Age‐Related Trends in the Predictive Value of Carotid Intima‐Media Thickness for Cardiovascular Death: A Prospective Population‐Based Cohort Study

Abstract

Background

The age‐related trends in the predictive ability of carotid intima‐media thickness (CIMT) for cardiovascular risk remain unclear. We aimed to identify the age‐related trends in the predictive value of CIMT for cardiovascular death.

Methods and Results

In a prospective cohort of adults aged 35 to 75 years without history of cardiovascular disease who were enrolled between 2014 and 2020, we measured CIMT at baseline and collected the vital status and cause of death. We divided the study population into 4 age groups (35–44, 45–54, 55–64, and 65–75 years). Competing risk models were fitted to estimate the associations between CIMT and cardiovascular death. The added values of CIMT in prediction were assessed by the differences of the Harrell's concordance index and the net reclassification improvement index. We included 369 478 adults and followed them for a median of 4.7 years. A total of 4723 (1.28%) cardiovascular deaths occurred. After adjusting for the traditional risk factors, the hazard ratios for CIMT_mean per SD decreased with age, from 1.27 (95% CI, 1.17–1.37) in the 35 to 44 years age group to 1.14 (95% CI, 1.10–1.19) in the 65 to 75 years age group (P for interaction <0.01). Meanwhile, the net reclassification improvement indexes for CIMT_mean were attenuated with age, from 22.60% (95% CI, 15.56%–29.64%) in the 35 to 44 years age group to 7.00% (95% CI, −6.82% to 20.83%) in the 65 to 75 years age group. Similar results were found for maximum CIMT in all age groups.

Conclusions

CIMT may improve cardiovascular risk prediction in the young and middle‐aged populations, rather than those aged ≥55 years.

Nonstandard Abbreviations and Acronyms

ΔC index

difference of the Harrell's concordance index

CCA

common carotid artery

CIMT

carotid intima‐media thickness

NRI

net reclassification improvement index

Clinical Perspective.

What Is New?

• The associations of carotid intima‐media thickness with the risk of cardiovascular death decrease with age.

• Carotid intima‐media thickness has added predictive value for cardiovascular death based on established traditional risk factors in people aged <55 years rather than those ≥55 years.

What Are the Clinical Implications?

• Carotid intima‐media thickness can improve prediction for cardiovascular death beyond the established risk factors in people aged <55 years but not in older adults.

As a surrogate of atherosclerosis, carotid intima‐media thickness (CIMT) is measured by B‐mode ultrasound, and has the advantages of noninvasive, easy, and reliable measurement. ¹ Although previous studies have showed that increased CIMT was associated with cardiovascular events, ² , ³ , ⁴ , ⁵ and the interventions on CIMT progression could reduce cardiovascular risk, ⁶ the added predictive value of CIMT for cardiovascular outcomes remains uncertain. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² Current guidelines have not recommended measuring CIMT to improve cardiovascular risk stratification. ¹³ , ¹⁴ , ¹⁵

The association of CIMT with cardiovascular risk and the predictive value of CIMT may be related with age, but evidence is lacking. The Carotid Atherosclerosis Progression Study and ARIC (Atherosclerosis Risk in Communities) Limited Access Data study have found that the association between CIMT and cardiovascular outcomes was stronger in younger adults (aged <50 years in the former study, and aged ≤54 years in the latter study), but both studies only divided the population into 2 broad age groups. ¹⁶ , ¹⁷ The age‐related trends in the association between CIMT and cardiovascular risk remain unclear, let alone the added predictive ability of CIMT for cardiovascular risk.

Using a prospective cohort of nearly 0.37 million people aged 35 to 75 years, we had the power to investigate the age‐related trends in the association between CIMT and the risk of cardiovascular death, and in the added predictive value of CIMT for cardiovascular risk.

METHODS

The data underlying this article currently cannot be shared publicly. The first author had full access to all the data in the study and takes responsibility for their integrity and the data analysis.

