Cardiovascular disease (CVD) is the leading cause of mortality globally [1]. China has a high burden of CVD, with a prevalence of 290 million, representing about 40% of the country’s mortality [2, 3]. Although treatments and secondary prevention improve the patient prognosis, primary prevention remains the best strategy but requires the early identification of individuals at a high risk of CVD. Over the last few decades, numerous CVD risk prediction models have been developed for the general population [4], including several well-known models, such as the Framingham model [5], SCORE [6], and QRISK3 [7]. Although many models exist for CVD risk prediction, their value remains unknown in China because of the differences in etiology and risk factors between Western and Chinese patients [8, 9]. More importantly, a notable characteristic of the CVD epidemiology in China is the rapidly increased mortality due to coronary heart disease (CHD) and the increasing age of the patients with CVD [10]. The analysis of Chinese data in the Global Burden of Disease study shows that fatal or nonfatal CHD events increased mainly in older adults aged 70–84 years [11]. Unfortunately, most CVD risk prediction models are not suitable for individuals aged $>$74 years. In addition, the association between traditional risk factors and CVD in older adults weakens with advancing age [12]. Therefore, an effective predictive biomarker beyond traditional risk factors needs to be explored to optimize the risk prediction model of CVD [13].

Recently, the studies by Saeed et al. [14] and Mehta et al. [15] reported that adding biomarkers (N-terminal pro-B-type natriuretic peptide, high-sensitivity cardiac troponin T, and high-sensitivity C-reactive protein) to the traditional risk factors improved CVD risk prediction in older adults. Increasing evidence shows that cardiac myosin-binding protein-C (cMyBP-C), a myocardium-restricted protein, can be used as an early diagnostic biomarker of acute myocardial infarction (MI) [16]. Large quantities of cMyBP-C are released into the bloodstream once a MI or ischemic injury occurs. Moreover, cMyBP-C is detectable in all healthy individuals independent of ischemic injury [17]. Incorporating plasma cMyBP-C levels in the analysis of CVD status and future events has been suggested as the most promising approach to achieve high prediction accuracy [18]. Still, whether cMyBP-C can be used as a predictive biomarker of CVD in Chinese patients is unknown.

Therefore, this study aimed to construct a nomogram for CVD risk prediction in the general population. This optimized model might demonstrate good performance for future CVD risk prediction in the general population, including young adults.

A nomogram based on cMyBP-C, age, DBP, cigarettes per day, and family history of CVD was constructed. The model displayed a high predictive ability. Therefore, the nomogram may be a promising way for CVD risk prediction.

CVD is a widespread public health concern. Its silent development until the sudden attack poses a significant threat to the lifespan, emphasizing the crucial role of early prevention. Clinical guidelines recommend using risk prediction models to identify individuals at a high risk of CVD [22]. Over the past three decades, numerous prediction models have been developed based on various traditional risk factors. The Framingham model is the most widely used [23, 24], but a recent study demonstrated that the model overestimated the risk of CVD for the Chinese population [8]. Besides, China is experiencing an increasingly aging population, and the growing annual burden of CHD weighs mainly on the older adults aged 70–84 years [25], who are not included in the popular models such as the Framingham model [5], SCORE [6], and QRISK3 [7]. Therefore, applying these models to patients aged $\ge$75 years is inadequate. Furthermore, only one or no traditional risk factors are found in many patients with CVD [26]. Hence, new biomarkers are needed to improve the information gained from traditional predictors.

