Table 3. Risk of bias assessment
Gleckman et al30Iravani31van Nieuwkoop et al29
Sequence g eneration Low: table of random digitsLowLow: randomised stratified per centre and sex. Computer-generated randomisation list
Allocation c oncealment Unclear: no information providedNo commentLow: treatment allocation completed after urine culture results. Restricted access to independent pharmacy
Blinding of participants and personnel for all outcomes Unclear: no description of the blinding providedHigh: no blinding of the participants or personnel only investigatorsLow: double blinding
Blinding of outcome assessors for all outcomes Unclear: no description of the blinding providedLow: investigators were blinded through third party. The drugs were dispensed by an independent third party to ensure investigator blindingLow: analysis based on intention to treat population
Incomplete outcome data for all outcomes High: data not reported for two patients in each group suffering adverse eventsHigh: the main outcome (clinical recovery) is reported for 436/727 patients onlyLow
Selective outcome reporting LowUnclear: both outcomes assessed were reported, but no pre-published protocol is available to control this with the initial designUnclear: all outcomes described in methods chapter are reported
Other sources of bias LowHigh: men are just 5% of population and a subgroup of the study. Dropout is about 50% for bacteriological cure and unclear for clinical cure.None identified