Associations of Modified Healthy Aging Index With Major Adverse Cardiac Events, Major Coronary Events, and Ischemic Heart Disease

Abstract

Background

The Healthy Aging Index (HAI) has been regarded as useful in capturing the health status of multiple organ systems. However, to what extent the HAI is associated with major cardiovascular events remains largely unknown. The authors constructed a modified HAI (mHAI) to quantify the association of physiological aging with major vascular events and explored how the effects of a healthy lifestyle can modify this association.

Methods and Results

The participants with either missing values of any individual mHAI component or major illnesses such as heart attack, angina and stroke, and self‐reported cancer at baseline were excluded. The mHAI components include systolic blood pressure, reaction time, forced vital capacity, serum cystatin c, and serum glucose. The authors used Cox proportional hazard models to quantify the association of mHAI with major adverse cardiac events, major coronary events, and ischemic heart disease. Cumulative incidence at 5 and 10 years was estimated, and joint analyses were stratified by age group and 4 mHAI categories. The mHAI was significantly correlated with major cardiovascular events, which is a better reflection of the aging level of the body than chronological age. An mHAI was calculated in 338 044 participants aged 38 to 73 years in the UK Biobank. Each point increase in the mHAI was associated with a 44% higher risk of major adverse cardiac events (adjusted hazard ratio [aHR], 1.44 [95% CI, 1.40–1.49]), 44% higher risk of major coronary events (aHR, 1.44 [95% CI, 1.40–1.48]), and 36% higher risk of ischemic heart disease (aHR, 1.36 [95% CI, 1.33–1.39]). The percentage of population‐attribution risk was 51% (95% CI, 47–55) for major adverse cardiac events, 49% (95% CI, 45–53) for major coronary events, and 47% (95% CI, 44–50) for ischemic heart disease, which means that a substantial portion of these events could be prevented. Systolic blood pressure was the factor most significantly associated with major adverse cardiac events (aHR, 1.94 [95% CI, 1.82–2.08]; percentage of population‐attribution risk, 36%), major coronary events (aHR, 2.01 [95% CI, 1.85–2.17]; percentage of population‐attribution risk, 38%), and ischemic heart disease (aHR, 1.80 [95% CI, 1.71–1.89]; percentage of population‐attribution risk, 32%). A healthy lifestyle significantly attenuated mHAI associations with incidence of vascular events.

Conclusions

Our findings indicate that higher mHAI is associated with increased major vascular events. A healthy lifestyle may attenuate these associations.

Nonstandard Abbreviations and Acronyms

aHR

adjusted hazard ratio

HAI

Healthy Aging Index

IHD

ischemic heart disease

MACE

major adverse cardiac events

MCE

major coronary events

mHAI

modified Healthy Aging Index

OPCS‐4

Office of Population Censuses and Surveys Classification of Interventions and Procedures, version 4

PAR%

percentage of population‐attributable risk

RT

reaction time

SBP

systolic blood pressure

UKB

UK Biobank

Clinical Perspective.

What Is New?

• The current study illustrates that a high modified Healthy Aging Index is associated with a higher risk of vascular events, and a healthy lifestyle can attenuate it.

What Are the Clinical Implications?

• We provide a rapid assessment of physiological aging to help identify individuals at high risk for cardiovascular disease in the clinic.

Vascular events remain a major contributor to the global burden of disease. ¹ , ² The prevalence of vascular events in the general population is increasing rapidly, affecting ≈300 million individuals. ³ , ⁴ Multiple studies have supported that aging and vascular events have shared causes and consequences. ⁵ , ⁶ , ⁷ , ⁸ Therefore, it is difficult to disentangle age‐related changes from vascular events. Due to the strong association, it is possible that the significant biomarkers of aging may also overlap with changes seen in vascular events.

The Healthy Aging Index (HAI) is a composite index of biomarkers taken from several systems. ⁹ , ¹⁰ The HAI is constructed using easily assessable examinations to represent an integrated state of health. These proxy indicators, serving as good predictors of the physiological state of individual systems, ¹¹ , ¹² , ¹³ have emerged as a powerful tool in predicting fatal or nonfatal vascular events. ¹² , ¹⁴ , ¹⁵ Such an index may advance the study of aging by serving as a predictor of adverse outcomes such as cardiovascular events.

The acquisition of factors and indicators associated with major vascular events are of great importance in screening high‐risk individuals, promoting healthy aging and increasing the lifespan of patients with vascular events. ¹⁶ , ¹⁷ Furthermore, the potential to tease out individuals who have aged well across systems makes it useful for identifying factors associated with healthy aging and developing strategies for prevention of cardiovascular events.

Therefore, the present study aimed to construct the modified HAI (mHAI) based on the UK Biobank (UKB), and explore the associations of the mHAI with major adverse cardiac events (MACEs), major coronary events (MCEs), and ischemic heart disease (IHD). The contribution of cardiovascular disease attributable to mHAI was also quantified. Finally, the ability of a healthy lifestyle to reduce the incidence of vascular events has also been explored.

