We thank Drs. Zwart and Sachs for their interest in our paper. We largely agree with the overall tenor of their comments. However, we do not believe that any good evidence supports a benefit from administering antibiotics to patients with non-group A streptococcal pharyngitis. In Zwart and colleagues' study (1), the results are neither clinically nor statistically compelling; indeed, in the Discussion section of that paper, the authors are (appropriately) much more cautious about non-group A streptococci than they are in their letter. We are therefore not convinced of the need for "better tests" to identify such infections.

With regard to testing in a more global sense, we tried to have our recommendations reflect the tradeoffs between a test-and-treat strategy and a strategy consisting entirely of empiric treatment of pharyngitis based on the validated clinical score described by Centor. We agree with Zwart and Sachs and the authors of another recent review (2) that this clinical screen provides an excellent basis for decision making. The particular test-and-treat strategy that we also recommended as an alternate appropriate approach is likely to lead to antibiotic treatment of somewhat fewer patients without group A streptococcal pharyngitis than will the purely empiric strategy; how-ever, this will come at the cost of testing a very large number of patients. It will also result in failure to treat some highly symptom-atic patients with group A streptococcal pharyngitis who have a false-negative result on a rapid antigen screening test. We believe that available evidence supports either of these strategies as reasonable and as justifiable in current clinical practice.

Finally, we stress that all evidence-based guidelines should be reevaluated periodically and revised as appropriate in the light of new information. If new evidence convincingly demonstrates the value of identifying and treating non-group A ß-hemolytic streptococcal in-fections, it should certainly be incorporated into future guidelines.