Journal Logo

Revision Risk for Total Knee Arthroplasty Converted from Medial Unicompartmental Knee Arthroplasty

Comparison with Primary and Revision Arthroplasties, Based on Mid-Term Results from the Danish Knee Arthroplasty Registry

El-Galaly, Anders MD1,2; Kappel, Andreas MD1,2; Nielsen, Poul Torben MD1; Jensen, Steen Lund MD, PhD1,2

The Journal of Bone and Joint Surgery: November 20, 2019 - Volume 101 - Issue 22 - p 1999-2006
doi: 10.2106/JBJS.18.01468
Scientific Articles
Free
Disclosures
Infographic

Background: Medial unicompartmental knee arthroplasties (UKAs) have good clinical outcomes but implant survival is inferior to that of total knee arthroplasties (TKAs). Conversion to a TKA is a reliable option when UKA fails. However, there is controversy regarding these conversions. The aim of this study was to analyze the survival of TKAs converted from UKAs when compared with both primary and revision TKAs.

Methods: On the basis of registrations in the Danish Knee Arthroplasty Registry from 1997 to 2017, 1,012 TKAs converted from UKAs were compared with 73,819 primary TKAs and 2,572 revision TKAs. The primary outcome was the risk of revision. Secondarily, the study analyzed the influence of different implants, the indication for the UKA conversion, and surgical volume on the survival of TKA converted from UKA. Third, the study compared the indications for revision.

Results: The converted UKAs were mainly mobile-bearing (85%) and, at the time of conversion, the patients were younger (mean [standard deviation], 66 ± 10 years) and more were Charnley class A (55%) compared with patients with primary TKA (70 ± 9 years and 35% class A) or revision TKA (70 ± 10 years and 42% class A) (all p < 0.001). The survival of TKAs converted from UKAs was comparable with that of revision TKAs (p = 0.42) and significantly inferior to the survival of primary TKAs (p < 0.001). This relationship was unaffected by differences between the groups, with an adjusted hazard ratio of 0.94 (95% confidence interval [CI]: 0.74 to 1.19) compared with revision TKAs and 3.00 (95% CI: 2.47 to 3.66) compared with primary TKAs. The survival of TKA converted from UKA was unaffected by differences in the conversion implants (all p ≥ 0.47), experience with revision surgery (all p ≥ 0.06), and the indications for the UKA-to-TKA conversion (all p ≥ 0.27). Instability (26%) and unexplained pain (13%) were more frequent indications for revisions of TKA converted from UKA (p < 0.001).

Conclusions: TKA converted from medial UKA has a 3-fold higher risk of revision when compared with primary TKA. The implant survival resembled that of revision TKA but with a higher prevalence of unexplained pain and instability.

Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

1Orthopaedic Research Unit, Aalborg University Hospital, Aalborg, Denmark

2Department of Clinical Medicine, Aalborg University, Aalborg, Denmark

Email address for A. El-Galaly: galaly@me.com

Investigation performed at the Department of Orthopaedic Surgery, Aalborg University Hospital, Aalborg, Denmark

Disclosure: The authors indicated that no external funding was received for any aspect of this work. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article (http://links.lww.com/JBJS/F452).

Medial unicompartmental knee arthroplasty (UKA) is a well-established treatment for primary osteoarthritis in the medial compartment1. When UKA fails, conversion to total knee arthroplasty (TKA) is a reliable solution, and often the conversion can be conducted with primary implants2. The limited surgical trauma and the ease of converting a UKA to a TKA when compared with revising a TKA might lead to the assumption that the outcome of a TKA converted from a UKA mimics the outcome of a primary TKA. However, this assumption is still debated, with authors of clinical studies reporting that the survival of a TKA converted from a UKA is like that of a primary TKA3-5 while registry studies have shown a survival similar to that of a revision TKA6,7. Our study aim was to analyze the survival of TKA converted from medial UKA in a large unselected cohort from the Danish Knee Arthroplasty Registry (DKR). The main objective was to analyze whether the survival of TKA converted from UKA was comparable with that of primary or revision TKA. Secondary objectives were to analyze the indications for revision and the influence of surgical differences on the estimated survival.

