Knee Replacement for Hemophilia Shows Strong Gains but Higher Transfusion Risk

For people with severe hemophilia, decades of recurrent bleeding into the knees can destroy the joint long before old age arrives. A new retrospective cohort study now offers one of the more careful head-to-head comparisons of what happens when those damaged knees are finally replaced. Researchers followed 22 total knee arthroplasties performed in 20 men with hemophilic arthropathy and compared them with 66 knee replacements in 66 male osteoarthritis patients operated on by the same senior surgeon during the same period, from September 2021 through May 2023. The findings, published in the journal Arthroplasty, paint a nuanced picture: substantial functional recovery in the hemophilia group, blood loss no greater than in ordinary arthritis patients, but a clear excess of transfusions and warning signals on wound healing and readmissions that the authors say demand larger studies to confirm.

The study, conducted at a single center and reported according to STROBE guidelines, was designed to address a persistent gap in the surgical literature. Most published series of knee replacement in hemophilia lack a contemporaneous osteoarthritis comparison group, making it hard to know whether poorer outcomes reflect the underlying bleeding disorder or simply the greater technical difficulty of these operations. Patients with hemophilic arthropathy tend to be much younger, arrive with more severe deformity and stiffness, and more often need constrained implants. Crude comparisons of postoperative scores may therefore capture differences in baseline characteristics and procedural complexity rather than the true effect of the diagnosis itself.

The age gap between the groups was striking. Men undergoing surgery for hemophilic arthropathy averaged 41.86 years of age, compared with 68.33 years in the osteoarthritis group, a difference of more than 26 years. This reflects the distinct natural histories of the two conditions: hemophilia, an X-linked recessive bleeding disorder caused by deficiency of coagulation factor VIII or IX, sends blood repeatedly into joints, with approximately 80 percent of cases involving joint bleeding. Recurrent hemarthrosis drives synovial hypertrophy, cartilage loss, and progressive joint destruction, so end-stage disease arrives decades early. The hemophilia cohort included 11 patients with hemophilia A and 9 with hemophilia B, and higher-constraint or stemmed implant constructs were used in 45.5 percent of their procedures versus just 1.5 percent of osteoarthritis procedures, underscoring how much more complex these reconstructions are.

Perioperative care in the hemophilia group followed a comprehensive, WFH-informed institutional protocol managed jointly by orthopedic and hematology teams. Patients with hemophilia A received recombinant factor VIII, while those with hemophilia B received high-purity factor IX concentrate, with target factor activity of 80 to 100 IU/dL before incision and intraoperatively, stepping down to 30 to 50 IU/dL by days 7 through 14. All patients underwent preoperative Bethesda inhibitor testing, and none was classified as inhibitor-positive. Both groups received tranexamic acid, one gram intravenously half an hour before incision and two grams injected into the joint before closure, and both followed the same standardized rehabilitation program, delivered twice daily, five days per week. One key difference: the hemophilia patients received factor replacement before each physiotherapy session to maintain factor levels of at least 40 percent, and they received mechanical prophylaxis with graduated compression stockings without routine pharmacologic thromboprophylaxis, unlike the osteoarthritis patients who received daily enoxaparin.

The primary functional outcome was the Knee Society clinical score at 12 months, analyzed with adjustment for its preoperative value, age, and body mass index, with standard errors clustered by patient to account for two staged bilateral procedures. Both groups improved dramatically. The mean increase in the Knee Society clinical score was 49.64 points in the hemophilia group and 53.24 points in the osteoarthritis group, an unadjusted difference of minus 3.61 points that was not statistically significant. After adjustment, the 12-month group effects were 0.54 points for the clinical score and 3.02 points for the Knee Society function score, neither remotely close to significance. In plain terms, once baseline differences were accounted for, the hemophilia patients recovered function about as well as the ordinary arthritis patients, a result the authors describe as substantial within-group improvement under comprehensive blood management.

Not every measure equalized, however. At 12 months, the hemophilia group had lower adjusted knee flexion than the osteoarthritis group, a mean difference of minus 19.64 degrees, with a confidence interval of minus 38.29 to minus 0.99 degrees. The authors attribute this to the greater prevalence of longstanding contracture, fixed deformity, soft-tissue fibrosis, and complex reconstruction in hemophilia knees, and note that additional adjustment for implant constraint did not eliminate the difference. Interestingly, pain during movement was actually lower in the hemophilia group, with an adjusted VAS difference of minus 0.54 points, though the researchers caution this observed outcome should not be read as evidence of greater pain tolerance, since pain reporting may be influenced by preoperative disease experience, activity level, and postoperative range of motion.