Study Cohort

In this analysis, we included the participants of the China Health Evaluation and risk Reduction through nationwide Teamwork Project (formerly named China Patient‐Centered Evaluative Assessment of Cardiac Events Million Persons Project) who underwent carotid ultrasonography and without prior cardiovascular disease at baseline (Figure S1). The project is a government‐funded public health screening program designed with a focus on cardiovascular risk across China. The details of the project were described previously. ¹⁸ , ¹⁹ Briefly, from November 2014 through December 2020, a total of 254 sites from the 31 provinces in mainland China were selected to provide diversity in geographic distribution, socioeconomic status structure, and exposure to risk factors and diseases. We invited the community residents, aged 35 to 75 years, who had lived in the community for at least 6 of the preceding 12 months to participate and have their cardiovascular risk evaluated. The central ethics committee at the China National Center for Cardiovascular Diseases approved this project. All enrolled participants provided written informed consent.

Data Collection and Variables

We conducted standardized in‐person interviews with all participants to collect their demographic status, lifestyle, physical examination results, medical history, and medication use. Their locations were divided by the Yangtze River into northern or southern, which is consistent with the Prediction for Atherosclerotic Cardiovascular Disease Risk in China Project. ²⁰ Information on the urban and rural areas was obtained from the local administrative code. Current cigarette smoking was self‐reported. Waist circumference was measured at 1 cm above the navel at minimal respiration. Blood pressure was measured twice, at an interval of 1 minute, on the right upper arm after 5 minutes of rest in a seated position with a standardized electronic blood pressure monitor (Omron HEM‐7430; Omron Corporation, Kyoto, Japan). If the difference between the 2 systolic blood pressure readings was >10 mm Hg, a third measurement was obtained, and the average of the last 2 readings was used. Blood lipids of participants were tested using the whole blood samples after fasting for at least 8 hours by a rapid lipid analyzer (CardioChek PA Analyzer; Polymer Technology Systems, Indianapolis, IN). Total cholesterol, high‐density lipoprotein cholesterol, and triglycerides were measured at baseline. When triglycerides were <4.5 mmol/L (<400 mg/dL), low‐density lipoprotein cholesterol was calculated by the Friedewald method ²¹ ; otherwise, it was measured directly. We defined diabetes as self‐reported history of diabetes or antidiabetic medication use during the past 2 weeks. We also asked about the use of antihypertensive and lipid‐lowering medications during the past 2 weeks. We defined family history of premature atherosclerotic cardiovascular disease as at least a first‐degree relative with known premature (men aged <55 years; women aged <65 years) coronary heart disease or stroke.

Trained sonographers in the local hospitals performed the carotid ultrasonography for participants according to the standard protocol. We required the ultrasonography machines used at local hospitals to be equipped with a 5‐ to 12‐MHz probe. There was no requirement on the types of ultrasonography machines. We measured the intima‐media thickness (IMT) in the far walls of common carotid artery (CCA) over 3 segments at both left and right sides: proximal CCA (1–1.5 cm distal to the origin of CCA), middle CCA, and distal CCA (1–1.5 cm proximal to the carotid bulb). We estimated CIMT in the regions free of plaque. The reasons for measuring the IMT of the CCA were as follows: (1) CCA‐IMT is more easily assessable and reproducible than other segments, because the CCA is perpendicular to the ultrasound beam ²² , ²³ ; and (2) previous studies have found that CCA‐IMT was as good as CIMT in all the carotid artery segments in improving prediction of cardiovascular risk. ²⁴ We defined the mean of CIMT as the average values of 6 measurements at left and right CCA, and the maximum of CIMT as the maximal values of 6 measurements at left and right CCA. Because the American Society of Echocardiography consensus statement recommends the use of CIMT >75th percentile for age, race, and sex as being abnormal, we defined the age‐ and sex‐specific 75th percentile as the cutoff value of CIMT to identify people at high cardiovascular risk in each age group. ²⁵ To ensure the quality of ultrasound examination, we required that sonographers be licensed, and we trained them at the beginning. Furthermore, the first 10 to 20 carotid ultrasound measurements for each study site were checked by senior sonographers from Fuwai Hospital. And if the sonographer in the study site was changed, we would conduct training and check again. However, we did not perform interrater and intrarater analysis among the ultrasound operators.

