Article Text
PDF
Abstract
Introduction Prognosis following surgical rotator cuff repair (RCR) is often established through the assessment of non-modifiable biomedical factors such as tear size. This understates the complex nature of recovery following RCR. There is a need to identify modifiable psychosocial and sleep-related variables, and to find out whether changes in central pain processing influence prognosis after RCR. This will improve our knowledge on how to optimise recovery, using a holistic rehabilitation approach.
Methods and analysis This longitudinal study will analyse 141 participants undergoing usual care for first time RCR. Data will be collected 1–21 days preoperatively (T1), then 11–14 weeks (T2) and 12–14 months (T3) postoperatively. We will use mixed-effects linear regression to assess relationships between potential prognostic factors and our primary and secondary outcome measures—the Western Ontario Rotator Cuff Index; the Constant-Murley Score; the Subjective Shoulder Value; Maximal Pain (Numeric Rating Scale); and Quality of Life (European Quality of Life, 5 dimensions, 5 levels). Potential prognostic factors include: four psychosocial variables; pain catastrophising, perceived stress, injury perceptions and patients’ expectations for RCR; sleep; and four factors related to central pain processing (central sensitisation inventory, temporal summation, cold hyperalgesia and pressure pain threshold). Intercorrelations will be assessed to determine the strength of relationships between all potential prognostic indicators.
Our aim is to explore whether modifiable psychosocial factors, sleep-related variables and altered central pain processing are associated with outcomes pre-RCR and post-RCR and to identify them as potential prognostic factors.
Ethics and dissemination The results of the study will be disseminated at conferences such as the European Pain Congress. One or more manuscripts will be published in a peer-reviewed SCI-ranked journal. Findings will be reported in accordance with the STROBE statement and PROGRESS framework. Ethical approval is granted by the Ethical commission of Canton of Zurich, Switzerland, No: ID_2018-02089
Trial registration number NCT04946149.
- shoulder
- pain management
- psychiatry
- rehabilitation medicine
- sleep medicine
- neurophysiology
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
This will be the first study adequately powered to identify modifiable psychosocial factors as potential prognostic factors of outcome after rotator cuff repair (RCR).
This study will also be the first to assess the complex interplay of psychosocial factors, sleep-related variables and central pain processing measures as potential prognostic factors of outcome following RCR.
The prospective longitudinal study design includes three measurement points, starting preoperatively, at 12 weeks postoperative, and following up for 12 months post RCR.
The questionnaires for sleep and patients’ expectations were translated to German for the purposes of this study. Therefore, their validity in our population (German-speakers) has yet to be validated.
Tear size is a known prognostic indicator of how well recovery will go following RCR. We will not account for tear size in our prognostic model which may bias our results.
Introduction
Prognostic factor research most often focuses on biomarkers, including biological, clinical or physiological factors. Prognostic factors help us predict the likely outcome of a patient undergoing a procedure, given the presence of certain behaviours or characteristics.1 For patients undergoing rotator cuff repair (RCR) for shoulder pain, prognostic factors often include: patient’s age; fatty infiltration into the rotator cuff muscles; quantified tendon tear size or multiple tendon involvement; and the presence of a confirmed diabetes diagnosis.2 These are all non-modifiable biomedical markers with established capability to predict worse outcomes for patients following RCR.2
Despite these biomarkers being recognised prognostic factors for RCR, we are still not able to fully predict who will recover successfully. A person’s perception of shoulder pain is far more complex than structural changes in the rotator cuff (RC) tendons. More information is required to gain a comprehensive understanding of all factors that influence recovery.3–7 Yet, the number of RC repairs in Europe and the USA continues to grow,4 8–10 in spite of this lack of knowledge on the odds of success. Current evidence suggests satisfactory outcomes post RCR range from 38% to 95%. This means surgical repair is either very successful or potentially a large waste of resources.11–14
There is growing evidence that psychosocial factors impact persistent shoulder pain.4 15–18 Factors such as: high distress; maladaptive beliefs;17 the perception of high-demand at work; and a lack of social support18 can influence whether persistent shoulder pain and disability occur. Patients with existing preoperative (RCR) psychological conditions like: depression and anxiety;14 who exhibit pain catastrophising and kinesiophobia;19 or suffer psychological distress14 20 may demonstrate greater preoperative shoulder pain.14 19 In the reverse, patients who anticipate a good recovery (positive expectations) post-RCR show independent and strong associations with satisfactory outcomes (good prognosis) for pain and disability measured 1 year post surgery.11 12 21 Prior research on psychosocial factors post RCR has been restricted to: preoperative measures;19 has lacked statistical power;14 20 or has failed to investigate potential psychosocial prognostic factors altogether.11 12 21
