Timing of orthostatic hypotension and its relationship with falls in older adults

Abstract

Background:

There is inconsistent evidence on the optimal time after standing to assess for orthostatic hypotension. We determined the prevalence of orthostatic hypotension at different time points after standing in a population of older adults, as well as fall risk and symptoms associated with orthostatic hypotension.

Methods:

We performed a secondary analysis of the Study to Understand Fall Reduction and Vitamin D in You (STURDY), a randomized clinical trial funded by the National Institute on Aging, testing the effect of differing vitamin D3 doses on fall risk in older adults. STURDY occurred between July 2015 and May 2019. Secondary analysis occurred in 2022. Participants were community-dwelling adults, 70 years or older. In the orthostatic hypotension assessment, participants stood upright from supine position and underwent six standing blood pressure measurements (M1–M6) in two clusters of three measurements (immediately and 3 min after standing). Cox proportional hazard models were used to examine the relationship between orthostatic hypotension at each measurement and subsequent falls. Participants were followed until the earlier of their 24-month visit or study completion.

Results:

Orthostatic hypotension occurred in 32% of assessments at M1, and only 16% at M5 and M6. Orthostatic hypotension from average immediate (M1-3) and average delayed (M4-6) measurements, respectively, predicted higher fall risk (M1-3 = 1.65 [1.08, 2.52]; M4-6 = 1.73 [1.03, 2.91]) (hazard ratio [95% confidence interval]). However, among individual measurements, only orthostatic hypotension at M5 (1.84 [1.16, 2.93]) and M6 (1.85 [1.17, 2.91]) predicted higher fall risk. Participants with orthostatic hypotension at M1 (3.07 [1.48, 6.38]) and M2 (3.72 [1.72, 8.03]) were more likely to have reported orthostatic symptoms.

Conclusions:

Orthostatic hypotension was most prevalent and symptomatic immediately within 1–2 min after standing, but more informative for fall risk after 4.5 min. Clinicians may consider both intervals when assessing for orthostatic hypotension.

Editor’s Note

This is a carefully performed and clearly written analysis on the timing of orthostatic hypotension (OH) and its relationship to postural symptoms and falls. Hypotension and OH causing near-syncope, syncope, and falls have been the most common iatrogenic conditions I have recognized over my 40 years of caring for older patients. It can be very difficult to diagnose, and I have often had to surmise this was the problem because of the clinical presentation and the medication list. It is very difficult to take blood pressure (BP) in an older person who just fell, particularly if they are in pain. Two people are often required. When an older person falls at home, paramedics do not generally check for OH, and/or may give fluids in the ambulance, which could mask the cause of the fall.

Colleagues in Ireland have published data in JAGS on beat-to-beat BP measurements. They found that some older people have rapid, relatively large drops in BP within 15–30 s of standing, and a subgroup of these have a slow recovery in BP which may underly some falls. This rapid drop mimics what is commonly observed in clinical practice. Beat-to-beat BP is, however difficult to detect without a device that costs as much as $40,000. More practical methods are needed.

The findings of the current study are somewhat surprising. Although OH was more common at 1 min after standing, OH at 3–4 min was more predictive of subsequent falls. Checking for delayed OH may be challenging in a busy office practice. Thus, it is possible that both rapid and delayed OH can cause falls, and we are just not detecting them.

The current study highlights the need for future research on the timing of checking for OH in clinical practice. Such research is likely to improve detection of OH and guide clinicians in identifying a major risk factor for falls and syncope and related morbidity and costs in the older population.

Joseph G. Ouslander, MD

INTRODUCTION

Orthostatic hypotension (OH) is common among older adults in the community and long-term care settings, and is a predictor of falls and death.^1–5 Clinicians screen for OH through one or multiple blood pressure (BP) measurements, taken after the patient stands from a supine position.⁶ However, there is limited agreement across guidelines as to when to perform standing BP assessments.⁷ While a number of guidelines recommend waiting approximately 3 min after standing, recent evidence suggests that more immediate assessments may detect both symptomatic OH and OH associated with adverse events.^8,9 However, these findings have not been replicated.

