Effectiveness of postdischarge interventions for reducing the severity of chronic pain after total knee replacement: systematic review of randomised controlled trials

Objective Approximately 20% of patients experience chronic pain after total knee replacement (TKR). The aim of this systematic review was to evaluate the effectiveness of postdischarge interventions commenced in the first 3 months after surgery in reducing the severity of chronic pain after TKR. Design The protocol for this systematic review was registered on PROSPERO (registration number: CRD42017041382). MEDLINE, Embase, CINAHL, PsycINFO and The Cochrane Library were searched from inception to November 2016. Randomised controlled trials of postdischarge intervention which commenced in the first 3 months after TKR surgery were included. The primary outcome of the review was self-reported pain severity at 12 months or longer after TKR. Risk of bias was assessed using the Cochrane risk-of-bias tool. Results Seventeen trials with data from 2485 randomised participants were included. The majority of trials evaluated physiotherapy interventions (n=13); other interventions included nurse-led interventions (n=2), neuromuscular electrical stimulation (n=1) and a multidisciplinary intervention (n=1). Opportunities for meta-analysis were limited by heterogeneity. No study found a difference in long-term pain severity between trial arms, with the exception of one trial which found home-based functional exercises aimed at managing kinesiophobia resulted in lower pain severity scores at 12 months postoperatively compared with advice to stay active. Conclusion This systematic review and narrative synthesis found no evidence that one type of physiotherapy intervention is more effective than another at reducing the severity of chronic pain after TKR. Further research is needed to evaluate non-physiotherapy interventions, including the provision of care as part of a stratified and multidisciplinary care package. PROSPERO registration number CRD42017041382.

This systematic review found evidence that post discharge physiotherapy interventions delivered uniformly to all patients in the first three months after TKR do not appear to be effective at reducing the severity of chronic pain. Further research is needed to evaluate whether stratified physiotherapy care and multidisciplinary interventions can reduce the severity of chronic pain after TKR.
( Total knee replacement, chronic post surgical pain, prevention, systematic review This is the first systematic review to evaluate the effectiveness of post discharge interventions delivered in the first three months after surgery in reducing the severity of chronic pain after total knee replacement.
Synthesis of adverse events data was not possible because assessment and reporting was variable and often poor.
We did not include studies that used a composite pain and function measure to assess pain outcome.

)* #$'+ )$*
Total knee replacement (TKR) is a common operation to provide pain relief, predominately due to osteoarthritis. Despite good outcomes for many, some patients report chronic pain in the months and years after TKR. Chronic post surgical pain is defined as pain that is present or increases in intensity at ≥3 months after surgery [1]. In representative populations, unfavourable long term pain outcomes have been reported by 10 34% of patients with TKR [2]. Patients with bothersome pain at ≥3 months after surgery are disappointed with their outcome [3 4]. Given the prevalence and impact of chronic pain, it is important to evaluate interventions that may optimise patients' outcomes after TKR.
During the hospital stay after TKR, rehabilitation focuses on regaining range of motion and improving mobility. After discharge, rehabilitation aims to enhance recovery, through supporting a person to regain function and quality of life, optimising pain relief and re integration into social and personal environments [5]. While physiotherapy often focusses on functional health, another key outcome is the prevention of long term pain [6]. Post operative physiotherapy may be combined with other interventions to provide comprehensive, multidisciplinary and holistic rehabilitation [7]. Therefore, a key step to improving patients' outcomes after TKR is to evaluate if early post operative rehabilitation interventions can reduce the severity of chronic pain after TKR.
The aim of this systematic review was to evaluate the effectiveness of post discharge interventions delivered in the first three months after surgery for reducing the severity of chronic pain after TKR.

, -$'
The review was registered on the international prospective register of systematic reviews (PROSPERO) on 17 th January 2017 (registration number CRD42017041382). The review was conducted following guidance from the Cochrane Handbook [8] and reported in accordance with PRISMA guidelines [9] (Appendix 1).

