INFECTIVE ENDOCARDITIS: DIAGNOSTIC CRITERIA
The diagnosis of bacterial endocarditis has long been based on a constellation of history, physical examination findings, laboratory data, including blood cultures, and an assessment of the patient’s risk factors and underlying diseases. Many patients complain only of fever and fatigue and are found to have a new cardiac murmur. The disease, however, presents in a myriad of ways. Friable vegetations may result in embolic features, such as stroke, meningitis, blindness, myocardial infarction, or arterial occlusion. Some patients will initially appear septic, some will present with autoimmune disease, and others with congestive heart failure secondary to rapid destruction of a heart valve. Embolic signs on physical examination, such as Osler’s nodes, Janeway lesions, Roth’s spots, and splinter hemorrhages are less often seen today because of the rapid institution of antibiotic therapy in most patients. Although positive results on blood cultures are very helpful in diagnosis, some patients with endocarditis will have persistently negative cultures. Other patients with positive blood cultures may appear to have endocarditis but be found to have some other focus of infection instead. For these reasons, standardized diagnostic criteria for endocarditis have long been sought.
In 1981, von Reyn et al. published criteria for the diagnosis of endocarditis based on clinical-pathologic criteria. These criteria were helpful particularly because of the wider spectrum of presentation of the disease in recent decades, including the more subtle presentation that often occurs in the elderly. The von Reyn criteria are listed in Table 10-1. Endocarditis is categorized as definite only when the characteristic histology is demonstrated on a surgical specimen or at autopsy, or when bacteria are cultured from a heart valve or peripheral embolus. Cases are further categorized as probable endocarditis, possible endocarditis, or diagnosis rejected.
Table 10-1. The von Reyn criteria for diagnosis of infective endocarditis
During the past several years, the von Reyn classification of diagnosis has become less useful for several reasons. Most importantly, the use of transthoracic echocardiography has become a major tool in the diagnosis of endocarditis, and results of this test need to be included in the overall assessment of the likelihood of endocarditis. Echocardiography is not part of the von Reyn diagnostic criteria. The von Reyn criteria were not studied prospectively, but recent reports suggest that some patients in whom the diagnosis is rejected by these criteria do in fact have the disease. The von Reyn criteria do not emphasize the importance of intravenous drug abuse as an extremely important predisposing factor.
In 1992, Lukes et al. proposed an endocarditis classification system that has come to be known as the Duke criteria. This classification system was published in 1994 and included an analysis of 67 patients with pathologically proven endocarditis, in whom the system proved to be very sensitive. The original article did not study the specificity of the classification method. Prospective studies have come to prove that this system is more sensitive and specific than the von Reyn system. The Duke approach classifies cases as definite, possible, or rejected on the basis of a scoring system of major and minor criteria (Table 10-2). For endocarditis to be diagnosed definitively by the system, a patient must show either histologic or pathologic evidence of the disease or exhibit definitive clinical criteria. Two major criteria, one major and three minor criteria, or five minor criteria provide sufficient evidence for a definitive diagnosis. Table 10-3 delineates the definitions of major and minor criteria.
Table 10-2. Proposed new criteria for diagnosis of infective endocarditis (Duke University)
Table 10-3. Definition of terms used in the proposed diagnostic criteria (Duke University)
Much like the von Reyn system, the Duke system classifies as definite any case of endocarditis for which there is evidence based on histology from surgery or autopsy or on direct culture from a vegetation or peripheral embolus. Unlike the von Reyn system, the Duke system can be used to make a definite diagnosis of endocarditis even without direct histologic or culture evidence. This definitive diagnosis requires either two major criteria, one major and three minor criteria, or five minor criteria. It should be noted that major criteria are related either to blood culture data or echocardiographic data. Positive results on blood cultures are not considered major criteria per se. Typical microorganisms, such as a-hemolytic streptococci, isolated from two separate cultures stand as a major criterion. Persistently positive blood cultures with an organism that can cause endocarditis would also count as a major criterion. Three other major criteria are based on data available from echocardiography. They are (a) an oscillating intracardiac mass in the absence of an alternative anatomic explanation (i.e., other than endocarditis); (b) abscess; and (c) a new partial dehiscence of a prosthetic valve or new valvular regurgitation. Hence, the definitive diagnosis of endocarditis in more easily made in the Duke system than in the von Reyn system because of the use of echocardiographic data.
The sensitivity and specificity of transesophageal echocardiography for the diagnosis of endocarditis was established between 1981 and 1994, when the two classification schemes were established. The Duke system, unlike the von Reyn, can accept a definitive diagnosis based on clinical criteria as long as enough clinical criteria are available. Five minor criteria establish the diagnosis as definitive, whereas the same clinical data would result only in a classification of possible endocarditis by the von Reyn system. Hence, according to the Duke system, a patient having blood cultures positive for a characteristic organism and positive findings on echocardiogram would fulfill two major criteria and be given a definitive diagnosis. Similarly. a patient with Staphylococcus aureus in the bloodstream (with no other focus of infection) and new valvular regurgitation by echocardiography would also fulfill two major criteria and be given a definitive diagnosis. Five minor criteria—such as fever, Janeway lesion, intravenous drug abuse, Osler node, and glomerulonephritis—would also be a basis for a definitive diagnosis.
