Improvement of Cardiovascular Prognostic Value of Endothelial Function Tests by Repeated Measurements

Abstract

Background: The clinical significance of monitoring endothelial function in predicting cardiovascular events has not yet been fully explored.

Methods and Results: Patients with diabetes or hypertension were enrolled in this study. Three flow-mediated dilation (FMD) and 2 EndoPAT reactive hyperemia index (RHI) were measured periodically in 3 years at 1.5-year intervals. Patients were followed up for 10 years after the examination period. In the FMD study, 136 patients were classified as those with consistently good endothelial function, as defined by FMD >7% on 3 occasions (n=33), and those without (n=103). In multivariate Cox analysis, patients who maintained high FMD had less thromboembolic major cardiovascular events or angina pectoris (n=24; hazard ratio [HR] 0.216; 95% confidence interval [CI] 0.047–0.985; P=0.048). In the EndoPAT study, 120 patients were classified as those with consistently abnormal endothelial function, as defined by RHI <1.67 on 2 occasions (n=34), and those without (n=86). There were 9 all-cause deaths and 10 hospitalizations for heart failure. Patients with consistent RHI <1.67 had a higher mortality (HR 10.794; 95% CI 1.520–76.629; P=0.017) and incidence of heart failure (HR 5.356; 95% CI 1.301–22.052; P=0.020).

Conclusions: Repeated measurements improved the predictive performance and revealed differences between FMD and EndoPAT RHI, which were better at predicting coronary events and heart failure, respectively.

What Is New?

Repeated measurements of EndoPAT were more reliable than single measurements in predicting 10-year mortality and heart failure. Repeated measurements of FMD predicted future incidence of a composite endpoint comprising thromboembolic major cardiovascular events and angina pectoris.

What Is Relevant?

Monitoring health status with FMD and EndoPAT help prevent cardiovascular events. Repeated measurements reveal the difference in cardiovascular events predictable by FMD and EndoPAT.

Vascular endothelial function has a significant impact on the development of cardiovascular disease.^1–3 Physiological tests have shown that vascular endothelial function can be improved by lifestyle modification and other factors.⁴

Several tests for vascular endothelial function have been devised, some of which already have sufficient evidence to support their use.^5,6 Flow-mediated dilation (FMD) evaluates the dilatation response of conduit arteries after the release of upper limb cuff occlusion. The EndoPAT test observes restoration of blood flow in the finger arterioles. The Framingham Heart Study found no statistically significant relationship between FMD and EndoPAT reactive hyperemia index (RHI), suggesting that these tests measure different aspects of vascular function.⁷ Previous research has established the value of EndoPAT RHI in predicting cardiovascular events, with an RHI value of 1.67 now accepted as the appropriate cut-off point.^8,9 In contrast, endothelial function is considered good with FMD >7%.^8,9

However, our expectation of these non-invasive tests is that regular testing will allow us to assess current status and modify treatment strategies as needed. Almost all of the evidence to date has been obtained with a 1-time measurement, and the clinical significance of monitoring in longitudinal observations has not yet been fully explored.

The aim of this study was to clarify whether repeated measurement of FMD, EndoPAT RHI is effective for evaluating cardiovascular health compared with the single measurement. A multicenter registry, known as the FMD-J, was previously established to record data obtained using simultaneous FMD and EndoPAT measurements.¹⁰ In this protocol, repeated measurements were performed at 1.5-year intervals. We analyzed the 10-year outcomes of patients who participated in this study at National Defense Medical College Hospital (Tokorozawa, Japan) and those who were followed with the same protocol.

Methods

Participants

The study included participants in the Flow-Mediated Dilation Japan (FMD-J) B trial enrolled at National Defense Medical College Hospital from 2010 to 2011 and non-participants who started prospective observation at the same time according to the same protocol. FMD-J is a prospective multicenter study conducted at 22 university hospitals and affiliated clinics in Japan to examine the usefulness of FMD assessment in the management of patients at risk of cardiovascular disease.¹⁰ The FMD-J study consisted of 3 registries according to the stage of cardiovascular risk. The FMD-J B study targeted patients with mild cardiovascular risk and included those aged 20–74 years with controlled hypertension or diabetes who had been receiving antihypertensive or antidiabetic treatment for at least 6 months. Detailed information on the FMD-J study protocol and its participants is publicly available.¹¹ The study protocol was approved by the ethics committee of National Defense Medical College (2010. KAN-13) and was conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from all participants.

Measurement Protocol

The method for simultaneous measurement of the FMD and RHI is described elsewhere.^12–17 Participants fasted from the night before the vascular endothelial examinations and abstained from alcohol, smoking, caffeine, and antioxidant vitamins on the day of the examination. The participants were kept in the supine position in a quiet, dark, air-conditioned room (maintained at a temperature of 23–26℃) throughout the study. A 23-gauge polyethylene catheter was inserted into the left deep antecubital vein to obtain blood samples. Blood pressure was measured in the left arm using a mercury sphygmomanometer with an appropriately sized cuff and recorded to the nearest 2 mmHg. The participant was in the supine position for at least 20 min before tests.

FMD was measured using a semi-automated ultrasound device (UNEXEF18G, UNEX, Nagoya, Japan). The ultrasound probe was placed on the brachial artery, and a cuff was placed on the forearm and inflated for 5 min. After the cuff was released, the vasodilator response was recorded.¹⁸ A peripheral arterial tonometry device (Endo-PAT2000, Itamar Medical, Caesarea, Israel) was placed on the index finger of each hand at the same time. The pulse amplitude detected by this device was then electronically recorded. The extent of reactive hyperemia was calculated as the ratio of the average pulse amplitude of the device signal over a 1-min time interval starting 1.5 min after cuff deflation to the average pulse amplitude of the device signal over a 2.5-min time period before cuff inflation. The RHI was calculated as the ratio of the reactive hyperemia between the 2 hands. The augmentation index (AI) was also measured using the Endo-PAT device. The result was further normalized to a heart rate of 75 beats/min (AI@75 beats/min). Lower AI values reflect better arterial elasticity. All examiners were blinded to the clinical information of patients. In addition, we measured brachial ankle pulse wave velocities, and intima thickness of common carotid arteries.