Sleep disturbances are also highly prevalent (up to 89%) in patients undergoing RCR. Sleep disturbance has been attributed to the presence of shoulder pain.22–24 RCR seems to reduce this interplay between shoulder pain and sleep disturbances as findings demonstrate an overall post RCR improvement of sleep quality.14 25 Yet, 41% of patients with RCR still suffer from sleep disturbances at 24 months follow-up.23 Investigations of sleep disturbances in relation to shoulder pain and RCR are incomplete with multiple factors affecting the relationship.26
Central pain processing (CPP) changes are measured via assessments for central sensitisation. Assessments of CPP are almost absent in studies of patients undergoing RCR.15 16 27 Two trials28 29 investigated the role of central sensitisation, measured with quantitative sensory testing (QST) on outcome (pain and disability) after different shoulder surgeries (RCR, superior labrum from anterior to posterior (SLAP) repair, shoulder arthroscopy (SA) and subacromial decompression). Both studies found small effects of CPP on postoperative outcomes. If a high amount of CPP was present preoperatively, it was related to a worse outcome 3 months postsubacromial decompression.28 In contrast, if a small amount of CPP was present 3 months postoperatively (RCR, SLAP-Repair, SA) it was associated to better functioning at 6 months postsurgery.29
The existence of potential modifiable prognostic indicators related to psychosocial factors, sleep and CPP and their effects on, shoulder function, disability, pain, quality of life and satisfaction following RCR require further investigation.4 19 Neither the local tissue pathology-pain model nor the growing knowledge about local biochemical changes in RC tendons sufficiently describe the relationship between tissue changes and patients’ perceived shoulder pain.3 5 15 30 Studying the relationship of psychosocial factors, sleep and CPP with RCR would improve our prognosis for outcomes post RCR. This holds the potential to improve treatment selection choices and reduce unnecessary surgical interventions.3 4 16 20 31
This study aims to answer the following questions:
Do psychosocial factors such as pain catastrophising, perceived stress, injury perceptions, patients’ expectations of surgery, sleep-related variables and measures of CPP obtained pre RCR (baseline), influence baseline shoulder function, disability, pain and quality of life and their evolution over time (1 year post-surgery)?
How do potential prognostic factors such as psychosocial indicators, sleep-related variables and CPP intercorrelate at baseline and over time?
Methods
Study design and setting
The longitudinal cohort study will be implemented and reported in line with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for observational studies32 informed and completed by the framework ‘prognosis research strategy’ (PROGRESS).1 33
Data will be obtained in a single shoulder and elbow surgery unit in the clinic of orthopaedic surgery and traumatology in alliance with the institute of therapy and rehabilitation of the acute care hospital, canton hospital Winterthur, Switzerland.
The current research project will analyse data from three time points in the routine clinical management post-RCR: 1–21 days preoperatively (T1); 11–13 weeks postoperatively (T2); and 12–14 months postoperatively (T3). Data from July 2019 onwards will be considered. Data collection including 12 months follow-up is estimated to be complete in Summer 2022.
See tables 1 and 2 for overview of measurement points.
Outcome measures
Potential prognostic factors
Participants
The population of interest includes adult patients undergoing elective RCR, for tears of traumatic and non-traumatic origin. To avoid selection bias, we will include data from consecutive patient consultations.
Eligibility criteria
Inclusion criteria
Adult men or women ≥18 years of age.
Scheduled for elective arthroscopic RCR.
First time RCR on the target shoulder.
Exclusion criteria
Changes of intraoperative procedure (eg, anything but RCR).
Re-repair of tendon.
No surgery.
No preoperative data available; for example, fast track patients with trauma.
Outcome measures and prognostic factors
Our outcome measures are consistent with those used in the existing literature. We consulted the guidelines from the OMERACT 2016 Shoulder Core Outcome Set Special Interest Group.34
Our dependant variables are the primary outcome measure Western Ontario Rotator Cuff Index (WORC) for disease-specific function, disability and quality of life. The secondary outcome measures are: Constant-Murley Score (CMS) and Subjective Shoulder Value (SSV) for shoulder function; maximum pain over the last 7 days on Numeric Rating Scale (NRS); European Quality of Life, 5 dimensions, 5 levels (EQ-5D-5L) for quality of life and health status; and a satisfaction measure developed by Swarup et al.35
A detailed description and overview about primary and secondary outcome measures and their psychometric properties is presented below in table 1.
Potential prognostic factors for postoperative outcome are:
psychosocial factor
spain catastrophising,
perceived stress,
injury perceptions,
patients’ expectations for RCR
sleep-related variables
sleep quality
sleep efficiency
sleep disturbance since when
No of awakenings per night
measures of CPP
the central sensitisation inventory (CSI) to assess self-reported somatic and emotional complaints associated to CPP
temporal summation (TS)
cold hyperalgesia (CH)
pressure pain threshold (PPT)
the Douleur Neuropathique-four assessment (DN4) to detect the possible presence of neuropathic pain and
pain surface/distribution on the body chart
Further factors include patient-related characteristics such as; demographics: (15) age and (16) sex; clinical variables (17) trauma vs non-traumatic tendon tear; and health status such as (18) body mass index. These characteristics are all handled as potential prognostic factors to ensure a correct estimation of our primary prognostic factors.
Table 2 presents an overview of the potential prognostic factors including a detailed description of all measurement tools and test methodology.
