Utilizing biomarkers in atrial fibrillation: the pros and cons

This editorial refers to ‘Using multimarker screening to identify biomarkers associated with cardiovascular death in patients with atrial fibrillation’ by T. Pol et al., https://doi.org/10.1093/cvr/cvab262. The interplay of atrial fibrillation (AF) with various conditions such as chronic kidney disease, obstructive sleep apnoea, and chronic obstructive pulmonary disease is well recognized, reinforcing the notion that AF rarely occurs in isolation and that this is in fact a multisystem disorder. Accordingly, the approach to AF care has moved beyond mere therapeutics aimed at the disease itself to greater focus on holistic and integrated management, as well as the upstream management of risk factors, given the improved outcomes with such an approach.1,2 A real-world observational study previously reported an annual mortality rate in patients with AF of 5.5%, the majority of which were attributable to cardiovascular causes3; however, a significant proportion of this cohort of patients were not receiving anticoagulation therapy. In contrast, landmark clinical trials on non-vitamin K antagonist oral anticoagulation found that the residual risk of mortality despite anticoagulation ranged from 1.90% to 4.35% per year.4 This highlights the ongoing need for better treatment in patients affected by AF. In this regard, identification of novel biomarkers may assist the pursuit of improved risk stratification, greater monitoring of response to therapy and enhanced understanding of the pathophysiological pathways which are influenced by AF. Pol et al.5 describe their comprehensive study of 268 biomarkers with targeted proteomic methodology using the Proximity Extension Assay technology, which allows the simultaneous measurement of multiple biomarkers with a high degree of specificity and accuracy, and conventional immunoassays to identify those associated with cardiovascular death among anticoagulated patients with AF. The study consisted of two phases: identification of relevant biomarkers using a cohort of 4574 patients from the ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) trial; and validation of these biomarkers using a separate cohort of 1319 patients from the RE-LY (Randomized Evaluation of Long-Term Anticoagulation Therapy) trial. Here, the primary outcome of cardiovascular death was defined as any vascular death excluding death from non-stroke-related haemorrhage. Cause of death was centrally adjudicated in both trials. The authors found that the biomarkers most strongly associated with cardiovascular death were N-terminal pro-B-type natriuretic peptide, cardiac troponin T, interleukin-6, growth differentiation factor-15, fibroblast growth factor 23, urokinase receptor, trefoil factor 3, tumour necrosis factor receptor 1, tumour necrosis factor-related apoptosis-inducing ligand receptor 2, and cathepsin L1.5 The findings suggest that cardiorenal dysfunction, oxidative stress, inflammation, calcium haemostasis, fibrinolysis, and apoptosis are among the mechanisms linked to cardiovascular death in AF. The use of biomarkers to improve risk stratification of diseases such as AF has been a longstanding topic of discussion. The concept is not new, with the first study >15 years ago showing that when added to clinical risk stratification schemes, plasma von Willebrand factor levels (a biomarker of endothelial damage/dysfunction) refined clinical risk stratification for stroke and vascular events among AF patients.6 Since then a plethora of new or novel biomarkers have been proposed, such that biomarkers seem to begat new biomarkers. While the addition of biomarkers to clinical risk scores typically results better prediction of outcomes, it introduces an additional step in the busy clinical environment which could result in a delay in treatment. In this regard, point-of-care-testing may offer a potential solution. There are several other disadvantages to consider with the use of biomarkers in clinical practice (Figure 1).7 Most biomarkers are non-specific and therefore a mere reflection of an unwell patient with multiple health issues.8 For example, elevated levels of natriuretic peptide are observed with a whole host of conditions such as heart failure, sepsis, hyperthyroidism, chronic kidney disease, advanced age, pulmonary embolism, and liver cirrhosis.9 Therefore, the timing of samples for biomarker testing is critical, notwithstanding the fact that some biomarkers are subject to diurnal variation. There is also the issue with inter- and intra-individual variability for each biomarker which complicates the definition of normal ranges, and cut-offs for significant deviation in these values need to be established. Benefits and limitations of biomarker testing in atrial fibrillation. Created using BioRender.com. Moreover, interpretation of the significance of abnormal biomarker levels is challenging as many studies rely on a one-off measurement at baseline to predict outcomes several years later, failing to account for fluctuations over time. Among patients with AF, repeated measures are needed to account for the dynamic nature of risk in this cohort,10 though this approach may have significant cost implications. Despite the limitations above, there are possible benefits to biomarker testing in AF. For instance, improved risk stratification may facilitate further research into whether or not certain high-risk subgroups of patients with AF may benefit from more aggressive interventions (e.g. catheter AF ablation, left atrial appendage occlusion) and frequent follow-up. The unanswered question here is whether poor outcomes in patients identified with abnormal biomarkers may be preventable. Biomarkers may also assist in the monitoring of response to treatment and dose-titration, along with providing an objective assessment of the patient’s general health given the non-specific nature of biomarkers. They also provide valuable insight into the pathophysiological processes that are involved in the development of AF and helps delineate the relationship between AF and other conditions. Overall, there may be a case for the use of biomarkers in AF though their implementation in clinical risk stratification schemas need to be balanced with simplicity and practicality. Further research is needed to define the cut-offs of established biomarkers such as N-terminal pro-B-type natriuretic peptide and cardiac troponins in the context of AF, and to provide a better understanding of the implications of abnormal biomarker levels in terms of prognosis and potential treatment strategies. Conflict of interest: G.Y.H.L.: consultant and speaker for BMS/Pfizer, Boehringer Ingelheim, and Daiichi-Sankyo. No fees are directly received personally. The funders had no role in the submitted paper. All other authors have declared no conflict of interest.