The Study to Understand Fall Reduction and Vitamin D in You (STURDY) was a double-blind, randomized trial, testing the effects of four doses of vitamin D3 (200, 1000, 2000, and 4000 IU/day) on fall risk in older adults (70 years and older) with low serum 25-hydroxyvitamin D [25(OH) D] levels (10–29 ng/mL). STURDY did not find any benefit from higher doses of vitamin D3 supplementation on the risk of falls or seated OH.^10,11 In an ancillary study to STURDY, we implemented a standardized supine OH protocol with standing measurements timed both immediately and 3 min after standing.

The purpose of the present study was to (1) characterize changes in BP upon standing, (2) determine associations between OH detected at different times after standing and fall risk, and (3) examine the association between OH detected at different times after standing and orthostatic symptoms. We hypothesized that OH would be more prevalent shortly after standing, and that older adults with OH would have a higher risk of falls and orthostatic symptoms regardless of timing.

METHODS

STURDY trial and ancillary studies

STURDY was a National Institute on Aging-funded randomized trial, occurring between July 2015 and May 2019 in Maryland at two community-based research clinics (Hagerstown, MD; Woodlawn, MD). A detailed description of the trial’s design and results are elsewhere (STURDY is registered on clinicaltrials.gov under identifier NCT02166333).^10,12 Beginning in July 2017, participants began participating in a supine OH ancillary study with timed standing BP assessments, funded by the National Heart, Lung, and Blood Institute. All participants provided written, informed consent; both the parent trial and the OH ancillary study were approved by a Johns Hopkins University Institutional Review Board.

Participants

STURDY enrolled community-dwelling adults, who were aged 70 years or older. Participants were required to be at increased fall risk, defined based on the presence of one of the following: two or more falls in the preceding year, at least one injurious fall in the preceding year, a fear of falling, difficulty maintaining balance, or use of an assistive walking device. Participants were also required to have a serum 25-hydroxyvitamin D [25(OH)D] level between 10 and 29 ng/mL. Older adults with cognitive impairment, hypercalcemia, kidney stones, consuming >1000 IU/day of vitamin D3 supplements, or consuming >1200 mg/day of calcium supplements were excluded from the study.

OH assessment

The supine orthostatic ancillary study assessment was performed during the randomization, 3-, 12-, and 24-month visits, using a standardized protocol.¹³ Participants could have multiple OH assessments over up to four study visits, depending on when they began or ended the trial. Time of day for OH assessment was not standardized, though participants could not undergo assessment less than 30 min following a previous meal. No explicit instructions were given in regard to the use of prescription medications, including antihypertensives, prior to the assessment.

All participants lied supine for 5 min with their arm by their side and then underwent three BP measurements, each of which took approximately 30–45 s from the beginning of automated BP cuff inflation to when the device reported BP and heart rate data. A 5-s pause took place between the completion of each BP measurement (after the 30–45 s inflation-deflation period) and the initiation of the subsequent measurement. After the three supine measurements, participants were asked to stand and place their arm on a bedside table at heart level at a 70–80 degree angle from their torso. After standing, participants underwent an “immediate” cluster of three BP measurements (with the same 30–45 s result time and 5-s pause between measurements). At 3 min after standing, participants underwent another cluster of three measurements (with the same 30–45 s result time and 5-s pause between measurements). The time required to stand, and the time of each BP measurement, was recorded.

The mean of the three supine BP measurements was used to determine change in BP at each of the six standing BP measurements (M1–M6). The average of the first three (M1–M3) and second three (M4–M6) standing BP measurements were also examined relative to the supine measurements to determine the difference in BP after standing, both immediate (i.e., within 3 min) and delayed (i.e., after 3 min). OH was defined as a drop in systolic BP (SBP) of at least 20 mm Hg or a drop in diastolic BP (DBP) of at least 10 mm Hg, based on thresholds established by the consensus OH definition.¹⁴ In addition, we also defined: (1) systolic OH as a drop in SBP of at least 20 mm Hg regardless of change in DBP and (2) diastolic OH as a drop in DBP of at least 10 mm Hg regardless of change in SBP. Heart rate was measured at the time of all BP assessments. Orthostatic tachycardia was defined as an increase in heart rate of at least 20 beats per minute after standing.

