline and postural haemodynamic characteristics for the full cohort, as well as for sex- and race-stratified subgroups, are provided in Supplementary data online, Tables S1–S3. Despite this consistent group-average pattern, analysis of individual changes in CO upon standing (Figure 1B) revealed a bimodal distribution (K-means cluster analysis, F = 5532.2, P < .001), with approximately 34% of participants experiencing a fall in CO, designated as CO_F. The CO_F phenotype was more prevalent in females and Asian subjects (see Supplementary data online, Table S4). Notably, the CO_R and CO_F phenotypes demonstrated high consistency within individuals when measurements were repeated either on the same day or 2 weeks later (Figure 1C and Supplementary data online, Table S5).

Figure 1.

Physiological characterization of cardiac output responses during orthostatic challenge. (A) Group-averaged time course tracings of continuously recorded haemodynamic parameters during an orthostatic challenge in the Melbourne cohort (n = 3074). (B) Frequency histogram depicting the distribution of individual changes in cardiac output from supine to standing in the Melbourne cohort. Among these, 34.2% exhibit a predominant fall in cardiac output (CO_F, n = 1,050, purple), while the remaining individuals show a rise in cardiac output (CO_R, n = 2,024, blue). (C) Consistency of the CO_R and CO_F phenotypes in individuals undergoing two orthostatic challenges, either on the same day (short-term) or two weeks apart (medium-term). (D) Haemodynamic tracings from the Melbourne cohort in CO_R and CO_F groups. (E) Haemodynamic tracings from the TILDA wave 1 screening phase in CO_R (n = 4184) and CO_F (n = 415) groups. CO, cardiac output; HR, heart rate; SBP, systolic blood pressure; SBP_LF, low frequency variability of systolic blood pressure; SV, stroke volume; SVR, systemic vascular resistance

Haemodynamic characteristics of the CO_F phenotype

In the Melbourne cohort, averaged cardiovascular tracings revealed clear distinctions between CO_R and CO_F groups in response to standing (Figure 1D). While both groups demonstrated an initial rise in CO in the first seconds of standing (associated with increasing HR), the CO_F group exhibited a subsequent and predominant fall in CO. This was accompanied by a larger drop in SBP, SV, and dp/dt, a lesser increase in HR, and a smaller fall in SVR (Figure 1D, Supplementary data online, Tables S6 and S7). There were also distinct haemodynamic differences at rest, characterized by lower SBP, DBP, pulse pressure (PP), SVR and dp/dt, higher HR in both lying and standing positions, and lower standing SV in CO_F subjects (Figure 1D, Supplementary data online, Table S6).

As CO is the product of SV and HR, we investigated the relationship between SV and HR from supine to standing. Across all participants, these changes were inversely correlated. In other words, the more the SV fell, the higher the HR rose, presumably in response to the falling BP. Interestingly, this correlation was weaker (z-statistic = −2.11, P = .017) in CO_F subjects (r = −0.19, 95% CI: −0.13 to −0.25) than in CO_R (r = −0.27, 95% CI: −0.23 to −0.31), suggesting reduced HR compensation in response to SV decline. This blunted response likely amplifies the CO drop in CO_F individuals and contributes to the overall dichotomy in CO responses.

To assess whether differences in sodium or potassium intake contributed to these patterns, we compared 24-hour urinary electrolyte excretion. No significant differences were observed between the CO_R and CO_F groups (see Supplementary data online, Table S8).

To evaluate the presence and relevance of the CO_F phenotype in older individuals, we analysed cardiovascular responses to active standing in 3025 older participants from TILDA. Among these, 22.2% exhibited the CO_F phenotype with a predominant fall in CO upon standing during at least one assessment wave (waves 1 or 3). Figure 1E illustrates the orthostatic cardiovascular response profiles of subjects with CO_R and CO_F phenotypes in wave 1 (similar tracings were observed for wave 3). These profiles reflected our original observations in young adults (Figure 1D). Before standing, CO_F individuals displayed lower SBP and higher HR; upon standing, they experienced a larger drop in SBP and a blunted HR rise. Differences in responses in SV and SVR were more pronounced than in younger adults, such that SV increased in CO_R but not in CO_F subjects, while SVR rose initially in CO_F (Figure 1E), potentially reflecting age-related shifts in orthostatic compensation.²

