Intervention Review

You have free access to this content

Antibiotic therapy for Shigella dysentery

  1. Prince RH Christopher1,*,
  2. Kirubah V David2,
  3. Sushil M John2,
  4. Venkatesan Sankarapandian2

Editorial Group: Cochrane Infectious Diseases Group

Published Online: 4 AUG 2010

Assessed as up-to-date: 17 JUL 2009

DOI: 10.1002/14651858.CD006784.pub4


How to Cite

Christopher PRH, David KV, John SM, Sankarapandian V. Antibiotic therapy for Shigella dysentery. Cochrane Database of Systematic Reviews 2010, Issue 8. Art. No.: CD006784. DOI: 10.1002/14651858.CD006784.pub4.

Author Information

  1. 1

    Christian Medical College, Family Medicine, Vellore, Tamilnadu, India

  2. 2

    Christian Medical College, LCECU, Vellore, Tamil Nadu, India

*Prince RH Christopher, Family Medicine, Christian Medical College, Vellore, Tamilnadu, 632004, India. prince.christopher@gmail.com.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 4 AUG 2010

SEARCH

 
Characteristics of included studies [ordered by study ID]
Alam 1994

MethodsRandomized controlled trial
Generation of allocation sequence: block randomization technique
Allocation concealment: drugs were stored in bottles, identical in appearance
Blinding: participants and provider blinded
Inclusion of all randomized participants: inadequate, 89%

Duration: not mentioned


ParticipantsNumber of participants enrolled: 80
Number of participants analysed: 71
Loss to follow up: none
Inclusion criteria: children of both sexes between 1 and 8 years of age; having bloody diarrhoea lasting less than 72 hours
Exclusion criteria: taken drugs for shigellosis; with systemic illnesses; severe malnutrition;


Interventions(1) Pivmecillinam (50 mg/kg/day, by mouth, in 4 divided doses, for 5 days)
(2) Nalidixic acid (60 mg/kg/day, by mouth, in 4 divided doses, for 5 days)


Outcomes(1) Treatment failure (diarrhoea at follow up) by day 5
(2) Bacteriological failure on day 5
(3) Temperature > 37.8 ºC (fever on day 5)

Not included in this review:
(4) Abdominal pain or tenderness on day 5


NotesLocation: Bangladesh

Setting: all patients hospitalized in the study ward for the study period

Follow-up period: 6 days

Antibiotic sensitivity pattern of Shigella isolates: 71/71, 100%, were sensitive to pivmecillinam; 26/37, 45%, in the nalidixic group sensitive to nalidixic acid. Nalidixic acid sensitivity is not reported in the pivmecillinam group.

Funding source(s):

  1. United States Agency for International Development (USAID);
  2. International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B);
  3. Leo Pharmaceutical Products, Copenhagen and M/S Opsonin Chemical industries Ltd., Bangladesh provided the study drugs.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Block randomisation technique". Probably done.

Allocation concealment?Low risk"...patients were randomly allocated to treatment groups". There is no clear mention that allocation was concealed. Probably done as drugs were stored in serially numbered bottles (see below).

Blinding?
All outcomes
Low risk"Drugs were stored in bottles, identical in appearance, flavour and weight; labels on the bottles contained only the name of the study and the serial number of the patient who used the bottle." Participant and assessor blinding.

Incomplete outcome data addressed?
All outcomes
High risk80 entered the study; 71 had Shigella in culture; no data regarding participants with non-Shigella dysentery (9) who were randomized according to the inclusion criteria. Outcomes reported only for all 71 (89%) with culture confirmed Shigella dysentery.

Free of selective reporting?Low riskThe study's prespecified outcomes which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Bennish 1990

MethodsRandomized controlled trial
Generation of allocation sequence: block randomization, random number table
Allocation concealment: medications and placebo packaged in identical appearing capsules
Blinding: participants, investigators, and assessor blinded
Inclusion of all randomized participants: inadequate, 75%

Duration: 1 year and 3 months, from June 1986 to September 1987


ParticipantsNumber of participants enrolled: 161
Number of participants analysed: 121
Loss to follow up: 6
Inclusion criteria: dysentery less than 72 hours duration; adult males; age 18 to 60 years; no prior treatment with antimicrobial agent effective against shigellosis; absence of trophozoites of Entamoeba histolytica on stool microscopy
Exclusion criteria: any other systemic illness additional to diarrhoea


Interventions(1) Ciprofloxacin (500 mg orally every 12 hours for 5 days)
(2) Ampicillin (500 mg orally every 6 hours for 5 days)


Outcomes(1) On day 5, resolution of illness (patients with less than 3 stools, none watery, afebrile)
(2) On day 5, marked improvement (patients with less than 6 stools, less than 1 watery stool)
(3) On day 5, slight improvement (less than 9 stools, less than 2 watery stools)
(4) On day 5, treatment failure (febrile, less than 10 stools, less than 3 watery stools)
(5) Bacteriological cure (if Shigella species could not be cultured from a stool or rectal swab on study day 3 or after)
(6) Mean stool frequency on day 3
(7) Adverse events (those that required discontinuation of the drug)
(8) Other adverse events


NotesLocation: Bangladesh

Setting: all patients hospitalized in the study ward for 6 days after the first dose of medication

Follow-up period: 13 days

Antibiotic sensitivity pattern of Shigella isolates: 121/121, 100%, were sensitive to ciprofloxacin; 34/60, 56.6%, in the ciprofloxacin group and 26/61, 42.6%, in the ampicillin group was sensitive to ampicillin

Funding source(s):

  1. Danish International Development Agency
  2. Applied Diarrheal Disease Research Project of the United States Agency for International Development (to M.L. Bennish)
  3. Miles Pharmaceuticals.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Randomisation was done with block randomisation technique using a random number table and block size four". Probably done.

Allocation concealment?Low riskNot described but both drugs were identically packaged (see below); possibly concealed

Blinding?
All outcomes
Low risk"...both medications and placebo were packaged in identical-appearing capsules, and patients, physicians, and nursing staff were blinded to their contents". Participant, investigator and assessor blinded.

Incomplete outcome data addressed?
All outcomes
High riskTotal randomized 161. Outcomes reported only for all 121 (75%) with culture confirmed Shigella dysentery. No data regarding participants with non-Shigella dysentery (34) who were randomized according to the inclusion criteria.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Bibile 1961

MethodsRandomized controlled trial
Generation of allocation sequence: previously prepared list of random numbers
Allocation concealment: no
Blinding: not blinded
Inclusion of all randomized participants: unclear

Duration: not mentioned


ParticipantsNumber of participants enrolled: unclear
Number of participants analysed: 80
Loss to follow up: unclear
Inclusion criteria: 3 or more unformed stools per day with blood and mucus; tenesmus; no previous treatment; macroscopic and microscopic appearance of the stool comparable with bacillary not amoebic dysentery
Exclusion criteria: amoebic dysentery


Interventions(1) Sulphadimidine (2 g immediately, followed by 1 g every 6 hours orally for 5 days)
(2) Sulpha methoxy pyridazine (1 g on first day and 0.5 g daily orally for a further 4 days)
(3)Tetracycline (250 mg orally every 6 hours for 5 days)
(4) "Strepto triad" (3 tablets three times daily, orally for 5 days; each tablet of streptotriad contains streptomycin 65 mg, sulphamerazine 65 mg, sulphadiazine 100 mg, and sulphathiazole 100 mg). This group was not included in the analysis (sulphonamides versus tetracycline) as it contains a non-sulphonamide drug - streptomycin.

Other interventions: Injection pethidine given to one participant for severe tenesmus


Outcomes(1) Number clinically cured by day 5
(2) Number bacteriologically cured
(3) Mean duration of fever in days
(4) Mean duration of abnormal stool in days


NotesLocation: Sri Lanka

Setting: not reported

Follow-up period: 8 days

Antibiotic sensitivity pattern of Shigella isolates: not reported

Funding source(s): Supplies of drugs from:

  1. Imperial Chemical Industries for sulphadimidine ("Sulphamethazine");
  2. Lederle Laboratories for tetracycline ("Achromycon") and sulphamethoxazole ("Lederkyn");
  3. May & Baker Ltd. for "Streptotriad".


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"...listed in a random order"

Allocation concealment?High risk"...previously prepared list of random numbers". Probably not done.

Blinding?
All outcomes
High riskNot mentioned; probably not done

Incomplete outcome data addressed?
All outcomes
Low riskNo missing outcome data

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Dutta 1995

MethodsRandomized controlled trial
Generation of allocation sequence: random number table; permuted blocks of block length 8
Allocation concealment: sealed envelopes
Blinding: outcome assessor blinded; others unclear
Inclusion of all randomized participants: inadequate, 88%

Duration: 8 months, from December 1992 to July 1993


ParticipantsNumber of participants enrolled: 72
Number of participants analysed: 63
Loss to follow up: 9
Inclusion criteria: children; both sexes; aged up to 5 years; with clinical diagnosis of dysentery (loose stool more than 3 times per day)
Exclusion criteria: no prior antibiotic therapy, no systemic illness


Interventions(1) Furazolidone (7.5 mg/kg/day orally in 4 divided doses for 5 days)
(2) Nalidixic acid (55 mg/kg/day orally in 4 divided doses for 5 days)


Outcomes(1) Clinical cure on day 3 and day 5 (no blood in stool, no fever, semisolid stools less than 3 times for last 24 hours, or no stool for last 18 hours)
(2) Treatment failure on day 3 or day 5 (deterioration or no improvement in clinical parameters, for example fever, presence of blood, and mucus in stool or frequency of stool on day 5)


NotesLocation: India

Setting: participants were hospitalized during the trial period

Follow-up period: 5 days

Antibiotic sensitivity pattern of Shigella isolates: not reported

Funding source(s): none mentioned


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Patients were randomised into two treatment groups...... in accordance with a random number table, using permuted block of block length eight"

Allocation concealment?Low risk"...sealed envelopes were used for treatment allocation"

Blinding?
All outcomes
Low risk"One of the investigators who had no knowledge of the drug administered monitored the clinical response"; only outcome assessor blinded.

Incomplete outcome data addressed?
All outcomes
High risk"Two patients in furazolidone group and seven patients in the nalidixic acid group dropped out"; no reasons given. 87.4% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias.