Clinical Data

We defined hypertension as blood pressure >140/90 mmHg or the use of antihypertensive medication. We diagnosed diabetes as fasting blood glucose >126 mg/dL, or the use of insulin or oral hypoglycemic agents. We defined hyperlipidemia as total cholesterol >220 mg/dL, low-density lipoprotein cholesterol >140 mg/dL, or the use of anti-hyperlipidemic medication. We calculated the estimated glomerular filtration rate using the Modification of Diet in Renal Disease equation modified for the Japanese population.¹⁹ The estimated salt intake was calculated from a spot urine specimen with a previously reported formula.²⁰

Patient Grouping and Follow-up

All patients first underwent a FMD test at enrollment. After that, FMD and EndoPAT RHI were measured simultaneously at 1.5-year intervals (Figure 1). Some patients underwent only FMD, but none underwent only EndoPAT. Two cohort studies were conducted using the 3-year results. Patients who completed the 3 tests were divided into 2 groups, with a threshold of 7% for the FMD study. Patients who completed the 2 tests were divided into 2 groups, wtih a threshold of 1.67 for the EndoPAT study. The patients were followed from the date of their last examination until November 2024. At least 2 authors assessed each patient’s condition from hospital records. The primary endpoints were all-cause mortality and cardiovascular death. The secondary endpoint was hospitalization for heart failure, thromboembolic major cardiovascular events (MACEs), consisting of cardiovascular death, acute myocardial infarction (AMI), and stroke. In addition, we also evaluated angina pectoris (AP) requiring coronary intervention and created composite endpoints of AP/AMI, and AP/AMI plus MACEs.

Figure 1.

Flowchart showing enrollment of patients. FMD, flow-mediated dilation; RHI, reactive hyperemia index.

Statistical Analysis

We evaluated the distribution of continuous clinical characteristics and measurements by examining a histogram and performing the Shapiro-Wilk test. Summary data are presented as the mean±standard deviation (SD) with a 95% confidence interval (CI) for a normal distribution or median (1st, 3rd quartile) for a non-normal distribution. We used the independent samples t-test or Mann-Whitney U test for 2-group comparisons, as appropriate. We performed cross-table analyses using the chi-squared test or Fisher’s exact test, as appropriate. We assessed the correlation between 2 variables using Pearson’s method or Spearman’s method.

We used Kaplan-Meier curves to show the unadjusted event-free rate among groups. We assessed differences using the log-rank test. We performed multivariate Cox regression analysis to exclude the influence of the variance of clinical features and laboratory parameters from the prediction of event-free survival. In the assessment, we adjusted for age, sex, body mass index (BMI; Model 1) or, diabetes, non-smokers, estimated glomerular filtration rate, in addition to Model 1 (Model 2). We also adjusted for hyperlipidemia and hypertension in addition to Model 1 (Model 3). We performed most statistical analyses using SPSS version 22.0 (SPSS Japan, Tokyo, Japan). For all analyses, we considered P<0.05 to be statistically significant.

Results

Patient Classification by Repeated EndoPAT Measurements

The FMD study analyzed data from Japanese adults who underwent all 3 sessions of FMD, and the EndoPAT study analyzed data from patients who underwent all 2 sessions of EndoPAT separately (Figure 1). We binary classified patients by the test results. In the FMD study, 136 patients were classified as those with consistently good endothelial function, as defined by FMD >7% on 3 occasions (n=33) and those without (n=103). In the EndoPAT study, 120 patients were classified as those with consistently abnormal endothelial function, as defined by RHI <1.67 on 2 occasions (n=34) and those without (n=86). The time interval between tests was 1.44±0.38 years.

Baseline Clinical Characteristics

The baseline clinical characteristics and laboratory tests are summarized in Table 1. In the FMD study, the prevalence of hypertension, serum glucose levels, hemoglobin A1c, and BMI was low in the consistently FMD >7% 3-times group. The Framingham score tends to be lower in this group. Persistent and paroxysmal atrial fibrillation (AF) was not different. In the EndoPAT study, the rate of men, persistent AF, alcohol consumption and serum uric acid level were higher in the RHI <1.67 twice group. The percentage of patients who never smoked was lower in the RHI <1.67 twice group. There were no differences in arterial stiffness or carotid atherosclerosis by both FMD and EndoPAT classifications (Table 2).

Table 1.

Demographic and Clinical Characteristics

	FMD >7%, 3 times			EndoPAT RHI <1.67, twice
	− (n=103)	+ (n=33)	P value	− (n=86)	+ (n=34)	P value
Age (years)	67±8	65±9	0.195	66±9	68±7	0.352
Men (%)	60 (58)	13 (39)	0.059	39 (45)	24 (71)	0.013*
BMI	24±3	22±3	<0.001*	23±3	24±3	0.424
Diabetes	12 (12)	1 (3)	0.143	8 (9)	2 (6)	0.541
Hypertension	90 (87)	23 (70)	0.018*	70 (81)	29 (85)	0.613
Hyperlipidemia	47 (46)	17 (52)	0.556	41 (48)	15 (44)	0.725
Persistent atrial fibrillation	9 (9)	3 (9)	1.000	5 (6)	5 (15)	0.144
Paroxysmal atrial fibrillation	22 (21)	12 (36)	0.129	27 (31)	5 (15)	0.070
Habit of alcohol drink	53 (51)	19 (58)	0.540	30 (35)	18 (53)	0.069
Habit of exercise	65 (63)	21 (64)	0.694	49 (57)	24 (71)	0.169
Estimated dietary salt intake (g/day)	10.5 (8.7, 12.3)	10.5 (8.8, 11.9)	0.700	10.5 (9.4, 12.0)	11.1 (8.5, 13.3)	0.521
Alcohol intake (g/day)	0 (0, 25.1)	0 (0, 5.5)	0.247	0 (0, 6.3)	0 (0, 44.0)	0.014*
Smoking status			0.735			0.013*
Current smoking	8 (8)	4 (12)		7 (8)	5 (15)
Past smoking	31 (30)	9 (27)		21 (24)	16 (47)
Non-smoking	64 (62)	20 (61)		58 (67)	13 (38)
Medications
β-blocker	30 (29)	10 (30)	0.897	21 (24)	15 (44)	0.034*
ARB/ACE inhibitor	42 (41)	11 (33)	0.445	31 (36)	14 (41)	0.601
Calcium channel blocker	71 (69)	15 (45)	0.015*	54 (63)	21 (62)	0.917
Diuretics for hypertension	6 (6)	5 (15)	0.087	7 (8)	3 (9)	0.903
Aldosterone antagonists	1 (1)	0 (0)	0.570	1 (1)	0 (0)	0.528
Statin	46 (45)	15 (45)	0.936	39 (45)	14 (41)	0.678
Fibrate	1 (1)	0 (0)	0.611	0 (0)	0 (0)	N/A
Insulin	0 (0)	0 (0)	N/A	0 (0)	0 (0)	N/A
Aspirin	42 (41)	9 (27)	0.163	28 (33)	17 (50)	0.075
Clopidogrel/ticlopidine	3 (3)	0 (0)	0.321	1 (1)	0 (0)	0.528
Laboratory data
Total cholesterol (mg/dL)	194±34	204±37	0.180	200±35	188±27	0.082
LDL cholesterol (mg/dL)	105±32	115±36	0.142	111±32	98±29	0.053
HDL cholesterol (mg/dL)	59±16	62±15	0.330	60±16	60±18	0.906
Triglyceride (mg/dL)	130 (99, 179)	113 (74, 189)	0.148	129 (95, 179)	131 (97, 186)	0.718
Small-dense LDL cholesterol (mg/dL)	36±16	34±17	0.478	35±16	34±14	0.881
MDA-LDL cholesterol (U/L)	107±39	101±41	0.462	105±39	104±35	0.919
Glucose (mg/dL)	116±31	101±15	0.010*	109±27	119±33	0.083
HbA1c (%)	6.0±0.8	5.7±0.6	0.033*	5.9±0.6	6.1±1.0	0.233
Uric acid (mg/dL)	5.6±1.2	5.2±1.3	0.064	5.4±1.3	6.0±1.2	0.030*
High-sensitive CRP (ng/dL)	414 (212, 929)	317 (131, 963)	0.288	321 (191, 800)	440 (162, 1,050)	0.771
eGFR (mL/min/1.73 m2)	68.9±15.6	71.1±12.4	0.460	69.8±16.2	69.1±11.2	0.823
Framingham score	7.5±3.6	6.1±2.9	0.061	7.1±3.4	7.6±3.3	0.567

