Why are patients dissatisfied following a total knee replacement? A systematic review

Background Although total knee replacement (TKR) is an effective intervention for end-stage arthritis of the knee, a significant number of patients remain dissatisfied following this procedure. Our aim was to identify and assess the factors affecting patient satisfaction following a TKR. Materials and methods In accordance with the PRISMA guidelines, two reviewers searched the online databases for literature describing factors affecting patient satisfaction following a TKR. The research question and eligibility criteria were established a priori. Any clinical outcome study that described factors relating to overall satisfaction after primary TKR was included. Quality assessment for the included studies was performed by two accredited orthopaedic surgeons experienced in clinical research. Results The systematic review identified 181 relevant articles in total. A history of mental health problems was the most frequently reported factor affecting patient satisfaction (13 reportings). When the results of the quality assessment were taken into consideration, a negative history of mental health problems, use of a mobile-bearing insert, patellar resurfacing, severe pre-operative radiological degenerative change, negative history of low back pain, no/less post-operative pain, good post-operative physical function and pre-operative expectations being met were considered to be important factors leading to better patient satisfaction following a TKR. Conclusion Surgeons performing a TKR should take these factors into consideration prior to deciding whether a patient is suitable for a TKR. Secondarily, a detailed explanation of these factors should form part of the process of informed consent to achieve better patient satisfaction following TKR. There is a great need for a unified approach to assessing satisfaction following a TKR and also the time at which satisfaction is assessed. Electronic supplementary material The online version of this article (10.1007/s00264-020-04607-9) contains supplementary material, which is available to authorized users.


Introduction
Total knee replacement (TKR) is one of the most effective surgical interventions for relief of pain and functional recovery in patients with advanced osteoarthritis (OA) of the knee. Management of OA costs the UK economy equivalent to 1% of its gross national product per year [1]. In the USA, the annual number of TKRs has been projected to rise by over 670% to 3.48 million cases by 2030 [2]. Outcomes of TKR are traditionally assessed by survival analysis with revision as the end point, and technical outcomes of this intervention are excellent. According to the UK National Joint Registry (NJR) annual report, the survival rate has been reported to be over 99.5% after one year and 95.6% at ten years [3].
A revision TKR is most commonly performed for loosening, fracture or infection. However, survival analysis tends to underestimate poor function, pain or dissatisfaction because these problems do not necessarily lead to a revision and are not recorded in the registry. Another issue is that reporting of the outcome of a TKR has predominantly been based on surgeon-derived outcome measures, which include range of movement (ROM), joint stability and post-operative alignment [4][5][6]. However, a report identified a poor correlation between surgeon-derived and patient-reported outcomes, with surgeons overestimating outcomes in comparison with the patients' [7]. This correlates well with the fact that a significant number of patients experience continual pain and functional disability and therefore remain dissatisfied following the procedure [8][9][10].
In the largest ever reported series on satisfaction following a TKR, which included a survey of 27,372 patients, 17% of the unrevised patients were either dissatisfied or uncertain regarding their outcome [11]. Baker et al. [12] also reviewed the data from Table 1 Inclusion and exclusion criteria applied to articles identified in the literature Inclusion criteria 1. All levels of evidence 2. Written in the English language 3. Studies on humans 4. Studies reporting factors affecting overall satisfaction and/or dissatisfaction following a primary total knee replacement 5. Operative procedure consisted solely of total knee replacement 6. Total knee replacement irrespective of any pathology Exclusion criteria 1. Studies whose results included other procedures 2. Studies reporting satisfaction/dissatisfaction for only a small part of the procedure (e.g. 'satisfaction in either pain control, skin closure, range of motion, nursing quality, anaesthesia, nerve block or physiotherapy' was excluded) 3. Studies not reporting patient's satisfaction (e.g. 'studies on family's or carer's satisfaction' were excluded) 4. Studies describing trial protocols without any results 5. Studies with follow-up period of 3 months or less 6. Revision total knee replacement 7. Unicompartmental knee replacement 8. Patellofemoral knee replacement 9. Cadaveric or radiological studies 10. Reviews, systematic reviews the NJR in the UK and reported that 71% of the patients experienced improvement of knee symptoms, but only 22% of them rated the results as excellent. Therefore, although the surgeonreported outcomes may be good and the patient has no indication for a revision, they may still be dissatisfied following their index TKR. This may be due to a multitude of reasons, but to the best of our knowledge, there has been no systematic review which has specifically focused on the factors that affect patient satisfaction following a TKR. The aim of this systematic review, therefore, was to identify and assess the factors affecting patient satisfaction following a TKR.