Multiple biomarkers that reflect myocardial injury, heart failure, and inflammation (e.g., troponin I, N-terminal pro-B-type natriuretic peptide, and C-reactive protein, respectively) have been investigated and proved to improve the predictive performance of models that include traditional risk factors of CVD [27, 28, 29, 30]. Unfortunately, a report from the Framingham Heart Study suggested that the currently selected biomarkers had only a small incremental effect on the accuracy compared with the prediction model based on traditional risk factors alone [31]. Hence, biomarker selection is crucial for improving the models based on traditional risk factors of CVD. With a lower detection limit of 0.05 ng/mL, the current cardiac troponin assay lacks the sensitivity to detect minor myocardial injury because of the low prevalence of elevated levels among healthy people. N-terminal pro-B-type natriuretic peptide is relatively well studied in primary prevention and predicts cardiovascular events; nevertheless, age contributes a powerful influence on its levels which could influence the applicable population of the prediction model. Compared with previously investigated biomarkers, cMyBP-C is produced exclusively by the myocardium and is elevated in acute MI [16] but is also detectable in all healthy individuals independent of ischemic injury [17].

The present study established a CVD risk prediction model suitable for all age groups. As expected, the nomogram incorporating age, DBP, cigarettes per day, family history of CVD, and cMyBP-C demonstrated adequate predictive ability in the training cohort (AUC = 0.816) and the validation cohort (AUC = 0.774). The novel biomarker cMyBP-C also displayed an independent association with the development of CVD. The risk prediction model consisting of cMyBP-C and traditional risk factors proved to be statistically significant; however, the AUC is not perfect, suggesting that the model can still be improved. The present study is the first to include cMyBP-C levels in a predictive model of CVD, and no comparison with previous studies was possible. Still, Tong et al. [18] reported an AUC of 0.629–0.638 for cMyBP-C used alone to predict future CVD, with a sensitivity of 81%–96%, specificity of 27%–48%, and NPV of 93%–97%. Next, we have proposed that combining cMyBP-C with other protein biomarkers above conventional cardiovascular risk factors could further improve risk assessment.

This study had certain limitations. First, the nomogram was derived based on a 7-year follow-up and was thus not comparable to the Framingham score. Secondly, individuals with missing data were directly excluded, which reduced the available sample size and could lead to the inaccurate predictive performance of the model [32]. Besides, the patients were all from the same hospital, inevitably leading to bias due to the relatively consistent external environment and treatment methods. A multicenter population and external validation are necessary for future derivation and validation. Another limitation was that the cMyBP-C levels were detected by ELISA. A more convenient high-sensitivity assay is essential for clinical practice. Future studies will expand the population sources and conduct external validation to optimize the prediction model in the future.

To the authors’ knowledge, this study is the first to propose a CVD risk prediction model including traditional risk factors and the novel biomarker cMyBP-C. The nomogram provides a quantitative approach for estimating future CVD events. Physicians and individuals can conveniently perform a CVD risk prediction through the website (https://data15651725761.shinyapps.io/DynNomapp/).

AUC, area under the receiver operating characteristic curve; Β, regression coefficient; BMI, body mass index; CHD, coronary heart disease; CI, confidence interval; Cr, creatinine; CVD, cardiovascular disease; cMyBP-C, cardiac myosin-binding protein-C; DBP, diastolic blood pressure; ELISA, enzyme-linked immunosorbent assay; FBS, fasting blood sugar; HDL-C, high-density lipoprotein cholesterol; ICD-10, International Classification of Diseases, 10th revisions; LASSO, least absolute shrinkage and selection operator; MI, myocardial infarction; NPV, negative predictive value; OR, odds ratio; SBP, systolic blood pressure; TC, total cholesterol; TG, triglyceride; UA, uric acid.

The datasets are available from the corresponding author on reasonable request.

XCS—methodology, software, formal analysis and writing. WH—investigation and data curation. JYS—software anf formal analysis. ZHZ—Conceptualization and investigation. QSL—investigation. HYZ—Conceptualization, formal analysis, investigation, data curation and project administration. MZL—Conceptualization, data curation and project administration.

The study was conducted according to the guidelines of the Declaration of Helsinki. This retrospective study was approved by the ethics committee of The Second Affiliated Hospital of Nanjing Medical University (No. [2021]-KY-078-01) with a waiver of informed consent.

No applicable.

This research received no external funding.

QSL is an employee of Nanjing Bottests Biotechnology Co, Ltd. The authors declare no conflict of interest.