Intervention Review

You have free access to this content

Antimicrobials for treating symptomatic non-typhoidal Salmonella infection

  1. Ifeanyi A Onwuezobe1,*,
  2. Philip O Oshun2,
  3. Chibuzo C Odigwe3

Editorial Group: Cochrane Infectious Diseases Group

Published Online: 14 NOV 2012

Assessed as up-to-date: 6 AUG 2012

DOI: 10.1002/14651858.CD001167.pub2


How to Cite

Onwuezobe IA, Oshun PO, Odigwe CC. Antimicrobials for treating symptomatic non-typhoidal Salmonella infection. Cochrane Database of Systematic Reviews 2012, Issue 11. Art. No.: CD001167. DOI: 10.1002/14651858.CD001167.pub2.

Author Information

  1. 1

    University of Uyo Teaching Hospital, Medical Microbiology and Parasitology, Uyo, Akwa Ibom State, Nigeria

  2. 2

    Lagos University Teaching Hospital, Department of Medical Microbiology, Lagos, Nigeria

  3. 3

    Institute of Tropical Disease Research & Prevention, Nigeria Effective Health Care Alliance Programme, Calabar, Cross River State, Nigeria

*Ifeanyi A Onwuezobe, Medical Microbiology and Parasitology, University of Uyo Teaching Hospital, Uyo, Akwa Ibom State, 1136, Nigeria. ifezobe@yahoo.com.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 14 NOV 2012

SEARCH

 
Characteristics of included studies [ordered by study ID]
Butler 1993

MethodsMulticentre RCT in six countries between July 1987 and September 1989.

Oral or written informed consent was obtained prior to entry into the study.


ParticipantsAdults with acute diarrhoea. Excluded if had previously taken antibiotics, unwell, or with other gut pathogen.

508 randomized; 46 culture positive for NTS included.


Interventions1: Oral fleroxacin 400 mg single dose

2: Oral fleroxacin 400 mg daily for three days

3. Placebo

110 patients were randomized to the placebo and single dose arms and 112 patients randomized to the multiple dose arm. 176 randomized patients were excluded.


OutcomesStool cultures were done on day 3 and 5 after start of treatment.

Outcomes of interest were: time to cessation of diarrhoea, mean number of loose stools per day, number of stool cultures negative for initial pathogens on day 3 and 5 after start of therapy.

Reported adverse events.

Blood and urine adverse events - renal and hepatic function tests, and crystalluria.


NotesNo specific serotype was indicated.

Countries: Italy, Thailand, Ivory Coast, Indonesia, Mexico and Israel.

Ethical approval was not reported.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskTreatment regimens were randomized from computer generated numbers.

Allocation concealment (selection bias)Unclear riskWe do not know how allocation was concealed.

Blinding (performance bias and detection bias)
All outcomes
Low riskInvestigators were blinded. Placebo tablets were used.

Incomplete outcome data (attrition bias)
All outcomes
High riskPatients with incomplete outcomes were excluded from the analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, we do not believe the outcomes were selectively reported.

Other biasUnclear riskFunding: no statement was made.

Chiu 1999

MethodsRCT conducted between August 1995 and March 1996.

No information regarding the source of ethical approval.


Participants42 patients were randomized. Patients were included in the trial if they were aged over 6 months and presented with blood or mucoid diarrhoea with or without fever, and had a positive stool culture.

Excluded if they had a toxic appearance, were vomiting, had abdominal distension, had taken antibiotics in the past 72 hours, and had a negative stool culture.


InterventionsThis was a three arm trial:

1: 10mg/kg per day of oral azithromycin once daily

2: Cefixime 10 mg/kg per day of oral cefixime in two divided doses

Control: No antibiotic.

Treatment was administered for 5 days.

14 patients were randomized to each arm.


OutcomesThis trial assessed number of days of fever and diarrhoea.


NotesDone in Taiwan. Funded by the National Health Research Institute, Department of Health, National Science Council.

There was resistance to azithromycin in two cases treated with cefixime.

No specific serotype of Salmonella was referred to in the study.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAllocation sequence was generated by computer randomization.

Allocation concealment (selection bias)Unclear riskNo information was supplied regarding concealment.

Blinding (performance bias and detection bias)
All outcomes
High riskPatient and clinician were not blinded, no information regarding blinding of outcome assessor.

Incomplete outcome data (attrition bias)
All outcomes
High riskThe patients who were lost to follow-up were excluded from analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, there was no evidence of selective reporting.

Other biasLow riskNo suggestion of other sources of bias.

Garcia de Olarte 1974

MethodsDouble blind RCT conducted between January 1970 and January 1972.

