In Reply:-Kleinman suggests a solution to the therapeutic dilemma that arises when faced with the preoperative patient who has a previously undetected systolic murmur due to any source. Although that topic was beyond the scope of the review, it does represent a scenario occasionally encountered in practice. Whereas a fraction of systolic murmurs may be diagnostically challenging, the majority of murmurs can be designated appropriately based on the history, epidemiology, and physical examination. For example, it is unlikely that the asymptomatic 80-yr-old patient with a soft crescendo-decrescendo murmur of aortic sclerosis radiating into the neck, or the 5-yr-old patient with a soft, vibratory functional murmur of childhood has a significant valvular lesion.

Clearly, for those patients who present with symptoms compatible with cardiac disease and a newly detected systolic murmur, postponement and cardiologic evaluation is appropriate. For those asymptomatic patients whose murmur cannot be defined clearly, we agree that cancellation is unnecessary, although a later cardiologic referral may be indicated. With the potential for anaphylaxis in mind, and based on available evidence. Kleinman's suggestion of the preoperative use of orally administered erythromycin for prophylaxis against bacterial endocarditis (BE) during dental and respiratory tract surgery is reasonable. Such a preoperative antibiotic regimen is appropriate whenever you have the advance time of at least 2 h to give it. It should be noted that a theoretical disadvantage of erythromycin is that it is bacteriostatic and not bacteriocidal, as is ampicillin. In patients who have a history of multiple and recent use of ampicillin or amoxicillin without reaction, the recommendation of oral amoxicillin or intravenous ampicillin remains valid. Intravenous use has the advantage of avoiding any delay, as well as avoiding the gastrointestinal distress that frequently accompanies oral erythromycin taken without food. We also note that routine prophylactic use of cephalosporin-based antibiotics, which are frequently given in certain of these settings to prevent wound infections, is generally adequate for BE prophylaxis in respiratory/dental cases.

With regard to patients undergoing gastrointestinal/genitourinary surgery, we agree that use of vancomycin is unnecessary, unless the patient is allergic to penicillin. In addition to the well-known cardiovascular side effects of intravenous vancomycin administration, at our institution, the emergence of vancomycin-resistant enterococcus has necessitated guidelines to limit vancomycin usage; this is one case of appropriate use, namely, BE prophylaxis, when the patient is allergic to penicillin. Again, the frequent use, in these settings, of preoperative intravenous antibiotics such as mezlocillin for surgical prophylaxis in the penicillin-tolerant patient may provide appropriate bacteriocidal activity. Referral to the individual hospital's current drug sensitivities for certain infections may avoid use of additional antibiotics.

However, when a penicillin-based drug seems indicated based on the presence of a significant murmur, that is, any readily audible murmur implying significant turbulence (benign flow murmurs rarely fall into this category), we see little need to cancel surgery that has a risk of bacteremia to obtain from a cardiologist a more accurate diagnosis of the type of cardiac lesion. Regardless of the origin in almost all cases, the presence of such a murmur is the criterion for antibiotic prophylaxis. Patients with cardiac valve lesions, as indicated by a significant murmur (and including those with a history of BE or a prosthetic valve), who are having surgery with a high-risk of transient bacteremia should typically receive antibiotic prophylaxis. We suspect that in the presence of a significant murmur, most cardiologists would, likewise, recommend antibiotic prophylaxis. Clearly, the patient must be informed of the evidence, risks, and various options. A more difficult and perplexing situation would occur if a patient preferred no antibiotic prophylaxis, when he/she is fully aware of the risk of BE (probably small) and anaphylaxis (very small)? In addition to the thorough documentation of discussions with the patient and decisions made as emphasized by Kleinman, one must always recognize the potential for an anaphylactic reaction to the antibiotics, and be prepared for appropriate treatment. In most preoperative settings, we, as anesthesiologists, are probably far better prepared to treat anaphylaxis than those in an outpatient or many other hospital settings.

Finally, we would emphasize that those patients who have mitral valve prolapse diagnosed by history and/or echocardiography, or by the presence of an isolated, intermittent click, but do not have an accompanying murmur, do not need to receive BE-prophylactic antibiotics, with its attendant risks. Our impression of current clinical practice is that this latter point is largely ignored.

Eric Hanson, M.D., Resident in Anesthesiology; Carl Lynch III, M.D., Ph.D., Professor of Anesthesiology, Department of Anesthesiology, University of Virginia Health Sciences Center, Charlottesville, Virginia 22908.

(Accepted August 28, 1996.)