Keywords total knee arthroplasty - American Joint Replacement Registry - cementless total knee - registry
Although cemented total knee arthroplasty (TKA) remains standard practice, cementless fixation in TKA continues to be an intriguing option. Potential advantages of cementless TKA include bone preservation in the setting of revision surgery, the elimination of third body wear from cement debris, and a shorter operative time.[1 ] Unfortunately, this enthusiasm has been tempered by early generations of cementless TKA designs that were plagued by a high associated risk of failure and poor clinical outcomes.[2 ]
[3 ]
Early generations of cementless TKA failures were attributed to failure of the cementless patellar or tibial components.[4 ]
[5 ]
[6 ] Retrieval analyses of early generation tibial components demonstrated levels of osseointegration of less than 30%.[7 ] Improvement in the design of the patellofemoral articulation, as well as the elimination of thin polyethylene and sharp metal borders, has led to improved results.[8 ] The decreasing popularity of patellar resurfacing as well as the introduction of modern highly porous surfaces has tempered concerns for patellar- or tibial-related failure. These design improvements have led to a relatively rapid increase in the use of cementless TKA in the United States, increasing from 1.9% in 2012 to 20.5% in 2022 in the American Joint Replacement Registry (AJRR).[9 ]
Recent prospective and retrospective investigations utilizing modern cementless TKA have demonstrated no differences in survivorship and lower rates of aseptic loosening in select populations when compared with cemented TKA.[1 ]
[10 ]
[11 ]
[12 ]
[13 ] Unfortunately, these studies are limited in their power to detect revision risk, specific mechanisms of failure including mechanical loosening, and in their generalizability to the U.S. experience. In contrast to these investigations are international registry data that indicate an increased risk of revision for cementless TKA.[14 ]
[15 ]
[16 ] Yet these international registry analyses are limited by a lower utilization of cementless fixation, a lack of availability of modern cementless designs, and other confounding variables that limit their applicability to the U.S. cementless TKA experience. In contrast, the AJRR collects data from over 1,400 sites in the United States on greater than 2 million arthroplasty procedures with a gradually increasing use of cementless TKA with modern tibial and patellar designs. The AJRR provides a unique opportunity to understand the TKA experience in the United States and to evaluate if there is evidence regarding the increased risks of revision with modern cementless TKA. The primary aim of the present study was thus to identify if there are differences in implant survivorship between modern cementless and cemented TKA designs using the AJRR dataset while controlling for potential confounding variables including age, obesity, sex, comorbidity burden, and femoral design.
Materials and Methods
We collaborated with the American Academy of Orthopaedic Surgeons Registry Analytics Institute to perform a retrospective analysis of AJRR data on primary TKA cases in the AJRR database from January 2012 to March 2020. We identified 1,098,792 primary TKAs over the study period and linked cases to U.S. Centers for Medicare & Medicaid Services (CMS) data to supplement outcome events occurring outside of AJRR reporting institutions where available. Catalog numbers from each TKA included were identified and recorded for the classification of cases into target component categories.
To facilitate linkage of cases to CMS data, we included only patients aged ≥65 years. We excluded 315,507 patients with either hybrid (cemented tibial component and cementless femoral component) or reverse hybrid (cementless tibial component and cemented femoral component) fixation or those with missing data on cement fixation. We excluded 2,956 patients with supplemental stemmed tibial or femoral fixation, use of tibial augmentation, or use of highly constrained implants. We only included modern generation tibial and patellar components, defined as those with highly porous metals,[17 ]
[18 ] including those with porous tantalum,[19 ] highly porous titanium,[20 ] and those with 3D printing or electron beam melting.[21 ]
[22 ] We excluded 834 patients who received older generation tibial components, including those with cobalt-chromium sintered beads, titanium fiber metal mesh, cancellous-structured titanium, or titanium plasma spray. This left us with a final cohort of 442,745 cemented TKAs and 19,841 modern cementless TKAs with a minimum of 2-year follow-up ([Fig. 1 ]).
Fig. 1 Flowchart demonstrating exclusion and allocation of primary total knee arthroplasty (TKA) cases into cohorts, cemented versus cementless. CR, cruciate retaining; PS, posterior stabilized.