Autonomic characteristics of the CO_F phenotype

We next examined whether the CO_F phenotype was associated with altered autonomic regulation. In all CO groups, in both Melbourne and TILDA, the orthostatic activation of the sympathetic nervous system was evident with the rise in SBP_LF just before standing, peaking within 15 s of standing, then declining to resting standing levels that were higher than supine (Figure 2A, Supplementary data online, Table S9). Notably, SBP_LF levels were significantly higher in the CO_F group in all postural phases in the young and upon standing in older subjects. Consistent with previous literature,²⁶ xBRS levels were overall lower in the older subjects, and the profile of change upon standing was slightly different. In the young, xBRS levels were significantly lower in the CO_F subjects across all postural phases. In the older group, xBRS was also nominally lower throughout, but the difference reached statistical significance only while standing (see Supplementary data online, Table S9).

Figure 2.

Autonomic dynamics during orthostatic challenge and interventions in CO_R vs CO_F. (A) Group-averaged time courses of sympathetic vascular activity (indexed by SBP_LF) and baroreflex function (xBRS) during orthostatic challenge in the Melbourne (CO_F: n = 1050; CO_R: n = 2024) and TILDA (CO_F: n = 291; CO_R: n = 3190) cohorts. TILDA data are from wave 3 only, as this wave provided unfiltered signals suitable for spectral analysis. (B) Cardiac output responses to standing before and after sublingual GTN (300 µg) in CO_R (n = 205) and CO_F (n = 185), and before and after sympathetic augmentation (handgrip + phenylephrine) in CO_R (n = 349) and CO_F (n = 170) in the Melbourne cohort. (C) Linear regression adjusted group differences (CO_R vs CO_F) in changes in haemodynamics during standing, before and after GTN or sympathetic augmentation in the Melbourne cohort. *P < .001 for within-group comparisons; ^†P < .001 for between-group comparisons

Effects of vasodilation and sympathetic augmentation

The administration of the vasodilator GTN before standing altered the cardiovascular responses in CO_R subjects to resemble patterns seen in the CO_F group (Figure 2B). Specifically, CO_R individuals developed a dip in CO accompanied by greater reductions in SV and SBP (Figure 2C, Supplementary data online, Table S10A). GTN also slightly exaggerated the CO drop in CO_F subjects (Figure 2B). Notably, GTN reduced supine and standing CO only in CO_F subjects, suggesting heightened sensitivity to GTN vasodilation even under resting conditions.

Conversely, augmentation of sympathetic activation had minimal impact on CO_R subjects (Figure 2B, Supplementary data online, Table S11A) but significantly counteracted the fall in CO upon standing in CO_F subjects, and resulted in higher SV and SBP in this group (Figure 2C, Supplementary data online, Table S11B).

Clinical correlates of the CO_F phenotype in older individuals

To assess clinical relevance, we compared outcomes between CO_R and CO_F groups using data pooled from both waves. The two groups differed in several baseline characteristics (Table 1). Therefore, appropriate adjustments were made in regression analyses for all group comparisons as detailed in the Supplementary Methods. After adjustment, the CO_F phenotype was independently associated with increased fall incidence, longer standing time, slower TUG and gait speeds, lower cognitive scores (MMSE and MoCA), higher frailty indices, and more frequent composite cardiovascular outcomes (Table 2). Grip strength did not differ significantly between groups. Moreover, there was no significant difference in the combined endpoint of falls and/or blackouts, suggesting the inclusion of blackouts may dilute the predictive signal for falls alone. Similarly, self-reported symptoms such as dizziness or light-headedness after standing did not differ significantly between CO_R and CO_F participants (chi-squared test, P = .454). Analyses stratified by sex reflected the same CO group differences for males and females as for the whole group (see Supplementary data online, Table S12) except that falls were significantly more frequent in CO_F males but not so in females.

Table 1.