Another, seated OH protocol was performed at baseline, 3-month, 12-month, and 24-month visits for nearly all participants. In contrast to the supine protocol, this seated protocol assessed seated-to-standing BP (with a total of three standing measurements) and did not report the time for individual measurements. The seated protocol was introduced to STURDY prior to the supine protocol. The supine assessment was performed during the randomization visit (about 2 weeks after the baseline visit), 3-month visit, 12-month visit, and 24-month visit. As a result of this later implementation, a number of participants did not have the time-stamped, supine assessment. In cases where participants underwent both the supine and the seated protocol, these protocols were performed the same day (all visits except the baseline visit) with the seated protocol occurring first, followed by the supine protocol.¹³

Falls

Falls were defined using the World Health Organization (WHO) definition of a fall: “an event which results in a person coming to rest inadvertently on the ground or floor or other lower level.”¹⁵ In STURDY, falls were ascertained three ways over the up to two-year follow-up period: monthly calendars, scheduled clinic visits and telephone calls, and ad hoc telephone contacts (participants called the clinic if they fell). Each month participants received a calendar and were asked to indicate whether a fall had occurred daily. Participants continued in surveillance for recurrent falls, even after an incident fall occurred, for up to 2 years or until the study ended.

Associated symptoms

Before undergoing supine BP measurements, participants answered questionnaires about the frequency with which they experienced specific symptoms upon standing over the prior 30 days. The symptoms included: lightheadedness, dizziness, fainting, blacking out, seeing spots, imbalance, room spinning, heart racing, sweating episodes, vision changes, nausea, trouble concentrating, shortness of breath, headache, muscle weakness, and fatigue. For each symptom, participants rated symptom frequency upon standing on a 9-point Likert scale, from never to every time. Responses were dichotomized as either no symptoms (score 1) or any symptoms (scores 2–9). Participants completed these questionnaires at the randomization visits as well as the 3-, 12-, and 24-month visits.

In addition to the symptoms above, between the immediate and the delayed BP measurement clusters during the supine OH protocol, participants were asked to answer the following question on a 9-point Likert scale: “As you stood or as you are standing right now, on a scale from 1 to 9, with 1 being ‘no symptoms’ and 9 being the ‘worst possible,’ please rate if you feel (or felt): dizziness, lightheadedness, faint, or like you might black out.”

STURDY participants also underwent a seated OH assessment as part of the primary trial protocol that included three standing BP measurements performed after 1 min of standing, separated by a 30-s time interval. The result times were not recorded during the seated protocol. However, both immediately after standing and after the three standing BP measurements, participants were asked if they felt lightheaded or dizzy.¹³ As described above, participants who underwent both seated and supine OH assessments, the seated protocol was always performed first. A comparison of the seated and supine protocols was described in depth elsewhere.¹³

Other covariates

Participants reported age, sex, and race (Black or non-Black). Seated SBP and DBP were based on the average of three measurements. Body mass index (BMI) was calculated from measured height (m) and weight (kg). Participants also reported the presence or absence of the following conditions: cardiovascular disease, diabetes, Parkinson’s disease, high blood pressure or hypertension, and fall in the prior year. They also reported active medication use, including antidepressants, Parkinson’s disease-related medications, antipsychotics, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, beta-blockers, calcium channel blockers, diuretics, central alpha agonists, and alpha blockers. Participants were categorized according to their frailty status as “robust,” “pre-frail,” or “frail,” according to a standard definition.^16,17

Statistical analysis

Baseline characteristics for the overall population were characterized, using means and proportions. Generalized estimating equations (GEE; normal family, identity link, exchangeable correlation matrix with a robust variance estimator) were used to determine mean measurement time after standing as well as SBP, DBP, and HR after standing from the supine position for each of the six individual BP measurements (M1-6). Mean time poststanding did not include times prior to standing, but was referenced to 0 (i.e., the time both feet were planted on the ground). Models were repeated using the means of the immediate (M1-3) and delayed (M4-6) standing measurements, rather than individual measures. Model-derived means for both individual and mean standing measures were depicted via line graphs.