Each identified procedure in the AJRR was assessed for a linked revision surgery in AJRR or CMS databases. A surgery was labeled as a revision if it met four criteria: surgical date after the index procedure, matched patient ID, matched surgical site, and matched laterality. Codes from the 10th revision of the International Statistical Classification of Diseases and Related Health Problems Procedure Coding System (ICD-10-PCS) were used to infer laterality along with a discrete laterality field in the AJRR database for older cases.
Data collected included patient age, sex, body mass index (BMI), Charlson's comorbidity index (CCI), femoral component design (cruciate retaining [CR], or posterior stabilized [PS]), and patellar resurfacing. Reason for revision was also recorded, and outcomes considered were all-cause revision and revision for aseptic mechanical loosening. Outcomes were captured through March 2022 to align with end of Medicare claims and to allow 2-year minimal potential follow-up.
We summarized the characteristics of the two fixation groups (cemented vs. cementless) using counts and percentages for categorical variables and mean (standard deviation [SD]) and median (quartile 1, quartile 3) for continuous variables. We compared groups using chi-squared test, student's t -test, and Wilcoxon rank-sum test as appropriate. Cumulative incident function (CIF) curves and cause-specific Cox models evaluated the risk of all-cause revision and revision for mechanical loosening, adjusting for BMI, sex, age, CR versus PS femoral design, patellar resurfacing, and CCI, reporting hazard ratios (HRs) and their 95% confidence intervals (CIs). Analyses were conducted using SAS Enterprise Guide v.7.15 (Cary, NC), and statistical significance evaluated at p < 0.05.
Results
Patients with cementless compared with cemented TKA were younger (mean age: 71.9 vs. 73.2 years; p < 0.001) and more likely to be aged 65 to 74 versus 75+ years (74.6 vs. 65.2%, p < 0.001). They were more likely to be male sex (48.8 vs. 39.0%, p < 0.001), and had a lower comorbidity burden with a lower CCI (mean: 2.9 vs. 3.1; p < 0.001). Patients with cementless compared with cemented TKA had a higher mean BMI (31.6 vs. 31.3%, p < 0.001) and were more likely to have a BMI greater than 35 (26.7 vs. 25.3%, p < 0.001). Furthermore, compared with cemented TKAs, cementless TKAs were more likely to have a CR femoral design (81.1 vs. 45.7%, p < 0.001) and less likely to have patellar resurfacing (92.7 vs. 95.0%, p < 0.001; [Table 1 ]).
Table 1
Demographic characteristics for total knee arthroplasty with cementless and cemented fixation
Fixation
Cemented
Cementless
Total
p value
(N = 442,745)
(N = 19,841)
(N = 462,586)
Body mass index (BMI)
Mean (SD)
31.4 (6.3)
31.6 (6.1)
31.4 (6.3)
<0.001
Median [Q1, Q3]
30.7 [27, 35]
30.9 [27.3, 35.2]
30.7 [27, 35]
<0.001
N (N missing)
187,080 (255,665)
12,050 (7,791)
199,130 (263,456)
Categories of BMI
35+
047,344 (25.31%)
003,213 (26.66%)
050,557 (10.93%)
<0.001
<35
139,736 (74.69%)
008,837 (73.34%)
148,573 (32.12%)
Missing
255,665 (57.75%)
007,791 (39.27%)
263,456 (56.95%)
Age (y)
Mean (SD)
73.3 (5.8)
71.9 (5.3)
73.2 (5.8)
<0.001
Median [Q1, Q3]
72.2 [68.5, 77.1]
70.8 [67.7, 75.1]
72.2 [68.4, 77]
<0.001
Categories of age
65–74 y
288,673 (65.2%)
014,802 (74.6%)
303,475 (65.6%)
<0.001
75+ y
154,072 (34.8%)
005,039 (25.4%)
159,111 (34.4%)
Charlson's comorbidity index (age adjusted)
Mean (SD)
3.1 (1.1)
2.9 (1.0)
3.1 (1.1)
<0.001
Median [Q1, Q3]
3 [2, 4]
3 [2, 3]
3 [2, 4]
<0.001
Femoral component design
Cruciate retaining
202,426 (45.7%)
016,087 (81.1%)
218,513 (47.2%)
<0.001
Posterior stabilized
240,319 (54.3%)
003,754 (18.9%)
244,073 (52.8%)
Patella
Resurfaced
402,962 (95.0%)
017,539 (92.7%)
420,501 (90.9%)
<0.001
Unresurfaced
021,440 (5.1%)
001,389 (7.3%)
022,829 (4.9%)
Missing
018,343 (4.1%)
0000913 (4.6%)
019,256 (4.2%)
Sex
Female
270,317 (61.1%)
010,169 (51.3%)
280,486 (60.6%)
<0.001
Male
172,428 (39.0%)
009,672 (48.8%)
182,100 (39.4%)
Abbreviation: SD, standard deviation.
