Donnez 2012.

Methods	Multi‐centre blinded placebo RCT
Participants	Participants randomised to 1 of 3 groups in a 2:2:1 ratio (n = 242) ‐ 96 ulipristal acetate 5 mg: 5 withdrawn, 91 analysed ‐ 98 ulipristal acetate 10 mg: 6 withdrawn, 92 analysed ‐ 48 placebo: 1 withdrawn, 47 analysed Randomisation stratified according to ‐ Baseline haematocrit (≤ 28% or > 28%) ‐ Race (black or other) Baseline demographics similar in each group Inclusion criteria ‐ 18 to 50 years old ‐ PBAC score > 100 during day 1 to 8 of menstruation ‐ Fibroid‐related anaemia (Hgb ≤ 10.2 g per decilitre without macrocytosis) ‐ Fibroid uterus with size ≤ 16 weeks ‐ At least 1 fibroid ≥ 3 cm and no fibroid > 10 cm in diameter (ultrasonography) ‐ BMI 18 to 40 Exclusion criteria ‐ History of uterine surgery (except for cesarean section or cervical conisation) ‐ Endometrial ablation or uterine artery embolisation ‐ History of concurrent gynaecological cancer ‐ Current endometrial hyperplasia ‐ Hgb ≤ 6 g/L or any condition requiring immediate blood transfusion ‐ Known haemoglobinopathy ‐ Known severe coagulation disorder ‐ Large uterine polyp (> 2 cm) ‐ 1 or more ovarian cyst ≥ 4 cm in diameter (ultrasonography) ‐ Previous or current fibroid treatment with an SPRM or GnRH agonist ‐ Treatment with agents known to affect hepatic cytochrome CYP3A4 ‐ Progestins, acetylsalicylic acid, mefenamic acid, anticoagulants, antifibrinolytic drugs or systemic glucocorticoid treatments Setting: 38 centres in 6 European countries
Interventions	Ulipristal acetate 5 and 10 mg orally daily vs placebo Duration: 13 weeks starting on cycle day 1 to 4 ‐ After 13 weeks, participant could undergo surgery ‐ No further pharmacological treatment of fibroid administered All participants received 80 mg oral iron supplementation once daily
Outcomes	Outcomes assessed at baseline and at 13 weeks Co‐primary efficacy endpoints (baseline and 13 weeks) ‐ Reduction in bleeding (PBAC < 75) ‐ Total fibroid volume (MRI, sum of all fibroid volumes) Secondary endpoints ‐ Bleeding pattern (change in PBAC) ‐ Amenorrhoea ‐ Reduction in uterine and fibroid volume ‐ Change in Hgb, Hct and ferritin levels ‐ Pain (Short‐Form McGill Questionnaire and VAS) ‐ Discomfort questionnaire ‐ Adverse events (up to week 38) Safety endpoints ‐ Endometrial thickness (MRI at week 13, also at weeks 26 and 38 if no surgery) ‐ Endometrial biopsy (screening, week 13, week 38 if no hysterectomy or ablation). Used standard definition of SPRM‐associated endometrial changes (PAEC) described by Mutter 2008 ‐ Serum oestradiol, progesterone, corticotropin, thyrotropin, prolactin (screening and weeks 5, 9, 13 and 17) ‐ Haematological, coagulation, biochemical variables, lipids, glucose (all visits) ‐ FSH (baseline and 13 weeks)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Patients were randomly assigned, in a 2:2:1 ratio, to receive 5 mg of ulipristal acetate per day, 10 mg of ulipristal acetate per day, or placebo. Randomization was stratified according to the hematocrit level at screening” From protocol: "a randomization list will be generated by a designated statistician from MDSL to be transmitted to the assigned clinical packaging organization for labelling"
Allocation concealment (selection bias)	Low risk	“The investigator assigned patients to a study group with the use of a Web‐integrated interactive voice‐response system”
Blinding of participants and personnel (performance bias) All outcomes	Low risk	“Study materials and medication packaging were identical for all three groups”
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participant outcomes were blinded “Change in total fibroid volume as assessed by MRI…by a radiologist who was unaware of study group assignments” “Endometrial biopsy samples were assessed by three independent pathologists who were unaware of the study‐group assignments, the visit sequence” “Data were collected by an independent contract research organization and were handled and analyzed by an independent data management organization”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for drop‐out are given. See Figure 1. "Intention to treat" analysis performed. Missing data have been imputed using appropriate methods: "In general, missing values were imputed for the statistical analyses with the use of the last available post‐baseline value up to the time point of interest. We performed sensitivity analysis that included four patients (all in the 10mg UPA group), who did not have any efficacy data while receiving treatment, using baseline data carried forward"
Selective reporting (reporting bias)	Low risk	Protocol available on clinicaltrials.gov NCT00755755
Other bias	Low risk	None identified