GEE (Poisson family, log link, exchangeable correlation matrix with a robust variance estimator) were also used to determine both the proportion of participants with consensus OH, systolic OH, diastolic OH, and orthostatic tachycardia at each time point and the change in proportion over time for both the individual standing measures (M1-6) as well as the mean immediate and delayed measures.

The primary outcome was time to new or recurrent fall. Absolute risk of falls from randomization was visualized with cumulative incidence plots; moreover, we determined the relative risk of falls associated with consensus OH, systolic OH, and diastolic OH at respective times post-standing, using Cox proportional hazards models adjusted for age, sex, race, study center, and time-to-stand. Our Cox models treated OH as a time-varying covariate. This allowed us to maximize data availability given later implementation of the protocol by allowing participants to enter surveillance at later visits in the study. Furthermore, for those with repeated OH assessments, OH status could be updated at the scheduled 3-month, 12-month, and 24-month visits, which is consistent with the intermittent presentation of OH observed clinically. We assessed the relation of OH exposures with multiple falls, following the Andersen-Gill approach for multiple failure survival analysis.^14,15 Fall surveillance for each person began with their first supine OH assessment. The randomization visit was the first possible supine OH assessment; however, a number of participants began assessments at later visits (see Supplemental Table 1). All models were adjusted for age, sex, race, and study center and included a robust variance estimator. We also used adjusted restricted cubic splines to characterize the continuous relationship between changes in standing BP measured at different time points and fall risk.

Finally, we determined the relationship between OH and participant-reported orthostatic symptoms experienced in the preceding 30 days or identified during the protocol, using GEE (binomial family, logit link, robust variance estimator, exchangeable correlation matrix).

Our analyses were restricted to participants with: (1) all nine planned BP measurements from the supine orthostatic protocol; (2) a documented time required to stand; and (3) no discrepancies recorded between the Omron-reported average and a calculated average of each cluster of the three BP measurements (Supplemental Figure 1). We excluded individuals where the third standing BP measurement resulted more than 180 s after standing or the first of the 3-min delayed measurements was more than 195 s after standing.

All analyses were performed using Stata 15.1 (Stata Corp, College Station, TX). A two-tailed p-value <0.05 was considered statistically significant without adjustment for multiple comparisons.

RESULTS

Baseline characteristics

Of the 491 participants with all six standing measurements, the mean age was 76.1 (standard deviation [SD] = 5.2) years; 58.5% were men and 16.9% reported Black race (Table 1). The average seated SBP was 129.1 (SD = 18.6) mm Hg and DBP was 67.8 (SD = 10.7) mm Hg. There were 61% of participants who reported a history of high BP and 63% of participants reported a history of a fall in the previous year. Frailty was detected in 8.4% of participants, while 58.9% were pre-frail and 32.8% were robust.

TABLE 1.

Baseline characteristics (N = 491).

	Mean (SD) or %
Age, years	76.1 (5.2)
Men, %	58.5
Black, %	16.9
Seated blood pressure, mm Hga
Systolic	129.1 (18.6)
Diastolic	67.8 (10.7)
Seated heart rate, beats per minute	65.7 (10.8)
Body mass index, kg/m2	30.1 (5.2)
History of cardiovascular disease, %	27.1
History of diabetes, %	25.9
History of Parkinson’s disease, %	1.4
History of high blood pressure, %	60.7
Antidepressant use, %	17.3
Parkinson’s disease medication use, %	3.1
Antipsychotic use, %	0.4
ACE inhibitor/ARB use, %	46.8
Beta blocker use, %	27.9
Calcium channel blocker use, %	22.4
Diuretic use, %	19.6
Central alpha agonist use, %	0.6
Alpha blocker use, %	9.0
Fall in the past year, %	62.5
Frailty, %
Robust	32.8
Pre-frail	58.9
Frail	8.4

Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; SD, standard deviation.

