Subendocardial Viability Ratio Is Associated With Target Organ Damage and Hints at a Potential Independent Predictor of Cardiovascular Mortality in Older Adults: A Prospective Cohort Study

Abstract

Background

The subendocardial viability ratio (SEVR) reflects myocardial perfusion relative to workload. This study explored the association of SEVR with mortality and subclinical target organ damage in an older adult population.

Methods

We analyzed the data from the Northern Shanghai Study, a community‐based cohort of older adults aged over 65 years. SEVR was measured with arterial tonometry. Cross‐sectional associations were assessed between SEVR levels and target organ damage (including arterial stiffness, peripheral artery disease, carotid plaque, left ventricular hypertrophy, left ventricular diastolic dysfunction, chronic kidney dysfunction, and microalbuminuria). Longitudinal associations between SEVR and mortality were evaluated using Cox and Fine‐Gray models.

Results

Among 3237 participants (mean age 71±6 years, 57% female), 233 deaths occurred over a median follow‐up of 5.7 years, including 94 cardiovascular deaths. Participants were divided into 2 groups by the median value of SEVR (129%). The group with a lower SEVR (≤129%) was associated with higher risk of cardiovascular death (adjusted _CoxHazard Ratio [HR]=1.70 [1.05–2.75]; P=0.03), but not with all‐cause death (adjusted _CoxHR=1.30 [0.98–1.73]; P=0.07). As for target organ damage, participants with more damaged organs had lower SEVR values (P for trend <0.001). Specifically, the occurrence of arterial stiffness and left ventricular diastolic dysfunction increased the potential of worse SEVR, with odds ratio 1.46 (1.17–1.83); P<0.001, and 1.90 (1.35–2.67); P<0.001, respectively.

Conclusions

Lower SEVR is independently associated with both increased cardiovascular mortality and organ damage, including arterial stiffness and left ventricular diastolic dysfunction in older adult populations. SEVR may be a convenient marker for early cardiovascular risk stratification in aging populations.

Registration

URL: https://ClinicalTrial.gov; Unique Identifier: NCT02368938.

Nonstandard Abbreviations and Acronyms

cf‐PWV

carotid‐femoral pulse wave velocity

DPTI

diastolic pressure–time index

LVEDP

LV end‐diastolic pressure

SEVR

subendocardial viability ratio

TOD

target organ damage

Clinical Perspective.

What Is New?

• The study revealed that decreased subendocardial viability ratio and the presence of target organ damage in the older adult population is prevalent.

• Among the target organ damage, both arterial stiffness and left ventricular diastolic dysfunction are significantly associated with subendocardial viability ratio decrease, and lower subendocardial viability ratio values (≤129%) were predictive of cardiovascular mortality but not all‐cause mortality in the older adult population.

What Are the Clinical Implications?

• The noninvasive nature and the predictive value of subendocardial viability ratio may enhance the application of pulse wave analysis in both pre‐ and postmyocardial infarction estimation, particularly among older adults who are more vulnerable to myocardial oxygen supply and demand imbalance.

Older adults have a complex cardiovascular risk profile, influenced by both aging‐related physiological changes and multiple comorbidities. ¹ With advancing age, endothelial dysfunction and arterial stiffness progressively impair vascular function, increasing susceptibility to cardiovascular events. ² The presence of multiple comorbidities and polypharmacy further complicates disease management, necessitating a nuanced and individualized approach. ³

Ischemic heart disease remains the leading cause of mortality in older adults, ⁴ , ⁵ with type 2 myocardial infarction (MI) constituting a significant proportion of cases. Unlike type 1 MI, which results from atherosclerotic plaque rupture, type 2 MI is primarily driven by an imbalance between myocardial oxygen supply and demand. ⁶ , ⁷ Despite its clinical significance, direct evaluation or precise measurement of myocardial oxygen supply–demand mismatch remains challenging.

The subendocardial viability ratio (SEVR) is a surrogate index of oxygen supply–demand mismatch and can indicate myocardial ischemia, even in the coronary microcirculation. ⁸ Proposed in the 1970s by Buckberg et al, ⁹ SEVR is calculated as the ratio of the diastolic pressure–time index, representing myocardial oxygen supply, to the systolic pressure–time index, representing myocardial oxygen demand. Currently, SEVR can be measured noninvasively using applanation arterial tonometry devices, which show acceptable agreement with the corresponding invasive measurements. ¹⁰ , ¹¹ Numerous studies have explored the factors influencing SEVR and its potential clinical applications. Key determinants such as sex, age, heart rate, blood pressure, and ejection time have significantly affected SEVR. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Moreover, SEVR has been investigated for its predictive value in cardiovascular events across various diseases and age spectrum, ¹⁰ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ including patients with hypertension, diabetes, peripheral arterial disease, and chronic kidney disease (CKD). Nevertheless, limited research has been conducted on the impact of preclinical organ damage on SEVR. Additionally, large‐scale epidemiological studies in both specific and general populations are necessary to further validate its predictive value for future cardiovascular events and mortality.

Therefore, in this study, we used data from over 3000 community‐dwelling older participants enrolled in the Northern Shanghai Study (NSS), which systemically evaluated cardiovascular risk factors and preclinical damage in the arteries, heart, and kidneys. We aimed to examine the cross‐sectional relationships between SEVR and various forms of target organ damage (TOD) and investigate the longitudinal associations of SEVR with cardiovascular and all‐cause mortalities.