%
Studies were eligible for inclusion in the review if they met the following criteria: : Adults discharged from hospital after primary TKR predominantly for osteoarthritis.
: Any post discharge intervention which commenced in the first three months after TKR surgery.
: Any, including no intervention, usual care, placebo or an alternative intervention.
The primary outcome was pain severity at 12 months or longer after TKR, as patient reported levels of pain plateau by this time point [10 11]. Pain severity could be assessed using a patient reported joint specific pain measure (e.g. WOMAC or KOOS pain domains), a quality of life measure (e.g. SF 36 or SF 12) or a Visual Analogue Scale (VAS).
The secondary outcome was serious adverse events.
: Randomised controlled trials (RCTs). only as abstracts or that were unobtainable as full text copies using inter library loans or email contact with authors were excluded. Citations of key reviews and studies were checked in ISI Web of Science.
Records identified by searches were imported into Endnote X7 (Thomson Reuters) and duplicates removed. From the searches, an Endnote database of all RCTs and systematic reviews in TKR was established. Within this database, interventions conducted during the post operative period were identified. An initial screen for potential eligibility was undertaken by one reviewer (ADB) to exclude articles that were clearly not relevant.
Subsequently, abstracts and full text articles were screened independently by two reviewers (VW and ADB or JD) and reasons for exclusion recorded.

' .
Data from studies that met the eligibility criteria were extracted onto a standardised proforma by one reviewer (VW). Data extraction was checked against source articles by a second reviewer (JD). Extracted data comprised: country, date, participant characteristics, selection criteria; intervention and control treatment; follow up intervals; losses to follow up; outcome data for pain (means and standard deviations or medians and ranges); serious adverse events and information for risk of bias assessment. Any disagreements between reviewers were discussed with a third reviewer (ADB) and consensus reached.
A single e mail was sent to authors of studies with an appropriate follow up period but no pain outcome to enquire if an appropriate outcome was available. If a combined pain and function outcome was reported, such as the OKS or the total WOMAC score, separate pain subscale data were requested. Authors were contacted when necessary for clarification purposes or to request unpublished relevant data. If a study reported data that were combined  Table 1 provides an overview of study characteristics. Included studies were from Australia (n=3), Canada (n=2), Finland (n=2), Germany (n=2), United Kingdom (n=2), China (n=1), Denmark (n=1), Italy (n=1), Norway (n=1) and United States of America (n=1). The number of centres was reported for 14 studies: seven studies were conducted in a single centre, three studies were conducted in two centres, and four studies were conducted in ≥4 centres. Sample sizes ranged from 47 to 422 participants, with a median of 138. All studies had two arms, with the exception of one three arm trial [19]. Three studies were described as pilot or feasibility studies [15 17 23].

0
Risk of bias assessments for individual studies are displayed in Table 2. All studies were at high risk of bias for blinding of participants and pain outcome assessment due to the nature of the intervention and the self reporting of pain. Six studies were at high risk of bias due to incomplete outcome data. Possible risk of selective outcome reporting was identified for two studies. The majority of studies evaluated physiotherapy interventions (n=12); other interventions evaluated included nurse led interventions (n=2), neuromuscular electrical stimulation (NMES) (n=1) and a multidisciplinary intervention (n=1).

