C-reactive protein and 90-day adverse clinical outcomes in acute myocarditis: a retrospective database analysis of patients requiring hospitalisation

Abstract

Background

C-reactive protein (CRP) is commonly used to assess systemic inflammation, yet its prognostic value in acute myocarditis remains uncertain. Evidence is limited by heterogeneous populations, variable CRP thresholds and differing outcome definitions. We therefore aimed to evaluate whether CRP levels measured at presentation are associated with 90-day clinical outcomes in adults hospitalised with acute myocarditis.

Methods

Using the TriNetX Research Network, we identified 3615 adults hospitalised with acute myocarditis between 2005 and 2025. Patients were stratified according to baseline CRP <10 mg/L (n=1001) vs ≥10 mg/L (n=2614). Propensity score matching (1:1) was performed using age, sex and race. The primary outcome was a 90-day composite of all-cause death and tachyarrhythmias (atrial fibrillation, ventricular tachycardia, ventricular fibrillation or torsades de pointes). Secondary outcomes included the individual components.

Results

In the unmatched cohort, patients with CRP <10 mg/L had a numerically higher 90-day incidence of the primary composite outcome compared with those with CRP ≥10 mg/L (16.4% vs 14.0%; log-rank p=0.087; HR 1.17, 95% CI 0.98 to 1.41), driven by a borderline higher incidence of tachyarrhythmias (14.1% vs 11.7%; log-rank p=0.059; HR 1.21, 95% CI 0.99 to 1.48), while all-cause mortality was similar between groups. After propensity score matching, the 90-day incidence of the primary composite outcome was virtually identical in the CRP <10 mg/L and CRP ≥10 mg/L groups (16.4% vs 16.9%; log-rank p=0.735; HR 0.96, 95% CI 0.78 to 1.19). Individual endpoints (tachyarrhythmias and all-cause mortality) were also similar, with no statistically significant differences.

Conclusion

In this large dataset analysis, elevated baseline CRP was frequent in patients with acute myocarditis and was not consistently associated with 90-day adverse outcomes. A single CRP measurement at presentation appears to have limited value for risk stratification in this setting.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• In a cohort of 3615 adults hospitalised with acute myocarditis, the level of C-reactive protein at presentation using thresholds of 10 mg/L was not associated with 90-day death or tachyarrhythmias after propensity score matching.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• C-reactive protein at presentation should not be used alone for risk stratification in acute myocarditis and should be interpreted in the context of overall clinical presentation.

Introduction

Myocarditis, an inflammatory condition of the myocardium, results from various infectious and non-infectious causes.¹ Its clinical presentation varies from mild, self-resolving cases to fulminant myocarditis with rapid haemodynamic collapse and high mortality.² Most uncomplicated cases resolve spontaneously, whereas complicated presentations characterised by severe ventricular dysfunction, electrical instability or cardiogenic shock often require advanced therapies, including mechanical circulatory support, as a bridge to recovery or transplantation. Mortality depends on presentation, aetiology and population characteristics; registry data show an in-hospital death or heart transplantation rate of approximately 3%, occurring exclusively in patients with complicated disease.³ In biopsy-proven cases, early adverse outcomes were seen in 28% of fulminant vs ~2% of non-fulminant myocarditis.⁴ Consequently, identifying patients at increased risk of adverse outcomes remains challenging. Reliable prognostic biomarkers that allow early discrimination between patients with non-fulminant myocarditis who will deteriorate and those with an uncomplicated course are currently unavailable.

C-reactive protein (CRP), an acute-phase protein and biomarker of systemic inflammation, has been associated with different cardiovascular diseases⁵ and is reported to be elevated in up to 80% of acute myocarditis.⁶ Because inflammation is central to the pathogenesis and clinical progression of myocarditis, CRP has been proposed as a biologically plausible marker for early risk stratification.^{7 8} However, the available evidence remains inconsistent. Schwuchow-Thonke et al⁹ observed that higher CRP, troponin I and impaired global longitudinal strain were associated with myocardial inflammation in non-ischaemic heart failure. Similarly, in 31 patients with lymphocytic myocarditis, admission CRP correlated with New York Heart Association (NYHA) functional class, suggesting a potential prognostic value.¹⁰ In contrast, Baritussio et al,¹¹ using a machine-learning approach in a large cohort of 409 patients with biopsy-proven or clinically suspected myocarditis, reported that higher CRP did not reliably identify patients with worse clinical features at presentation. Notably, higher CRP levels were found in male patients with clinically suspected myocarditis and higher troponin I values. In comparison, lower CRP levels were found among patients with biopsy-proven myocarditis, arrhythmic presentation, more advanced NYHA class at presentation and a higher likelihood of anti-heart autoantibody positivity.

