Background The iron overload disorder hereditary haemochromatosis is most commonly caused by HFE p.Cys282Tyr homozygosity. In the absence of results from any randomised trials, current evidence is insufficient to determine whether individuals with hereditary haemochromatosis and moderately elevated serum ferritin, should undergo iron reduction treatment. This trial aimed to establish whether serum ferritin normalisation in this population improved symptoms and surrogate biomarkers. Methods This study was a multicentre, participant-blinded, randomised controlled trial done at three centres in Australia. We enrolled people who were homozygous for HFE p.Cys282Tyr, aged between 18 and 70 years, with moderately elevated serum ferritin, defined as 300–1000 μg/L, and raised transferrin saturation. Participants were randomly assigned, via a computer-generated random number, to undergo either iron reduction by erythrocytapheresis (treatment group) or sham treatment by plasmapheresis (control group). Randomisation was stratified by baseline serum ferritin (<600 μg/L or ≥600 μg/L), sex, and study site. Erythrocytapheresis and plasmapheresis were done every 3 weeks, the number of procedures and volume of red cells or plasma removed determined on the basis of each patient's haemoglobin, haematocrit, and serum ferritin concentration, as well their height and weight. In the erythrocytapheresis group, the target was to reduce serum ferritin to less than 300 μg/L. The number of procedures for the control group was based on the initial serum ferritin and prediction of decrease in serum ferritin of approximately 120 μg/L per treatment. The primary outcome was patient-reported Modified Fatigue Impact Scale (MFIS) score, measured at baseline and before unblinding. Analyses were by intention to treat, including the safety analysis. The trial is registered with ClinicalTrials.gov, number NCT01631708, and has been completed. Findings Between Aug 15, 2012, and June 9, 2016, 104 participants were randomly assigned to the treatment (n=54) and control (n=50) groups, of whom 94 completed the study (50 in the treatment group and 44 in the control group). Improvement in MFIS score was greater in the treatment group than in the control group (mean difference −6·3, 95% CI −11·1 to −1·4, p=0·013). There was a significant difference in the cognitive subcomponent (–3·6, −5·9 to −1·3, p=0·0030), but not in the physical (–1·90 −4·5 to 0·63, p=0·14) and psychosocial (–0·54, −1·2 to 0·11, p=0·10) subcomponents. No serious adverse events occurred in either group. One participant in the control group had a vasovagal event and 17 participants (14 in the treatment group and three in the control group) had transient symptoms assessed as related to hypovolaemia. Mild citrate reactions were more common in the treatment group (32 events [25%] in 129 procedures) compared with the control group (one event [1%] in 93 procedures). Interpretation To our knowledge, this study is the first to objectively assess the consequences of iron removal in individuals with hereditary haemochromatosis and moderately elevated serum ferritin. Our results suggest that serum ferritin normalisation by iron depletion could be of benefit for all individuals with hereditary haemochromatosis and elevated serum ferritin levels. Funding National Health and Medical Research Council (Australia).