METHOD

Individual participant data from our study will not be publicly available. For information on data access policies and procedures, please contact the corresponding authors.

Study Design

This study is based on the data from the NSS, which is a registered prospective cohort study (ClinicalTrials.gov Identifier: NCT02368938). The study enrolled a community‐dwelling population of older Chinese adults to assess the profile of cardiovascular risk factors and organ damages, to inform future prevention and intervention for cardiovascular diseases (CVDs). Local residents aged 65 years or older from several urban communities of northern Shanghai were invited. Participants with severe cardiovascular diseases (New York Heart Association Classification IV), stroke within 3 months, CKD stage 4 and 5, or malignant tumors, were excluded. The detailed rationale and design of the NSS have been described previously. ²² From June 2014 to October 2022, a total of 3363 responders (response rate 93.7%) were enrolled in the study. For the present study, further exclusions were made, finally leaving 3237 participants for analysis (Figure S1).

The study was carried out according to the principles of the Declaration of Helsinki and approved by the local ethics committees (ethical approval number: 22K148). Written informed consent was obtained from all participants. Data were obtained and analyzed according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Clinical Covariates Measurement

We collected demographic data by face‐to‐face interview using a standard questionnaire, including age, sex, living habits, and medical history. Arterial hypertension was defined by ≥140/90 mm Hg according to the mean value of 3 measurements, or the use of antihypertensive drugs. Diabetes and hyperlipidemia were also recorded. Antihypertensive drugs included angiotensin‐converting enzyme inhibitors, angiotensin II receptor blockers, calcium blockers, diuretics, or beta blockers. Anthropometric data, including body weight, body height and waist circumference, were measured using standardized methods. Technicians conducted laboratory tests of blood and urine samples in the Department of Laboratory Medicine of Shanghai Tenth People’s Hospital. Electrocardiograms (ECGs) were processed by the Modular ECG Analysis System (MEANS) to obtain Minnesota codes. Atrial fibrillation was diagnosed according to the Minnesota codes. Valvular disease was diagnosed based on echocardiography report. The details of clinical covariates are provided in Supplemental Material: eMethods.

SEVR and Hemodynamic Parameters Measurement

All hemodynamic parameters were measured by physicians at a standard room temperature (22 °C–27 °C). Physicians received standardized training to ensure the consistency and accuracy of the output. SEVR and other hemodynamic parameters, including aortic pulse pressure, aortic augmentation pressure, and augmentation index, were measured by a proprietary validated algorithm, using the applanation arterial tonometry device (SphygmoCor, AtCor Medical, Australia). SEVR was calculated as the ratio of the area under the central aortic pressure wave during diastole (surrogate for the diastolic pressure–time index) to that during systole (surrogate for the systolic pressure–time index), detailed in Supplemental Material: eMethods. We defined SEVR 129% as a binary cut‐off, corresponding to the median value observed in our cohort. This threshold also aligns with the manufacturer’s operator manual and prior studies using the same device. ¹⁹ , ²³ , ²⁴ The pulse wave analysis quality control was performed automatically by the device. Correspondingly, the quality indices were as follows: 80<T1<150 milliseconds, operator index ≥80%, average pulse height ≥80%, diastolic variation ≤5%, and pulse height variation ≤5%.

TOD Measurement

According to the latest European Society of Cardiology hypertension guideline, ²⁵ we assessed 8 subtypes of hypertension‐mediated organ damage across 3 organ systems: heart (echocardiographic and electrocardiographic left ventricular hypertrophy [LVH]; LV diastolic dysfunction), kidneys (CKD defined by decreased estimated glomerular filtration rate; microalbuminuria), and arteries (large artery stiffness indicated by increased carotid‐femoral pulse wave velocity [cf‐PWV]; presence of carotid plaque; and a low ankle‐brachial index indicating peripheral artery disease [PAD]). We use the term “target organ damage” to indicate the presence of subclinical complications, given that factors besides blood pressure (ie, dyslipidemia, hyperglycemia) can contribute to this organ damage, ²⁵ and not all participants had hypertension. The acquisition details of each TOD are provided in Supplemental Material: eMethods.

Study Outcomes

The outcomes of this study consist of cardiovascular death and all‐cause death, which were confirmed by reviewing the Shanghai medical insurance system and the Shanghai mortality record system, and telephone follow‐up, in October 2022. Cardiovascular death encompassed death coded using the International Classification of Diseases, Tenth Revision (ICD‐10) (I00‐I09, I11, I13, I20‐I51, I60‐69 used for cardiovascular disease deaths).

Statistical Analysis

Continuous variables were presented as mean±SD or median (interquartile range) according to the data distribution. Categorical variables were presented as absolute numbers and the corresponding percentages in round brackets. The differences in characteristics between groups were detected applying the 2‐sample t‐test or Mann–Whitney U test, whenever appropriate, for continuous variables, and the Chi‐square test for categorical variables.

