This study investigates the impact of breathing motion on proton breast treatment plans. Twelve patients with CT datasets acquired during breath-hold-at-inhalation (BHI), breath-hold-at-exhalation (BHE) and in free-breathing (FB) were included in the study. Proton plans were designed for the left breast for BHI and subsequently recalculated for BHE or designed for FB and recalculated for the extreme breath-hold phases. The plans were compared from the point of view of their target coverage and doses to organs-at-risk. The median amplitude of breathing motion determined from the positions of the sternum was 4.7 mm (range 0.5-14.6 mm). Breathing motion led to a degradation of the dose coverage of the target (heterogeneity index increased from 4-7% to 8-11%), but the degraded values of the dosimetric parameters of interest fulfilled the clinical criteria for plan acceptance. Exhalation decreased the lung burden [average dose 3.1-4.5 Gy (RBE)], while inhalation increased it [average dose 5.8-6.8 Gy (RBE)]. The individual values depended on the field arrangement. Smaller differences were seen for the heart [average dose 0.1-0.2 Gy (RBE)] and the LAD [1.9-4.6 Gy (RBE)]. Weak correlations were generally found between changes in dosimetric parameters and respiratory motion. The differences between dosimetric parameters for various breathing phases were small and their expected clinical impact is consequently quite small. The results indicated that the dosimetric parameters of the plans corresponding to the extreme breathing phases are little affected by breathing motion, thus suggesting that this motion might have little impact for the chosen beam orientations with scanned proton beams.