Methods

Study Population

The authors declare that all supporting data are available within the article. The UKB, a large prospective population cohort, was established in 22 assessment centers geographically distributed across the United Kingdom. ¹⁸ In brief, individual‐level data of included participants were collected between 2006 and 2010 in the UKB. Personal health information was obtained from self‐administered touchscreen questionnaires and physical measurements and was linked with health records. Various biomarkers were measured by collecting biological samples, such as blood, saliva, and urine. In the present study, participants with either missing values of any individual mHAI component or major illnesses such as heart attack, angina and stroke, and self‐reported cancer at baseline were excluded. UKB has approval from the North West Multi‐centre Research Ethics Committee as a Research Tissue Bank approval. All participants signed informed consent forms. The censoring date of vascular events was determined by 3 main sources (Hospital Episode Statistics for England, Scottish Morbidity Record, and Patient Episode Database for Wales), whose detailed information can be seen on the UKB official website (https://biobank.ctsu.ox.ac.uk/). Follow‐up time was defined as the time from the first measurement to the incidence of vascular events, death, or censoring, whichever came first.

Of the 433 219 participants enrolled in the UKB after excluding diseases as described above, 338 044 participants (78.0%) had complete data on mHAI components. Outcome information and all other mHAI components with a sample size of >400 000 were available for all participants, except for the number of serum glucose (n=370 882), with ≈14.4% missing data. The excluded population had a lower proportion of men (43.7% versus 45.3%) and a higher percentage of smokers (11.6% versus 10.2%) than those included in the study (Table S1).

Definition for Components of mHAI

Previous studies have defined mHAI, ⁹ , ¹⁰ but this is the first time it has been used in the UKB. The physiologic index, including carotid thickness and white matter grade, was regarded as a major predictor of a common age‐related fatal or nonfatal cardiovascular disease. ¹⁹ , ²⁰ However, tests such as color‐flow Doppler ultrasound are not only expensive but also not suitable for screening of vascular disease. Then, we used several proxy indicators for expensive and impractical medical examinations. These proxies can substitute for the health status typically of their respective physiological systems. ⁹ Each indicator was divided into 3 different groups, scoring 0, 1, and 2 points. A smaller score indicates that the participant was healthier than those who received a higher score (Figure 1 and Table S2).

Figure 1. The components of the modified Healthy Aging Index.

The Healthy Aging Index is generated by 5 indicators that represent various physiological systems, including reaction times, systolic blood pressure, forced vital capacity, serum glucose, and cystatin C. Except for serum glucose, which was classified by clinical cutoff value, all of the other items were divided into 3 groups with sex‐specific tertiles.

Systolic Blood Pressure

As a proxy indicator of the circulatory system, the systolic blood pressure (SBP) of participants was measured with a standard automated device. If an automatic device was not available, a manual mercury sphygmomanometer was used instead. Two adjacent measurements of blood pressure were taken a few moments apart. The average SBP were calculated. The analytic sample was divided into 3 groups using sex‐specific tertiles: 0 = ≤134 mm Hg for men and ≤127 mm Hg for women; 1 = 135 to 149 mm Hg for men and 128 to 144 mm Hg for women; and 2 = ≥150 mm Hg for men and ≥145 mm Hg for women. Participants who were taking antihypertensive drugs and had hypertension were divided into the unhealthy group (score=2).

Reaction Time

Previous studies have utilized the digit symbol substitution test to measure the health status of cognitive function. ¹⁰ , ²¹ The bias caused by sampling error with small sample sizes may occur when applying the digit symbol substitution test in the UKB. Therefore, we substituted the digit symbol substitution test with reaction time (RT) in the present study. Participants were shown in front of two random cards, and the RT was the mean duration of the first press of snap butter in which both cards matched. Times <50 ms and >2000 ms were ignored to ensure the fairness of the experiment. Several observational studies have suggested that the symptoms of cognitive deficiency can be obvious with the augmentation of RT. ²² , ²³ , ²⁴ There is evidence in UKB to support RT as a suitable proxy for cognitive function. ²⁵ In addition, RT was significantly correlated with the digit symbol substitution test in our study population (P<0.001). We classified participants into 3 groups using sex‐specific tertiles of the samples, with 507 590 being regarded as cutoff values for women and 489 572 for men.

Forced Vital Capacity

Most participants were blowing using standardized spirometry to measure forced vital capacity at study baseline. A marginal decline in forced vital capacity, which was expressed as a percentage, was associated with a poor prognosis in lung function. Sex‐specific tertiles were used to divide analytic participants into 3 groups: 0 = ≥3.41% for men and ≥4.77% for women; 1 = 2.88% to 3.40% for men and 4.02% to 4.76% for women; and 2 = ≤2.87% for men and ≤4.01% for women. ⁹

Serum Cystatin C

Serum cystatin C is a cysteine protease inhibitor that appears to be a greater predictor of glomerular function than serum creatinine in patients with chronic kidney disease. ²⁶ It was measured by latex enhanced immunoturbidimetric analysis using a Siemens ADVIA 1800. Participants were classified into 3 groups using sex‐specific tertiles, with 0.87 mg/L and 0.97 mg/L being regarded as cutoff values for men and 0.80 mg/L and 0.91 mg/L for women.