Given these conflicting findings, increasing attention has shifted towards high-sensitivity CRP (hsCRP), a precise marker of systemic inflammation and a validated predictor of cardiovascular risk across multiple conditions.^{12 13} Whether inflammatory activity across the full spectrum, from marked elevations typical of acute-phase responses to the lower grade inflammation detectable by hsCRP assays, carries prognostic relevance in acute myocarditis remains unknown. To address this gap, we sought to determine whether inflammatory markers such as CRP can identify patients with acute myocarditis who are at increased risk for subsequent deterioration. Specifically, we evaluated whether both higher levels of systemic inflammation and milder, low-grade elevations measured at presentation are associated with adverse short-term outcomes across multiple US centres.

Methods

We used the TriNetX Research Network, which provides longitudinal, real-time, deidentified data from electronic medical records across different International Classification of Diseases (ICD) coding systems. All data and materials used in this study are publicly accessible through the TriNetX database platform, and the authors did not have direct access to individual-level raw data. This retrospective study was exempt from institutional review board oversight and informed consent requirements. The TriNetX network was searched on 19 November 2025 covering 70 US healthcare organisations. We included adults (≥18 years) hospitalised for acute myocarditis (ICD-10th Revision-Clinical Modification (ICD-10-CM) code I40) between 2005 and 2025. Individuals with diagnosis of sepsis or septic shock were excluded to limit confounders. Patients were required to have at least one CRP measurement obtained within 3 days of the index myocarditis diagnosis during the index hospitalisation; for each individual, the first CRP value obtained at presentation was used to capture early systemic inflammatory activity. Patients were then stratified into two groups according to a prespecified CRP threshold of 10 mg/L, a widely used cut-off indicating clinically significant inflammation in acute settings, as referenced in current guidelines.^{6 14} In a sensitivity analysis, we also applied a lower threshold of 3 mg/L to examine whether milder elevations consistent with low-grade inflammation were associated with adverse outcomes.

The primary outcome was a 90-day composite of all-cause death and tachyarrhythmias. Tachyarrhythmias were defined as atrial fibrillation, ventricular tachycardia, ventricular fibrillation or torsades de pointes. Secondary outcomes included the individual 90-day components of the primary outcome (all-cause death and tachyarrhythmias) analysed separately. Exploratory analyses evaluated the same endpoints over 1-year (1–365 days) follow-up using the 10 mg/L CRP threshold. For each patient, the starting point of the analysis (index date) corresponded to the first hospitalisation in which the acute myocarditis was diagnosed. Data analysis was conducted using the TriNetX analytics platform. Continuous variables were presented as mean±SD and compared using the independent samples t-test, while categorical variables were reported as counts and percentages and compared using the χ² test. Patients were 1:1 propensity score matched (PSM) by age at index event, sex and race. These variables were selected because they are determinants of both inflammatory profiles and clinical outcomes, and matching on them reduces baseline demographic confounding while preserving comparability between groups. Propensity scores were generated via logistic regression, and patients were matched 1:1 using a greedy nearest-neighbour algorithm with a calliper set at 0.1 of the pooled SDs and including a randomisation step to minimise selection bias. Baseline balance before and after matching was assessed using standardised mean differences (SMD), with values below 0.1 indicating adequate balance. Both unmatched and PSM analyses were conducted. Cumulative incidence curves were derived from Kaplan-Meier survival curves and compared using the log-rank (Mantel-Cox) test. HRs and 95% CIs were estimated using Cox proportional hazards regression analysis. For each outcome, time-to-event analyses were performed within a prespecified 1–90 days’ follow-up window.

Results

We identified 3615 adults hospitalised with acute myocarditis who met all inclusion criteria. Of these, 1001 (28%) had a CRP value <10 mg/L at presentation and 2614 (72%) had a CRP value ≥10 mg/L (figure 1).

