The effectiveness of physiotherapy interventions on pain and quality of life in adults with persistent post-surgical pain compared to usual care: A systematic review

Background Persistent post-surgical pain (PPSP) is a common condition following surgery, resulting in persistent pain and reduced quality of life (QoL). While pharmacological management is common, its effectiveness remains equivocal. This systematic review investigated the effectiveness of physiotherapy management in adults with PPSP in comparison to usual care. Methods A systematic search of six electronic databases was conducted. Studies of human adults (>18 years) with PPSP localised or directly referred from the surgical site, pain persisting for at least two months post-surgery and with physiotherapy as the intervention were included. This review was conducted and reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The McMaster critical review form for quantitative studies was utilised to assess the methodological quality. A descriptive synthesis was undertaken due to the heterogeneity of the included studies. Results Of the 1395 articles that were screened, eight studies met the inclusion criteria. A diverse range of physiotherapy interventions were utilised, and effectiveness was measured through diverse outcomes and measures. Summarised findings from the heterogenous evidence base indicated that physiotherapy interventions for PPSP has a positive impact across a range of outcomes, including pain, quality of life (QoL), physical function and depression. While these are encouraging findings, the current evidence base lacks uniformity with regards to participant characteristics, time periods since diagnosis, interventions delivered, and its parameters, and outcomes measured. Conclusion Due to ongoing challenges in the management of PPSP, alternate treatment strategies such as physiotherapy are being trialled. Despite a number of methodological constraints, current evidence indicates that physiotherapy could play a role in the management of PPSP. PROSPERO registration CRD42019129580

Wylde and colleagues [17] investigating the management of PPSP, was unable to conclude the effectiveness of any modality, pharmaceutical or medical, to reduce pain in PPSP patients. Additionally, a systematic review by Chaparro and colleagues [18] investigated different pharmaceutical approaches to prevent the occurrence of PPSP. They found the incidence of chronic pain post-surgery was reduced with ketamine administration; however, multiple other pharmaceutical approaches showed no reduction [18]. Andreae and Andreae [19] investigated the effectiveness of anaesthetic and observed minor pain reduction in PPSP following thoracotomy and mastectomy. Medical interventions are currently accepted for management of PPSP, despite ambiguous evidence to support its use [20].
Some non-pharmacological interventions have also been utilised in the management of PPSP, including psychotherapy, acupuncture, osteopathy and physiotherapy [21][22][23]. Emerging studies investigating physiotherapy management are shown to have some effectiveness in reducing pain, improving QoL, physical functioning and depression [22][23][24][25][26][27][28][29]. However, there is a limited evidence base; and to date, no systematic reviews have been published evaluating the effectiveness of physiotherapy for PPSP. Therefore, the present review aims to evaluate the effectiveness of physiotherapy interventions as compared to usual care on pain and quality of life in PPSP subjects.

Search protocol and registration
The protocol for this systematic review was registered with the international prospective register of systematic reviews-PROSPERO (Registration # CRD42019129580).

Search strategy
This review was conducted and reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and correlates with the PRSIMA checklist (S1 Appendix). The PICO format was utilised in the development search strategy with search terms and limits relating to PPSP (population of interest) and physiotherapy (intervention). The development of the search strategy was informed by discussions with a University of South Australia Academic Librarian. As means of trialling the search strategy, two reviewers (HP and GB) searched a database independently. The results from the search was then cross-referenced to ensure there was consistency in the search process. If the outcomes of the search results were different, the two reviewers met to discuss these differences and conflicts were resolved through discussion. Only after consistency in the search process was achieved, the reviewers commenced formal searching of the databases. A comprehensive search was developed, and six electronic databases were searched between 26 th -28 th of March 2019. These databases include: Ovid Medline, Ovid Embase, Ovid Emcare, Cochrane (all available dates), PEDro, PsycINFO. Grey literature was additionally searched for further relevant information. As no systematic reviews were previously published, no date restrictions were applied. The following search terms were used as MESH headings: physical therapy modalities, physiotherapy, postoperative pain, chronic pain. The full search strategy for a database is outlined in S3 Appendix and syntax in S4 Appendix. The reference lists of all included studies were reviewed to obtain any relevant studies that were not located by electronic search. experimental studies. Secondary research designs were included for pearling purposes and qualitative research designs were excluded. The eligibility criteria for the population-intervention-comparator-outcome (PICO) is outlined below.