1
Twelve studies with 1,846 randomised patients evaluated the effectiveness of post discharge physiotherapy interventions. There was considerable variation in the interventions evaluated.
Seven studies compared physiotherapy interventions with usual or minimal care; these interventions included a walking skills programme [13], group based circuit exercise classes [16], erogometer cycling [21], home based functional rehabilitation [23], clinic based functional rehabilitation [24], home based functional exercises aimed at managing kinesiophobia [25], and delayed monitored home exercises [28]. Five studies compare two forms of treatment including inpatient rehabilitation compared with home exercise [14], home based functional exercise and home based traditional exercise [17], 1:1 physiotherapy and home based rehabilitation [20], early aquatic therapy and late aquatic therapy [22], and a three arm trial comparing 1:1 physiotherapy, group based circuit classes and a monitored home exercise programme [19]. All interventions started within two months of surgery, with the majority commencing within two weeks of surgery. Of the 12 studies, only one trial reported a difference in pain outcomes between groups; patients randomised to home based exercises aimed at managing kinesiophobia had lower pain scores at 12 months post operative compared with patients randomised to usual care [25].
Meta analysis was conducted with six studies that reported relevant pain outcome data to compare the effectiveness of physiotherapy interventions compared with usual care for reducing the severity of chronic pain after TKR [13 16   outcomes were observed between treatments, with a standardised mean difference of 0.14 (95% CI 0.41, 0.13; I 2 69%). Similar results were obtained in sensitivity analysis excluding two studies at high risk of bias from incomplete outcome data [21 24] (Appendix 5).
Due to the small number of studies, subgroup analyses to explore the effectiveness of different intervention content and intensity, and different comparator interventions, were not possible.

*
Two studies with 319 randomised patients reported evaluation of a nurse led intervention compared with no care or usual care. Except for issues relating to blinding, both studies were at low risk of bias. Both studies evaluated nurse led structured telephone follow up; one aimed to improve adherence to home exercise [15] and the other to provide information regarding well being, integrity, prophylaxis, safety and other issues relevant to patients after TKR [27]. Pain outcome data (mean and standard deviations) were not available for latter study and therefore meta analysis was not possible. Neither study found a difference in pain scores at 12 months post operative between the intervention and control group.

$
Two studies reported evaluations of other interventions. Except for issues relating to blinding, both studies were at low risk of bias. One trial involving 86 patients compared a group based multidisciplinary programme with usual care [18]. This 10 day programme involved physiotherapy, Nordic walking, relaxation strategies, and sessions with a psychologist, social   There are a number of strengths and limitations to this systematic review. The main outcome of interest in this review was pain severity at ≥12 months after TKR. Although the primary outcome of many of the included studies was function, pain severity was an important secondary outcome in these studies. Studies that used a composite pain and function measure to assess outcome, for example the OKS or WOMAC, were excluded if authors were unable to provide pain subscales scores. Although this reduced the number of studies eligible for inclusion, this approach was taken because pain and function are distinct outcome domains, with different predictors and recovery trajectories [34 35]. The secondary outcome of this review was adverse events, to allow the synthesis of harms data. However, synthesis was not possible because assessment and reporting of adverse events was variable and often poor. The quality of adverse events reporting is a common issue in surgical trials [36], and evidence based recommendations are needed to promote standardisation, improve quality and reduce heterogeneity of adverse events reporting in orthopaedic studies. A potential limitation of the included studies was that they were all at high risk of bias due to the lack of participant blinding for self report pain. However, blinding of participants is rarely possible in RCTs of this nature. Also, it would be expected that the risk would arise from participants in the intervention group reporting less pain, which may potentially be an issue with shorter term outcomes, but this was not evident from the longer term follow up of the studies included in this review.
This systematic review took a broad approach by evaluating the effectiveness of any type of post discharge intervention that aimed to reduce the severity of chronic pain after TKR.
( Total knee replacement, chronic post surgical pain, prevention, systematic review This is the first systematic review to evaluate the effectiveness of post discharge interventions delivered in the first three months after surgery in reducing the severity of chronic pain after total knee replacement.
Synthesis of adverse events data was not possible because assessment and reporting was variable and often poor.
We did not include studies that used a composite pain and function measure to assess pain outcome.