Methods
The protocol of this systematic review was developed and has been registered in the International Prospective Register of S y s t e m a t i c R e v i e w s ( P R O S P E R O 2 0 1 7 CRD42017084659). The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were used for designing this study [13].

Search strategy
Two accredited orthopaedic surgeons experienced in clinical research searched the online database Medline, Embase, BNI, AMED, Cochrane and Google Scholar for literature relating to satisfaction following a TKR. The PICO (participants, interventions, comparators, outcomes) tool was adopted and modified to formulate the research question and establish the inclusion and exclusion criteria. Selected articles were then exported to Mendeley reference manager software to organise screen and select articles.

Study screening and selection
Clinical outcome studies that described the factors relating to the overall or general satisfaction/dissatisfaction following a primary TKR irrespective of any pathology were included. The inclusion and exclusion criteria are described in Table 1. Any discrepancies at the title and abstract revision stage were resolved by automatic inclusion to ensure thoroughness. Any discrepancies at the full-text stage were resolved by consensus between the two reviewers. If a consensus could not be reached, a third, more senior reviewer was consulted to resolve the discrepancy.

Data extraction and analysis
The two reviewers independently extracted relevant study data from the final pool of included articles and recorded this data on a spreadsheet designed a priori in Microsoft Excel 2013 (Microsoft Corporation, Redmond, WA, USA). The quality of studies including bias was then analysed and               for cohort studies, case-control studies, cross-sectional studies and case series [14]. For RCTs, a modified version of critical appraisal checklist by van Tulder et al. was used [15].

Statistical methods
Statistical analysis in this study focused on descriptive statistics. After assessing the quality of each study, the score was converted into a percentage from the full score (%), which was then considered to be the 'strength' of that particular study. Microsoft Excel 2013 was used for our analysis in reporting the factors affecting patient satisfaction following a TKR, based on the strength of studies as per the type of evidence. The potential factors were then categorised into seven groups designed from the findings of the studies included. The strength of each factor was presented, regardless of whether it was a FACTOR ('it is a factor for patient satisfaction') or a Not-FACTOR ('it is a factor which does NOT relate to patient satisfaction'-in other words, 'researcher X found Factor Z was irrelevant to patient satisfaction'). Details are described in Electronic Supplementary Material 1 and Table 2.

Results
A total of 5635 articles were found following the initial search of the electronic databases and citation tracking, followed by removing 2424 duplicate articles. After review by title and abstract, 2977 articles were excluded and 234 potential articles remained for a full-text review. After application of the inclusion and exclusion criteria, a further 53 articles were discarded, leaving 181 relevant articles for the final inclusion, analysis and assessment. The study finally included 40 RCTs (22.1%), 93 cohort studies (51.4%), nine case-control studies (5.0%), 37 cross-sectional studies (20.4%) and 2 case series (1.1%) (Electronic Supplementary Material 2). Flowchart for the review is shown in Fig. 1 and the details of all the 181 studies are shown in Table 3. A total of 22 authors were found to have written several papers. To ensure that duplicate numbers were not included in our analysis, we contacted all these authors and reminder emails were sent as well to ensure a reply. Only five authors replied back with no overlap in their studies, three authors said that there was an overlap and 14 did not reply back. Those who did not reply back were treated as if it was an overlap and, thus, not considered. Due to the lack of homogeneity between studies, a meta-analysis was deemed unsuitable for this study.
The quality of all the 181 studies was assessed and the results are shown in Tables 6,7,8,9 and 10. The strength of each factor was described using the sum of percentage in each type of study (RCT, cohort study, case-control study, crosssectional study and case series) (Fig. 3). RCTs were considered to be the strongest (deep colour in Fig. 3) and this was followed by cohort study, case-control study and crosssectional study, respectively. Case series was considered to be the weakest (light colour in Fig. 3).
When the results of the quality assessment were taken into consideration, a negative history of mental health problems, use of a mobile-bearing insert, patellar resurfacing, severe preoperative radiological degenerative change, negative history of low back pain, no/less post-operative pain, good postoperative physical function and pre-operative expectations being met were considered to be important factors. Significant factors affecting patient satisfaction are summarised in Tables 11, 12 and 13.

Discussion
The dissatisfaction rate following a TKR remains around 20% and is a constant source of frustration for the patient and the surgeon [11,12]. Our study has systematically reviewed all the articles looking at satisfaction following a TKR to determine the factors, which could be responsible for this issue. Several factors were deemed to be important in affecting patient satisfaction based on the number of studies in which they were reported as well as the results of the quality assessment of the study (Tables 11, 12 and 13).