^aThere were only 490 participants with blood pressure measured in the seated position at baseline.

Timing of OH assessments

Participants took an average of 10.1 s (standard error [SE] = 0.3) to stand from the supine position. The immediate three BP measurements resulted within an average of less than 2 min 15 s (Table 2). For this cluster, BP and heart rate data resulted at 43.6 s (42.8 s, 44.5 s) (mean [95% confidence interval (CI)]) after standing for M1; at 1 m 27.4 s (1 m 26.3 s, 1 m 28.4 s) for M2; and at 2 m 11.9 s (2 m 10.6 s, 2 m 13.3 s) for M3. Delayed BP measurements resulted, on average, between 3 min 45 s, and 5 min 20 s. For this second cluster, BP and heart rate data resulted at 3 m 46.3 s (3 m 45.1 s, 3 m 47.5 s) for M4; at 4 m 31.7 s (4 m 30.2 s, 4 m 33.2 s) for M5; and at 5 m 16.9 s (5 m 15.1 s, 5 m 18.8 s) for M6.

TABLE 2.

Prevalence^a of orthostatic hypotension and tachycardia, % (95% confidence interval) (N = 491 participants with 5190 measurements).

	Time post-standing, mean (SE)	Consensus OH	Systolic OH	Diastolic OH	Orthostatic tachycardiab
M1	43.6 s (42.8 s, 44.5 s)	31.5 (28.1, 35.4)	25.2 (22.1, 28.7)	19.9 (17.0, 23.2)	7.3 (5.6, 9.5)
M2	1 m 27.4 s (1 m 26.3 s, 1 m 28.4 s)	19.3 (16.5, 22.6)	15.5 (13.0, 18.5)	12.8 (10.5, 15.6)	7.5 (5.8, 9.8)
M3	2 m 11.9 s (2 min 10.6 s, 2 m 13.3 s)	18.0 (15.2, 21.3)	12.2 (9.9, 15.0)	12.7 (10.4, 15.5)	5.8 (4.3, 7.8)
Mean earlyc	1 m 27.6 s (1 m 26.5, 1 m 28.6 s)	20.6 (17.7, 24.0)	15.3 (12.8, 18.4)	13.8 (11.4, 16.7)	5.9 (4.4, 7.9)
M4	3 m 46.3 s (3 m 45.1 s, 3 min 47.5 s)	17.1 (14.4, 20.2)	13.0 (10.6, 15.8)	10.7 (8.6, 13.3)	3.7 (2.5, 5.4)
M5	4 m 31.7 s (4 min 30.2 s, 4 m 33.2 s)	16.0 (13.4, 19.1)	10.4 (8.3, 13.0)	12.2 (10.0, 14.9)	4.2 (2.9, 6.0)
M6	5 m 16.9 s (5 m 15.1 s, 5 m 18.8 s)	16.4 (13.7, 19.5)	9.6 (7.6, 12.2)	13.0 (10.7, 15.9)	3.3 (2.2, 5.0)
Mean Latec	4 m 31.6 s (4 m 30.1 s, 4 m 33.1 s)	14.7 (12.2, 17.7)	9.9 (7.8, 12.5)	11.1 (8.9, 13.8)	3.3 (2.2, 5.1)

Abbreviations: OH, orthostatic hypotension; SE, standard error.

^aEstimated using generalized estimating equations. Total measurements differ from Supplemental Table 1 because supine measurements were used to calculate postural changes in BP used in determining OH.

^bOnly 488 participants, total of 5110 measurements.

^cFor the mean estimates there were 491 participants with 1730 measurements for blood pressure and 488 participants and 1703 measurements for heart rate.