SEVR was analyzed both as a continuous variable and a categorical variable binary classified by median value. In cross‐sectional analysis, crude and adjusted linear and logistic regression analysis were conducted, with each TOD component as the independent variable, and SEVR as the dependent variable. Multivariable models adjusted for age, sex, body mass index, smoking status, diabetes, hypertension, hypertension treatment, smoking, prevalent CVDs, hyperlipidemia, hemoglobin, blood creatinine, heart rate, and central pulse pressure. In the adjusted mediation analysis, the PROCESS macro in SPSS, modeled with ordinary least squares regression analysis, was used to estimate the mediation effect. Details are described in Supplemental Material: eMethods. Subgroup analyses were also conducted. Sensitive analysis was performed in participants without CVDs history.

In survival analysis, Cox proportional hazards regression and competing risk models were conducted. Univariable and multivariable Cox regression models were used to assess the hazard ratio (HR) and 95% CI, while the multivariable model adjusted for traditional cardiovascular risk factors, including age, sex, body mass index, physical exercises, smoking, diabetes, hypertension, prevalent CVDs, hyperlipidemia, blood creatinine, and pulse pressure. The proportional hazards assumption, influential observations, and nonlinearity of the Cox regression model was checked or detected by using Schoenfeld, Deviance, and Martingale residuals, respectively. Sensitivity analysis after excluding outliers was also performed. Power analysis for Cox models was also performed using R package “powerSurvEpi” and “powerEpiCont”. Nonlinearity in the relationship between continuous SEVR measures and mortality risk was assessed using restricted cubic splines (RCS) regression model. Data were fitted by a restricted cubic spline Cox proportional hazards regression model, with 3 knots at the 10th, 50th, and 90th percentiles of SEVR.

Given the presence of the competing risk of cardiovascular mortality, cumulative incidence curves of binary SEVR were estimated using Gray’s test. These curves were the marginal event rates in the presence of the competing risks of death when cardiovascular mortality is the event of interest. Meanwhile, Fine‐Gray subdistribution models were conducted to evaluate the hazard ratio of cardiovascular mortality.

A 2‐sided P value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS v.27 statistical software package (SPSS Inc., Chicago, IL) and R Software (version 4.4.1).

RESULT

Baseline Characteristics of Participants

Of the 3237 participants in our study, 56.6% were women. The median (IQR) age was 69 (66−74) years, the median SEVR value was 129% (115−145), and a total of 7187 TOD were recorded. The distribution of SEVR and normality test results are shown in Figure S2. Other baseline characteristics are shown in Table 1, with participants divided into 2 groups classified by the median SEVR value. SEVR ≤129% subgroup was characterized by older age, a higher percentage of women (nearly 20% more), a higher heart rate, a greater burden of TOD, and more underlying diseases. The atrial fibrillation was not shown in Table 1 because there were only 2 cases in each group. Over a median follow‐up time of 5.7 (4.4–7.3) years, 233 all‐cause deaths (12.6 per 1000 person‐years) occurred, including 94 cardiovascular deaths (5.1 per 1000 person‐years).

Table 1.

