Hypothesis The purposes of this study were to determine the incidence of blood transfusion after revision shoulder arthroplasty and to assess risk factors associated with an increased risk of transfusion. Materials and methods Between 1994 and 2008, 566 consecutive revision shoulder procedures were performed at our institution, which formed the basis of this study. The patient's age, sex, body mass index, comorbidities, preoperative and postoperative hemoglobin level, details of the surgery, operative time, and transfusion details were documented retrospectively from medical records. Results Overall, 11.3% of patients (64 of 566) required a transfusion. An increased transfusion rate was associated with age (odds ratio OR per 10 years, 1.5 95% confidence interval (CI), 1.2 to 2.0; P = .002), operative time (≤5 hours vs >5 hours) (OR, 3.3 95% CI, 1.9 to 5.8; P < .001), diabetes (OR, 2.3 95% CI, 1.2 to 4.4; P = .01), and cardiac disease (OR, 2.7 95% CI, 1.5 to 5.0; P < .001). There were significant associations between preoperative hemoglobin level (OR, 0.4 per 1 point 95% CI, 0.3 to 0.5; P < .001) and a decreased odds of transfusion. The type of surgery (surgery on humeral component) also had an impact on the need for transfusion ( P < .001). Conclusions Older age, low preoperative hemoglobin level, increased operative time, diabetes, presence of cardiac disease, and type of revision surgery are associated with higher postoperative transfusion rates. These factors should be taken into consideration to more accurately predict the need for transfusion and modify preoperative blood-ordering protocols.