Changes in standing blood pressure and OH

OH was most prevalent at M1 (31.5%; 95% CI: 28.1, 35.4), but only observed among 16.4% (95% CI: 13.7, 19.5) at M6 (Figure 1, Table 2). The greatest change from supine BP occurred at the first BP measurement immediately after standing (M1; mean = 43.6 s), when the mean change in SBP was −9.5 mm Hg (−10.8, −8.3) and mean change in DBP was −2.6 mm Hg (95% CI: −3.3, −1.9; Supplemental Figure 2, Supplemental Table 2). These changes stabilized with time, such that by M3 (mean- = 2 m 11.9 s), standing SBP was within −3.0 (−4.2, −1.8) mm Hg of supine SBP and standing DBP was within −0.7 (−1.4, −0.1) mm Hg of supine DBP.

FIGURE 1.

Orthostatic hypotension (OH) prevalence after standing from the supine position. Horizontal axis represents average time after standing for respective measurements (e.g., M1 = 0 m 44 s). Respective colors indicate different OH prevalence when using different diagnostic criteria. Black diamond = consensus OH (cOH). Blue circle = systolic OH (sOH). Green square = diastolic OH (dOH). Gray triangle = OH based on heart rate (hOH). N = 491 participants.

OH timing and new or recurrent falls

There were 425 new and recurrent falls over a median follow-up of 275 days. Cumulative incidences of falls were similar by OH status at each measurement (Supplemental Figures 3 and 4). OH detected within the first 3 min (M1, M2, M3) was not associated with falls; however, M5 and M6 were associated with falls with HRs of 1.84 (95% CI: 1.16, 2.93) and 1.85 (95% CI: 1.17, 2.91), respectively (Table 3). However, mean immediate and mean delayed measurements were associated with falls with HRs of 1.65 (95% CI: 1.08, 2.52) and 1.73 (95% CI: 1.03, 2.91), respectively. Greater decreases in SBP and DBP were associated with a higher risk of falls, especially for the delayed measurements (Supplemental Figures 5–8). Increases in DBP upon standing were also associated with higher risk of falls.

TABLE 3.

Association^a between orthostatic hypotension and first or recurrent falls, hazard ratio (95% confidence interval).

	Consensus OH	Systolic OH	Diastolic OH
M1	1.35 (0.93, 1.95)	1.36 (0.92, 2.00)	1.26 (0.80, 1.99)
M2	1.30 (0.81, 2.09)	1.42 (0.84, 2.39)	1.34 (0.73, 2.46)
M3	1.51 (0.94, 2.42)	1.86 (1.08, 3.19)	1.51 (0.84, 2.70)
Mean immediate	1.65 (1.08, 2.52)	1.72 (1.05, 2.79)	1.63 (0.96, 2.77)
M4	1.48 (0.90, 2.42)	1.57 (0.92, 2.68)	1.44 (0.74, 2.80)
M5	1.84 (1.16, 2.93)	1.82 (1.01, 3.28)	2.04 (1.23, 3.38)
M6	1.85 (1.17, 2.91)	1.99 (1.13, 3.51)	2.09 (1.29, 3.40)
Mean delayed	1.73 (1.03, 2.91)	1.71 (0.91, 3.23)	2.08 (1.19, 3.63)

Note: Bold numbers represent P < 0.05.

Abbreviations: CI, confidence interval; OH, orthostatic hypotension; SE, standard error.

^aValues generated with Cox proportional hazards models adjusted for age, sex, race, study center, and time-to-stand. There were 491 participants included in this analysis with 425 fall events. Median follow-up time was 275 days.

OH timing and symptoms

Participants who reported OH after completing all three BP measurements in the seated protocol were more likely to have OH during immediate BP measurements in the supine protocol (Table 4). However, OH was not associated with most of the reported orthostatic symptoms reported by participants 30 days prior to BP assessment protocols (Supplemental Tables 3 and 4).

TABLE 4.

Association^a of orthostatic hypotension detected^b at measurements with participant-reported symptoms (odds ratio, 95% confidence interval).