Gotuzzo 1989

MethodsRandomized controlled trial
Generation of allocation sequence: block randomization with a random number table
Allocation concealment: unclear
Blinding: nil
Inclusion of all randomized participants: inadequate, 32%

Duration: not reported


ParticipantsNumber of participants enrolled: 174
Number of participants analysed: 55
Loss to follow up: 7
Inclusion criteria: adults; dysentery; duration of illness less than 24 hours; informed consent
Exclusion criteria: antibiotic therapy within 48 hours


Interventions(1) Cotrimoxazole (160/800 mg twice a day for 5 days)
(2) Norfloxacin (800 mg single dose)


Outcomes(1) Days to last unformed stool
(2) Number of culture positive follow up


NotesLocation: Peru

Setting: participants were not hospitalized but followed up in the out-patients

Follow-up period: 2 weeks

Antibiotic sensitivity pattern of Shigella isolates: 84% in the cotrimoxazole group and 86% in the norfloxacin group were sensitive to cotrimoxazole; 100% sensitivity in both groups to norfloxacin

Funding source(s): in part by the International Collaboration in Infectious Disease Research grant 5 P01 A120130 from the National Institute of Allergy and Infectious Diseases


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low riskBlock randomization with a random number table

Allocation concealment?Unclear riskNot mentioned

Blinding?
All outcomes
High riskNot mentioned but unlikely to be blinded as the dosage regimens of interventions were different

Incomplete outcome data addressed?
All outcomes
High risk174 entered the study; analysis was done on 55 (32%) patients; 62 had Shigella in culture; no data regarding participants with non-Shigella dysentery (112) who were randomized according to the inclusion criteria. 7 patients were excluded from the culture Shigella positive 62 (5 from cotrimoxazole group due to drug resistance to the allocated drug and 2 others not mentioned).

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Haltalin 1973

MethodsRandomized controlled trial
Generation of allocation sequence: unclear
Allocation concealment: no
Blinding: unclear
Inclusion of all randomized participants: adequate, 100%

Duration: not reported


ParticipantsNumber of participants enrolled: 36
Number of participants analysed: 36
Loss to follow up: nil
Inclusion criteria: infants and children; acute diarrhoeal disease; presumptive bacteriologic diagnosis of shigellosis; written informed consent from responsible legal guardian
Exclusion criteria: infants under 1 month of age; known drug allergy; requiring specific antimicrobial therapy for concurrent infection


Interventions(1) Nalidixic acid (13.75 mg/kg, orally, every 6 hours for 5 days)
(2) Ampicillin (25 mg/kg, orally, every 6 hours for 5 days)

Other interventions:
Symptomatic treatment for fever and convulsions was ordered as necessary and was similar for both groups
Fluid and electrolyte therapy and oral alimentation were given according to ward routine and was similar for both groups


Outcomes(1) Number culture positive > 48 hours after start of treatment
(2) Number culture positive > 5 days after start of treatment
(3) Relapse
(4) Number of days until culture negative
(5) Diarrhoea > 5 days after start of treatment
(6) Removed from protocol due to worsening
(7) Number of days diarrhoea after start of treatment
(8) Days until afebrile after start of treatment


NotesLocation: United States of America

Setting: hospital, in-patient based trial

Follow-up period: 5 days

Antibiotic sensitivity pattern of Shigella isolates: 17/17, 100%, in the nalidixic acid group were sensitive to nalidixic acid and 19/19, 100%, in the ampicillin group were sensitive to ampicillin. Nalidixic acid sensitivity in the ampicillin group and ampicillin sensitivity in the nalidixic group is not reported.

Funding source(s):

  1. John A. Hartford Foundation
  2. Sterling-Winthrop Research Institute.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Unclear risk"...randomly assigned"; but the method of sequencing not mentioned. In a previous trial done by the same author (Haltalin 1967) randomization was done based on the terminal digit number of the hospital record. The author could not be contacted for details since there was no mail ID. The journal's present editorial team did not have any details of the study.

Allocation concealment?Unclear riskNot mentioned

Blinding?
All outcomes
Unclear riskNot mentioned

Incomplete outcome data addressed?
All outcomes
Low riskNo missing outcome data

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Islam 1994

MethodsRandomized controlled trial
Generation of allocation sequence: block randomization
Allocation concealment: was done by sequentially numbered identical containers. "Test drug and the standard drug were packed in identical bottles, were identical in appearance, flavour, and weight; the label of the bottles contained only the name of the study and the serial number of the patient for whom the bottle was used".
Blinding: participant and provider blinded
Inclusion of all randomized participants: inadequate, 89%

Duration: 2 years, from January 1989 to December 1990


ParticipantsNumber of participants enrolled: 79
Number of participants analysed: 69
Loss to follow up: 10
Inclusion criteria: children between 1 and 8 years; bloody diarrhoea; duration of illness, less than 72 hours; absence of trophozoites of E. histolytica; with informed consent
Exclusion criteria: systemic illness; severe malnutrition; taken effective anti-Shigella drugs before coming to hospital


Interventions(1) Gentamicin (30 mg/kg, orally in 4 divided doses for 5 days)
(2) Nalidixic acid (60 mg/kg, orally in 4 divided doses for 5 days)


Outcomes(1) Temperature > 37.8 ºC on post treatment days 1
(2) Temperature > 37.8 ºC on post treatment days 3
(3) Temperature > 37.8 ºC on post treatment days 5
(4) Isolation rates of Shigella species from stool/rectal swabs on post treatment days 1
(5) Isolation rates of Shigella species from stool/rectal swabs on post treatment days 2
(6) Isolation rates of Shigella species from stool/rectal swabs on post treatment days 3
(7) Isolation rates of Shigella species from stool/rectal swabs on post treatment days 4
(8) Isolation rates of Shigella species from stool/rectal swabs on post treatment days 5
(9) Bacteriologic relapse
(10) Lack of clinical improvement
(11) Lack of bacteriologic cure


NotesLocation: Bangladesh

Setting: participants were admitted in the study ward during the follow-up period

Follow-up period: 5 days

Antibiotic sensitivity pattern of Shigella isolates: all in both groups were sensitive to gentamicin; 26/37, 70%, in the nalidixic acid group were sensitive to nalidixic acid. Nalidixic acid sensitivity in the gentamicin group was not reported.

Funding source(s):

  1. United States Agency for International Development (USAID);
  2. International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B);
  3. M/S Opsonin Chemical industries Ltd., Bangladesh provided the study drugs.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"...randomly allocated to two treatment groups using a block randomisation technique."

Allocation concealment?Low risk"...packaged in identical bottles... The labels on the bottles contained only the name of the study and the serial number of the patient for whom the bottle was used"

Blinding?
All outcomes
Low riskParticipant and provider

Incomplete outcome data addressed?
All outcomes
High risk7/40 missing from the gentamicin group (5 failed to grow Shigella species; 1 developed severe broncho pneumonia and another required blood transfusion for severe anaemia and were excluded from the study); 3/39 missing from nalidixic acid group since they failed to grow Shigella species); 87% follow up

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Kabir 1986

MethodsRandomized controlled trial
Generation of allocation sequence: random numbers table
Allocation concealment: no
Blinding: participant and provider blinded
Inclusion of all randomized participants: adequate, 100%

Duration: not reported


ParticipantsNumber of participants enrolled: 94
Number of participants analysed: 94
Loss to follow up: nil
Inclusion criteria: adult males; dysentery duration of illness less than 48 hours, more than 20 fecal leukocytes per high powered field; no trophozoites of E. histolytica in stool
Exclusion criteria: other illnesses; history of allergy to penicillin; history of recent antibiotic therapy


Interventions(1) Ceftriaxone (1 g, intravenous, single dose)
(2) Ampicillin (4 g, intravenous, single dose)
(3) Placebo


Outcomes(1) Mean duration in days of diarrhoea
(2) Mean duration in days of blood in stool
(3) Mean duration in days of fever
(4) Mean duration in days of positive stool culture


NotesLocation: Bangladesh

Setting: patients were requested to stay in the hospital for 7 days

Follow-up period: 7 days

Antibiotic sensitivity pattern of Shigella isolates: all were sensitive to ceftriaxone; 34/34, 100%, in the ceftriaxone group, 24/30, 80%, in the ampicillin group and 28/30, 93%, in the placebo group were sensitive to ampicillin

Funding source(s):

  1. United Nations Development Program;
  2. The World Health Organization
  3. Roche Research Foundation Far East


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"...randomly allocated"

Allocation concealment?Unclear riskNot mentioned

Blinding?
All outcomes
Low riskParticipants and provider blinded

Incomplete outcome data addressed?
All outcomes
Low riskNo missing outcome data. Outcomes reported for all 94 with culture confirmed Shigella dysentery.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Khan 1997a

MethodsRandomized controlled trial
Generation of allocation sequence: random number table; block randomization with a block size of 6
Allocation concealment: adequate; the randomization list was developed and kept by a person not involved in the care or evaluation or in data analysis
Blinding: participant, provider and outcome assessor blinded
Inclusion of all randomized participants: inadequate, 83%

Duration: not reported


ParticipantsNumber of participants enrolled: 85
Number of participants analysed: 70
Loss to follow up: 6
Inclusion criteria: adult men aged 18 to 60 years; grossly bloody-mucoid stool, tenesmus; duration of illness less than 72 hours; informed consent
Exclusion criteria: taken an effective antimicrobial agent for current illness; co-existing illness requiring antimicrobial therapy; had trophozoites of E. histolytica


Interventions(1) Azithromycin (500 mg, orally on day 1 followed by 250 mg orally for next 4 days)
(2) Ciprofloxacin (500 mg, orally, every 12 hours for 5 days)


Outcomes(1) Clinical failure
(2) Bacteriologic failure
(3) Fever > 24 hours


NotesLocation: Bangladesh

Setting: patients were asked to stay in the hospital for a period of 6 days

Follow-up period: 6 days

Antibiotic sensitivity pattern of Shigella isolates: all were sensitive to both antibiotics in both groups

Funding source(s):

  1. International Centre for Diarrhoeal Disease Research, Bangladesh
  2. Pfizer, Inc.
  3. Dr Seas was supported by a fellowship from the Swedish Agency for Research Cooperation with Developing Countries


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"...patients were given a consecutive study number to which treatment had been randomly pre assigned by use of a random number table...block randomisation method with a block size six was used"

Allocation concealment?Low risk"...randomisation list was developed and kept by a person not involved study"

Blinding?
All outcomes
Low risk"...double dummy technique"; participants, provider and outcome assessor blinded

Incomplete outcome data addressed?
All outcomes
High risk9/85 participants were excluded from analysis as their rectal swab cultures did not grow Shigella; further, 6 of the remaining 76 were removed due to withdrawal from study (4 in the azithromycin group and 2 in the ciprofloxacin group). 83% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Leibovitz 2000

