Biomarkers in Primary Prevention: Steps beyond Established Risk Scores

Cardiovascular risk stratification is essential in order to assess prognosis and tailor specific preventive and/or therapeutic interventions (1, 2). Besides exploration and establishment of risk assessment scores such as SCORE2 by the European Society of Cardiology (3), this has prompted the search for novel blood-based biomarkers for prediction of cardiovascular events. Since the event rate in most populations is relatively low during intermediate-term follow-up and circulating levels of most candidate biomarkers show a narrow distribution range in (apparently) healthy individuals, prediction of adverse cardiovascular events in primary prevention has always been challenging. Among multiple biomarker candidates, C-reactive protein (CRP) and cardiac troponins I (cTnI) and T (cTnT), measured by high-sensitivity assays, have received widespread interest and a large database has been accumulated on their potential role as predictors of cardiovascular risk (4, 5). In addition, natriuretic peptides—most prominently N-terminal pro-B-type natriuretic peptide (NTproBNP) and growth differentiation factor-15 (GDF-15)—have also been evaluated in various studies for their predictive value (6, 7). The observed associations between these biomarkers and future coronary and cardiovascular events are not deemed to be causal, but the biomarkers have been selected as reflectors of different underlying pathophysiological features of cardiovascular disease (CVD). CRP is used as indicator of acute-phase inflammation (5); troponins are used for their potential to mirror myocardial injury (4), natriuretic peptides as markers of remodeling (8), and GDF-15 as an indicator of tissue damage and inflammatory conditions (7). In secondary prevention, multiple studies have documented a statistically significant and moderately clinically relevant contribution of the aforementioned biomarkers to cardiovascular risk prediction independent of established risk scores. However, such data is scarce in the setting of primary prevention. In the current issue of Clinical Chemistry, Welsh and coinvestigators evaluated the use of a combined biomarker approach for prediction of major adverse cardiovascular events (MACE) in the general population (9). Most of the biomarkers under investigation—namely NTproBNP, cTnI, cTnT, GDF-15, and CRP—have already been studied at an individual level in disease cohorts. The utility of a multimarker approach in a primary prevention setting has been performed in only a few studies, and the biomarkers of choice investigated across these studies seem repetitive, mainly troponins and natriuretic peptides (10). The present study adds to this knowledge and provides additional evidence for the use of a combined biomarker approach for MACE prediction in both females and males. The study population comprised a large sample size of more than 18 000 individuals of mainly European ethnicity from the Generation Scotland Scottish Family Health Study, with an age range of 35 to 65 years at baseline and a median follow-up of 11.6 years. The authors investigated relevant clinical outcomes and used appropriate statistical methodology. During the follow-up, MACE occurred in a total of 5.1% of the study participants. In unadjusted and SCORE2-adjusted single biomarker models, NT-proBNP had numerically the strongest association with MACE in both sexes. The predictive value and discrimination of a base model using the SCORE2 cardiovascular risk factors was already excellent for both women and men (c-indices 0.826 and 0.795, respectively). While in C-statistics all biomarkers individually failed to improve discrimination in both sexes, a combined biomarker approach was able to slightly improve discrimination and thus lead to a small but statistically significant improvement when added to SCORE2 factors. The study by Welsh and coworkers is one of a few that have investigated both cardiac troponins, hsTnI and hsTnT. It seems somewhat of a surprise that the correlation between these 2 markers is only moderate and that cTnT appears to be a weaker predictor of cardiovascular risk (9). Troponins are cornerstones in the diagnosis of acute myocardial infarction, and there is great interest in the expansion of their clinical use towards CVD prognosis as well as CVD risk prediction. Most often either cTnI or cTnT are measured individually—rarely are both measured (11). Although both troponins perform similarly in the acute clinical care setting, it has already been reported previously that they show a low correlation and that cardiac troponin T appears to be a stronger predictor of mortality—in particular of non-cardiovascular mortality. In contrast, cTnI seems to be more sensitive to myocardial ischemia and coronary artery disease (12). Thus, predictive differences between cTnI and cTnT are present, and further studies should ultimately assess the importance of such differences and whether the choice of the troponin subunit to be measured is critical for real-life clinical application and risk management or whether they have additive value. It is much appreciated that Welsh and colleagues touched on the rising importance of sex in preventive cardiology. CVD in females is still highly understudied (13). Despite a lower prevalence of CVD in women and the fact that women tend to develop CVD later in life, females show increased rates of relevant CVD comorbidities and worse mortality rates (14). Therefore, improved knowledge in relation to sex and gender differences