Characteristics of Participants at Baseline

Variables	No. (%)			P value*
	Total (n=3237)	SEVR≤129 (n=1621, 50.1%)	SEVR>129 (n=1616, 49.9%)
Age, median (IQR), y	69 [66–74]	70 [67–75]	69 [66–73]	<0.001
Sex
Female	1832 (56.6)	1079 (66.6)	753 (46.6)	<0.001
Male	1404 (43.4)	541 (33.4)	863 (53.4)
BMI,† median (IQR), kg/m2	24 [22–26]	24 [22–26]	24 [22–26]	0.85
Current smoking	799 (24.7)	287 (17.7)	512 (31.7)	<0.001
Physical activity, median (IQR), h/wk	3.5 [0–7]	3.5 [0–7]	3.5 [0–7]	0.60
Hypertension	1711 (52.9)	925 (57.1)	786 (48.6)	<0.001
Antihypertensive treatment	1607 (93.9)	873 (94.4)	734 (93.4)	<0.001
Beta blocker	162 (10.1)	98 (11.2)	64 (8.7)	0.007
Calcium channel blockers	795 (49.5)	422 (48.3)	373 (50.8)	0.05
ACEIs and ARBs	814 (50.7)	443 (50.7)	371 (50.5)	0.004
Diuretics	159 (9.9)	87 (10.0)	72 (9.8)	0.23
Valvular disease‡	251 (7.8)	106 (6.5)	145 (9.0)	0.01
Aortic disease§	41 (1.3)	18 (1.1)	23 (1.4)	0.26
Diabetes	629 (19.4)	388 (23.9)	241 (14.9)	<0.001
Dyslipidemia	1066 (32.9)	562 (34.7)	504 (31.2)	0.04
Stroke	582 (18.0)	319 (19.7)	263 (16.3)	0.03
CVDs||	1018 (31.4)	521 (32.1)	497 (30.8)	0.40
Laboratory test, median (IQR)
HDL, mg/dL	52.1 [44.0–62.5]	52.9 [44.8–63.3]	51.7 [42.8–61.4]	0.004
LDL, mg/dL	119.7 [98.4–142.8]	121.6 [99.6–144.8]	117.9 [97.3–140.5]	0.003
Total cholesterol, mg/dL	195.7 [171.3–221.6]	199.9 [173.6–223.9]	192.2 [168.6–218.5]	<0.001
Total triglyceride, mg/dL	122.1 [92.0–167.3]	124.8 [94.3–171.7]	119.5 [90.3–164.6]	0.007
Hemoglobin, g/dL	13.6 [12.9–14.6]	13.5 [12.7–14.3]	13.8 [13.0–14.8]	<0.001
Arterial hemodynamics, median (IQR)
SEVR, %	129 [115–145]	115 [105–122]	145 [137–158]	<0.0001
Heart rate, beats/min	70 [63–77]	76 [70–83]	63 [59–69]	<0.0001
Central SBP, mm Hg	128 [116–140]	130 [118–143]	125 [115–137]	<0.001
Central PP, mm Hg	55 [45–65]	58 [48–69]	52 [43–62]	<0.0001
Central AP, mm Hg	16 [11–23]	16 [11–22]	17 [12–23]	0.003
Central AIx	31 [24–37]	29 [22–36]	33 [27–39]	<0.0001
Target organ damage
Carotid‐femoral PWV, median (IQR), m/s	9.0 [7.8–10.6]	9.6 [8.3–11.2]	8.5 [7.3–9.8]	<0.0001
Minimum ABI value, median (IQR)	1.05 [0.97–1.11]	1.04 [0.96–1.10]	1.06 [0.98–1.12]	<0.001
LVMI, median (IQR), g/m2	83.3 [68.4–102.1]	82.3 [66.9–100.2]	84.5 [69.7–103.9]	0.002
SV1+RV5, median (IQR), mm	21.3 [16.4–26.5]	21.0 [16.4–26.0]	21.8 [16.5–27.4]	0.009
E/e’ ratio, median (IQR)	8.8 [6.7–11.3]	9.0 [6.9–11.6]	8.4 [6.5–11.0]	<0.001
UACR, median (IQR), mg/g	30.4 [14.4–58.4]	32.9 [15.3–62.8]	28.8 [13.8–53.4]	0.001
eGFR, median (IQR), mL/min per 1.73 m2	84.7 [73.9–91.7]	85.0 [73.8–91.8]	84.7 [74.4–91.6]	0.62
Arterial stiffness	1021 (31.5)	679 (41.9)	342 (21.2)	<0.001
Carotid plaque	2135 (66.0)	1078 (66.5)	1057 (65.4)	0.51
Peripheral artery disease	407 (12.6)	234 (14.4)	173 (10.7)	<0.001
LV hypertrophy by echo	761 (23.5)	387 (23.9)	374 (23.1)	0.62
LV hypertrophy by ECG	310 (9.6)	145 (8.9)	165 (10.2)	0.22
LV diastolic dysfunction	271 (8.4)	161 (9.9)	110 (6.8)	0.001
Microalbuminuria	1603 (49.5)	839 (51.8)	764 (47.3)	0.01
Chronic kidney disease	269 (8.3)	149 (9.3)	120 (7.5)	0.07
No. of TOD ≥1	2988 (92.3)	1536 (96.2)	1452 (91.3)	<0.001
No. of TOD ≥2	2239 (69.2)	1225 (76.7)	1014 (63.7)	<0.001

Sl conversion factor: To convert total cholesterol, LDL, and HDL to milligrams per liter, multiply by 0.1; to convert triglycerides to millimoles per liter, multiply by 0.0113; to convert hemoglobin to grams per liter, multiply by 10.

ACEIs indicates angiotensin‐converting enzyme inhibitors; AIx, argumentation index; AP, argumentation pressure; ARBs, angiotensin II receptor blockers; BMI, body mass index; CVDs, cardiovascular diseases; Echo, echocardiographic; HDL, high‐density lipoprotein cholesterol; IQR, interquartile range; LDL, low‐density lipoprotein cholesterol; LV, left ventricular; PP, pulse pressure; SBP, systolic blood pressure; SEVR, subendocardial viability ratio; and TOD, target organ damage.

^* P value for comparison between SEVR ≤129% subgroup and SEVR >129% subgroup.

^† Calculated as weight in kilograms divided by height in meters squared.

^‡ Valvular disease indicates aortic, mitral, or tricuspid regurgitation and stenosis reported by echocardiography.

^§ Indicates clinically significant aortic regurgitation or stenosis (moderate or severe).

^|| Contains history of any of myocardial infarction, coronary artery disease, arrhythmia, clinically significant aortic disease, and heart failure.

Association of SEVR and TOD Prevalence

Table 1 presents the prevalence of each TOD component and cumulative TOD number per person according to binary SEVR classification. Broadly, in the lower SEVR subgroup, the prevalence of most TOD was higher, with significant differences observed in arterial stiffness (41.9% versus 21.2%), PAD (14.4% versus 10.7%), LV diastolic dysfunction (9.9% versus 6.8%), and microalbuminuria (51.8% versus 47.3%). The average cumulative number of TOD per person was 2.7±1.3 in the lower SEVR subgroup, compared with 2.5±1.3 in the higher SEVR subgroup. Additionally, Figure 1 shows the trend between the cumulative number of TOD per person and SEVR value both as a continuous variable (P for trend <0.001) and as a categorical variable, further suggesting that lower SEVR was closely associated with a higher prevalence of TOD.

