Antibiotics for non-typhoidal Salmonella diarrhoea

Non-typhoidal Salmonella (NTS) can cause diarrhoea in people. In this review, we investigated the benefits and safety of antibiotics for treatment of NTS versus placebo or no antibiotic treatment. We found that in otherwise healthy people, treatment with antibiotics did not have any benefit over treatment with no antibiotics. Furthermore, treatment with antibiotics made it more likely that patients would continue to excrete the same organisms for up to one month after treatment. We are unable to comment on the use of antibiotics in very young people, very old people and people who are unable to fight off infection because the trials we identified did not include these patients.

Authors' conclusions: 

There is no evidence of benefit for antibiotics in NTS diarrhoea in otherwise healthy people. We are uncertain of the effects in very young people, very old people, and in people with severe and extraintestinal disease. A slightly higher number of adverse events were noted in people who received antibiotic treatment for NTS.

Read the full abstract...

Non-typhoidal Salmonella (NTS) commonly causes diarrhoea, and is usually self-limiting, although sometimes people become ill with sepsis and dehydration. Routine antibiotic use for this infection could result in persistent colonization and the spread of resistant bacterial strains.


To assess the efficacy and safety of giving antibiotics to people with NTS diarrhoea.

Search strategy: 

We searched the Cochrane Infectious Diseases Group trials register (up to August 2012), the Cochrane Controlled Trials Register (CENTRAL) published in The Cochrane Library (up to Issue 8 2012); and MEDLINE, African Index Medicus, CINAHL, EMBASE, LILACS, and the Science Citation Index, all up to 6 August 2012. We also searched the metaRegister of Controlled Trials (mRCT) for both completed and on going trials and reference lists of relevant articles.

Selection criteria: 

Randomized controlled trials (RCTs) comparing any antibiotic treatment for diarrhoea caused by NTS species with placebo or no antibiotic treatment. We selected trials that included people of all ages who were symptomatic for NTS infection. Examples of symptoms included fever, abdominal pain, vomiting and diarrhoea. We excluded trials where the outcomes were not reported separately for the NTS subgroup of patients. Two review authors independently applied eligibility criteria prior to study inclusion.

Data collection and analysis: 

Two review authors independently extracted data on pre-specified outcomes and independently assessed the risk of bias of included studies. The primary outcome was the presence of diarrhoea between two to four days after treatment. The quality of evidence was assessed using the GRADE methods.

Main results: 

Twelve trials involving 767 participants were included. No differences were detected between the antibiotic and placebo/no treatment arms for people with diarrhoea at two to four days after treatment (risk ratio (RR) 1.75, 95% confidence interval (CI) 0.42 to 7.21; one trial, 46 participants; very low quality evidence). No difference was detected for the presence of diarrhoea at five to seven days after treatment (RR 0.83, 95% CI 0.62 to 1.12; two trials, 192 participants; very low quality evidence), clinical failure (RR 0.88, 95% CI 0.62 to 1.25; seven trials, 440 participants; very low quality evidence). The mean difference for diarrhoea was 0 days (95% CI -0.54 to 0.54; 202 participants, four studies; low quality evidence);for fever was 0.27 days (95% CI -0.11 to 0.65; 107 participants, two studies; very low quality evidence); and for duration of illness was 0 days (95% CI -0.68 to 0.68; 116 participants, two studies; very low quality evidence). Quinolone antibiotic treatment resulted in a significantly higher number of negative stool cultures for NTS during the first week of treatment (microbiological failure: RR 0.33, 95% CI 0.20 to 0.56; 166 participants, four trials).

Antibiotic treatment meant passage of the same Salmonella serovar one month after treatment was almost twice as likely (RR 1.96, 95% CI 1.29 to 2.98; 112 participants, three trials), which was statistically significant. Non-severe adverse drug reactions were more common among the patients who received antibiotic treatment.