Skip to main content
Download PDF
Download ePub

Long-term trajectories of apolipoprotein A1 and major adverse cardiovascular events and mortality in a community cohort

Lipids in Health and Disease volume 24, Article number: 137 (2025) Cite this article

Abstract

Background

Apolipoprotein A1 (ApoA1) is a major component of high-density lipoprotein cholesterol and plays a critical role in reverse cholesterol transport. Dynamic changes in ApoA1 levels may be associated with major adverse cardiovascular events. This study aimed to evaluate the impact of ApoA1 trajectories over three early assessments.

Methods

Participants in the Chin-Shan Community Cardiovascular Cohort with dyslipidemia and receiving three early ApoA1 assessments were enrolled. Group-based multivariate trajectory modeling was used to classify participants into distinct trajectories after multivariable adjustment. The follow-up duration was from April 1990 to August 2022, and the long-term outcomes of major adverse cardiovascular events (MACE) and death outcomes were evaluated.

Results

A total of 1,080 participants were included (median [interquartile range] age 66.14 [57.93–75.04] years, 43.2% males). Participants were classified into four ApoA1 trajectories: Trajectory 1 (low-level persistence pattern); Trajectory 2 (fall-then-rise pattern); Trajectory 3 (rise-then-fall pattern); and Trajectory 4 (elevated stable pattern). The cumulative incidence of MACE was ranked as Trajectory 4 (7.9%) < Trajectory 2 (9.3%) < Trajectory 3 (9.4%) < Trajectory 1 (12.7%). Comparing to Trajectory 4, both Trajectory 1 and Trajectory 2 had significantly higher risks of MACE (Trajectory 1: hazard ratio [HR] = 2.06, 95% confidence interval [CI] 1.10–3.86; Trajectory 2: HR = 2.38, 95% CI 1.03–5.48). For cardiovascular death, similar results were present. There were no significant differences in composite outcome, all-cause death, non-cardiovascular death across ApoA1 trajectories.

Conclusion

The trajectory changes of ApoAI levels significantly influences MACE risk during long-term follow-up, particularly in the low-stable and J-shaped trajectories. Dynamic monitoring of ApoAI may serve as a valuable tool for early risk stratification in high-risk populations, facilitating more individualised interventions.

Introduction

Cardiovascular disease (CVD) remains one of the leading causes of death, disability, and excess health spending worldwide, with the number of cardiovascular deaths steadily increasing from 12.1 million in 1990 to 18.6 million in 2019 [1]. Despite advancements in medical technology and treatments, the burden of CVD continues to grow with population ageing, lifestyle changes and the rising prevalence of metabolic diseases. Between 2025 and 2050, the crude mortality rate for CVD is projected to rise by more than 70%, while the prevalence of CVD is expected to increase by approximately 90% [2]. Therefore, it is crucial to explore risk factors for the incidence and death of CVD in order to effectively prevent and reduce their global burden [3].

Apolipoprotein A1 (ApoA1) is the major protein component of high-density lipoprotein (HDL), which is formed in the liver and intestinal endoplasmic reticulum [4]. ApoA1 plays a critical role in reverse cholesterol transport and HDL metabolism, which helps to transport excess cholesterol from peripheral tissues to the liver for excretion, reducing the formation of atherosclerotic plaques [5]. In patients without CVD at baseline and without statins, the risk of developing composite non/fatal/fatal CVD events was reduced by approximately 19% per one standard deviation (SD) of elevated ApoA1 [6]. Additionally, elevated ApoA1 significantly reduces the risk of cardiovascular death [7]. For specific atherosclerotic CVD, a Mendelian randomization analysis has showed that elevated ApoA1 are associated with a reduced risk of stroke [8]. However, another cross-sectional study showed no significant correlation between ApoA1 and obstructive CAD (over 70% of lesions in at least one vessel) [9]. These findings emphasize the inconsistency of the associations between ApoA1 and CVD and necessitate further investigations.

One recent study has shown the importance of assessing dynamic changes in lipid levels for disease risk prediction [10]. Nevertheless, most of the existing studies have been limited to single baseline measurements of ApoA1 levels, failing to assess its dynamic changes over time. The dynamic trajectory of ApoA1 may better reflect its long-term impact on risk of major adverse cardiovascular events (MACE) than single time-point measurements. Therefore, this study aimed to investigate the relationship between dynamic trajectory of ApoA1 and MACE in a community based cohort study, with long term followup.

Methods

Study design and study population

The Chin-Shan Community Cardiovascular Cohort (CCCC) Study is a prospective, population-based cohort designed to investigate the incidence of CVD and associated risk factors in individuals aged 35 years and older. Participants were recruited between 1990 and 1991 from the Chin-Shan community in New Taipei City, Taiwan, using household registration lists. A total of 3,602 participants were enrolled in the original cohort [11,12,13].

For the current analysis, specific exclusion criteria were applied to refine the study population. Participants were excluded if they met any of the following conditions: (i) a history or incidence of coronary artery disease (CAD) or cerebrovascular accident (CVA) before the 2006 re-survey; (ii) failure to complete three ApoA1 tests prior to the 2006 re-survey; (iii) absence of dyslipidemia at baseline, defined as having any of the following: low-density lipoprotein cholesterol (LDL-C) ≥ 100 mg/dL, high-density lipoprotein cholesterol (HDL-C) < 40 mg/dL for males or < 50 mg/dL for females, triglycerides (TG) ≥ 150 mg/dL, or total cholesterol (TC) ≥ 200 mg/dL; and (iv) abnormal ApoA1 levels, defined as values outside the mean ± three times of SD.