Population
The studies were included if participants were adults (>18 years) of either gender who have PPSP. PPSP was defined by persistent pain from two months following [7]; pain-free before surgery or any preoperative pain must have varying characteristics or increased intensity; pain must be localised or directly referred from the surgical site; exclusion of pain caused by other factors, such as, infection or malignancy [4,7]. PPSP Studies were excluded if participants had undergone amputations (phantom limb pain), had pre-existing pain that had not changed or intensified since surgery, acute pain (less than two months post-surgery) or other causes of pain, such as infection or malignancy.

Intervention
Studies were included if the physiotherapeutic intervention was prescribed, or could be performed, by a Physiotherapist (or an equivalent professional group such as physical therapist). Physiotherapeutic interventions included but were not limited to electrotherapy, exercise, education, manual therapy (manipulation, mobilisation, massage) etc. Studies were excluded if the intervention was self-prescribed by the patient without the direct involvement of a physiotherapist (such as topical creams or gels) or had a pharmaceutical focus. As physiotherapy is commonly provided as a package of care (comprising of multiple interventions), the review did not limit to a single intervention or set specific parameters of intervention.

Comparator
The acceptable comparators were control (no intervention provided) or 'usual care' which typically focusses on pharmacological (non-opioid analgesics, opioids, and adjuvant analgesics) and other conservative interventions.

Outcome
The search was not limited to any specific outcomes, due to the multidimensional nature of pain and the variety of outcomes related to the effects of physiotherapy interventions for pain. Outcomes of interest included but were not limited to: pain, QoL, depression, anxiety and activities of daily living.

Literature search
Once the search strategy was developed, a review protocol was established (S2 Appendix), and databases were searched. Search results from each of the electronic databases were transported into the referencing management program, Endnote TM to organise and sort identified studies. Duplicate results found during the search process were then deleted through Endnote TM . From the Endnote TM library, all records were uploaded into Covidence TM , where studies were accepted or removed through screening the title and abstract against the PICO criteria. The subsequent relevant studies' full texts were then independently analysed by two reviewers to determine their eligibility for the PICO criteria. Any disagreements were resolved by discussion or with a third reviewer, where required.

Methodological quality
The included studies were independently reviewed and ranked using the 'Intervention category' of the National Health and Medical Research Council (NHMRC) hierarchy of evidence [30], by all six reviewers. Any disagreements were resolved by discussion or with the independent reviewer (SK), where required. A modified McMaster Critical Review Form for Quantitative Studies [31] was used following the guidelines to assess the methodological quality of the included studies, shown in appendix 2. This tool assessed eight main components: study purpose; literature review; study design (all experimental designs); sample (participants description, size justification, ethics and consent); outcomes (reliability and validity, outcome areas and measures used); intervention (description, contamination and co-intervention); results (statistical and clinical significance, analysis methods and drop outs) and conclusion with implications to practice (limitations and biases). The individual components were rated as 'yes', 'no', 'not addressed' or 'NA-Not Applicable'. A score of '1' was given to 'yes', '0' to 'no and not-addressed'; if 'NA' category applied then the total scoring was changed accordingly. Dependent on the research design and applicable components, the maximum total score of a study was 14. Each study was independently rated by each reviewer and any disagreements were resolved by discussion or with independent reviewer (SK), where required.

Data management
A customised data extraction was developed by the reviewers on Microsoft Excel TM Spreadsheets, including information related to: study design, participants' information (age, sex, diagnosis), intervention components, intervention duration, dose and frequency of intervention, experimental design, method of randomisation, outcome measures, statistical analysis and results of pre-and post-analysis or other appropriate results. All six reviewers independently extracted data using this tool to ensure reliability and consistency. Any disagreement was resolved through discussion and consultation with another reviewer (SK), where needed.

Synthesis of results
Due to heterogeneity in the included studies, a meta-analysis was not conducted. Instead, a descriptive synthesis of results was undertaken, following the NHMRC FORM framework [32]. This framework has been widely used in previous systematic reviews successfully [33][34][35]. The framework consists of five main components: 1. Quantity and quality of evidence; 2. consistency; 3. clinical impact; 4. generalisability; and 5. applicability to the Australian health care setting. The final component was not used in the present systematic review due to its geographical focus.

