Predicting New-Onset Atrial Fibrillation in Patients with Arrhythmias Using High-Sensitivity Troponin I

1. Introduction

Atrial fibrillation (AF) is the most common sustained arrhythmia, significantly increasing the risk of stroke, heart failure, and death [1,2,3]. Early detection and risk stratification are crucial for improving patient outcomes [4]. While various risk factors and biomarkers for AF have been identified [5], their predictive value remains a subject of ongoing research, particularly in patients already presenting with other forms of arrhythmia.

Troponin I (TnI) is a well-established biomarker for myocardial injury and is central to the diagnosis of acute coronary syndromes [6,7]. The development of high-sensitivity troponin I (hs-TnI) assays has enabled the detection of even minute elevations in troponin levels, reflecting subtle subclinical myocardial damage [6]. Recent studies have suggested that these low-level elevations of hs-TnI, even in the absence of acute ischemia, may be indicative of chronic atrial remodeling and fibrosis - pathophysiological processes known to be fundamental to the development and maintenance of AF [8,9].

Current literature has explored the association between elevated hs-TnI and the incidence of new-onset AF in the general population and in specific high-risk groups, such as patients with heart failure or post-cardiac surgery [10,11,12,13,14]. Several studies have shown a positive correlation, with higher baseline hs-TnI levels being independently associated with a greater risk of developing AF [10,12,13,14,15]. However, a significant gap remains in understanding the specific role of hs-TnI as a prognostic marker for the development of AF in a cohort of patients with pre-existing arrhythmias other than AF. This specific population is at an inherently higher risk for arrhythmia-related complications, and a predictive biomarker could be invaluable for targeted monitoring and prophylactic strategies.

The primary objective of this study is to investigate the predictive value of high-sensitivity Troponin I for the development of new-onset atrial fibrillation in patients with other existing arrhythmias. By evaluating the relationship between baseline hs-TnI levels and subsequent AF events, this research aims to determine whether hs-TnI can serve as a useful tool for early risk stratification and guide management strategies for this vulnerable patient population.

2. Materials and Methods

3. Results

4. Discussion

The primary objective of this prospective, single-center study was to investigate the predictive value of baseline hs-TnI for the development of new-onset AF in a vulnerable cohort of 232 patients presenting with other non-AF arrhythmias. Our central hypothesis was that elevated hs-TnI levels, reflecting subclinical myocardial injury and remodeling, would be independently associated with a higher risk of incident AF during the 12-month follow-up.

While 16 cases (6.9%) of new-onset AF were observed, and initial Kaplan-Meier analysis showed a notable trend - the highest hs-TnI quartile had a cumulative AF rate of approximately 15% compared to near-zero rates in the lowest quartiles - this association did not reach statistical significance in the adjusted Cox regression model (p = 0.844). Instead, the multivariate analysis revealed three strong, independent predictors of new-onset AF: Left Atrial (LA) size (HR = 1.27, p < 0.001), History of Diabetes Mellitus (HR = 5.99, p = 0.006), and History of Stroke (HR = 10.18, p = 0.032).

Our findings both align with and diverge from established literature regarding AF prediction. The high predictive power of LA size is consistent with numerous studies recognizing left atrial enlargement as the single strongest morphological substrate for AF due to increased wall stress, fibrosis, and conduction heterogeneity [19,20,21,22,23]. Similarly, the significant hazard ratios observed for Diabetes Mellitus (HR ≈ 6) and History of Stroke (HR ≈ 10) reinforce their roles as powerful systemic and embolic risk factors, likely mediated through mechanisms such as diabetic cardiomyopathy and atrial stunning post-stroke [24,25,26,27,28,29,30,31].

Conversely, our result showing no statistically significant independent association between baseline hs-TnI and incident AF contrasts with several large population studies that reported a clear link between even low-level troponin elevation and future AF risk. For instance, some studies found hs-TnI to be an independent predictor in cohorts with heart failure or general cardiovascular disease [10,12,13,15]. The disparity in our results, however, may be partially explained by the specific nature of our cohort—patients with pre-existing, often symptomatic, non-AF arrhythmias. This patient group likely already harbors a high baseline burden of atrial remodeling that might override the incremental prognostic value of a single baseline hs-TnI measurement over a short follow-up period.

