Introduction

The medial unicompartmental knee replacement (UKR) initially presented with unfavourable post-operative results and corresponding high failure rates [₁₂₃₄]. However, with the refinement of surgical techniques and implant design, and with patient selection becoming more defined, UKR is presently regarded as a clinically successful procedure to improve function, reduce pain and provide excellent implant survival in patients with unicompartmental arthritis of the knee [_5678910].

Previous literature has identified prognostic indicators for total knee replacements [_111213141516]. These have identified variables such as age, gender, weight, smoking and level of activity, as important pre-operative factors on outcome. To our knowledge, only age and, as a secondary evaluation, body mass index (BMI) have been assessed as potential prognostic indicators for UKR populations [₁₇₁₈₁₉]. These have been investigated using small samples, with the exception of Price et al’s study [₁₉]. Furthermore, although pre-operative functional level and degree of pre-operative angular deformity have been suggested as possible prognostic indicators [_20212223], a review of the literature did not identify any studies to validate such statements.

In response to this, the purpose of this study was to determine whether gender, age, BMI, smoking history, uni- or bilateral procedures, pre-operative location, operative location, pre-operative Oxford Knee Score (OKS), clinical and functional American Knee Society Score (AKS), Short Form-12 (SF-12) correlate with length of stay or 1 year OKS, clinical AKS, SF-12, knee range of motion, and maximum and minimum visual analogue scale (VAS) pain scores, in a cohort of medial UKR patients. This is warranted, as we hypothesis that by identify demographic and pre-operative factors at the time of consultation, the surgeon may better able to predict the outcome of UKR.

Patients and Methods

We retrospectively reviewed a prospective database of UKR surveillance. All UKRs were performed between the 1st January 2004 and 31st December 2006. Patients were considered for a UKR if they presented with: significant and functionally limiting pain related to medial compartment osteoarthritis; with no significant anterior or lateral pain; varus deformity less than 15°; fixed flexion deformity not greater than 10°; knee flexion of at least 110º under anaesthetic to allow access for preparation of the femoral condyle; correctable varus deformity; and an intact anterior and posterior cruciate ligaments and collateral ligaments. Exclusion criteria included: evidence of medial or lateral subluxation or posterior tibial bone loss to strongly suggest damage to the cruciate mechanism; inflammatory arthritis; previous high tibial osteotomy with overcorrection; sepsis; tibial or femoral shaft deformity; lateral compartment or severe patellofemoral joint osteoarthritis.

In all cases the Oxford Phase 3 UKR (Biomet, Warsaw, USA) was used. The operations were performed by 10 different consultant orthopaedic surgeons. All performed a minimal access medial parapatellar approach, using a skin incision of 8 to 10 centimetres. Femoral and tibial tray components were cemented. Intra-operative flexion and extension testing was performed prior to wound closure. In addition, prior to closure, local anaesthetic was infiltrated into the knee. Either clips or dissolvable sutures were used for skin closure and covered with a gauge, wool and crepe dressing.

All patients were prescribed paracetomol and codeine, and, if required diclofenac or oramorph in severe pain. Venous thromboprophylaxis involved intra-operative calf stimulation and early mobilisation. Low molecular weight heparin was only used in patients with risk factors for venous thromboembolism. Patients began mobilising and knee exercises after two hours post-operatively, supervised by physiotherapists. The aim was to discharge patients on the first post-operatively day.

Data was collected pre-operatively by Consultant Orthopaedic teams and the institute’s UKR Review programme. Pre-operatively, data collected included: gender, age, BMI, smoking history, uni- or bilateral procedures, pre-operative location (either Orthopaedics surgeon clinic or Specialist nurse and physiotherapist clinics), operative location (either National Health Service acute hospital or Independent hospital), pre-operative OKS, clinical and functional AKS and SF-12. At 1 year, data was collected by the institute’s UKR review programme. Data collected included: length of hospital stay, OKS, the clinical AKS, SF-12, knee range of motion, and maximum and minimum VAS pain scores.

Since the data collected was retrospectively analysed as part of an ongoing hospital audit, ethical approval was not sought.

Analysis

Results

Two hundred and twenty three unicompartmental knee replacements were reviewed, including 34 bilateral procedures between 1st January 2004 and 31st December 2006 in two hospitals. The demographic details and length of stay and 1 year outcome data is presented (Table 1 to 4).

Age

Gender

Procedure

There was no statistically significant difference at 1 year for any outcome measure between those subjects who had unilateral, bilateral, or TKR and UKR procedures during the same operation. There was a difference in length of stay where bilateral procedures stayed longer in hospital than unilateral procedures (p<0.001).

BMI

Pre-operative OKS

Pre-operative clinical AKS

Pre-operative functional AKS

Pre-operative SF-12

Pre-operative Location

Operative location

Smoking incidence

Discussion

The objective of this study was to determine whether gender, age, BMI, smoking history, uni- or bilateral procedures, pre-operative location, operative location, pre-operative OKS, clinical and functional AKS or SF-12 scores correlated to length of stay or to post-operative outcomes at 1 year following medial UKR. The findings of this study suggested that a number of pre-operative characteristics and assessment scores may be associated with 1 year outcomes and length of stay.