Characteristics of CO_F and CO_R groups in the TILDA cohort

	COR	COF	P-value
N (%)	2353 (77.8)	672 (22.2)
Age, years	63.0 (58.0–69.0)	64.0 (58.0–71.0)	.220
Females, n/N (%)	1241/2353 (52.7)	455/672 (67.7)	<.001
Body weight, kg	76.8 (67.3–87.3)	79.4 (68.7–92.6)	<.001
BMI, kg/m2	27.6 (25.1–30.4)	29.4 (26.1–33.6)	<.001
Education			<.001
Primary/none, n/N (%)	345/2353 (14.7)	177/672 (26.3)
Secondary, n/N (%)	949/2353 (40.3)	256/672 (38.1)
Third/higher, n/N (%)	1059/2353 (45.0)	239/672 (35.6)
Smoking status, n/N (%)	319/2353 (13.6)	138/672 (20.5)	<.001
Hypertension, n/N (%)	1157/2353 (49.2)	316/671 (47.1)	.342
Diabetes, n/N (%)	149/2353 (6.33)	84/672 (12.5)	<.001
CVD, n/N (%)	440/2353 (18.7)	187/672 (27.8)	<.001
Antihypertensive use, n/N (%)	843/2353 (35.8)	336/672 (50.0)	<.001
Negative chronotropic medications, n/N (%)	339/2353 (14.4)	148/672 (22.0)	<.001
Antidepressant use, n/N (%)	189/2353 (8.0)	102/672 (15.2)	<.001
Benzodiazepine use, n/N (%)	65/2353 (2.8)	38/672 (5.7)	<.001

Phenotypes are summarized as n (proportions) or median values (interquartile range). BMI, body mass index; CVD, cardiovascular disease, defined as ‘Yes’ if the participant reported any of the following conditions in any wave: angina, heart attack, heart failure, abnormal heart rhythm, heart murmur, stroke, and/or transient ischemic attack; otherwise, ‘No’.

Table 2.

Comparative analysis of clinical outcomes between CO_F and CO_R groups in TILDA

Clinical measures	COR	COF	Odds ratio or B	95% CI	P-value
Falls, n/N (%)	789/2350 (33.6)	278/619 (44.9)	1.26	1.04 to 1.52	.016
Falls and/or blackouts, n/N (%)	1086/2325 (46.7)	354/617 (57.4)	1.17	0.97 to 1.40	.099
Standing time, s	6.5 (5.5–8.0)	8.0 (6.5–10.0)	0.92	0.69 to 1.14	<.001
Timed Up and Go, s	8.3 (7.6–9.2)	8.9 (7.8–10.4)	0.67	0.45 to 0.88	<.001
Normal gait velocity, cm/s	139.7 (128.7–150.2)	131.5 (115.9–145.2)	−4.19	−5.62 to −2.77	<.001
Maximum gait velocity, cm/s	174.5 (158.3–190.9)	165.2 (145.3–182.9)	−4.21	−6.43 to −1.99	<.001
Grip strength, kg	25.8 (20.3–34.0)	22.6 (18.0–29.1)	−0.30	−0.81 to 0.21	.254
MMSE score	29.5 (28.5–30.0)	29.0 (28.0–29.5)	−0.25	−0.38 to −0.12	<.001
MOCA score	26.5 (24.5–28.0)	25.5 (23.0–27.5)	−0.64	−0.89 to −0.39	<.001
Combined CVD outcomes, n/N (%)	440/2353 (18.7)	187/672 (27.8)	1.39	1.11 to 1.75	.005
Fried frailty category	0.0 (0.0–0.5)	0.5 (0.0–1.0)	0.19	0.13 to 0.25	<.001

Phenotypes are summarized as n (proportions) or median values (interquartile range). Summary values from TILDA waves 1 and 3 are presented except for maximum gait velocity, which is derived solely from wave 3. The combined CVD outcomes considered were angina, heart attack, heart failure, abnormal heart rhythm, heart murmur, stroke, and/or transient ischemic attack.

MMSE, Mini-Mental State Examination; MOCA, Montreal Cognitive Assessment.