Serum Glucose

The regulation of glucose is a central component of living systems, and serum glucose is a crucial indicator of the homeostasis of body metabolism. We used cutoff values of 140 mg/dL and 200 mg/dL to divide the participants into 3 groups according to the criterion of serum random glucose draw by the American Diabetes Association. ²⁷ Participants who were taking a hypoglycemic drug or having an insulin injection or had a diabetes diagnosis were assigned to the group scoring 2.

A scoring system of mHAI was established to grade the aging of various organs and systems by summing each mHAI indicator mentioned above. We further split it into 4 groups as follows: 0 to 2 (the healthiest group), 3 or 4 (the healthy group), 5 or 6 (the unhealthy group), and 7 to 10 (the unhealthiest group). We examined the comparability of baseline characteristics between the 4 groups (Table 1).

Table 1.

Baseline Characteristics of Participants Stratified by mHAI

Characteristics	All	mHAI*				P value†
		Healthiest group (0–2)	Healthy group (3 or 4)	Unhealthy group (5 or 6)	Unhealthiest group (7–10)
No. of patients	338 044	83 135	110 060	97 739	47 142
Age at baseline, y	55.8±8.1	49.7±6.8	54.8±7.4	59.1±6.8	61.9±5.9	<0.0001
Men	153 095 (45.29)	35 661 (42.9)	51 660 (46.9)	44 632 (45.7)	21.142 (44.9)	<0.0001
Smoking status						<0.0001
Current	34 270 (10.2)	7504 (9.1)	11 185 (10.2)	10 249 (10.5)	5332 (11.4)
Previous	112 283 (33.3)	25 182 (30.4)	35 688 (32.5)	34 069 (35.0)	17 344 (37.0)
Never	190 338 (56.5)	50 272 (60.6)	62 881 (57.3)	53 024 (54.5)	24 161 (51.6)
BMI, kg/m2	27.2±4.7	25.2±3.7	26.8±4.2	28.1±4.7	29.9±5.4	<0.0001
Daily alcohol consumption	15.0±18.3	15.4±16.8	15.9±18.9	14.7±18.7	12.6±18.2	<0.0001
Physical activity, 1000 MET	2.67±2.45	2.69 (2.47)	2.70 (2.50)	2.67 (2.45)	2.54 (2.31)	<0.0001
Family history of diabetes	63 332 (20.2)	13 692 (18.4)	19 915 (19.5)	18 730 (20.2)	10 995 (24.5)	<0.0001
Family history of CVD	96 677 (30.8)	19 289 (26.0)	31 315 (30.6)	30 811 (33.3)	15 262 (34.0)	<0.0001
Family history of cancer	89 444 (28.5)	20 818 (28.0)	29 233 (28.6)	26 866 (29.0)	12 527 (27.9)	<0.0001
Single item of mHAI
Forced vital capacity, L	3.76±1.06	4.33±1.06	3.91±1.02	3.45±0.90	3.02±0.81	<0.0001
Systolic blood pressure, mm Hg	139.4±19.5	125.8±13.7	137.5±17.8	146.9±18.5	152.5±18.2	<0.0001
Cystatin C, mg/L	0.89±0.16	0.80±0.09	0.87±0.12	0.94±0.15	1.03±0.21	<0.0001
Reaction time, ms	554.8±115.2	485.2±68.5	539.2±100.5	586.9±116.9	647.6±123.3	<0.0001
Serum glucose, mg/dL	91.5±20.8	87.1±11.5	89.6±15.3	92.6±20.7	101.3±35.9	<0.0001

Values are expressed as number (percentage) or mean±SD. BMI indicates body mass index; MET, metabolic equivalent; and mHAI, modified Healthy Aging Index.

^*Patients with cardiovascular disease (CVD) were excluded.

^† P value were obtained from F test for continuous variable and χ² test for categorical variables.

Definition for Healthy Lifestyle Score

As described in the previous literature, the lifestyle score is classified by 5 lifestyle factors: body mass index (BMI), physical activity, diet, smoking status, and alcohol intake. ²⁸ A BMI <25 kg/m² is defined as healthy; ≥150 minutes of moderate activity per week or ≥75 minutes vigorous activity per week, or a combination thereof, is defined as healthy; a healthy diet was described as an appropriate intake of 4 types of food (vegetables, fruits, fish, processed meat and nonprocessed meat); ²⁹ , ³⁰ never smoking is defined as healthy; and alcohol intake ≤14 g/d for women and ≤28 g/d for men is defined as healthy. A score of one point was given if “healthy” was met. A score of 0 or 1 was defined as the unhealthy lifestyle group, a score of 2 or 3 was defined as the intermediate group, and a score of 4 or 5 was defined as the healthy lifestyle group. A joint effect of 3 healthy lifestyle score groups (healthy, intermediate, and unhealthy) and mHAI groups (healthiest, healthy, unhealthy, and unhealthiest) on the risk of incident vascular events were explored.