We identified a total of 9,013 all-cause revisions, with 377 noted in the cementless cohort and 8,636 in the cemented group. Of these, we identified 1,242 revisions for mechanical loosening, with 45 noted in the cementless group and 1,197 in the cemented group. Adjusted HRs showed no difference in associated risk for all-cause revision (HR: 1.07; 95% CI: 0.92–1.24; p = 0.382) or revision for mechanical loosening (HR: 1.38; 95% CI: 0.9–2.12; p = 0.14) for cementless versus cemented TKA. The adjusted CIF curve showed no difference in the associated risk of all-cause revision across the study period comparing cementless to cemented TKA ([Fig. 2 ]). The adjusted CIF curve also showed no difference in the associated risk of mechanical loosing across the study period comparing cementless to cemented TKA ([Fig. 3 ]).
Fig. 2 Cumulative incident function (CIF) curve comparing cemented to cementless total knee arthroplasty (TKA) for all-cause revision, adjusting for body mass index (BMI), sex, age, cruciate retaining (CR) versus posterior stabilized (PS) femoral design, patellar resurfacing, and Charlson's comorbidity index (CCI) across the study period.
Fig. 3 Cumulative incident function (CIF) curve comparing cemented to cementless total knee arthroplasty (TKA) for revision for mechanical loosening, adjusting for body mass index (BMI), sex, age, cruciate retaining (CR) versus posterior stabilized (PS) femoral design, patellar resurfacing, and Charlson's comorbidity index (CCI) across the study period.
Discussion
In this analysis of data from the AJRR linked to supplementary CMS data, we found no associated risk of all-cause revision or revision for mechanical loosening when comparing modern cementless to cemented TKA controlling for age, BMI, sex, CCI, and use of the CR design. This lack of associated risk was noted to be unchanged over the 8 years of the study period despite the rapidly increasing use of cementless TKA in the AJRR over that time. In addition, as a function of this increasing adoption, it is important to note that our results are weighed toward more recent short-term data as opposed to being balanced throughout the study period. Thus, although reassuring, as cementless TKA adoption continues to increase, caution should be exercised in extrapolating the results of this study widely across this cohort without more robust long-term survivorship data.
Our study serves as a balancing counterpoint to existing contradictory literature regarding the risk of revision when comparing cementless and cemented TKA. In a single-center prospective trial, Nam et al randomized patients to cementless or cemented fixation for TKA; this investigation was powered to evaluate the primary outcome of the Oxford Knee Score and only evaluated loosening as a secondary outcome measure.[10 ] They found no instances of loosening in either cohort at 2 years, but this investigation was likely underpowered to evaluate the risk of revision and lacked generalizability as a single-center study examining only one implant design. In a separate prospective investigation, Kamath et al evaluated 100 consecutive patients who received cementless TKA and compared this to a control group of 312 concurrent cemented controls.[13 ] They found no failures of aseptic loosening in the cementless group and two in their cemented cohort with an overall rate of aseptic loosening of 0.6%. This investigation also lacked generalizability, as they only included one implant in the investigation. In another retrospective matched case-control study of 400 TKAs comparing cementless versus cemented TKA, Miller et al found a similar incidence of postoperative complications with one mechanical loosening (0.5%) in the cementless group and five in the cemented group (2.5%).[1 ] The overall rate of mechanical loosening noted in these studies is consistent with the findings noted in our investigation of around 0.27% in both cohorts. As the overall incidence of mechanical loosening is relatively rare, these prior investigations were likely underpowered to detect differences between cohorts. In addition, the prior studies included only a single implant in each, thereby limiting their generalizability. Our investigation expands the power and generalizability of these prior results by using the robust AJRR dataset with 19,841 modern cementless TKAs of various manufactures.