To investigate relevant medication effects, we stratified analyses by antihypertensive usage. Differences in mobility and cognitive outcomes between the CO_R and CO_F groups were independent of antihypertensive usage (see Supplementary data online, Table S13). However, the association between CO_F and falls was only significant among those taking antihypertensives, and particularly those on negative chronotropic agents (e.g. beta-blockers, non-dihydropyridine calcium channel blockers) (see Supplementary data online, Table S14). In subjects taking negative chronotropic drugs (n = 487), the rise in HR from supine to peak upon standing was significantly lower in the CO_F subjects (median: 10.9, IQR: 7.6) than in the CO_R group (median: 12.1, IQR: 6.4) (Kruskal–Wallis test, P = .006).

Discussion

In this study, we identified a distinct and reproducible subgroup of individuals, observed consistently across both young and older populations, who exhibited a transient fall in CO upon active standing, a response that diverges from the expected physiological increase. This novel phenotype (CO_F) appears to reflect a peripheral impairment in sympathetic cardiovascular regulation, with a dissociation between autonomic activation and its vascular and cardiac effects. Importantly, in older adults, the CO_F phenotype was associated with a range of adverse clinical features, including increased risk of falls, impaired mobility, lower cognitive performance, and frailty (Structured Graphical Abstract). These associations point to a broader dysfunction underlying CO_F, potentially extending beyond autonomic regulation, and underscore its significance as a physiologically distinct and clinically informative phenotype.

Mechanistic basis

Detailed characterization of the CO_F phenotype across both the Melbourne and TILDA cohorts revealed a consistent haemodynamic pattern (Figure 1D and E). The core physiological abnormality appears to be an exaggerated fall in SV that implies reduced venous return to the heart, most likely the result of pooling in the splanchnic capacitance veins. Under normal conditions, standing triggers a rapid, baroreflex-mediated sympathetic response, which constricts splanchnic capacitance veins, reduces venous compliance,²⁷ and maintains preload. In CO_F individuals, SV falls disproportionately despite enhanced markers of sympathetic activation, suggesting an impairment in the sympathetic signal transmission.

In addition to preload failure, CO_F individuals show a blunted increase in HR, impaired BRS and reduced cardiac contractility (as indexed by dp/dt), pointing to a broader deficit in efferent sympathetic effect. These abnormalities are evident not only during orthostatic challenge but also in the supine state, where lower BP, cardiac contractility, and SVR are observed. While some reduction in dp/dt can be explained by reduced preload through the Frank-Starling mechanism, the pattern of recovery provides further insight. The exaggerated fall in dp/dt in CO_F subjects (Figure 1D and E) mirrors the initial changes in SV upon standing. However, after this nadir, CO_R subjects show an increase in dp/dt above supine values, but CO_F subjects do not. SV alone is not an explanation for this difference as it remains low in the younger CO_R subjects and is comparable between older CO_R and CO_F subjects. This would suggest that the CO_R group is more responsive to the sympathetic input influencing cardiac contractility than are CO_F subjects. Together, these findings are consistent with impaired sympathetic transmission at the peripheral level.

Support for this interpretation comes from pharmacological manipulations. Augmenting sympathetic tone with combined handgrip and phenylephrine partially restored the haemodynamic profile in CO_F individuals, consistent with a functionally recruitable—but suboptimal—sympathetic effector axis. Conversely, venous dilation with GTN²⁸ induced a CO_F-like pattern in otherwise normal responders, reinforcing the critical role of venous tone and preload preservation in orthostatic CO maintenance (Figure 2B and C).

These proposed mechanisms are broadly aligned with classical models of orthostatic cardiovascular control, which emphasize the role of baroreflex-mediated sympathetic vasoconstriction and tachycardia in buffering postural circulatory shifts. However, CO_F diverges from other recognized syndromes. Unlike clinical OH, where BP falls after 2 min standing by more than 20/10 mmHg,¹ BP in CO_F subjects is only 8/5 mmHg lower than CO_R individuals. Similarly, while neurally mediated vasovagal syncope is typically characterized by paradoxical bradycardia, the CO_F pattern is one of a blunted orthostatic tachycardia. Notwithstanding, a recent study identified low PP (likely reflecting low SV) during adulthood as a typical profile in those experiencing neurally mediated reflex syncope.²⁹ These subjects shared relevant haemodynamic characteristics with our CO_F subjects. The authors proposed a tendency for blood to pool in subdiaphragmatic regions or low blood volume as possible contributory factors to reflex syncope. Although we did not measure blood volume, our studies of 24-hour urinary sodium excretion gave no indication of reduced sodium intake (and by implication extracellular fluid volume) in CO_F subjects.