Study selection
The literature search returned 2157 total results, 762 duplicates were removed. The title and abstract of 1395 articles were screened following pooling of searches and of 1371 irrelevant articles were removed. 24 studies were reviewed in full, of which 16 studies were excluded: seven due to the study population not matching the aforementioned inclusion/ exclusion criteria; six for the incorrect study design and three for the incorrect intervention. Eight studies successfully met the eligibility criteria and were included in this study. The PRISMA flowchart is outlined in Fig 1. Table 1 shows an overview of the NHMRC levels of evidence and allocated scores based on the McMaster Quantitative Critical Appraisal Tool [31] of the eight included studies. The study designs comprised of one RCT [24], one CCT [25] one pre-post study [22], three case series [23,[26][27] and two case reports [28][29]. The two highest critical appraisal scores, 100% and 91.6% were allocated respectively to the studies by Nunez-Cortez [26] and Albayrak [25] and colleagues. The lowest score was allocated to Vas and colleagues with 55.5%, suggesting the lowest level of study quality out of the eight included studies. Overall, what was done well amongst seven of the included studies were the discussion of relevant background information and justifying the research [22][23][24][25][26][27]29]. In addition, all studies used valid and reliable outcome measures [22][23][24][25][26][27][28][29].

Risk of bias
The predominant methodological concerns between the included studies were: lack of justification of the sample size (one study did a power calculation [24]), lack of detailed descriptions regarding intervention (six studies utilised multiple interventions and provided inadequate details of the same) [22,[25][26][27][28][29], lack of randomisation (only one study randomised Bailey [22] Pre-post Wong [29] Case Report participants [24]), lack of reporting statistical significance (six of the eight studies included statistical significance [22][23][24][25][26]28]) and lack of reporting clinical significance (three studies reported clinical significance [22,26,29], with one reporting effect size [26]). Additionally, three studies failed to detail ethics approval and participants' consent [22,[28][29]. Table 2 provides a summary of the study characteristics 2. A variety of research designs were identified for this systematic review; which included a range of participants from differing countries. The studies were conducted in the United States of America [22,24,29], Turkey [25], Sweden [23], Chile [26], New Zealand [28] and India [27]; published between 2010-2017. All studies investigated the effects of physiotherapy interventions on PPSP in relation to a range of outcome measures. Despite having the same patient population, there was a large distribution of the duration of PPSP ranging from three months to seven years.

Participant characteristics
There was a combined total of 166 participants, with the number in each study ranging from one to 58, aged between 28 and 74 years. Of total participants, there were more female (97) participants than male (69). Ethnicity of participants was reported in one study [22]. Preceding surgical procedures differed between studies, which included total knee arthroplasty [25][26][27], laminectomy [22], thoracoabdominal oesophagus resection [23], multi-level lumbar fusion [28] and abdominal and pelvic adhesion [29].

Outcome measures (OMs)
The types of outcomes and OMs varied between studies to evaluate the effectiveness of physiotherapy interventions. Studies utilised a combination of subjective measures: including pain; QoL and depression; and objective measures, such as the timed up and go test (TUG), the sixminute walk test (6MWT) and the 30-second chair stand test (30-CST). The time point measurements differed between studies with one measuring the immediate effect (same day) [29], and all eight-measuring the medium-term effects (one month to four months) [22][23][24][25][26][27][28][29]. No studies measured, nor reported, adverse outcomes because of physiotherapy. The range of outcome domains and measures used for each study are identified in Table 3.

QOL
Two studies measured the effects of physiotherapy management on QOL using self-administered questionnaires, including the WOMAC and the 36-item short form survey (SF-36) [22,25]. Studies by Albayrak and colleagues [25] provided a package of care to participants, comprising of: TENS, exercise and PRF; as well as cognitive behavioural therapy, exercise and education [22]. When using WOMAC, Albayrak and colleagues [25] reported a significant (p<0.05) reduction in total scores between pre-treatment, 15 days post-treatment and onemonth post-treatment; between groups. Bailey and colleagues [22] also discovered a statistically significant difference between admission and discharge scores on the SF-36 (p<0.05).

Physical function
Two studies measured the effect on physical function and utilised the 6MWT, the TUG and the 30-CTS to measure the same [22,26]. Similarly, Nunez-Cortes and colleagues [26] used a package of care in the management of the participants, comprising of dry needling and exercise. Bailey and colleagues [22] used the 6MWT to assess physical function and found a statistically significant improvement (p<0.05) between admission and discharge; however, participants were still below normative values for their age group. Nunez-Cortes and colleagues [26] used the 6MWT, TUG and 30-CST and found a statistically significant improvement (p<0.05) in the 6MWT pre-and post-intervention. However, they found no statistically significant improvement in the TUG and 30-CST scores [26].