This is in contrast to international registry investigations that suggest an increased risk of failure with cementless fixation. Most similar to the United States in terms of the use of cementless fixation for TKA is the Australian experience with approximately 20% use in 2022.[15 ] The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) 2023 annual report notes increased cumulative percent revision with cementless compared with cemented TKA when using a CR femoral component.[15 ] While the AOANJRR specifies the analyses is restricted to “modern prostheses,” only one femoral prosthesis is noted to be excluded from analyses and there is no formal delineation of what compromises modern prostheses. This may explain the discordant findings compared with our study as the registry includes various older generation tibial components including those with cobalt-chromium sintered beads, cancellous-structured titanium, or titanium plasma spray. The National Joint Registry (NJR) of England, Wales, Northern Ireland, the Isle of Man and Guernsey, 20th annual report Kaplan–Meier estimates demonstrate increased cumulative revision risk for cementless compared with the cemented TKA at 3, 5, and 10 years.[16 ] It can be difficult to contextualize the results from the NJR as the overall use of cementless TKA represents only 2.1% of TKAs in 2022. In addition to the low overall use of cementless fixation, the use of older generation tibial components and lack of uniform adjusting for confounding variables may also be a contributing factor to the differences in revision risk found in the current analyses.
Over the course of the study period, there has been a relatively rapid increase in the use of cementless TKA in the United States. Its use has increased from 1.9% in 2012 to 20.5% in 2022 in the AJRR dataset. This is likely driven by several factors: multiple manufactures have recently released newer designs utilizing modern highly porous tibial and patellar components, and prospective investigations have demonstrated equivalent patient-reported outcome measures[10 ] and noted no differences in terms of survivorship and clinical or radiographic outcomes at midterm follow-up[23 ] when comparing modern cementless to cemented TKA. Even with evolving indications and expanded use, it is reassuring that our CIF curves demonstrated no changing risk of either all-cause revision or revision for mechanical loosening over the study period. Cementless fixation can achieve osseointegration and has the potential to improve long-term survivorship by reducing the risk of aseptic loosening. It is important to note that the study period was only 8 years and only required a minimum of 2 years of follow-up; thus, even though cementless TKA use has increased throughout the study period, we cannot yet evaluate any long-term potential survivorship benefits for cementless fixation.
There are several limitations to this study. This is a retrospective study and therefore has the potential for selection bias. Our cementless group was younger, more likely to be male, more likely to have a CR femoral component, and had a lower comorbidity burden. Our comparison groups were different and we do not attempt to claim equivalence. However, we did attempt to control for these factors in our adjusted analyses. Nevertheless, with the selection bias present in this study, we emphasize that our conclusions are limited in scope: with current indications, the use of cementless TKA does not appear to be associated with an increased revision risk. Furthermore, this investigation is potentially limited in generalizability when considering international experiences. Despite this, the AJRR contains data representing over 2 million procedures from more than 1,300 sites in the United States and is the largest orthopaedic registry in the world.[24 ] Many types of practice settings are captured in this dataset. In addition to the size and variety of institutional capture in the AJRR, a recent study found distributions across hospital volume, age, and geography to be proportionally similar between the AJRR and the National Inpatient Sample (NIS) database further supporting the generalizability of AJRR data to the overall United States cohort.[24 ] Additionally, we limited our population to patients aged ≥65 years to allow linkage to CMS data. It is possible that younger patients may have different results, and our results may not be generalizable to a younger patient population. Despite this, we captured over 442,745 cemented TKAs and 19,841 cementless TKAs for our analysis, making this a very robust and encompassing group size.
Conclusion
Advancements in implant design with the availability of modern highly porous surfaces for osseointegration into patellar and tibial components have given surgeons confidence to increase the use of cementless TKA in the United States. It is encouraging that despite this relatively rapid transition we found no associated risk of all-cause revision or revision for mechanical loosening with the use of cementless TKA when controlling for age, BMI, sex, CCI, and use of CR design. Nevertheless, we believe caution is required prior to broad expansion of cementless TKA utilization in the population aged ≥65 years as appropriate indications for use are established, and future investigations demonstrating robust mid- and long-term outcomes of this relatively new technology are established.