Figure 1. SEVR distribution by cumulative number of TOD per person.

*In TOD=7 subgroup, the percentage of SEVR >129% is 0.03%, and the percentage of SEVR ≤129% is 0.03%. A, The box‐and‐whisker plot shows continuous SEVR distribution among different cumulative numbers of TOD per person. B, The bar graph shows binary SEVR distribution among different cumulative numbers of TOD per person. SEVR indicates subendocardial viability ratio; and TOD, target organ damage.

Independent Effect of Each TOD Component on SEVR

In the age‐, sex‐, and heart rate‐adjusted correlation analysis (Table S1), SEVR was significantly correlated with cf‐PWV (r=−0.12; P<0.001), E/e’ (r=−0.13; P<0.001), SV1+RV5 (r=−0.04; P=0.026), and estimated glomerular filtration rate (r=−0.07; P<0.001). We then conducted univariable/multivariable linear and logistic regression analysis, shown in Table 2 and Table S2. SEVR was negatively associated with cf‐PWV (β=−0.42, P<0.001), E/e’ (β=−0.37, P<0.001), estimated glomerular filtration rate (β=−0.09, P<0.001), but positively associated with LVMI (β=0.03, P=0.002). Accordingly, lower SEVR level was associated with arterial stiffness (adjusted odds ratio [OR], 1.46; 95% CI, 1.17–1.83; P<0.001) and LV diastolic dysfunction (adjusted OR, 1.90; 95% CI, 1.35–2.67; P<0.001). These associations remained consistent in sensitive and subgroup analyses (Tables S3 through S5). These findings suggest that the odds of SEVR decrease were 46% higher in participants with arterial stiffness and nearly 2‐fold higher in those with LV diastolic dysfunction than their counterparts. Further mediation analyses revealed that the effect of arterial stiffness on SEVR was mediated by the augmentation of arterial pulse waveform (Table S6; Figure S3). Moreover, although a decreased SEVR was associated with LV hypertrophy assessed by echocardiography (adjusted OR, 0.77; 95% CI, 0.61–0.97; P=0.03), this association was not observed when LV hypertrophy was assessed by ECG (adjusted OR, 0.97; 95% CI, 0.70–1.35; P=0.86).

Table 2.

Association of Decreased SEVR With TOD Components^*

TOD components	Unadjusted		Model 1		Model 2
	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
Arterial variables
Arterial stiffness (Yes=1)	2.69 (2.30–3.14)	<0.001	2.61 (2.21–3.08)	<0.001	1.46 (1.17–1.83)	<0.001
Carotid plaque (Yes=1)	1.05 (0.91–1.21)	0.51	1.06 (0.91–1.24)	0.43	1.14 (0.93–1.40)	0.20
Peripheral artery disease (Yes=1)	1.41 (1.14–1.74)	0.001	1.37 (1.09–1.71)	0.007	1.04 (0.77–1.40)	0.82
Heart variables
LV hypertrophy by echo (Yes=1)	1.04 (0.89–1.23)	0.62	0.79 (0.66–0.94)	0.007	0.77 (0.61–0.97)	0.03
LV hypertrophy by ECG (Yes=1)	0.86 (0.68–1.09)	0.22	0.85 (0.67–1.09)	0.20	0.97 (0.70–1.35)	0.86
LV diastolic dysfunction (Yes=1)	1.51 (1.17–1.94)	0.001	1.30 (1.00–1.69)	0.05	1.90 (1.35–2.67)	<0.001
Kidney variables
Microalbuminuria (Yes=1)	1.20 (1.04–1.37)	0.01	1.08 (0.93–1.24)	0.32	0.94 (0.77–1.14)	0.50
Chronic kidney diseases (Yes=1)	1.26 (0.98–1.62)	0.07	1.01 (0.77–1.33)	0.96	0.87 (0.55–1.38)	0.54

Model 1 was adjusted for: age, sex, body mass index, current smoking; Model 2 was adjusted for: covariates in model 1 plus hypertension, hypertension treatment, diabetes, cardiovascular disease history, hyperlipidemia, heart rate, and blood creatinine.

Echo indicates echocardiographic; LV, left ventricular; SEVR, subendocardial viability ratio; and TOD, target organ damage.

^*Odds ratios (ORs) and 95% CIs are from a logistic regression model and represent the odds of SEVR below 129% in TOD subgroups compared with no TOD subgroups.

Association of SEVR With All‐Cause and Cardiovascular Deaths

We examined the association between decreased SEVR and the risk of all‐cause and cardiovascular mortality using Cox proportional hazards models and Fine‐Gray subdistribution hazard models, with SEVR evaluated both as a continuous and a binary variable (threshold: SEVR ≤129%) (Table 3; Figure 2). As a continuous variable, each SD decrease in SEVR was significantly associated with an increased risk of both all‐cause death (adjusted _CoxHR, 1.16, 95% CI, 1.01–1.35, P=0.04) and cardiovascular death (adjusted _CoxHR, 1.31, 95% CI, 1.02–1.68, P=0.03). Nonlinearity in the relationship between continuous measures of SEVR and unadjusted Cox HR of mortality was shown in RCS curves in Figures 2A and 2B. Consistently, the Fine‐Gray model accounting for non‐cardiovascular death as a competing risk showed that decreased SEVR was significantly associated with higher cardiovascular mortality (adjusted sHR, 1.29, 95% CI, 1.03–1.62, P=0.03).

