Secondary Logo

Journal Logo


Unicompartmental or Total Knee Arthroplasty?

Results from a Matched Study

Amin, Anish K, MRCS; Patton, James T, FRCS; Cook, Robert E, FRCS; Gaston, Mark, MRCS; Brenkel, Ivan J, FRCS

Author Information
Clinical Orthopaedics and Related Research®: October 2006 - Volume 451 - Issue - p 101-106
doi: 10.1097/01.blo.0000224052.01873.20
  • Free


Surgical options for unicompartmental arthritis of the knee include high tibial osteotomy, unicompartmental knee arthroplasty (UKA), and total knee arthroplasty (TKA). Results of high tibial osteotomy tend to decline with time.5,11,22 Compared with high tibial osteotomy, UKA has a greater initial success rate, fewer complications, and the difference in the survival rate increases with time.7,30,33 Total knee arthroplasty is a well-established procedure because of consistent, reproducible, long-term results from large survivorship series.9,10,27,28 Initial reports of high failure rates with UKA13,15,20,21 were followed by good 10-year results. Ten-year survivorship rates for UKAs based on revision for any reason vary from 84% to 98%, with the best results obtained using strict selection criteria.2,4,24,31

Evidence for UKA or TKA in treating unicompartmental arthritis is based on several large noncomparative survivorship series for both types of prostheses, but there are few direct comparative studies between UKA and TKA,1,6,17,25,29,32 and even fewer in comparable patients,6,17,25 to avoid bias. Because of the renewed interest in UKA, we evaluated results of UKA and TKA in comparable patients.

We determined whether the active range of motion (ROM), Knee Society score (KSS),12,14 and 5-year survivorship rate after 54 consecutive unilateral UKAs were similar to those of 54 unilateral TKAs performed in patients matched for age, gender, body mass index (BMI), preoperative active ROM, and preoperative KSS.


From 1995 to 1999, more than 600 consecutive patients had UKAs or TKAs at one orthopaedic institution. Patients were followed up prospectively at 6, 18, 36, and 60 months postoperatively. The followup was performed independently by a dedicated knee audit service led by an arthroplasty nurse practitioner. Preoperative patient data, postoperative complications, and subsequent followup KSS (knee and function scores) were recorded prospectively in a database maintained by the knee audit service.

We identified 54 consecutive unilateral UKAs performed for patients with osteoarthritis (OA) using our prospective database. Bilateral procedures were excluded to ensure independence of measurements recorded for each knee. The 54 patients who had UKAs were matched individually with 54 patients who had uni-lateral TKAs, for age, gender, BMI, preoperative active ROM, and preoperative Knee Society knee and function scores. The matched patients in the TKA group were identified retrospectively by one author (JTP) using only the preoperative data recorded in the database maintained at the institution. We performed the matching process blinded to the eventual outcomes of UKAs and TKAs. We matched preoperative variables using a spreadsheet program in the following order: age, BMI, preoperative knee score, preoperative function score, preoperative active ROM, and gender. There were no differences in the preoperative characteristics of the patients in the UKA and TKA groups (Table 1).

Preoperative Characteristics for the UKA and TKA Groups

Patients selected for UKAs had anteromedial OA (with a normal lateral compartment; patellofemoral OA was not an exclusion for UKA), intact cruciate ligaments, flexion deformity of 15° or less, and varus deformity of 15° or less. Patients selected for TKAs had unicompartmental or more extensive OA. We used a standard surgical approach with a medial parapatellar incision and arthrotomy for UKAs and TKAs. Osteophytes were removed in each group. We used the Oxford unicompartmental prosthesis (Biomet Ltd, Bridgend, UK) for all UKAs and the Press-Fit Condylar (PFC) prosthesis (DePuy International, Leeds, UK) for all TKAs. We followed manufacturers' instructions regarding the use of implants and instruments. We used low molecular weight heparin for prophylaxis against deep vein thrombosis (DVT). The postoperative rehabilitation program was identical in both groups. Procedures were performed by five surgeons with previous experience using both implants.

