In 1952, Harry Eagle demonstrated the failure of penicillin treatment to cure S. pyogenes(STREPA) myositis in a mouse model. Subsequent animal model studies have confirmed this “Eagle effect” and have shown that clindamycin is a more effective treatment, however this has never been validated in humans. We conducted a retrospective review of the outcomes of all patients at The Children's Hospital of Denver from 1983-1997 treated for invasive (not pharyngitis or impetigo) STREPA infection. A favorable outcome was defined as lack of progression of disease (or improvement) after 24 hours of treatment. Fifty-six children were included: 37 with initially superficial disease (24 cellulitis, 13 abscess), and 18 with deep or multiple-tissue infection (6 abscess, 5 septic arthritis, 3 necrotizing fasciitis, 4 other). The median number of antibiotic exposures was 3 per patient (range 1-6) with clindamycin predominating in 39/45 courses of protein-synthesis-inhibiting antibiotics (PSI) and beta-lactams (BL) in 125/126 of the remainder. Clindamycin was often used in combination with a beta-lactam antibiotic. Overall, there was a 68% failure rate of BL antibiotics when used alone - verifying the Eagle effect. Patients with deep infection were more likely to have a favorable outcome if initial treatment included a PSI as compared to exclusive treatment with BL (83% vs. 14%, p=0.006) with a similar trend in those with superficial disease (83% vs 48%, p=0.07). For those children initially treated with BL alone, surgical drainage or debridement increased the probability of favorable outcome in patients with superficial disease (100% vs 41%, p=0.04)with a similar trend in a smaller number of deep infections (100% vs 0%, p=0.14). Of 20 initial treatment failures for whom follow-up information was available, 19 improved on PSI (12 with surgical intervention). The “Eagle effect” is relevant in children with invasive STREPA infections and clindamycin in combination with a beta-lactam (with surgery if indicated) appears to be the most effective treatment.