On the safety side, the results were more mixed. Calculated blood loss, estimated with the Nadler-Gross method from preoperative and postoperative hematocrit values, averaged 532.53 milliliters in the hemophilia group versus 628.00 milliliters in the osteoarthritis group, a difference that was not statistically significant, and an alternative weight-based calculation yielded consistent findings. Yet perioperative allogeneic red blood cell transfusion occurred in 4 of 22 hemophilia procedures and in none of the 66 osteoarthritis procedures, a crude difference the authors flag as clinically important. The exploratory adjusted analysis, using Firth-penalized logistic regression, produced an odds ratio of 20.73 with a confidence interval stretching from 0.47 to 913.62, a reminder of just how little data four events can support. The authors distinguish the two measures carefully: hematocrit-based estimates are influenced by fluid administration and laboratory timing, while transfusion decisions also reflect symptoms, hemodynamic status, and clinical concern about continued bleeding.

Two other signals warrant attention. Wound-healing complications occurred after 6 of 22 hemophilia procedures, or 27.3 percent, versus 3 of 66 osteoarthritis procedures, or 4.5 percent, an absolute risk difference of 22.7 percent. Unplanned 90-day readmission occurred after 9 of 22 hemophilia procedures, or 40.9 percent, versus 1 of 66, or 1.5 percent, a risk difference of 39.4 percent. But in adjusted models these associations became extremely imprecise, with odds ratios of 2.63 for wound complications and 2.92 for readmission and confidence intervals spanning nearly two orders of magnitude. The authors suggest the crude excess may reflect a combination of bleeding-related concerns, wound surveillance, factor-replacement requirements, operative complexity, and different thresholds for hospital reassessment, rather than a proven independent effect of the diagnosis. Notably, no periprosthetic joint infection occurred in either group, and calf muscular-vein thrombosis rates were similar, detected in 2 of 22 and 6 of 66 procedures respectively.

Satisfaction was high in both cohorts, reported after 90.9 percent of hemophilia procedures and 97.0 percent of osteoarthritis procedures. The study’s limitations are considerable and the authors enumerate them candidly: the retrospective single-center design, the small hemophilia sample of 22 procedures in 20 patients, the pronounced age and implant imbalances that limited covariate overlap, the absence of consistently retained untreated baseline factor activity for severity classification, and functional assessment standardized only to 12 months, with later follow-up used solely for complication surveillance. No prospective sample-size calculation was performed, and analyses of rare complications were explicitly exploratory, with no adjustment for multiple comparisons.

Its central message is that, within a multidisciplinary pathway combining monitored factor replacement, tranexamic acid, standardized rehabilitation, and close hematologic oversight, total knee arthroplasty can deliver meaningful clinical benefit to appropriately selected patients with end-stage hemophilic arthropathy, with functional gains that look much like those achieved in ordinary osteoarthritis. At the same time, the transfusion signal and the unadjusted wound and readmission excesses reinforce that these operations are not routine knee replacements, and the authors call for confirmation in larger, prospectively defined multicenter cohorts before the comparative safety picture can be considered settled.

Subject of Research: Clinical outcomes and perioperative safety of total knee arthroplasty in hemophilic arthropathy compared with osteoarthritis

Article Title: Clinical outcomes and perioperative safety of primary total knee arthroplasty for hemophilic arthropathy versus osteoarthritis: a retrospective cohort study

Article References: Ma, R., Liu, Q., Hu, X., Sun, H., Zhao, J., Tong, J., Su, Z., & Zhu, C. (2026). Clinical outcomes and perioperative safety of primary total knee arthroplasty for hemophilic arthropathy versus osteoarthritis: a retrospective cohort study. Arthroplasty, 8(1), Article 71. https://doi.org/10.1186/s42836-026-00431-5

Image Credits: AI Generated

DOI: 10.1186/s42836-026-00431-5

Keywords: total knee arthroplasty, hemophilic arthropathy, hemophilia, osteoarthritis, knee replacement, blood transfusion, wound complications, Knee Society score, factor replacement, retrospective cohort, tranexamic acid, multidisciplinary care