Ascertainment of Vascular Disease Events

The vascular events we used were segmented into 3 relative events: MACE, MCE, and IHD. Information on prevalent and incident MACE, MCE and IHD was regularly obtained by constantly updated medical records of physicians' diagnoses. Both the International Classification of Diseases, Ninth Revision (ICD‐9), and the International Statistical Classification of Diseases, Tenth Revision (ICD‐10), were utilized to define medical records. The operative procedures we used were from the Office of Population Censuses and Surveys Classification of Interventions and Procedures, version 4 (OPCS‐4). The definition of MCE was composed of the first occurrence of nonfatal myocardial infarction, coronary revascularization, or coronary death. ³¹ We defined MACE as MCE plus ischemic stroke. ³¹ IHD was defined using ICD‐9 code 410–414 and ICD‐10 code I20‐I25. Detailed information on the outcome definition can be obtained in Table S3.

Statistical Analysis

Several covariates measured at baseline were included and divided into 3 models. Specifically, we considered the following 4 Cox models: model 0 did not adjust for any covariates; model 1 adjusted for sex and age; model 2 additionally adjusted for BMI (kg/m²), smoking status (never smoking, previous smoking, current smoking), daily alcohol intake (measured by drink equivalent, 1 drink equivalent described as containing 14 g of pure alcohol), and family history of disease (diabetes, cardiovascular disease, cancer). In addition, the World Health Organization age classification criteria were used to group age ranges (≤44 years, 45 to 60 years, and ≥60 years). ³²

We used the Cox proportional hazard model to evaluate the association between mHAI and the 3 vascular outcomes and compared hazard ratios for each mHAI group with those in the healthiest groups. The relationship between each individual mHAI item and the 3 outcomes was also tested, and the percentage of population‐attribution risk (PAR%) was calculated for the unhealthy group. We examined mHAI distributions in different age groups separately. Potential nonlinear associations between mHAI and outcomes were assessed by utilization of restricted cubic spline analysis, with a significance test of nonlinear associations. Cox regression was also conducted to calculate the cumulative incidence of events and 95% CIs during the ≈10‐year follow‐up. We further examined the association between mHAI and the 5‐ and 10‐year cumulative incidence of MACE, MCE, and IHD, owing to the long follow‐up period.

To ensure the robustness of the results, we performed a sensitivity analysis. We checked the association between single items of mHAI and risk of MACE, MCE, and IHD in different models. For the covariate we used, we performed a subgroup analysis to explore the potential association and examined its interaction effect with the mHAI (P for interaction). A 2‐sided P value of <0.05 was considered to indicate a significant difference. All statistical analyses were performed in Stata/SE, version 15.0 (StataCorp LLC).

Results

The mean mHAI in all 338 044 participants was 4.11 (SD, 2.11) (Table 1). A magnitude of variation existed in the mHAI groups among different baseline characteristics. The distribution of mHAI was roughly normal, but the distribution in the older age group (age ≥60 years) was negatively skewed compared with that in the young age groups (Figure 2). Participants with the unhealthy group tended to be older, have higher BMIs, and have lower rates of nonsmokers. As mHAI increased, participants seemed to have a larger proportion of family history of CVD and diabetes.

Figure 2. Distribution of the modified Healthy Aging Index between age groups.

The distribution of the Healthy Aging Index among 3 age groups (≤ 44 years, 45–60 years, and ≥60 years).

The average follow‐up of all participants was 11.03 years, and the number of incident cases of MACE, MCE, and IHD were 8039 (2.38%), 5968 (1.77%), and 12 864 (3.81%), respectively (Table 2). A linear and positive association between the mHAI and risk of vascular events using restricted cubic spline regression was also detected (for nonlinearity, P<0.001 for MACE, MCE, and IHD) (Figure 3A, 4). Each point increase in the mHAI was associated with a 44% higher risk of MACE (adjusted hazard ratio [aHR], 1.44 [95% CI, 1.40–1.49]), 44% higher risk of MCE (aHR, 1.44 [95% CI, 1.40–1.48]), and 36% higher risk of IHD (aHR, 1.36 [95% CI, 1.33–1.39]). Similar positive correlations between mHAI and 5‐ and 10‐year cumulative incidence of vascular events were found (Table S4).

Table 2.