Figure 1. Study flow diagram. Flow diagram illustrating the identification and selection of adult patients (≥18 years) hospitalised for acute myocarditis within the TriNetX US Collaborative Network between 2005 and 2025. A total of 3615 patients with at least one CRP measurement during the index hospitalisation were included. Patients were stratified into two groups according to baseline CRP levels (≥10 mg/L vs <10 mg/L). Propensity score matching (1:1) was performed using demographic covariates, yielding two matched cohorts of 997 patients each. CRP, C-reactive protein; HCOs, healthcare organisations.

Before PSM, patients with CRP <10 mg/L were slightly older (42 (17) vs 39 (18) years; p<0.001) and more often female (37% vs 27%; p<0.001), and there was no difference in racial distribution compared with those with CRP ≥10 mg/L. Baseline troponin levels and left ventricular ejection fraction were also similar between groups. Complete baseline characteristics are reported in table 1.

Table 1. Baseline characteristics of patients with acute myocarditis according to CRP levels before and after PSM.

Before PSM	Low CRP group (<10 mg/L) n=1001	High CRP group (≥10 mg/L) n=2614	P value	SMD
Age	42±17	39±18	<0.001	0.178
Female	370 (37%)	706 (27%)	<0.001	0.215
White	618 (62%)	1666 (64%)	0.230	0.045
Black or African-American	164 (16%)	407 (16%)	0.567	0.021
LVEF (%)	49±17	50±17	0.726	0.057
Troponin I (ng/mL)	7.2±19.3	5.9±11.7	0.647	0.032
CRP (mg/L)	4.4±3.0	73.4±73.7	<0.001	1.322
ESR (mm/hour)	17.5±19.2	27.2±25.6	0.001	0.430
Heart rate (bpm)	81±18	85±19	<0.001	0.239
QTc (ms)	425±31	423±40	0.691	0.046

After PSM	Low CRP group (<10 mg/L) n=997	High CRP group (≥10 mg/L) n=997	P value	SMD
Age	42±17	42±18	0.864	0.008
Female	368 (37%)	370 (37%)	0.926	0.004
White	618 (62%)	619 (62%)	0.963	0.002
Black or African-American	164 (16%)	165 (17%)	0.952	0.003
LVEF (%)	49±18	48±19	0.783	0.053
Troponin I (ng/mL)	7.2±19.1	5.8±10.5	0.758	0.033
CRP (mg/L)	4.4±3.0	82.6±79.5	<0.001	1.389
ESR (mm/hour)	17.5±19.2	32.7±27.4	<0.001	0.643
Heart rate (bpm)	81±18	86±18	<0.001	0.310
QTc (ms)	425±31	428±43	0.544	0.079

Values are reported as mean±SD or number (percentage). P values and SMD refer to comparisons between the low CRP group (<10 mg/L) and the high CRP group (≥10 mg/L) before and after 1:1 PSM.

BPM, beats per minute; CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; LVEF, left ventricular ejection fraction; PSM, propensity score matching; QTc, corrected QT interval; SMD, standardised mean difference.

Following 1:1 PSM, 997 patients were included in each group, and all matched variables achieved good balance (SMD <0.1). Age was nearly identical between groups (42±17 vs 42±18 years; p=0.864; SMD=0.008), as were sex distribution (female 37% vs 37%; p=0.926; SMD=0.004) and race (white 62% vs 62%, p=0.963, SMD=0.002; black/African-American 16% vs 17%, p=0.952, SMD=0.003). Complete matched characteristics are presented in table 1.

In the unmatched cohort, patients with CRP <10 mg/L and those with CRP ≥10 mg/L had comparable 90-day risk of the composite outcome (164 of 1001 (16.4%) vs 365 of 2614 (14.0%); log-rank p=0.087; HR 1.17, 95% CI 0.98 to 1.41) (figure 2).

Figure 2. Cumulative incidence of 90-day outcomes according to baseline CRP levels in the unmatched cohort. Cumulative incidence curves (Kaplan-Meier estimates) at 90 days for the primary composite outcome (all-cause death and tachyarrhythmias), all-cause death and tachyarrhythmias in the unmatched cohort, stratified by baseline CRP <10 mg/L (red) vs CRP ≥10 mg/L (blue). Log-rank p values and HRs with 95% CIs are shown for each comparison. CRP, C-reactive protein.

