Fiscella 2006.

Methods	Single‐centre blinded placebo RCT
Participants	Participants randomised to 1 of 2 groups in a 1:1 ratio (n = 42) ‐ 22 mifepristone 5 mg, 2 withdrawn, 20 analysed ‐ 20 placebo, 3 withdrawn, 17 analysed Randomisation stratified according to ‐ Uterine Fibroid Symptom Quality of Life (UFS‐QoL) symptom severity (> 64 vs ≤ 63) Baseline demographics similar in each group except for ‐ BMI and baseline uterine volume (higher in treatment group) Inclusion criteria ‐ Age ≥ 18 years ‐ Premenopausal ‐ At least moderately severe fibroid‐related symptoms (> 39 on UFS‐QoL symptom severity subscale) ‐ Total uterine volume ≥ 160 mL ‐ At least 1 fibroid ≥ 2.5 cm ‐ Had not used short‐acting hormone analogues or other long‐acting hormone medications in past 6 months Exclusion criteria ‐ Pregnant or intending pregnancy during next 6 months ‐ Major medical morbidity ‐ Severe anaemia ‐ Active mental illness ‐ Elevated liver enzymes ‐ Substance abuse Setting: New York, USA
Interventions	Mifepristone 5 mg orally daily vs placebo Duration: 26 weeks ‐ Participants agreed to use barrier contraception and not to use hormonal or surgical treatments for fibroids during the study ‐ Participants were allowed to use analgesics and were asked to record use
Outcomes	Outcomes assessed at baseline and at 1, 3 and 6 months Primary outcome ‐ Fibroid‐specific overall QoL (UFS‐QoL) Secondary outcomes ‐ Global health status (Medical Outcomes 36‐item Short Form (SF‐36) survey) ‐ Global pain (McGill Pain Questionnaire) assessed monthly ‐ Bleeding (daily menstrual logs and pictorial charts) ‣ Amenorrhoea ‣ Mean monthly blood loss index ‣ Haemoglobin ‐ Uterine/fibroid (5 largest summed) ultrasound volumes (measured in 3 planes) ‐ Fibroid symptoms (5‐point Likert scale) assessed monthly ‣ Pelvic pain ‣ Pelvic pressure ‣ Bladder pressure ‣ Urinary frequency ‣ Low back pain ‣ Rectal pain ‣ Pain with intercourse ‐ Adverse effects (5‐point Likert scale) assessed monthly ‐ Liver function tests ‐ Endometrial biopsy assessed at baseline and at 6 months. Did not report on PAEC
Notes	19/20 women in the treatment group correctly guessed that they had been receiving mifepristone because of cessation of bleeding
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Using random numbers generated with SAS 9, women were randomly assigned in blocks of four, stratified by Uterine Fibroid Symptom Quality of Life”
Allocation concealment (selection bias)	Low risk	“Study assignments were placed in opaque sealed envelopes that were opened by the study pharmacist once the participant was fully qualified”
Blinding of participants and personnel (performance bias) All outcomes	Low risk	“The study pharmacist prepared mifepristone 5 mg and placebo in capsules that were identical in appearance and weight” “None of the study personnel, with the exception of the pharmacist, were aware of treatment assignments”
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"None of the study personnel, with the exception of the pharmacist, were aware of treatment assignments" However, blinding of outcome assessors was not specifically described
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing outcome data relatively balanced in numbers across intervention groups and unlikely to be related to the true outcome “3 [women] completed the intake measures and were randomised (two to treatment and one to placebo) but then declined to participate, and 39 began the trial and participated for at least one month” “Two other women (from placebo group) dropped out during the course of the study. Neither reported leaving due to adverse effects”
Selective reporting (reporting bias)	Low risk	Main outcomes reported. Protocol available at clinicaltrials.gov NCT00133705
Other bias	Low risk	None identified