Negative history of mental health problems
A negative history of mental health problems was the most frequently reported factor affecting patient satisfaction (Table 11) and also scored the highest sum of percentage of FACTOR based on the quality assessment for RCT + cohort study (± case-control study ± cross-sectional study ± case series study) (Table 12). In addition, it was ranked first in terms of the highest sum of percentage of FACTOR and Not-FACTOR based on the quality assessment for all types of the studies (Table 13). Depressive symptoms and anxiety were reported to be predictive of long-term pain and functional impairment as measured by the Knee Society Score in 83 patients at 5 years [16]. In addition, it was reported that preoperative anxiety/depression is an independent risk for severe post-operative pain and may explain as to why there is a subset of patients with unexplained pain after surgery [17]. Moreover, Macleod et al. report that patients with mental disability suffered a greater level of comorbidity and were socially deprived, which is also related to poorer physical health which then has an impact on satisfaction [18]. Finally, another study reported that patients with poor mental health, which can impair coping mechanisms for pain, might present with Fig. 2 Number of reportings in seven groups of factors for patient satisfaction following total knee replacement. Blue bar means FACTOR ('it is a factor for patients' satisfaction') and orange bar means Not-FACTOR ('it is a factor which does NOT relate to patients' satisfaction') less severe disease, and this could also influence their satisfaction [19].

Use of a mobile-bearing insert
The use of a mobile-bearing insert had the highest sum of percentage of FACTOR based on the quality assessment for RCTs. Also, it had the second highest sum of percentage of FACTOR based on the quality assessment for RCT + cohort study (± case-control study) ( Table 12). The rationale behind the design of a mobile-bearing insert is to solve the kinematic conflict between low-stress articulation and free axial femoral-tibial rotation by allowing rotation of a highly conforming polyethylene insert [20]. Theoretically, the design of the mobile-bearing insert could lead to better ROM especially during flexion [21]. A greater loss of flexion was reported after 12 months in patients with a TKR with a fixed-bearing prosthesis in comparison with a mobile-bearing prosthesis [22]. It is quite intuitive to comprehend that a good post-operative ROM relates to patient satisfaction, and our results support this (improvement in ROM was the 4th most frequently reported factor for patient satisfaction). Kim et al. suspect the low constraint of mobile-bearing insert may restore normal kinematics of the knee and it contributes to favourable clinical outcomes compared with a fixed-bearing insert [23]. Price et al. in a prospective multicentre trial of 39 simultaneous bilateral procedures also found that patients with a mobile-bearing insert had significantly better clinical results than patients with a fixed-bearing insert [21].

Patellar resurfacing
Patellar resurfacing has the second highest sum of percentage of FACTOR based on the quality assessment for RCTs (Table 12). Four studies showed patients with patella resurfacing were more satisfied than those without it [11,[24][25][26]. Amongst them, one study focused on only knees with no exposed bone on the undersurface of the patella to determine the potential advantages of leaving the patella non-resurfaced [25]. Dissatisfaction in patella non-resurfaced patients may be due to the higher rate of post-operative anterior knee pain, and patients whose patella was not resurfaced at the index TKR tended to have a higher revision rate as well [25][26][27][28]. However, it should be noted that this issue may be strongly related to the design of the implant. There have also been abundant literature that showed that the patellofemoral design in TKR is critical and can vary the forces on the patellofemoral joint as well as patellofemoral tracking [29][30][31]. Two of the 4 studies relate to a specific prosthesis (PFC) which is notoriously patella unfriendly [25,26], so this relationship may therefore not necessarily hold true for the newer implants with patella-friendly designs.
Severe pre-operative radiological degenerative change Severe pre-operative radiological degenerative change has the fifth highest sum of percentage of FACTOR and Not-FACTOR based on the quality assessment for all types of studies (Table 13). Although the classic indication for replacing a patient's knee is end-stage arthritis (Kellgren-Lawrence grade IV [32]), there are a number of patients who have a TKR much before grade IV radiological changes have set in and it is dependent on the symptoms of the patient. The individual indication is complex and involves multiple factors [33]. Studies are described using serial numbers in Table 3. The number of studies in each group is shown in parentheses

No low back pain
No low back pain has the sixth highest sum of percentage of FACTOR and Not-FACTOR based on the quality assessment for all types of the studies (Table 13). The prevalence of chronic low back pain in the UK has been reported to range from 6 to 11% [29], and this is increased to 55% in patients with OA of the knee [30]. Furthermore, low back pain has been demonstrated to be three to four times more likely to be present in patients with a history of depression [37]. Also, patients with chronic low back pain have a higher rate of musculoskeletal and neuropathic pain conditions, depression, anxiety and sleep disorders [31]. In addition, patients with low back pain reported to have more symptoms from their osteoarthritic knee which may suggest a lower threshold for pain in this cohort leading to dissatisfaction [30].