Back to Top | Article Outline

Materials and Methods

The study was designed as a cohort study based on the DKR following approval from the Danish Data Protection Agency (entry number 2008-58-0028, ID number 2016-138) and is presented in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement8. The DKR was established in 1997 to monitor the survival of knee arthroplasties performed in Denmark (estimated population in 2017: 5.8 million). Reporting is mandatory, and data are collected prospectively through standardized forms. The registry has been reported to be suitable for epidemiological studies9, with 91% of the knee arthroplasties conducted since 1997 registered10. The DKR records patient characteristics such as vital status, sex, age, side of surgery, weight, previous surgical procedures on the knee, and indication for surgery as well as information regarding duration of surgery, need for additional components (stem, augments, or cones), and perioperative complications (fractures, rupture of patellar ligaments, etc.). Comorbidity is depicted by the Charnley classification, which has been associated with the outcomes of arthroplasties11. Charnley class A indicates unilateral arthritis; B1, bilateral arthritis; B2, contralateral arthroplasty; and C, severe limitations in activities of daily living. In the DKR, revision is defined as the removal, exchange, or addition of any part of the implant, and multiple indications can be registered for a single revision. In a recent study12, we defined a clinical hierarchy of indications and, on basis of that hierarchy, considered only the most clinically important indication for each revision in the present study. The type of TKA was sorted by level of constraint (cruciate-retaining, posterior stabilized, constrained condylar, hinged, and undefined [when the implant was unknown]).

All TKAs converted from medial UKAs (TKA converted from UKA) from January 1, 1997, through December 31, 2017, were retrieved from the DKR. For comparison, TKAs performed for primary osteoarthritis in knees without prior surgery (de novo TKA) and their first-time revisions (TKA revised from TKA) were retrieved as well. Bias from bilateral observations has been reported to be negligible in large arthroplasty registry studies; thus, each knee was considered individually with the possibility of 1 patient contributing with 2 observations13,14.

Back to Top | Article Outline

Statistical Analysis

The TKA survival was estimated by Kaplan-Meier analysis with origin at the performance of the TKA converted from UKA, de novo TKA, or TKA revised from TKA. The end point was revision for any indication, and unrevised knees were censored at the time of death, emigration from Denmark, or the end of the study period (December 31, 2017), whichever came first. The estimated implant survival was compared between groups using the log-rank test, and multivariate Cox regression was used to estimate the influence of preoperative patient characteristics (age in years, sex, Charnley class, and weight in kg). The assumption of proportional hazards was checked by the Schoenfeld residuals test for all covariates, and variates violating this assumption were included as a time-dependent covariate. Survival estimations were limited to 15 years of follow-up because of the small number of observations exceeding this. Missing variables were estimated by multiple imputation based on either predictive mean matching or ordered logistic regression. Categorical variables were compared using the chi-square test or Fisher exact test, and continuous variables were compared using the Wilcoxon rank-sum test or Student t test. The data are presented as a 95% confidence interval (CI), median with interquartile range (IQR), or mean and standard deviation (mean ± SD), and p < 0.05 was considered significant. All analyses were conducted in Stata 15 (StataCorp).