The final analysis included a subset of 1,080 participants with complete medical records across three distinct time periods: 1990–1991, 1992–1993, and 1997–2006. These participants were selected based on the availability of sufficient longitudinal data for assessing cardiovascular outcomes. Ethical approval for the study was obtained from the National Taiwan University Hospital Institutional Review Board (IRB No: 201003001R). Written informed consent was obtained from each participant recruited in this study.

Data collection

During the baseline survey, all participants were individually interviewed using a structured questionnaire to collect comprehensive data on various factors, including sociodemographic characteristics, lifestyle habits, dietary patterns, and personal and family histories of diseases and hospitalisations. Trained medical students, supported by local community leaders, conducted door-to-door visits to invite participants to the study. Participants who consented to the study also underwent physical examinations and laboratory tests, conducted by physicians and medical students at a clinic established specifically for this research [11, 13].

A specialised cardiology clinic was set up at the Chin-Shan Community Health Center, staffed by a team comprising 20 senior medical students, two assistant nurses, 10 cardiologists, and local practitioners. The medical students received three months of intensive training, which covered essential communication skills, home visit protocols, basic medical knowledge, practice evaluations, history-taking techniques, physical examination skills, and guidance on completing the study questionnaire. Follow-up visits were conducted biennially to monitor participant health status.

For this analysis, covariates were derived from participant characteristics collected at the time of the third ApoA1 measurement in 2006, which served as the final ApoA1 assessment. As 2006 was established as the starting point for follow-up, the patient characteristics from that year were used as baseline variables to evaluate the trajectory of ApoA1 levels and their association with cardiovascular outcomes.

The covariates included in this study encompassed a range of demographic data such as age, sex, and body mass index (BMI), along with key comorbidities, including hypertension and diabetes mellitus (DM). Laboratory test results were also extracted, capturing biomarkers such as TG, HDL-C, LDL-C, glucose, uric acid, and blood cell counts (white and red blood cells). These covariates were essential for adjusting the statistical models and ensuring a comprehensive analysis of the relationship between ApoA1 levels and cardiovascular outcomes.

Blood sampling and analytical methods

Venous blood samples were collected after a 12-hour overnight fast and were immediately refrigerated. Within 6 h of collection, the samples were transported to National Taiwan University Hospital. Serum samples were stored at − 70 °C until analysis, ensuring stability of the biochemical markers [12, 13].

ApoA1 levels, along with other lipid parameters, were measured through standardized methods. ApoA1 concentrations were quantified using turbidimetric immunoassay techniques with commercially available kits (Sigma). This method ensures accurate measurement of ApoA1, providing critical data for evaluating cardiovascular risk profiles in the study population. Total cholesterol and triglyceride levels were determined using enzymatic assays (Merck 14354 and 14366, respectively). For HDL-C, levels were measured in the supernatant after precipitation of specimens using magnesium chloride phosphotungstate reagents (Merck 14993). LDL-C concentrations were calculated by subtracting the cholesterol in the supernatant from the total cholesterol, based on the precipitation method (Merck 14992).

Follow-up strategy and outcome confirmation

The follow-up strategy involved a comprehensive approach to ensure accurate identification and verification of mortality events. A dedicated outcome adjudication committee, composed of two senior physicians with expertise in clinical medicine, was responsible for reviewing and confirming all study endpoints. The committee’s adjudication process involved cross-referencing multiple data sources, including: (I) Taiwan’s national death registry, with classification of causes of death based on ICD-9 codes; (II) Hospital medical records; (III) Death certificates from local statistical offices; (IV) Bi-annual household follow-up visits and family interviews [11, 13].

Regular follow-up occurred every two years, with household visits conducted to track participant status. In addition, death certificates were reviewed monthly at the local vital statistics office to ensure the timely recording of mortality events. To determine the cause of death and other major health events, cardiologists held blinded case conferences to discuss and verify diagnoses. If disagreements arose, a neurologist was consulted to help reach a consensus. When consensus was not achieved, the diagnosis was classified as unclassified. Diagnostic consistency was maintained through regular committee discussions and cross-checking with hospital records.

In this study, outcomes were defined as MACE, which included cardiovascular death, CAD, and CVA. The composite outcome encompassed all-cause mortality, CAD, and CVA. CAD in our study encompassed nonfatal myocardial infarction, unstable angina requiring hospitalisation, and revascularisation procedures such as percutaneous coronary intervention or coronary artery bypass grafting. CVA was defined as a neurological deficit of vascular origin persisting for more than 24 h, with confirmation through neuroimaging studies. This identification of CAD/CVA aligns with previous studies on cardiovascular event adjudication [14]. All-cause mortality was defined as death due to any cause, while cardiovascular death was specifically attributed to causes such as heart failure or myocardial infarction. Non-cardiovascular deaths included deaths from other causes, such as cancer or infections.

