Table 1 Studies which were excluded and the principle reason for exclusion

Unkila-Kallio 1993

Mainly epidemiologic data. Appears to be the cohort from which Peltola 1997 is drawn.

Dones 1994

Small prospective cohort of 4 patients evaluating teicoplanin.

Nelson 1982

Study addressed bactericidal titers. Follow-up did not extend to 6 months although reported cure rate was 95% with an average duration of 5 days of intravenous antimicrobials.

Beauvais 1981

Addressed the use of pristinamycin orally. Data was not usable because group could not be reliably separated for route and type of antimicrobial.

Fleming 1970

No follow-up data available. Trial focussed on microbiology and side effects of treatment.

Nussinovitch 1997

Data on osteomyelitis cases alone could not be extracted. Maximum follow-up was about 4 months. Cure appeared to be 95% with mean duration of IV antibiotics of 13.7 days.

Walker 1973

Follow-up was 4 months. Mean duration of intravenous antimicrobials in 14 children was 4.21 days (range 2–21 days) and cure was 100%.

Bandelon 1988

Group of osteomyelitis and septic arthritis could not be separated with respect to length of therapy or etiology.

Kulhanjianv 1989

Only 3 cases of osteomyelitis due to S. aureus.

Aronoff 1986

Of 9 children only 1 child had osteomyelitis alone. No follow-up.

Learmonth 1984

Combined patients with septic arthritis and osteomyelitis. Could not extract data with respect to S. aureus alone.