At inclusion, patients had both a stool sample and a rectal swab collected, rectal swabs for culture were collected daily and after 10 days every 3 days.


Participants282 children were randomized. 110 children in this trial were culture positive for Salmonella.

Children were excluded from the trial if they had other illnesses requiring antibiotics, if they were under 6 weeks of age, and had a history of allergy to penicillin or its derivatives.


InterventionsIntervention: ampicillin 100 mg/kg in equally divided doses every six hours.

Control: matching placebo.

Intervention was administered intramuscularly in the first year of the trial and orally in the second. Intervention was administered for 5 days.

Among the Salmonella culture positive children, 57 were randomized to antibiotic and 53 randomized to placebo.


OutcomesOutcomes assessed in this trial include: number of patients excreting pathogens for more than or equal to 48 hours, number of days till culture negative, number of patients excreting pathogens after 5 days of therapy, number of days till diarrhoea improved, number of days to cessation of diarrhoea, number of days till patients were afebrile, number of patients with diarrhoea for more than 5 days. Incidence of bacteriologic relapse and all cause mortality were also assessed.


NotesTwo patients in this trial had positive culture for S. Typhi, but outcomes are not reported separately for them.

45 different serotypes of Salmonella were identified among the 280 patients who excreted Salmonella. S. enteritidis ser London, S. enteritidis ser Muenchen, S. enteritidis ser Typhimurium were the most frequently isolated serotypes.

Conducted in Colombia. Ethical approval was received from the Human Experimentation Committee in Dallas, USA and the Concejo Normativo in Medellin, Colombia.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAllocation sequence was generated by random number table.

Allocation concealment (selection bias)Unclear riskNo information as to how allocation was concealed in trial report.

Blinding (performance bias and detection bias)
All outcomes
Low riskTrial report describes the investigators and participants as blinded. There was use of a matching placebo, and dosing frequency was similar.

Incomplete outcome data (attrition bias)
All outcomes
Low riskPatients appear to have been analysed as per intention to treat although the trial report does not specifically say so.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, there is no reason to believe trial was selectively reported.

Other biasHigh riskThe study was funded by a drug company grant - from the International Division of Bristol Myers Company. The trial report does not say whether they played any role in study design, analysis and reporting.

Goodman 1990

MethodsDouble blind, placebo controlled RCT, conducted between June 1985 and September 1987. Follow-up was for 14 days.

Stool samples were collected at each visit for microbiological assessment.


ParticipantsPatients were included if they were 18 years or more and had an acute diarrhoeal illness lasting 7 days or less. Patients were excluded if they were pregnant, had a history of allergy to nalidixic acid, trimethoprim or sulphamethoxazole, had received antibiotics within the preceding two weeks had a history of significant renal or hepatic dysfunction, were using cathartics or could not give consent.

This trial randomized 202 patients in total and had 15 Salmonella patients whose ages ranged between 20 and 46 years. Among these 15 patients, 2 were randomized to ciprofloxacin, 4 to sulphamethoxazole trimethoprim, and 7 to placebo.

None of the participants were immunosuppressed.


Interventions1: Ciprofloxacin 500 mg.

2: Trimethoprim-Sulphamethoxazole 160-800 mg.

Control: Placebo.

Treatment was administered twice daily for 5 days.


OutcomesClinical outcomes: cure, improvement, relapse, and failure. Microbiological failure, adverse events.


NotesThe reporting of adverse events was not separate for the Salmonella subgroup of patients. There are a few inconsistencies in the report regarding number of patients with Salmonella. Table 2 lists 13 Salmonella patients, but elsewhere the text refers to 15.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information as to how the sequence was generated.

Allocation concealment (selection bias)Unclear riskNo information as to how the allocation was concealed.

Blinding (performance bias and detection bias)
All outcomes
Low riskTrial report describes the study as blinded, and there was use of matching placebo.

Incomplete outcome data (attrition bias)
All outcomes
High riskThe trial report has not specified use of intention to treat. Patients with incomplete outcome data were excluded from the analysis.

Selective reporting (reporting bias)Low riskAlthough the trial protocol is not available to us, we have no reason to suspect the trial was selectively reported.

Other biasHigh riskFunded by a grant from Miles Pharmaceutical Co, New Haven CT, USA. The trial report has not made any statement as regards their involvement with the design, conduct, analysis and reporting of the trial.

Joint Project ASID 1970

MethodsDouble blind, placebo controlled RCT. Trialists attempted to follow up the patients for up to 6 weeks.


Participants239 patients were randomized, but analysis was only possible for 168 patients. Age: 0 to > 65 years.