A review committee was also established to ensure the accuracy for diagnoses of major cardiovascular events. All preliminary diagnoses, medical records, death certificates, and interviews with relatives were presented to this committee for verification. Final confirmation of stroke cases was conducted by internists, with regular discussions to maintain diagnostic consistency across the study. Outcomes were tracked till August 2022.

Statistical analysis

In this study, we chose group-based trajectory modelling using ‘traj’ plugin in STATA to identify specific trajectories of ApoA1, which has been previously applied in studies of the medical field [15, 16]. This approach is suitable for revealing heterogeneity in longitudinal data and understanding the relationship between dynamic changes in biomarkers and clinical outcomes. To select the best-fitting trajectory model, this study used a combination of criteria [17]. First, the absolute values of the Bayesian Information Criterion (BIC) values for the models were used to compare the goodness-of-fit of the models for the different trajectories (groups 2, 3, 4, and 5) of first-order/second-order/third-order equations, selecting the model with the lowest BIC value. Second, to ensure that the trajectories were statistically significant, the sample proportion for each group had to be at least 5%. Third, the average posterior probability (AveP%) of each trajectory is required to be greater than 70% to ensure the reliability of correct classification. Fourth, the odds of correct classification (OCC) needs to be greater than 5 to further enhance the accuracy of classification. Finally, the visual display of trajectories ensures that different trajectories are clearly distinguished morphologically. Therefore, the best model is finally confirmed by these criteria.

The proportions of missing values of the included covariates were less than 5% (Supplementary Table S1). Therefore, to avoid losing samples, we interpolated the missing values using “miceforest” in Python (number of interpolations was 20). Continuous variables were expressed using mean (SD) or median (interquartile range [IQR]), depending on the distribution of variable, and compared among trajectories using one-way analysis of variance or Kruskal-Wallis H Test tests. Categorical variables were expressed as number and percentage and compared among trajectories by Chi-square or Fisher’s exact tests.

We assessed the incidence of cumulative composite outcomes and MACE across trajectories using Kaplan-Meier analyses. The incidences of outcomes were compared among trajectories by generating survival curves, testing the differences across trajectories over the follow-up period with Log-rank tests.

Next, we used Cox proportional hazards model to calculate hazard ratios (HR) and 95% confidence intervals (CI) for evaluating the risk of each outcome event among the trajectories, and we selected the trajectory with the lowest incidence of each outcome as the reference group. Cox proportional hazards models were adjusted for age, sex, BMI, comorbidities (hypertension, diabetes mellitus) and biomarkers (glucose, uric acid, white blood cell, red blood cell, TG, and LDL-C). Additionally, follow-up for study outcomes began after the third ApoA1 measurement. Therefore, the trajectory groups reflect longitudinal patterns prior to the start of follow-up, and baseline ApoA1 levels at third time point were not included as an independent covariate in the adjusted model to avoid overadjustment. The proportional hazards assumptions were checked using the Schoenfeld residuals.

Furthermore, subgroup analyses were performed based on sex (male and female), age (≥ 65 years and < 65 years), BMI (≥ 23 kg/m2 and < 23 kg/m2), and DM status (No DM and DM). Additionally, to compare differences in the performance of dynamic changes of ApoA1 versus ApoA1 levels measured at a single time point on the risk stratification of study outcomes. We also grouped the study subjects by quartiles of last ApoA1 measurement, and performed Cox proportional hazards model analyses to assess the relationships of quartiles and the risk of study outcomes.

All statistical analyses in this study were performed using STATA version 17.0 (StataCorp LLC, College Station, Texas, USA), R version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria), and Python version 3.11.1 (Python Software Foundation, Wilmington, Delaware, USA). A two-sided P-value of less than 0.05 was considered statistically significant.

Results

Baseline characteristics of the study population

A total of 1080 participants were included in this study (Supplementary Figure S1). The median age was 66.14 years (IQR: 57.93–75.04), and 467 (43.2%) were male. The median BMI was 24.33 kg/m² (IQR: 22.28–26.44). The prevalence of hypertension and DM was 23.6% and 33.7%, respectively. Laboratory tests showed median TG levels of 116.00 mg/dL (IQR: 81.00–169.00), TC of 202.00 mg/dL (IQR: 178.00–230.00), LDL-C of 127.00 mg/dL (IQR: 101.00–149.00), and HDL-C of 43.00 mg/dL (IQR: 36.00–52.00). Further details of the baseline characteristics for the whole cohort are presented in Table 1.

Table 1 Baseline characteristics according to trajectory groups
Full size table

Trajectory modeling and classification

Supplementary Table S2 shows the metrics for evaluating the goodness-of-fit of different group-based trajectory models. In selecting the optimal model, the four-trajectory model with a second-order equation had the lowest absolute BIC value (611.3). The average posterior probability (AvePP) for each group exceeded 80%, the proportion of participants in each trajectory was over 10%, and the OCC for each trajectory was greater than 5. Therefore, the models with four trajectories was finally selected as the best-fitting model. There were 4 ApoA1 trajectories: Trajectory 1 (low-level persistence pattern), Trajectory 2 (fall-then-rise pattern), Trajectory 3 (rise-then-fall pattern), and Trajectory 4 (elevated stable pattern) [Figure 1]. The four ApoA1 trajectories identified were: Trajectory 1: 504 (46.7%); Trajectory 2: 118 (10.9%); Trajectory 3: 307 (28.4%); and Trajectory 4: 151 (14.0%). The median (IQR) age of all participants was 66.14 years (57.93, 75.04), with Trajectory 2 having the lowest age (median age: 60.95 years, IQR: 55.16–67.86) [P < 0.001]. Males for all individuals were 467 (43.2%), with lower proportions of males in Trajectory 2 (33.1%) and Trajectory 4 (33.1%) [P < 0.001]. Detailed characteristics are shown in Table 1.

