Low Bone Density Tracks With Higher Postoperative Coagulation Signals After UKA
A retrospective study of 97 unicompartmental knee arthroplasty cases found that low bone density was associated with higher postoperative coagulation markers, including D-dimer on postoperative day 3 and a higher Wells score. The signal was strongest when low bone density was defined using the lowest T-score from the lumbar spine or femoral neck, and it weakened when the femoral neck alone was used to classify bone status.
The finding is biologically plausible. Lower skeletal mineral density often travels with older age, frailty, reduced mobility, and a more inflammatory perioperative milieu, all of which can tilt the balance toward venous stasis and coagulation activation after knee surgery. In this cohort, those factors may have converged on laboratory and clinical risk surrogates even though no duplex ultrasonography endpoint was available to confirm thrombosis.
Surrogate Markers Rose in Step With Bone Loss
Dual-energy X-ray absorptiometry divided patients into normal bone density, osteopenia, and osteoporosis. Low bone density was common: 75 of 97 patients, or 77.3%, fell into either osteopenia or osteoporosis. Serum D-dimer was measured before surgery and again on postoperative day 3, and the Wells score was tracked as a clinical estimate of thrombotic risk.
The primary multivariable Gamma regression found a significant association between low bone density and postoperative D-dimer, with an adjusted mean ratio of 1.39. After additional adjustment for age and sex, the estimate remained similar at 1.40. Those results were not isolated one-off findings; bootstrap resampling and inverse probability of treatment weighting both pointed in the same direction, suggesting the pattern was not simply an artifact of one statistical specification.
A single sentence captures the central clinical tension: the signal was present, but the endpoint remained indirect.
Wells Scores Shifted Toward Higher Risk in the Osteoporosis Group
The clinical scoring data moved in parallel with the laboratory results. The Kruskal-Wallis test showed significant differences in Wells score distribution across the bone-density strata, and pairwise testing after Holm-Bonferroni correction identified a significant difference between the osteoporosis and osteopenia groups.
The proportion of patients meeting the study’s composite DVT risk proxy was highest in the osteoporosis group, at 47.4%, compared with 25.0% in osteopenia and 13.6% in the normal bone density group. In a postoperative orthopedic setting, that pattern matters because D-dimer elevations are common after joint surgery, yet persistent elevation combined with a higher Wells score can sharpen suspicion when imaging access is limited or delayed. The study did not include imaging confirmation, so the composite endpoint functions only as a practical surrogate rather than proof of thrombosis.
Statistical Strength Improved, Then Frayed Under Reclassification
The analysis used several methods to test whether the association would hold under different assumptions. Investigators applied multivariable Gamma regression with log link, bootstrap resampling, permutation testing, a female-only sensitivity analysis, ordinary least squares on log-transformed D-dimer with heteroscedasticity-consistent errors, quantile regression, and propensity score inverse probability of treatment weighting. That breadth strengthens confidence in the initial pattern, but it also exposes the fragility of the result when bone density is defined differently.
When bone density was reclassified using the femoral neck T-score alone, the association with postoperative D-dimer fell away and no longer reached significance. Median quantile regression, logistic regression for Wells score, and the adjusted composite proxy analysis also did not achieve significance. This is the critical fault line in the paper: the association depends on how skeletal status is measured, which means the apparent signal may reflect a broader systemic vulnerability captured by the lowest T-score rather than a femoral neck-specific phenomenon.
Bone loss at the spine and hip may not carry identical perioperative meaning. Lumbar spine values can be influenced by degenerative change, while femoral neck assessment may better reflect cortical bone loss and overall skeletal fragility in some patients. The fact that the signal attenuated under one definition suggests the relationship is not yet stable enough to guide practice on its own.
What the Data Suggest About Thrombotic Risk After UKA
Unicompartmental knee arthroplasty is often viewed as less invasive than total knee arthroplasty, but the procedure still occurs in a postoperative window where inflammation, reduced ambulation, and transient hypercoagulability are expected. In that context, a higher postoperative D-dimer in patients with low bone density could reflect more than clot turnover alone. It may also mark an older, less mobile, more physiologically vulnerable subgroup in whom thrombotic risk deserves closer attention.
The study does not prove that osteoporosis or osteopenia causes venous thromboembolism after UKA. It does, however, raise a testable question: whether preoperative bone-density status can help identify patients who warrant more careful surveillance, more aggressive mobilization, or more selective use of diagnostic imaging when postoperative symptoms appear.
Because all outcomes were surrogate markers, the results should not be converted into a new clinical rule. D-dimer rises after orthopedic surgery for many reasons, and Wells scoring was never designed to replace imaging in this setting. Without duplex ultrasonography data, the study cannot distinguish biochemical noise from true thrombotic events.
A Hypothesis Worth Testing Prospectively
The most defensible reading is narrow. Low bone density, when defined by the lowest T-score from the lumbar spine or femoral neck, tracked with higher postoperative D-dimer and higher Wells scores after unicompartmental knee arthroplasty, but the association weakened under alternative classification and several supporting analyses fell short of significance.
This is an exploratory signal, not a practice-changing directive. A prospective study with imaging-confirmed venous thromboembolism would be needed to determine whether bone density belongs in perioperative thrombosis risk assessment for UKA, or whether it is simply a marker of overall patient frailty that happens to travel with higher coagulation markers.