Risk of MACE, MCE, and IHD According to mHAI Categories

Outcome	mHAI				P trend	Each point increment	PAR%§
	Healthiest group (0–2)	Healthy group (3 or 4)	Unhealthy group (5 or 6)	Unhealthiest group (7–10)
MACE
Cases, PYs	478 (728 932)	1478 (955 447)	2221 (837 678)	1791 (396 152)
Model 0*	Reference	2.38 (2.15–2.64)	4.12 (3.73–4.54)	7.08 (6.40–7.84)	<0.001	1.83 (1.79–1.88)	68% (66%–70%)
Model 1†	Reference	1.81 (1.63–2.02)	2.67 (2.40–2.97)	4.15 (3.71–4.63)	<0.001	1.56 (1.51–1.61)	57% (54%–60%)
Model 2‡	Reference	1.63 (1.46–1.82)	2.26 (2.02–2.52)	3.24 (2.88–3.65)	<0.001	1.44 (1.40–1.49)	51% (47%–55%)
MCE
Cases, PYs	650 (728 327)	1927 (953 760)	2995 (834 880)	2467 (393 637)
Model 0*	Reference	2.28 (2.08–2.49)	4.07 (3.74–4.43)	7.17 (6.58–7.82)	<0.001	1.86 (1.82–1.90)	68% (66%–70%)
Model 1†	Reference	1.69 (1.54–1.85)	2.51 (2.29–2.75)	3.93 (3.57–4.32)	<0.001	1.55 (1.51–1.59)	55% (52%–58%)
Model 2‡	Reference	1.53 (1.39–1.68)	2.14 (1.95–2.35)	3.10 (2.80–3.44)	<0.001	1.44 (1.40–1.48)	49% (45%–53%)
IHD
Cases, PYs	1088 (726 024)	3291 (946 250)	4839 (823 667)	3646 (385 537)
Model 0*	Reference	2.32 (2.16–2.48)	3.90 (3.66–4.17)	6.28 (5.87–6.72)	<0.001	1.76 (1.73–1.79)	66% (65%–68%)
Model 1†	Reference	1.77 (1.65–1.90)	2.51 (2.34–2.70)	3.63 (3.37–3.91)	<0.001	1.49 (1.46–1.52)	54% (52%–57%)
Model 2‡	Reference	1.60 (1.49–1.72)	2.09 (1.94–2.25)	2.73 (2.52–2.96)	<0.001	1.36 (1.33–1.39)	47% (44%–50%)

All hazard ratios (HRs) (95% CIs) were derived from Cox proportional hazards regression. IHD indicates ischemic heart disease; MACE, major adverse cardiac events; MCE, major coronary event; PAR%, percentage of population‐attributable risk; and PYs, person‐years.

^*Model 0 was unadjusted.

^† Model 1 was adjusted for age, sex, and recruitment assessment center.

^‡ Model 2 was additionally adjusted for body mass index, smoking status, alcohol daily intake, family disease history (diabetes, cardiovascular disease, and cancer).

^§ Percentage (95% CI) of incident cases theoretically attributable to disadvantageous modified Healthy Aging Index (mHAI) group (score ≥3).

Figure 3. Relationship of the modified Healthy Aging Index to major adverse cardiac events (MACEs), major coronary events (MCEs), and ischemic heart disease (IHD).

A, The bar chart shows the number and proportion of participants for each Healthy Aging Index (HAI), and the cubic spline modes for the association of HAI with the incidence risks of MACE, MCE, and IHD. Models were adjusted by sex, age, body mass index (kg/m²), smoking status (never smoking, previous smoking, current smoking), alcohol daily intake (measured by drink equivalent, 1 drink equivalent described as containing 14 g of pure alcohol), and family history of disease (diabetes, cardiovascular disease, cancer). Solid line denotes adjusted hazard ratio (HR), and red area denotes 95% CIs. B, The standardized cumulative incidence of vascular events in score 0 to 2 (reference group), 2 to 4, 5 or 6, 7 to 10 groups in the UK Biobank during ≈10 years of follow‐up; the shaded regions represent the 95% CIs. The cumulative 5‐ and 10‐year incidence rates for each age group are marked by the dashed line.

Figure 4. The joint association of the modified Healthy Aging Index (mHAI) and lifestyle with major adverse cardiac events, major coronary events, and ischemic heart disease.

The lifestyle score was comprised of 5 lifestyle factors: body mass index (BMI), physical activity, diet, smoking status, and alcohol intake. A BMI <25 kg/m² was defined as healthy; ≥150 minutes moderate activity per week or ≥75 minutes vigorous activity per week or mixed was defined as healthy; a healthy diet was defined as an appropriate intake of 4 types of food (vegetables, fruits, fish, processed and nonprocessed meats); never smoking was defined as healthy; and alcohol intake ≤14 g/d for women and ≤28 g/d for men was defined as healthy. Score one point when “healthy” is met. A score of 0 or 1 was defined as an unhealthy lifestyle group; a score of 2 or 3 was defined as an intermediate group; and a score of 4 or 5 was defined as a healthy lifestyle group.