Data are presented as n (%), or mean±SD with a 95% confidence interval (CI) for a normal distribution or median (1st, 3rd quartile) for a non-normal distribution. *P<0.05 is considered statistically significant. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; FMD, flow-mediated dilation; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MDA, malondialdehyde; RHI, reactive hyperemia index; TG, triglycerides.

Table 2.

Results of Vascular Function Tests

	FMD >7%, 3 times			EndoPAT RHI <1.67, twice
	− (n=103)	+ (n=33)	P value	− (n=86)	+ (n=34)	P value
Vascular examination
Systolic BP, P1 (mmHg)	131±21	130±18	0.787	132±17	127±23	0.232
Diastolic BP, P1 (mmHg)	78±11	79±10	0.716	79±10	77±12	0.328
Heart rate, P1 (beats/min)	67±11	69±10	0.375	67±10	70±11	0.137
baPWV, left (m/s)	1,746±374	1,662±309	0.245	1,691±288	1,810±502	0.108
baPWV, right (m/s)	1,746±384	1,640±303	0.151	1,683±294	1,797±516	0.128
Carotid artery max-IMT (mm)	1.1±0.2	1.1±0.2	0.381	1.1±0.2	1.2±0.3	0.166
FMD
FMD, P1 (%)	5.91±2.54	9.65±2.40	<0.001*	7.05±3.07	6.13±4.01	0.175
FMD, P2 (%)	6.58±2.25	9.86±2.50	<0.001*	7.73±3.04	6.64±1.87	0.053
FMD, P3 (%)	7.63±3.58	9.52±1.80	0.003*	7.93±2.66	8.54±4.80	0.378
Diameter at baseline, P1 (mm)	4.20±0.70	3.68±0.60	<0.001*	4.01±0.67	4.24±0.77	0.097
Diameter at baseline, P2 (mm)	4.19±0.69	3.72±0.59	<0.001*	4.02±0.65	4.28±0.74	0.061
Diameter at baseline, P3 (mm)	4.14±0.71	3.58±0.55	<0.001*	3.95±0.70	4.20±0.67	0.072
Diameter at maximum dilation, P1 (mm)	4.45±0.72	4.03±0.62	0.003*	4.28±0.67	4.50±0.82	0.129
Diameter at maximum dilation, P2 (mm)	4.46±0.72	4.08±0.61	0.005*	4.32±0.67	4.56±0.77	0.095
Diameter at maximum dilation, P3 (mm)	4.46±0.74	3.92±0.57	<0.001*	4.25±0.71	4.55±0.70	0.040*
Increase of blood flow velocity, P1 (fold)	2.84±1.83	2.68±1.72	0.655	3.00±2.08	2.45±1.34	0.156
Increase of blood flow velocity, P2 (fold)	3.05±1.97	3.07±1.80	0.964	3.24±1.96	2.99±2.10	0.540
Increase of blood flow velocity, P3 (fold)	2.99±1.48	2.94±1.50	0.853	3.17±1.59	2.75±1.46	0.194
EndoPAT
RHI, P2	1.80±0.51	1.94±0.50	0.175	2.02±0.48	1.40±0.20	<0.001*
RHI, P3	1.82±0.49	1.85±0.48	0.769	1.96±0.46	1.45±0.15	<0.001*
AI, P2	31 (18, 47)	36 (26, 51)	0.390	36 (24, 54)	28 (11, 36)	0.014*
AI, P3	29 (20, 43)	36 (25, 50)	0.221	33 (23, 45)	25 (10, 42)	0.082
AI@75 beats/min, P2	24 (13, 39)	28 (16, 42)	0.317	30 (17, 42)	21 (8, 32)	0.012*
AI@75 beats/min, P3	21 (13, 35)	31 (20, 40)	0.063	29 (15, 37)	21 (9, 28)	0.054

Data are presented as mean±SD, or median (1st, 3rd quartile). EndoPAT measurements were taken at 1.5 years (Period 2) and 3 years (Period 3) after enrollment. Please also see Figure 1. max-IMT represents the highest value in the bilateral common carotid arteries. *P<0.05 is considered statistically significant. AI, augmentation index; AI@75 beats/min, AI normalized to a heart rate of 75 beats/min; baPWV, brachial-ankle pulse wave velocity; BP, blood pressure; max-IMT, maximum intima media thickness; RHI, reactive hyperemia index.