Regarding individual outcomes, tachyarrhythmias showed a borderline higher incidence in the CRP <10 mg/L group, although this difference did not reach statistical significance (141 of 1001 (14.1%) vs 305 of 2614 (11.7%); log-rank p=0.059; HR 1.21, 95% CI 0.99 to 1.48). All-cause mortality remained comparable (29 of 1001 (2.9%) vs 87 of 2614 (3.3%); p=0.488; HR 0.86, 95% CI 0.57 to 1.31) (figure 2).

After PSM, at 90 days, the composite outcome of all-cause death and tachyarrhythmias occurred in 164 of 997 (16.4%) patients in the CRP <10 mg/L group and 168 of 997 (16.9%) in the CRP ≥10 mg/L group (log-rank p=0.735; HR 0.96, 95% CI 0.78 to 1.19), indicating no significant difference between groups (figure 3). Individual outcomes showed similar consistency. All-cause mortality occurred in 29 of 997 (2.9%) patients with CRP <10 mg/L and 36 of 997 (3.6%) with CRP ≥10 mg/L (log-rank p=0.369; HR 0.80, 95% CI 0.49 to 1.30). Rates of tachyarrhythmias were similar between groups, with no statistically significant difference (141 of 997 (14.1%) vs 139 of 997 (13.9%); log-rank p=0.970; HR 1.00, 95% CI 0.79 to 1.27) (figure 4).

Figure 3. Study cohort, propensity score-matched groups and 90-day cumulative incidence of the primary composite outcome according to baseline CRP levels. Schematic representation of the study cohort and propensity score-matched analysis. From 3615 adults hospitalised with acute myocarditis and available CRP values in the TriNetX US Collaborative Network (70 HCOs, 2005–2025), two groups were created based on baseline CRP <10 mg/L and ≥10 mg/L. A 1:1 propensity score matching was performed on age, sex and race, yielding 997 patients with CRP <10 mg/L and 997 with CRP ≥10 mg/L. The right panel shows the 90-day cumulative incidence curve for the primary composite outcome (all-cause death and tachyarrhythmias) in the matched cohorts, with corresponding log-rank p value and HR with 95% CI. CRP, C-reactive protein; HCOs, healthcare organisations.

Figure 4. Cumulative incidence of 90-day individual outcomes in the propensity score-matched cohort according to baseline CRP levels. Cumulative incidence curves (Kaplan-Meier estimates) at 90 days for all-cause death (left) and tachyarrhythmias (right) in the propensity score-matched cohort, stratified by baseline CRP <10 mg/L (red) vs CRP ≥10 mg/L (blue). For each endpoint, log-rank p values and HRs with 95% CIs are reported. CRP, C-reactive protein.

In the exploratory 1-year analysis using the 10 mg/L threshold in the unmatched cohort, patients with CRP <10 mg/L had a similar 1-year risk of the primary composite outcome compared with those with CRP ≥10 mg/L (17.2% vs 15.3%; log-rank p=0.277; HR 1.10, 95% CI 0.92 to 1.32). Rates of tachyarrhythmias (14.5% vs 12.5%; log-rank p=0.165; HR 1.15, 95% CI 0.94 to 1.40) and all-cause mortality (3.8% vs 4.0%; log-rank p=0.644; HR 0.92, 95% CI 0.63 to 1.33) were likewise similar between groups, with no statistically significant differences (online supplemental figure 1).

Interestingly, in the exploratory 1-year analysis in the PSM cohort (997 vs 997 patients), the primary composite outcome of all-cause death and tachyarrhythmias occurred less frequently in the CRP <10 mg/L group than in the CRP ≥10 mg/L group (17.3% vs 20.5%; log-rank p=0.059; HR 0.82, 95% CI 0.67 to 1.01), although this difference did not reach statistical significance. Tachyarrhythmias showed a similar pattern (14.5% vs 16.9%; log-rank p=0.138; HR 0.85, 95% CI 0.68 to 1.06), and all-cause mortality remained low and comparable (3.9% vs 5.0%; log-rank p=0.269; HR 0.79, 95% CI 0.51 to 1.21) (online supplemental figure 1).