Outcomes

We ascertained the vital status and the causes of death of each enrolled participant from the China National Mortality Surveillance System, with annual active confirmation from local residential, medical, health insurance, and administrative records. Cardiovascular death was defined as International Classification of Diseases, Tenth Revision (ICD‐10) code I00 to I99.

Statistical Analysis

We divided all the included participants into 4 age groups (35–44, 45–54, 55–64, and 65–75 years). We described the characteristics of the study participants as mean±SD for continuous variables and frequencies with percentages for categorical variables.

Accounting for death from other causes as a competing risk, we developed the Fine‐Gray competing risk models to estimate the risk of cardiovascular death. We calculated hazard ratios (HRs) and 95% CIs per SD increment of CIMT and abnormal CIMT with adjustment sets as follows: (1) age, sex, current smoking status, systolic blood pressure, low‐density lipoprotein cholesterol, and diabetes in model 1; (2) additionally adjusted for the use of antihypertensive or lipid‐lowering or antidiabetic medications in model 2. The above potential covariates were selected on the basis of the published literature and clinical significance. We further tested the interaction between age and CIMT, and estimated the age‐specific HRs based on age‐specific SD of CIMT and age‐ and sex‐specific 75th percentile of CIMT. We also examined the interaction between sex and CIMT. Because the P value for interaction term was >0.05, we did not construct sex‐specific models. And because the baseline blood glucose data were not available to define diabetes in this study, we fitted models after excluding diabetes from the aforementioned models to evaluate the influence of the underestimated rate of diabetes on the association between CIMT and cardiovascular death. Moreover, we calculated the 5‐year cardiovascular death risk with adjustment sets as covariates of model 1 for each age and CIMT group.

Furthermore, we calculated the difference of the Harrell's concordance index (ΔC index) and the net reclassification improvement index (NRI) to evaluate the improvement in risk prediction per SD increment of CIMT for cardiovascular death based on the established risk factors. ²⁶ , ²⁷ Two reference models were constructed on the basis of the 2 sets of established risk factors. One was Framingham risk factors (ie, age, sex, current smoking status, systolic blood pressure, total cholesterol, high‐density lipoprotein cholesterol, diabetes, and antihypertensive medication). ⁸ , ²⁸ Another was Prediction for Atherosclerotic Cardiovascular Disease Risk in China risk factors (ie, age, sex, current smoking status, systolic blood pressure, total cholesterol, high‐density lipoprotein cholesterol, diabetes, antihypertensive medication, waist circumference, geographic region, urbanization, and family history of premature atherosclerotic cardiovascular disease). ²⁰ We used risk factors instead of the scores as their algorithms, which were for predicting cardiovascular events rather than cardiovascular death. We performed bootstrapping to obtain 95% CIs for the ΔC indexes and NRIs and to adjust for the overoptimism (bootstrapping numbers=100). We calculated the ΔC indexes and NRIs for the median and 5‐year follow‐up because these were between the 50th and 75th percentiles of follow‐up duration. ²⁷

We imputed the missing data using the Markov Chain Monte Carlo method for continuous variables and the mode for categorical variables (the highest percentage of missing variable, 7.2%). ²⁹ As the mortality data were available up to December 31, 2021, we censored the follow‐up at this date or the date of death, whichever occurred first.

We considered P<0.05 as statistically significant without addressing multiple testing, given that our report focused on the estimation and modeling. All analyses were conducted with SAS 9.4 (SAS Institute Inc, Cary, NC).

RESULTS

Baseline Characteristics and Cardiovascular Mortality by Age Group

Among the 369 478 participants included, the average age was 57.81±9.44 years, and 58.63% were women. Overall, 36.17% were living in urban areas, 53.10% were living in northern China, 21.22% were current smokers, 10.99% had diabetes, and 10.62% had a family history of premature atherosclerotic cardiovascular disease. Among all included participants, 33.98% were taking antihypertensive medications, 3.33% were taking lipid‐lowering medications, and 8.60% were taking antidiabetic medications. The levels of systolic blood pressure and low‐density lipoprotein cholesterol and the proportions of diabetes and medication use were higher in the older groups, whereas the prevalence of smoking was higher in the younger groups.