VAS NRS PSEQ MPI-LC MPI-I BPI-SF WOMAC-Pain PCS OKS CMS M-CMS WOMAC SF-36
6MWT TUG PHQ-9 CES-D Albayrak [  Depression Two studies measured changes in depression symptoms and utilised the Centre for Epidemiologic Studies Depression Scale (CES-D) and the Patient Health Questionnaire-9 (PHQ-9) [22,27]. Vas and colleagues [27] provided a package of care in managing the participants, comprising of: PRF, TENS, dry needling and exercises. They reported a positive impact in the PHQ-9 scores; however, did not report statistical significance. Bailey and colleagues [22] used the CES-D and reported a positive statistical significance (p<0.05) between pre-and post-discharge, indicating a reduction in depressive symptoms and negative pain-related cognition.

Summary of results
The results from the included studies, which highlights the range of outcomes measured (16 different outcomes) within four domains, are summarised in Table 4. Irrespective of this, collective findings indicate that physiotherapy management has a positive impact across all domains. There is consistent evidence to suggest that physiotherapy had a positive impact in symptom reduction (pain), which is the most commonly measured outcome. While QoL, physical function and depression were measured by only a handful of studies (two studies each for QoL and depression and three studies for physical function), available evidence points to a positive impact from physiotherapy. Regarding physical function, two studies reported statistically significant improvements in the 6MWT [22,26] and one study reported improvement in the TUG and 30-CST [26]. While these are encouraging findings, given the small number of studies, especially for outcome domains such as QoL, physical function and depression, caution is required when interpreting these results.

NHMRC FORM framework
The synthesis of results using the NHMRC FORM framework is shown in Table 5. Despite positive outcomes reported in some studies, especially the positive impact of physiotherapy for

Bailey [22]
pain, methodological concerns of the evidence base lowered the overall grade of recommendation. While these findings may be supportive of physiotherapy, caution is required when implementing these recommendations.

Discussion
This systematic review aimed to address the evidence in the emerging area of physiotherapy management of PPSP. Previous systematic reviews have addressed the effectiveness of pharmaceutical management, which have resulted in equivocal recommendations. There was a modest evidence base, consisting of eight studies representing varying research designs. The summarised findings from this review indicate that physiotherapy interventions are likely to have a positive impact across a range of outcomes including pain, QoL, physical function and depression. This was particularly evident for pain, which indicates that physiotherapy interventions are likely to be effective in the management of adults with PPSP. Despite these positive findings, caution is required when interpreting these findings due to the methodological issues within, and lack of uniformity of, the evidence base. As this is the first systematic review to map the literature on the effectiveness of physiotherapy interventions for adults with PPSP, it

Clinical impact D-Poor Slight or poor
Consistent findings for primary and secondary outcomes: pain, QoL, physical function and depression No adverse effects reported Intervention protocol adequately described however, no justification for parameters or development process was outlined

Generalisability B-Good Population(s) studied in body of evidence are similar to the target population
Population of studies is similar to the target population Age range: 28-74 years The current evidence base lacks clarity in terms of reporting of comorbidities and co-interventions and its impact on outcomes Studies conducted in six different countries that have different health care contexts

D-Poor
Body of evidence is weak, and recommendation should be applied with caution Overall, these studies were low level of evidence and were of moderate methodological quality. While there were generally consistent positive findings, the current evidence base lacks uniformity with regards to participant characteristics, varied time periods since diagnosis, interventions delivered, and its parameters and outcomes measured for adults with PPSP. https://doi.org/10.1371/journal.pone.0226227.t005 Effectiveness of physiotherapy interventions in persistent post-surgical pain showcases where there is consistent evidence (pain), emergent evidence (QoL, physical function and depression) and areas for future research.