)* #$'+ )$*
Total knee replacement (TKR) is a common operation to provide pain relief, predominately due to osteoarthritis. Despite good outcomes for many, some patients report chronic pain in the months and years after TKR. Chronic post surgical pain is defined as pain that is present or increases in intensity at ≥3 months after surgery [1]. In representative populations, unfavourable long term pain outcomes have been reported by 10 34% of patients with TKR [2]. Patients with bothersome pain at ≥3 months after surgery are disappointed with their outcome [3 4]. Given the prevalence and impact of chronic pain, it is important to evaluate interventions that may optimise patients' outcomes after TKR.
During the hospital stay after TKR, rehabilitation focuses on regaining range of motion and improving mobility. After discharge, rehabilitation aims to enhance recovery, through supporting a person to regain function and quality of life, optimising pain relief and re integration into social and personal environments [5]. While physiotherapy often focusses on functional health, another key outcome is the prevention of long term pain [6]. Post operative physiotherapy may be combined with other interventions to provide comprehensive, multidisciplinary and holistic rehabilitation [7]. A key step to improving patients' outcomes after TKR is to evaluate if early post operative rehabilitation interventions can reduce the severity of chronic pain after TKR. Chronic pain is difficult to treat once established [8], and therefore it is important to evaluate the effectiveness of early post operative interventions in reducing the severity of chronic pain.
The aim of this systematic review was to evaluate the effectiveness of post discharge interventions delivered in the first three months after surgery for reducing the severity of chronic pain after TKR.

, -$'
The review was registered on the international prospective register of systematic reviews (PROSPERO) on 17 th January 2017 (registration number CRD42017041382). The review was conducted following guidance from the Cochrane Handbook [9] and reported in accordance with PRISMA guidelines [10] (Appendix 1).

%
Studies were eligible for inclusion in the review if they met the following criteria: : Adults discharged from hospital after primary TKR predominantly for osteoarthritis.
: Any post discharge intervention which commenced in the first three months after TKR surgery.
: Any, including no intervention, usual care, placebo or an alternative intervention.
The primary outcome was pain severity at 12 months or longer after TKR, as patient reported levels of pain plateau by this time point [11 12]. Pain severity could be assessed using a patient reported joint specific pain measure (e.g. WOMAC or KOOS pain domains), a quality of life measure (e.g. SF 36 or SF 12) or a Visual Analogue Scale (VAS).
The secondary outcome was adverse events.
: Randomised controlled trials (RCTs). only as abstracts or that were unobtainable as full text copies using inter library loans or email contact with authors were excluded. Citations of key reviews and studies were checked in ISI Web of Science.
Records identified by searches were imported into Endnote X7 (Thomson Reuters) and duplicates removed. From the searches, an Endnote database of all RCTs and systematic reviews in TKR was established. Within this database, interventions conducted during the post operative period were identified. An initial screen for potential eligibility was undertaken by one reviewer (ADB) to exclude articles that were clearly not relevant.
Subsequently, abstracts and full text articles were screened independently by two reviewers (VW and ADB or JD). Results of screening were compared, and any discrepancies were resolved through further review of the full text articles and discussion between reviewers.
Reasons for exclusion were recorded.

' .
Data from studies that met the eligibility criteria were extracted onto a standardised proforma by one reviewer (VW). Data extraction was checked against source articles by a second reviewer (JD). Extracted data comprised: country, date, participant characteristics, selection criteria; intervention and control treatment; follow up intervals; losses to follow up; primary outcome; outcome data for pain; adverse events (any untoward medical occurrence in a clinical study participant regardless of the causal relationship with the study treatment) and information for risk of bias assessment. Any disagreements between reviewers were discussed with a third reviewer (ADB) and consensus reached.
A single e mail was sent to authors of studies with an appropriate follow up period but no pain outcome to enquire if an appropriate outcome was available. If a combined pain and function outcome was reported, such as the OKS or the total WOMAC score, separate pain subscale data were requested. Authors were contacted when necessary for clarification purposes or to request unpublished relevant data. If a study reported data that were combined for knee and hip replacement patients, then disaggregated data for patients with TKR were requested. If this was not available, then the study was excluded.