InterventionsIntervention: Oral neomycin 50 mg/kg body weight daily in divided doses.

Control: Placebo.

Treatment lasted for 5 days.

Among the 168 that were analysed, 78 were randomized to antibiotic and 90 to placebo.


OutcomesThe duration of clinical illness and the incidence of negative stool cultures, incidence of relapse.


NotesWe attempted to extract outcome measures in this trial from the graphs, but have not used it in a meta-analysis because of concerns that not all the patients in the trial are accounted for and lack of information as to the reasons for the exclusions.

The trialists referred to non-invasive Salmonella and specifically excluded the typhoidal Salmonella. No specific serotypes of NTS were mentioned.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo statement was made as to how allocation sequence was generated.

Allocation concealment (selection bias)Unclear riskNo statement was made as to allocation concealment.

Blinding (performance bias and detection bias)
All outcomes
Low riskThe trial describes the outcome assessors as blinded, and there was use of placebo.

Incomplete outcome data (attrition bias)
All outcomes
High riskTrialists excluded all the patients who did not complete the trial from the analysis. They did not do and did not report an intention to treat analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, we have no reason to believe the outcomes were selectively reported.

Other biasLow riskNothing to suggest other bias.

Kazemi 1973

MethodsRCT.


Participants36 children with Salmonella gastroenteritis aged between 10 months and 15 years were randomized. None of the patients was immunosuppressed.

Patients were included if they had a history of fever and diarrhoea lasting more than 3 days, mucus or blood in diarrhoeal stools and if they had NTS proven by culture.

Exclusion criteria - received antibiotics within the preceding 5 days, renal or hepatic disease, blood dyscrasia and Salmonella bacteraemia.

Participants were followed up for 6 months.


Interventions1: Sulphamethoxazole-trimethoprim (100 mg/20 mg per 24 hours) given orally in 4 divided doses.

2 : Ampicillin (100 mg per kg/day).

CONTROL: No treatment.

Treatment was administered for 7 days.

14 children were randomized to intervention arm 1, 10 to intervention arm 2, 12 to control arm.


OutcomesOutcomes of interest included duration of diarrhoea, duration of fever, duration of illness, duration of hospitalization, microbiological failure, adverse events (generalized maculopapular rash and vomiting), convalescent fecal carriage of Salmonella.


NotesConducted in Canada. 6 patients were lost to follow up and these were excluded from the analysis.

The Salmonella serotypes in the study were Salmonella Typhimurium (19), Salmonella Heidelberg (2), Salmonella Blockley (5), Salmonella Montevideo (1), Salmonella Newport (5), Salmonella Infantis (1), Salmonella Enteritidis (1), Salmonella Java (1) and Salmonella Thompson (1).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskThe trialists describe the trial as randomized but do not say how the sequence was generated.

Allocation concealment (selection bias)Unclear riskThe trialists do not state how allocation was concealed.

Blinding (performance bias and detection bias)
All outcomes
High riskThe trial had an arm where participants were not treated. There is no statement as to whether outcome assessors were blinded to the treatment patients received in the two treatment arms and the fact that the patients in the control arm were not treated.

Incomplete outcome data (attrition bias)
All outcomes
Low riskThe trialists accounted for all the patients excluded from analysis although they did not do an intention to treat analysis.

Selective reporting (reporting bias)Low riskWe are not able to verify the protocol but do not have any reason to believe that there was selective outcome reporting.

Other biasHigh riskThe study was supported by a grant from a pharmaceutical company, Hoffman La Roche Ltd. The trialists have not made any statement regarding the extent of their involvement in the design, conduct, analysis and reporting of the trial.

Macdonald 1954

MethodsRCT conducted between November 1951 and March 1953.


Participants51 children was randomized. Children included were less than 2 years of age, presented with diarrhoea, and had a positive culture for NTS without any coexisting parenteral infection. Enteritis was less than one week in duration. Exclusion criteria - anorexia and severe dehydration.


InterventionsIntervention: Oral chloramphenicol 120 mg/kg 6-8 hourly for 10 days.

Control: No specific treatment.

25 children were randomized to antibiotic and 26 to no treatment.


OutcomesThis trial evaluated duration of illness, clinical and microbiological failure, and duration of diarrhoea.


NotesThe isolated Salmonellae were Salmonella Typhimurium (48), Salmonella Adelaide (2), Salmonella Derby (1). No information was provided regarding the source of ethical approval.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAllocation sequence was generated by a table of random numbers and allocation was done on the basis of odds and evens.

Allocation concealment (selection bias)Unclear riskThe trial report does not provide information regarding allocation concealment.