Back to Top | Article Outline

Results

Study Population

There were 1,012 TKAs converted from UKA, 73,819 de novo TKAs, and 2,572 TKAs revised from TKA included in the study (Fig. 1). Multiple imputation was conducted for missing values for weight (n = 2,159), Charnley class (n = 262), and duration of surgery (n = 187). The baseline patient and surgery characteristics are shown in Table I. The duration of follow-up for TKA converted from UKA (4.3 years, IQR: 1.8 to 7.1 years) was comparable with that for TKA revised from TKA (4.1 years, IQR: 1.9 to 7.1 years; p = 0.50); however, it was significantly shorter than that for de novo TKA (6.1 years, IQR: 3.1 to 9.4 years; p < 0.001). The distribution of sex was similar in all 3 groups (p > 0.24), with a minority of men (range, 35% to 37% men). Patients with TKA converted from UKA were significantly younger (66 ± 10 years) and more of them were classified as Charnley class A when compared with both TKA revised from TKA (70 ± 10 years and 42% class A) and de novo TKA (70 ± 9 years and 35% class A) (all p < 0.001). The types of TKA implants differed significantly among the groups (p < 0.001), with cruciate-retaining implants being the most commonly used for both TKA converted from UKA (63%) and de novo TKA (84%) and constrained condylar implants being the most frequently used for TKA revised from TKA (47%). Use of posterior stabilized implants was significantly more frequent in TKA converted from UKA (30%) than in either de novo TKA (13%) or TKA revised from TKA (15%) (p < 0.001 for both). The median procedure time for TKA converted from UKA (90 minutes, IQR: 75 to 120 minutes) was significantly longer than that for de novo TKA (65 minutes, IQR: 58 to 80 minutes; p < 0.001) but significantly shorter than the time for TKA revised from TKA (115 minutes, IQR: 80 to 141 minutes; p < 0.001). Similarly, the need for additional components was more frequent in TKA converted from UKA (30%) than in de novo TKA (1%, p < 0.001) but less frequent than in TKA revised from TKA (61%, p < 0.001).

Fig. 1

Fig. 1

TABLE I - Baseline Characteristics
TKA Converted from UKA De Novo TKA TKA Revised from TKA
Observations (no.) 1,012 73,819 2,572
Revisions (no. [%]) 105 (10%) 2,572 (3%)* 244 (9%)
Median follow-up (IQR) (yr) 4.3 (1.8-7.1) 6.1 (3.1-9.4)* 4.1 (1.9-7.1)
Patient characteristics
 Sex (no. [%])
  Male 355 (35%) 26,467 (36%) 956 (37%)
  Female 657 (65%) 47,352 (64%) 1,616 (63%)
 Mean age (SD) (yr) 66 (10) 70 (9)* 70 (10)*
 Mean weight (SD) (kg) 84 (18) 84 (SD:19) 86 (20)*
 Charnley class (no. [%])
  A 560 (55%) 26,079 (35%)* 1,091 (42%)*
  B1 176 (17%) 25,595 (35%)* 429 (17%)*
  B2 221 (22%) 18,172 (25%)* 811 (32%)*
  C 55 (5%) 3,973 (5%)* 241 (9%)*
Surgical characteristics
 TKA type (no. [%])
  Cruciate-retaining 633 (63%) 62,051 (84%)* 647 (25%)*
  Posterior stabilized 302 (30%) 9,443 (13%)* 386 (15%)*
  Hinged 11 (1%) 129 (0%)* 182 (7%)*
  Constrained condylar 64 (6%) 274 (0%)* 1,212 (47%)*
  Undefined 2 (0%) 1,922 (3%)* 145 (6%)*
 Median procedure time (IQR) (min) 90 (75-120) 65 (58-80)* 115 (80-141)*
 Perioperative complications (no. [%]) 12 (1%) 462 (1%) 56 (2%)
 Additional components used (no. [%]) 301 (30%) 724 (1%)* 1,578 (61%)*
*
P < 0.001 for the difference compared with TKA converted from UKA.
P ≤ 0.05 for the difference compared with TKA converted from UKA.

