Background: Antibiotic treatment of salmonella infections aims to shorten illness and prevent serious complications. There are also concerns about increasing antibiotic drug resistance.
Objectives: The objective of this review was to assess the effects of antibiotics in adults and children with diarrhoea who have salmonella.
Search strategy: We searched the Cochrane Infectious Diseases Group trials register, the Cochrane Controlled Trials Register, Medline, Science Citation Index, African Index Medicus, Lilacs, Extra Med and reference lists of relevant articles. We also contacted experts in the field.
Selection criteria: Randomised and quasi-randomised trials comparing antibiotic therapy with placebo or no antibiotic therapy for salmonella infections in symptomatic or asymptomatic adults or children. Typhoid and paratyphoid salmonella infections were excluded.
Data collection and analysis: Trial quality assessment and data were extracted independently by two reviewers.
Main results: Twelve trials involving 778 participants (with at least 258 infants and children) were included. There were no significant differences in length of illness, diarrhoea or fever between any antibiotic regimen and placebo. The weighted mean difference for length of illness was -0.07 days, 95% confidence interval -0.55 to 0.40; diarrhoea -0.03 days, 95% confidence interval -0.53 to 0.48; fever -0.45 days, 95% confidence interval -0. 98 to 0.08. Antibiotic regimens resulted in more negative cultures during the first week of treatment. Relapses were more frequent in those receiving antibiotics, and there were more cases with positive cultures in the antibiotic groups after three weeks. Adverse drug reactions were more common in the antibiotic groups (odds ratio 1.67, 95% confidence interval 1.05 to 2.67).
Reviewer's conclusions: There appears to be no evidence of a clinical benefit of antibiotic therapy in otherwise healthy children and adults with non-severe salmonella diarrhoea. Antibiotics appear to increase adverse effects and they also tend to prolong salmonella detection in stools.