Blinding (performance bias and detection bias)
All outcomes
Low riskOutcome assessors were blind as to which arm a patient belonged.

Incomplete outcome data (attrition bias)
All outcomes
Low riskThe trial is not reported as intention to treat but all patients were accounted for in the analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, we have no reason to believe outcomes were selectively reported.

Other biasUnclear riskNo information provided as regards the source of funding for the trial.

Neil 1991

MethodsPlacebo controlled, double blind RCT.


Participants15 patients were randomized in this trial. Trial commenced on day 9 after disease outbreak. To be eligible for inclusion: acute onset of abdominal pain or diarrhoea with at least one of fever, headache, nausea or vomiting, informed consent. All participants had a positive stool culture for S. java.

Pregnant women and people receiving previous antibiotic therapy were excluded.


InterventionsIntervention: Oral ciprofloxacin 750 mg twice daily.

Control: Matching placebo.

Treatment was administered for 14 days.

8 patients were randomized to the intervention arm and 7 patients to the control arm.


OutcomesThe outcomes assessed in this trial included: duration of stool culture positivity, incidence of adverse events, incidence of relapse after treatment.


NotesThe strain of Salmonella in the study was Salmonella java. Ethical clearance from Brown University, USA.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskTrial report does not state how allocation sequence was generated.

Allocation concealment (selection bias)Unclear riskTrial report does not state how allocation was concealed.

Blinding (performance bias and detection bias)
All outcomes
Low riskTrial was described as blinded and there was use of matching placebo.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskWe are unable to assess how the trialists dealt with incomplete outcome data. They do not say they did an intention to treat analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the protocol, we have no reason to believe trial was selectively reported.

Other biasLow riskThe study was funded by a drug company, Miles Pharmaceuticals although this was retrospective.

Nelson 1980

MethodsRCT.


Participants45 children were randomized in this trial.

Children in this trial had uncomplicated Salmonella gastroenteritis (no extraintestinal infection, no high fever, no toxic appearance suggesting bacteraemia).

Children were excluded if there was a history of adverse reactions to penicillins, if there was another focus of infection like otitis media, pneumonia, and if the child was aged less than 6 weeks of age.


Interventions1: Ampicillin.

2: Amoxicillin.

Control: Matching placebo.

Treatment was administered for 5 days.


OutcomesThis trial assessed the number of days to first negative culture, incidence of bacteriologic relapse and number of days until last positive culture, number of days to clinical improvement, number of days to cessation of diarrhoea, and incidence of relapse.


NotesConducted in USA. 1 of the patients in this trial was positive for S. paratyphi B which is not the subject of this review.

The isolated Salmonella were categorized by serogroup as follows: Salmonella Typhimurium (14), Salmonella Heidelberg (7), Salmonella Agona (2), Salmonella Newport (1) Salmonella Manhattan (1) Salmonella Rubislaw (7), Salmonella Oranienburg (1), Salmonella Anatum (1), Salmonella Mississippi (1), Salmonella Infantis (1), Salmonella Javiana (1).

No information as regards the source of ethical approval.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer generated random number lists. Separate lists were used for randomization of patients less than and more than 1 year of age.

Allocation concealment (selection bias)Unclear riskNo information could be extracted regarding allocation concealment.

Blinding (performance bias and detection bias)
All outcomes
Low riskTrial is described as double blind, and matching placebo was used.

Incomplete outcome data (attrition bias)
All outcomes
Low riskTrial appears to have conducted an intention to treat analysis, as all randomized patients are in the analysis except for one randomized patient who was excluded from analysis and this was because of a false inclusion as he did not have a positive stool culture at the point of randomization even though he had 5 days earlier.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, we have no reason to believe trial was selectively reported.

Other biasUnclear riskNo information regarding funding.

Pitkajarvi 1996

MethodsRCT. Conducted between October 1989 and May 1992.


Participants100 patients were randomized (47 to intervention and 45 to control group). 8 patients were lost to follow up.

Inclusion criteria - age between 18 and 60 years of age, acute gastroenteritis, bacteriologically verified S almonella in their stool in the 4 days prior to the study. Exclusion criteria - pregnancy or lactation, hypersensitivity to quinolones, additional antibacterial treatment during the 14 days preceding the trial entry, proven or suspected gall bladder disease or gall stones, impaired kidney function, severe illness or nausea, known HIV infection, or were handling perishable food.


InterventionsIntervention: Oral norfloxacin 400 mg twice daily.

Control: Matching placebo.

Treatment lasted for 10 days.