We performed statistical analyses using SPSS version 13.0 computer software (SPSS Inc, Chicago, IL). A repeated measures analysis of variance, combining between-subject and within-subject variables, was used to compare the active ROM and KSS scores after UKA and TKA. The between-subjects independent variable was the type of prosthesis (UKA or TKA), the within-subjects independent variable was time (6, 18, 36, and 60 months followup), and the dependent variables were the active ROM knee score and function score recorded at each of the four followups. The preoperative active ROM, knee scores, and function scores were included in the repeated measures models as covariates. Observations for patients who had revision surgery were excluded from the repeated measures analysis. All available observations recorded for patients who died or were lost to followup were included in the repeated measures analysis. Observations recorded for each knee arthroplasty were assumed as independent. There were no serious violations of the Levene's test of equality of error variances used to test the assumption of homogeneity of variance. We used tests of between-subjects effects to assess changes in the dependent variables (active ROM, knee score, or function score) with time. We used the Kaplan-Meier method to construct 5-year survivorship curves for the two groups. The primary end point for the survivorship analyses was revision for any reason. A secondary end point used for the survivorship analyses was worst-case, which considered any patient lost to followup as having failed results and assumed these patients had revision surgery. We compared proportions using the chi square test, continuous variables using Student's t test or Mann-Whitney U test, and survivorship using the log-rank test. The level of significance was set at p < 0.05.


The mean followup was 59 months (range, 24-71 months) in the UKA group and 59 months (range, 7-72 months) in the TKA group. At 5 years followup, the numbers of deaths in the two groups were similar (six of 54 patients in the UKA group [11%] and six of 54 patients in the TKA group [11%]). Followup was incomplete at 5 years for two patients (4%) in the UKA group and one patient (2%) in the TKA group. One patient declined additional followup, one patient was unable to attend after having a stroke, and one could not be traced despite attempts to contact the patient by telephone and letter, but the patient was confirmed as being alive according to general practitioner records. All three patients lost to followup had a satisfactory KSS at the last recorded followup, which was at 6 months for two patients (UKAs) and 18 months for one patient (TKA). Observations from patients who had revision surgery were excluded, but observations from patients who died or were lost to followup were included in the repeated measures analysis. The numbers of patients evaluated at each followup were as follows: 99 of 108 patients (92%) at 6 months, 97 of 108 patients (90%) at 18 months, 94 of 108 patients (87%) at 36 months, and 88 of 108 patients (81%) at 60 months.

The mean postoperative active ROM was greater (p = 0.001) in the UKA group compared with the TKA group (Fig 1). However, there were no differences between the UKA and TKA groups for the mean knee or function scores (Fig 2).

Fig 1
Fig 1:
A graph shows postoperative active ROM at different times after UKA and TKA. Bars show means, error bars show 95% CI of mean. Dotted bars represent TKA and lined bars represent UKA.
Fig 2A
Fig 2A:
B. A graph shows the Knee Society (A) knee and (B) function scores at different times after UKA and TKA. Bars show means, error bars show 95% CI of mean. Dotted bars represent TKA and lined bars represent UKA.

Five-year survivorship based on revision for any reason was lower (p = 0.01) in the UKA group (88%; 95% confidence interval [CI], 79-97%) than in the TKA group (100%; 95% CI, 100%) (Fig 3). The worst-case 5-year survivorship, assuming all patients lost to followup had revision surgery, also was lower (p = 0.01) in the UKA group (85%; 95% CI 75-95%) than in the TKA group (98%; 95% CI, 94-100%).

Fig 3
Fig 3:
A graph shows the 5-year survivorship curves for UKA and TKA based on revision for any reason.