MethodsRandomized controlled trial
Generation of allocation sequence: computer list of random numbers
Allocation concealment: the list of random numbers was created by a person uninvolved in the study
Blinding: participant, provider and outcome assessor blinded
Inclusion of all randomized participants: adequate, 91%

Duration: 1 year and 6 months, from July 1996 to December 1997


ParticipantsNumber of participants enrolled: 221
Number of participants analysed: 201
Loss to follow up: 5
Inclusion criteria: ambulatory infants and children; 6 months to 11 years; community acquired; acute invasive diarrhoea; illness that started less than 7 days before enrolment; grossly bloody-mucoid stools on examination; more than or equal to soft or liquid stools within the last 24 hours; temperature more than or equal to 38 ºC, more than 15 white blood cells/high-power microscopic field; able to take oral medications
Exclusion criteria: were unable to take oral drugs; were receiving antibiotic therapy for the current illness, unless clinical failure was documented; were receiving antimicrobial treatment for more than 3 days for a concomitant infectious disease; needed hospitalization; had a known previous history of renal impairment, liver damage, cardiac disease or seizures; had a known hypersensitivity to either of the study drugs


Interventions(1) Ciprofloxacin suspension (10 mg/kg, every 12 hours for 3 days + placebo intramuscular injection, one shot per day for 3 days)
(2) Ceftriaxone (intramuscular injection, 50 mg/kg/day, once daily for 3 days, maximal dose of 1 g per day + placebo suspension, one dose every 12 hours for 3 days)


Outcomes(1) Failure at end of therapy (day 4 to 5)
(2) Relapse at end of follow up (day 21 +/- 5)


NotesLocation: Israel

Setting: not reported

Follow-up period: 21 +/- 5 days

Antibiotic sensitivity pattern of Shigella isolates: all were sensitive to both antibiotics

Funding source(s): in part by Bayer Corp., USA


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Patients were randomly assigned to one of the therapies according to the computerized list provided by Pharma clinical limited"

Allocation concealment?Low risk"The randomisation list was developed and kept by a person not involved in the care or evaluation of the patients or in data analysis"

Blinding?
All outcomes
Low riskBlinding was done by "double dummy technique". Participant, provider and outcome assessor blinded.

Incomplete outcome data addressed?
All outcomes
High risk"Sixteen and four patients from the ciprofloxacin and ceftriaxone group respectively, were excluded from the efficacy analysis because they are withdrawn from the study before its completion". 91% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Nelson 1976a

MethodsRandomized controlled trial
Generation of allocation sequence: random number tables
Allocation concealment: no
Blinding: nil
Inclusion of all randomized participants: adequate, 100%

Duration: not reported


ParticipantsNumber of participants enrolled: 28
Number of participants analysed: 28
Loss to follow up: nil
Inclusion criteria: infants and children, diarrhoeic form of shigellosis (abrupt onset with high fever, prostration followed by large volume watery stools containing mucus, no blood); dysenteric form of shigellosis (onset is less abrupt, with a 1- to 3-day period of increasing loose stools with blood, abdominal cramps and tenesmus)
Exclusion criteria: none reported


Interventions(1) Cotrimoxazole suspension (40 mg trimethoprim and 200 mg sulphamethoxazole in each 5 ml, by mouth 1.25 ml/kg, daily in 2 doses every 12 hours for 5 days, total 10 doses)
(2) Ampicillin trihydrate suspension, by mouth, 100 mg/kg/day in divided doses every 6 hours for 5 days, total 20 doses

Other interventions:
Fluid and electrolyte therapy and diet were given according to ward routine
Drugs were used in the management of high fever or convulsions


Outcomes(1) Culture positive after > 48 hours
(2) Diarrhoea > 5 days
(3) Number of days until diarrhoea stopped
(4) Adverse events


NotesLocation: United States of America

Setting: participants were admitted in the hospital for 5 days and then discharged and followed up in the out-patients

Follow-up period: 14 to 21 days

Antibiotic sensitivity pattern of Shigella isolates: all in both groups were sensitive to cotrimoxazole; 10/14, 71%, in the ampicillin group and 9/14, 64%, in the cotrimoxazole group were sensitive to ampicillin

Funding source(s): Hoffmann-La Roche, Inc.


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Assignment was made according to a list generated from random number tables"

Allocation concealment?Unclear riskNot mentioned

Blinding?
All outcomes
High riskAmpicillin was given 4 times a day and cotrimoxazole was given 2 times a day without dummies

Incomplete outcome data addressed?
All outcomes
Low riskNo missing outcome data. All randomized participants were used in analysis.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Prado 1993

MethodsRandomized controlled trial
Generation of allocation sequence: randomization list
Allocation concealment: randomization list was kept with WHO, Geneva and was broken only after the study was completed
Blinding: participant, investigator, and outcome assessor blinded by double dummy technique
Inclusion of all randomized participants: inadequate, 40%

Duration: 2 years and 3 months, from November 1989 to January 1992


ParticipantsNumber of participants enrolled: 150
Number of participants analysed: 59
Loss to follow up: 2
Inclusion criteria: acute diarrhoea less than 3 days; children, age range 6 months to 13 years; clinical syndrome of dysentery (visible blood in stools and presence of sheets of polymorphonuclear white cells on stool examination or acute diarrhoea (passage of 3 liquid motions within 24 hours) with the presence of polymorphonuclear white cells on stool microscopy); weight for height index above 70%
Exclusion criteria: treatment with antibiotics within 2 days prior to entry into the study; any life threatening illness due to Shigella; any concurrent disease that required treatment with antibiotics other than the drugs being studied; known hypersensitivity to penicillin or cotrimoxazole; presence of trophozoites of Entamoeba histolytica in stools


Interventions(1) Pivmecillinam (40 mg/kg/day in 4 doses per day)
(2) Cotrimoxazole (40 mg/kg/day in 4 doses per day)

Other interventions:
Dehydration was corrected with orally administered fluids as recommended by WHO


Outcomes(1) Treatment failure
(2) Duration of diarrhoea
(3) Duration of fever
(4) Duration of grossly visible blood in stools
(5) Duration of positive stool culture
(6) Adverse events


NotesLocation: Guatemala

Setting: participants were hospitalized for 5 days and then followed up in the out-patients

Study period: 11 to 13 days

Antibiotic sensitivity pattern of Shigella isolates: 26/29 in pivmecillinam group and 25/30 in the cotrimoxazole group were sensitive to pivmecillinam; 23/29 in the pivmecillinam group and 24/30 in the cotrimoxazole group were sensitive to cotrimoxazole

Funding source(s): World Health Organization


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Randomisation list"

Allocation concealment?Low risk"...randomisation list was kept with WHO, Geneva and was broken only after the study was completed"

Blinding?
All outcomes
Low riskParticipant and provider blinded by "double dummy technique"

Incomplete outcome data addressed?
All outcomes
High risk59/150 (39%) of randomized participants were not included in the analysis as Shigella strains not isolated. 2 patients who withdrew from the study on first day of treatment were not included in the analysis.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Rodriguez 1989

MethodsRandomized controlled trial
Generation of allocation sequence: unclear
Allocation concealment: no
Blinding: nil
Inclusion of all randomized participants: adequate, 100%

Duration: 1 year and 7 months, from January 1987 to July 1988


ParticipantsNumber of participants enrolled: 125
Number of participants analysed: 123
Loss to follow up: nil
Inclusion criteria: children, aged 2 months to 59 months; passage of 3 or more watery stools in the previous 24 hours; history of diarrhoea up to 5 days before admission; and polymorphonuclear leucocytes and blood in stool samples

Exclusion criteria: received in the previous 48 hours any antimicrobials, antidiarrhoeals or any other drug capable of modifying the course of the disease; who had amoeba in stools; any severe concomitant disease; any intolerance to the drug; any known allergy to the study drugs


Interventions(1) Furazolidone (7.5 mg/kg/day, in 4 equally divided doses)
(2) Cotrimoxazole (Trimethoprim (8 mg/kg/day) + sulphamethoxazole (40 mg/kg/day)) in 2 equally divided doses
(3) Control group (no antimicrobials)


Outcomes(1) Cure/treatment success (in initial culture positive cases it is defined as both clinical cure, absence of diarrhoea and alleviation of all signs and symptoms by day 3 plus a bacteriological cure, a negative stool culture; in initial culture negative patients only clinical cure on day 3)
(2) Adverse events


NotesLocation: Mexico

Setting: out-patient study

Follow-up period: 6 days

Antibiotic sensitivity pattern of Shigella isolates: not reported

Funding source(s): Norwich Eaton Pharmaceuticals, Inc. (a Proctor and Gamble company)


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Unclear risk"...randomised into three groups" but the method not mentioned. Neither the author nor the journal could be contacted for clarifications.

Allocation concealment?Unclear riskNot mentioned

Blinding?
All outcomes
High risk"Single blind"; not mentioned which group was blinded; blinding of the dosage schedules of the trial drugs in the 3 arms not done

Incomplete outcome data addressed?
All outcomes
High risk"...two patients in the control group were voluntarily withdrawn from the study". They were not included in the analysis. 98% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?High riskBaseline imbalance, patients in furazolidone group had fewer days with diarrhoea (P value < 0.02)

Salam 1988

MethodsRandomized controlled trial
Generation of allocation sequence: random number table; block randomization with block size of 16
Allocation concealment: unclear in the published data but a personal communication from the author revealed that allocation concealment was done
Blinding: participant, provider, and outcome assessor blinded
Inclusion of all randomized participants: inadequate, 71%

Duration: not reported


ParticipantsNumber of participants enrolled: 90
Number of participants analysed: 64
Loss to follow up: 5
Inclusion criteria: age between 6 months and 12 years; history of blood, mucoid diarrhoea and a stool specimen that had grossly visible blood and mucus; illness duration less than 72 hours
Exclusion criteria: severe malnutrition; with systemic illnesses in addition to shigellosis; who had received allopathic medications other than anti pyretics


Interventions(1) Nalidixic acid (55 mg/kg/day, in 4 equally divided doses for 5 days)
(2) Ampicillin (100 mg/kg/day in 4 equally divided doses for 5 days)


Outcomes(1) Stool frequency
(2) Clinical cure
(3) Rectal prolapse
(4) Fever
(5) Bacteriological failure on day 3
(6) Bacteriological failure on day 6
(7) Adverse events


NotesLocation: Bangladesh

Setting: participants were hospitalized for 6 days

Follow-up period: 6 days

Antibiotic sensitivity pattern of Shigella isolates: all in both groups were sensitive to nalidixic acid. 25/40 in the ampicillin group were sensitive to ampicillin. Ampicillin sensitivity in the nalidixic acid group is not reported.

