By Michael H. Crawford, MD, Editor

SYNOPSIS: This multicenter study revealed cardiac MRI may be useful in the diagnosis of difficult pericarditis cases, especially if pericardial edema and late gadolinium enhancement are found. Pericardial thickening and elevated C-reactive protein were found to be predictive of recurrent pericarditis and other complications.

SOURCE: Imazio M, Pivetta E, Palacio Restrepo S, et al. Usefulness of cardiac magnetic resonance for recurrent pericarditis. Am J Cardiol 2020;125:145-151.

In situations where clinical, electrocardiographic, and echocardiographic criteria cannot establish the diagnosis of pericarditis, cardiac MRI has been recommended. Diagnosis may be especially challenging in recurrent pericarditis.

Investigators from Italy conducted a multicenter observational study of consecutive patients with suspected recurrent pericarditis who were evaluated by cardiac MRI. The diagnosis of acute pericarditis was made when two or more of four clinical criteria were met. Also, C-reactive protein (CRP) was measured to assess its value. These patients were followed for at least 18 months for the adverse events of recurrences, tamponade, and constriction. Cardiac MRI was performed as soon as possible after the onset of symptoms to assess pericardial thickness, pericardial edema/inflammation, myocardial and pericardial fibrosis by late gadolinium enhancement (LGE), and pericardial effusion. In addition, all cardiac MRI cases were matched with patients without pericarditis who had undergone cardiac MRI on the same day.

Inclusion criteria were met in 128 patients who underwent a cardiac MRI a mean of 12 days after the onset of symptoms. In 92 of the 128 patients, two or more of the four cardiac MRI diagnostic findings were found. Pericardial edema and LGE had an area under the curve (AUC) for diagnosing pericarditis of 0.80 and 0.76, respectively. Pericardial thickening and effusion were less predictive at AUCs of 0.64 and 0.71, respectively. The combination of pericardial edema and LGE had a sensitivity of 73% and a specificity of 99%. During a mean follow-up of 34 months, 52% of the patients had a recurrence, 6% tamponade, and 11% constriction. In a multivariate model, an elevated CRP (hazard ratio [HR], 11.7; 95% confidence interval [CI], 5-27.2) and cardiac MRI pericardial thickening (HR, 2.6; 95% CI, 1.6-4.4) were predictors of adverse events during follow-up. LGE predicted a lower risk (HR, 0.3; 95% CI, 0.1-0.7). A prognostic model using sex, age, CRP, and all four cardiac MRI variables had a C-index of 0.84. The authors concluded cardiac MRI has a high diagnostic accuracy for the diagnosis of acute pericarditis and may identify patients at risk for complications during follow-up.


Acute and recurrent pericarditis can be difficult to diagnose, and pharmacologic treatment can produce considerable side effects. Consequently, accurate diagnosis is desirable. Currently, diagnosis is based on four clinical criteria, with two of four required to make the diagnosis. Chest pain is present in 40% of cases, but it is nonspecific. Only one-third of patients have pericardial rubs auscultated and < 60% have classical ECG changes. Many do not exhibit detectable pericardial effusions by echocardiography. CRP often is elevated, but also is nonspecific and not considered a key diagnostic finding. Thus, another diagnostic technique such as CMR could be useful.

In this study, all cardiac MRI criteria showed a high specificity (90-100%) and positive predictive value (84-100%). However, their sensitivity (30-73%) and negative predictive value (58-78%) were lower. However, these cardiac MRI criteria are better than almost all the clinical criteria. Also, cardiac MRI pericardial thickening plus an elevated CRP was the best cardiac MRI-clinical combination for identifying future complications due largely to enhanced prediction of pericardial constriction. Pericardial thickening was observed in 80% of those destined to experience pericardial constriction. Interestingly, the combination of older age and pericardial LGE were associated with less complications.

The major limitation of this study was the delay in obtaining cardiac MRI. The mean time from first symptoms to cardiac MRI was 12 days, which may have mitigated some of the diagnostic features. Cardiac MRI was most helpful for diagnosis and prognosis if performed within two weeks. However, this was a real-world study, and cardiac MRI is expensive, time-consuming, and unpleasant. It often is avoided unless really needed or indicated. At this time, the classic four clinical plus echocardiographic findings make up the initial evaluation of suspected pericarditis, with the possible addition of CRP. Cardiac MRI should be considered early if the diagnosis is unclear or if recurrent pericarditis is suspected. Cardiac MRI can help make the diagnosis and is predictive of complications. In classic cases that respond well to therapy, cardiac MRI probably is unnecessary.