At 5 years postoperatively, six UKAs had been revised to TKAs whereas none of the TKAs required revision. One patient who had a UKA had revision surgery for infection, four had revision surgery for pain, and one had revision surgery for recurrent dislocation of the mobile meniscal implant (Table 2). The median time to revision was 19.5 months (range, 4-59 months). In the four patients with UKAs who had revision surgery for persistent pain, none of their implants was loose or infected at the revision procedure. It was unclear from review of operation records whether these four patients had any progression of arthritis in other compartments or impingement of the meniscal bearing on the femur causing pain. An additional analysis comparing the 5-year survivorship rate based on revision for pain as an end point (excluding the one infection and one meniscal dislocation in the UKA group) was still lower (p = 0.04) in the UKA group (92%; 95% CI, 85-99%) compared with the TKA group (100%; 95% CI, 100%). Pain and function improved in three patients (two who had revision surgery for pain, one who had revision surgery for meniscal dislocation) after revision of the UKAs to TKAs. This was confirmed by an improvement in the KSS recorded after revision when compared with the prerevision KSS for these three patients (Table 2). One patient with infection continued to have pain and limited function despite revision and microbiologic confirmation of eradication of infection. One patient who had revision surgery because of pain remained symptomatic despite revision of the UKA to a TKA. The postrevision KSS was not able to be determined in one patient who had died.

Revision Details

The percentage of TKAs performed by the five surgeons, A, B, C, D, and E, were 28%, 13%, 35%, 15%, and 9% respectively. The percentage of UKAs performed by surgeons A, B, C, D, and E were 15%, 43%, 22%, 15%, and 5% respectively. Unicompartmental knee arthroplas-ties that required revision were performed by surgeons A (one UKA), B (three UKAs), and C (two UKAs).


We determined the results of UKAs and TKAs performed in patients with similar preoperative knee and functional characteristics to avoid bias. Survivorship remained superior for the TKA at 5 years postoperatively. Although the active ROM was greater for the UKA group, there was no difference in the overall clinical outcome based on the KSS at 6, 18, 36, and 60 months postoperatively.

We evaluated survivorship and clinical outcome after UKA or TKA in comparable patients from one orthopaedic unit, with multiple surgeons, using the same prosthetic design and using a formal outcome score measured prospectively and based on a strict repeated followup protocol. Our study design avoided problems associated with retrospective measurement of outcome scores,6,17 comparing patients with different preoperative knee and functional characteristics,1,32 or comparing multiple prosthetic designs.29 However, the comparison was not randomized, and a potential criticism of the study design is the retrospective identification of a matched TKA group for comparison with a consecutive series of UKAs. However, matched subjects were identified by one observer using only the preoperative data for each knee recorded in a database, and who was blinded to the eventual outcome after TKA. Data for all patients were collected prospectively and independently by a knee audit service at our institution.

Results from some direct comparative studies investigating the survivorship and clinical outcome after UKAs and TKAs1,25,32 have supported the renewed interest in UKAs. The clinical outcome based on disease-specific or patient-based outcome measures were not different after UKA or TKA (Table 3). Ten-year survivorship rates were reported as similar after UKA or TKA in one direct comparative study1 and other individual, noncomparative survivorship series after UKA.4,24 Surgeons who favor the UKA argue that the clinical outcome and 10-year survivorship rate seem to be similar after UKA or TKA, and UKA has the additional benefits of providing patients with a faster postoperative recovery, shorter hospital stay, and better postoperative ROM.17,25,29

Comparative Studies of Unicompartmental and Total Knee Arthroplasties

Our data suggest UKA may be associated with a better postoperative ROM than TKA, although the overall clinical outcome was similar for UKA and TKA. However, in contrast to published findings from comparative and non-comparative survivorship series, we found inferior mid-term survivorship after UKA. One must question why the 5-year survivorship rate after UKA (88%) compared poorly with the 10-year survivorship rate of 87% in a direct comparative study.1 There also was an excellent 10-year survivorship rate of 98% in a noncomparative series on the same prosthesis used in our study.24

There are few direct comparisons of survivorship rates after UKA and TKA.1,29 These studies however, have compared patients with different preoperative knee and functional characteristics, with TKAs performed in patients with more advanced disease. This may have biased the survivorship in favor of the UKA; a problem we avoided with preoperatively matched patients. The 5-year survivorship rate (100%) (based on revision for any reason) for TKA was high. However, the 5-year survivorship rate for 478 consecutive TKAs from which the study sample was selected was 98%, and compares well with rates in noncomparative survivorship series for TKAs from other centers.10,19,34 It may be argued that the inferior results after UKA reflect poor surgical technique compared with the results reported by the prosthesis designers.24 However, the results of TKAs compare favorably with reported results from other centers, which suggests the surgical team was adequately experienced in knee arthroplasties. There were multiple surgeons, all with previous experience in the use of the implant, and failures occurred among surgeons performing the majority of the arthroplasties. Therefore, a learning curve cannot explain the high failure rate in the UKA group. Additionally, the availability of a dedicated knee audit service would have led to better outcomes than the average community practice because of feedback from the service. We think the strict indications for performing the UKA mean the procedure is performed much less frequently than the TKA, and consistent results from different centers therefore may be more difficult to reproduce.