Using the healthiest group (mHAI 0–2) as the reference, the unhealthiest group (mHAI 7–10) was associated with a significantly increased risk of 3 outcomes after multivariable adjustments, with an aHR of 3.24 (95% CI, 2.88–3.65) for MACE, 3.10 (95% CI, 2.80–3.44) for MCE, and 2.73 (95% CI, 2.52–2.96) for IHD (Table 2). We noted a float in the vicinity of 50% for PAR% after adjustment, which suggests that approximately half of the relation of risks was attributed to the nonhealthy group (mHAI ≥3). Figure 3B illustrates that the trend of cumulative incidence of 3 vascular events in the unhealthiest group was steeper than that in the other groups. The participants with the highest mHAI (mHAI 7–10) had a higher 5‐year cumulative incidence (2.38% for MACE, 1.63% for MCE, and 4.66% for IHD) than those with the lowest mHAI (0.33% for MACE, 0.24% for MCE, and 0.65% for IHD). A significant difference in the 10‐year cumulative rates of the 3 vascular events was found, reaching approximately 7% for IHD.

Each single item of the mHAI was strongly associated with MACE, MCE, and IHD. SBP was the most prominent factor associated with vascular events among the 5 individual items, with an aHR of 1.94 (95% CI, 1.82–2.08) for MACE, 2.01 (95% CI, 1.85–2.17) for MCE, and 1.80 (95% CI, 1.71–1.89) for IHD (Table 3, Table S5). The PAR% of SBP ranged from 32% in IHD to 38% in MCE. The aHRs of the second strongest factor, forced vital capacity, were 1.49 (95% CI, 1.40–1.59) for MACE, 1.55 (95% CI, 1.44–1.67) for MCE, and 1.49 (95% CI, 1.42–1.57) for IHD. Forced vital capacity also contributed a respectable PAR%, which ranged from 20% to 23%. RT, which represents the state of cognitive function, was slightly associated with the risk of vascular events. The PAR% of serum glucose associated with vascular events was low compared with other factors.

Table 3.

Associations Between Single Item of mHAI and risk of MACE, MCE, and IHD

mHAI items	MACE		MCE		IHD
	HR (95% CIs)*	PAR%†	HR (95% CIs)	PAR%	HR (95% CIs)	PAR%
FVC
0	1.00 (reference)	20% (17%–23%)	1.00 (reference)	23% (19%–26%)	1.00 (reference)	22% (19%–24%)
1	1.16 (1.09–1.24)		1.22 (1.14–1.32)		1.24 (1.18–1.30)
2	1.49 (1.40–1.59)		1.55 (1.44–1.67)		1.49 (1.42–1.57)
SBP
0	1.00 (reference)	36% (33%–39%)	1.00 (reference)	38% (34%–41%)	1.00 (reference)	32% (30%–35%)
1	1.26 (1.16–1.36)		1.35 (1.23–1.47)		1.24 (1.17–1.32)
2	1.94 (1.82–2.08)		2.01 (1.85–2.17)		1.80 (1.71–1.89)
Cystatin C
0	1.00 (reference)	20% (17%–24%)	1.00 (reference)	20% (16%–24%)	1.00 (reference)	15% (13%–18%)
1	1.16 (1.08–1.24)		1.17 (1.08–1.26)		1.13 (1.08–1.19)
2	1.51 (1.41–1.60)		1.49 (1.39–1.61)		1.34 (1.28–1.41)
RT
0	1.00 (reference)	8% (4%–11%)	1.00 (reference)	8% (4%–12%)	1.00 (reference)	6% (3%–9%)
1	1.07 (1.01–1.14)		1.08 (1.00–1.15)		1.08 (1.03–1.13)
2	1.16 (1.09–1.23)		1.16 (1.08–1.24)		1.10 (1.05–1.15)
SG
0	1.00 (reference)	5% (4%–5%)	1.00 (reference)	5% (4%–5%)	1.00 (reference)	4% (4%)
1	1.33 (1.04–1.69)		1.29 (0.97–1.71)		1.05 (0.85–1.30)
2	1.68 (1.55–1.81)		1.70 (1.56–1.86)		1.61 (1.51–1.71)

All hazard ratios (HRs) (95% CIs) were derived from Cox proportional hazards regression. FVC indicates forced vital capacity; IHD, ischemic heart disease; MACE, major adverse cardiac events; MCE, major coronary events; PAR%, percentage of population‐attributable risk; RT, reaction time; SBP, systolic blood pressure; and SG, serum glucose.

^*Model adjusted for age, sex, recruitment assessment center, body mass index, smoking status, daily alcohol intake, family disease history (diabetes, cardiovascular disease, and cancer).

^† Percentage (95% CIs) of incident cases theoretically attributable to disadvantageous modified Healthy Aging Index (mHAI) group (score ≥3).