Funding source(s):

  1. United Nations Children Fund (UNICEF);
  2. Dr Bennish is supported by grants from the Danish International Developmental Agency (DANIDA) and the U.S. Agency for International Development (UASAID)
  3. Chinoin Pharmaceutical and Chemical Works Ltd., Budapest, Hungary and Ambee Pharmaceuticals Ltd., Dhaka, Bangaldesh supplied the study drugs


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"...random number table and block randomisation method with block size of 16".

Allocation concealment?Low risk"...patients were randomly assigned to receive either nalidixic acid or ampicillin" but the concealment method was not mentioned in the published data. Personal communication from the author revealed that allocation concealment was done. The drug was administered to the participating children by the research ward nurses, and the investigators only knew the random
number pre-assigned to one of the 2 drugs, by the randomization process.

Blinding?
All outcomes
Low risk"...drugs were administered as syrups that had similar colour, consistency, and flavour, and the concentration of each drug was adjusted so that patients received the same volume... patients, staff and investigators were unaware of which drug was being given."

Incomplete outcome data addressed?
All outcomes
High risk"data were analysed only from patients with culture-confirmed cases of shigellosis who remained in the study for at least 24 hours." 90 enrolled, 74 eligible for analysis, 64 analysed. 10 drop-outs - 6 withdrawn by their parents, reasons not provided, 4 withdrawn because of lack of clinical improvement. 82% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Salam 1998

MethodsRandomized controlled trial
Generation of allocation sequence: computer generated list of random numbers
Allocation concealment: allocated by Bayer AG Pharma and not available to researchers, double dummy technique
Blinding: participants, providers and outcome assessor blinded
Inclusion of all randomized participants: inadequate, 84%

Duration: 1 year and 8 months, from August 1995 to March 1997


ParticipantsNumber of participants enrolled: 143
Number of participants analysed: 120
Loss to follow up: 10
Inclusion criteria: children aged 2 years to 15 years; dysentery (passage of grossly bloody-mucoid stools for 72 hours or less); who had not received any antimicrobial treatment (agent known to be effective in vivo against shigellosis and active in vitro against the Shigella strain isolated from the patient); gave informed consent
Exclusion criteria: co-existing disorders that required antimicrobial therapy


Interventions
  1. Ciprofloxacin suspension (10 mg/kg, every 12 hours, maximum of 500 mg, for 5 days, total 10 doses with placebo of pivmecillinam)
  2. Pivmecillinam (15 to 20 mg/kg, maximum of 300 mg, every 8 hours for 5 days, total 15 doses with placebo of ciprofloxacin)


Outcomes(1) Clinical failure (if patient did not have persistent dysentery on day 3, and if on day 5 a patient had 6 stools or less, no bloody-mucoid stools, no more than 1 watery stool and no fever)
(2) Bacteriological failure (bacteriological success: if the initial Shigella species could not be identified in culture on day 3 or later)
(3) Fever less than 24 hours
(4) Number of patients with bloody-mucoid stools more than 3 days
(5) Relapse
(6) Adverse event - limp (one of the adverse reactions to the antibiotic therapy could be a LIMP on walking due to joint pain caused by the antibiotics)
(7) All adverse events


NotesLocation: Bangladesh

Setting: participants were hospitalized for 6 days after the first dose and then discharged for follow up

Follow-up period: 180 days

Antibiotic sensitivity pattern of Shigella isolates: all in both groups were sensitive to ciprofloxacin. 58/60, in the ciprofloxacin group and 57/60 in the pivmecillinam group were sensitive to pivmecillinam.

Funding source(s):

  1. Bayer AG, Wuppertal, Germany
  2. ICDDR, Bangladesh


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Low risk"Drug allocation used a computer-generated list of random numbers".

Allocation concealment?Low risk"...list of random numbers, which was not available to the researchers".

Blinding?
All outcomes
Low risk"...double dummy technique". Participant, provider and outcome assessor blinded.

Incomplete outcome data addressed?
All outcomes
High risk13/143 (6 in the ciprofloxacin group and 7 in the pivmecillinam group) were excluded from analysis because they were found not eligible (12 did not grow Shigella in their stool culture and 1 had taken nalidixic acid before study entry). Further 10 (5 in each group) withdrew before study completion. 84% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

Shanks 1999

MethodsRandomized controlled trial
Generation of allocation sequence: unclear
Allocation concealment: not mentioned
Blinding: participants, providers and outcome assessor blinded; double dummy
Inclusion of all randomized participants: inadequate, 87%

Duration: not reported


ParticipantsNumber of participants enrolled: 137
Number of participants analysed: 113
Loss to follow up: 17
Inclusion criteria: adults; acute dysentery (visible blood on recently passed unformed stools); not receiving antibiotics likely to be effective against Shigella species; if female and not pregnant as confirmed by urine testing; able to take oral medications; no study drug allergy; no alternative cause for dysentery; informed consent

Exclusion criteria: not reported


Interventions
  1. Azithromycin (1 g single dose with placebo of ciprofloxacin
  2. Ciprofloxacin (500 mg twice a day with placebo of azithromycin)


Outcomes
  1. Time to clearance of dysentery
  2. Number of participants with dysentery on day 10
  3. Number of days until resolution of dysentery
  4. Number of days until resolution of fever
  5. Number of days of therapeutic support
  6. Relapse after 10 days
  7. Adverse events


NotesLocation: Kenya

Setting: participants were hospitalized for 3 days after the first dose and then discharged for follow up in out-patients

Follow-up period: 10 days

Antibiotic sensitivity pattern of Shigella isolates: not reported

Funding source(s): none mentioned


Risk of bias

BiasAuthors' judgementSupport for judgement

Adequate sequence generation?Unclear risk"Volunteers were... randomised to receive...". Mentioned randomized but not how generated. Author could not be contacted via e-mail.

Allocation concealment?Unclear riskNot mentioned

Blinding?
All outcomes
Low riskParticipants, providers and outcome assessor blinded; double dummy

Incomplete outcome data addressed?
All outcomes
High risk17/130 were withdrawn as they left the hospital before completion of the study drug regimen. 87% follow up.

Free of selective reporting?Low riskThe study's prespecified outcomes, which were of interest in this review, have been reported

Free of other bias?Low riskThe study appears to be free of other sources of bias

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Aoki 1987Not dysentery

Aoki 1989Not dysentery

Ashkenazi 1993Not dysentery

Barada 1980Not dysentery

Bassily 1994Not dysentery

Basualdo 2003Not dysentery

Bennish 1992Same antibiotic in all arms; quinolone, ciprofloxacin; 3-arm trial, 1 g single dose versus 1 g at admission and 2nd dose at 24 hours versus 500 mg twice daily for 5 days

Bezjak 1966Not a RCT

Bhattacharya 1991Same class of drugs in all arms; quinolones; norfloxacin versus nalidixic acid

Bhattacharya 1992Same class of drugs in all arms; quinolones; norfloxacin versus nalidixic acid

Bhattacharya 1997Same class of drugs in all arms; quinolones; norfloxacin versus nalidixic acid

Bogaerts 1985Not a RCT

Browne 1983Not dysentery

Brugel 1950Not a RCT

Butler 1993Not dysentery

Cabada 1992Not dysentery

Camacho 1989Not dysentery

CDC 2006Not a RCT

Chang 1977Not dysentery

de Olarte 1974Not dysentery

Dryden 1996Not dysentery

Dumitriu 1992Not dysentery

DuPont 1973Not dysentery

DuPont 1982Not dysentery

DuPont 1983Not dysentery

DuPont 1984Not dysentery

DuPont 1986Not dysentery

DuPont 1992Not dysentery

DuPont 1992aNot dysentery

Ekwall 1984Not dysentery

Ericsson 1983Not dysentery

Ericsson 1992Not dysentery

Fakouhi 1971Not a RCT

Gendrel 1997Not a RCT

Gilman 1980Same antibiotic in all arms; beta-lactams; ampicillin, high-dose (150 mg/kg/day) versus low-dose (50 mg/kg/day)

Gilman 1981Same antibiotic in all arms; beta-lactam; ampicillin, single dose (150 mg/kg; 1 dose) versus multiple doses (150 mg/kg/day for 5 days)

Goodman 1990Not dysentery

Ha 2008Same class of drugs in all arms; quinolones; ciprofloxacin versus gatifloxacin

Haltalin 1967Not dysentery

Haltalin 1968Not dysentery

Haltalin 1968aNot a RCT

Haltalin 1969Not a RCT

Haltalin 1972Not a RCT

Haltalin 1972aNot a RCT

Han 1998Same class of drugs in all arms; quinolones; rufloxacin versus homefloxacin

Hansson 1981Not dysentery

Helvaci 1998Same class of drugs in all arms; beta-lactam; cefixime versus ampicillin-sulbactam

Hiraishi 1980Not dysentery

Imagawa 1988Not dysentery

Iushchuk 2007Not a RCT

Jiang 1994Not a RCT

Jiang 2000Not a RCT

Jinhua 1992Not a RCT

Kabir 1984Same class of drugs in all arms; beta-lactam; pivmecillinam versus ampicillin

Legros 2004Not a RCT

Lexomboon 1972Not dysentery

Lionel 1969Same antibiotic in all arms; macrolide; tetracycline; single-dose (2.5 g single-dose) versus multiple doses (250 mg, 6-hourly for 5 days)

Lolekha 1988Not dysentery

Lolekha 1991Not dysentery

Mabadeje 1974Not dysentery

Mahllooji 2004Not dysentery

Martin 2000Not dysentery

Matsuoka 1995Not a RCT

Miles 1977Not a RCT

Mol 1987Same class of drugs in all arms; quinolones, enoxacin versus nalidixic acid

Moolasart 1999Not dysentery

Morisawa 1970Not dysentery

Motohiro 1982Not dysentery

Nelson 1975Not dysentery

Nelson 1976Not dysentery

Nikorowitsch 1978Not a RCT

Oldfield 1987Not dysentery

Orenstein 1981Not dysentery

Ostrower 1979Not dysentery

Petruccelli 1992Not dysentery

Pichler 1986Not dysentery

Pichler 1987Not dysentery

Prado 1981Not dysentery

Prado 1992Not dysentery

Rabbani 1982Not a RCT

Rakhmanova 1996Not a RCT

Raqib 2008Not antibiotics

Rogerie 1986Not a RCT

Sagara 1993Not a RCT

Sagara 1994Not a RCT

Saito 1983Not dysentery

Saito 1984Not dysentery

Salam 1995Same class of drugs in all arms; beta-lactams, cefixime versus pivamdinocillin