	Dizziness/lightheadedness, fainting, blacking out during supine protocolc, N = 490 (862 measurements)	Lightheaded or dizzy immediately after standing from a seated positiond, N = 484 (850 measurements)	Lightheaded or dizzy after the three standing blood pressure measures of the seated protocol were performedd, N = 489 (858 measurements)
M1	1.26 (0.91, 1.74)	0.74 (0.37, 1.51)	3.07 (1.48, 6.38)
M2	1.10 (0.73, 1.66)	0.79 (0.33, 1.89)	3.72 (1.72, 8.03)
M3	0.92 (0.59, 1.44)	0.57 (0.21, 1.51)	1.30 (0.52, 3.25)
Mean immediate	1.10 (0.73, 1.64)	1.18 (0.55, 2.55)	2.85 (1.31, 6.19)
M4	1.03 (0.68, 1.58)	0.62 (0.23, 1.64)	1.43 (0.57, 3.61)
M5	1.19 (0.78, 1.81)	0.99 (0.41, 2.39)	1.54 (0.58, 4.11)
M6	1.19 (0.79, 1.78)	0.79 (0.31, 1.98)	2.44 (1.02, 5.86)
Mean delayed	0.89 (0.56, 1.41)	0.74 (0.28, 1.98)	1.42 (0.52, 3.93)

Note: Several participants were missing different symptom data. Bold numbers represent P < 0.05.

Abbreviations: CI, confidence interval; OR, odds ratio.

^aValues generated with generalized estimating equations (binomial family, logit link, robust variance estimator) adjusted for age, sex, race, study center, and time-to-stand.

^bDuring the supine protocol.

^cRefers to any time during the supine protocol, not limited to periods in which the participant was in the supine position.

^dThese symptoms were asked as part of a seated OH protocol both immediately after standing and after three standing blood pressures were measured. During the seated OH protocol, BP measurements were not timed and occurred after a 1 min delay with a programmed 30-s pause between measurements.

DISCUSSION

In this population of older adults at risk of falls, OH was most prevalent within 1 min of standing from the supine position, and OH within 1.5 min was most closely associated with lightheadedness and dizziness. However, participants with OH after 4.5 min of standing were at higher risk of falls. Furthermore, continuous decrease in BP was increasingly predictive of fall risk based on restricted cubic splines, and this effect was most strongly noted with delayed BP measurements.

There is ongoing debate as to the clinical significance of OH detected immediately after standing.^18–22 Recent guidelines recommend waiting 1 min before assessing OH to avoid transient BP fluctuations, which corresponds to protocols used in ACCORD and SPRINT trials.^19,20,23 However, the ARIC cohort study, a community-based population of over 11,000 middle-aged adults, found that measurements within 1 min of standing were strongly associated with falls (hazard ratio of 1.29).²⁴ Similarly, in the present study, the magnitude of the association between immediate OH and falls was comparable to ARIC (hazard ratio of 1.35); however, the association in STURDY was not significant, possibly due to its smaller sample size. Notably, ARIC did not include measurements after 2 min. However, in the present study, OH after 3 min of standing was more predictive of fall risk than OH prior to 3 min of standing.

In previous work, dizziness after standing was a strong predictor for incident dementia and stroke, even stronger than measured OH.²⁵ While neither of these events were ascertained in the STURDY trial, it remains noteworthy that more immediate OH was associated with lightheadedness or dizziness after standing from the seated position. The implications of such results are unclear. Future research should examine the association between patient-reported histories of lightheadedness and dizziness while standing, and fall risk.

Several limitations warrant discussion. First, while this study included nearly 500 older adults, we may have been underpowered to detect associations with falls, particularly with the more immediate time measurements. Second, participants were enrolled from a population at high risk of falls, not from a community-based sample. This could limit generalizability of our results to a more fall-prone subset of the older adult population. Frailty may influence BP regulation upon standing, and OH may indeed be a clinical indicator of frailty.²⁶ However, only 8.4% of participants in the present analysis were frail, and nearly one-third were robust, which suggests that our results are generalizable at least to non-frail older adults. Third, standing measurements were separated by a 5-s pause, which may not have been adequate for resolution of the effects of cuff inflation. Fourth, symptoms reported during the prior 30 days, may be subject to recall bias and less informative than symptoms reported during the seated or supine OH protocols. Fifth, we restricted analyses to participants with BP results that met our protocol requirements, which could exclude patients with delayed timing due to underlying differences in physiology. Sixth, we performed multiple comparisons, particularly with analyses examining associations between OH and symptoms. While our smaller sample increases the possibility of underpowered associations (false negatives), multiple comparisons increases the possibility of false associations. Finally, this was an observational study and may be impacted by residual confounding.