The association between mHAI and vascular events stratified by potential risk factors is shown in Table 4. The mHAI was more pronounced associated with vascular events among participants in the younger group (age ≤44 years) (aHR, 1.31 [95% CI, 1.21–1.42] for MACE; aHR, 1.35 [95% CI, 1.24–1.48] for MCE; and aHR, 1.33 [95% CI, 1.25–1.41] for IHD) than in the older group (age ≥60 years) (aHR, 1.22 [95% CI, 1.20–1.25] for MACE; aHR, 1.22 [95% CI, 1.20–1.24] for MCE; and aHR, 1.17 [95% CI, 1.15–1.18] for IHD). Such associations were stronger among women than among men (P<0.001 for interaction). Stratified analyses by BMI group showed that the association of the mHAI with vascular events persisted in the 4 BMI groups but was more significant for underweight participants (BMI <18.5 kg/m²) than for normal weight and above participants (BMI ≥18.5 kg/m²) in the MACE and MCE groups (Table 4).

Table 4.

Subgroup Analyses for the Associations Between 1‐Factor Increment of mHAI and Risks of MACE, MCE, and IHD

Subgroup	MACE	MCE	IHD
	HR (95% CIs)*	HR (95% CIs)	HR (95% CIs)
Total	1.21 (1.20–1.23)	1.22 (1.20–1.24)	1.18 (1.16–1.19)
Age groups, y
≤44	1.31 (1.21–1.42)	1.35 (1.24–1.48)	1.33 (1.25–1.41)
>45 and <60	1.26 (1.24–1.29)	1.26 (1.23–1.29)	1.25 (1.23–1.27)
≥60	1.22 (1.20–1.25)	1.22 (1.20–1.24)	1.17 (1.15–1.18)
P for interaction	0.002†	<0.001†	<0.001†
Sex
Women	1.26 (1.22–1.29)	1.28 (1.24–1.32)	1.21 (1.19–1.24)
Men	1.19 (1.17–1.21)	1.20 (1.17–1.22)	1.16 (1.14–1.17)
P for interaction	<0.001†	<0.001†	<0.001†
Smoking status
Never	1.20 (1.18–1.23)	1.21 (1.18–1.24)	1.18 (1.16–1.20)
Previous	1.22 (1.19–1.25)	1.23 (1.20–1.26)	1.17 (1.15–1.19)
Current	1.22 (1.18–1.26)	1.21 (1.16–1.25)	1.17 (1.13–1.20)
P for interaction	0.076	0.014†	<0.001†
Family history of CVD
Without	1.23 (1.20–1.25)	1.23 (1.21–1.26)	1.19 (1.18–1.21)
With	1.19 (1.16–1.22)	1.19 (1.16–1.22)	1.15 (1.13–1.17)
P for interaction	0.019	0.008†	0.001†
Family history of cancer
Without	1.21 (1.19–1.23)	1.22 (1.20–1.24)	1.18 (1.17–1.20)
With	1.21 (1.18–1.24)	1.21 (1.17–1.24)	1.16 (1.14–1.18)
P for interaction	0.828	0.575	0.061
Family history of diabetes
Without	1.20 (1.18–1.22)	1.20 (1.18–1.23)	1.16 (1.15–1.18)
With	1.25 (1.21–1.28)	1.26 (1.22–1.30)	1.22 (1.19–1.25)
P for interaction	0.222	0.322	0.024
Healthy lifestyle groups
Unhealthy	1.22 (1.19–1.25)	1.22 (1.18–1.25)	1.20 (1.18–1.23)
Intermediate	1.24 (1.22–1.27)	1.25 (1.22–1.28)	1.19 (1.17–1.21)
Healthy	1.21 (1.15–1.28)	1.22 (1.14–1.30)	1.21 (1.16–1.26)
P for interaction	0.007†	0.001†	0.010†
BMI
Underweight, <18.5	1.57 (1.17–2.10)	1.49 (1.01–2.20)	1.14 (0.92–1.42)
Normal weight, 18.5–25	1.26 (1.22–1.29)	1.27 (1.23–1.31)	1.19 (1.17–1.22)
Overweight, 25.0–30.0	1.21 (1.19–1.24)	1.22 (1.19–1.24)	1.18 (1.16–1.20)
Obese, ≥30	1.18 (1.15–1.21)	1.18 (1.15–1.22)	1.17 (1.15–1.19)
P for interaction	<0.001†	<0.001†	<0.001†

CVD indicates cardiovascular disease; HR, hazard ratio; IHD, ischemic heart disease; MACE, major adverse cardiac event; MCE, major coronary event; and mHAI, modified Healthy Aging Index.

^*Model adjusted for age, sex, recruitment assessment center, body mass index (BMI), smoking status, daily alcohol intake, family disease history (diabetes, cardiovascular disease, and cancer).

^† P <0.05.

