The use of samples from routine microbiology and "nonstreptococcal" samples will tend, contrary to the prospective open inclusion used in the present study, to over-report the pathogen studied.