Back to Top | Article Outline

Survival Analyses

The survival estimates for TKA converted from UKA, TKA revised from TKA, and de novo TKA are shown in Figure 2 and Cox regressions, in Tables II and III. The 5, 10, and 15-year estimated implant survival rates (95% CI) for TKA converted from UKA were 0.89 (0.87 to 0.91), 0.83 (0.79 to 0.87), and 0.78 (0.71 to 0.83). The respective values for TKA revised from TKA were 0.90 (0.88 to 0.91), 0.86 (0.84 to 0.88), and 0.78 (0.71 to 0.84). Thus, the survival of TKAs converted from UKAs did not differ from that of TKAs revised from TKAs (p = 0.42). Similarly, Cox regression showed no difference between TKA converted from UKA and TKA revised from TKA, with an estimated crude hazard ratio (HR) of 1.10 (95% CI: 0.87 to 1.38, p = 0.42) and a multivariate adjusted HR of 0.94 (95% CI: 0.74 to 1.19, p = 0.58). The 5, 10, and 15-year estimated survival rates for de novo TKA were 0.97 (0.97 to 0.97), 0.96 (0.95 to 0.96), and 0.94 (0.94 to 0.94). The survival rate of TKA converted from UKA was significantly lower than that of de novo TKA (p < 0.001), which corresponded to an unadjusted HR for revision of 3.65 (95% CI: 3.00 to 4.44, p < 0.001). After multivariate adjustment, the risk of revision decreased to 3.00 (95% CI: 2.47 to 3.66, p < 0.001).

Fig. 2

Fig. 2

TABLE II - Risk of Revision for TKA Converted from UKA Versus TKA Revised from TKA*
Covariates HR (95% CI)
Unadjusted Adjusted for Sex Adjusted for Sex, Age Adjusted for Sex, Age, Weight Adjusted for Sex, Age, Weight, Charnley Class
TKA from UKA 1.10 (0.87-1.38) 1.11 (0.88-1.40) 0.95 (0.75-1.20) 0.95 (0.75-1.21) 0.94 (0.74-1.19)
Female sex Ref. Ref. Ref. Ref.
Male sex 1.50 (1.18-1.79) 1.48 (1.19-1.82) 1.46 (1.17-1.82) 1.45 (1.16-1.80)
Age per year 0.97 (0.96-0.98) 0.97 (0.96-0.98) 0.97 (0.96-0.98)
Weight per kg 1.00 (0.99-1.01) 1.00 (0.99-1.01)
Charnley class
 A Ref.
 B1 1.05 (0.79-1.39)
 B2 0.71 (0.53-0.93)§
 C 1.18 (0.81-1.74)
*
Cox regressions with successive adjustments for patient characteristics.
P < 0.001.
P < 0.01.
§
P < 0.05.

TABLE III - Risk of Revision for TKA Converted from UKA Versus De Novo TKA*
Covariates HR (95% CI)
Unadjusted Adjusted for Sex Adjusted for Sex, Age Adjusted for Sex, Age, Weight Adjusted for Sex, Age, Weight, Charnley Class
TKA from UKA 3.65 (3.00-4.44) 3.65 (3.01-4.44) 3.04 (2.50-3.71) 3.08 (2.53-3.75) 3.00 (2.47-3.66)
Female sex Ref. Ref. Ref. Ref.
Male sex 1.15 (1.06-1.24) 1.11 (1.03-1.20) 1.08 (0.99-1.17) 1.07 (0.99-1.16)
Age per year (baseline) 0.98 (0.97-0.98) 0.98 (0.98-0.99) 0.98 (0.98-0.99)
Age (yearly) 0.99 (0.99-0.99) 0.99 (0.99-0.99) 0.99 (0.99-0.99)
Weight per kg 1.00 (1.00-1.01) 1.00 (1.00-1.01)
Charnley class
 A Ref.
 B1 0.90 (0.83-0.99)§
 B2 0.71 (0.64-0.80)
 C 0.99 (0.83-1.18)
*
Cox regressions with successive adjustments for patient characteristics. Age violated the assumption of proportional hazard and thus was included as a time-varying covariate.
P < 0.001.
P < 0.01.
§
P < 0.05.

Back to Top | Article Outline

Type of Implant

Cruciate-retaining (n = 511) and posterior stabilized (n = 187) implants without additional components were the most frequent implants used for the TKA converted from UKA and accounted for 69% of all implants in this group. Cruciate-retaining (n = 122) and posterior stabilized (n = 115) implants with the need for additional components accounted for 23%, and constrained implants (n = 75) accounted for 7%. Undefined implants (n = 2) were omitted from this analysis. The need for additional components with cruciate-retaining or posterior stabilized implants did not affect the risk of revision, with an adjusted HR of 1.18 (95% CI: 0.75 to 1.88, p = 0.47). Similarly, constrained implants were not associated with a significant increase in the risk of revision when compared with cruciate-retaining or posterior stabilized implants, with an adjusted HR of 1.17 (95% CI: 0.57 to 2.42, p = 0.66).