# %
Potential sources of bias were assessed using the Cochrane risk of bias tool [9]. At the protocol stage, analysis was planned which included all studies, with sensitivity analyses conducted to exclude studies judged to be at high risk of bias.
At the protocol stage, meta analysis using RevMan 5 [13] was planned if two or more studies were identified with similar interventions and comparator groups and appropriate outcome data. If continuous pain outcomes were measured differently across studies, overall standardised mean differences and 95% confidence intervals would be calculated and presented alongside measures of heterogeneity (I 2 ). Where possible, subgroup analyses were planned to explore the effectiveness of different intervention content and intensity, and different comparator interventions.
Opportunities for pooling outcome data in meta analysis were limited by heterogeneity. This included the content, duration and intensity of both the treatments in both the intervention and comparison group. For example, a number of the trials were pragmatic with the control group receiving 'usual care', which varied considerably between studies. Therefore, a narrative synthesis was performed.   [20]. Three studies were described as pilot or feasibility studies [16 18 24].

0
Risk of bias assessments for individual studies are displayed in Figure 2. All studies were at high risk of bias for blinding of participants and pain outcome assessment due to the nature of the intervention and the self reporting of pain. Five studies were at high risk of bias due to incomplete outcome data and one due to selective outcome reporting.

$
The primary outcome was specified for 13 trials; this was function in eight trials; a composite of pain and function in four trials, and pain in one trial (Appendix 3). Pain severity was most commonly assessed using the WOMAC pain scale (n=9); other tools included the KOOS pain   physiotherapy, group based circuit classes and a monitored home exercise programme [20].

') + )$*
This systematic review aimed to evaluate the effectiveness of post discharge interventions delivered in the first three months after surgery for reducing the severity of chronic pain after TKR. Interventions that predominately comprise physiotherapy have been evaluated in RCTs.
In most studies, the control group received some form of physiotherapy care and therefore the aim of the trials was to compare the effectiveness of different types of physiotherapy, rather than comparing the effectiveness of physiotherapy to no care. A narrative synthesis of the evidence suggests that no physiotherapy intervention appears to be more effective than another at reducing the severity of chronic pain after TKR. However, findings from the trial of a 6 month home based functional exercise programme aimed at managing kinesiophobia [26] compared to advice to stay active were encouraging and warrant further evaluation. Few studies have been conducted to evaluate the effectiveness of non physiotherapy interventions at reducing chronic pain after TKR, and further research is needed. based recommendations are needed to promote standardisation, improve quality and reduce heterogeneity of adverse events reporting in orthopaedic studies. A potential limitation of the included studies was that they were all at high risk of bias due to the lack of participant blinding for self report pain. However, blinding of participants is rarely possible in RCTs of this nature. Also, it would be expected that the risk would arise from participants in the intervention group reporting less pain, which may potentially be an issue with shorter term outcomes, but this was not evident from the longer term follow up of the studies included in this review.
This systematic review took a broad approach by evaluating the effectiveness of any type of post discharge intervention that aimed to reduce the severity of chronic pain after TKR.
Interventions that span the post operative period may be delivered as part of a comprehensive peri operative package of care, and these would not have been identified in this review; however, evaluations of the effectiveness of pre operative and peri operative interventions for reducing chronic pain severity are being conducted separately (CRD42017041382). severity was assessed as a secondary outcome in these trials and therefore it was expected that the intervention may reduce long term pain. All but one study found that the intervention did not provide any benefit on long term pain severity compared to the control group.
However, the treatment received in the control group, particularly in the physiotherapy trials, varied considerably between studies, including a different form or intensity of physiotherapy, provision of physiotherapy based on a needs assessment, delayed treatment or no treatment.
Therefore, it is not appropriate to draw conclusions on the effectiveness of any particular type

! (*$ / '3 , *
We would like to thank all the study authors who took the time to reply to our requests for further clarification or additional data.
Usual care, which involved no additional guidance from 2 months post operative.