Salam 1999Not a RCT

Sepp 1995Not dysentery

Seto 1992Not dysentery

Soares 1994Same class of drugs in all arms; quinolones; ciprofloxacin, short course (2 days) versus long course (5 days)

Soares 1996Same class of drugs in all arms; quinolones; 3-arm trial, ciprofloxacin versus lomefloxacin long course versus lomefloxacin short course

Study Group 2002Same antibiotic in all arms; quinolone; ciprofloxacin 15 mg/kg/every 12 hours, short course (3 days) versus standard course (5 days)

Tian 1986Not a RCT

Tong 1970Not dysentery

Varsano 1991Not dysentery

Vinh 2000Same class of drugs in all arms; quinolones, ofloxacin versus nalidixic acid

Wistrom 1992Not dysentery

Xiouying 1986Not a RCT

Yamamoto 1973Not dysentery

Ye 1990Not a RCT

Yin 1998Same class of drugs in all arms; beta-lactams; ceftriaxone made in China versus ceftriaxone made outside China

Yunus 1982Not dysentery

Yuying 1995Not a RCT

Zhang 1991Not dysentery

 
Characteristics of studies awaiting assessment [ordered by study ID]
Carbo 1981

MethodsRandomized controlled trial (used a "randomisation table")

Allocation concealment: not described
Blinding: not specified
Inclusion of all randomized participants: not reported

Duration: unclear

ParticipantsNumber of participants enrolled: not reported
Number of participants analysed: not reported
Loss to follow up: unclear
Inclusion criteria: children over 6 years of age (age limit not mentioned); symptoms and positive bacterial culture
Exclusion criteria: prior renal or hepatic disease

InterventionsAmpicillin: variable doses according to body weight for 7 days; number allocated not reported

Ro-12-2510: 2 tablets every 24 hours; duration unclear; number allocated not reported

OutcomesClinical failure
Microbiological failure
Relapse

NotesNo numerical data provided on number randomized to each arm or for outcomes

Further details from author awaited

 
Comparison 1. Antibiotic versus no drug or placebo

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

 1 Diarrhoea on follow up1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    1.1 Furazolidone versus no drug
173Risk Ratio (M-H, Fixed, 95% CI)0.21 [0.09, 0.48]

    1.2 Cotrimoxazole versus no drug
176Risk Ratio (M-H, Fixed, 95% CI)0.30 [0.15, 0.59]

 2 Time to cessation of fever (in days)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    2.1 Ceftriaxone (IV) versus placebo
164Mean Difference (IV, Fixed, 95% CI)-1.20 [-2.20, -0.20]

    2.2 Ampicillin (IV) versus placebo
160Mean Difference (IV, Fixed, 95% CI)-1.50 [-2.41, -0.59]

 3 Time to cessation of diarrhoea (in days)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    3.1 Ceftriaxone (IV) versus placebo
164Mean Difference (IV, Fixed, 95% CI)-0.30 [-1.41, 0.81]

    3.2 Ampicillin (IV) versus placebo
160Mean Difference (IV, Fixed, 95% CI)-0.30 [-1.37, 0.77]

 4 Time to cessation of blood in stools (in days)1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    4.1 Ceftriaxone (IV) versus placebo
164Mean Difference (IV, Fixed, 95% CI)-0.30 [-1.43, 0.83]

    4.2 Ampicillin (IV) versus placebo
160Mean Difference (IV, Fixed, 95% CI)-0.30 [-1.41, 0.81]

 5 Other adverse events194Risk Ratio (M-H, Fixed, 95% CI)1.43 [0.06, 34.13]

 
Comparison 2. Fluoroquinolones versus beta-lactams

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

 1 Diarrhoea on follow up6Risk Ratio (M-H, Random, 95% CI)Subtotals only

    1.1 All trials
6686Risk Ratio (M-H, Random, 95% CI)1.03 [0.45, 2.37]

    1.2 Adults (subgroup)
1127Risk Ratio (M-H, Random, 95% CI)0.14 [0.04, 0.44]

    1.3 Children (subgroup)
5559Risk Ratio (M-H, Random, 95% CI)1.46 [0.64, 3.34]

    1.4 Confirmed Shigella > 90% (subgroup)
2257Risk Ratio (M-H, Random, 95% CI)4.68 [1.74, 12.59]

    1.5 Confirmed Shigella < 90% (subgroup)
4429Risk Ratio (M-H, Random, 95% CI)0.65 [0.29, 1.42]

 2 Fever at follow up2191Risk Ratio (M-H, Random, 95% CI)0.87 [0.25, 3.06]

 3 Relapse3Risk Ratio (M-H, Random, 95% CI)Subtotals only

    3.1 All trials
3357Risk Ratio (M-H, Random, 95% CI)0.91 [0.11, 7.55]

    3.2 Confirmed Shigella > 90% (subgroup)
2237Risk Ratio (M-H, Random, 95% CI)0.91 [0.11, 7.55]

    3.3 Confirmed Shigella < 90% (subgroup)
1120Risk Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

 4 Bacteriological failure51350Risk Ratio (M-H, Random, 95% CI)0.74 [0.50, 1.11]

    4.1 All trials
5450Risk Ratio (M-H, Random, 95% CI)0.73 [0.33, 1.62]

    4.2 Adults (subgroup)
1127Risk Ratio (M-H, Random, 95% CI)0.28 [0.08, 0.95]

    4.3 Children (subgroup)
4323Risk Ratio (M-H, Random, 95% CI)0.95 [0.43, 2.09]

    4.4 Confirmed Shigella > 90% (subgroup)
136Risk Ratio (M-H, Random, 95% CI)5.56 [0.29, 108.16]

    4.5 Confirmed Shigella < 90% (subgroup)
4414Risk Ratio (M-H, Random, 95% CI)0.65 [0.29, 1.43]

 5 Development of severe complications290Risk Ratio (M-H, Fixed, 95% CI)0.89 [0.28, 2.85]

 6 Serious adverse events1221Risk Ratio (M-H, Fixed, 95% CI)10.90 [0.61, 194.82]

 7 Adverse events leading to discontinuation of treatment1127Risk Ratio (M-H, Fixed, 95% CI)1.02 [0.27, 3.89]

 8 Other adverse events4570Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.77, 1.39]

 
Comparison 3. Fluoroquinolones versus macrolides

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

 1 Diarrhoea on follow up2189Risk Ratio (M-H, Fixed, 95% CI)0.6 [0.24, 1.49]

 2 Fever at follow up2189Risk Ratio (M-H, Fixed, 95% CI)0.33 [0.08, 1.35]

 3 Time to cessation of blood in stools1113Mean Difference (IV, Fixed, 95% CI)-0.20 [-0.68, 0.28]

 4 Bacteriological failure176Risk Ratio (M-H, Fixed, 95% CI)0.33 [0.07, 1.55]

 5 Other adverse events176Risk Ratio (M-H, Fixed, 95% CI)1.33 [0.32, 5.56]

 
Comparison 4. Cotrimoxazole versus beta-lactams

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

 1 Diarrhoea on follow up289Risk Ratio (M-H, Fixed, 95% CI)0.59 [0.23, 1.49]

 2 Bacteriological failure128Risk Ratio (M-H, Fixed, 95% CI)0.75 [0.20, 2.75]

 3 Time to cessation of diarrhoea (hours)161Mean Difference (IV, Fixed, 95% CI)-0.20 [-15.10, 14.70]

 4 Time to cessation of fever (hours)161Mean Difference (IV, Fixed, 95% CI)5.90 [-5.30, 17.10]

 5 Time to cessation of visible blood in stools161Mean Difference (IV, Fixed, 95% CI)2.80 [-12.71, 18.31]

 6 Other adverse events289Risk Ratio (M-H, Fixed, 95% CI)0.81 [0.27, 2.45]

 
Comparison 5. Cotrimoxazole versus fluoroquinolones (norfloxacin)

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

 1 Bacteriological failure162Risk Ratio (M-H, Fixed, 95% CI)1.69 [0.64, 4.47]

 2 Other adverse events162Risk Ratio (M-H, Fixed, 95% CI)2.82 [0.12, 66.62]

 
Comparison 6. Cotrimoxazole versus furazolidone

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

 1 Diarrhoea on follow up1101Risk Ratio (M-H, Fixed, 95% CI)0.71 [0.27, 1.84]

 
Comparison 7. Oral gentamicin versus nalidixic acid

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

 1 Diarrhoea at follow up179Risk Ratio (M-H, Fixed, 95% CI)1.71 [0.98, 2.97]

 2 Fever at follow up179Risk Ratio (M-H, Fixed, 95% CI)2.37 [1.11, 5.07]

 3 Bacteriological relapse179Risk Ratio (M-H, Fixed, 95% CI)1.95 [0.64, 5.95]

 4 Bacteriological failure179Risk Ratio (M-H, Fixed, 95% CI)2.1 [1.29, 3.42]

 
Comparison 8. Sulphonamides versus tetracycline

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

 1 Diarrhoea at follow up160Risk Ratio (M-H, Fixed, 95% CI)7.68 [0.46, 128.12]

 2 Bacteriological failure160Risk Ratio (M-H, Fixed, 95% CI)11.78 [0.73, 190.30]

 
Summary of findings for the main comparison. Antibiotic versus no drug or placebo for Shigella dysentery

Antibiotic versus no drug or placebo for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Mexico and Bangladesh
Intervention: Antibiotic versus no drug or placebo

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

Assumed riskCorresponding risk

ControlAntibiotic versus no drug or placebo

Diarrhoea on follow up - Furazolidone versus no drug
clinical criteria
Follow-up: 6 days
58 per 10012 per 100
(5 to 28)
RR 0.21
(0.09 to 0.48)
73
(1 study)
⊕⊝⊝⊝
very low1,2,3
Antibiotic sensitivity of Shigella isolates not reported; Trial done in 1989

Diarrhoea on follow up - Cotrimoxazole versus no drug
clinical criteria
Follow-up: 6 days
58 per 10017 per 100
(9 to 34)
RR 0.3
(0.15 to 0.59)
76
(1 study)
⊕⊝⊝⊝
very low1,2,4
Same trial as above; had three arms

Relapse - not reportedSee commentSee commentNot estimable-See commentThe two trials for this comparison were too short in follow up duration (6-7 days) to estimate relapses and none were reported.