Effect of physiotherapy interventions on various outcomes
The use of physiotherapy interventions for managing PPSP identified positive effects across several outcome domains including subjective and objective measures, highlighting the potential role of physiotherapy for these patients. The findings of this systematic review are supported by previously published research on the positive effects of physiotherapy for persistent pain. For example, physiotherapy interventions have been shown to reduce pain across a range of conditions through diverse interventions [36]. This includes patient population which are similar to PPSP such complex regional pain syndrome [37]. As reported in this systematic review, physiotherapy interventions are often offered as a package of care which in turn may address diverse and multifactorial components which underpin PPSP. Two studies reported improvements in depression in patients with PPSP [22,27]. These positive results are likely as physiotherapy management often includes exercise prescription, which has previously shown to improve anxiety and depression symptoms for a range of conditions, including breast cancer and spinal disorders [38][39]. Exercise has previously been shown to be effective in improving QoL and psychological functioning [40] as well. Furthermore, in the general adult population, higher levels of exercise are associated with decreased symptoms of anxiety and depression, in addition to an overall increased health related QoL [41][42][43]. The benefits exercise has on depression and anxiety may occur due to reducing cortisol levels, increasing a patient's self-esteem and acting as a diversion for negative thoughts [43]. In addition, physiotherapists often incorporate group classes when exercise is included in the management of patients and this social contact may be beneficial in reducing the above symptoms [43]. It has been suggested that physiotherapy should be considered an adjunct to antidepressants and/or psychotherapy in moderate to severe depression [44].
Two studies also reported statistically significant improvement in QoL after physiotherapy management of PPSP [22,25]. As PPSP is associated with hospital stays, physiotherapy has been shown to effective in improving QoL in other patient population, such has post-cancer treatment or intensive care unit stays, who have also experienced hospital stays [45][46]. One possible explanation for these findings is that hospital-based physiotherapy may better prepare patients to transition into the community and subsequently by focussing on improved physical activity, function and decreasing troublesome symptoms, physiotherapy may be effective in improving QoL [45]. Similarly, two studies [22,26] reported statistically significant improvements in physical function by increasing the distance walked in the 6MWT and decreasing time taken to perform the TUG. These findings are supported by previous research which has demonstrated that exercise programs are beneficial in increasing physical function in frail older adults by increasing gait speed, improving balance and activities of daily living [47]. Physiotherapy may improve physical functioning by addressing psychological factors (through education), deconditioning (through exercises) and pain (through a package of care consisting of education, manual therapy and exercises).

Limitations
This systematic review was underpinned by best practice in the conduct of systematic reviews (PRISMA). However, as with any research, there are limitations to this systematic review as well. In order to avoid publication bias, grey literature and secondary searching strategies were implemented. However, unlike black literature search, which can be standardised and applied across multiple databases, grey literature searching is complex and imprecise. Therefore, some publications may have been missed. Due to access and resource limitations, only publications in English language were included and hence language bias should be acknowledged. While this systematic review utilised current and universally agreed definition of PPSP as part of its literature selection process, there may have been other studies which were excluded as they utilised other definitions of PPSP. By using the current and universally agreed definition of PPSP, this systematic review was able to compare similar evidence.
While the systematic searching of the literature identified a modest body of evidence with consistent positive findings, there was some concerns with the methodological quality of the included studies. The sample sizes were generally small and lacking justification, possibly due to the fact that many of the research studies were low quality studies, for which sample size calculations/justifications may be seen to be unnecessary. Due to the nature of the intervention, blinding of patient and therapist was not possible, although the assessor could be blinded. While placebo and Hawthorne effects may not be avoided, measurement bias could be. Given the nature of physiotherapy intervention, which is often provided as a package of care, in some studies physiotherapy interventions were complemented with other interventions (including medications). While this may reflect what occurs in clinical practice settings, the role of co-intervention bias should be considered. Due the complex nature of pain, there was a variety of outcome measures utilised across the studies. This made direct comparison between the studies difficult. Given that more than half of the included studies were case studies/series, generalisability of the findings to a wider population context is limited. While physiotherapy may play a role in the management of PPSP, access to physiotherapy services is likely to be influenced by a number of factors such as private health insurance, local health care models of service delivery, funding models and requirements. These factors need to be considered when considering physiotherapy as an option in the management of PPSP.

Implications for practice
There is an emerging body of evidence to support the use of physiotherapy as a treatment option for adults with PPSP. There were positive impacts across a range of outcomes including pain, QoL, physical function and depression. However, while physiotherapy may be considered in the management of PPSP in adults, it must be recognised that the current evidence base suffers from several methodological concerns and therefore implementation of these recommendations should be made with caution.

Implications for practice and future research
A modest body of evidence has been identified to support the use of physiotherapy interventions for adult patients with PPSP. However, methodological concerns within the evidence base have also been recognised, highlighting the need for further research. While physiotherapy is commonly offered as a package of care in clinical practice settings, in the research context there is variability in results in terms of intervention parameters. Future research would benefit from studies that focus on developing standardised intervention parameters. Similarly, future research could also add to the current evidence base by developing and implementing standardised outcome measures for PPSP. Finally, methodologically sound randomised controlled trials that are conducted with larger sample sizes, with clear control and intervention groups with randomisation, using power calculations, and include long term follow-up would assist in identifying the sustained impact of physiotherapy on adults with PPSP. This could be complemented with information regarding clinical significance. This will increase generalisability, clinical impact and provide unequivocal evidence for the effectiveness of physiotherapy for adults with PPSP.