Kaplan-Meier Curves for Clinical Endpoints

In the FMD study, over 9.5±3.0 years, the primary endpoint of all-cause death was reached in 9 (7%) patients, and included 5 (4%) cardiovascular deaths. The events are detailed in Table 3. We confirmed 17 (13%) MACEs, and 10 (7%) AP/AMI. A composite endpoint of AP/AMI plus MACEs was reached in 24 (18%) patients. Heart failure was observed in 11 (8%) patients. The Kaplan-Meier curves showed the FMD >7% 3-times group had a tendency for low MACEs and AP/AMI, and significantly lower rates of AP/AMI plus MACEs (Figure 2).

Table 3.

Clinical Events

	FMD >7%, 3 times		EndoPAT <1.67, twice
	− (n=103)	+ (n=33)	− (n=86)	+ (n=34)
All-cause death	9 (9)	0 (0)	2 (2)	7 (21)
Cardiovascular death	5 (5)	0 (0)	2 (2)	3 (9)
Sudden death	3 (3)	0 (0)	1 (1)	2 (6)
Heart failure	2 (2)	0 (0)	1 (1)	1 (3)
Other cause of death	3 (3)	0 (0)	0 (0)	4 (12)
Cancer	2 (2)	0 (0)	0 (0)	2 (6)
Pneumonia	1 (1)	0 (0)	0 (0)	1 (3)
Senility	1 (1)	0 (0)	0 (0)	1 (3)
MACEs	15 (15)	2 (6)	12 (14)	5 (15)
Stroke	8 (8)	2 (6)	8 (9)	2 (6)
AMI	2 (2)	0 (0)	2 (2)	0 (0)
Other cardiac events
AP/AMI	10 (10)†	0 (0)	5 (6)†	3 (9)
AP/AMI plus MACEs	22 (21)	2 (6)	14 (16)	8 (24)
Heart failure	9 (9)	2 (6)	3 (3)	7 (21)

Data are presented as n (%). ^†One patient who underwent coronary intervention died due to heart failure. AMI, acute myocardial infarction; AP, angina pectoris; MACEs, major adverse cardiovascular events.

Figure 2.

Kaplan-Meier curves of repeated FMD measurements. The survival curves of all-cause death (A), cardiovascular death (B), MACEs (C), heart failure (D), AP/AMI (E), and AP/AMI plus MACEs (F) are shown. AMI, acute myocardial infarction; AP, angina pectoris; FMD, flow-mediated dilation; MACEs, major adverse cardiovascular events.

In the EndoPAT study, over 10.4±1.7 years, the primary endpoint of all-cause death was reached in 9 (8%) patients, and included 5 (4%) cardiovascular deaths. We confirmed 17 (14%) MACEs, and 8 (7%) AP/AMI. A composite endpoint of AP/AMI plus MACEs was reached in 22 (18%) patients. Heart failure was observed in 10 (8%) patients. The Kaplan-Meier curves showed the RHI <1.67 twice group had significantly lower rates of all-cause death and heart failure (Figure 3), but no predictive effect was found for MACEs and/or AP/AMI.

Figure 3.

Kaplan-Meier curves of repeated EndoPAT RHI measurement. The survival curves of all-cause death (A), cardiovascular death (B), MACEs (C), heart failure (D), AP/AMI (E), and AP/AMI plus MACEs (F) are shown. AMI, acute myocardial infarction; AP, angina pectoris; MACEs, major adverse cardiovascular events; RHI, reactive hyperemia index.

Figure 4 shows survival curves according to the frequency of FMD >7% or RHI <1.67, demonstrating that consistent test results correlate with clinical outcome.

Figure 4.

Kaplan-Meier curves grouped by frequency of flow-mediated dilation (FMD) >7% or EndoPAT reactive hyperemia index (RHI) <1.67. The survival curves of all-cause death (A), heart failure for RHI <1.67 (B), and AP/AMI plus MACEs (C) for FMD >7% are shown. The events rate of RHI <1.67 twice (n=34), once (n=34), and none (n=52) were 7 (21%), 0 (0%), and 2 (4%) for all-cause death, and 7 (21%), 1 (3%), and 2 (4%) for heart failure, respectively. The events rate of FMD >7% 3 times (n=33), twice (n=43), once (n=32), and none (n=28) were 2 (6%), 9 (21%), 7 (22%), and 6 (21%) for AP/AMI plus MACEs, respectively. *P<0.05. ^¶P<0.01. AMI, acute myocardial infarction; AP, angina pectoris.

Cox Regression Analysis

Analysis of a single FMD performed in phase 2 showed predictive power of FMD with a borderline of 7% for AP/AMI plus MACEs, but the results were not consistent in the 3 periods (Table 4). Multivariate analysis using all 3 FMD measurements demonstrated a predictively of keeing FMD higher than 7% for lower incidence of AP/AMI plus MACEs (P=0.048; Table 4A, Model 3).

Table 4.

