Second-knee transfusion burden persists under a high-dose intravenous iron protocol without tranexamic acid in one-week staged bilateral total knee arthroplasty: a paired cohort study

The 1-week staged interval for bilateral total knee arthroplasty (TKA) is favored for cost-efficiency, but the second knee operates on an incompletely recovered hematologic reserve. Whether high-dose perioperative intravenous iron without tranexamic acid (TXA) eliminates the inter-knee transfusion gap is unknown. We retrospectively analyzed 210 consecutive patients undergoing 1-week staged bilateral primary TKA under a high-dose ferric carboxymaltose protocol (1,500 mg cumulative) without TXA; the transfusion trigger was hemoglobin ≤ 8.0 g/dL or symptomatic anemia. Three hypotheses were pre-specified: that the within-patient second-knee transfusion rate exceeds the first-knee rate; that the lowest tertile of inter-knee hemoglobin recovery has a higher second-knee rate; and that the immediate postoperative hemoglobin after the second operation discriminates second-knee transfusion. The second-knee transfusion rate was nearly twice the first-knee rate within the same patient (59.5% vs. 31.0%); inter-knee hemoglobin recovery did not predict second-knee transfusion. The immediate postoperative hemoglobin after the second operation discriminated transfusion (area under the curve 0.718), comparably to the preoperative value. The exploratory Youden-optimal cutoff of 8.8 g/dL had a prevalence-dependent positive predictive value and a sensitivity of only 0.62, missing 38% of transfused patients. Preoperative hemoglobin and body mass index were independently associated with second-knee transfusion, and no surgical site infection or thromboembolic complication occurred within 30 days. High-dose intravenous iron did not eliminate the second-knee transfusion burden, and inter-knee hemoglobin recovery did not reliably identify high-risk patients. Preoperative hemoglobin was at least as discriminating and is the more practical screen; the 8.8 g/dL cutoff is exploratory — not a decision threshold for use outside this cohort — and requires external validation. Because no TXA was used, the absolute transfusion rates exceed contemporary TXA-based practice and are not directly generalizable; adding TXA is the priority next step.