Normal BMI
Normal BMI was the fifth most frequently reported factor for patient satisfaction (Table 11). BMI greater than 30 kg/m 2 was reported to be associated with a higher rate of revision and poorer functional outcomes as well which again contributes to dissatisfaction [38]. In addition, morbidly obese patients are likely to suffer from wound problems, ligament injuries and infections peri-operatively which lead to dissatisfaction [22]. Another study showed that despite lower pre-and post-operative WOMAC and SF-36 scores, obese patients experienced similar improvements compared with non-obese patients, although levels of satisfaction in the obese group were lower than those in the non-obese group [39]. The authors stated that one explanation for this might be that satisfaction was more closely associated with the absolute post-operative functional level rather than the magnitude of any improvement, as the rate of satisfaction mirrored absolute values of post-operative WOMAC and SF-36 scores.

Other factors
Other than factors discussed in the previous section, no/less post-operative pain, good post-operative physical function, improvement in ROM and pre-operative expectations being met were considered to be important for patient satisfaction based on the number of reportings and the results of quality assessment (Tables 11, 12 and 13). TKR is a painful procedure and it does take at least six to 12 months to get the maximum benefit from this procedure [40], and therefore, setting realistic expectations with the patient in the pre-operative clinic is essential to avoid dissatisfaction.

Limitations and strengths of the study
Our study has several limitations. Firstly, the method of measuring satisfaction is different in each study, and therefore, a  Fig. 3 Sum of percentage from full score (%) based on the quality assessment in each type of study for each factor. Blue bar means FACTOR ('it is a factor for patients' satisfaction') and orange bar means Not-FACTOR ('it is a factor which does NOT relate to patients' satisfaction'). a Patients' demographical factors. b Patients' non-knee factors. c Patients' knee factors. d Factors of implants/prostheses. e Intra-operative technical factors. f Post-operative outcome factors. g Surgeon and healthcare factors c d e Fig. 3 (continued) uniform way of assessing satisfaction is essential for the orthopaedic community. Secondly, the timing of assessment of satisfaction after the index TKR varied amongst studies and this again requires standardisation. Thirdly, in many of the studies included in this review, the authors have only focused on one factor and the mutual or overall effect of multiple factors was not assessed. Fourthly, no statistical tests of intra-class correlation coefficients, inter-rater reliability and heterogeneity amongst the studies were performed in this systematic review. Finally, there are several studies in which patients are duplicated amongst studies and our review was Table 11 List of frequently reported factors as FACTOR ('it is a factor for patient satisfaction') Factors (number of reportings) 1st place No mental health problems (13 reportings) 2nd place No/less post-operative pain (7 reportings) 2nd place Good post-operative physical function (7 reportings) 4th place Improvement in ROM (6 reportings) 5th place Normal BMI (5 reportings) 5th place Pre-operative expectations met (5 reportings) BMI body mass index, ROM range of motion f g Fig. 3 (continued) limited to publications in English, so there is a possibility of publication bias. However, despite all these limitations, the main strength of this study lies in its broad and comprehensive initial literature search as well as complete and in-depth quality assessment for each study and the factors. We have determined all the factors which could potentially affect patient satisfaction following a TKR which have been reported in the literature thus far.

Conclusion
No history of mental health problems, use of a mobile bearing insert, patellar resurfacing, severe pre-operative radiological degenerative change, no low back pain, normal BMI, no/less post-operative pain, good physical function post-operatively, improvement in ROM and pre-operative expectations being met were considered to be significant factors leading to better patient satisfaction following a TKR.
Surgeons performing a TKR should take these factors into consideration prior to deciding whether a patient is suitable for a TKR. Secondarily, a detailed explanation of these factors should form part of the process of informed consent to achieve better patient satisfaction following TKR.
There is great need for a unified approach to assessing satisfaction following a TKR and also the time at which satisfaction is assessed.
Moreover, further studies and ideally larger RCTs focusing on each of these factors are required to determine the exact correlation of these factors with satisfaction.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. RCT randomised controlled trial Table 13 List of factors which have the highest sum of percentage score (a percentage from full score) of FACTOR ('it is a factor for patient satisfaction') and Not-FACTOR ('it is a factor which does NOT relate to patient satisfaction') based on the quality assessment for all type of the studies Percentage score of Not-FACTOR was calculated as negative value