Background 

The ability to identify which patients are at a greater risk of early revision knee arthroplasty has important practical and resource implications. Many international arthroplasty registries administer patient-reported outcome measures (PROMs) to provide a holistic assessment of pain, function, and quality of life. However, few PROM scores have been evaluated as potential indicators of early revision knee arthroplasty, and earlier studies have largely focused on knee-specific measures.

Questions/purposes 

This national registry-based study asked: (1) Which 6-month postoperative knee-specific and generic PROM scores are associated with early revision knee arthroplasty (defined as revision surgery performed 6 to 24 months after the primary procedure)? (2) Is a clinically important improvement in PROM scores (based on thresholds for the minimal important change) after primary knee arthroplasty associated with a lower risk of early revision?

Methods 

Preoperative and 6-month postoperative PROM scores for patients undergoing primary knee arthroplasty were sourced from the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) and Arthroplasty Clinical Outcomes Registry National. Between January 2013 and December 2020, PROM data were available for 19,402 primary total knee arthroplasties; these data were linked to AOANJRR data on revision knee arthroplasty. Of these, 3448 procedures were excluded because they did not have 6-month PROM data, they had not reached the 6-month postoperative point, they had died before 24 months, or they had received revision knee arthroplasty before the 6-month PROMs assessment. After these exclusions, data were analyzed for 15,954 primary knee arthroplasties. Associations between knee-specific (knee pain, Oxford Knee Score, and 12-item Knee injury and Osteoarthritis Outcome Score [KOOS-12]) or generic PROM scores (5-level EuroQol quality of life instrument [EQ-5D], EQ VAS, perceived change, and satisfaction) and revision surgery were explored using t-tests, chi-square tests, and regression models. Ninety-four revision procedures were performed at 6 to 24 months, most commonly for infection (39% [37 procedures]). The early revision group was younger than the unrevised group (mean age 64 years versus 68 years) and a between-group difference in American Society of Anesthesiologists (ASA) grade was noted. Apart from a small difference in preoperative low back pain for the early revision group (mean low back pain VAS 4.2 points for the early revision group versus 3.3 points for the unrevised group), there were no between-group differences in preoperative knee-specific or generic PROM scores on univariate analysis. As the inclusion of ASA grade or low back pain score did not alter the model results, the final multivariable model included only the most clinically plausible confounders (age and gender) as covariates. Multivariable models (adjusting for age and gender) were also used to examine the association between a clinically important improvement in PROM scores (based on published thresholds for minimal important change) and the likelihood of early revision.

Results 

After adjusting for age and gender, poor postoperative knee pain, Oxford, KOOS-12, EQ-5D, and EQ VAS scores were all associated with early revision. A one-unit increase (worsening) in knee pain at 6 months was associated with a 31% increase in the likelihood of revision (RR 1.31 [95% confidence interval (CI) 1.19 to 1.43]; p < 0.001). Reflecting the reversed scoring direction, a one-unit increase (improvement) in Oxford or KOOS-12 score was associated with a 9% and 5% reduction in revision risk, respectively (RR for Oxford: 0.91 [95% CI 0.90 to 0.93]; p < 0.001; RR for KOOS-12 summary: 0.95 [95% CI 0.94 to 0.97]; p < 0.001). Patient dissatisfaction (RR 6.8 [95% CI 3.7 to 12.3]) and patient-perceived worsening (RR 11.7 [95% CI 7.4 to 18.5]) at 6 months were also associated with an increased likelihood of early revision. After adjusting for age and gender, patients who did not achieve a clinically important improvement in PROM scores had a higher risk of early revision (RR 2.9 for the knee pain VAS, RR 4.2 for the Oxford Knee Score, RR 6.3 to 8.6 for KOOS-12, and RR 2.3 for EQ-5D) compared with those who did (reference group).

Conclusion 

Knee-specific and generic PROM scores offer an efficient approach to identifying patients at greater risk of early revision surgery, using either the 6-month score or the magnitude of improvement. These data indicate that surgeons can use single- and multi-item measures to detect a patient-perceived unsuccessful surgical outcome at 6 months after primary knee arthroplasty. Surgeons should be alert to poor PROM scores at 6 months or small improvements in scores (for example, less than 2 points for knee pain VAS or less than 10.5 points for Oxford Knee Score), which signal a need for direct patient follow-up or expedited clinical review.

Level of Evidence 

Level III, therapeutic study.

