On the track of new approaches to myocarditis

A 2014 study that evaluated a panel of biomarkers to diagnose patients with suspected myocarditis looked at hs-TnT, N-terminal B-type natriuretic peptide (NT-proBNP), copeptin, and mid-regional pro-adrenomedullin (MR-proADM) (Ukena C, et al. Clin Res Cardiol. 2014;103[9]:743–751). “In patients with biopsy-proven acute myocarditis, high-sensitivity troponin concentration is dramatically higher compared with the chronic myocarditis and nonmyocarditis groups,” Dr. Li said of the study findings. “However, none of the other biomarkers showed any predictive value for the diagnosis of myocarditis.”

In their study published in 1997, Smith, et al., also analyzed the troponin levels in 32 of the 53 patients with myocarditis who had information about the duration of their heart failure symptoms, Dr. Li said. “The troponin level was significantly higher in the 20 patients with symptoms lasting less than one month compared with the 12 patients with symptom duration greater than one month. And 11 of the 20 patients with symptoms for less than one month had the value of elevated troponin level, but only one of the 12 patients with the longer duration of symptoms had elevated troponin. So troponin might be an early indicator of heart failure caused by myocarditis.”

In examining the relationship between the histopathologic pattern and severity of myocarditis to troponin elevation, the authors found that the mean troponin level was increased in the diffuse myocarditis group, but the difference did not achieve statistical significance, Dr. Li said. “So troponin elevations do not always correlate with histological severity of myocarditis.”

Authors of another earlier study reported that when a 40-year-old patient diagnosed with acute myocarditis was prescribed prednisolone, her troponin level decreased from 10 ng/mL to 0.5 ng/mL (Quiroz R, et al. J Heart Lung Transplant. 2010;29[7]:820–822). “However, when prednisolone was stopped, the troponin level increased to 12 ng/mL,” Dr. Li said, adding that troponin levels did not correlate well with BNP levels.

When the patient’s prednisolone was restudied three weeks after discharge, her symptoms had improved and her troponin levels had declined to 0.5 ng/mL. “The troponin levels may not indicate the ongoing cardiomyocyte necrosis, but rather imply a poor prognosis,” Dr. Li said.

In another example of the use of serial troponin and BNP levels to monitor myocarditis therapy, authors of a case report found the troponin and BNP levels of an acute myocarditis patient declined after conventional therapy for heart failure (Kadota S, et al. J Cardiol. 2008;52[2]:154–158). However, “six months later, the patient showed clinical symptoms—general fatigue and dyspnea—and was readmitted,” she said. “The echocardiogram showed severe left ventricular dysfunction, and both troponin and BNP levels were significantly elevated again.”

There was neither elevation of creatine kinase nor ST-T change on the EKG, she said, and the endomyocardial biopsy showed severe lymphocytic infiltration, which indicated active myocarditis. After treatment with immunosuppressant drugs, troponin levels decreased dramatically. BNP levels also decreased, though much more slowly. The second endomyocardial biopsy showed mild lymphocytic infiltration. “So the dynamic change of troponin and BNP in a patient following acute myocarditis might be the indicator for the patient’s response to treatment,” Dr. Li said.

The same finding was achieved in COVID-19 patients at OSU (Scarl RT, et al. Ann Cardiol Vasc Med. 2021;4[1]:1041). In 81 patients hospitalized for COVID-19, Dr. Li and colleagues observed the serial measured troponin levels were dramatically elevated in nonsurvivors compared with survivors. “Thus, this could be a good marker for severe myocarditis cases,” Dr. Fairweather says. “However, less severe cases either recover or may progress to DCM and may not be captured by high-sensitivity troponin as a biomarker.”

The OSU findings are consistent with those of others, Dr. Li said (Sandoval Y, et al. J Am Coll Cardiol. 2020;76[10]:1244–1258; Zhou F, et al. Lancet. 2020;395[10229]:1054–1062). Sandoval, et al., write in their review of cardiac troponin in assessing myocardial injury in COVID-19, “The structured use of serial cardiac troponin has the potential to facilitate risk stratification, help make decisions about when to use imaging, and inform stage categorization and disease phenotyping among hospitalized COVID-19 patients.”

“So serial measured troponin levels might be a good approach for the diagnosis and prognosis of myocardial injury or myocarditis caused by the SARS-CoV-2 virus,” Dr. Li said. “Theoretically,” she tells CAP TODAY, “elevation of blood troponin levels indicates ongoing cardiomyocyte necrosis. It might be elevated with lymphocytic infiltration in cardiac tissues resulting from a virus attack, but it has not been confirmed yet.”

