Hajifaraji 2017.

Study characteristics
Methods	Double‐blind, placebo‐controlled randomised trial
Participants	Pregnant women with GDM (n = 64) Setting Gynecology and Endocrinology clinics at Tabriz Al‐Zahra referral University hospital, Tabriz, Iran Dates of study Spring and summer 2014 Inclusion criteria Nulliparity, newly diagnosed gestational diabetes at 24 and 28 weeks + 6 days, GDM diagnosed by a fasting BG of 92 mg/dL to125 mg/dL at time of diagnosis. Age 18 to 45 years, with BMI ≥ 18.5 kg/m2. No history of IGT in early pregnancy, T2DM, chronic diseases, smoking, alcohol, ingestion of probiotics (including yoghurt, kefir, other fermented food) within 2 weeks of intervention, or antibiotic use 1 month prior to intervention. No acute GI problems for a month prior to intervention. Exclusion criteria Needing insulin therapy or other medications (fasting BG > 105 and BG postprandial > 120 mg/dL) during intervention. Ingestion of other forms of probiotics (yoghurt, kefir, fermented foods), antibiotics, glucocorticoids, immunosuppressive drugs, having acute GI problems, and not taking adequate number of tablets.
Interventions	Probiotic: capsules (4Biocap) containing 180 mg (> 4 x 109 CFU) standard power including freeze dried cultures of Lactobacillus acidophilusLA‐5, Bifibacterium BB‐12,Streptococcus thermophiles STY‐31, and Lactobacillus delbrueckii bulgaricus LBY‐27 + dextrose anhydrate filler and magnesium stearate lubricant (n = 32). Placebo: capsules similar design, shape and colour to the probiotics (n = 32).
Outcomes	Systolic and diastolic BP changes from baseline, to 2 weeks, 4 weeks, 6 weeks and 8 weeks after recruitment. Serum inflammatory markers: serum hs‐CRP, lnTNF‐α, Serum IL‐6. Oxidative stress markers: serum TAC, serum MDA, serum GSHR, erythrocyte GPX, serum uric acid, erythrocyte SOD. Weight changes, FPG, serum insulin levels, HOMA‐IR, QUICKI index
Notes	Funding: Tehran, Shahid Beheshti University Medical Sciences. Tehran Darou pharmaceuticals, Tehran, Iran – provided the 4‐Biocap (probiotic supplement). Conflicts of interest: authors declare no conflict of interest.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random sequence. To confirm double‐blinding, a coder secretly labelled the capsule packages A or B.
Allocation concealment (selection bias)	Unclear risk	All 64 females were allocated using block randomisation techniques to either probiotic or placebo; it is unclear whether the allocation was concealed.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and clinicians were blinded to the intervention.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors were blinded to the intervention.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Of the 64 women who were randomised, 56 were analysed in the final analysis. Reasons for discontinuation: probiotic arm, (3 loss to follow‐up); declined to continue (1); needing drug therapy (2): n = 29 (91%). In the placebo arm: (5 loss to follow‐up); declined to continue (2), preterm pregnancy (1), needing drug therapy (2).
Selective reporting (reporting bias)	Unclear risk	No retrospective trial registration. Insufficient information to judge if pre‐specified outcomes are reported in full.
Other bias	Low risk	Baseline characteristics were balanced across groups.