Cementless Knee Implants Cut Poor Outcomes Without Raising Average Scores

A 3D-printed, fully titanium cementless knee replacement did not outperform its identical cemented twin on average joint awareness scores, but it sharply reduced the share of patients left with the poorest results, a prospective two-cohort study has found. In 109 patients followed for two years, 96.3 percent of those given the cementless implant reached the threshold for an acceptable symptom state, compared with 76.4 percent of those given the cemented version, an absolute difference of 19.9 percentage points. The finding, published in Archives of Orthopaedic and Trauma Surgery, suggests that the value of modern cementless fixation may lie not in lifting average outcomes but in raising the floor for the patients who struggle most.

The study, led by Luca De Berardinis and Andrea Baldini, set out to test a hypothesis that earlier comparisons had never examined. Most trials comparing cemented and cementless total knee arthroplasty, or TKA, have analysed mean clinical scores and generally reported equivalence between the two fixation methods. The researchers suspected that averaging might be hiding something. If biological osseointegration at the bone-implant interface improves load transfer, they reasoned, its benefit should be greatest in patients for whom that interface limits joint perception, and negligible in those who already achieve near-complete joint unawareness, where a ceiling effect would be expected.

To test this, the team compared two versions of the same implant, the GKS Prime Flex CR device made by Permedica. Both share an identical articular geometry and a titanium-niobium nitride articulating surface coating, so the only meaningful difference is fixation. The cemented femoral component is made of cobalt-chromium alloy and fixed with simplex bone cement. The cementless version is fully 3D-printed using selective laser melting of Ti6Al4V titanium alloy, producing a highly porous bone-facing surface with 70 percent porosity and randomly irregular 520-micrometre pores engineered to maximize osseointegration. The tibial components in both versions are titanium, but the cementless one is printed with a short keel, four bullet-shaped porous pegs, and a modular polished stem.

The design was a prospective, non-randomised two-cohort study of consecutive primary unilateral TKAs performed at a single institution between January 2021 and July 2022. All 109 patients had end-stage bone-on-bone knee osteoarthritis and complete 24-month follow-up, with none lost. Patients with valgus deformity or bilateral surgery were excluded. Allocation was determined by surgeon preference rather than randomisation: the cementless design was preferentially selected for younger patients, while the cemented version was preferred for older patients. As a result, the cementless group was younger, averaging 66.5 years against 74.6 years in the cemented group, contained more men, 48 percent versus 24 percent, and had a lower body mass index.

The primary outcome was the Forgotten Joint Score-12, or FJS-12, a patient-reported measure of how aware patients are of their replaced joint during daily life, with higher scores meaning the joint is more forgotten. The researchers analysed it in two ways: at the mean, using multivariable linear regression adjusted for age, sex, BMI, and preoperative Oxford Knee Score, and across its full distribution, using quantile regression at the 10th, 25th, 50th, 75th, and 90th percentiles. The distributional analysis was pre-specified but exploratory, and p-values were corrected for multiple comparisons using the Benjamini-Hochberg method. A propensity-score overlap weighting analysis served as a sensitivity check on the mean comparison.

At the mean, the two fixation methods were statistically indistinguishable. Cementless fixation was not a significant predictor of FJS-12 at 24 months, with an adjusted difference of 7.4 points favouring cementless, but a 95 percent confidence interval running from minus 1.8 to 16.6 and a p-value of 0.115. Overlap weighting, which achieved exact balance of all covariates, produced a closely concordant estimate of 7.3 points, indicating that the demographic imbalance between the groups did not account for the result. The authors caution, however, that the confidence interval did not exclude a clinically relevant difference. With 55 and 54 patients per group, the study could detect a difference of roughly 10.7 FJS-12 points at 80 percent power, so it was able to exclude large mean differences but was not powered to establish formal equivalence at the minimal clinically important difference.