METHODS
Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number.

5
Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale.

5-6
Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched. 6 Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated.

Appendix 2
Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis).

6-7
Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators.

6-7
Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made.

6-7
Risk of bias in individual studies 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis.

7
Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.

RESULTS
Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.

Figure 1
Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. Table 1 Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Table 2 Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.  Table 2 Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]).

DISCUSSION
Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).

12-15
Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).

12-13
Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research.

Objective
Approximately 20% of patients experience chronic pain after total knee replacement (TKR).
The aim of this systematic review was to evaluate the effectiveness of post-discharge interventions commenced in the first three months after surgery in reducing the severity of chronic pain after TKR

Design
The protocol for this systematic review was registered on PROSPERO (CRD42017041382).
MEDLINE, Embase, CINAHL, PsycINFO and The Cochrane Library were searched from inception to November 2016. Randomised controlled trials of post-discharge intervention which commenced in the first three months after TKR surgery were included. The primary outcome of the review was self-report pain severity at 12 months or longer after TKR. Risk of bias was assessed using the Cochrane risk-of-bias tool.

Results
Seventeen trials with data from 2,485 randomised participants were included. The majority of trials evaluated physiotherapy interventions (n=13); other interventions included nurse-led interventions (n=2), neuromuscular electrical stimulation (n=1) and a multidisciplinary intervention (n=1). Opportunities for meta-analysis were limited by heterogeneity. No study found a difference in long-term pain severity between trial arms, with the exception of one trial which found home-based functional exercises aimed at managing kinesiophobia resulted in lower pain severity scores at 12 months post-operative compared to advice to stay active.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   3 This systematic review and narrative synthesis found no evidence that one type of physiotherapy intervention is more effective than another at reducing the severity of chronic pain after TKR. Further research is needed to evaluate non-physiotherapy interventions,

Conclusion
including the provision of care as part of a stratified and multidisciplinary care package.
Key words: Total knee replacement, chronic post-surgical pain, prevention, systematic review

Strengths and limitations of this study
This is the first systematic review to evaluate the effectiveness of post-discharge interventions delivered in the first three months after surgery in reducing the severity of chronic pain after total knee replacement.
Synthesis of adverse events data was not possible because assessment and reporting was variable and often poor.

INTRODUCTION
Total knee replacement (TKR) is a common operation to provide pain relief, predominately due to osteoarthritis. Despite good outcomes for many, some patients report chronic pain in the months and years after TKR. Chronic post-surgical pain is defined as pain that is present or increases in intensity at ≥3 months after surgery [1]. In representative populations, unfavourable long-term pain outcomes have been reported by 10-34% of patients with TKR [2]. Patients with bothersome pain at ≥3 months after surgery are disappointed with their outcome [3 4]. Given the prevalence and impact of chronic pain, it is important to evaluate interventions that may optimise patients' outcomes after TKR.
During the hospital stay after TKR, rehabilitation focuses on regaining range of motion and improving mobility. After discharge, rehabilitation aims to enhance recovery, through supporting a person to regain function and quality of life, optimising pain relief and reintegration into social and personal environments [5]. While physiotherapy often focusses on functional health, another key outcome is the prevention of long-term pain [6]. Post-operative physiotherapy may be combined with other interventions to provide comprehensive, multidisciplinary and holistic rehabilitation [7]. A key step to improving patients' outcomes after TKR is to evaluate if early post-operative rehabilitation interventions can reduce the severity of chronic pain after TKR. Chronic pain is difficult to treat once established [8], and therefore it is important to evaluate the effectiveness of early post-operative interventions in reducing the severity of chronic pain.

METHODS
The review was registered on the international prospective register of systematic reviews (PROSPERO) on 17 th January 2017 (registration number CRD42017041382). The review was conducted following guidance from the Cochrane Handbook [9] and reported in accordance with PRISMA guidelines [10] (Appendix 1).