In the overall study population, the average and SD of CIMT_mean and maximum CIMT (CIMT_max) were 0.76±0.16 and 0.91±0.25 mm, respectively. Both CIMT_mean and CIMT_max increased with age. The CIMT_mean increased from 0.65±0.13 mm in the 35 to 44 years age group to 0.83±0.17 mm in the 65 to 75 years age group. The CIMT_max increased from 0.76±0.18 mm in the 35 to 44 years age group to 1.00±0.26 mm in the 65 to 75 years age group (Table 1). The age‐ and sex‐specific 75th percentile of both CIMT_mean and CIMT_max increased with age, and the values were higher in the men than women (Table S1).

Table 1.

Baseline Characteristics of Participants

Characteristic	Overall (N=369 478)	Age groups, y
		35–44 (n=35 462)	45–54 (n=103 170)	55–64 (n=128 537)	65–75 (n=102 309)
Age, y	57.81 (9.44)	40.68 (2.63)	50.10 (2.81)	59.81 (2.88)	69.01 (3.01)
Female sex, n (%)	216 615 (58.63)	18 166 (51.23)	61 282 (59.40)	78 854 (61.35)	58 313 (57.00)
Urban location, n (%)	133 643 (36.17)	14 206 (40.06)	37 562 (36.41)	46 967 (36.54)	34 908 (34.12)
Northern China, n (%)	196 177 (53.10)	20 560 (57.98)	56 817 (55.07)	68 308 (53.14)	50 492 (49.35)
Current smoking, n (%)	78 400 (21.22)	9415 (26.55)	23 296 (22.58)	26 300 (20.46)	19 389 (18.95)
Waist circumference, cm	85.96 (9.94)	85.03 (10.85)	85.90 (9.95)	86.16 (9.67)	86.10 (9.90)
SBP, mm Hg	154.44 (23.66)	139.70 (24.27)	149.39 (23.80)	156.56 (22.63)	161.96 (20.97)
Total cholesterol, mmol/L	4.93 (1.33)	4.51 (1.26)	4.88 (1.34)	5.05 (1.34)	4.98 (1.30)
LDL‐C, mmol/L	2.76 (1.11)	2.46 (1.05)	2.73 (1.12)	2.85 (1.13)	2.78 (1.08)
HDL‐C, mmol/L	1.40 (0.45)	1.26 (0.45)	1.35 (0.45)	1.42 (0.44)	1.46 (0.45)
Triglycerides, mmol/L	1.72 (0.97)	1.77 (1.07)	1.78 (1.02)	1.73 (0.95)	1.64 (0.89)
Diabetes, n (%)	40 605 (10.99)	1451 (4.09)	7660 (7.42)	15 642 (12.17)	15 852 (15.49)
Antihypertensive medication, n (%)	125 555 (33.98)	4830 (13.62)	27 417 (26.57)	47 560 (37.00)	45 748 (44.72)
Lipid‐lowering medication, n (%)	12 294 (3.33)	595 (1.68)	2561 (2.48)	4786 (3.72)	4352 (4.25)
Antidiabetic medication, n (%)	31 767 (8.60)	1007 (2.84)	5690 (5.52)	12 233 (9.52)	12 837 (12.55)
Family history of premature ASCVD, n (%)	39 254 (10.62)	4687 (13.22)	13 126 (12.72)	13 758 (10.70)	7683 (7.51)
CIMTmean, mm	0.76 (0.16)	0.65 (0.13)	0.71 (0.14)	0.78 (0.16)	0.83 (0.17)
CIMTmax, mm	0.91 (0.25)	0.76 (0.18)	0.84 (0.21)	0.93 (0.24)	1.00 (0.26)
Cardiovascular mortality, n (%)	4723 (1.28)	95 (0.27)	524 (0.51)	1360 (1.06)	2744 (2.68)

Data are given as mean (SD) unless otherwise indicated. ASCVD indicates atherosclerotic cardiovascular disease; CIMT_max, maximum carotid intima‐media thickness; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; and SBP, systolic blood pressure.