The distributional analysis told a different story. Cementless patients scored substantially higher at the lower end of the FJS-12 range, by 15.4 points at the 10th percentile and 17.4 points at the 25th percentile, with both differences remaining significant after correction, at adjusted p-values of 0.028 and 0.021. No significant difference was detected at the median or at the upper percentiles. In other words, the advantage of cementless fixation was concentrated among the patients at greatest risk of poor joint awareness, while those who already achieved high scores did so regardless of which implant they received. The authors note that applying quantile regression to this comparison is, to their knowledge, novel in the cemented versus cementless literature.

The secondary outcomes reinforced the pattern. A greater proportion of cementless patients achieved the Patient Acceptable Symptom State for the FJS-12, defined as a score of at least 33.3 according to Singh and colleagues: 96.3 percent versus 76.4 percent, an absolute difference of 19.9 percentage points with a 95 percent confidence interval of 7.1 to 32.9, corresponding to a number needed to treat of 5. The difference persisted after covariate adjustment using Firth penalized logistic regression, with a p-value of 0.005, and age itself was not significantly associated with PASS achievement, indicating that the younger profile of the cementless group did not explain the result. By contrast, the Oxford Knee Score at 24 months showed no significant difference between groups, and time to functional satisfaction, which appeared slower in the cementless group on unadjusted comparison, was not associated with fixation after adjustment, suggesting that the apparent delay reflected the demographic profile of the cementless group rather than the implant itself.

Safety outcomes were reassuring in both arms. Complications occurred in four cementless patients, 7.4 percent, including one suture-material reaction, one manipulation under anaesthesia, one joint effusion, and one lower-limb oedema, and in seven cemented patients, 12.7 percent, including two manipulations under anaesthesia, one further case of stiffness managed conservatively, two wound-related complications, one lower-limb oedema, and one reactive cutaneous eruption. All complications resolved without implant failure, revision, or loosening, and no revisions or loosening occurred during the 24-month follow-up. Radiographic assessment was performed at each visit to evaluate implant positioning, radiolucent lines, and signs of loosening, though the authors note that radiographs were obtained as a routine safety assessment rather than as a scored study endpoint.

The authors are careful about what their findings can and cannot support. This study alone will not change clinical practice, they write, but it generated a testable hypothesis: that the benefit of modern cementless TKA may lie not in improving average outcomes, but in raising the floor of patient experience. Confirmation through adequately powered randomized trials or large registry analyses is warranted before these findings can inform clinical decision-making. The limitations are substantial: allocation by surgeon preference leaves open the possibility of residual confounding from unmeasured determinants of joint awareness such as activity level, mental health, comorbidity, bone quality, and expectation, and overlap weighting balances only measured covariates. The open-label design left patient-reported endpoints susceptible to an expectation effect, though patients were unaware of which implant they had received. The quantile regression was exploratory, with extreme-percentile estimates based on few patients, and the single-surgeon, single-centre design aids internal consistency but limits generalizability. One author discloses royalties from the manufacturer, though the authors note that both implants come from the same company and differ only in fixation. Within those constraints, the message is striking: the cementless implant did not prove to be a better implant on average, but a more reliable one, producing fewer of the subjective failures that leave patients persistently feeling their prosthesis.

Subject of Research: Comparison of cemented and cementless total knee arthroplasty outcomes at 24 months

Article Title: Fewer poor joint awareness results with cementless TKA compared with the same-design cemented implant

Article References: De Berardinis, L., Leggieri, F., Salari, P., Polizzotti, G., & Baldini, A. (2026). Fewer poor joint awareness results with cementless TKA compared with the same-design cemented implant. Archives of Orthopaedic and Trauma Surgery, 146(1), Article 365. https://doi.org/10.1007/s00402-026-06531-w

Image Credits: AI Generated

DOI: 10.1007/s00402-026-06531-w

Keywords: total knee arthroplasty, cementless fixation, 3D-printed titanium implant, joint awareness, Forgotten Joint Score-12, quantile regression, patient acceptable symptom state, osseointegration, orthopaedic surgery, patient-reported outcomes, Fewer, poor