Prognostic Value of Cardiovascular Biomarkers in the Population

Abstract

Importance: Identification of individuals at high risk for atherosclerotic cardiovascular disease within the population is important to inform primary prevention strategies.

Objective: To evaluate the prognostic value of routinely available cardiovascular biomarkers when added to established risk factors.

Design, setting, and participants: Individual-level analysis including data on cardiovascular biomarkers from 28 general population-based cohorts from 12 countries and 4 continents with assessments by participant age. The median follow-up was 11.8 years.

Exposure: Measurement of high-sensitivity cardiac troponin I, high-sensitivity cardiac troponin T, N-terminal pro-B-type natriuretic peptide, B-type natriuretic peptide, or high-sensitivity C-reactive protein.

Main outcomes and measures: The primary outcome was incident atherosclerotic cardiovascular disease, which included all fatal and nonfatal events. The secondary outcomes were all-cause mortality, heart failure, ischemic stroke, and myocardial infarction. Subdistribution hazard ratios (HRs) for the association of biomarkers and outcomes were calculated after adjustment for established risk factors. The additional predictive value of the biomarkers was assessed using the C statistic and reclassification analyses.

Results: The analyses included 164 054 individuals (median age, 53.1 years [IQR, 42.7-62.9 years] and 52.4% were women). There were 17 211 incident atherosclerotic cardiovascular disease events. All biomarkers were significantly associated with incident atherosclerotic cardiovascular disease (subdistribution HR per 1-SD change, 1.13 [95% CI, 1.11-1.16] for high-sensitivity cardiac troponin I; 1.18 [95% CI, 1.12-1.23] for high-sensitivity cardiac troponin T; 1.21 [95% CI, 1.18-1.24] for N-terminal pro-B-type natriuretic peptide; 1.14 [95% CI, 1.08-1.22] for B-type natriuretic peptide; and 1.14 [95% CI, 1.12-1.16] for high-sensitivity C-reactive protein) and all secondary outcomes. The addition of each single biomarker to a model that included established risk factors improved the C statistic. For 10-year incident atherosclerotic cardiovascular disease in younger people (aged <65 years), the combination of high-sensitivity cardiac troponin I, N-terminal pro-B-type natriuretic peptide, and high-sensitivity C-reactive protein resulted in a C statistic improvement from 0.812 (95% CI, 0.8021-0.8208) to 0.8194 (95% CI, 0.8089-0.8277). The combination of these biomarkers also improved reclassification compared with the conventional model. Improvements in risk prediction were most pronounced for the secondary outcomes of heart failure and all-cause mortality. The incremental value of biomarkers was greater in people aged 65 years or older vs younger people.

Conclusions and relevance: Cardiovascular biomarkers were strongly associated with fatal and nonfatal cardiovascular events and mortality. The addition of biomarkers to established risk factors led to only a small improvement in risk prediction metrics for atherosclerotic cardiovascular disease, but was more favorable for heart failure and mortality.

Figure 1.. Flow of Publications, Cohorts, and Individuals by Exclusion Criteria

BNP indicates B-type natriuretic peptide; CRP, C-reactive protein; NT-proBNP, N-terminal pro-BNP.

Figure 2.. Association of Biomarkers With Incident Cardiovascular Events

The data were plotted based on results from adjusted Cox models or Fine and Gray subdistribution hazard models for the log_e-transformed biomarker concentrations using penalized cubic splines. The x-axes contain the back-transformed values on a log scale (range, lowest reported value [or the limit of detection in the case of high-sensitivity cardiac troponin I and T] to the 99th percentile). The shading indicates the 95% CIs. The corresponding reference values vary per graph and are used specifically for each biomarker and event combination. The median follow-up time was 11.8 years (IQR, 6.2-18.0 years) for atherosclerotic cardiovascular disease, 13.0 years (IQR, 7.9-19.2 years) for all-cause mortality, 12.8 years (IQR, 6.2-19.3 years) for heart failure, 11.9 years (IQR, 6.3-17.7 years) for ischemic stroke, and 11.6 years (IQR, 6.2-17.8 years) for myocardial infarction. Additional data appear in eTable 6 and eTables 10-13 in Supplement 1.

Figure 3.. Association of the Biomarkers of Interest With Incident Cardiovascular Events

NT-proBNP indicates N-terminal pro-B-type natriuretic peptide. ^aHazard ratio (HR) per unit increase = exponential function(log[HR per 1-SD increase]/SD). ^bFirst possible/definite coronary heart disease or ischemic stroke event; coronary revascularization; or died of coronary heart disease, ischemic stroke, or unclassifiable. ^cThe median follow-up was 11.8 years (IQR, 6.2-18.0 years) for atherosclerotic cardiovascular disease; all-cause mortality, 13.0 years (IQR, 7.9-19.2 years); heart failure, 12.8 years (IQR, 6.2-19.3 years); ischemic stroke, 11.9 years (IQR, 6.3-17.7 years); and myocardial infarction, 11.6 years (IQR, 6.2-17.8 years). ^dThese 3 biomarkers were combined into 1 model to investigate their independent effect on association.

Figure 4.. Heat Maps Displaying the Changes in the C Statistic After the Addition of the Biomarker to the Base Model

The changes in the C statistic were based on Cox regression models for all-cause mortality and based on Fine and Gray models for the other outcomes. The C statistics were computed based on the 1-year, 5-year, and 10-year probabilities of any events. Additional data appear in eTable 6 and eTables 10-13 in Supplement 1. The analyses stratified by age appear in eTable 16 in Supplement 1.