The strong, albeit non-significant, trend observed in the Kaplan-Meier analysis (Figure 4) suggests that while hs-TnI may reflect underlying pathology, its predictive utility in this specific high-risk arrhythmia cohort is eclipsed by established structural (LA size) and systemic (Diabetes, Stroke) risk factors.

The model’s preference for LA size and Diabetes as predictors suggests that the final common pathway to new-onset AF in this group is dominated by advanced atrial remodeling. A key hypothesis emerging from this work is the potential role of dynamic troponin changes. While baseline hs-TnI may capture static myocardial health, a new model could investigate whether the rate of change or serial measurements of hs-TnI over time better reflect rapid or progressive atrial injury that precipitates AF. In this already compromised population, small, acute spikes in hs-TnI may be more relevant than a single baseline value.

Furthermore, the non-significant but notable high Hazard Ratio for History of Myocardial Infarction (HR ≈ 5.1, p = 0.094) suggests that with a larger sample size, this factor, which is clearly linked to myocardial stress and fibrosis, may reach significance, supporting a broader mechanism of global myocardial strain leading to AF.

Our findings have crucial implications for the risk stratification of patients presenting with non-AF arrhythmias. The results strongly suggest that risk assessment should be driven primarily by structural parameters, particularly LA size, and major cardiovascular comorbidities like Diabetes and prior Stroke. Clinicians should aggressively screen for AF in arrhythmia patients who present with any of these three factors. Given the 10-fold increased risk associated with a history of stroke, enhanced and continuous rhythm monitoring (e.g., implantable loop recorders) is warranted in this specific subset to prevent recurrent stroke.

While hs-TnI did not emerge as an independent predictor in this specific model, the finding that 80 patients (over one-third of the cohort) had hs-TnI levels above the 99th percentile cutoff is clinically relevant, indicating a high prevalence of subclinical myocardial injury in patients with non-AF arrhythmias. This underscores the need to address underlying cardiovascular comorbidities vigorously, even if the troponin value itself does not independently predict AF in this model.

This study possesses several strengths, including its prospective cohort design and the focus on a highly specific, high-risk patient population (non-AF arrhythmia patients). We used a validated high-sensitivity assay for hs-TnI and employed a robust multivariate Cox regression model adjusted for major confounders. The clear exclusion of patients with pre-existing AF and severe renal impairment (eGFR<30 ml/min/m2) ensures a purer assessment of new-onset AF risk.

The study also has limitations. First, the sample size was relatively small (N = 232), and the number of incident AF events was low (n=16), which may have contributed to the borderline statistical significance observed for variables like hs-TnI and Myocardial Infarction history. Second, the median follow-up period was relatively short (12 months), meaning late-onset AF may have been missed. Third, the diagnosis of AF relied primarily on scheduled follow-up ECGs and Holter monitoring, potentially missing asymptomatic, paroxysmal AF episodes between visits. Lastly, the study’s single-center origin may limit the generalization of the findings to more diverse populations

5. Conclusions

In this prospective cohort of patients with non-AF arrhythmias, baseline hs-TnI was not found to be an independent predictor of new-onset atrial fibrillation (AF) during the 12-month follow-up period. Instead, the risk of developing AF was overwhelmingly predicted by established clinical and structural factors. Specifically, Left Atrial (LA) size, a History of Diabetes Mellitus, and a History of Stroke emerged as powerful, independent predictors of incident AF.

These findings suggest that in patients already presenting with other arrhythmias - a population likely to have significant underlying atrial pathology - risk stratification for future AF should prioritize structural changes and major comorbidities over a single baseline measurement of hs-TnI. Clinically, aggressive screening and management strategies should be focused on arrhythmia patients exhibiting LA enlargement, diabetes, or a history of stroke to mitigate the substantial associated AF risk.