During the median follow‐up of 4.7 years (interquartile range, 3.5–5.7 years), 4723 participants died of cardiovascular disease (event rate, 1.28%). The rates of cardiovascular mortality increased with age, from 0.27% in the 35 to 44 years age group to 2.68% in the 65 to 75 years age group (Table 1).

Associations of Carotid Measurements With the Risk of Cardiovascular Death

In the overall study population and different age groups, both CIMT_mean and CIMT_max were significantly associated with the risk of cardiovascular death after adjustment for covariates. The interactions between both age‐specific SD of CIMTs and age groups were significant (P for interaction <0.01). The adjusted HRs per SD of CIMT_mean decreased with age, from 1.27 (95% CI, 1.17–1.37) in the 35 to 44 years age group to 1.14 (95% CI, 1.10–1.19) in the 65 to 75 years age group. As for CIMT_max, the HRs per SD decreased with age, from 1.16 (95% CI, 1.08–1.25) in the 35 to 44 years age group to 1.12 (95% CI, 1.09–1.15) in the 65 to 75 years age group in model 1. Moreover, the HRs of both CIMTs remained significant after adjustment for the use of medications in model 2 (Figure 1). And after excluding diabetes from the aforementioned models, the point estimates and 95% CI of HR were consistent with the prior results of the models, including diabetes. And the age‐related relationship between CIMT and cardiovascular death still existed (Figure S2). Furthermore, the HRs of abnormal CIMTs were also decreased with age (P for interaction <0.05). The adjusted HRs of abnormal CIMT_mean decreased from 1.69 (95% CI, 1.11–2.55) in the 35 to 44 years age group to 1.30 (95% CI, 1.20–1.42) in the 65 to 75 years age group. The adjusted HRs of abnormal CIMT_max decreased from 2.30 (95% CI, 1.53–3.45) in the 35 to 44 years age group to 1.31 (95% CI, 1.21–1.41) in the 65 to 75 years age group in model 1. Also, the HRs of both abnormal CIMTs remained significant after adjustment for the use of medications in model 2 (Table S2).

Figure 1. The hazard ratio (HR) per SD increment of carotid intima‐media thickness (CIMT) for cardiovascular death in the overall population and different age groups.

Model 1, adjusted for age, sex, current smoking status, systolic blood pressure, low‐density lipoprotein cholesterol, and diabetes. Model 2, adjusted for variables in model 1 and the use of antihypertensive, lipid‐lowering, and antidiabetic medications. The P value for interaction term between CIMT and age groups <0.01 in both models 1 and 2. CIMT_max indicates maximum CIMT; CIMT_mean, mean CIMT.

Adjusted 5‐Year Risk of Cardiovascular Death in Different Age and CIMT Groups

In all age groups, people with ≥75th percentile CIMT had higher adjusted 5‐year risk compared with those with normal CIMT. The adjusted 5‐year risk of cardiovascular death in people with abnormal CIMT_mean was 0.51%, 0.92%, 1.68%, and 3.99% for those aged 35 to 44, 45 to 54, 55 to 64, and 65 to 75 years, respectively. The absolute differences in the adjusted 5‐year risk between <75th percentile CIMT groups and ≥75th percentile CIMT groups increased with age, but the relative differences in the adjusted 5‐year risk decreased with age. The adjusted 5‐year risk of ≥75th percentile CIMT groups was around twice as high as <75th percentile CIMT groups in the 35 to 44 and 45 to 54 years age groups. However, the relative differences in the adjusted 5‐year risk were lower in the 55 to 64 and 65 to 75 years age groups (Table 2).

Table 2.

Adjusted 5‐Year Risk for Cardiovascular Death by Age and CIMT Groups

Population age, y	Adjusted 5‐y risk, %*
	CIMTmean <75th percentile	CIMTmean ≥75th percentile	CIMTmax <75th percentile	CIMTmax ≥75th percentile
35–44	0.24	0.51	0.20	0.51
45–54	0.46	0.92	0.45	0.82
55–64	1.05	1.68	1.06	1.53
65–75	2.77	3.99	2.75	3.81

CIMT indicates carotid intima‐media thickness; and CIMT_max, maximum CIMT; CIMT_mean, mean CIMT.

^*Adjusted for age, sex, current smoking status, systolic blood pressure, low‐density lipoprotein cholesterol, and diabetes.