Fig. 1
figure 1

Longitudinal trajectories of apolipoprotein A1

Full size image

Study outcomes across ApoA1 trajectories

Within median follow-up time (IQR) of 16.36 years (10.14–17.31), there were 447 (41.4%) cases of the composite outcome in the overall population, with 413 (38.2%) all-cause deaths, 82 (7.6%) cardiovascular deaths, 331 (30.6%) non-cardiovascular deaths, and 116 (10.7%) cases of MACE. Overall, there were differences in outcome events across trajectories (Fig. 2 & Supplementary Table S3). Specifically, Trajectory 2 had the lowest incidences of composite outcome (27.1%), all-cause death (24.6%) and non-cardiovascular death (17.8%). Besides, Trajectory 4 had the lowest incidences of cardiovascular death (6.0%), MACE (7.9%), and CAD/CVA (2.0%).

Fig. 2
figure 2

Outcome distribution by longitudinal trajectories of apolipoprotein A1. CAD: coronary artery disease; CVA, cerebrovascular accident; MACE, major adverse cardiovascular events

Full size image

The results of the Kaplan-Meier survival curves showed differences in study outcomes across trajectories during follow-up (Fig. 3). For MACE and cardiovascular death, the cumulative risk of events decreased in the order of Trajectory 2 > Trajectory 1 > Trajectory 3 > Trajectory 4 (Log-rank P < 0.05). However, for the composite outcome, all-cause death and non-cardiovascular death, no significant differences in survival probabilities were observed accross trajectories (Log-rank P > 0.05).

Fig. 3
figure 3

Kaplan-meier curves for study outcomes across longitudinal trajectories of apolipoprotein A1

Full size image

Associations of ApoA1 trajectories and study outcomes

Figure 4 presents the results of the Cox proportional hazards models fully adjusted for confounders. Given that Trajectory 2 exhibited the lowest risk for composite outcomes, all-cause death, and non-cardiovascular death, it served as the reference group in these models. For cardiovascular death and MACE, Trajectory 4 was used as the reference group, as it had the lowest incidence of MACE and similar cardiovascular death rates compared with Trajectory 3. For MACE and cardiovascular death, compared with Trajectory 4, Trajectory 1 had significantly increased risk of MACE (HR = 2.06, 95% CI: 1.10–3.86) and cardiovascular death (HR = 2.28, 95% CI: 1.09–4.75). Similarly, Trajectory 2 had higher risk of MACE (HR = 2.38, 95% CI: 1.03–5.48) and cardiovascular death (HR = 2.68, 95% CI: 1.01–7.13).

Fig. 4
figure 4

The results of Cox proportional hazards model analyses for study outcomes across longitudinal trajectories of apolipoprotein A1. CI, confidence interval; HR, hazard ratio; MACE, major adverse cardiovascular events

Full size image

For the composite outcome, all-cause death and non-cardiovascular death, there were no significant differences accross these trajectories. For CAD/CVA, Cox proportional hazards model analyses were not performed because the number of events was only 3 in Trajectory 2 and Trajectory 4. In addition, adjusted ApoA1 per one SD revealed decreased risk of composite outcome (HR = 0.90, 95% CI: 0.82-1.00) and MACE (HR = 0.78, 95% CI: 0.63–0.96). However, quartiles grouped based on single time point levels of ApoA1 did not demonstrate significant associations with risk of all study outcomes (Supplementary Table S4).

Subgroup analyses

Supplementary Table S5 shows the results of subgroup analyses after stratification by sex, age, and BMI. Although the associations of the ApoA1 trajectories with MACE and cardiovascular death differed across subgroups from those in overall populations, no significant interactions were found (all P for interaction > 0.05).

Discussion

In this study, we analysed the dynamic changes of multiple ApoA1 measurements in the early life of 1080 older adults by using a group-based trajectory modelling approach, and identified four specific patterns of trajectories, including Trajectory 1 (low-level persistence pattern), Trajectory 2 (fall-then-rise pattern), Trajectory 3 (rise-then-fall pattern), and Trajectory 4 (elevated stable pattern). We further found that there were significant differences in the risk of MACE and cardiovascular death across ApoA1 trajectories. Trajectory 1 (low-level persistence pattern) and trajectory 2 (fall-then-rise pattern) significantly increased the risk of MACE and cardiovascular death compared with Trajectory 4 (elevated stable pattern).

These findings highlight the potential role of dynamic changes in ApoA1 levels in cardiovascular risk stratification. In contrast, the associations of single time-point ApoA1 measurements with MACE and cardiovascular death were unstable, further supporting the strength of the dynamic trajectory model and suggesting that assessing trends in ApoA1 over time more accurately reflects the risk of cardiovascular events than a single measurement.