Cox Regression Analyses of Repeated FMD and EndoPAT RHI

(A)	MACEs		AP/AMI		AP/AMI plus MACEs
Category	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
FMD >7%, 3 times
Unadjusted HR	0.367 (0.084–1.606)	0.183	0.031 (0.000–12.011)	0.253	0.239 (0.052–1.111)	0.068
Adjusted HR, Model 1	0.356 (0.076–1.674)	0.191	N/A		0.250 (0.059–1.064)	0.061
Adjusted HR, Model 2	0.332 (0.069–1.598)	0.169	N/A		0.227 (0.050–1.019)	0.053
Adjusted HR, Model 3	0.468 (0.096–2.273)	0.346	N/A		0.216 (0.047–0.985)	0.048*
FMD >7%, P1
Unadjusted HR	0.793 (0.319–1.974)	0.619	0.709 (0.199–2.526)	0.596	0.847 (0.372–1.933)	0.694
Adjusted HR, Model 1	0.828 (0.322–2.130)	0.695	0.748 (0.203–2.758)	0.663	0.856 (0.363–2.018)	0.722
Adjusted HR, Model 2	0.793 (0.306–2.054)	0.633	0.853 (0.230–3.161)	0.812	0.874 (0.361–2.119)	0.766
Adjusted HR, Model 3	1.135 (0.425–3.028)	0.800	0.648 (0.165–2.544)	0.534	1.135 (0.425–3.028)	0.800
FMD >7%, P2
Unadjusted HR	0.389 (0.148–1.025)	0.056	0.834 (0.240–2.895)	0.775	0.376 (0.155–0.915)	0.031*
Adjusted HR, Model 1	0.415 (0.147–1.173)	0.097	0.941 (0.246–3.595)	0.929	0.364 (0.140–0.951)	0.039*
Adjusted HR, Model 2	0.411 (0.143–1.184)	0.100	0.911 (0.240–3.463)	0.891	0.356 (0.133–0.950)	0.039*
Adjusted HR, Model 3	0.542 (0.183–1.603)	0.268	0.810 (0.215–3.048)	0.756	0.396 (0.153–1.027)	0.057
FMD >7%, P3
Unadjusted HR	0.646 (0.262–1.591)	0.342	0.848 (0.239–3.007)	0.799	0.634 (0.279–1.437)	0.275
Adjusted HR, Model 1	0.564 (0.217–1.467)	0.240	0.838 (0.230–3.052)	0.788	0.529 (0.220–1.268)	0.153
Adjusted HR, Model 2	0.548 (0.208–1.447)	0.225	0.867 (0.236–3.187)	0.830	0.525 (0.211–1.309)	0.167
Adjusted HR, Model 3	0.751 (0.278–2.031)	0.573	0.677 (0.179–2.563)	0.566	0.765 (0.322–1.815)	0.543
Scale	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
FMD, P1
Unadjusted HR	0.925 (0.810–1.055)	0.245	0.907 (0.796–1.033)	0.142	0.905 (0.783–1.045)	0.174
Adjusted HR, Model 1	0.908 (0.776–1.062)	0.227	0.892 (0.769–1.035)	0.131	0.886 (0.754–1.042)	0.144
Adjusted HR, Model 2	0.906 (0.776–1.058)	0.213	0.891 (0.765–1.038)	0.138	0.888 (0.754–1.046)	0.155
Adjusted HR, Model 3	0.932 (0.786–1.104)	0.414	0.854 (0.721–1.012)	0.068	0.885 (0.796–0.983)	0.022*
FMD, P2
Unadjusted HR	0.888 (0.745–1.059)	0.188	0.945 (0.743–1.201)	0.643	0.854 (0.725–1.005)	0.057
Adjusted HR, Model 1	0.913 (0.750–1.113)	0.369	0.974 (0.752–1.260)	0.839	0.856 (0.712–1.030)	0.100
Adjusted HR, Model 2	0.917 (0.750–1.122)	0.401	0.972 (0.758–1.246)	0.821	0.853 (0.704–1.034)	0.106
Adjusted HR, Model 3	0.984 (0.802–1.208)	0.880	0.934 (0.718–1.214)	0.609	0.901 (0.751–1.082)	0.265
FMD, P3
Unadjusted HR	0.948 (0.812–1.107)	0.499	0.947 (0.764–1.173)	0.616	0.951 (0.826–1.094)	0.479
Adjusted HR, Model 1	0.949 (0.811–1.112)	0.519	0.947 (0.766–1.170)	0.613	0.946 (0.817–1.095)	0.457
Adjusted HR, Model 2	0.955 (0.822–1.110)	0.548	0.946 (0.754–1.187)	0.632	0.949 (0.816–1.103)	0.492
Adjusted HR, Model 3	1.010 (0.860–1.187)	0.902	0.911 (0.723–1.148)	0.431	0.986 (0.857–1.133)	0.839
(B)	All-cause death		Heart failure
Category	HR (95% CI)	P value	HR (95% CI)	P value
EndoPAT <1.67, twice
Unadjusted HR	9.464 (1.965–45.585)	0.005*	6.076 (1.440–25.638)	0.014*
Adjusted HR, Model 1	6.848 (1.345–34.871)	0.021*	6.676 (1.725–25.833)	0.006*
Adjusted HR, Model 2	10.794 (1.520–76.629)	0.017*	5.356 (1.301–22.052)	0.020*
Adjusted HR, Model 3	6.119 (1.237–30.276)	0.026*	5.311 (1.218–23.149)	0.026*
EndoPAT RHI <1.67, P2
Unadjusted HR	5.180 (1.076–24.938)	0.040*	6.218 (1.320–29.286)	0.021*
Adjusted HR, Model 1	4.273 (0.853–21.393)	0.077	5.615 (1.155–27.290)	0.032*
Adjusted HR, Model 2	3.972 (0.709–22.255)	0.117	4.973 (0.978–25.295)	0.053
Adjusted HR, Model 3	3.577 (0.732–17.482)	0.115	6.053 (1.158–31.648)	0.033*
EndoPAT RHI <1.67, P3
Unadjusted HR	5.029 (1.042–24.263)	0.044*	3.428 (0.885–13.275)	0.075
Adjusted HR, Model 1	3.783 (0.727–19.692)	0.114	2.601 (0.624–10.849)	0.190
Adjusted HR, Model 2	4.594 (0.787–26.815)	0.090	2.674 (0.608–11.755)	0.193
Adjusted HR, Model 3	3.618 (0.707–18.530)	0.123	2.698 (0.631–11.533)	0.181
Scale	HR (95% CI)	P value	HR (95% CI)	P value
EndoPAT RHI, P2
Unadjusted HR	0.453 (0.096–2.128)	0.316	0.339 (0.074–1.543)	0.162
Adjusted HR, Model 1	0.613 (0.131–2.872)	0.535	0.408 (0.087–1.920)	0.257
Adjusted HR, Model 2	0.807 (0.163–4.009)	0.794	0.478 (0.096–2.371)	0.366
Adjusted HR, Model 3	0.767 (0.172–3.424)	0.728	0.438 (0.095–2.026)	0.291
EndoPAT RHI, P3
Unadjusted HR	0.887 (0.191–4.127)	0.878	0.419 (0.083–2.105)	0.291
Adjusted HR, Model 1	1.117 (0.215–5.793)	0.895	0.471 (0.084–2.650)	0.393
Adjusted HR, Model 2	1.086 (0.217–5.422)	0.920	0.591 (0.117–2.994)	0.525
Adjusted HR, Model 3	1.448 (0.247–8.497)	0.681	0.453 (0.077–2.677)	0.383

*P<0.05 is considered statistically significant. Data for FMD (A) and EndoPAT (B) represent univariate and multivariate analyses subsequently adjusted for age, sex, BMI (Model 1) or, diabetes, non-smokers eGFR, in addition to Model 1 (Model 2), or hypertension, hyperlipidemia, in addition to Model 1 (Model 3). HRs for scales are expressed as a unit increase. CI, confidence interval; HR, hazard ratio; N/A, analysis was not available. Other abbreviations as in Table 1.