Our study demonstrates that while immediate assessments may be more sensitive for OH, OH detected beyond 3 min of standing was a stronger predictor for fall risk. Thus, when screening for falls in older adults, clinicians may consider tailoring the timing of BP measurements to the pertinent clinical question. Clinicians who wish to simultaneously screen for OH and assess for fall risk may consider obtaining two standing BP measurements—the first occurring immediately after standing, for instance, and a second occurring at approximately 4 min. Of particular importance in this study, is the observation that multiple measurements were required to detect clinically relevant OH, which may represent a more effective screening approach for patients with a higher index of suspicion for OH as a cause of falls.

In conclusion, in this study of older adults at risk for falls, BP measurements immediately after standing were more sensitive for detecting OH, while BP measurements delayed by 3 min after standing were more predictive of falls. These results suggest that clinicians screening for both the presence of OH and fall risk associated with OH, should perform both immediate and delayed BP measurements. However, further studies are needed to determine the reliability and clinical applicability of these findings.

Supplementary Material

Supplement Material

Supplemental Table 1. Tabulation of supine-to-standing blood pressure assessments during the study.

Supplemental Table 2. Changes in blood pressure after standing from the supine position.

Supplemental Table 3. Association of orthostatic hypotension detected at measurements performed immediately after standing with participant-reported symptoms.

Supplemental Table 4. Association of orthostatic hypotension detected at measurements performed after 3 min of standing with participant-reported symptoms.

Supplemental Figure 1. Study population reported in the blood pressure and orthostatic hypotension analysis.

Supplemental Figure 2. Changes in systolic blood pressure, diastolic blood pressure, and heart rate over 6 measurements after standing from supine position.

Supplemental Figure 3. Presence of orthostatic hypotension during immediate measurements and cumulative incidence of falls.

Supplemental Figure 4. Presence of orthostatic hypotension during delayed measurements and cumulative incidence of falls.

Supplemental Figure 5. Adjusted restricted cubic splines of the relationship between changes in standing and supine systolic blood pressure and fall events, immediate blood pressure measurements.

Supplemental Figure 6. Adjusted restricted cubic splines of the relationship between changes in standing and supine systolic blood pressure and fall events, delayed blood pressure measurements.

Supplemental Figure 7. Adjusted restricted cubic splines of the relationship between changes in standing and supine diastolic blood pressure and fall events, immediate blood pressure measurements.

Supplemental Figure 8. Adjusted restricted cubic splines of the relationship between changes in standing and supine diastolic blood pressure and fall events, delayed blood pressure measurements.

Key points

• Orthostatic hypotension was twice as prevalent 45 s after standing than 4–6 min after standing.

• Orthostatic hypotension detected between 4 and 6 min was more predictive of fall risk than orthostatic hypotension detected before this time period.

• Lightheadedness and dizziness were more strongly associated with orthostatic hypotension within 90 s than beyond this time period.

Why does this paper matter?

Geriatricians, and clinicians in general, lack sufficient evidence for the ideal timing for blood pressure measurement after standing to detect orthostatic hypotension among older adults, as well as information about the association between orthostatic hypotension and fall risk or orthostatic symptoms. The present report suggests that clinicians should consider both immediate and delayed BP measurements when assessing for OH, depending on the clinical situation. Sensitivity is higher for orthostatic hypotension when measuring blood pressure more immediately after standing from the supine position, but older adults with orthostatic hypotension detected 4–6 min after standing are at higher risk of subsequent falls. Therefore, earlier measurements may be more useful for the diagnosis of orthostatic hypotension, but later measurements may be more useful if orthostatic hypotension is suspected as a contributor to falls among older adult patients.