OutcomesClinical outcomes were assessed in terms of the time to disappearance of clinical symptoms (loose stools, abdominal cramps, vomiting and fever). Bacteriological efficacy was assessed in terms of elimination, persistence (growth of original pathogen without previous post treatment elimination), relapse (growth of original pathogen after previous post treatment elimination) and reinfection (growth of a new pathogen post-treatment). Incidence of adverse events was also assessed.


NotesAs regards the Salmonella serotypes, the trialists confirmed that all the patients has Salmonella infection and only excluded Salmonella Typhi. Patients were followed up for 6 months. Ethical clearance from City of Tampere, Finland.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information as to how the allocation sequence was generated.

Allocation concealment (selection bias)Unclear riskNo information as to how allocation was concealed.

Blinding (performance bias and detection bias)
All outcomes
Low riskTrial is described as double blind, placebo controlled.

Incomplete outcome data (attrition bias)
All outcomes
High riskTrialists have excluded patients who were lost to follow up from the analysis. They have not done an intention to treat analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the protocol, nothing in the trial report suggests selective outcome reporting.

Other biasHigh riskFunding was provided by Astra Arcus AB Sodertalje, Sweden. Trial report has not stated their extent of involvement with design, conduct analysis and reporting of the trial.

Sanchez 1993

MethodsRCT. conducted between June 1988 and September 1990. Patients were followed up for 42 days.

Patients gave written informed consent.


Participants78 patients were randomized, but outcomes were analysed for 65 patients. Inclusion criteria - age > 14 years old, diarrhoea < 48 hours and an axillary temperature of more than or equal to 37.5 degrees celsius. Exclusion criteria - pregnancy, >50 years of age, chronic renal insufficiency, Diabetes Mellitus, Cirrhosis, of the liver, neoplasia, immunodeficiency, gastrectomy, antibacterial drug ingestion during the 72 hours prior to admission, had severe gastroenteritis, or had a negative stool culture or a positive culture with an organism other than Salmonella.


Interventions1: Oral ciprofloxacin 500 mg twice daily.

2: Oral trimethoprim-sulphamethoxazole (160/800) mg twice daily.

Control: Matching placebo.

Treatment was administered for 5 days.

23 patients were randomized to intervention arm 1, 26 to intervention arm 2 and 16 to the control arm.


OutcomesDuration of clinical symptoms, excretion of Salmonella in stool, incidence of adverse events (liver enzymes).


NotesThe serotypes in this study were Salmonella Typhimurium in 4 cases and Salmonella Enteritidis in the rest. Approved by the ethical committees of the respective hospitals.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskAllocation was generated by computer random number programme.

Allocation concealment (selection bias)Unclear riskNo information is provided regarding allocation concealment.

Blinding (performance bias and detection bias)
All outcomes
Low riskThe trial used identical drug and placebo.

Incomplete outcome data (attrition bias)
All outcomes
High riskTrialists excluded patients from analysis if they were lost to follow up, if the protocol was violated, if they withdrew consent, or were unable to tolerate the study drug (vomiting). Patients were also excluded from analysis because they had incomplete outcome data as a result of their not being evaluated clinically on days 3 and 4 of treatment.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, we have no reason to believe trial was selectively reported.

Other biasLow riskFunding was provided by the Department of Medicine of the Hospital de Mutua de Terrassa.

Wistrom 1992

MethodsMulticentre RCT conducted between September 1989 - April 1991.


Participants82 patients in this trial had stool culture positive for Salmonella. This study randomized 598 patients.

Patients were included if they had a history of diarrhoea lasting up to 5 days, fever, vomiting and abdominal cramps in the past 24 hours.

Exclusion criteria were pregnancy, nursing, quinolone hypersensitivity, antibiotic treatment within the three preceding weeks, suspected renal failure, concomitant treatment with drugs known to interact with norfloxacin, non infectious diarrhoea, suspected Clostridium difficile infection, food poisoning, severe vomiting or suspected septicaemia, HIV infection and previous inclusion in the study. Patients had to be over 12 years of age.


InterventionsIntervention - 400 mg of norfloxacin twice daily taken orally.

Control - matching placebo.

Treatment was administered for 5 days.

45 patients were randomized to the intervention arm and 37 to the control arm.


OutcomesClinical outcomes - cure (< or = 1 loose stool per 24 hours without additional symptoms), improvement (two loose stools per 24 hours without additional symptoms or one loose stool per 24 hours with accompanying symptoms) or failure. Recurrence was defined as return to inclusion criteria within 7 days after the last treatment dose. Early treatment failure was defined as discontinued treatment after 7 or fewer doses and appropriate antibiotic treatment due to diarrhoeal disease.