Eligibility criteria
Studies were eligible for inclusion in the review if they met the following criteria: Population: Adults discharged from hospital after primary TKR predominantly for osteoarthritis.
Intervention: Any post-discharge intervention which commenced in the first three months after TKR surgery.
Control: Any, including no intervention, usual care, placebo or an alternative intervention.

Outcomes:
The primary outcome was pain severity at 12 months or longer after TKR, as patient-reported levels of pain plateau by this time point [11 12]. Pain severity could be assessed using a patient-reported joint-specific pain measure (e.g. WOMAC or KOOS pain domains), a quality of life measure (e.g. SF-36 or SF-12) or a Visual Analogue Scale (VAS).

Screening
Records identified by searches were imported into Endnote X7 (Thomson Reuters) and duplicates removed. From the searches, an Endnote database of all RCTs and systematic reviews in TKR was established. Within this database, interventions conducted during the post-operative period were identified. An initial screen for potential eligibility was undertaken by one reviewer (ADB) to exclude articles that were clearly not relevant.
Subsequently, abstracts and full-text articles were screened independently by two reviewers (VW and ADB or JD). Results of screening were compared, and any discrepancies were resolved through further review of the full text articles and discussion between reviewers.
Reasons for exclusion were recorded.

Data extraction
Data from studies that met the eligibility criteria were extracted onto a standardised proforma by one reviewer (VW). Data extraction was checked against source articles by a second reviewer (JD). Extracted data comprised: country, date, participant characteristics, selection criteria; intervention and control treatment; follow up intervals; losses to follow-up; primary outcome; outcome data for pain; adverse events (any untoward medical occurrence in a clinical study participant regardless of the causal relationship with the study treatment) and information for risk of bias assessment. Any disagreements between reviewers were discussed with a third reviewer (ADB) and consensus reached.
A single e-mail was sent to authors of studies with an appropriate follow-up period but no pain outcome to enquire if an appropriate outcome was available. If a combined pain and  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   7 function outcome was reported, such as the OKS or the total WOMAC score, separate pain subscale data were requested. Authors were contacted when necessary for clarification purposes or to request unpublished relevant data. If a study reported data that were combined for knee and hip replacement patients, then disaggregated data for patients with TKR were requested. If this was not available, then the study was excluded.

Risk of bias assessment
Potential sources of bias were assessed using the Cochrane risk of bias tool [9]. At the protocol stage, analysis was planned which included all studies, with sensitivity analyses conducted to exclude studies judged to be at high risk of bias.

Strategy for data synthesis
At the protocol stage, meta-analysis using RevMan 5 [13] was planned if two or more studies were identified with similar interventions and comparator groups and appropriate outcome data. If continuous pain outcomes were measured differently across studies, overall standardised mean differences and 95% confidence intervals would be calculated and presented alongside measures of heterogeneity (I 2 ). Where possible, subgroup analyses were planned to explore the effectiveness of different intervention content and intensity, and different comparator interventions.

Study quality
Risk of bias assessments for individual studies are displayed in Figure 2. All studies were at high risk of bias for blinding of participants and pain outcome assessment due to the nature of the intervention and the self-reporting of pain. Five studies were at high risk of bias due to incomplete outcome data and one due to selective outcome reporting.

Outcomes assessment
The primary outcome was specified for 13 trials; this was function in eight trials; a composite of pain and function in four trials, and pain in one trial (Appendix 3). Pain severity was most commonly assessed using the WOMAC pain scale (n=9); other tools included the KOOS pain