One argument made in favor of accepting a greater revision rate for the UKA is that the prosthesis is easier to revise than that of a TKA. Unicompartmental knee arthroplasty has been recommended for the young active patient32 in whom the prosthesis would be expected to fail within the lifetime, or in the elderly with low functional demands,29 with revision to a TKA should the UKA fail. However, the complexity of the revision procedure after UKA and long-term functional outcome remain controversial with favorable18,23 and unfavorable results.3,26 The need for a second operation in young active patients and the elderly can be questioned when long-term results of primary TKAs in both age groups have been excellent.8,16 The small number of patients requiring revision surgery makes reliable findings difficult, but the improvement in symptoms after revision surgery in 1/2 of the patients suggests an initial TKA may have allowed them to avoid the need for a revision procedure.

For comparable patients with OA of the knee, the survivorship rate remained superior for TKAs at 5 years. The active ROM provided with the UKA was greater, but there were no differences in the overall clinical outcomes after UKA and TKA. We think the TKA is a more reliable procedure. Midterm clinical outcomes were similar after UKA or TKA, but the complication rate may be greater for the UKA.


We thank Lorraine McComiskie, Ann Simpson, Janette McDonald, and Wilma Addison for data collection.


1. Ackroyd CE, Whitehouse SL, Newman JH, Joslin CC. A comparative study of the medial St Georg sled and kinematic total knee arthroplasties: ten-year survivorship. J Bone Joint Surg Br. 2002; 84:667-672.
2. Ansari S, Newman JH, Ackroyd CE. St Georg sledge for medial compartment knee replacement: 461 arthroplasties followed for 4 (1-17) years. Acta Orthop Scand. 1997;68:430-434.
3. Barrett WP, Scott RD. Revision of failed unicondylar unicompart-mental knee arthroplasty. J Bone Joint Surg Am. 1987;69: 1328-1335.
4. Berger RA, Nedeff DD, Barden RM, Sheinkop MM, Jacobs JJ, Rosenberg AG, Galante JO. Unicompartmental knee arthroplasty: clinical experience at 6-to 10-year followup. Clin Orthop Relat Res. 1999;367:50-60.
5. Berman AT, Bosacco SJ, Kirshner S, Avolio A Jr. Factors influencing long-term results in high tibial osteotomy. Clin Orthop Relat Res. 1991;272:192-198.
6. Cameron HU, Jung YB. A comparison of unicompartmental knee replacement with total knee replacement. Orthop Rev. 1988;17: 983-988.
7. Coventry MB. Upper tibial osteotomy for osteoarthritis. J Bone Joint Surg Am. 1985;67:1136-1140.
8. Duffy GP, Trousdale RT, Stuart MJ. Total knee arthroplasty in patients 55 years or younger: 10-to 17-year results. Clin Orthop Relat Res. 1998;356:22-27.
9. Font-Rodriguez DE, Scuderi GR, Insall JN. Survivorship of cemented total knee arthroplasty. Clin Orthop Relat Res. 1997;345: 79-86.
10. Gill GS, Joshi AB. Long term results of Kinematic Condylar knee replacement: an analysis of 404 knees. J Bone Joint Surg Br. 2001; 83:355-358.
11. Insall JN, Joseph DM, Msika C. High tibial osteotomy for varus gonarthrosis: a long-term follow-up study. J Bone Joint Surg Am. 1984;66:1040-1048.
12. Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res. 1989;248: 13-14.
13. Insall JN, Walker PS. Unicondylar knee replacement. Clin Orthop Relat Res. 1976;120:83-85.