Back to Top | Article Outline

Revision Load

During the study period, 5 centers each performed >190 TKA-from-TKA revisions, which accounted for 1,386 (54%) of the 2,572 TKA-from-TKA revisions. Thus, these 5 centers were regarded as being experienced in revision surgery and denoted as revision centers. In total, 469 (46%) of the 1,012 TKA-from-UKA revisions were conducted at revision centers. At these centers, more females were treated (69% compared with 61% at the other centers, p = 0.01), the mean age was lower (64 ± 10 compared with 66 ± 10 years, p = 0.001), and the median procedure time was shorter (94 minutes [IQR: 72 to 110 minutes] compared with 101 minutes [IQR: 75 to 120 minutes], p = 0.002). Within the first 3 years, there was no difference in the estimated survival rate of TKA from UKA between revision and non-revision centers (p = 0.99), but thereafter the survival rate was significantly lower at revision centers (p = 0.04). Multivariate regression showed no significant difference following the third year (HR = 1.96, 95% CI: 0.97 to 3.95; p = 0.06).

Back to Top | Article Outline

UKA Characteristics

Predominantly, mobile-bearing UKAs were inserted in the study period (88%) and they accounted for 85% (n = 861) of the UKAs converted to TKAs. The risk of revision of the subsequent TKAs was not affected by the type of UKA bearing (p = 0.94). Aseptic loosening (27%, n = 277), unexplained pain (25%, n = 250), and progression of arthritis (18%, n = 186) were the most frequent indications for conversion of the UKA to the TKA. None of these indications were associated with an increased risk of revision of the subsequent TKA (all p ≥ 0.27).

Back to Top | Article Outline

Indications for Revision

The indications for revision of TKA converted from UKA differed from the indications for revision of both de novo TKA and TKA revised from TKA as shown in Table IV (all p < 0.001). Revision for instability was more frequently in the group of TKAs converted from UKAs (26%) than in either the TKA-from-TKA (18%) or the de novo TKA (16%) group. Revision for infection was less frequent in the group with TKA converted from UKA (10%) than in either the group with de novo TKA (16%) or the group with TKA revised from TKA (23%). Fourteen (13%) of the TKAs converted from UKA were revised because of unexplained pain, which was double the frequency observed in the other 2 groups (6%). Five of these TKAs were previously converted from UKA because of explained pain, which corresponded to 2% of the UKAs converted because of unexplained pain.

TABLE IV - Indications for Revision of TKA Converted from UKA, De Novo TKA, and TKA Revised from TKA
Hierarchy Indications TKA Converted from UKA De Novo TKA* TKA Revised from TKA*
1 Infection 11 (10%) 409 (16%) 56 (23%)
2 Aseptic loosening 40 (38%) 685 (27%) 66 (27%)
3 Wear 0 (0%) 55 (2%) 2 (1%)
4 Instability 27 (26%) 405 (16%) 44 (18%)
5 Patellar resurfacing 2 (2%) 51 (2%) 1 (0%)
6 Unexplained pain 14 (13%) 151 (6%) 14 (6%)
7 Others 3 (3%) 151 (6%) 17 (7%)
8 Undefined 8 (8%) 665 (26%) 44 (18%)
Total 105 (100%) 2,572 (100%) 244 (100%)
*
P < 0.001 for the difference compared with TKA converted from UKA.