中文翻译：

---

膝关节特异性和一般患者报告的 6 个月结果测量分数较差与早期膝关节置换术修复相关：来自澳大利亚骨科协会国家关节置换登记处的一项研究

背景 

确定哪些患者接受早期膝关节置换术的风险更大的能力具有重要的实践和资源意义。许多国际关节置换术登记处都采用患者报告结果测量 (PROM)，以提供对疼痛、功能和生活质量的全面评估。然而，很少有 PROM 评分被评估为早期膝关节置换术的潜在指标，早期的研究主要集中在针对膝关节的具体测量上。

问题/目的 

这项基于国家登记的研究询问：(1) 哪些术后 6 个月膝关节特异性和通用 PROM 评分与早期膝关节置换术（定义为初次手术后 6 至 24 个月进行的翻修手术）相关？(2) 初次膝关节置换术后 PROM 评分（基于最小重要变化的阈值）的临床重要改善是否与较低的早期翻修风险相关？

方法 

接受初次膝关节置换术的患者术前和术后 6 个月的 PROM 评分来源于澳大利亚骨科协会国家关节置换登记处 (AOANJRR) 和国家关节置换术临床结果登记处。2013 年 1 月至 2020 年 12 月期间，可获得 19,402 例初次全膝关节置换术的 PROM 数据；这些数据与膝关节翻修术的 AOANJRR 数据相关。其中，3448 例手术被排除，因为他们没有 6 个月 PROM 数据、未达到术后 6 个月点、在 24 个月前死亡，或者在 6 个月 PROM 评估之前接受了膝关节翻修术。在这些排除之后，对 15,954 例初次膝关节置换术的数据进行了分析。膝关节特异性（膝关节疼痛、牛津膝关节评分、12 项膝关节损伤和骨关节炎结果评分 [KOOS-12]）或通用 PROM 评分（5 级 EuroQol 生活质量工具 [EQ-5D]、EQ VAS 之间的关联使用 t 检验、卡方检验和回归模型对修复手术进行了探索。在 6 至 24 个月内进行了 94 次翻修手术，最常见的是感染（39% [37 次手术]）。早期修复组比未修复组年轻（平均年龄 64 岁 vs 68 岁），并且注意到美国麻醉医师协会 (ASA) 等级的组间差异。除了早期翻修组术前腰痛有微小差异（早期翻修组平均腰痛 VAS 为 4.2 分，未翻修组为 3.3 分）外，术前膝关节特异性或单变量分析的通用 PROM 分数。由于纳入 ASA 分级或腰痛评分不会改变模型结果，因此最终的多变量模型仅包含临床上最合理的混杂因素（年龄和性别）作为协变量。多变量模型（根据年龄和性别进行调整）也用于检查 PROM 评分的临床重要改善（基于已发布的最小重要变化阈值）与早期修正的可能性之间的关联。

结果 

调整年龄和性别后，术后膝关节疼痛不佳、Oxford、KOOS-12、EQ-5D 和 EQ VAS 评分均与早期翻修相关。6 个月时膝盖疼痛增加 1 个单位（恶化）与翻修可能性增加 31% 相关（RR 1.31 [95% 置信区间 (CI) 1.19 至 1.43]；p < 0.001）。反映反向评分方向的是，Oxford 或 KOOS-12 分数每提高一个单位（提高），修订风险分别降低 9% 和 5%（牛津的 RR：0.91 [95% CI 0.90 至 0.93] ；p < 0.001；KOOS-12 总结的 RR：0.95 [95% CI 0.94 至 0.97]；p < 0.001）。6 个月时患者不满意（RR 6.8 [95% CI 3.7 至 12.3]）和患者感知的病情恶化（RR 11.7 [95% CI 7.4 至 18.5]）也与早期翻修可能性增加相关。调整年龄和性别后，PROM 评分未实现临床重要改善的患者早期修正的风险较高（膝关节疼痛 VAS 的 RR 2.9，牛津膝关节评分的 RR 4.2，KOOS 的 RR 6.3 至 8.6） 12，EQ-5D 的 RR 2.3）与那些这样做的人（参考组）相比。

结论 

膝关节特异性和通用 PROM 评分提供了一种有效的方法，可以使用 6 个月评分或改善幅度来识别早期翻修手术风险较高的患者。这些数据表明，外科医生可以使用单项和多项测量来检测初次膝关节置换术后 6 个月时患者感知的不成功手术结果。外科医生应警惕 6 个月时 PROM 评分不佳或评分小幅改善（例如，膝关节疼痛 VAS 评分低于 2 分或牛津膝关节评分低于 10.5 分），这表明需要对患者进行直接随访或加快临床审查。

证据水平 

III级，治疗研究。