Malkawi 2003.

Methods	RCT Setting: Jordan Method of randomisation: unclear Blinding: none Number randomised: 161
Participants	Summary: metformin vs LOD in CC‐resistant PCOS women Inclusion criteria: PCOS, diagnosed if polycystic ovaries on US and ≥ 3 of oligomenorrhoea, hirsutism, hyperandrogenism, elevated LH, LH:FSH ratio > 2, CC resistance (failure to ovulate or conceive after CC treatment up to a daily dose of 150 mg in at least 3 consecutive cycles), normal uterine cavity, normal tubal patency, normal semen parameters Exclusion criteria: congenital adrenal hyperplasia, Cushing's syndrome, hyperprolactinaemia, thyroid disease Baseline characteristics of each group: metformin (n = 64) vs LOD (n = 97) Mean age (SD) 27.4 (3) vs 27.1 (4.4) Mean BMI (SD) 27.5 (4.1) vs 26.6 (2.3) Mean fasting blood sugar, mg/dL (SD) 80.6 (5.6) vs 81 (5.2) Mean fasting insulin, micU/mL (SD) 20.5 (4.2) vs 19 (7.2) Mean testosterone, ng/mL (SD) 330 (48) vs 290 (88) Regular menstrual cycles, number (%) 17 (26.6) vs 27 (27.8) Dropouts: none
Interventions	Main intervention: metformin 850 mg twice daily; LOD (8‐10 punctures per ovary, each for 2‐3 s with insulated needle adjusted at 40 watts, ovaries then washed with crystalloid solution) Duration: 3 months then if no ovulation, CC was added to both groups Co‐interventions: CC 50 mg/d starting on days 5‐9 of cycle. If no ovulation, dose increased to 100 mg/d then 150 mg/d in each consecutive cycle
Outcomes	Primary: live birth rate, gastrointestinal side effects Secondary: clinical pregnancy, menstrual frequency, ovulation: serum progesterone > 10 ng/mL, BMI, fasting blood glucose, fasting insulin, serum testosterone, miscarriage, multiple pregnancy, other adverse effects: ectopic pregnancy
Notes	No information on method of randomisation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No method of randomisation
Allocation concealment (selection bias)	Unclear risk	No information
Blinding of participants and personnel (performance bias) All outcomes	High risk	No information however the study compares LOD vs metformin therefore blinding not achievable
Blinding of outcome assessment (detection bias) All outcomes	High risk	No information however the study compares LOD vs metformin therefore blinding not achievable
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts
Selective reporting (reporting bias)	Unclear risk	Insufficient information in the study
Other bias	Low risk	No evidence of other bias however limited reported methodology