Improvement of Discrimination and Reclassification by Different Carotid Measurements

After adding CIMT into the Framingham risk factors, the NRIs for CIMT_mean were significant, except in the 65 to 75 years age group, and attenuated with age, from 22.60% (95% CI, 15.56%–29.64%) in the 35 to 44 years age group to 7.00% (95% CI, −6.82% to 20.83%) in the 65 to 75 years age group, resulting in nonsignificance in the overall population. Similarly, the NRIs for CIMT_max were significant, except in the 55 to 64 and 65 to 75 years age groups, and attenuated with age, from 21.41% (95% CI, 12.16%–30.66%) in the 35 to 44 years age group to 8.61% (95% CI, −5.78% to 23.00%) in the 65 to 75 years age group, resulting in nonsignificance in the overall population. However, the ΔC indexes of both CIMTs showed similar trends with age, which, however, were not significant except in the 65 to 75 years age group (Figure 2).

Figure 2. The improvement in discrimination and reclassification for the added value of carotid intima‐media thickness (CIMT) based on the Framingham risk factors at the median follow‐up.

ΔC index indicates difference of the Harrell's concordance index; CIMT_max, maximum CIMT; CIMT_mean, mean CIMT and NRI, net reclassification improvement index.

The findings mentioned were also supported by adding CIMT into the Prediction for Atherosclerotic Cardiovascular Disease Risk in China risk factors. Significant NRIs for CIMT_mean were obtained in the 35 to 44 years (16.74% [95% CI, 10.39%–23.09%]) and 45 to 54 years (19.41% [95% CI, 7.96%–30.86%]) age groups, which were higher than those in the 55 to 64 and 65 to 75 years age groups. And the NRIs for CIMT_mean were not significant in the 55 to 64 and 65 to 75 years age groups. The NRIs for CIMT_max attenuated with age, from 22.46% (95% CI, 11.57%–33.35%) in the 35 to 44 years age group to 8.19% (95% CI, −6.30% to 22.69%) in the 65 to 75 years age group, and these were significant in the 2 younger groups, but not in the 2 older ones. The ΔC indexes of both CIMTs also showed decreasing trends with age, whereas they were not significant in all age groups (Figure 3).

Figure 3. The improvement in discrimination and reclassification for the added value of carotid intima‐media thickness (CIMT) based on the Prediction for Atherosclerotic Cardiovascular Disease Risk in China risk factors at the median follow‐up.

ΔC index indicates difference of the Harrell's concordance index; CIMT_max, maximum CIMT; CIMT_mean, mean CIMT and NRI, net reclassification improvement index.

The aforementioned ΔC indexes and NRIs were calculated for the median follow‐up. Similarly, the results of 5‐year follow‐up showed that the discrimination and reclassification abilities of both CIMTs were attenuated with age (Figures S3 and S4).

DISCUSSION

Using the data from a prospective cohort including nearly 0.37 million community residents without prior cardiovascular disease, we first demonstrated that the association between CIMT and cardiovascular death was attenuated with age. In addition to the established risk factors, CIMT can improve prediction for cardiovascular death in people aged <55 years rather than in older ones. These findings indicate that among the young people, CIMT measurements may help improve their cardiovascular risk assessment.

The association between CIMT and cardiovascular death was attenuated with age. Although many studies demonstrated thickened CIMT was associated with increased cardiovascular risk, none reported the age‐related trends of this association. ⁷ , ³⁰ , ³¹ , ³² Lorenz et al reported that the association between CIMT and cardiovascular risk was stronger in people aged <50 years. ¹⁶ Eigenbrodt et al found similar phenomena in men (ie, a stronger association in men aged <55 years). ¹⁷ But both studies only analyzed 2 broad age groups. ¹⁶ , ¹⁷ On the basis of the largest cohort studying CIMT, we were able to divide the study population into 4 groups, and demonstrated a clear age‐related trend. The strength of the associations between CIMT and cardiovascular death decreased with age. This trend could be explained by the differences in the levels of cardiovascular risk among different age groups. The average level of cardiovascular risk of the younger people was generally lower than that of the older ones, which may lead to higher relative risk of the increased CIMT in the younger people compared with that in the older ones. Currently, the pathophysiological mechanism of the age‐related differences in the increased CIMT prediction is still not clear.

