Résumé : Questions: When added to decongestive lymphatic therapy (DLT), what is the effect of fluoroscopy-guided manual lymphatic drainage (MLD) versus traditional MLD or placebo MLD for the treatment of breast cancer-related lymphoedema (BCRL)? Design: Multicentre, three-arm, randomised controlled trial with concealed allocation, intention-to-treat analysis and blinding of assessors and participants. Participants: At five hospitals in Belgium, 194 participants with unilateral chronic BCRL were recruited. Intervention: All participants received standard DLT (education, skin care, compression therapy and exercises). Participants were randomised to also receive fluoroscopy-guided MLD (n = 65), traditional MLD (n = 64) or placebo MLD (n = 65). Participants received 14 sessions of physiotherapy during the 3-week intensive phase and 17 sessions during the 6-month maintenance phase. Participants performed self-management on the other days. Outcome measures: All outcomes were measured: at baseline; after the intensive phase; after 1, 3 and 6 months of maintenance phase; and after 6 months of follow-up. The primary outcomes were reduction in excess volume of the arm/hand and accumulation of excess volume at the shoulder/trunk, with the end of the intensive phase as the primary endpoint. Secondary outcomes included daily functioning, quality of life, erysipelas and satisfaction. Results: Excess lymphoedema volume decreased after 3 weeks of intensive treatment in each group: 5.3 percentage points of percent excessive volume (representing a relative reduction of 23.3%) in the fluoroscopy-guided MLD group, 5.2% (relative reduction 20.9%) in the traditional MLD group and 5.4% (relative reduction 24.8%) in the placebo MLD group. The effect of fluoroscopy-guided MLD was very similar to traditional MLD (between-group difference 0.0 percentage points, 95% CI –2.0 to 2.1) and placebo MLD (–0.2 percentage points, 95% CI –2.1 to 1.8). Fluid accumulated at the shoulder/trunk in all groups. The average accumulation with fluoroscopy-guided MLD was negligibly less than with traditional MLD (–3.6 percentage points, 95% CI –6.4 to –0.8) and placebo MLD (–2.4 percentage points, 95% CI –5.2 to 0.4). The secondary outcomes also showed no clinically important between-group differences. Conclusion: In patients with chronic BCRL, MLD did not provide clinically important additional benefit when added to other components of DLT. Registration: NCT02609724.