In the sensitivity analysis using a lower CRP threshold of 3 mg/L, 457 patients (12.6%) had CRP <3 mg/L, whereas 3158 (87.4%) had CRP ≥3 mg/L. In this unmatched cohort, the composite outcome of death and tachyarrhythmias occurred with virtually identical frequency in the two groups (67 of 457 (14.7%) vs 464 of 3158 (14.7%); log-rank p=0.947; HR 0.99, 95% CI 0.77 to 1.28). Rates of tachyarrhythmias alone were likewise similar (56 of 457 (12.3%) vs 393 of 3158 (12.4%); log-rank p=0.893; HR 0.98, 95% CI 0.74 to 1.30). All-cause mortality also did not differ significantly according to the 3 mg/L threshold (12 of 457 (2.6%) vs 103 of 3158 (3.3%); log-rank p=0.470; HR 0.80, 95% CI 0.44 to 1.46) (online supplemental figure 2).

After PSM (457 vs 457 patients), outcomes were reassessed in the balanced cohorts. At 90 days, the composite outcome of all-cause death and tachyarrhythmias occurred in 67 of 457 (14.7%) patients in the CRP <3 mg/L group and 60 of 457 (13.1%) in the CRP ≥3 mg/L group (log-rank p=0.513; HR 1.12, 95% CI 0.79 to 1.59), indicating no significant difference between groups. Individual endpoints showed a similar pattern. Tachyarrhythmias occurred in 56 of 457 (12.3%) patients with CRP <3 mg/L and 49 of 457 (10.7%) with CRP ≥3 mg/L (log-rank p=0.468; HR 1.15, 95% CI 0.78 to 1.69). All-cause mortality also remained low and comparable between groups (12 of 457 (2.6%) vs 14 of 457 (3.1%); log-rank p=0.709; HR 0.86, 95% CI 0.40 to 1.87) (online supplemental figure 2).

Discussion

In this multicentre real-world cohort of more than 3500 adults hospitalised with acute myocarditis, we found that CRP measured at presentation was not consistently associated with 90-day risk of death or tachyarrhythmias. In the unmatched cohort using a conventional threshold of 10 mg/L, patients with CRP <10 mg/L had a numerically higher incidence of the composite outcome and tachyarrhythmias, but these differences were of borderline statistical significance. After PSM, 90-day rates of the composite outcome and its individual components were virtually identical across CRP categories. Exploratory 1-year analyses using the same 10 mg/L threshold showed a similarly inconsistent pattern. In the PSM cohort, there was an unexpected borderline trend towards a higher incidence of the composite outcome in patients with CRP ≥10 mg/L, although this difference did not reach statistical significance. Sensitivity analyses using a lower threshold of 3 mg/L also did not reveal any consistent association between baseline CRP and short-term outcomes in either the unmatched or matched cohorts. Collectively, these results add important information to the long-standing uncertainty surrounding the prognostic value of CRP in acute myocarditis.

The difficulties in establishing a consistent association between CRP levels and clinical outcomes across studies likely arise from several factors, including substantial patient heterogeneity, variation in the timing and sensitivity of CRP measurements and differences in outcome definitions. From a pathophysiological perspective, these findings are plausible. CRP is a downstream marker of systemic inflammation, predominantly driven by interleukin 6 (IL-6) and IL-1 activity, and therefore may not fully reflect the diversity of immune mechanisms involved in myocarditis.^{7 15} The disease encompasses a broad spectrum of inflammatory endotypes, including viral, autoimmune, lymphocytic and eosinophilic forms, each driven by distinct immune pathways that do not uniformly translate into proportional elevations in circulating CRP.² Moreover, the relationship between inflammation and CRP can be modulated by cytokine pathways that diverge from the classical IL-6 and IL-1-driven response. Notably, high levels of type I interferons are known to suppress IL-6-induced CRP production in the liver. This phenomenon, well described in autoimmune diseases such as systemic lupus erythematosus, results in a ‘muted’ or paradoxically low CRP response despite substantial underlying inflammation.^{16 17}

This biological heterogeneity, together with the imperfect correlation between CRP and the underlying inflammatory processes, also helps explain the long-standing lack of solid evidence and the absence of large randomised trials evaluating immunosuppressive therapy in myocarditis.^{6 14} Unlike pericarditis, where inflammation is more uniform and closely mirrored by conventional biomarkers, myocarditis comprises multiple immune-driven phenotypes that do not consistently map onto systemic inflammatory markers, making patient selection, risk stratification and trial design substantially more challenging.