Bacteriologic outcomes were elimination, persistence (identification of the same pathogen before and after treatment at the first follow up), relapse (bacteriologic recurrence with the initial pathogen) or re-infection (clinical recurrence with a new pathogen). Median time to cure was incorporated into duration of illness.


NotesThe isolated organisms were Salmonella Enteritidis (38), Salmonella Typhimurium(20), other Salmonella species(24). Conducted in Sweden.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskTrial does not mention how the sequence was generated.

Allocation concealment (selection bias)Unclear riskTrial does not mention how allocation was concealed.

Blinding (performance bias and detection bias)
All outcomes
Low riskThe trial is described as double blind and used a matching placebo.

Incomplete outcome data (attrition bias)
All outcomes
Low riskAll the patients randomized were subsequently accounted for in the analysis.

Selective reporting (reporting bias)Low riskAlthough we do not have the trial protocol, we have no reason to believe trial was selectively reported.

Other biasHigh riskFunding was from Astra Arcus AB, Sodertalje, Sweden. No statement as regards their role in the design, conduct, analysis, and reporting of the trial.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Bessudo 1972This trial investigated rifampicin and no treatment. They gave their intervention group the antibiotic and the control group was untreated. However they report that they gave 3 of the patients in the control group another antibiotic which they do not name and they reported everything together.

Secondly they report that of the patients in their intervention group, some of them had been treated with another antibiotic which they do not name just prior to the trial.

Thirdly we are unable to extract any data of relevance to the review because of the nature of reporting.

Carlstedt 1990This was a quasi RCT.

Dryden 1996This study did not report separate outcome assessment for the NTS patients.

Ericsson 1983This study investigated a drug, bicozamycin, no longer utilized in medical practice.

Hatalin 1972This was a quasi RCT.

Lolekha 1988This study did not report separate outcome assessment for the NTS patients in the study.

Mattila 1993This study did not report its data clearly enough to permit meaningful interpretation and inclusion in meta-analysis.

Mensa 1989The study was not actually randomized.

Noguerado 1995This study did not report separate outcome assessment for the NTS patients.

Pichler 1986This study did not report separate outcome assessment for the NTS patients

Pichler 1987This study did not report separate outcome assessment for the NTS patients.

Robins-Browne 1983This study did not report separate outcome assessment for the NTS patients.

Sirinavin 2003This was a RCT, but it evaluated asymptomatic patients.

Svenungsson 1990This trial investigated asymptomatic patients.

Taylor 2006This study did not report a separate outcome assessment for the NTS patients.

Wolfsdorf 1973This trial had only one patient who had Salmonella but the patient became culture positive after treatment and was not ab initio.

 
Characteristics of ongoing studies [ordered by study ID]
Tsai 2012

Trial name or titleThe Role of Short-course Ceftriaxone Therapy in the Treatment of Severe Nontyphoidal Salmonella Enterocolitis.

MethodsRandomized Controlled Trial

ParticipantsChildren aged between 3 months and 18 years with suspected severe Salmonella enterocolitis

- defined as those with a high fever (core body temperature more than 38.5) persisting for
longer than 48 hours

- diarrhoea with mucous and bloody-tinged stool.

EXCLUSION CRITERIA

- Children with a toxic appearance, severe vomiting and abdominal distension

- suggestive of sepsis or toxic megacolon, those with an increased risk of invasive NTS
diseases

- immunosuppressive illnesses

- had taken antibiotics during the 7 days before the visit will be excluded

InterventionsCeftriaxone

OutcomesTo evaluate if short-course of ceftriaxone therapy could shorten the clinical courses of severe NTS enterocolitis in children and the excretion of Salmonella in faeces.

Starting dateAugust 2010

Contact informationMing_Han Tsai, MD,
tel: 886-2-24313131 ext.: 2626


Chang Gung Memorial Hospital
Taoyuan
333

drtsai1208@gmail.com

NotesNCT01278017

 
Comparison 1. Antibiotics versus placebo or no treatment

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Presence of diarrhoea at 2 to 4 days146Risk Ratio (M-H, Fixed, 95% CI)1.75 [0.42, 7.21]

 2 Duration of diarrhoea4202Mean Difference (IV, Fixed, 95% CI)1.33 [-0.54, 0.54]

 3 Presence of diarrhoea at 5 to 7 days2192Risk Ratio (M-H, Fixed, 95% CI)0.83 [0.62, 1.12]

 4 Clinical failure7440Risk Ratio (M-H, Fixed, 95% CI)0.88 [0.62, 1.25]

 5 Duration of fever2107Mean Difference (IV, Fixed, 95% CI)0.27 [-0.11, 0.65]

 6 Duration of illness2116Mean Difference (IV, Fixed, 95% CI)-0.00 [-0.68, 0.68]

 7 Microbiological failure8Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    7.1 Quinolones versus placebo
4166Risk Ratio (M-H, Fixed, 95% CI)0.33 [0.20, 0.56]