There were significant differences in trends across ApoA1 trajectories, and the relevant mechanisms behind these may cause different patterns of cardiovascular risk. Trajectory 1 manifested as persistently low ApoA1 levels, which could lead to inefficient reverse cholesterol transport and imbalanced cholesterol homeostasis, increasing the risk of atherosclerosis formation [18]. In addition, low levels of ApoA1 are associated with inflammatory immune response and oxidative stress [19, 20]. Long-term low levels of ApoA1 may reflect higher degrees of systemic inflammation and oxidative stress, the progression of atherosclerosis and cardiovascular risk. In contrast, Trajectory 4 showed a sustained rise in ApoA1 levels and maintained them at high levels, indicating lower adverse cardiovascular risk. Trajectory 2 and Trajectory 3 revealed fluctuating levels of ApoA1 variations. Trajectory 2 exhibited a fall-then-rise pattern, although ApoA1 rose to a higher level in the later period, the low ApoA1 levels experienced in the early period might still have caused early vascular damage and increased long-term cardiovascular risk. In contrast, in Trajectory 3, despite the fact that levels of ApoA1 declined later in life, patients might have had better cardiovascular health earlier in life, and therefore did not exhibit significantly higher cardiovascular risk. However, the specific lipid metabolic mechanisms behind this still require further investigation to better understand their impact on cardiovascular prognosis.

What is the correlation between ApoA1 and cardiovascular risk? Churchill et al. found that adjusted ApoA1 per one SD showed a 10% decreased risk for the composite outcome (MACE and all-cause mortality) over a 15-year follow-up period in a community-based population (HR = 0.90, 95% CI: 0.81–0.99), but no significant association with stroke/MI risk [21]. Bodde and colleagues reported that adjusted ApoA1 per one SD revealed a 76% reduction in the risk of first-ever ST-segment elevation myocardial infarction in a general population (HR = 0.24, 95% CI: 0.18–0.33) [22]. However, in this study, we similarly observed that adjusted ApoA1 per one SD was associated with reduced risk of MACE, but quartile grouped ApoA1 was not, and that continuous or quartile grouped ApoA1 was not associated with cardiovascular death. These findings demonstrated that the correlations between single ApoA1 measures and cardiovascular risk was unreliable, and were also consistent with other studies not finding association between ApoA1 and cardiovascular risk in different populations. Indeed, Akiyama et al. showed no significant correlation between ApoA1 and MACE in patients undergoing percutaneous coronary intervention (P = 0.852) [23]. Moreover, Faaborg-Andersen et al. found a U-shaped relationship between ApoA1 and cardiovascular death during a median of 12.1 years of follow-up in more than 400,000 general participants [24]. Overall, these studies further demonstrated the limitations of single-measurement ApoA1 as a predictor of cardiovascular risk.

A previous review have highlighted important sex differences in CVD pathophysiology, risk stratification, and treatment response [25]. In particular, women may experience distinct pattern of lipid metabolism that could influence the relationship between ApoA1 and CVD outcomes. In our analysis, we did not detect a significant interaction between sex and ApoA1 trajectories with respect to risk of MACE and cardiovascular death. Nonetheless, the underrepresentation of women in cardiovascular clinical research, as reported in previous literature, underscores the need for further sex-specific investigations to confirm our findings and guide personalised preventive strategies.

Our study is the first to explore the dynamic pattern of ApoA1 and the relationship with cardiovascular risk, supporting the potential role of ApoA1 as a biomarker for cardiovascular risk stratification. While traditional lipid parameters (e.g., TC, HDL-C) remain the cornerstone of atherosclerotic cardiovascular disease risk prediction [26], integrating ApoA1—particularly its longitudinal patterns—may enhance current risk models. However, further validation is needed before ApoA1 trajectories can be incorporated into clinical practice. Future studies should focus on exploring the underlying mechanisms behind the ApoA1 trajectory pattern. This will not only facilitate deeper understanding of the role of dynamic changes in ApoA1 in the development and progression of atherosclerosis, but may also provide a theoretical basis for the development of new preventive and therapeutic strategies. Additionally, some interventions may have potential clinical applications for high-risk populations with low-level ApoA1 persistence trajectory pattern, including reconstituted human ApoA1 (CSL112) infusion therapy [27], and HDL mimetic (CER-001) infusion [28]. Moreover, adherence to a Mediterranean diet [29], increased intake of omega-3 fatty acids [30], and regular physical activity [31] are associated with higher ApoA1 levels. Further studies are warranted to determine whether increasing ApoA1 levels through these interventions can lead to a reduction in cardiovascular events.

Strength and limitations

Our study has several advantages. We used a large cohort of older adults with early-life dyslipidemia in Taiwan to explore the dynamics of ApoA1 at multiple time points in early life. By identifying different ApoA1 patterns, such as low-level persistence pattern and fall-then-rise pattern, we added valuable insights for long-term cardiovascular risk stratification. This approach allowed us to gain a more nuanced understanding of the ApoA1 dynamic trajectories (rather than single measurements) associated with cardiovascular risk.