Serious adverse events - not reportedSee commentSee commentNot estimable-See commentNone of the two trials for this comparison reported serious adverse events

Other adverse events
clinical criteria
Follow-up: 7 days
0 per 1000 per 100
(0 to 0)
RR 1.43
(0.06 to 34.13)
94
(1 study)
⊕⊝⊝⊝
very low5,6,7
Data from a three armed trial; only one non-serious adverse event in the antibiotic arms and none in placebo arm

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Very serious limitations: The method of randomization was not described and there were baseline imbalances in duration of diarrhoea. Allocation concealment and blinding were not reported and this increases the risk of bias in the detection and reporting of some adverse events, though not for other primary outcomes that were objectively ascertained.
2 Serious indirectness: The single trial included only children and hence the evidence for effectiveness of antibiotics over no antibiotics in adults is uncertain. Though the trial did not exclude participants who were malnourished, it is unclear if any participant was malnourished.
3 No imprecision: Both limits of the point estimate of the trial indicated benefit with furazolidine over not receiving an antibiotic
4 No imprecision: Both limits of the point estimate showed appreciable benefit with cotrimoxazole over not receiving an antibiotic
5 Very serious limitations: Allocation was not concealed and there were baseline imbalances in antibiotic sensitivity to those allocated to ceftriaxone (100%) and ampicillin (80%)
6 Serious indirectness: The trial randomized only adults. The antibiotics assessed were ceftriaxone and ampicillin.
7 Very serious imprecision: The 95% CI of the point estimate of the trial includes appreciable risk of adverse events for antibiotics over placebo with no significant differences between interventions.
 
Summary of findings 2. Fluoroquinolones versus beta-lactams for Shigella dysentery

Fluoroquinolones versus beta-lactams for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Bangladesh (4 trials), Israel (1 trial), USA (1 trial)
Intervention: Fluoroquinolones versus beta-lactams

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

Assumed riskCorresponding risk

ControlFluoroquinolones versus beta-lactams

Diarrhoea on follow up - All trials
clinical criteria
Follow-up: 5 to 180 days
251 per 1000259 per 1000
(113 to 595)
RR 1.03
(0.45 to 2.37)
686
(6 studies)
⊕⊝⊝⊝
very low1,2,3,4
One trial from 1973; four trials in the 1990s; only one trial after 2000. The fluoroquinolones evaluated were nalidixic acid, and ciprofloxacin and the beta-lactams evaluated were ampicillin, (intra-muscular) ceftriaxone and pivmecillinam.

Relapse - All trials
clinical criteria
Follow-up: 5 to 180 days
70 per 100064 per 1000
(8 to 529)
RR 0.91
(0.11 to 7.55)
357
(3 studies)
⊕⊝⊝⊝
very low5,6,7,8
One trial from 1973, one from 1990 and one from 2000. Only two reported relapse.

Serious adverse events
clinical criteria
Follow-up: 16 to 21 days
0 per 1000 per 100
(0 to 0)
RR 10.9
(0.61 to 194.82)
221
(1 study)
⊕⊝⊝⊝
very low9,10,11
Only seen in 4.5% of those allocated to fluoroquinolones and not in those given beta-lactams

Adverse events leading to discontinuation of treatment62 per 100064 per 1000
(17 to 245)
RR 1.02
(0.27 to 3.89)
127
(1 study)
⊕⊝⊝⊝
very low12,13,14

Other adverse events
clinical criteria
Follow-up: 5 to 180 days
177 per 1000182 per 1000
(136 to 246)
RR 1.03
(0.77 to 1.39)
570
(4 studies)
⊕⊝⊝⊝
very low15,16,17,18

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 No serious limitations: Four of the six trials in this comparison had limitations in reporting outcomes for some participants but a sensitivity analysis did not appreciably alter the results
2 Serious inconsistency: I squared for the pooled data from six trials was 83% but could be partially explained by subgroup analyses of adults and children and by culture-confirmed versus unconfirmed diagnosis of Shigella dysentery and resultant sensitivity patterns. The one trial in adults showed that a fluoroquinolone (ciprofloxacin) was superior (no imprecision) to a beta-lactam (ampicillin) when sensitivity of the Shigella isolates was 100% for the former and 43% for the latter . Homogenous data (I squared 0%) from two trials in children showed that beta-lactams (ampicillin and intra-muscular ceftriaxone) were superior to fluoroquinolones (nalidixic acid and ciprofloxacin) when >90% of participants had culture-confirmed Shigella dysentery with 100% sensitivity to the antibiotic used (no imprecision).
3 No serious indirectness: The six trials included children and adults and only two excluded severely malnourished children. The fluoroquinolones used included nalidixic acid and ciprofloxacin and the macrolides used included ampicillin, ceftriaxone and pivmecillinam.
4 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with both interventions.
5 No serious limitations: One of the three trials for this comparison had limitations in reporting the method of randomization and allocation concealment but exclusion of this trial in sensitivity analysis did not alter results.
6 Serious inconsistency: The I squared for the pooled data was 63% and could not explained by subgroup analyses.
7 Serious indirectness: The trials that reported this outcome only included children; hence the effects of antibiotics in preventing relapses in adults is unclear.
8 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with both interventions
9 No serious limitations: There were imbalances in those excluded from analysis in the single trial but randomization, allocation concealment and blinding were free of the risk of bias and follow up included 91% of participants
10 Serious indirectness: The trial included only infants and children and the applicability of the results for this outcome in adults is uncertain.
11 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with ceftriaxone and ciprofloxacin.
12 Serious limitations: This outcome was reported only for 75% of randomized participants with culture-confirmed Shigella dysentery.
13 Serious indirectness: The trial that reported this outcome included only adults
14 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with ampicillin and ciprofloxacin in this single trial.
15 Serious limitations: Three of the four trials that reported this outcome reported on less than 85% of those randomised.
16 No inconsistency: I squared was 0%
17 No serious indirectness: The four trials included adults and children and two did not specifically exclude malnourished children.
18 Very serious imprecision: The 95% CI of the pooled estimate indicated appreciable harm and non-appreciable benefit with beta-lactams (ampicillin, ceftriaxone and pivmecillinam) over fluoroquinolones (ciprofloxacin and nalidixic acid)
 
Summary of findings 3. Fluoroquinolones versus macrolides for Shigella dysentery

Fluoroquinolones versus macrolides for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Bangladesh and Kenya
Intervention: Fluoroquinolones versus macrolides

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

Assumed riskCorresponding risk

ControlFluoroquinolones versus macrolides

Diarrhoea on follow up
clinical criteria
Follow-up: 6 to 10 days
105 per 100063 per 1000
(25 to 156)
RR 0.6
(0.24 to 1.49)
189
(2 studies)
⊕⊝⊝⊝
very low1,2,3,4
One trial reported that none of the participants had diarrhoea on day 10 and in the other 16/76 had diarrhoea on the sixth day

Relapse - not reportedSee commentSee commentNot estimable-See commentDuration of follow up in both trials were too short (6 to 10 days) to assess relapse and none were reported.

Serious adverse events - not reportedSee commentSee commentNot estimable-See commentNone of the two trials reported that any participant developed serious adverse events

Other adverse events
clinical criteria
Follow-up: 6 days
Study populationRR 1.33
(0.32 to 5.56)
76
(1 study)
⊕⊝⊝⊝
very low5,6,7

79 per 1000105 per 1000
(25 to 439)

Medium risk population

79 per 1000105 per 1000
(25 to 439)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Serious limitations: One of the two included trials had limitations in allocation concealment and both reported outcomes for less than 90% of those randomized (82% and 87%)
2 No serious inconsistency: One of the trials had no participants with this outcome and hence risk ratios were estimated for only one trial.
3 Serious indirectness: Both trials randomized only adults. Effects of fluoroquinolones over macrolides in children, especially those who are malnourished are unclear. Antibiotics used were azithromycin and ciprofloxacin in both trials.
4 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with ciprofloxacin and azithromycin.
5 Very serious limitation: The trial reported this outcome only for 82% of randomized participants.
6 Serious indirectness: The trial included only adults. The antibiotics studied were ciprofloxacin and azithromycin.
7 Very serious imprecision:The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with ciprofloxacin and azithromycin.
 
Summary of findings 4. Cotrimoxazole versus beta-lactams for Shigella dysentery

Cotrimoxazole versus beta-lactams for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Guatemala and USA
Intervention: Cotrimoxazole versus beta-lactams

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

Assumed riskCorresponding risk

ControlCotrimoxazole versus beta-lactams

Diarrhoea on follow up
clinical criteria
Follow-up: 11 to 21 days
227 per 1000134 per 1000
(52 to 338)
RR 0.59
(0.23 to 1.49)
89
(2 studies)
⊕⊝⊝⊝
very low1,2,3,4
One trial was reported in 1976 and the other in 1993. The antibiotics compared with cotrimoxazole were ampicillin and pivmecillinam respectively.

Relapse - not reportedSee commentSee commentNot estimable-See commentThe two trials followed participants for 11 to 21 days but did not report any relapses in this time.

Serious adverse events - not reportedSee commentSee commentNot estimable-See commentNo serious adverse events leading to death or hospitalization were reported in either trial.

Other adverse events
clinical criteria
Follow-up: 11 to 21 days
136 per 1000110 per 1000
(37 to 333)
RR 0.81
(0.27 to 2.45)
89
(2 studies)
⊕⊝⊝⊝
very low1,2,3,5

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Serious limitations: Inadequate allocation concealment in one trial and inadequate outcome data reporting (for 39% of randomized participants whose cultures were negative for Shigella) in the other
2 No inconsistency: I squared was 0% and the direction of effect favoured cotrimoxazole in both trials.
3 Serious indirectness: Both trials included only infants and children. The antibiotics compared were cotrimoxazole versus ampicillin and pivmecillinam.
4 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with cotrimoxazole and ampicillin and pivmecillinam.
5 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with beta-lactams and cotrimoxazole.
 
Summary of findings 5. Cotrimoxazole versus fluoroquinolones (norfloxacin) for Shigella dysentery

Cotrimoxazole versus fluoroquinolones (norfloxacin) for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Peru
Intervention: Cotrimoxazole versus fluoroquinolones (norfloxacin)

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

Assumed riskCorresponding risk

ControlCotrimoxazole versus fluoroquinolones (norfloxacin)

Diarrhoea on follow up - not reportedSee commentSee commentNot estimable-See commentOutcome assessed as number of days to last unformed stool. Data not available for proportions with diarrhoea on follow up.