Looking ahead, these findings underscore the need for more refined inflammatory markers to improve risk stratification in acute myocarditis. While hsCRP may offer greater sensitivity in detecting low-grade inflammatory activity, its prognostic value in this context remains uncertain and requires dedicated investigation. In parallel, composite or cellular markers of immune activation, such as the neutrophil-to-lymphocyte ratio, may provide additional insight into inflammatory endotypes that are not adequately captured by CRP alone.¹⁸ Future prospective studies should evaluate the additive value of combining CRP with troponin, N-terminal pro-B-type natriuretic peptide and other markers of immune activation in structured multibiomarker panels.

Limitations

This study has several limitations that must be acknowledged. First, its retrospective design introduces the possibility of selection bias, misclassification and incomplete capture of clinical information. The identification of acute myocarditis relied on ICD-10-CM codes, which lack the diagnostic accuracy of endomyocardial biopsy or advanced imaging such as cardiac magnetic resonance and do not include a specific code for myopericarditis. As a result, the cohort likely includes a mix of confirmed, probable and suspected cases, including possible myopericarditis, thereby introducing dilution bias and potentially attenuating true associations between CRP levels and adverse outcomes.

Second, although PSM was performed to balance key demographic variables, this statistical technique cannot eliminate hidden or paradoxical biases, and residual confounding from unmeasured covariates (including haemodynamic status at presentation, aetiology of myocarditis, immunosuppressive therapy and detailed imaging parameters) remains unavoidable. Additionally, the stringent 1:1 matching algorithm reduced the analytical sample from 3615 to 1994 patients, which may limit the statistical power to detect modest differences in outcomes.

Third, the reliance on CRP values extracted from a large, multi-institutional network introduces methodological variability. Consequently, differences in assay type, analytical sensitivity, reporting units and timing of measurement across sites may have affected the accuracy and comparability of inflammatory profiling. In routine clinical practice, CRP may not be universally obtained in all patients presenting with acute myocarditis; therefore, our cohort may be enriched with patients of moderate severity requiring hospitalisation, predominantly stage C by the 2024 American College of Cardiology Expert Consensus Decision Pathway on Myocarditis,¹⁴ and the findings may not be generalisable to patients at the extremes of the clinical spectrum. Moreover, only a single baseline CRP value was extracted for each patient; the dynamic trajectory of inflammation, which may carry prognostic significance, could not be evaluated.

Fourth, the ascertainment of clinical outcomes was limited by the use of administrative data. Although arrhythmias were included as endpoints in the primary analysis, the capture of arrhythmic events through ICD codes lacks clinical granularity and may misrepresent event burden or timing.

Fifth, variations in follow-up completeness across institutions and the absence of longitudinal data on medication use, immunosuppressive therapy, device implantation or rehospitalisations may have affected outcome classification.

Finally, the observational nature of this analysis precludes any inference of causality. Given the heterogeneity of myocarditis and the limited precision of CRP as a biomarker of immune activity, the absence of a strong association between CRP and outcomes in this study should be interpreted with caution. Furthermore, while the TriNetX Research Network provides a large, multi-institutional dataset, our cohort reflects a very strong representation of US healthcare organisations. Given that the specific distribution and identities of the contributing hospitals and medical centres remain undisclosed, the potential for geographical, institutional and insurance-based selection biases cannot be excluded. Patients with limited access to healthcare, or those managed at community hospitals not represented within the network, may be systematically under-represented. These limitations should be considered when interpreting the generalisability of our findings.¹⁹ Prospective studies with adjudicated diagnoses, standardised biomarker measurement and detailed phenotyping are required to validate these findings and better define the prognostic role of inflammatory markers in acute myocarditis.

Conclusion

In this large, multicentre dataset of adults hospitalised with acute myocarditis, baseline CRP levels were not consistently associated with short-term (90-day) or exploratory 1-year risk of all-cause death or tachyarrhythmias. Using either conventional (10 mg/L) or lower (3 mg/L) thresholds, only modest, non-significant differences in arrhythmic risk were observed in unmatched analyses, and these were no longer evident after PSM on demographic characteristics. These findings suggest that a single CRP measurement at presentation provides limited prognostic discrimination and should not be used in isolation for risk stratification in acute myocarditis. Future studies integrating serial CRP trajectories, more specific inflammatory biomarkers and comprehensive clinical and imaging phenotyping are warranted to refine risk assessment and guide targeted management in this population.

Data availability statement

Data may be obtained from a third party and are not publicly available.