Back to Top | Article Outline

Sensitivity Analyses

In this study, there was bilateral observation for 14,348 patients (23 with bilateral TKA converted from UKA, 14,151 with bilateral de novo TKA, and 174 with TKA converted from UKA on 1 side and de novo TKA on the other). Repeated analyses including only the first knee to receive surgery did not change the adjusted HRs for TKA converted from UKA compared with either de novo TKA (HR: 3.05, 95% CI: 2.50 to 3.73; p < 0.001) or TKA revised from TKA (HR: 1.02, 95% CI: 0.79 to 1.30; p = 0.90).

To account for death as a competing risk, we repeated the calculations using a Fine-Gray proportional subhazards model15,16. A small, clinically unimportant difference from the presented analysis was found in the adjusted HR for TKA converted from UKA compared with both TKA revised from TKA (HR: 0.94, 95% CI: 0.74 to 1.19; p = 0.59) and de novo TKA (HR: 2.96, 95% CI: 2.43 to 3.60; p < 0.001).

Follow-up studies of arthroplasties often include a minimum of 2 years of follow-up. In the current study, 73% (742) of the TKAs converted from UKA, 73% (1,879) of the TKAs revised from TKA, and 83% (61,632) of the de novo TKAs were followed for >2 years. Repeated analyses of only those arthroplasties with a minimum of 2 years of follow-up did not change the adjusted HRs for TKA converted from UKA compared with TKA revised from TKA (1.01, 95% CI: 0.71 to 1.44; p = 0.94) or compared with de novo TKA (3.46, 95% CI: 2.60 to 4.60; p < 0.001).

To account for changes over time, the observations were split into 2 decades at 2007. We found no difference in survival of TKAs converted from UKA in or before 2007 (n = 144) when compared with those performed after 2007 (n = 868), with a multivariate adjusted HR of 1.37 (95% CI: 0.86 to 2.18, p = 0.18).

Back to Top | Article Outline

Discussion

In the future, more patients will need conversion from UKA to TKA given the increased popularity and inferior survival of UKAs17,18. This nationwide registry study compared the survival of TKA converted from medial UKA with survival of TKA revised from TKA and primary TKA. TKA converted from UKA had a survival comparable with that of TKA revised from TKA and thus had a 3 times increased risk of revision when compared with primary TKA. Our results support studies from the Norwegian and Swedish knee registries. Leta et al. reported similar survival between aseptic TKA conversions from UKA and aseptic TKA revisions from TKA6. Similarly, Robertsson and W-Dahl reported a 2.8-fold increase in the risk of revision for TKA converted from UKA when compared with primary TKA7. In contrast, clinical studies showed similar survival rates for TKA from UKA and primary TKA4,5, which has been supported by a recent review3. In general, the survival of TKA converted from UKA reported in nationwide registries has been inferior to that in single-center studies. This might be a result of superior survival at high-volume clinics19. In our study, the surgical centers were divided according to their experience with revision surgery and, surprisingly, centers with limited TKA revisions tended to have superior survival of TKA converted from UKA. This tendency might be explained by differences in the case mix between revision and non-revision centers. Similarly, previous studies have shown that, when bone and soft-tissue conditions allow for use of primary implants, survival comparable with primary TKA can be expected4,5,20. However, in our study the survival of TKA converted from UKA was unaffected by use of constrained implants or the need for additional components (such as stems, augments, and cones) during the conversion from UKA.

In our study, the indications for UKA-to-TKA conversion were in accordance with those in previous publications6,7, with aseptic loosening, unexplained pain, and progression of arthritis as major indications. Revision due to unexplained pain without demonstrable abnormality is questionable21 and has, in UKA, been associated with inferior outcome scores22. We did not find conversion due to unexplained pain to be associated with inferior survival of the subsequent TKA or subsequent revision due to unexplained pain. However, in general, revision due to unexplained pain was more frequent in the group that had TKA converted from UKA than in the other groups (13% compared with 6%). Likewise, revision due to instability was more frequent (26%) with TKA converted from UKA whereas revision because of infection was less frequent (10%). The increased risk of instability and the longer duration of surgery observed in the group with TKA converted from UKA might be an indication of more complex surgery. Instability should be addressed intraoperatively, and the high prevalence of instability after these UKA conversions calls for thorough preoperative planning to ensure the best implant selection and surgical expertise23. The lower rate of infection in the group with TKA converted from UKA is surprising but might be explained by a healthier group of patients, depicted by the Charnley class and age selected for UKA.

