Favrat 2014.

Methods	RCT; parallel‐group (phase III). Undertaken in 21 centres across Europe
Participants	Premenopausal, menstruating women > 18 years of age with symptomatic fatigue but normal or borderline Hb (> 115 g/L) at screening. Iron deficiency was defined as ferritin < 50 mcg/L with TSAT < 20%, or ferritin < 15 mcg/L.
Interventions	Single dose of intervention (ferric carboxymaltose, 1000 mg) or an equivalent volume of placebo (250 mL 0.9% sodium chloride)
Outcomes	Primary endpoint was proportion of women achieving ≥ 1 improvement in Piper Fatigue Score at end of follow‐up (day 56). Secondary endpoints included changes in Piper Fatigue Score, SF‐12 and computerised cognitive tests. Changes in Hb, ferritin and TSAT were also recorded.
Study funding arrangements	Vifor Pharma Pty Ltd sponsored this study and supported the study design. Funding was provided for a clinical research organisation, statistical analysis and manuscript preparation.
Author conflicts of interest	The study authors declare multiple conflicts of interest, having accepted honoraria from Vifor Pharma Pty Ltd and other organisations.
Sample size	290 participants were randomised to receive the intervention (n = 144) or the control (n = 146) treatment.
Notes	Study registered in a public trials register (NCT 01110356)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Treatment allocation followed a computer‐generated list of random numbers that has been prepared by the clinical research organization using block randomisation with variable block length."
Allocation concealment (selection bias)	Low risk	Comment: on the basis of the randomisation method described, there is little interaction between the clinical research organisation and the researchers involved in the screening and enrolment of patients prior to randomisation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "investigators received a set of sealed envelopes that corresponded to a randomisation number and contained the identity of the study drug, and prepared and administered the study drug."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: no comment is made on the blinding of outcome assessors, and the risk of bias is therefore unclear. In particular, there is no specific separation mentioned on separating investigators responsible for preparation of study drug and those responsible for assessing outcome.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: consequences of protocol analysis not reported. Different metrics used when reporting haematinic data
Selective reporting (reporting bias)	High risk	Quote: "The most common major protocol deviations were ‘disallowed concurrent medications’ (15 patients in the placebo and 19 patients in the FCM group) and ‘selection criteria not met’ (5 patients in the placebo and 8 patients in the FCM group)." Comment: whilst a per protocol analysis was performed, this was not reported. In addition, haematinic data were reported differently in the participating characteristics and outcomes assessment, precluding extraction. Poor reporting of haematinic follow‐up data. First study author contacted for access to raw data ‐ no response. Data reported as mean (IQR) at baseline, and as mean (SD), or mean (range) at day 56. Poor reporting of haematinic follow‐up data
Other bias	High risk	Comment: study author (AM) an employee of supporting pharmaceutical company