We further found that CIMT could improve the prediction for cardiovascular death in people aged <55 years rather than in older ones. A meta‐analysis evaluated the added predictive value of CIMT, which was too small to be of clinical importance in the general population aged 35 to 75 years. ⁷ We obtained consistent results in the overall population. However, after the stratification by age, we demonstrated the significant cardiovascular risk predictive value of CIMT on top of the traditional risk factors in the younger population rather than the older ones. Without stratification by age, the results would be dominated by the older people, because they have a higher event rate than the younger ones, which would obscure the effects of CIMT in the young people.

Our findings may facilitate more targeted strategies for preventing cardiovascular disease in the young and middle‐aged populations. Cardiovascular disease was the leading cause of death in people aged 35 to 54 years worldwide, and the proportion of cardiovascular death has been increasing during the past decades. ³³ However, this population has been neglected because of their lower absolute cardiovascular risk than the elders. ³⁴ On the basis of traditional risk factors, CIMT can further refine the predication of cardiovascular risk. With the use of age‐ and sex‐specific 75th percentile as the cutoff value, CIMT can clearly identify the people at the higher relative risk in people aged <55 years. Particularly, measuring CIMT is simple, fast, and noninvasive, which is applicable in both developed and less developed countries.

This study has the following limitations. First, the follow‐up period was relatively short, which limited our statistical power to estimate the predictive abilities for cardiovascular death. However, we still observed consistent trends with age in all analyses. Second, the study covered people aged 35 to 75 years, and therefore we were unable to evaluate the predictive ability of CIMT out of the range, which needs to be explored in future studies. Third, we conducted the study only in the Chinese population. CIMT varies across race populations. ²⁵ The uses of antihypertensive and lipid‐lowering medications were lower than in populations from the United States and Europe. ³⁵ , ³⁶ Therefore, the results should be interpreted with cautions when applying to other races. Fourth, the outcome of this study was cardiovascular mortality rather than incident cardiovascular events, so the findings need to be validated to further illustrate the role of CIMT in cardiovascular risk assessment. Fifth, the study sample is not a nationally or regionally random sample, which would limit the representativeness of the results. Nevertheless, our population was widely distributed throughout China, which could ensure a good generalizability of the findings.

In conclusion, the association between CIMT and cardiovascular death is attenuated with age. Measuring CIMT could improve the prediction of cardiovascular death based on the established risk factors in people aged <55 years. This is particularly important for dealing with the challenge of increasing burden of cardiovascular disease in young and middle‐aged adults.

Sources of Funding

This work was supported by the China Academy of Chinese Medical Sciences Innovation Fund for Medical Science (2021‐I2M‐1‐009 and 2021‐I2M‐1‐011), the National Key Research and Development Program from the Ministry of Science and Technology of China (2018YFC1311205 and 2020YFC2004703), and the Major Public Health Service Project from the Ministry of Finance and National Health and Family Planning Commission of China.

Disclosures

Dr Jing Li reported receiving research grants, through Fuwai Hospital, from the Chinese government for work to improve the management of hypertension and blood lipids, to improve care quality and patient outcomes of cardiovascular disease, and to improve care for COVID‐19 infection; receiving research agreements, through the National Center for Cardiovascular Diseases and Fuwai Hospital, from Amgen for a multicenter clinical trial assessing the efficacy and safety of omecamtiv mecarbil and for dyslipidemic patient registration; receiving a research agreement, through Fuwai Hospital, from Sanofi for a multicenter clinical trial on the effects of sotagliflozin; receiving a research agreement, through Fuwai Hospital, with the University of Oxford for a multicenter clinical trial of empagliflozin; receiving a research agreement, through the National Center for Cardiovascular Diseases, from AstraZeneca for clinical research method training outside the submitted work; and receiving a research agreement, through the National Center for Cardiovascular Diseases, from Lilly for physician training outside the submitted work. The remaining authors have no disclosures to report.

Supporting information

Tables S1–S2

Figures S1–S4