in CVD risk factors and related biomarkers is needed. In line with this, Welsh et al. applied sex-specific analyses and report sex differences in biomarker concentrations and in their association with MACE. These tended to be numerically stronger in women than men, showing once more that sex-specific differences are important factors that must be acknowledged. The authors recognized some limitations of their work, for example the limited generalizability of their findings to ethnicities others than those of European descent and shortcomings related to biomarker measurements. Additionally, some further issues should be considered. First, the number of endpoints is fairly low, thus overestimation of the results might be an issue. In this context, an even longer follow-up period would also have been desirable. While the combined biomarker approach further refined cardiovascular risk estimates on top of SCORE2 factors, the improvement ranged from 0.011 to 0.014 in C-statistics and thereby seems small from a clinical point of view. Looking into the baseline characteristics, this is not surprising, since there are clear differences in subjects with and without incident MACE during follow-up: median age (61 vs 47 years), mean systolic blood pressure (142 vs 131 mmHg), mean total cholesterol (5.4 vs 5.1 mmol/L), prevalence of diabetes (7.7% vs 2.1%), and the number of subjects receiving antihypertensive medication (18.8% vs 6.3%) is higher in the event group. This, in turn, results in excellent discrimination by the clinical score in the first place. An aim of future research should be to identify and develop novel tools that are able to improve risk prediction beyond established risk factors and clinical risk scores even more. Proteomics could represent such a novel approach—for example, Hoogeveen et al. (15) showed that a proteome-based model outperformed a model comprising clinical risk factors in predicting the risk of cardiovascular events. Nevertheless, the fundamental task in risk prediction is not to identify biomarkers associated with incident disease but rather to identify those markers sufficiently unrelated to established risk factors to be used in addition to these factors. Finally, to ensure broad implementation into clinical pathways all preventive diagnostics and applications need to be evaluated in a cost-benefit analysis. This is a particular challenge in primary prevention, where the measurement of biomarkers in clinical practice often comes at increased cost to patient care without known therapeutic or interventional value. In the specific setting of the Generation Scotland Scottish Family Health Study, where the application of a simple anamnestic tool, SCORE2, already performs excellently and cannot be further improved to a significant extent by blood-based biomarkers, the cost-benefit relationship is more of academic interest. Although the current study by Welsh included relatively young subjects (age range 35 to 65 years at baseline), the data already show increased levels of biomarkers, which in turn might reflect pathophysiological conditions that are already present. To achieve even more effective risk prediction, one should consider the application of preventive strategies and schemes much earlier in the life course. The rise of polygenic risk scores demonstrates the power of genetics to improve risk prediction, as subjects at risk can be identified with a higher precision than conventional risk scores even before the onset of traditional risk factors and disease manifestations (16). Quantification of CVD risk on a genetic basis raises the possibility not only of personalized prevention but also of personalized treatment strategies and can be used to better connect CVD and other disease traits on the level of genetics. Whether genetic information and polygenic risk scores can be introduced into clinical practice on a broad scale needs to be evaluated by future studies. In summary, Welsh and coworkers confirm and extend our knowledge regarding multiple cardiac biomarkers and their combination in primary prevention and risk stratification. One may hope that this study will not only further boost modern research paths toward the clinical utility of blood-based biomarkers but might also revive awareness of the (high) value of established clinical risk scores. The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list. Tanja Zeller (Conceptualization-Lead, Writing—original draft-Equal, Writing—review & editing-Equal), Apurva Shrivastava (Conceptualization-Supporting, Writing—original draft-Supporting, Writing—review & editing-Supporting), and Mahir Karakas (Conceptualization-Lead, Writing—original draft-Equal, Writing—review & editing-Equal). Upon manuscript submission, all authors completed the author disclosure form. T. Zeller is supported by the German Center for Cardiovascular Research (FKZ 81Z1710101 and FKZ 81Z0710102). M. Karakas is supported by the Else Kroener-Fresenius-Stiftung (W3 Else Kroener Clinician Scientist Professorship). T. Zeller is co-founder and shareholder of the ART-EMIS Hamburg GmbH, which holds an international patent application on the use of a computing device to estimate the probability of myocardial infarction (Publication Numbers WO2022043229A1, TW202219980A). M. Karakas is a part-time employee of 4TEEN4 Pharmaceuticals GmbH and reports grant and nonfinancial support from Adrenomed AG and CSL Vifor, as well as personal fees from Adrenomed AG, Sphingotec, Vifor, Daiichi-Sankyo, Pharmacosmos, and 4TEEN4—all outside the submitted work.