The EndoPAT study showed significant predictive ability for all-cause death and heart failure in every single measurement, but analysis of 2 measurements improved the predictive performance. Patients with RHI <1.67 twice had a higher mortality (HR 10.794; 95% CI 1.520–76.629; P=0.017) and incidence of heart failure (HR 5.356; 95% CI 1.301–22.052; P=0.020) after adjusting for age, sex, renal function, diabetes, non-smokers, and body mass index (Table 4B, Model 2).

Thus, even after adjusting for background and laboratory findings related to FMD and EnodPAT, the prognostic value of endothelial function testing remained significant. In contrast to the categorical data analyses, the actual scales of FMD and RHI were less effective in predicting events. For objectivity, additional analyses including all patients who had at least 1 FMD or EndoPAT test are shown in Supplementary Table 1. The patients with missing values were classified as controls regardless of their results. In addition, all events occurring after enrollment were included. However, no differences in outcomes were found (Supplementary Figures 1,2; Supplementary Tables 2,3).

Discussion

In this study, patients with mild cardiovascular risk were followed up with endothelial function tests. Patients with an EndoPAT RHI of <1.67 on 2 consecutive tests had significantly higher rates of all-cause mortality and heart failure hospitalization, while those with FMD over 7% on 3 tests had a lower incident rate of MACEs or coronary events. We found that single FMD tests are variable in their predictions. A prognostic value in mortality, heart failure, by only 1 measurement with an EndoPAT RHI <1.67 was effective, but repeated measurements improved the predictive accuracy.

Although several studies have evaluated the reproducibility of endothelial function tests, few have aimed for monitoring.^21–24 Their testing intervals were short and not predetermined. In contrast, the present study was designed to evaluate changes in endothelial function and patient outcomes. The examination schedule was predetermined every 1.5 years assuming regular outpatient monitoring. A previous study suggested the impact of persistent impairment of FMD on clinical outcome.²⁵ However, to our knowledge, this study is the first to demonstrate the validity of repeated measurements with respect to EndoPAT.

It is difficult to determine whether the endothelial function is good or bad by whether it falls below 1 borderline in a single measurement. For this reason, the Japanese Society for Vascular Failure and the Japanese Society of Hypertension have established intermediate ranges for FMD or EndoPAT RHI.^8,9 The cut-off values for distinguishing abnormal and normal endothelial function are 4%, and 7% for FMD, and 1.67 and 2.00 for EndoPAT RHI, respectively. The range between these 2 boundary values is defined as moderate risk. Some studies also use the logarithmic value of RHI because of its skewed distribution.^5,26 In the present study, the crude values of RHI were less effective for prediction of cardiovascular events.

Analysis of repeated measurements is another way to improve risk stratification by distinguishing between patients with constantly poor endothelial function and those likely to recover. In the present study, a single RHI measurement diagnosed abnormal endothelial function in approximately 40–50% of all patients. However, analysis of 2 examinations allowed us to narrow down the proportion of patients with severely abnormal vascular endothelial function to approximately 30%. The same is true for FMD. As only a small number of patients in our population had an FMD consistently below 4%, we used repeated FMD measurements to find patients who maintained good endothelial function. The percentage of FMD over 7% was approximately 50% in every period of test, but it decreased to 24% when 3 tests were analysed together.

EndoPAT mainly evaluates endothelium-dependent hyperpolarization (EDH).^27–29 EDH plays an important role in relaxation of resistance arteries.^30–32 Patients with consistently low RHI in the present study were predominantly men, current or former smokers, had significantly higher alcohol intake, more likely to be taking β-blockers, and had hyperuricemia. Low-density lipoprotein cholesterol tended to be lower in the low RHI group, although the rate of statin use remained the same. It is well known that EndoPAT RHI is associated with smoking²⁶ and vasospastic angina.³³ Recently, it has been reported that microvascular dysfunction assessed by EndoPAT RHI is associated with heart failure.^34–36 However, the predictivity of EndoPAT for thromboembolic MACEs remains controversial.^37–39

In contrast, FMD was more associated with coronary events. The FMD-J A study, which investigated patients with more severe atherosclerosis, found that an FMD of 7% was appropriate for prediction of cardiovascular events and could be a target for secondary prevention.⁴⁰ This is consistent with our results aimed at primary prevention. FMD was associated with hypertension and BMI as previously reported.⁷ FMD indicates endothelial function to secrete nitric oxide, but can be influenced by the baseline diameter of the artery.⁴¹ Therefore, FMD reflects more subtle atherosclerotic changes leading to the development of coronary artery disease.

The effectiveness of FMD and EndoPAT in predicting cardiovascular prognosis has already been recognized, and the next step is to find appropriate ways to use endothelial function testing for monitoring patient care. In our opinion, some patients may have a temporary decrease in endothelial function, and multiple results, rather than a single measurement, should be considered to determine the subsequent prognosis. If a consistent decline is observed with periodic measurements, some action should be taken. Repeated measurements can also provide accurate information on what can be done to improve endothelial function and which diseases can be predicted.

Study Limitations

There are several limitations to this study. First, the number of patients was small. Second, the cuff was placed on the forearm to allow simultaneous measurement of FMD and EndoPAT. Recent guidelines recommend the upper arm as the recommended cuff position for EndoPAT RHI.^42,43 However, the combination of vascular tests is becoming common.^26,44 Last, we did not consider gender differences in the test thresholds.

Conclusions

A persistently decreased RHI indicates a poor prognosis and a high risk of developing heart failure, whereas maintaining a high FMD value may help to avoid coronary events. Further studies are needed to establish the appropriate use of endothelial function tests in clinical practice.

Sources of Funding

This research was supported in part by a Grant-in-aid for Scientific Research from the Ministry of Education, Science, and Culture of Japan (24K11204).

Disclosures

The authors report no relationships that could be construed as a conflict of interest.

Author Contributions

All authors contributed to the conception and design of the study, and carried out the collection, analysis, and interpretation of data. N.M. wrote the first draft of the manuscript. B.T. was the supervisor of the present study and participated in drafting, revising the manuscript, and approving the final version. All authors read and approved the final version of the manuscript and declared that the contents have not been published elsewhere.