However, there are several limitations to our study. First, our cohort was only from Taiwan, such geographic specificity limits the extrapolation of the findings to other regional cohorts or other ethnicities. Second, covariates included at baseline may have been affected by earlier study designs that were insufficiently comprehensively considered. For example, lifestyle factors such as smoking status, physical activity, and dietary habits were inadequately captured in the analyses, which may have caused residual confounding affecting the interpretation of our findings. Third, the outcome definition of MACE neglected to include other important atherosclerotic diseases (e.g., myocardial infarction), which may underestimate the real effects of the ApoA1 trajectories on cardiovascular risk. Fourth, since our study was observational, causality could not be definitively confirmed, but as only 3 MACE events occurred in the overall population within the first year of follow-up, reverse causality is less likely. Fifth, medication use, particularly lipid-lowering drugs, was not included in our analysis, which may affect ApoA1. However, we initially included patients with dyslipidaemia who would receive long-term lipid-lowering therapy in routine clinical practice. Sixth, while our findings suggest that early low ApoA1 levels may have a lasting impact on cardiovascular outcomes, the potential role of prolonged exposure to low ApoA1 levels remains unclear. Due to the limited number of ApoA1 measurements (three time points) in our cohort, we were unable to directly assess the cumulative burden or duration of low ApoA1 exposure. Future studies with more frequent and continuous longitudinal measurements of ApoA1 are needed to clarify whether prolonged exposure to low ApoA1 levels confers additional cardiovascular risk beyond that associated with early deficits. Finally, the wide CI for the results of Cox proportional hazards models analyses reflect the limited sample size, therefore larger cohorts are required to further validate the robustness and broad applicability of our findings.

Conclusion

This study is the first to characterize longitudinal ApoA1 trajectory phenotypes over the life course from middle to old age in a cohort diagnosed with dyslipidemia. We identified four trajectories: Trajectory 1 (low-level persistence pattern), Trajectory 2 (fall-then-rise pattern), Trajectory 3 (rise-then-fall pattern), and Trajectory 4 (elevated stable pattern). These trajectory phenotypes were differentially associated with MACE and cardiovascular death during later life. These findings underscore the identification of the ApoA1 trajectory as an independent predictor of cardiovascular events, providing new insights into the stratified management of cardiovascular risk.

Data availability

Data are available on reasonable request to the corresponding author.

Abbreviations

ApoA1:

Apolipoprotein A1;

BIC:

Bayesian Information Criterion;

BMI:

Body mass index;

CAD:

Coronary artery disease;

CCCC:

Chin-Shan Community; Cardiovascular Cohort;

CI:

Confidence interval;

CVA:

Cerebrovascular accident;

CVD:

Cardiovascular disease;

HDL-C:

High-density lipoprotein cholesterol;

HDL:

High-density lipoprotein;

HR:

Hazard ratios;

IQR:

Interquartile range;

LDL-C:

Low-density lipoprotein cholesterol;

MACE:

Major adverse cardiovascular events;

OCC:

Odds of correct classification;

SD:

Standard deviation;

TC:

Total cholesterol;

TG:

Triglycerides

References

  1. Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al. Global burden of cardiovascular diseases and risk factors, 1990–2019: update from the GBD 2019 study. J Am Coll Cardiol. 2020;76(25):2982–3021.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Chong B, Jayabaskaran J, Jauhari SM, Chan SP, Goh R, Kueh MTW et al. Global burden of cardiovascular diseases: projections from 2025 to 2050. Eur J Prev Cardiol. 2024;zwae281.

  3. Joynt Maddox KE, Elkind MSV, Aparicio HJ, Commodore-Mensah Y, de Ferranti SD, Dowd WN, et al. Forecasting the burden of cardiovascular disease and stroke in the united States through 2050-Prevalence of risk factors and disease. Circulation: a Presidential Advisory From the American Heart Association; 2024.

    Google Scholar 

  4. Alvarez-Sola G, Uriarte I, Latasa MU, Fernandez-Barrena MG, Urtasun R, Elizalde M, et al. Fibroblast growth factor 15/19 (FGF15/19) protects from diet-induced hepatic steatosis: development of an FGF19-based chimeric molecule to promote fatty liver regeneration. Gut. 2017;66(10):1818–28.

    Article  PubMed  Google Scholar 

  5. Mehta A, Shapiro MD. Apolipoproteins in vascular biology and atherosclerotic disease. Nat Reviews Cardiol. 2022;19(3):168–79.

    Article  Google Scholar 

  6. Welsh C, Celis-Morales CA, Brown R, Mackay DF, Lewsey J, Mark PB, et al. Comparison of conventional lipoprotein tests and apolipoproteins in the prediction of cardiovascular disease. Circulation. 2019;140(7):542–52.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Zhang J, Song X, Li Z, Xu H, Shu H, Li J, Zhang Y. Association of apolipoprotein levels with all-cause and cardiovascular mortality. Eur J Prev Cardiol. 2024;31(9):1183–94.

    Article  PubMed  Google Scholar 

  8. Liu Y, Wang W, Cui X, Lyu J, Xie Y. Exploring genetic associations of 3 types of risk factors with ischemic stroke: an integrated bioinformatics study. Stroke. 2024;55(6):1619–28.

    Article  PubMed  Google Scholar 

  9. Natarajan P, Collier TS, Jin Z, Lyass A, Li Y, Ibrahim NE, et al. Association of an HDL apolipoproteomic score with coronary atherosclerosis and cardiovascular death. J Am Coll Cardiol. 2019;73(17):2135–45.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Yang T, Zhao B, Pei D. Relationship between the ApoB/ApoA-I ratio trajectory and risk of type 2 diabetes in China: a retrospective cohort study. Endocrine. 2022;76(1):36–43.