Because at least some patients will be available for study who have definitive evidence for infective endocarditis by surgical or autopsy material, the sensitivity of the von Reyn and Duke criteria can be assessed in these cases. At least six studies have compared the Duke criteria with the von Reyn criteria prospectively. By the Duke criteria, 83% of the confirmed cases were considered definitive. None of the confirmed cases was rejected. By the von Reyn methodology, 48% of the confirmed cases were classified only as probable. Twenty-one percent of the confirmed cases would have been rejected by this system.
The usefulness of the Duke criteria was evaluated by reviewing cases during a 3-year period at 54 hospitals in the Philadelphia area. The clinical judgment of three infectious disease experts who reviewed the records of 410 patients was compared with classification that would be generated by the Duke system. There was excellent agreement (91%) for possible and probable cases. However, the experts found 36 cases that they did not feel were likely to be endocarditis but that would have been classified as definite or probable by the Duke criteria. The authors warn that although the Duke criteria are very sensitive, they may result in overdiagnosis of infective endocarditis.
The percentage of endocarditis cases in patients with prosthetic heart valves is increasing. The Duke criteria were studied in 25 patients with pathologically confirmed prosthetic valve endocarditis. Seventy-six percent of these cases were classified as definitive by Duke criteria, and none were rejected. The von Reyn criteria rejected five cases, or 20% of the confirmed cases.
Additional prospective studies will be required to confirm the validity of the Duke diagnostic criteria. The specificity of the system, in particular, needs further evaluation. The sensitivity of the methodology seems certain. Standardized systems for the diagnosis of endocarditis will be extremely important in prospective clinical studies evaluating new diagnostic or treatment methods. However, for the diagnosis of endocarditis in any particular patient, no standardized methodology can replace a clinician’s skills and judgment. (S.L.B.)
Bayer AS, et al. Evaluation of new clinical criteria for the diagnosis of infective endocarditis. Am J Med 1994;96:211.
Sixty-three febrile patients with suspected infective endocarditis who had open heart surgery were evaluated, and the von Reyn and Duke criteria for endocarditis were compared. The Duke criteria were superior predominantly because of the use of transthoracic echocardiographic data.
Cecchi E, et al. New diagnostic criteria for infective endocarditis. A study of sensitivity and specificity. Eur Heart J 1997;18:1149.
Italian study in which 143 patients with suspected endocarditis had long-term follow-up. The sensitivity and specificity of the von Reyn and Duke criteria were compared. The Duke criteria were more sensitive and specific than the von Reyn criteria.
Durack DT, et al. New criteria for the diagnosis of infectious endocarditis: utilization of specific echocardiographic findings. Am J Med 1994;96:200.
Establishes the Duke criteria and uses the system to classify 400 patients as definite, possible, or rejected. The system was 80% sensitive in classifying 69 proven cases. No attempt to determine specificity was made.
Hoen B, et al. The Duke criteria for diagnosing endocarditis are specific: analysis of 100 patients with acute fever or fever of unknown origin. Clin Infect Dis 1996;23:298.
In a study of patients with acute fever admitted to medical wards, the Duke criteria were 99% specific (i.e., cases of acute fever were not misdiagnosed as endocarditis when the Duke criteria were used).
Nettles RE, et al. An evaluation of the Duke criteria in 25 pathologically confirmed cases of prosthetic valve endocarditis. Clin Infect Dis 1997;25:1401.
The authors used 25 cases of pathologically confirmed prosthetic valve endocarditis to compare the von Reyn and Duke criteria for diagnosis. When the Duke method was used, 76% of confirmed cases were considered definite. No cases were rejected. The von Reyn method would have rejected five cases or 20% of the total.
Sekeres MA, et al. An assessment of the usefulness of the Duke criteria for diagnosing active infective endocarditis. Clin Infect Dis 1997;24:1185.
Infectious disease experts reviewed the charts of 410 patients with suspected endocarditis for 3 years at 54 hospitals in Philadelphia. Cases were classified as definite, probable, or possible, and then results were compared with the Duke method of classification. The sensitivity of the Duke method was good to excellent, but some concern about specificity is expressed by the authors.
von Reyn, et al. Infective endocarditis: an analysis based on strict case definitions. Ann Intern Med 1981;94:505.
Established case definitions for endocarditis that have been widely used, especially for clinical studies. The system was not tested prospectively and was developed before breakthroughs in diagnosis by transesophageal echocardiography.