Our hypothesis of impaired sympathetic neurotransmission may seem at odds with two observations. The first is the higher resting HR in CO_F subjects when supine and standing at rest. However, this occurs in the context of lower BP, and it is arguable that the increased HR is inadequate given the reduced BP. In this regard, the impaired BRS observed in the CO_F group is likely relevant. The second is the higher SVR upon standing in the CO_F group. However, the factors affecting SVR here are complex and include local metabolic activity and beta-adrenergic mediated vasodilation in active skeletal muscle.¹ The elevated SVR observed in CO_F subjects may imply inadequate skeletal arteriolar dilation as a result of diminished response to adrenaline³⁰ or to tissue metabolites such as tissue oxygen, adenosine, and potassium levels.³¹

It is noteworthy that the CO_F phenotype was less prevalent in the older TILDA cohort (22%) than in the younger Melbourne cohort (34%), which may appear counterintuitive given the expected age-related decline in cardiovascular control. Among screened individuals who did not participate in the TILDA study, a higher proportion may have had the CO_F phenotype, as frailty, disease, or other health-related limitations could have prevented them from attending the study centre or completing the active stand. The exclusion of these individuals from testing may therefore have resulted in an underestimation of CO_F frequency in the analysed sample. Physiological factors may also be relevant, including age-related increases in basal sympathetic activity³² and reduced abdominal venous capacitance³³ in older adults. Future investigations in more diverse and representative ageing populations will be needed to clarify this finding.

Clinical implications

In the TILDA cohort, the presence of the CO_F phenotype in older adults was consistently observed and associated with a more complex and less healthy clinical profile (Table 1). For example, individuals with CO_F were more likely to be taking antihypertensive medications (50% vs 36%), including agents that reduce HR or vascular tone, potentially blunting compensatory responses to standing. In addition, they had higher rates of CVD (28% vs 19%), which may limit cardiac reserve, and a greater prevalence of diabetes (13% vs 6%), a condition known to impair autonomic regulation. These coexisting factors may help reveal or amplify an underlying predisposition to the CO_F phenotype, particularly in the context of diminished cardiovascular and autonomic resilience in later life.

Notably, even after adjusting for demographic and clinical differences, individuals with the CO_F phenotype demonstrated a higher frequency of reported falls, as well as significantly lower scores in both mobility and cognitive performance (Table 2). Falls are the leading cause of injury in older people³⁴ and a major health cost.³⁵ In CO_F participants, the observed reduction in CO per se,¹⁰ along with the associated low standing BP, frailty, impaired mobility, and cognition, are relevant risk factors for falls. Interestingly, IOH alone has been reported not to be associated with a higher risk of falls in the TILDA population.³ There was also poor agreement between IOH and OH, which was associated with an increased risk of falls.³ In the present analyses also, IOH was not associated with increased falls risk (P = .93, data not shown), but falls were associated with an initial fall in CO (44.9% vs 33.6%, P = .016, Table 2). Additionally, our stratified analyses revealed the relationship between CO_F and falls was statistically significant among individuals taking antihypertensive medications and particularly those on drugs with negative chronotropic effects (see Supplementary data online, Table S14). These observations highlight the importance of recognizing the CO_F phenotype in clinical settings. Identifying individuals with the CO_F phenotype could help guide medication strategies to minimize falls risk.