Relapse - not reportedSee commentSee commentNot estimable-See commentThe trial followed up participants for 14 days. Relapses were not reported in this time.

Serious adverse events - not reportedSee commentSee commentNot estimable-See commentNo participant is reported to have developed serious adverse events leading to death or hospitalisation.

Other adverse events
clinical criteria
Follow-up: 2 weeks
0 per 10000 per 1000
(0 to 0)
RR 2.82
(0.12 to 66.62)
62
(1 study)
⊕⊝⊝⊝
very low1,2,3

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Very serious limitations: Inadequate allocation concealment and blinding and very inadequate outcome data reporting (for only 32% of 174 randomized). Baseline imbalance in antibiotic sensitivity (100% sensitivity in norfloxacin arm and 84% in the cotrimoxazole arm).
2 Serious indirectness: The trial included only adults.
3 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with cotrimoxazole and norfloxacin.
 
Summary of findings 6. Cotrimoxazole versus furazolidone for Shigella dysentery

Cotrimoxazole versus furazolidone for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Mexico
Intervention: Cotrimoxazole versus furazolidone

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

Assumed riskCorresponding risk

ControlCotrimoxazole versus furazolidone

Diarrhoea on follow up
clinical criteria
Follow-up: 6 days
173 per 1000123 per 1000
(47 to 318)
RR 0.71
(0.27 to 1.84)
101
(1 study)
⊕⊝⊝⊝
very low1,2,3
Trial reported in 1989; antimicrobial sensitivity to Shigella isolates not reported

Relapse - not reportedSee commentSee commentNot estimable-See commentFollow up duration too short (6 days) in the sole trial for this comparison

Serious adverse eventsMedium risk populationRR 0
(0 to 0)
0
(0)
See commentNo participant is reported to have developed serious adverse events leading to death or hospitalization.


Other adverse events - not reportedSee commentSee commentNot estimable-See commentNo adverse effects reported; unclear if formally evaluated

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Very serious limitations: Risk of bias likely due to inadequate allocation concealment and blinding. Baseline imbalances in participant characteristics (significantly fewer days of diarrhoea in those allocated to furazolidone- P =0.02).
2 Serious indirectness: The single trial included only infants and children.
3 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with cotrimoxazole over furazolidone.
 
Summary of findings 7. Oral gentamicin versus nalidixic acid for Shigella dysentery

Oral gentamicin versus nalidixic acid for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Bangladesh
Intervention: Oral gentamicin versus nalidixic acid

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

Assumed riskCorresponding risk

ControlOral gentamicin versus nalidixic acid

Diarrhoea at follow up
clinical criteria
Follow-up: 5 days
308 per 1000527 per 1000
(302 to 915)
RR 1.71
(0.98 to 2.97)
79
(1 study)
⊕⊝⊝⊝
very low1,2,3
Data from a single trial reported in 1994. Antimicrobial sensitivity for Shigella isolates was 100% in those allocated to oral gentamicin and 70% to those allocated to nalidixic acid.

Relapse - not reportedSee commentSee commentNot estimable-See commentFollow up duration too short (5 days) to assess.

Serious adverse events - not reportedSee commentSee commentNot estimable-See commentNo participant is reported to have developed serious adverse events

Other adverse events - not reportedSee commentSee commentNot estimable-See commentNo adverse effects reported; unclear if systematically assessed.

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Very serious limitations: Though randomization, allocation and blinding were adequate, data were reported only for 87% randomized and there were baseline imbalances in antibiotic sensitivity (100% sensitive in gentamicin arm and 70% in nalidixic acid arm).
2 Serious indirectness: The trial randomized only infants and children and specifically excluded those severely malnourished.
3 Serious imprecision: The 95% CI for the point estimate from the trial includes appreciable and non-appreciable benefit for nalidixic acid over oral gentamicin.
 
Summary of findings 8. Sulphonamides versus tetracycline for Shigella dysentery

Sulphonamides versus tetracycline for Shigella dysentery

Patient or population: patients with Shigella dysentery
Settings: Sri Lanka
Intervention: Sulphonamides versus tetracycline

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

Assumed riskCorresponding risk

ControlSulphonamides versus tetracycline

Diarrhoea at follow up
clinical criteria
Follow-up: 8 days
0 per 10000 per 1000
(0 to 0)
RR 7.68
(0.46 to 128.12)
60
(1 study)
⊕⊝⊝⊝
very low1,2,3
Trial reported in 1961. Antimicrobial sensitivity not reported

Relapse - not reportedSee commentSee commentNot estimable-See commentDuration of follow up too short (8 days) to assess relapse

Serious adverse events - not reportedSee commentSee commentNot estimable-See commentNo participant is reported to have developed serious adverse events.

Other adverse events - not reportedSee commentSee commentNot estimable-See commentNot reported or pre-stated as an outcome; unclear if assessed.

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) 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 Very serious limitations: Risk of bias likely due to inadequate allocation concealment and blinding and unclear reporting of numbers randomized and numbers analysed.
2 Unclear indirectness: Unclear from report if trial included adults and children; malnourished participants were not specifically excluded.
3 Very serious imprecision: The 95% CI of the pooled estimate includes appreciable benefit and appreciable harm with tetracycline and sulphonamides.
 
Table 1. Known adverse effects of antibiotics used to treat Shigella dysentery^

AntibioticLife threateningDiscontinuation^^Other

TetracyclineAnaphylaxisOesophageal irritation, antibiotic-associated colitis, headache and visual disturbancesIn children under 12 years of age causes dental hypoplasia and staining, benign intracranial hypertension

ChloramphenicolBlood disorders, peripheral and optic neuritis, erythema multiformeDyspepsia

AmpicillinHypersensitivity reactionsDiarrhoea

Co-trimoxazole or trimethoprim - sulphamethoxazoleStevens-Johnson syndromeDiarrhoea, rash

FluoroquinolonesHypersensitivityDyspepsia, headache, hypotensionPruritis, tachycardia

NorfloxacinDyspepsia, headache, hypotensionEuphoria, tinnitus, polyneuropathy

CiprofloxacinDyspepsia, headache, hypotensionHot flushes, sweating, tenosynovitis

OfloxacinDyspepsia, headache, hypotensionAnxiety, unsteady gait

AzithromycinHypersensitivityDyspepsia, flatulence, headache

CeftriaxoneHypersensitivity reactionsDiarrhoea, headache, abdominal discomfort

Nalidixic acidSame as in fluoroquinolonesToxic psychosis, increased intracranial tension, cranial nerve palsy

RifaximinAllergic reactionsAllergic reactions

CefiximeHypersensitivity reactionsFlatulence, headache, abdominal pain, defecation urgency, nausea, constipation, pyrexia, vomiting

PivmecillinamSame as ampicillin, dyspepsia

 ^Source: BNF 2007.
^^Can result in discontinuation of treatment.
 
Table 2. Detailed search strategies

Search setCIDG SR^CENTRALMEDLINE^^EMBASE^^LILACS^^

1Shigell*Shigell*Shigell*Shigell$Shigell*

2DysenteryDYSENTERY, BACILLARYDYSENTERY, BACILLARYSHIGELLOSISDysentery

31 or 21 or 21 or 2DYSENTERY1 or 2

4antibiotic*ANTI-BACTERIAL AGENTS/THERAPEUTIC USEANTI-BACTERIAL AGENTS/THERAPEUTIC USE1 or 2 or 3antibiotic*

5tetracycline*ANTI-INFECTIVE AGENTS/THERAPEUTIC USEANTI-INFECTIVE AGENTS/THERAPEUTIC USEtetracycline$tetracycline*

6chloramphenicolantibiotic*antibiotic*chloramphenicolchloramphenicol

7ampicillin*tetracycline*tetracycline*ampicillinampicillin

8co-trimoxazolechloramphenicolchloramphenicolco-trimoxazoleco-trimoxazole

9fluoroquinolone*ampicillinampicillinfluoroquinolone$fluoroquinolone*

10quinolone*co-trimoxazoleco-trimoxazolequinolone$quinolone*

11norfloxacinfluoroquinolone*fluoroquinolone*norfloxacinnorfloxacin

12ciprofloxacinquinolone*quinolone*ciprofloxacinciprofloxacin

13ofloxacinnorfloxacinnorfloxacinofloxacinofloxacin

14azithromycinciprofloxacinciprofloxacinazithromycinazithromycin

15ceftriaxoneofloxacinofloxacinceftriaxoneceftriaxone

16nalidixic acidazithromycinazithromycinnalidixic acidnalidixic acid

17pivmecillinamceftriaxoneceftriaxonerifaximinrifaximin

184-17/ornalidixic acidnalidixic acidcefiximecefixime

193 and 18rifaximinrifaximintrimethoprim-sulfamethoxazoletrimethoprim-sulfamethoxazole

20cefiximecefiximeantibiotic$pivmecillinam

21trimethoprim-sulfamethoxazoletrimethoprim-sulfamethoxazolepivmecillinam4-20/or

22pivmecillinampivmecillinam5-21/or3 and 21

234-22/or4-22/orLimit 22 to human

243 and 233 and 23

25Limit 24 to human

 ^Cochrane Infectious Diseases Group Specialized Register.
^^Search terms used in combination with the search strategy for retrieving trials developed by The Cochrane Collaboration (Higgins 2006); upper case: MeSH or EMTREE heading; lower case: free text term.
 