    Article  PubMed  Google Scholar 

  11. Lee Y, Lin RS, Sung FC, Yang C, Chien K, Chen W, et al. Chin-Shan community cardiovascular cohort in Taiwan-baseline data and five-year follow-up morbidity and mortality. J Clin Epidemiol. 2000;53(8):838–46.

    Article  PubMed  Google Scholar 

  12. Chien KL, Sung FC, Hsu HC, Su TC, Lin RS, Lee YT. Apolipoprotein A-I and B and stroke events in a community-based cohort in Taiwan: report of the Chin-Shan community cardiovascular study. Stroke. 2002;33(1):39–44.

    Article  PubMed  Google Scholar 

  13. Chien KL. Mini-review of the Chin-Shan community cardiovascular cohort study in population health research in Taiwan. Acta Cardiol Sin. 2017;33(3):226–32.

    PubMed  PubMed Central  Google Scholar 

  14. Yang W-S, Chen Y-Y, Chen P-C, Hsu H-C, Su T-C, Lin H-J, et al. Association between plasma N-6 polyunsaturated fatty acids levels and the risk of cardiovascular disease in a community-based cohort study. Sci Rep. 2019;9(1):19298.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Schmidt JA, Woolpert KM, Hjorth CF, Farkas DK, Ejlertsen B, Cronin-Fenton D. Social characteristics and adherence to adjuvant endocrine therapy in premenopausal women with breast cancer. J Clin Oncol. 2024;42(28):3300–7.

    Article  PubMed  Google Scholar 

  16. Senior M, Pierce M, Taxiarchi VP, Garg S, Edge D, Newlove-Delgado T, et al. 5-year mental health outcomes for children and adolescents presenting with psychiatric symptoms to general practitioners in England: a retrospective cohort study. Lancet Psychiatry. 2024;11(4):274–84.

    Article  PubMed  Google Scholar 

  17. Nagin D. Group-based modeling of development. Harvard University Press; 2005.

  18. Andraski AB, Sacks FM, Aikawa M, Singh SA. Understanding HDL metabolism and biology through in vivo tracer kinetics. Arterioscler Thromb Vasc Biol. 2024;44(1):76–88.

    Article  Google Scholar 

  19. Busnelli M, Manzini S, Chiara M, Colombo A, Fontana F, Oleari R, et al. Aortic gene expression profiles show how ApoA-I levels modulate inflammation, lysosomal activity, and sphingolipid metabolism in murine atherosclerosis. Arterioscler Thromb Vasc Biol. 2021;41(2):651–67.

    Article  PubMed  Google Scholar 

  20. Morris G, Puri BK, Bortolasci CC, Carvalho A, Berk M, Walder K, et al. The role of high-density lipoprotein cholesterol, apolipoprotein A and paraoxonase-1 in the pathophysiology of neuroprogressive disorders. Neurosci Biobehav Rev. 2021;125:244–63.

    Article  PubMed  Google Scholar 

  21. Churchill RA, Gochanour BR, Scott CG, Vasile VC, Rodeheffer RJ, Meeusen JW, Jaffe AS. Association of cardiac biomarkers with long-term cardiovascular events in a community cohort. Biomarkers: Biochemical Indicators of Exposure, Response, and Susceptibility To Chemicals. 2024;29(4):161–70.

  22. Bodde MC, Hermans MPJ, Jukema JW, Schalij MJ, Lijfering WM, Rosendaal FR, et al. Apolipoproteins A1, B, and apoB/apoA1 ratio are associated with first ST-segment elevation myocardial infarction but not with recurrent events during long-term follow-up. Clin Res Cardiology: Official J German Cardiac Soc. 2019;108(5):520–38.

    Article  Google Scholar 

  23. Akiyama T, Ikegami R, Kubota N, Takano T, Yoneyama S, Okubo T et al. Serum apolipoprotein-A2 levels are a strong predictor of future cardiovascular events in patients undergoing percutaneous coronary intervention. Circulation Journal: Official J Japanese Circulation Soc. 2024;88(11):1770–1777.

  24. Faaborg-Andersen CC, Liu C, Subramaniyam V, Desai SR, Sun YV, Wilson PWF, et al. U-shaped relationship between apolipoprotein A1 levels and mortality risk in men and women. Eur J Prev Cardiol. 2023;30(4):293–304.

    Article  PubMed  PubMed Central  Google Scholar 

  25. Calabrò P, Niccoli G, Gragnano F, Grove EL, Vergallo R, Mikhailidis DP, et al. Are we ready for a gender-specific approach in interventional cardiology? Int J Cardiol. 2019;286:226–33.

    Article  PubMed  Google Scholar 

  26. Medina-Inojosa JR, Somers VK, Garcia M, Thomas RJ, Allison T, Chaudry R, et al. Performance of the ACC/AHA pooled cohort cardiovascular risk equations in clinical practice. J Am Coll Cardiol. 2023;82(15):1499–508.