In her group’s study, Dr. Li says, the dynamic change of elevation of troponin with hospitalization days has a similar trend with that of other inflammation biomarkers (IL-6, ferritin, PCT, and D-dimer), “indicating that the release of troponin in myocarditis/myocardial damage resulting from COVID-19 is closely associated with systemic inflammation.”

In a 2019 study of 56 children with viral myocarditis, the serum concentration of CK-MB, TNFα, and high-sensitivity CRP were found to be statistically different in the acute stage, recovery stage, and the control group (Chen J, et al. Open Life Sci. 2019;14:38–42). “The levels gradually decreased as the disease progressed,” Dr. Li said. “So the dynamic change in the inflammatory factors, like TNF-alpha and high-sensitivity CRP, might reflect the development of viral myocarditis in children. And they might be used as serology markers to identify viral myocarditis in the acute and recovery phases.”

A study published in the 2021 European Journal of International Medicine confirmed for troponin a well-defined diagnostic role in myocarditis, Dr. Li said (Scicchitano P, et al. Eur J Intern Med. 2021;85:56–62). The aim of this multicenter, retrospective study of 104 patients with myocarditis was to provide the DAME (Diagnosis of Acute Myocarditis in Emergency) score for the fast identification of such patients in the emergency department.

The determinants of the score were fever, chest pain, ESR > 20 mm/h, hs-CRP > 3 mg/L, hs-troponin serum levels > 3 ng/L, and left ventricular ejection fraction < 50 percent. Each received a score ranging from zero to four, and a final score of four or greater was related to a 75 percent probability of myocarditis. A score between one and four was related to a 57 percent probability.

Pietro Scicchitano, MD, PhD, a coauthor of the study and professor, cardiovascular diseases section, Department of Emergency and Organ Transplantation, University of Bari in Italy, tells CAP TODAY: “The DAME score better pointed out the relevance of a comprehensive evaluation of patients when admitted to the emergency department for a chest complaint. High-sensitivity troponin I is included in the score and acquires a more reliable use for the overall assessment of patients admitted to the ED, beyond acute coronary syndromes.” The study confers to troponins the role for excluding heart lesions, he says, “although they should be evaluated with further parameters.”

Dr. Li says the DAME score might help prevent delay in diagnosing myocarditis, but routine use of the score, Dr. Scicchitano says, hasn’t caught on in the scientific community. “We usually adopt it in our clinical practice,” he says, “but we hope to implement it beyond our hospital.”

Like the DAME score, Dr. Li says, machine learning can be of use in diagnosing or monitoring myocarditis, but here, too, as with serial troponin measurement, sample size is the challenge. “It is never easy to have a study on myocarditis with a large sample size,” she says, adding, “All the published studies just contribute the best they can.”

Mortality prediction in pediatric myocarditis was the subject of a study published last year in which the authors compared the performance of traditional logistic regression models with that of a machine-learning-based model. Logistic models were good at predicting patients who were likely to survive, they found, but performed poorly in predicting patients who were likely to die. “On the contrary,” they write, “the ML model had a good balance between sensitivity and specificity, and would be much more useful in clinical settings where predicting mortality is more crucial than predicting survival.”

The probability of mortality can be satisfactorily approximated when any of these variables are present, they say: mechanical ventilation, ECMO use, cardiac arrest, ventricular fibrillation, and acute kidney injury.

Fu-Sheng Chou, MD, PhD, a coauthor of the study and neonatologist at Kaiser Permanente in Riverside, Calif., says the ML model “exemplifies the power of data and data science.”

“Causal interference and predictive modeling are now at our fingertips, thanks to the computing power and the emergence of the data science discipline,” he tells CAP TODAY.

He and his coauthor reported in the study that it was “striking” to them that acute kidney injury carried considerable weight in putting patients at a significant risk for mortality. He says AKI in the context of pediatric myocarditis may be a result of poor vital organ perfusion because of poor contractility, or cardiogenic shock. “In the context of overwhelming sepsis, distributive shock may also contribute to acute kidney injury. We were surprised that its association with mortality has not been established in the literature despite it being a reasonable indicator of overall disease severity.” They speculate, Dr. Chou says, that it’s due to the rarity of the disease. Identifying AKI as an important predictor of mortality affirms the importance, he says, of having multicenter datasets and well-curated data in health care research.

He applauds the Healthcare Cost and Utilization Project for its efforts in publishing the Kids’ Inpatient Database used in his study to investigate risk factors associated with myocarditis “and to turn the identified risk factors into a prediction tool.” The database has its limitations, he says, but he hopes the findings of his study will serve as initial “observations” and prompt other retrospective and prospective studies with dedicated cohorts on the pathophysiology underlying the cause-effect link between the identified risk factors and mortality outcome.

“Pediatric myocarditis is a devastating yet understudied disease,” Dr. Chou says.

Amy Carpenter Aquino is CAP TODAY senior editor.