The research team, led by Sukruth Pradeep Kundur and Daniel I. Bromage of King’s College London, assembled a cohort of 243 consecutive patients aged 18 or older whose acute myocarditis was confirmed by cardiac magnetic resonance imaging according to the 2018 updated Lake Louise criteria, with endomyocardial biopsy supporting the diagnosis in a small proportion of cases. Patients were identified using CogStack, an open-source clinical data retrieval platform developed at King’s College London, supplemented by Hospital Episode Statistics data and dedicated inflammatory heart disease clinic records. Crucially, the investigators excluded patients with COVID-19 or vaccine-related myocarditis, cardiac sarcoidosis, and myocarditis secondary to immune checkpoint inhibitor use, because these represent distinct inflammatory myocardial syndromes with different pathophysiology. Obstructive coronary artery disease was ruled out by invasive coronary angiography in 112 patients and by stress perfusion cardiac magnetic resonance in one patient, ensuring that the diagnosis reflected genuine myocardial inflammation rather than ischaemic injury.

Of the 243 patients included, 63 received colchicine during their index admission while 178 did not, with colchicine status unavailable for two patients. The median follow-up extended to 53 months, and approximately two-thirds of the overall cohort were male with a median age of 36 years. The colchicine-treated patients were significantly younger, with a median age of 29 years compared with 39 years in the untreated group, although the sex distribution was similar between groups. Clinical presentation differed markedly: an acute coronary syndrome-like presentation with chest pain was far more frequent among colchicine-treated patients, occurring in 94 percent versus 72 percent of untreated patients, whereas heart failure and arrhythmia presentations were more common in those who did not receive the drug. Hypertension was also less frequent in the colchicine group, affecting just 3 percent compared with 18 percent of untreated patients.

Despite these demographic and phenotypic differences, baseline ventricular function was comparable between the two groups. The median cardiac magnetic resonance left ventricular ejection fraction was 57 percent in both groups, and the median admission echocardiographic ejection fraction was 55 percent versus 54 percent. Notably, no colchicine-treated patient had severely reduced admission ejection fraction below 30 percent, compared with 11 of 145 evaluable untreated patients. N-terminal pro-B-type natriuretic peptide was higher in colchicine-treated patients, although this biomarker was missing in a third of the cohort. Concomitant therapy also differed, with untreated patients more often receiving aspirin or non-steroidal anti-inflammatory drugs, statins, immunosuppressive therapy, and corticosteroids, reflecting the higher-risk phenotype concentrated in that arm.

The primary outcome was echocardiographic maintenance or recovery of left ventricular ejection fraction to at least 50 percent at follow-up. The investigators chose this dichotomous threshold rather than continuous ejection fraction change because continuous measures are subject to a ceiling effect, whereby patients with preserved or near-preserved admission function have limited scope for further gain. Among the 93 patients with evaluable follow-up echocardiography, the outcome occurred in 18 of 19 colchicine-treated patients, or 95 percent, compared with 52 of 74 untreated patients, or 70 percent, a difference that reached statistical significance with a p-value of 0.035. The association persisted after adjustment for age and sex using Firth penalised logistic regression, a technique that provides finite estimates when event counts are sparse, yielding an odds ratio of 5.10 with a 95 percent confidence interval of 1.14 to 48.7 and a p-value of 0.03.

Restricted to the 47 patients who had a reduced admission ejection fraction below 50 percent and therefore required genuine recovery rather than mere maintenance of function, recovery to at least 50 percent occurred in 7 of 8 colchicine-treated patients and 22 of 39 untreated patients. Although the unadjusted odds ratio of 5.41 in this subgroup did not reach conventional significance with a p-value of 0.13, the direction of association was consistent with the primary analysis. In a propensity score-matched sensitivity analysis using age and sex, which produced a matched cohort of 126 patients with all standardised mean differences below 0.1, follow-up ejection fraction of at least 50 percent occurred in 95 percent of colchicine-treated patients versus 71 percent of untreated patients, with an odds ratio of 6.55 that remained statistically significant. Expanded propensity score models incorporating clinical presentation, admission ejection fraction, hypertension, C-reactive protein, and peak troponin produced point estimates for recovery ranging from 3.89 to 5.00 across matching, inverse probability weighting, and covariate-adjusted regression methods.