    Article  PubMed  Google Scholar 

  27. Gibson CM, Duffy D, Korjian S, Bahit MC, Chi G, Alexander JH, et al. Apolipoprotein A1 infusions and cardiovascular outcomes after acute myocardial infarction. N Engl J Med. 2024;390(17):1560–71.

    Article  PubMed  Google Scholar 

  28. Nicholls SJ, Andrews J, Kastelein JJP, Merkely B, Nissen SE, Ray KK, et al. Effect of serial infusions of CER-001, a pre-β high-density lipoprotein mimetic, on coronary atherosclerosis in patients following acute coronary syndromes in the CER-001 atherosclerosis regression acute coronary syndrome trial: a randomized clinical trial. JAMA Cardiol. 2018;3(9):815–22.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Violante B, Gerbaudo L, Borretta G, Tassone F. Effects of extra virgin olive oil supplementation at two different low doses on lipid profile in mild hypercholesterolemic subjects: a randomised clinical trial. J Endocrinol Invest. 2009;32(10):794–6.

    Article  PubMed  Google Scholar 

  30. Merino J, Sala-Vila A, Kones R, Ferre R, Plana N, Girona J, et al. Increasing long-chain n-3PUFA consumption improves small peripheral artery function in patients at intermediate-high cardiovascular risk. J Nutr Biochem. 2014;25(6):642–6.

    Article  PubMed  Google Scholar 

  31. Rosenkilde M, Rygaard L, Nordby P, Nielsen LB, Stallknecht B. Exercise and weight loss effects on cardiovascular risk factors in overweight men. J Appl Physiol (1985). 2018;125(3):901–8.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

We thank the participants in the Chin-Shan community for their assistance in this study.

Funding

This work was supported by the Ministry of Science and Technology of Taiwan (MOST 110-2314-B-A49A-541-MY3); the TVGH (C19-027); the TVGH and NTUH (VN111-05); and the Research Foundation of Cardiovascular Medicine (110-02-006).

Author information

Author notes

  1. Kuo-Liong Chien, Yenn-Jiang Lin, Gregory Y. H. Lip and Shih-Ann Chen are joint senior authors.

Authors and Affiliations

  1. Liverpool Centre for Cardiovascular Science, University of Liverpool, Liverpool John Moores University and Liverpool Heart & Chest Hospital, Liverpool, UK

    Yang Chen & Gregory Y. H. Lip

  2. Department of Medical Research, Taichung Veterans General Hospital, No. 1650, Sec. 4, Taiwan Boulevard, Xitun District, Taichung, Taiwan

    Yun-Yu Chen, Fang-Yi Chen & Yu-Cheng Hsieh

  3. Cardiovascular Center, Taichung Veterans General Hospital, Taichung, Taiwan

    Yun-Yu Chen, Yenn-Jiang Lin, Yu-Cheng Hsieh & Shih-Ann Chen

  4. Heart Rhythm Center, Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan

    Yun-Yu Chen, Yenn-Jiang Lin, Fang-Yi Chen & Shih-Ann Chen

  5. Cardiovascular Research Center, College of Medicine, National Chung Hsing University, Taichung, Taiwan

    Yun-Yu Chen, Yu-Cheng Hsieh & Shih-Ann Chen

  6. Institute of Epidemiology and Preventive Medicine College of Public Health, National Taiwan University, Taipei, Taiwan

    Kuo-Liong Chien

  7. School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan

    Yenn-Jiang Lin, Yu-Cheng Hsieh & Shih-Ann Chen

  8. Department of Internal Medicine, Taichung Veterans General Hospital, Taichung, Taiwan

    Yenn-Jiang Lin

  9. Danish Center for Health Services Research, Department of Clinical Medicine, Aalborg University, Aalborg, Denmark

    Gregory Y. H. Lip

  10. Medical University of Bialystok, Bialystok, Poland

    Gregory Y. H. Lip

Authors
  1. Yang Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yun-Yu Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Kuo-Liong Chien
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Yenn-Jiang Lin
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Fang-Yi Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Yu-Cheng Hsieh
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Gregory Y. H. Lip
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Shih-Ann Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Yang Chen, Yun-Yu Chen, Gregory Y. H. Lip, and Shih-Ann Chen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Yang Chen, Yun-Yu Chen, Gregory Y. H. Lip, and Shih-Ann Chen. Acquisition, analysis, or interpretation of data: Yang Chen, Yun-Yu Chen. Drafting of the manuscript: Yang Chen, Yun-Yu Chen. Critical review of the manuscript for important intellectual content: All authors. Statistical analysis: Yang Chen, Yun-Yu Chen. Supervision: Yun-Yu Chen, Gregory Y. H. Lip, and Shih-Ann Chen.

Corresponding author

Correspondence to Yun-Yu Chen.

Ethics declarations

Ethics approval and consent to participate

Ethical approval for the study was obtained from the National Taiwan University Hospital Institutional Review Board (IRB No: 201003001R). Written informed consent was obtained from each participant recruited in this study.

Consent for publication

The submitted manuscript has been read and approved by all authors.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Chen, Y., Chen, YY., Chien, KL. et al. Long-term trajectories of apolipoprotein A1 and major adverse cardiovascular events and mortality in a community cohort. Lipids Health Dis 24, 137 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-025-02552-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-025-02552-3

Keywords

Lipids in Health and Disease

ISSN: 1476-511X