This study has several limitations. First, it lacked patient-reported outcomes. A relationship between survival and patient-reported outcomes has been reported24, but poorly performing TKAs are not always revised. Second, there are inherent limitations in retrospective registry studies. As they are not randomized, unmeasured confounders might affect the results. In addition, registry studies are prone to misclassifications. However, the potential misclassifications in this study are assumed to be evenly distributed among the 3 groups and therefore tend to underestimate the observed differences among them25. Finally, clinical and surgical information is limited in registry studies. For instance, the current study lacks information about specific comorbidities associated with inferior arthroplasty survival, such as diabetes26. Similarly, information about the congruency of the polyethene insert was unobtainable.

In summary, this nationwide registry study suggested that the survival of TKA converted from medial UKA resembles the survival of TKA revised from TKA and is inferior to that of primary TKA. On the basis of this study, we believe that careful consideration is necessary before using medial UKA as treatment for knee osteoarthritis, as a potential conversion to a TKA decreased implant survival when compared with that following primary TKA.

Note: The authors thank the Danish Knee Arthroplasty Registry for providing the data used in this study and the Danish orthopaedic surgeons for thorough registration of their surgical procedures.

Back to Top | Article Outline

References

1. Murray DW, Parkinson RW. Usage of unicompartmental knee arthroplasty. Bone Joint J. 2018 Apr 1;100-B(4):432-5.
2. Khan Z, Nawaz SZ, Kahane S, Esler C, Chatterji U. Conversion of unicompartmental knee arthroplasty to total knee arthroplasty: the challenges and need for augments. Acta Orthop Belg. 2013 Dec;79(6):699-705.
3. Thienpont E. Conversion of a unicompartmental knee arthroplasty to a total knee arthroplasty: can we achieve a primary result? Bone Joint J. 2017 Jan;99-B(1)(Supple A):65-9.
4. Lombardi AV Jr, Kolich MT, Berend KR, Morris MJ, Crawford DA, Adams JB. Revision of unicompartmental knee arthroplasty to total knee arthroplasty: is it as good as a primary result? J Arthroplasty. 2018 Jul;33(7S):S105-8. Epub 2018 Mar 17.
5. Sierra RJ, Kassel CA, Wetters NG, Berend KR, Della Valle CJ, Lombardi AV. Revision of unicompartmental arthroplasty to total knee arthroplasty: not always a slam dunk! J Arthroplasty. 2013 Sep;28(8)(Suppl):128-32. Epub 2013 Jul 23.
6. Leta TH, Lygre SHL, Skredderstuen A, Hallan G, Gjertsen JE, Rokne B, Furnes O. Outcomes of unicompartmental knee arthroplasty after aseptic revision to total knee arthroplasty a comparative study of 768 TKAs and 578 UKAs revised to TKAs from the Norwegian Arthroplasty Register (1994 to 2011). J Bone Joint Surg Am. 2016 Mar 16;98(6):431-40.
7. Robertsson O, W-Dahl A. The risk of revision after TKA is affected by previous HTO or UKA. Clin Orthop Relat Res. 2015 Jan;473(1):90-3.
8. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP, STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Int J Surg. 2014 Dec;12(12):1495-9. Epub 2014 Jul 18.
9. Pedersen AB, Mehnert F, Odgaard A, Schrøder HM. Existing data sources for clinical epidemiology: the Danish Knee Arthroplasty Register. Clin Epidemiol. 2012;4:125-35. Epub 2012 May 7.
10. Odgaard A, Madsen F, Rasmussen LE, Schrøder H, Østgaard SE, Troelsen A. Dansk Knæalloplastikregister. Årsrapport 2018. https://www.sundhed.dk/content/cms/99/4699_dkr-rapport-2018_til-offentligg%C3%B8relse.pdf. Accessed 2019 Jul 5.