14. Kreibich DN, Vaz M, Bourne RB, Rorabeck CH, Kim P, Hardie R, Kramer J, Kirkley A. What is the best way of assessing outcome after total knee replacement? Clin Orthop Relat Res. 1996;331: 221-225.
15. Laskin RS. Unicompartmental tibiofemoral resurfacing arthroplasty. J Bone Joint Surg Am. 1978;60:182-185.
16. Laskin RS. Total knee replacement in the patient older than 85 years. Clin Orthop Relat Res. 1999;367:43-49.
17. Laurencin CT, Zelicof SB, Scott RD, Ewald FC. Unicompartmental versus total knee arthroplasty in the same patient: a comparative study. Clin Orthop Relat Res. 1991;273:151-156.
18. Levine WN, Ozuna RM, Scott RD. Conversion of failed modern unicompartmental arthroplasty to total knee arthroplasty. J Arthroplasty. 1996;11:797-801.
19. Malkani AL, Rand JA, Bryan RS, Wallrichs SL. Total knee arthroplasty with the kinematic condylar prosthesis: a ten-year follow up study. J Bone Joint Surg Am. 1995;77:423-431.
20. Marmor L. Unicompartmental arthroplasty of the knee with a minimum ten-year follow-up period. Clin Orthop Relat Res. 1988;228: 171-177.
21. Marmor L. Unicompartmental knee arthroplasty: ten-to 13-year follow up study. Clin Orthop Relat Res. 1988;226:14-20.
22. Matthews LS, Goldstein SA, Malvitz TA, Katz BP, Kaufer H. Proximal tibial osteotomy: factors that influence the duration of satisfactory function. Clin Orthop Relat Res. 1988;229:193-200.
23. McAuley JP, Engh GA, Ammeen DJ. Revision of failed unicompartmental knee arthroplasty. Clin Orthop Relat Res. 2001;392: 279-282.
24. Murray DW, Goodfellow JW, O'Connor JJ. The Oxford medial unicompartmental arthroplasty: a ten-year survival study. J Bone Joint Surg Br. 1998;80:983-989.
25. Newman JH, Ackroyd CE, Shah NA. Unicompartmental or total knee replacement? Five-year results of a prospective, randomised trial of 102 osteoarthritic knees with unicompartmental arthritis. J Bone Joint Surg Br. 1998;80:862-865.
26. Padgett DE, Stern SH, Insall JN. Revision total knee arthroplasty for failed unicompartmental replacement. J Bone Joint Surg Am. 1991;73:186-190.
27. Ranawat CS, Boachie-Adjei O. Survivorship analysis and results of total condylar knee arthroplasty: eight-to 11-year follow-up period. Clin Orthop Relat Res. 1988;226:6-13.
28. Ritter MA, Campbell E, Faris PM, Keating EM. Long-term survival analysis of the posterior cruciate condylar total knee arthroplasty: a 10-year evaluation. J Arthroplasty. 1989;4:293-296.
29. Rougraff BT, Heck DA, Gibson AE. A comparison of tricompart-mental and unicompartmental arthroplasty for the treatment of gonarthrosis. Clin Orthop Relat Res. 1991;273:157-164.
30. Scott RD, Santore RF. Unicondylar unicompartmental replacement for osteoarthritis of the knee. J Bone Joint Surg Am. 1981;63: 536-544.
31. Tabor OB Jr, Tabor OB. Unicompartmental arthroplasty: a long-term follow-up study. J Arthroplasty. 1998;13:373-379.
32. Weale AE, Halabi OA, Jones PW, White SH. Perceptions of outcomes after unicompartmental and total knee replacements. Clin Orthop Relat Res. 2001;382:143-153.
33. Weale AE, Newman JH. Unicompartmental arthroplasty and high tibial osteotomy for osteoarthrosis of the knee: a comparative study with a 12-to 17-year follow-up period. Clin Orthop Relat Res. 1994;302:134-137.
34. Weir DJ, Moran CG, Pinder IM. Kinematic condylar total knee arthroplasty: 14-year survivorship analysis of 208 consecutive cases. J Bone Joint Surg Br. 1996;78:907-911.
© 2006 Lippincott Williams & Wilkins, Inc.