Table 3.

Associations of All‐Cause and Cardiovascular Mortality With Decreased SEVR^*

Outcomes	Unadjusted Cox model		Adjusted Cox model		Unadjusted Fine‐Gray model		Adjusted Fine‐Gray model
	CoxHR (95% CI)	P value	CoxHR (95% CI)	P value	SHR (95% CI)	P value	SHR (95% CI)	P value
As continuous variable (per SD decrease of SEVR)
All‐cause death	1.22 (1.06–1.40)	0.005	1.16 (1.01–1.35)	0.04	…	…	…	…
Cardiovascular death	1.39 (1.10–1.76)	0.006	1.31 (1.02–1.68)	0.03	1.38 (1.11–1.72)	0.004	1.29 (1.03–1.62)	0.03
As binary variable (SEVR >129% as reference)
All‐cause death (SEVR ≤129%)	1.34 (1.02–1.75)	0.04	1.30 (0.98–1.73)	0.07	…	…	…	…
Cardiovascular death (SEVR ≤129%)	1.76 (1.12–2.78)	0.02	1.70 (1.05–2.75)	0.03	1.74 (1.10–2.74)	0.02	1.65 (1.03–2.65)	0.04

Adjusted models are adjusted for age, sex, body mass index, current smoking, weekly exercise hours, blood creatinine, hypertension, diabetes, cardiovascular disease history, and hyperlipidemia.

_CoxHR indicates hazard ratio for Cox model; SEVR, subendocardial viability ratio; and _SHR, hazard ratio for Fine‐Gray subdistribution model.

^*SEVR is evaluated as a continuous and binary variable separately. Hazards ratios (HR) and 95% CIs are from a Cox proportional hazards model. For cardiovascular death, a Fine‐Gray subdistribution hazard model is added, accounting for non‐cardiovascular death as a competing risk.

Figure 2. RCS curve and cumulative incidence curves of all‐cause and cardiovascular mortalities by SEVR.

A, RCS curve for hazard ratios of all‐cause mortality by continuous SEVR according to the Cox regression model. B, RCS curve for hazard ratios of cardiovascular mortality by continuous SEVR according to the Cox regression model. C, Cumulative incidence curves (Gray’s test) for all‐cause mortality grouped by binary SEVR. D, Cumulative incidence curves (Gray’s test) for cardiovascular mortality grouped by binary SEVR. HR indicates hazard ratios; RCS, restricted cubic splines; and SEVR, subendocardial viability ratio.

When SEVR was a binary variable, individuals with SEVR ≤129% had a significantly higher risk of all‐cause death (unadjusted Gray’s test P=0.035) and cardiovascular death (unadjusted Gray’s test P=0.016) compared with those with SEVR >129%, shown in Figures 2C and 2D. Similarly, in hazard models analysis, participants with SEVR ≤129% had nearly 70% increased hazards of cardiovascular death (adjusted _CoxHR, 1.70, 95% CI, 1.05–2.75, P=0.03; adjusted sHR, 1.65, 95% CI, 1.03–2.65, P=0.04) compared with SEVR >129% subgroup. However, after adjustment, the association between low SEVR and all‐cause mortality did not reach statistical significance (adjusted _CoxHR, 1.30, 95% CI, 0.98–1.73, P=0.07).

Analyses of Schoenfeld residuals showed that all Cox regression models met the proportional hazards assumption (Figure S4). Deviance residual analysis identified some outliers deviance residuals >3 (Figure S5), which were further checked and found no accuracy of data. Martingale residual indicated the near linearity between the log hazard and SEVR (Figure S6). Sensitivity analysis excluding outliers showed similar results in most models (Table S7). Power analysis showed that study is somewhat underpowered, especially for the association of all‐cause mortality with SEVR (Table S8). Of note, results after excluding outliers showed enough power for examining the significant association between cardiovascular mortality and SEVR (Table S9).

DISCUSSION

In the present study, we assessed SEVR and its associations with both TOD cross‐sectionally and mortality longitudinally in a cohort of more than 3000 older Chinese adults. Our analysis yielded 3 noteworthy findings: First, a reduction in SEVR and the presence of subclinical TOD were common in our older adult participants. Second, decreased SEVR was significantly associated with the accumulation of TOD in arteries, heart, and kidneys; in specific, decreased SEVR was significantly and negatively associated with arterial stiffness, which was mediated by the augmentation of arterial pulse waveforms. Third, decreased SEVR emerged as a potential independent predictor of cardiovascular mortality.

First and foremost, we acknowledged that the measured SEVR can only serve as a reference or indicator of the real subendocardial viability. Because the intra‐ventricular diastolic pressure and left ventricular isometric duration were not considered, SEVR based on central aortic waveforms might cause overestimation. Despite more precise measures having been proposed, ²⁶ our equipment limitations prevent us from deriving a more accurate SEVR. Although the absolute SEVR value is difficult to generalize due to device variance, its distribution and trend within this cohort can give some insights. To mitigate the impact of potential inaccuracies, we conducted subgroup and sensitivity analyses and optimized covariate selection to ensure the reliability and accuracy of the results.

