Data-driven preoperative hemoglobin thresholds in patients undergoing Bilateral simultaneous total Knee arthroplasty are similar to patients undergoing unilateral total Knee arthroplasty

Document Type

Journal Article

Publication Date



The Knee






Anemia; Arthroplasty; Bilateral Knee; Hemoglobin


INTRODUCTION: Preoperative anemia is a risk factor for transfusions and complications following total knee arthroplasty (TKA). Bilateral TKA (bTKA) is more extensive compared to unilateral TKA (uTKA) and a higher preoperative hemoglobin threshold may be warranted. We hypothesized that the optimal hemoglobin cutoff value which predicts the need for postoperative transfusion would be higher for bTKA than uTKA. METHODS: We conducted a case control study using a national database and identified patients undergoing primary TKA from 2010-2020. 1:1, nearest-neighbor propensity-score matching was used to create a cohort of patients undergoing uTKA matched with patients undergoing bTKA based on age, gender, Charlston Comorbidity Index (CCI), and American Society of Anesthesiology (ASA) classification. After 2015, NSQIP discontinued collection of the variables MI, angina, and hemiplegia. Thus, the accuracy of CCI, which was used as a matching variable, will be less accurate after 2015. To explore this limitation further, a sensitivity analysis was performed excluding data after 2015 and there was no change in the significance of our primary outcomes. Hemoglobin thresholds which maximally predict postoperative transfusion risk and 30-day complications were identified using Youden's index. Significance was considered if 95% CI's were non-overlapping. RESULTS: 9,891 patients were included in each of the bTKA and uTKA cohorts with successful 1:1 matching (p > 0.05 for all criteria). 3.216 (16 %) of patients received a transfusion in the postoperative period. Hemoglobin values which predict postoperative transfusions were not significantly different between uTKA and bTKA for both female and male groups. In females, the preoperative hemoglobin threshold was 12.8 g/dL (95 % CI: 12.2-13.3) in patients undergoing bTKA and 12.7 g/dL (95 % CI: 12.2 - 13.2) in uTKA. In males, the threshold was 13.9 (95 % CI: 13.7-14.2) in patients undergoing bTKA and 13.1 g/dL (95 % CI: 12.5-13.8) in patients undergoing uTKA. CONCLUSIONS: Preoperative hemoglobin values which maximally predict postoperative transfusion risk following uTKA and bTKA are similar without significant differences.


School of Medicine and Health Sciences Resident Works