IRB Information

The Ethics Committee of Tokyo Medical University (the core center of the FMD-J study; No. 2456), and the Ethics Committee of National Defense Medical College (2010. KAN-13).

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circrep.CR-25-0042

References

• 1.   Deanfield JE, Halcox JP, Rabelink TJ. Endothelial function and dysfunction: Testing and clinical relevance. Circulation 2007; 115: 1285–1295.
• 2.   Celermajer DS, Sorensen KE, Gooch VM, Spiegelhalter DJ, Miller OI, Sullivan ID, et al. Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis. Lancet 1992; 340: 1111–1115.
• 3.   Takase B, Uehata A, Akima T, Nagai T, Nishioka T, Hamabe A, et al. Endothelium-dependent flow-mediated vasodilation in coronary and brachial arteries in suspected coronary artery disease. Am J Cardiol 1998; 82: 1535–1539, A7–A8.
• 4.   Matsuzawa Y, Sugiyama S, Sugamura K, Sumida H, Kurokawa H, Fujisue K, et al. Successful diet and exercise therapy as evaluated on self-assessment score significantly improves endothelial function in metabolic syndrome patients. Circ J 2013; 77: 2807–2815.
• 5.   Matsuzawa Y, Kwon TG, Lennon RJ, Lerman LO, Lerman A. Prognostic value of flow-mediated vasodilation in brachial artery and fingertip artery for cardiovascular events: A systematic review and meta-analysis. J Am Heart Assoc 2015; 4: e002270, doi:10.1161/JAHA.115.002270.
• 6.   Xu Y, Arora RC, Hiebert BM, Lerner B, Szwajcer A, McDonald K, et al. Non-invasive endothelial function testing and the risk of adverse outcomes: A systematic review and meta-analysis. Eur Heart J Cardiovasc Imaging 2014; 15: 736–746.
• 7.   Hamburg NM, Palmisano J, Larson MG, Sullivan LM, Lehman BT, Vasan RS, et al. Relation of brachial and digital measures of vascular function in the community: The Framingham Heart Study. Hypertension 2011; 57: 390–396.
• 8.   Tanaka A, Tomiyama H, Maruhashi T, Matsuzawa Y, Miyoshi T, Kabutoya T, et al.; Physiological Diagnosis Criteria for Vascular Failure Committee. Official announcement of physiological diagnostic criteria for vascular failure from the Japanese Society for Vascular Failure. Vasc Fail 2018; 2: 59–60.
• 9.   Tanaka A, Tomiyama H, Maruhashi T, Matsuzawa Y, Miyoshi T, Kabutoya T, et al.; Physiological Diagnosis Criteria for Vascular Failure Committee. Physiological diagnostic criteria for vascular failure. Hypertension 2018; 72: 1060–1071.
• 10.   Tomiyama H, Kohro T, Higashi Y, Takase B, Suzuki T, Ishizu T, et al. A multicenter study design to assess the clinical usefulness of semi-automatic measurement of flow-mediated vasodilatation of the brachial artery. Int Heart J 2012; 53: 170–175.
• 11.   Tomiyama H, Kohro T, Higashi Y, Takase B, Suzuki T, Ishizu T, et al. Reliability of measurement of endothelial function across multiple institutions and establishment of reference values in Japanese. Atherosclerosis 2015; 242: 433–442.
• 12.   Tomiyama H, Yoshida M, Higashi Y, Takase B, Furumoto T, Kario K, et al.; sub-group study of FMD-J. Autonomic nervous activation triggered during induction of reactive hyperemia exerts a greater influence on the measured reactive hyperemia index by peripheral arterial tonometry than on flow-mediated vasodilatation of the brachial artery in patients with hypertension. Hypertens Res 2014; 37: 914–918.
• 13.   Takase B, Higashimura Y, Hashimoto K. Disparity between EndoPAT measurement and brachial artery flow-mediated vasodilatation in hypertensive patients. Vasc Fail 2018; 2: 61–65.
• 14.   Fukumoto K, Takemoto Y, Norioka N, Takahashi K, Namikawa H, Tochino Y, et al. Predictors of the effects of smoking cessation on the endothelial function of conduit and digital vessels. Hypertens Res 2021; 44: 63–70.
• 15.   Tajima E, Sakuma M, Tokoi S, Matsumoto H, Saito F, Watanabe R, et al. The comparison of endothelial function between conduit artery and microvasculature in patients with coronary artery disease. Cardiol J 2020; 27: 38–46.
• 16.   Bonetti PO, Pumper GM, Higano ST, Holmes DR Jr, Kuvin JT, Lerman A. Noninvasive identification of patients with early coronary atherosclerosis by assessment of digital reactive hyperemia. J Am Coll Cardiol 2004; 44: 2137–2141.
• 17.   Rubinshtein R, Kuvin JT, Soffler M, Lennon RJ, Lavi S, Nelson RE, et al. Assessment of endothelial function by non-invasive peripheral arterial tonometry predicts late cardiovascular adverse events. Eur Heart J 2010; 31: 1142–1148.
• 18.   Masaki N, Adachi T, Tomiyama H, Kohro T, Suzuki T, Ishizu T, et al. Reduced reactive hyperemia of the brachial artery in diabetic patients assessed by repeated measurements: The FMD-J B study. Physiol Rep 2023; 11: e15786, doi:10.14814/phy2.15786.
• 19.   Imai E, Horio M, Nitta K, Yamagata K, Iseki K, Hara S, et al. Estimation of glomerular filtration rate by the MDRD study equitation modified for Japanese patients with chronic kidney disease. Clin Exp Nephrol 2007; 11: 41–50.
• 20.   Tanaka T, Okamura T, Miura K, Kadowaki T, Ueshima H, Nakagawa H, et al. A simple method to estimate populational 24-h urinary sodium and potassium excretion using a casual urine specimen. J Hum Hypertens 2002; 16: 97–103.
• 21.   Hansen AS, Butt JH, Holm-Yildiz S, Karlsson W, Kruuse C. Validation of repeated endothelial function measurements using EndoPAT in stroke. Front Neurol 2017; 8: 178, doi:10.3389/fneur.2017.00178.