As a surrogate index used to evaluate the myocardial ischemic susceptibility, SEVR’s prediction potential on adverse CVD outcomes has been reported in previous literature. Age‐related attribution was the first to study. The CARLA study suggested that SEVR ≤130% was an age‐dependent predictor for all‐cause mortality in men under 60 years. ²³ However, the association between SEVR and mortality was not observed in participants over 80 years, consistent with findings from our subgroup analysis. This attenuation is mainly due to age becoming a dominant risk factor in very old individuals, potentially overshadowing SEVR’s prognostic value. Moreover, the small sample size in this age stratum limits statistical power in both studies, so subgroup results should be interpreted cautiously. In contrast, the predictive values of blood pressure (BP) and ARTerial stiffness in institutionalized very AGEd population (PARTAGE) study, which involved a larger sample of older nursing home residents and used an improved measurement device, demonstrated a clear inverse association between SEVR and mortality. ²¹ Given its methodological strengths, we consider the PARTAGE results relatively compelling. Sex‐related differences of SEVR were discussed in the Anglo‐Cardiff Collaborative Trial ¹³ and the Wakuya study. ¹⁴ According to them, older women had more unfavorable SEVR than men because of differing effects of vascular aging and accelerated aortic diastolic pressure decay. Other researchers compared a serial of hemodynamic parameters and found that SEVR predicted cardiovascular outcome and improved risk discrimination in patients with type 2 diabetes ¹⁶ as well as CKD. ¹⁹ Our study provided additional evidence supporting the prognostic value of SEVR for cardiovascular mortality, with consistent results in traditional Cox regression analysis and competing risk analysis. In our cohort, SEVR ≤129% was characterized by older age, a higher percentage of women (nearly 20% more), a higher heart rate, a greater burden of TOD, and more underlying diseases. These findings reinforced the prognostic power of SEVR in cardiovascular mortality.

While both TOD and SEVR reductions may suggest a predisposition to adverse cardiovascular outcomes, ²⁵ to our knowledge, few studies have interpreted SEVR decrease from the perspective of organ damage. We found that SEVR declined with the increased cumulative TOD per person. Among the 8 subtypes studied, arterial stiffness and LV diastolic dysfunction exhibited a reliable association with SEVR.

Arterial stiffness in the large arteries modulates central hemodynamics and can predispose to myocardial ischemia irrespective of coronary artery status, especially in older adults. ¹³ The long‐term prognostic importance of arterial stiffness for multiple health outcomes has been established in the Framingham Heart Study. ²⁷ In a previous study, arterial stiffness was proved to contribute to a reduction in the myocardial supply–demand ratio. ²⁸ In terms of mechanism, arterial stiffness made early systolic pressure rise, which increased the afterload of the heart as well as peak myocardial wall stress, a key determinant of myocardial oxygen consumption. Moreover, the physiology of forward and reflected waves changed. In young healthy adults, wave reflections arrive at the aorta predominantly during diastole, augmenting diastolic pressure. When vessel wall stiffens, the premature wave reflections arrive at the aorta in mid‐to‐late systole, producing systolic pressure augmentation and widening pulse pressure. This led to a decrease in the area under the pressure curve in diastole, which is a key determinant of coronary blood flow. We further performed an exploratory mediation analysis. In the model, the adjusted mediation effect of augmentation pressure on cf‐PWV to SEVR was significantly close to 100%, with acceptable multicollinearity diagnosis. The conclusion remained consistent when aortic augmentation pressure was adjusted by HR at 75 beats per minute. Our study corroborated the influence of arterial stiffness on ventricular‐vascular interactions from a hemodynamic point of view.

Our previous analysis demonstrated an association between peripheral artery disease (PAD) and myocardial perfusion, although this relationship was attenuated after further adjustment for aortic augmentation pressure. ²⁹ In the present study, which included a larger sample of participants, we considered SEVR the dependent variable and observed a similar result. The impact of PAD on SEVR was significant only in the unadjusted and basic‐variable adjusted models. Previous studies have reported contradictory findings, with some showing a negative correlation between PAD severity and SEVR values, ¹⁸ , ³⁰ but others have not. ³¹ , ³² The divergent conclusions implicated the intricate relationship between peripheral artery condition and myocardial perfusion. We speculated that pathological change of peripheral arteries may impede pulse wave transmission, which altered central hemodynamics. Indeed, variables reflecting vascular health, such as cf‐PWV, ABI, and carotid plaque, were highly interrelated. Further analysis is needed to comprehensively evaluate these variables and their interactions.

In terms of heart TOD, we found a negative association of SEVR with LV diastolic dysfunction, but a dubious relationship with LVH. Previous studies have shown that the E/e’ ratio and e’ velocity, indicators of LV diastolic dysfunction, serve as independent predictors of primary cardiac events. ³³ , ³⁴ Increased E/e’ ratio partly indicates elevated LV end‐diastolic pressure. An elevated LV end‐diastolic pressure reduces coronary perfusion pressure, hence could decrease SEVR. Although our SEVR algorithm did not take LV end‐diastolic pressure into account, the solid correlation still existed, which implied that E/e’ ratio may have an inherent modulation to parameters in SEVR. Regarding LVH and SEVR analysis, it remains challenging to tell the cause from the effect. LVH is a compensatory response to increased hemodynamic demand. The rise of myocardium mass and workload increases coronary blood flow. ³⁵ However, once the autoregulation decompensates, myocardial blood supply may no longer meet metabolic demands, setting the stage for impaired oxygen and nutrients supplied and potential ischemic injuries. In our study, the inconsistency between continuous and binary SEVR analysis, as well as LVH measurement by ECG and echocardiography, further validate the ambiguity of their relationship.