    7.2 Other antibiotics versus placebo
4362Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.83, 1.11]

 8 Fecal carriage of the same Salmonella serovar after 1 month following the end of antibiotic treatment3112Risk Ratio (M-H, Fixed, 95% CI)1.96 [1.29, 2.98]

 
Summary of findings for the main comparison. Antibiotics versus placebo or no treatment for treating symptomatic NTS infection

Antibiotics versus placebo or no treatment for treating symptomatic NTS infection

Patient or population: patients symptomatic for NTS infection
Settings:
Intervention: Any antibiotic versus placebo or no treatment

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlAntibiotics versus placebo or no treatment

Duration of diarrhoea (days)The mean duration of diarrhoea (days) ranged across control groups from
3 to 13 days
The mean duration of diarrhoea (days) in the intervention groups was
0 higher
(0.54 lower to 0.54 higher)
202
(4 studies)
⊕⊕⊝⊝
low1,2,3

Duration of fever (days)The mean duration of fever (days) ranged across control groups from
1 to 2 days
The mean duration of fever (days) in the intervention groups was
0.27 higher
(0.11 lower to 0.65 higher)
107
(2 studies)
⊕⊝⊝⊝
very low1,4,5

Duration of illness (days)The mean duration of illness (days) ranged across control groups from
3 to 19 days
The mean duration of illness (days) in the intervention groups was
0 higher
(0.68 lower to 0.68 higher)
116
(2 studies)
⊕⊝⊝⊝
very low1,5,6

Clinical treatment failure
(Persistent or worsening symptoms at the end of treatment)
230 per 1000202 per 1000
(143 to 287)
RR 0.88
(0.62 to 1.25)
440
(7 studies)
⊕⊝⊝⊝
very low1,7,8

Presence of diarrhoea at 2-4 days77 per 1000135 per 1000
(32 to 555)
RR 1.75
(0.42 to 7.21)
46
(1 study)
⊕⊝⊝⊝
very low1,5,9

Presence of diarrhoea at 5-7 days456 per 1000378 per 1000
(282 to 510)
RR 0.83
(0.62 to 1.12)
192
(2 studies)
⊕⊝⊝⊝
very low1,5,10

*The basis for the assumed risk (eg the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio

GRADE Working group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 Downgraded by one for risk of bias: allocation concealment was not adequately described in any of these studies.
2 No serious indirectness: these four trials have tested standard doses of chloramphenicol, amoxicillin, ampicillin, ciprofloxacin, sulfadoxine-pyrimethamine, azithromycin, and cefixime. Most of the patients were children and severe cases have generally been excluded.
3 Downgraded by one for imprecision: each of these antibiotics have only been tested in a single trial, and these trials are too small to confidently exclude the possibility of an effect.
4 No serious indirectness: these two trials have tested four different antibiotics: ciprofloxacin, trimethaprim-sulfamethoxazole, azithromycin and cefixime. One trial was in adults and one in children.
5 Downgraded by two for imprecision: each antibiotic was evaluated in a single small trial, with too few patients to confidently detect or exclude clinically important benefits or harms
6 Downgraded by one for indirectness: one trial testing chloramphenicol is now over 50 years old. The remaining trial has only tested ciprofloxacin and trimethoprim-sulfamethoxazole.
7 Dowgraded by one for inconsistency: one trial of norfloxacin did show a statistically significant benefit compared to placebo. However this result should be repeated before concluding that norfloxacin is beneficial.
8 Downgraded by one for imprecision: these trials are individually small and tested different antibiotics. Even the cumulative sample size remains underpowered to confidently exclude benefits.
9 Downgraded by one for indirectness: these two studies have only tested norfloxacin and fleroxacin. Clinically important benefit with alternative antibiotics cannot be excluded.
10 Downgraded by one for indirectness: these two trials have only evaluated ampicillin and norfloxacin.
 