We observed a joint effect of lifestyle groups (healthy, intermediate, and unhealthy) and mHAI on the risk of incident vascular events (Figure 4). The aHR of 3 vascular events significantly decreased with increase in lifestyle score. The healthiest lifestyle and mHAI group had 18% in MACE, 16% in MCE, and 17% in IHD of the aHR when compared with the unhealthiest lifestyle state (lifestyle groups and mHAI groups were both unhealthy). A joint effect of age group and mHAI on the risk of incident vascular events were also detected. The overall risk of 3 vascular events increased as the unhealthier mHAI group increased (Figure S1). A significant interaction between age group and mHAI groups was found (P _interaction=0.002 for MACE; <0.001 for MCE and IHD). We conducted several sensitivity analyses by excluding participants who had onset of vascular events in the previous 3 or 5 years and who had missing covariable values (Table S6). Similar results were found in the sensitivity model, which showed the robustness of our results.

Discussion

The goal of the present study was to quantify the associations between mHAI and major vascular events. We found that each point increase in the mHAI was strongly associated with 44% higher risk of MACE, 44% higher risk of MCE, and 36% higher risk of IHD. Our findings further indicated that the PAR% of MACE, MCE, and IHD were 51%, 49%, and 47%, respectively, which means that about half of cases of the 3 vascular events can be preventable if the senescence status was reversed. Importantly, a healthy lifestyle significantly attenuated the risk between aging and vascular events.

Our results suggest that mHAI tended to increase with chronologic age. The mHAI captured the signs of organ or system dysfunction, and is identified as a marker of the biological aging process. Our findings further support the sensitivity of mHAI to healthy physiological changes. It is worth noting that previous studies have shown that mHAI predicted the decline of several age‐related functions, such as mortality, slow gait speed, and disability. ⁹ , ³³ , ³⁴ These findings confirm that the mHAI can distinguish individuals who stay in good condition from those who do not.

Interestingly, the association between mHAI and the risk of vascular events was stronger in the young age group than in the old age group. This finding illustrates that the strength of the association between combined exposure to mHAI and lifetime risk of vascular events may depend on both the magnitude and duration of the mHAI. People aged younger but unhealthier (with higher mHAI) were meant to endure the state of senescence for longer. For vascular events, the continued accumulation of organogenic or systemic senescence in early life may be more vital than short‐term exposure to unhealthy mHAI in later life. Our finding showed that SBP is the most associated risk aging factor for vascular events, and previous results have suggested that cumulative exposure to SBP, which is defined as a combination of the magnitude and duration of SBP, may be a crucial risk factor for vascular events in the rest of the lifetime. ³¹ The mHAI is more applicable for evaluating young people because of the integration of effect size and potential duration of multiple risk factors. Namely, the risk gradient of vascular events assigned by mHAI was more valuable in predicting fatal or nonfatal vascular events in elderly individuals with younger age. Because of the lack of dynamically repeated exposure measurements, further studies are needed to quantify the cumulative exposure of single mHAI items precisely. We also found that healthy lifestyle attenuated the observed associations of mHAI with the risk of vascular events, which was proved by previous studies, ³⁵ , ³⁶ , ³⁷ suggesting that a healthy lifestyle can improve the status of aging and help achieve healthy aging.

In addition, the risk between mHAI and MACE and MCE was higher than that in IHD, which implies that mHAI is more effective in predicting the risk of vascular disease than organic disease. A possible explanation is that the heart, compared with blood vessels, is better able to maintain its compensatory state under exposure to both internal and external environmental stresses during the aging process. ³⁸ Although the specific course of dysfunctional development cannot be assessed, the current mHAI may also provide sensitive and potential clinical assessment metrics for vascular senescence.

There were several considerable limitations in our study. First, because the UKB includes predominantly White individuals, our results need to be cautiously extrapolated to other ethnic populations. Second, while several covariables were adjusted for, there were unknown confounding effects. Third, misclassification bias occurred when we used baseline measures of 5 indexes as their lifetime exposure levels. Another limitation is that we lack causal inferences about associations. In spite of the fact that we found a favorable association between HAI and adverse vascular events, this does not imply that the association is causal. It has been shown that genetic risk factors can also lead to adverse vascular events. ³⁹ Further studies are required to explore innovative approaches that include more sensitive predictors, repeated follow‐up data on risk factors, and method of causal inference to capture and describe the dynamic progression of aging.

Conclusion

In summary, the findings indicate that the mHAI was closely associated with fatal or nonfatal vascular events, and the mHAI could be a promising tool to evaluate the cumulative effect of aging on health instead of chronologic age. Furthermore, our findings suggest that a healthy lifestyle may attenuate the associations of HAI with vascular events, underscoring the importance of targeted lifestyle intervention on prevention of vascular events, especially among those with high HAI.

Sources of Funding

The study was supported by grants from the National Key R&D Program of China (2020YFC2003401), and High‐performance Computing Platform of Peking University. The funders had no role in the study design, data collection, data analysis and interpretation, writing of the report, or the decision to submit the article for publication.

Disclosures

None declared.

Supporting information

Tables S1–S6

Figure S1

Reference 40