Table 3. Search strategy: proceedings, organizations, and pharmaceutical companies

TypeDetail

Conference proceeding- Commonwealth Congress on Diarrhoea and Malnutrition: 8th, Dhaka, Bangladesh, 6 to 8 February 2006 (searched on 12 April 2007)
- Asian Conference on Diarrhoeal Diseases and Nutrition: 10th, Dhaka, Bangladesh, 7 to 9 December 2003 (searched on 13 April 2007)
- Annual Scientific Conference: 10th Dhaka, Bangladesh, 11 to 13 June 2002 (searched on 13 April 2007)
- Annual Meeting of Infectious Disease Society of America: 44th, Toronto, Ontario, Canada, 12 to 15 October 2006; 43rd, San Francisco, California, 6 to 9 October 2005; 42nd, Boston, Massachusetts, USA, 30 September to 3 October 2004 (searched on 18 March 2008)
- Interscience Conference on Antimicrobial Agents and Chemotherapy: 46th, San Francisco, California, 27 to 30 September 2006; 45th, Washington DC, USA, 16 to 19 December 2005; 44th, Washington DC, USA, 30 October to 2 November, 2004 (searched on 18 March 2008)
- European Congress of Clinical Microbiology and Infectious Diseases: 16th, Nice, France, 1 to 4 April 2006; 15th, 2 to 5 April 2005 (searched on 18 March 2008)
- International Congress on Infectious Diseases: 12th, Lisbon, Portugal, 15 to 18 June 2006; 11th, Cancun, Mexico, 4 to 7 March 2004 (searched on 18 March 2008)
- Annual Meeting of The European Society for Paediatric Infectious Disease: 24th, Basel, Switzerland, 3 to 5 May 2006 (searched on 18 March 2008)
- Western Pacific Congress of Chemotherapy and Infectious Diseases: 10th, Fukuoka, Japan, 3 to 6 December 2006 (searched on 18 March 2008)
- European Congress of Chemotherapy and Infection: 8th, Budapest, Hungary, 25 to 28 October 2006 (searched on 18 March 2008)

Organizations- Liverpool School of Tropical Medicine (contacted on 11 April 2007)
- World Health Organization (contacted on 17 March 2008)
- American Society of Tropical Medicine and Hygiene (contacted on 15 April 2007)
- International Society of Tropical Pediatrics (contacted on 15 April 2007)
- South East Asian Ministers Education Organization (SEAMEO) TROPMED Network (contacted on 17 March 2008)
- International Center for Diarrhoeal Disease Research in Bangladesh (contacted on 21 April 2007)

Pharmaceutical companies- Goldshield Pharmaceuticals Ltd (tetracycline, Deteclo; chloramphenicol, Chloromycetin) - contacted on 17 March 2008
- Chemidex (ampicillin, Penbritin) - contacted on 17 March 2008
- GlaxoSmithKline (co-trimoxazole, Septrin) - contacted on 17 March 2008
- Merck Sharp & Dohme Ltd (norfloxacin, Utinor) - contacted on 17 March 2008
- Bayer (ciprofloxacin, Ciproxin) - contacted on 20 April 2007
- Aventis Pharma (ofloxacin, Tarivid) - contacted on 15 April 2007
- Pfizer (azithromycin, Zithromax) - contacted on 17 March 2008
- Roche (ceftriaxone, Rocephin) - contacted on 20 April 2007
- Rosemont Pharmaceuticals Ltd (nalidixic acid, Uriben) - contacted on 13 April 2007
- Salix Pharmaceuticals (rifaximin, Xifaxan) - contacted on 17 March 2008
- Rhone-Poulenc Rorer (cefixime, Suprax) - contacted on 17 March 2008
- LEO pharma (pivmecillinam, Selexid) - contacted on 17 March 2008

 
Table 4. Sensitivity patterns of the Shigella isolates reported in included trials

Study IDGroup 1Group 2Group 3

Alam 1994Pivmecillinam group:

All were sensitive to pivmecillinam

Nalidixic acid sensitivity not reported
Nalidixic acid group:

All were sensitive to pivmecillinam

26/37, 45%, were sensitive to nalidixic acid
Nil

Bennish 1990Ciprofloxacin group:

All were sensitive to ciprofloxacin; 34/60, 56.6%, were sensitive to ampicillin
Ampicillin group:

All were sensitive to ciprofloxacin; 26/61, 42.6%, were sensitive to ampicillin.
Nil

Bibile 1961This is a 4-armed trial with sulphadimidine, sulpha methoxy pyridazine, Strepto triad, and tetracycline in each group respectively

Sensitivity patterns not reported for any group

Dutta 1995Furazolidone and nalidixic acid

Sensitivity patterns not reported for any group

Gotuzzo 1989Cotrimoxazole group:

27/32, 84%, were sensitive to cotrimoxazole; all were sensitive to norfloxacin
Norfloxacin group:

26/30, 86%, were sensitive to cotrimoxazole; all were sensitive to norfloxacin
Nil

Haltalin 1973Nalidixic acid group:

All were sensitive to nalidixic acid; ampicillin sensitivity not reported.
Ampicillin group:

All were sensitive to ampicillin; nalidixic acid sensitivity not reported
Nil

Islam 1994Nalidixic acid group:

26/37, 70%, were sensitive to nalidixic acid; all were sensitive to gentamicin
Oral gentamicin:

Nalidixic acid sensitivity not reported; all were sensitive to gentamicin
Nil

Kabir 1986Ceftriaxone group:

All were sensitive to ceftriaxone; all were sensitive to ampicillin
Ampicillin group:

All were sensitive to ceftriaxone; 24/30, 80%, were sensitive to ampicillin
Placebo:

All were sensitive to ceftriaxone; 28/30, 93%, were sensitive to ampicillin

Khan 1997aAzithromycin group:

All were sensitive to both antibiotics
Ciprofloxacin group:

All were sensitive to both antibiotics
Nil

Leibovitz 2000Ciprofloxacin group:

All were sensitive to both antibiotics
Ceftriaxone group:

All were sensitive to both antibiotics
Nil

Nelson 1976aCotrimoxazole group:

All were sensitive to cotrimoxazole; 9/14, 64%, were sensitive to ampicillin
Ampicillin group:

All were sensitive to cotrimoxazole 10/14, 71%, were sensitive to ampicillin
Nil

Prado 1993Cotrimoxazole group;

24/30, 80%, were sensitive to cotrimoxazole; 25/30, 83.3%, were sensitive to pivmecillinam
Pivmecillinam group:

23/29, 79.3%, were sensitive to cotrimoxazole; 26/29, 89.7%, were sensitive to pivmecillinam
Nil

Rodriguez 19893-armed trial with furazolidone, cotrimoxazole and a control (no antimicrobials) respectively

Sensitivity patterns not reported for any group
Nil

Salam 1988Nalidixic acid group:

All were sensitive to nalidixic acid; ampicillin sensitivity not reported
Ampicillin group:

All were sensitive to nalidixic acid; 25/40, 62.5%, were sensitive to ampicillin
Nil

Salam 1998Ciprofloxacin group:

All were sensitive to ciprofloxacin; 58/60, 96.7%, were sensitive to pivmecillinam
Pivmecillinam group:

All were sensitive to ciprofloxacin; 57/60, 95%, were sensitive to pivmecillinam
Nil

Shanks 1999Azithromycin and ciprofloxacin

Sensitivity patterns not reported for any group
Nil

 Sensitivity patterns not reported by 4 trials (Bibile 1961; Rodriguez 1989; Dutta 1995; Shanks 1999).
 
Table 5. Suggestions for a trial of antibiotic for Shigella dysentery

MethodsParticipantsInterventionsOutcomesNotes

Allocation:
Centralized sequence generation with table of random numbers or computer generated lists

Stratified by severity of illness

Sequence concealed until interventions are assigned

Blinding:
Those recruiting and assigning participants, those administering the intervention, and those assessing the outcomes, must all be blind to the allocated group; the administered drugs have to be identical or a double dummy technique has to be used. Liquid medications have to be in similar looking bottles, identical in shape and weight; the medications must themselves be similar in colour and flavour.

Duration:
Minimum of 4 weeks after completion of therapy to assess relapse
Entry criteria can be clinical dysentery, i.e. acute onset frequent loose stools with blood or mucus or both lasting for less than 72 hours and at least 3 stools per day. Other features, such as fever and tenesmus at presentation, have to be recorded but need not be necessary for inclusion into study.

If it is possible to presumptively or decisively detect Shigella in stool before inclusion into study, it should be done. Real-time PCR is a rapid but expensive method to diagnose Shigella early (Legros 2004).

Sample size:
(See  Table 6). Age group: trials should be separately done for adults and children (less than 15 years of age) or at least presented separately if they are in the same trial. In children, infants must be a separate group.

Setting: in- or out-patients. The number of participants, if hospitalized for standardization of administration of the interventions, have to be reported separately from those hospitalized due to complications.

Sex: men and women.

Special groups (those who have higher risk of complications:

  • Malnourished children
  • HIV positive individuals
  • Adults more than 50 years of age
  • Infants


Exclusion criteria:
Allergy to the drug studied; history of antibiotic use for this episode of illness in the previous 48 hours; pregnant and lactating women; clinical presence of another infection needing antimicrobials
  1. Any antibiotic studied for efficacy and safety
  2. Any other antibiotic that is the standard for the treatment of Shigella dysentery at that period of time in that country


Others: placebos or probiotics to be studied only on those with no risk of complications and those who have mild illness
Primary outcomes:

  1. Number of patients with diarrhoea on follow up.
  2. Clinical relapse
  3. Adverse effects of antibiotics
    1. Life threatening adverse effects of the drug
    2. Those that require discontinuation of the drug
    3. MIld adverse events that need extra therapy but not discontinuation of the drug
  4. Duration of fever
  5. Duration of blood in stools


Secondary outcomes:

  1. Removed from study due to clinical worsening
  2. Fever on follow up
  3. Abdominal pain on follow up
  4. Bacteriological cure
  5. Bacteriological relapse
  6. Duration of diarrhoea
  7. Duration of abdominal pain
  8. Number of days of hospitalisation
Once patients are randomized into the treatment groups, they should not be removed until final analysis. The trial author(s) must publish the outcome findings of the whole group first and then present data for those positive for Shigella by stool or rectal swab culture or PCR and those negative for Shigella. The data have to be presented according to the severity of illness the patients presented with.

Antibiotic sensitivity patterns have to be reported for all antibiotics studied and in all groups

Response to treatment stratified by in vitro antibiotic sensitivity also needs to be reported

Drop-outs:
The patients who drop out after randomization due to loss of follow up, withdrawal from protocol or consent withdrawal etc have to be reported and accounted in the final analysis (intention-to-treat analysis).

 
Table 6. Sample size suggestions for trial of antibiotics in Shigella dysentery

SAMPLE SIZESAntibiotic versus no drug or placebo (placebo response at 45%)

or

Antibiotic versus another antibiotic

1 sided α10% difference: 310

20% difference: 75

25% difference: 50

30% difference: 30

40% difference: 15

2 sided α10% difference: 390

20% difference: 95

25% difference: 60

30% difference: 40

40% difference: 20

 1. The sample size required to detect the assumed difference in improvement or worsening with 80% power and 5% significance level using the outcome of 'diarrhoea at follow up' from this review using StatCalc 2006.
2. The sample size mentioned is for each arm of the study.