Previous studies have established a link between SEVR and CKD. Ekart et al have found that SEVR is reduced in patients with CKD with higher proteinuria levels and poor renal function but are not undergoing dialysis. ³⁶ , ³⁷ Prince et al believed that SEVR was associated with microalbuminuria and poor renal function in type 1 diabetes. ¹⁷ However, these findings contrast with ours, where we found insufficient evidence to support a negative correlation between SEVR and CKD. Upon tracing the potential mechanisms, we considered several factors besides sample size and subject selection differences. The change in threnin‐angiotensin‐aldosterone system may disrupt electrolyte and volume homeostasis ³⁸ ; the anemia caused by reduced erythropoietin production from kidneys may impair coronary oxygen supply ³⁹ ; uremic toxins accumulated in CKD could induce endothelial dysfunction and uremic cardiomyopathy ⁴⁰ ; and the dyslipidemia following CKD might accelerate atherosclerosis, thereby diminishing coronary blood flow. ⁴¹ The multiple effects may cover the influence of CKD and microalbuminuria on SEVR, especially in the elderly population.

To our knowledge, this is the biggest database evaluating subendocardial viability ratio in the older adult population, and the first report showing the association between subendocardial viability ratio and target organ damage. However, our research carried several limitations that must be considered when interpreting our data. First, the noninvasive measurement is challenging in accurately determining the SEVR value’s reflection of real subendocardial oxygen supply and demand. We approximated the diastolic pressure–time index and systolic pressure–time index as the areas under the aortic pressure curve during diastole and systole, respectively. This lack of accounting for the isovolumetric components of the cardiac cycle and the LV end‐diastolic pressure is estimated to overestimate diastolic pressure–time index and underestimate systolic pressure–time index. ⁴² Second, as a calculated clinical parameter, SEVR has some limitations in nature, including its variability and dependence on other hemodynamic parameters, such as heart rate. Third, we did not assess SEVR’s predictive ability for type 2 MI due to limited cases, as mismatch between subendocardial oxygen demand and supply was recognized as the underlying mechanism of type 2 MI. Fourth, given the cohort design of the study, the influence of residual and unmeasured confounders cannot be ruled out. Moreover, due to limitations in disease inclusion criteria and diagnostic equipment, several common clinical conditions, such as arterial calcification and comprehensive valvular evaluation, were not fully assessed. Fifth, the number of events is relatively small and power analysis showed that study is somewhat underpowered, especially in the association between all‐cause mortality and SEVR. Therefore, the longitudinal associations found in the present study should be taken as exploratory results and studies with more events and enough power are warranted to validate our findings in the future. Lastly, although our study focused on the elderly population, frailty status was not evaluated due to the lack of standardized frailty data. In addition, the inclusion of only older Chinese adults may limit the generalizability of our findings to other ethnicities and age groups.

CONCLUSION

In this community‐based cohort of older adults, we observed prevalent decreases in SEVR and the presence of TOD before the onset of adverse cardiovascular events, highlighting the importance of primary and secondary prevention in the older adult population. Among the TOD, both arterial stiffness and LV diastolic dysfunction emerged as significant contributors to the decrease in SEVR, and the impact of arterial stiffness was primarily mediated through the augmentation effect of arterial pulse waveforms. Additionally, lower SEVR values were predictive of cardiovascular mortality in the older population, indicating a potential pathophysiological pathway linking central hemodynamic changes to cardiovascular events irrespective of atherosclerosis and coronary stenosis. Further studies with a larger sample size and enough statistical power are needed to verify our findings and extend the results of our study to a younger population or other clinical settings.

APPENDIX

The Northern Shanghai Study investigators: Yi Zhang, Yawei Xu, Jing Xiong, Shikai Yu, Chen Chi, Yifan Zhao, Wenhui Yue, Jiadela Teliewubai, Song Zhao, Jiamin Tang, Haotian Yang, Jiawen Wu, Chong Xu, Rusitanmujiang Maimaitiaili, Jun Han, Weilun Meng, Moran Li, Jie Liu, Jieying Shi, Jingjing Hou, Yue Zhao, Minghui Chen, Qiaorui He, Lili Zhou, Haonan Li.

Sources of Funding

This study was financially supported by the Health Youth Talent Project of Shanghai Municipal Health Commission (2022YQ023), Shanghai Municipal Health Commission’s Seed Program for Medical New Technology Research and Translation (2024ZZ2053), the National Nature Science Foundation of China (82170388, 82370300), the Shanghai Shenkang Simulated RCT Research Project (SHDC2024CRI072).

Disclosures

None.

Supporting information

Tables S1–S9

eMethods

Figures S1–S6

References 43–59