Cerebral MRI in Endocarditis

Abstract & Commentary

By Michael H. Crawford, MD. Dr. Crawford is Professor of Medicine, Chief of Cardiology, University of California, San Francisco. Dr. Crawford is on the speaker's bureau for Pfizer. This article originally appeared in the June 2010 Clinical Cardiology Alert. It was peer reviewed by Ethan Weiss, MD. Dr. Weiss is Assistant Professor of Medicine, Division of Cardiology and CVRI, University of California, San Francisco. Dr. Weiss reports no financial relationships relevant to this field of study.

Source: Duval X, et al. Effect of early cerebral magnetic resonance imaging on clinical decisions in infective endocarditis. Ann Intern Med. 2010;152:497-504.

The presence of cerebral complications of infective endocarditis (IE) can affect diagnostic and treatment decisions. Thus, this group from Paris, France, evaluated whether early cerebral MRI would affect the diagnosis and management of hospitalized patients suspected of having IE. In addition to blood cultures, echocardiograms, and other routine tests, cerebral MRI was performed within seven days of admission and evaluated independently by blinded neuroradiologists. A cardiologist and an infectious-disease specialist evaluated the patients < 24 hours before the MRI, as well as after the MRI results were known. The patients were followed for six months. Of the 274 patients admitted for suspected IE, 150 were included in the study. Of these, 20 were excluded for inadequate MRI images or because the diagnosis of IE was immediately excluded. Of the remaining 130 patients, 59% had definite IE, 38% possible IE, and IE was ultimately excluded in 2%. Neurologic abnormalities were present in 16 patients; 13 of these 16 patients had CT scans that indicated lesions likely due to IE. MRI detected cerebral lesions in 106 (82%), including ischemic lesions in 68, microhemorrhages in 74, and aneurysms in 10. Based upon these findings, 17 of the 53 non-definite IE cases (32%) were upgraded to definite or possible IE, and therapeutic plans were changed in 24 of the 130 patients (18%). Overall, 36 of 130 patients (28%) had their diagnosis or therapeutic plan modified by the MRI results. The authors concluded that many patients with IE but no symptoms had cerebral lesions identified by MRI that modified their diagnosis and treatment.


Autopsy studies have shown a high incidence of cerebral lesions in confirmed IE cases, but most patients do not have neurologic symptoms (12% in this series). Given the difficulty in diagnosing IE and the high mortality despite modern diagnostic and therapeutic approaches, any test that would augment the diagnosis and refine therapy would be welcome. These investigators sought to evaluate whether cerebral MRI is such a test. Routine early MRIs in this study demonstrated lesions in 82% of suspected IE cases. The most common finding was microhemorrhages, which are not specific for IE. Microhemorrhages are seen in hypertension, amyloid vasculitis, and other conditions. Second most common were ischemic lesions, and third were aneurysms, both of which are more likely to be IE complications and are included as minor criteria in the Duke classification. Excluding microhemorrhages, and based upon the more specific MRI findings, almost one-third of patients were upgraded to definite or possible IE by the Duke criteria. Therapeutic plans were altered based upon the MRI in most of these patients. Although this would seem beneficial, there are no data to suggest that outcomes are different in this observational study. Also, the costs of routinely performing cerebral MRI in all suspected IE cases were not explored. Not only would the direct cost of the MRI be relevant, but the indirect costs of false-positive diagnoses, more surgery, etc., would be as well.

Another problem with an observational study is that there are no control groups and no age-matched normal groups. Consequently, we not only don't have relative outcome data, but we don't know how specific these MRI findings are for IE. Mycotic aneurysms would be expected to be specific, but only 8% had them. Cerebral abscesses would also likely be due to IE, but only 6% had them. Ischemic lesions were much more common, but their specificity is not completely known. On the other hand, all the findings on the MRI in this series could represent vascular phenomena due to embolic, generalized vasculitis or immune phenomena of IE. Since vascular phenomena are a minor criterion in the Duke criteria, these findings would skew the diagnosis toward IE and would likely alter therapy, as was seen in this study. Interestingly, the total mortality in this series was 15%, which is low for an IE series. However, patients who went straight to surgery, or were too sick to have an MRI, were excluded.

The diagnosis of IE has evolved over the last 40 years. In the mid-20th century, it was clinical exam and blood cultures; then, echocardiography, especially transesophageal, changed things. In most patients, the diagnosis is now made by echo before the blood cultures are back. MRI and other imaging modalities may now alter our diagnostic approach again, and we may soon need new diagnostic criteria.