• 22.   Moerland M, Kales AJ, Schrier L, van Dongen MG, Bradnock D, Burggraaf J. Evaluation of the EndoPAT as a tool to assess endothelial function. Int J Vasc Med 2012; 2012: 904141, doi:10.1155/2012/904141.
• 23.   Degnan AJ, Shah N, Carty DM, Petrie JR, Delles C, Schneider MP. Repeatability of peripheral artery tonometry in female subjects. Int Sch Res Notices 2013; 383624, doi:10.1155/2013/383624.
• 24.   Sauder KA, West SG, McCrea CE, Campbell JM, Jenkins AL, Jenkins DJ, et al. Test-retest reliability of peripheral arterial tonometry in the metabolic syndrome. Diab Vasc Dis Res 2014; 11: 201–207.
• 25.   Kitta Y, Obata JE, Nakamura T, Hirano M, Kodama Y, Fujioka D, et al. Persistent impairment of endothelial vasomotor function has a negative impact on outcome in patients with coronary artery disease. J Am Coll Cardiol 2009; 53: 323–330, doi:10.1016/j.jacc.2008.08.074.
• 26.   Hahad O, Arnold N, Prochaska JH, Panova-Noeva M, Schulz A, Lackner KJ, et al. Cigarette smoking is related to endothelial dysfunction of resistance, but not conduit arteries in the general population: Results from the Gutenberg Health Study. Front Cardiovasc Med 2021; 8: 674622, doi:10.3389/fcvm.2021.674622.
• 27.   Félétou M. Endothelium-dependent hyperpolarization and endothelial dysfunction. J Cardiovasc Pharmacol 2016; 67: 373–387.
• 28.   Shimokawa H, Sunamura S, Satoh K. RhoA/Rho-Kinase in the cardiovascular system. Circ Res 2016; 118: 352–366.
• 29.   Nagao T, Illiano S, Vanhoutte PM. Heterogeneous distribution of endothelium-dependent relaxations resistant to NG-nitro-L-arginine in rats. Am J Physiol 1992; 263: H1090-4.
• 30.   Hwa JJ, Ghibaudi L, Williams P, Chatterjee M. Comparison of acetylcholine-dependent relaxation in large and small arteries of rat mesenteric vascular bed. Am J Physiol 1994; 266: H952–H958.
• 31.   Shimokawa H, Yasutake H, Fujii K, Owada MK, Nakaike R, Fukumoto Y, et al. The importance of the hyperpolarizing mechanism increases as the vessel size decreases in endothelium-dependent relaxations in rat mesenteric circulation. J Cardiovasc Pharmacol 1996; 28: 703–711.
• 32.   Goto K, Ohtsubo T, Kitazono T. Endothelium-dependent hyperpolarization (EDH) in hypertension: The role of endothelial ion channels. Int J Mol Sci 2018; 19: 315, doi:10.3390/ijms19010315.
• 33.   Ohura-Kajitani S, Shiroto T, Godo S, Ikumi Y, Ito A, Tanaka S, et al. Marked impairment of endothelium-dependent digital vasodilatations in patients with microvascular angina: Evidence for systemic small artery disease. Arterioscler Thromb Vasc Biol 2020; 40: 1400–1412.
• 34.   Akiyama E, Sugiyama S, Matsuzawa Y, Konishi M, Suzuki H, Nozaki T, et al. Incremental prognostic significance of peripheral endothelial dysfunction in patients with heart failure with normal left ventricular ejection fraction. J Am Coll Cardiol 2012; 60: 1778–1786.
• 35.   Rajai N, Toya T, Sara JD, Rajotia A, Lopez-Jimenez F, Lerman LO, et al. Prognostic value of peripheral endothelial function on major adverse cardiovascular events above traditional risk factors. Eur J Prev Cardiol 2023; 3: 1781–1788.
• 36.   Fujisue K, Sugiyama S, Matsuzawa Y, Akiyama E, Sugamura K, Matsubara J, et al. Prognostic significance of peripheral microvascular endothelial dysfunction in heart failure with reduced left ventricular ejection fraction. Circ J 2015; 79: 2623–2631.
• 37.   Dirjayanto VJ, Pompei G, Rubino F, Biscaglia S, Campo G, Mihailidou AS, et al. Non-invasive vascular measures as prognostic predictors for older patients with non-ST elevation acute coronary syndrome. Coron Artery Dis 2024; 35: 368–381.
• 38.   Shechter M, Natanzon SS, Lerman A, Cohn H, Prasad M, Goitein O, et al. Endothelial function predicts 5-year adverse outcome in patients hospitalized in an emergency department chest pain unit. J Cardiovasc Med (Hagerstown) 2023; 24: 729–736.
• 39.   Cheng X, He Y, Fan H, Liu T, Pan W, Wang K, et al. Endothelial function as predictor in patients with coronary syndrome treated by percutaneous coronary intervention. Biosci Rep 2018; 38: BSR20180732, doi:10.1042/BSR20180732.
• 40.   Maruhashi T, Soga J, Fujimura N, Idei N, Mikami S, Iwamoto Y, et al. Endothelial dysfunction, increased arterial stiffness, and cardiovascular risk prediction in patients with coronary artery disease: FMD-J (Flow-Mediated Dilation Japan) Study A. J Am Heart Assoc 2018; 7: e008588, doi:10.1161/JAHA.118.008588.
• 41.   Maruhashi T, Soga J, Fujimura N, Idei N, Mikami S, Iwamoto Y, et al. Brachial artery diameter as a marker for cardiovascular risk assessment: FMD-J study. Atherosclerosis 2018; 268: 92–98, doi:10.1016/j.atherosclerosis.2017.11.022.
• 42.   Thijssen DH, Black MA, Pyke KE, Padilla J, Atkinson G, Harris RA, et al. Assessment of flow-mediated dilation in humans: A methodological and physiological guideline. Am J Physiol Heart Circ Physiol 2011; 300: H2–H12.
• 43.   Thijssen DHJ, Bruno RM, van Mil ACCM, Holder SM, Faita F, Greyling A, et al. Expert consensus and evidence-based recommendations for the assessment of flow-mediated dilation in humans. Eur Heart J 2019; 40: 2534–2547.
• 44.   Kelly AS, Marlatt KL, Steinberger J, Dengel DR. Younger age is associated with lower reactive hyperemic index but not lower flow-mediated dilation among children and adolescents. Atherosclerosis 2014; 234: 410–414.