11. Bjorgul K, Novicoff WM, Saleh KJ. Evaluating comorbidities in total hip and knee arthroplasty: available instruments. J Orthop Traumatol. 2010 Dec;11(4):203-9. Epub 2010 Nov 13.
12. El-Galaly A, Nielsen PT, Jensen SL, Kappel A. Prior high tibial osteotomy does not affect the survival of total knee arthroplasties: results from the Danish Knee Arthroplasty Registry. J Arthroplasty. 2018 Jul;33(7):2131-2135.e1. Epub 2018 Feb 28.
13. Robertsson O, Ranstam J. No bias of ignored bilaterality when analysing the revision risk of knee prostheses: analysis of a population based sample of 44,590 patients with 55,298 knee prostheses from the national Swedish Knee Arthroplasty Register. BMC Musculoskelet Disord. 2003 Feb 5;4:1.
14. Na YG, Kang YG, Chang MJ, Chang CB, Kim TK. Must bilaterality be considered in statistical analyses of total knee arthroplasty? Clin Orthop Relat Res. 2013 Jun;471(6):1970-81. Epub 2013 Jan 31.
15. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94(446):496-509.
16. Gillam MH, Ryan P, Graves SE, Miller LN, de Steiger RN, Salter A. Competing risks survival analysis applied to data from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop. 2010 Oct;81(5):548-55.
17. Badawy M, Fenstad AM, Bartz-Johannessen CA, Indrekvam K, Havelin Li, Robertsson O, W-Dahl A, Eskelinen A, Mäkelä K, Pedersen AB, Schrøder HM, Furnes O. Hospital volume and the risk of revision in Oxford unicompartmental knee arthroplasty in the Nordic countries -an observational study of 14,496 cases. BMC Musculoskelet Disord. 2017 Sep 7;18(1):388.
18. Chawla H, van der List JP, Christ AB, Sobrero MR, Zuiderbaan HA, Pearle AD. Annual revision rates of partial versus total knee arthroplasty: a comparative meta-analysis. Knee. 2017 Mar;24(2):179-90.
19. Liddle AD, Pandit H, Judge A, Murray DW. Effect of surgical caseload on revision rate following total and unicompartmental knee replacement. J Bone Joint Surg Am. 2016 Jan 6;98(1):1-8.
20. Johnson S, Jones P, Newman JH. The survivorship and results of total knee replacements converted from unicompartmental knee replacements. Knee. 2007 Mar;14(2):154-7. Epub 2007 Feb 1.
21. Toms AD, Mandalia V, Haigh R, Hopwood B. The management of patients with painful total knee replacement. J Bone Joint Surg Br. 2009 Feb;91(2):143-50.
22. Kerens B, Boonen B, Schotanus MGM, Lacroix H, Emans PJ, Kort NP. Revision from unicompartmental to total knee replacement: the clinical outcome depends on reason for revision. Bone Joint J. 2013 Sep;95-B(9):1204-8.
23. Lombardi AV Jr, Berend KR, Adams JB. Why knee replacements fail in 2013: patient, surgeon, or implant? Bone Joint J. 2014 Nov;96-B(11)(Supple A):101-4.
24. Maradit Kremers H, Kremers WK, Berry DJ, Lewallen DG. Patient-reported outcomes can be used to identify patients at risk for total knee arthroplasty revision and potentially individualize postsurgery follow-up. J Arthroplasty. 2017 Nov;32(11):3304-7. Epub 2017 May 31.
25. Sorensen HT, Sabroe S, Olsen J. A framework for evaluation of secondary data sources for epidemiological research. Int J Epidemiol. 1996 Apr;25(2):435-42.
26. Bolognesi MP, Marchant MH Jr, Viens NA, Cook C, Pietrobon R, Vail TP. The impact of diabetes on perioperative patient outcomes after total hip and total knee arthroplasty in the United States. J Arthroplasty. 2008 Sep;23(6)(Suppl 1):92-8.

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2019 by The Journal of Bone and Joint Surgery, Incorporated