Study designs of evaluations included in the review
Articles containing a series of children with Streptococcus pneumoniae bacteraemia treated as out-patients were included if they divided the patients either prospectively or retrospectively into parenteral and oral treatment groups and included a description of all patients on follow-up to determine whether serious bacterial infection was present. Articles were excluded if either an oral treatment or parenteral treatment group was absent. Reasons were given for exclusion of studies.
Specific interventions included in the review
Parenteral and oral antibiotics were studied. Parenteral antibiotics included: intra muscular ceftriaxone (50 or 75 mg/kg); and ceftriaxone, ampicillin, penicillin, ampicillin/sulbactam where dose and delivery method was not stated. Oral antibiotics included the following: amoxycillin 60mg/kg and dose not stated; amoxycillin/clavulanate 40mg/kg and dose not stated; trimethoprim-sulphamethoxazole; cefaclor; erythromycin/sulpha; and cefixime. Antibiotics therapy was initiated at the time the blood culture was obtained.
Participants included in the review
Children with Streptococcus (S) pneumoniae occult bacteraemia (defined as well-appearing, with no major focus of infection, recovery of S pneumoniae from a blood culture on initial physician visit) were studied. Excluded from analysis were the following children: immunocompromised; serious bacterial infection present at initial presentation; no antibiotics were administered; child was judged ill-appearing or the Yale observation score > 10; or child underwent lumbar puncture. Reasons were given for excluded cases.
Outcomes assessed in the review
Outcomes included meningitis and serious bacterial infection defined as pneumonia, meningitis, soft-tissue infection, bone or joint infection. The following were excluded from the definition of serious bacterial infection: persistent bacteraemia, persistent fever, otitis media, bronchitis, and sinusitis.
How were decisions on the relevance of primary studies made?
Retrieved articles were independently analysed by 3 senior authors for adherence to inclusion criteria. Disagreements were resolved by consensus.