Table 3.

The risk of key biases present within each study

Author, year	Random sequence generation	Allocation concealment	Selective reporting	Other biases	Blinding of participants and personnel	Blinding of outcome assessment	Incomplete data outcome	Additional information
Rudzki et al (2019)20	Low	Low	Unclear	Low	Low	Unclear	High	19 lost to follow-up; information concerning these individuals not included in report
Majeed et al (2018)23	Unclear	Unclear	Low	High	Unclear	Unclear	Low	Authors employed by manufacturer of the intervention—Sabinsa, USA. Sabinsa also funded the trial
Kazemi et al (2018)21	Low	Low	Low	Low	Low	Unclear	High	Attrition bias present; however, all subjects accounted for and details of loss to follow-up well explained
Miyaoka et al (2018)24	Unclear	High	Low	Unclear	High	High	Low	Random group allocation insufficiently detailed; open label design (no blinding).
Ghorbani et al (2018)25	Low	Low	Low	Low	Low	Unclear	Low	Independent company employed for randomisation protocols, no loss to follow-up. Blinding of results unspecified
Bambling et al (2017)26	N/A	N/A	Low	High	High	N/A	High	No randomisation or control group. Patients and investigators knew what intervention was. Patients lost to follow-up (accounted for). Potential conflict of interest—researchers work with companies that seed fund, produce and research probiotics
Akkasheh et al (2016)22	Low	Unclear	Low	Unclear	Unclear	Unclear	Low	Randomisation of groups and allocation claimed to be done; claimed outcomes were blinded, though no explicit details provided (therefore unclear). Attrition bias, participants included in analysis with intention-to-treat approach. Conflict of interest not stated