All-cause mortality, a secondary outcome, told a different story. Death occurred in 1 of 63 colchicine-treated patients, or 1.6 percent, compared with 12 of 178 untreated patients, or 6.7 percent. However, after age and sex adjustment the association was not statistically significant, with an odds ratio of 0.58 and a confidence interval of 0.06 to 2.76, reflecting the single event in the colchicine arm. The cohort was simply underpowered to detect any mortality difference. Age was independently associated with death, and univariable analysis identified arrhythmic presentation, hypertension, and diabetes as each associated with all-cause mortality, consistent with previous epidemiological studies. Absolute change in ejection fraction and improvement of at least 10 percentage points did not differ significantly between groups, with linear regression identifying lower admission ejection fraction as the dominant predictor of ejection fraction change rather than colchicine use itself.

The biological rationale for colchicine in myocarditis is grounded in its mechanism of action. Colchicine disrupts microtubule polymerisation and impairs NLRP3 inflammasome assembly, reducing downstream interleukin-1 beta and interleukin-18 mediated inflammation. Both processes are particularly relevant in acute myocarditis, where innate immune activation and inflammatory cell recruitment contribute to ongoing myocardial injury with direct implications for ventricular function. In a murine model of coxsackievirus B3 myocarditis, colchicine improved left ventricular function and reduced NLRP3 inflammasome activity without impairing viral clearance. Equally, acute myocarditis is thought to be driven largely by an adaptive CD4-positive Th1/Th17 response, which colchicine does not directly target, suggesting that any detectable effect would act on the innate inflammasome pathway rather than the immune trigger itself.

The findings are directionally consistent with emerging observational evidence. In a large TriNetX real-world analysis of adults hospitalised with acute myocarditis, Golino and colleagues reported lower 90-day rates of all-cause death, ventricular arrhythmias, and acute heart failure among colchicine-treated patients after propensity score matching, with a hazard ratio of 0.66. However, that analysis relied on administrative ICD codes and incomplete treatment information. In patients with myopericarditis, Collini and colleagues similarly found fewer recurrences after colchicine treatment, with recurrence rates of 19.2 percent versus 43.8 percent after matching. Colchicine’s broader cardiovascular credentials include the COLCOT trial in patients following myocardial infarction and the LoDoCo2 trial in chronic stable coronary artery disease, both of which demonstrated significant reductions in atherothrombotic events. The 2025 European Society of Cardiology guidelines assign colchicine a Class IIa recommendation in myocarditis when there is pericarditis with concomitant myocardial involvement.

The authors interpret their results cautiously as hypothesis-generating rather than causal inference. Colchicine-treated patients were younger, more frequently presented with chest pain, and had a lower comorbidity burden, and the higher-risk phenotype was underrepresented in the colchicine arm. There is also a possible confounding by indication attributable to concurrent pericarditis involvement, since colchicine already carries a guideline recommendation specifically in the context of myocarditis with concomitant pericarditis, and systematic data on pericardial assessment were not available in this cohort. The primary outcome was evaluable in only 93 of 243 patients, with follow-up echocardiography missing in 70 percent of colchicine-treated patients and 58 percent of untreated patients, and missingness was associated with baseline disease severity. Nevertheless, two ongoing prospective randomised controlled trials, the ARGO trial in hospitalised acute myocarditis and the CMP-MYTHiC trial in chronic inflammatory cardiomyopathy, are specifically evaluating colchicine across the spectrum of myocardial inflammation, and this study contributes patient-level data from a cohort with confirmed diagnosis to the evolving evidence base supporting their prospective evaluation.

Subject of Research: Colchicine and clinical outcomes in acute myocarditis

Article Title: Colchicine use and clinical outcomes in acute myocarditis: a retrospective cohort study

Article References: Pradeep Kundur, S., Sadler, M., Shamsi, A., Madaudo, C., Chick, W., Noakes, D., Manzoni, M., Perkan, B., Scott, P. A., Cannatà, A., & Bromage, D. I. (2026). Colchicine use and clinical outcomes in acute myocarditis: a retrospective cohort study. Frontiers in Cardiovascular Medicine, 13, Article 1930042. https://doi.org/10.3389/fcvm.2026.1930042

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1930042

Keywords: colchicine, acute myocarditis, left ventricular ejection fraction, cardiac magnetic resonance, anti-inflammatory therapy, heart failure, clinical, outcomes, acute, myocarditis, retrospective, cohort