The results of this study indicated that patient’s BMI influenced pain scores only. Kort et al [₁₈] reported little correlation between outcome and BMI and noted that BMI only related to range of motion. There was no correlation between BMI pre-operatively and post-operative scores in total knee replacement populations in Benjamin et al’s paper [₂₅]. However, this finding was not observed in studies by Amin et al [₁₁] or Franklin et al [₂₆] which noted that in total knee replacement populations functional outcomes differed in those with a higher than lower BMI patients when assessed with functional measurements. With an aging population, a rapid rise in average body-weight and the prevalence of obesity [₂₇], clinicians may encounter a greater proportion of morbidly obese patients requiring knee arthroplasty [₁₁]. The findings of Amin et al’s [₁₁] total knee replacement study suggested that patients should be advised to lose weight pre-operatively as a prerequisite for surgery, in order to optimise post-operative range of motion. We did not assess the revision or complication rate of our cohort at 1 year. Further study may be indicated to assess whether BMI or patient weight was related to post-operative complications such as implant loosening, or technical difficulties. This may be of particular importance given Kort et al’s [₁₈] statement that obesity can cause technical difficulties in UKR.

Smoking was not shown to be associated with post-operative outcome for any outcome measure or length of stay in this study. This finding is contrary to Møller et al’s [₁₂] findings which suggested that smoking was the single most important risk factors for the development of post-operative complications in total knee replacements. Our cohort indicated that based on 1 year post-operative outcomes, smoking may not necessarily be an important factor when estimating outcome in UKR populations.

Gender was shown to be associated with 1 year OKS and maximum VAS pain scores in this UKR cohort. Such limited association between gender and 1 year results have been reported in total knee replacement data, to suggest that gender is not a prognostic indicator for post-operative outcomes [₂₈]. Although the literature on total knee replacements has reported that post-operative pain may be greater for women, this was largely not reflected in UKR patients [₂₉]. Similarly, there was no difference seen in length of stay between genders in this UKR cohort. This finding was comparable with Smith et al’s [₃₀] cohort of total knee replacement and Forrest et al’s [₁₄] hip and knee arthroplasty population. However, this result is contrary to that of Bozic et al [₁₅], Liebergall et al [₃₁], Vincent et al [₃₂], Watkins et al [₃₃] and Hayes et al’s [₃₄] findings, which suggested that hospital length of stay was greater for women compared to men, following joint arthroplasty surgery.

Age has been purported to be one of the most important factors in the decision-making process for UKR [₃₅]. This case series suggested that with the exception of maximum VAS pain scores, there was no significant difference in 1 year outcomes for pre-operative assessments with age. This was in agreement with Price et al’s [₁₉] study for knee range of motion at 10 years following UKR. There was no correlation between age pre-operative and post-operative scores in total knee replacement populations [₂₅]. Age was shown to correlate with length of stay. This was supported by previous hip and knee arthroplasty case series [_1415303236]. Given that authors have suggested that the long-term survivorship of the Oxford UKR has given surgeons increased confidence to use this prosthesis in younger generations of patients [₃₇], it is important that results such as this, suggesting that age does not influence outcome at 1 year is important. In total knee replacement populations, age seems to be associated with post-operative complications and implant rates which is not reflected in unicompartmental cases [₁₆].

One year outcomes did not differ between those patients who underwent unilateral to bilateral procedures in this cohort. Similar results were noted by Sofat et al [₃₈] who reported no significant difference between the type of procedure and post-operative OKS values. The only difference being that those who underwent unilateral procedures found it less difficult to kneel than those with bilateral procedures.

Rees et al [₃₉] acknowledged that surgeon’s experience of UKR had an important influence on post-operative outcome. This variable was not assessed in our study, and it is suggested that further assessment of this factor is evaluated with a longer follow-up period that Rees et al’s [₃₉] one year period, to determine whether revision rates are influenced by surgical experience.

Pre-operative location and operative location were not associated with length of stay or post-operative outcome; with the exception of maximum VAS pain and clinical AKS values respectively. Similarly, the findings of this study suggested that pre-operative assessment of OKS, AKS, and most notably as Table 4 suggests, pre-operative SF-12 scores would be important variables to influence outcomes at 1 year. Further study is now required to compare whether these outcomes correlate to post-operative results in differing populations. This may be of particular note given that variations were noted between outcomes in total knee replacement cohorts when comparing patients from the United Kingdom, North America and Australia. Accordingly, an assessment of pre-operative data compared to post-operative outcomes may be warranted from additional countries to this UK population.

This paper presented some methodological limitations. Firstly, although the study followed patients for a minimum of one year, further evaluation is indicated to evaluate outcomes over a longer period in order to assess the effect of demographic variables on prosthesis wear and revision rates. It is our intention to re-evaluate this cohort at five years. Due to the retrospective nature of this case series, we did not assess our prognostic indicators against matched groups for each variable; this may have improved methodological rigor by controlling other variables to solely examine the effects of each variable on outcome. Further study is suggested to undertake such a design, to further develop the evidence-base on this topic. In addition, this study only assessed medial Oxford UKR. Further study is recommended to determine whether the findings of this study are transferable to patients following a lateral UKR, or other prosthesis designs such as to compare those of mobile or fixed bearings implants.

Conclusion

The findings of this study suggest that most pre-operative variables assessed correlated to maximum pain outcomes, and that the pre-operative clinical and functional AKS and SF-12 scores correlated with 1 year OKS and clinical AKS. Length of stay correlated with pre-operative OKS, SF-12, functional AKS, age and whether patients underwent uni- or bilateral procedures. Further study is recommended to evaluate this study’s findings to UKR populations in differing centres using mobile and fixed implants and of different surgical and post-operative regimes.

Acknowledgements

Correspondence to

Mr T O Smith Institute of Orthopaedics – Centre Block Norfolk and Norwich University Hospital Colney Lane Norwich Norfolk NR4 7UY Telephone Number 01603 286286 Email Address toby.smith@nnuh.nhs.uk