Background: We evaluate tranexamic acid use in high-risk shoulder arthroplasty patients, hypothesizing that tranexamic acid will decrease transfusion rates in patients with low preoperative haematocrit. Methods: A retrospective review of shoulder arthroplasty patients with preoperative anaemia compared those treated with and without perioperative tranexamic acid. Inclusion criterion was any shoulder arthroplasty with a preoperative haematocrit (Hct) <38%. Tranexamic acid was given generally as 1 g intravenously before incision and 1 g during wound closure; topical tranexamic acid was used in cases of contraindication to IV tranexamic acid. Preoperative Hct, postoperative Hct, estimated blood loss, preoperative anticoagulation use, American Society of Anesthesiologist score, transfusion rate and demographic data were compared between groups with and without tranexamic acid. Results: Of the 435 arthroplasties performed, 109 patients had preoperative Hct < 38% (haemoglobin of 12.7 g/dL); 69 had tranexamic acid perioperatively and 40 did not. Demographics, anticoagulation use, diagnosis, arthroplasty types, estimated blood loss and preoperative Hct were similar between groups. The tranexamic acid group had a higher postoperative Hct, significantly lower transfusion rate and significantly smaller drop in Hct. Hct levels, Hct change and transfusion rate for topical versus intravenous tranexamic acid were not significantly different. Discussion: Perioperative tranexamic acid in high-risk shoulder arthroplasty patients with preoperative Hct <38% is associated with higher postoperative Hct and lower transfusion rates.