Table 1. Detailed search strategies

Search setCIDG SRaCENTRALMEDLINEbEMBASEbLILACSbSCI

1Salmonell* NOT typhoid*Salmonell* NOT typhoid*Salmonell* NOT typhoid*Salmonell$Salmonell$

NOT typhoid$
Salmonell* NOT typhoid*

2antibiotic*SALMONELLA INFECTIONSSALMONELLA INFECTIONSSALMONELLOSISantibiotic$antibiotic*

3antimicrobial*1 or 21 or 21 or 2antimicrobial$antimicrobial*

4treat*ANTI-BACTERIAL AGENTSANTI-BACTERIAL AGENTSANTIBIOTIC-AGENTtreat$treat*

5therap*antibiotic*antibiotic*antibiotic$therap$therap*

62 or 3 or 4  or 5ANTI-INFECTIVE AGENTSANTI-INFECTIVE AGENTSANTIINFECTIVE- AGENT2 or 3 or 4  or 52 or 3 or 4  or 5

71 and 6ampicillin*ampicillin*ampicillin$1 and 61 and 6

8amoxicillin*amoxicillin*amoxicillin$randomized controlled trial*

9cotrimoxazolecotrimoxazolecotrimoxazolerandomised controlled trial*

10chloramphenicolchloramphenicolchloramphenicolcontrolled clinical trial*

11fluoroquinolone*fluoroquinolone*fluoroquinolone$double blind*

12quinolone*quinolone*quinolone$single blind*

13ofloxacinofloxacinofloxacinplacebo*

14norfloxacinnorfloxacinnorfloxacin8-13/or

15ciprofloxacinciprofloxacinciprofloxacin7 and 14

16fleroxacinfleroxacinfleroxacin

17cephalosporin*cephalosporin*cephalosporin$

18ceftriaxoneceftriaxoneceftriaxone

19cefotaximecefotaximecefotaxime

20cefiximecefiximecefixime

214-20/or4-20/or4-20/or

22Limit 21 to HumanLimit 21 to Humans

 aCochrane Infectious Diseases Group Specialized Register.
bSearch terms used in combination with the search strategy for retrieving trials developed by The Cochrane Collaboration (Higgins 2011); upper case: MeSH or EMTREE heading; lower case; free text term.
 
Table 2. Search strategy for the metaRegister of Controlled Trials

'(diarrhoea AND Salmonell*) NOT (typhi OR paratyphi)'


 
Table 3. Duration of Diarrhoea

Study IDInterventionEffect estimate (mean in days)Is Difference SignificantNumber of patients

Garcia de Olarte 1974Ampicillin5.2 in ampicillin arm versus 4.2 in placebo arm (effect measure reported - mean in days)No110

Kazemi 1973Co trimoxazole & ampicillin2.8 (range of 1 to 5) in the co trimoxazole group

3.1 (range of 1 to 7) in the ampicillin group

3 (range of 1 to 10) in the untreated group
 36

 
Table 4. Duration of Fever

Study IDInterventionEffect estimate (mean in days)Is difference significant?Number of patients

Kazemi 1973Co-Trimoxazole & ampicillin3.2 (range 2-7) in the sulphamethoxazole-trimethoprim group

1.6 (range 1-2) in the ampicillin group

2.6 (range 1-7) in the no treatment group
 36

Garcia de Olarte 1974 Ampicillin 0.8 in ampicillin arm vs 1.0 in placebo arm No110

 
Table 5. Duration of Illness

Study IDInterventionEffect estimate (mean in days)Is difference significant?

Kazemi 1973Ampicillin & co trimoxazole3.8 (range 2 to 7) in the sulphamethoxazole-trimethoprim group

2.6 (range 1 to 7) in the ampicillin group

4 (range 1 to 6) in the no treatment group
No

Nelson 1980Ampicillin & amoxicillin20.4 in the ampicillin group

17.6 in the amoxicillin group

16.5 in the placebo group
No

Wistrom 1992NorfloxacinMedian days of treatment

5 days in norfloxacin group

7 days in placebo group.
No (P > 0.2)

 
Table 6. Fecal carriage of same Salmonella serovar 1 month after treatment

Study IDInterventionEffect estimateIs difference significant?

 Pitkajarvi 1996Norfloxacin21% of patients relapsed in antibiotic arm, 16% relapsed in placebo armNo

 Nelson 1980Ampicillin and amoxicillinRelapse in 4 patients in both antibiotic arms. None in placeboYes (P = 0.003) 

 Sanchez 1993Ciprofloxacin and co trimoxazole3/45 antibiotic patients (2 in the ciprofloxacin group, and 1 in the trimethoprim sulphamethoxazole group) relapsed versus 1/12 placebo patients at 3 weeks

2/41 antibiotic patients relapsed versus 1/15 placebo patients at 6 weeks
No

 Kazemi 1973Co trimoxazole and ampicillinNo patients had positive cultures at 8 weeks.

One co-trimoxazole patient was positive at 6 months.
No