Physiotherapy interventions
Thirteen studies with 1,880 randomised patients evaluated the effectiveness of post-discharge physiotherapy interventions. All interventions started within two months of surgery, with the majority commencing within two weeks of surgery. In addition to all studies being at risk of bias due to issues with blinding, risk of bias due to incomplete outcome data was evident for four studies [18 22 23 25]. Seven studies compared physiotherapy interventions with usual care or minimal care; interventions included a walking skills programme [14], group-based circuit exercise classes [17], erogometer cycling [22], home-based functional rehabilitation , and a three-arm trial comparing 1:1 physiotherapy, group-based circuit classes and a monitored home exercise programme [20].  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60   F  o  r  p  e  e  r  r  e  v  i  e  w  o  n  l  y   10 Of the 13 studies, only one trial reported a difference in pain severity between groups; patients randomised to 6 months of home-based exercises aimed at managing kinesiophobia had lower pain severity scores at 12 months post-operative compared with patients who received general advice to stay active [26].

Nurse-led interventions
Two studies with 319 randomised patients reported evaluation of a nurse-led intervention compared with no care or usual care. Except for issues relating to blinding, both studies were at low risk of bias. Both studies evaluated nurse-led structured telephone follow-up; one aimed to improve adherence to home exercise [16] and the other to provide information regarding well-being, integrity, prophylaxis, safety and other issues relevant to patients after TKR [28]. Pain outcome data (mean and standard deviations) were not available for latter study and therefore meta-analysis was not possible. Neither study found a difference in pain severity scores at 12 months post-operative between the intervention and control group.

Other interventions
Two studies reported evaluations of other interventions. Except for issues relating to blinding, both studies were at low risk of bias. One trial involving 86 patients compared a group-based multidisciplinary programme with usual care [19].

DISCUSSION
This systematic review aimed to evaluate the effectiveness of post-discharge interventions delivered in the first three months after surgery for reducing the severity of chronic pain after TKR. Interventions that predominately comprise physiotherapy have been evaluated in RCTs.
In most studies, the control group received some form of physiotherapy care and therefore the aim of the trials was to compare the effectiveness of different types of physiotherapy, rather than comparing the effectiveness of physiotherapy to no care. A narrative synthesis of the evidence suggests that no physiotherapy intervention appears to be more effective than another at reducing the severity of chronic pain after TKR. However, findings from the trial of a 6 month home-based functional exercise programme aimed at managing kinesiophobia [26] compared to advice to stay active were encouraging and warrant further evaluation. Few studies have been conducted to evaluate the effectiveness of non-physiotherapy interventions at reducing chronic pain after TKR, and further research is needed. included studies was that they were all at high risk of bias due to the lack of participant blinding for self-report pain. However, blinding of participants is rarely possible in RCTs of this nature. Also, it would be expected that the risk would arise from participants in the intervention group reporting less pain, which may potentially be an issue with shorter-term outcomes, but this was not evident from the longer-term follow-up of the studies included in this review.
This systematic review took a broad approach by evaluating the effectiveness of any type of post-discharge intervention that aimed to reduce the severity of chronic pain after TKR.
Interventions that span the post-operative period may be delivered as part of a comprehensive peri-operative package of care, and these would not have been identified in this review; however, evaluations of the effectiveness of pre-operative and peri-operative interventions for reducing chronic pain severity are being conducted separately (CRD42017041382).  severity was assessed as a secondary outcome in these trials and therefore it was expected that the intervention may reduce long-term pain. All but one study found that the intervention did not provide any benefit on long-term pain severity compared to the control group.
However, the treatment received in the control group, particularly in the physiotherapy trials, varied considerably between studies, including a different form or intensity of physiotherapy, provision of physiotherapy based on a needs assessment, delayed treatment or no treatment.
Therefore, it is not appropriate to draw conclusions on the effectiveness of any particular type    Delayed monitored home exercises, with guidance from physiotherapist at 2, 3 and 6 months post-operative. Commenced at 2 months after surgery for 12 months.
Usual care, which involved no additional guidance from 2 months post-operative.

7
Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified.

RESULTS
Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.

Figure 1
Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. Table 1 Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Table 2 Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot.  Table 2 Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]).

DISCUSSION
Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers).

12-15
Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias).

12-13
Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research.