Interestingly, reported symptoms commonly associated with orthostatic intolerance, such as blackouts or dizziness, did not differ in CO_F subjects. This aligns with previous findings from the TILDA cohort, where only about one-third of individuals with OH reported symptoms like dizziness, unsteadiness, or light-headedness after standing. Notably, those with asymptomatic OH exhibited a higher risk of future unexplained falls.³⁶ Such symptom paucity poses a significant challenge for clinical identification of CO_F, highlighting the need for continuous, beat-to-beat cardiovascular monitoring to detect these covert haemodynamic abnormalities. Unfortunately, such monitoring is limited in routine clinical practice. This limitation is particularly important when considering that real-life orthostatic stress often involves prolonged (e.g. ≥3 min)³⁷ or repeated (e.g. 40–50 sit-to-stand transitions per day)³⁸ episodes of standing, which may impose greater cumulative haemodynamic strain than the early responses typically studied in laboratory settings. While our study focused on the immediate transition to upright posture, future work is needed to explore whether such covert impairments persist, and whether their clinical consequences accumulate, under conditions that better reflect everyday behaviour. Expanding access to such technologies, especially through advances in wearable devices and artificial intelligence, holds promise for earlier detection of covert abnormalities like CO_F, enabling timely intervention to reduce falls risk and preserve function in vulnerable individuals.³⁹

Strengths and limitations

The strengths of this study include the use of two large, well-characterized cohorts comprising over 6000 participants, including healthy young adults and community-dwelling older adults, enabling age-stratified comparisons with adequate statistical power and generalizability. Real-time, beat-to-beat haemodynamic monitoring during active standing enabled detailed characterization of early cardiovascular adjustments. Mechanistic insights were further strengthened by targeted physiological and pharmacological interventions, which helped probe the role of sympathetic vasoconstriction. The integration of clinical outcome measures, including falls, gait speed, frailty, and cognitive performance, provided a multidimensional perspective on the clinical relevance of these phenotypes. Nevertheless, several important limitations should be acknowledged. First, participant exclusions may have introduced selection bias, as individuals unable or unwilling to attend the centre or complete the active stand are plausibly less healthy. This could bias the analysed sample toward healthier participants, potentially underestimating the prevalence of abnormal haemodynamic responses and attenuating associations with clinical characteristics. Second, the reproducibility of phenotype appeared lower in the older TILDA cohort, assessed using data from two waves (Wave 1: 2009–13; Wave 3: 2014–16). This analysis was not pre-planned and was limited by incomplete follow-up. The long interval between measurements, typically exceeding one year, together with external factors such as vasoactive drug use, diabetes, and cardiovascular disease, likely contributed to the lower reproducibility. These factors should be considered when interpreting clinical associations in elderly adults. Third, sympathetic vascular modulation was inferred indirectly through time–frequency analysis of SBP variability and not from direct muscle sympathetic nerve activity recordings or plasma norepinephrine spillover measurements.³² Fourth, we did not measure other cardiovascular factors such as the activity of the renin–angiotensin–aldosterone system, blood volume, data on prolonged³⁷ or repeated³⁸ standing, hydration status, 24-hour ambulatory BP measures or echocardiographic assessments of cardiac structure and performance. Finally, data on cardiovascular comorbidities were based primarily on self-report, which introduces the potential for recall bias and diagnostic misclassification.

Conclusions

These results bring a new focus to the response of CO upon standing. A novel orthostatic CO phenotype (CO_F), characterized by a transient fall in CO upon active standing, was identified consistently in both young and older populations. Physiological analyses implicate impaired peripheral sympathetic regulation in its mechanism. Among older adults, the sentinel CO_F phenotype has revealed unexpected and far-reaching associations with conditions of major health relevance. Much is yet to be explored beyond this initial discovery. A deeper understanding of the molecular, cellular, physiological, and clinical aspects of this novel physiological paradigm offers the promising prospect of new means of risk prediction, preventive strategies, and therapeutic interventions.

Supplementary data

Supplementary data are available at the European Heart Journal online.

Declarations

Disclosure of Interest

Nothing to declare.

Data Availability

The data supporting this study are available from the corresponding author upon reasonable request.

Funding

The following funding sources are acknowledged: - Financial support for TILDA: Provided by the Irish Government's Department of Health, the Health Research Board, and the Atlantic Philanthropies.

Ethical Approval

The Melbourne cohort study was approved by the Human Research Ethics Committee of the University of Melbourne, and all participants provided written informed consent.

The Irish Longitudinal Study on Ageing (TILDA) was approved by the Faculty of Health Sciences Research Ethics Committee at Trinity College Dublin, and all participants provided written informed consent.

Pre-registered Clinical Trial Number

Not applicable.