Four (AS, JW, QdG, FM) experienced and trained (by the first author AS) shoulder specialists and physiotherapists will perform all the measurements (CMS, QST, SSV, NRS pain). To support the training, all participant assessment files will incorporate detailed descriptions with respect to how the assessor should formulate questions and offer answer suggestions.
Statistical methods and analysis
Statistical analyses will be performed using SAS (SAS V.9.4, SAS Institute Inc, Cary, North Carolina, USA). Level of significance is set at p=0.05. Measurements will take place at three time points in the perioperative management, as described above (T1=at baseline 2–3 weeks prior to RCR, T2=at 12 weeks post RCR and T3=at 12 months post RCR as follow-up).
The primary outcome (WORC) will be modelled using mixed-effects linear regression models for repeated (longitudinal) measures, using an unstructured covariance matrix. Dependent variables are the primary and secondary outcomes. Continuous secondary outcomes will be assessed in a similar way to the primary outcome. The models will be developed by stepwise reduction of the a priori determined potential prognostic factors (eg, psychosocial factors, sleep and CPP). A prognostic factor will be retained in the model if it has a significant effect on the initial outcome or on the outcome over time, or if the fit statistics (Deviance, AIC, BIC and R2) of the model improves after inclusion of the variable, in order to increase the precision of the fixed-effects estimates.36–38 This means that a prognostic factor may be retained in the final model, even if it is not significant (p>0.05), to ensure correct estimation of other (significant) prognostic factors.
Descriptive statistics will be performed for comorbidities such as obesity, diabetes and depression; for insurance status (healthcare vs accident insurance); and current profession.
Sample size
We will use a linear mixed-effects regression model for repeated measures. This will have a power of 90% to identify prognostic factors of both interindividual baseline differences in WORC score and a change in WORC score over time that are considered clinically relevant (we assume a SD of 300 points at baseline and a decline in WORC score of at least 15% over time on average),39 at a confidence level Alpha=0.05 (two-tailed). The required total sample size was calculated to be 125 subjects (R, Edland package).40 41 To account for an expected attrition rate of 12.5%, the final sample size was set at 141 participants.
The power is set at 90% to minimise the chance of making a type II error.
It is especially difficult to determine a correct sample size for a longitudinal exploratory study, as the final mixed model is likely to contain complex variance and correlation patterns that are not known beforehand. Therefore, we plan an interim analysis after the inclusion of the first 80 participants, to assess the drop-out rate, the achieved power and the potential futility of the a priori selected prognostic factors. Mixed models do not require complete datasets to produce accurate results, through correct specification of the likelihood function.37
Data security and management
Data generation, transmission, storage and analysis within this project strictly follow Swiss legal requirements for data protection. The electronic data capture software REDCap42 43 will be used for data processing and management. REDCap was developed by an informatics core at Vanderbilt University in 2004, with ongoing support from US National Center for Research Resources (NCRR) and US National Institute of Health (NIH), grants NIH/NCATS UL1 TR000445. REDCap was specifically developed around HIPAA security guidelines and is Good Clinical Practice-compliant and fulfils the regulatory requirements regarding the collection of patient data in clinical trials or non-interventional studies and patient registries and the EU data protections laws. Appropriate coded identification (eg, pseudonymisation) is used in order to enter subject data into the database. The coding list of target data is saved in a secured folder on the hospital’s server. Only the project leader, study nurses and principal investigator have access to it. Between the members of the research team only coded and deidentified data will be shared. Safe handling of the coded data will be covered by the software REDCap.
Study monitoring
An audit trail and history of data transmission are provided by REDCap. The steering committee of the research project will oversee all aspects of design, delivery, quality assurance and data analysis according to good clinical practice and local legislation.
Ethics
The study follows the principles of the Helsinki Declaration. Only data of patients who gave general consent to the hospital or informed written consent to the project will be considered for analysis. Ethical approval was received in January 2019 (ID 201802089) by the Ethical Committee of the Canton of Zurich, Switzerland.
Dissemination of results
The research team is committed to full disclosure of the results of the study. The results of the study will be disseminated for research purpose at different conferences and as published articles in peer-reviewed journals. Findings will be reported in accordance to the STROBE statement and we aim to publish in high impact journals.
Ethics statements
Patient consent for publication
Acknowledgments
The authors thank all the study participants in advance. The authors thank all the physiotherapy staff (Axel Boger, Julian Wiedenbach (JW), Quintin de Groot (QdG), Fabian Mottier (FM), Annina Banz, Nina Pietsch and Sarina Schär), surgical team (MD Alexa Schmied-Steinbach, MD Emanuel Benninger and assistant surgeons), administration officers (Beatrice Hurst, Vanessa Humair), nurses and research departments collaborating on this study. A special thanks is towards Michaela Hagen and Michelle Hitz, the outstanding study nurses. Also, thank you to Marlene Wegmann and Ursina Spörri, who supported the process of ethical approval and the usage of REDCap software. Susanna Weber from the Clinical Trials Center is also acknowledged for her great technical support with REDCap software. We thank the patients who supported our feasibility phase.
References
Footnotes
Twitter @arianeschwank, @FilipStruyf, @michelmertens4
Contributors AS is the project leader, who receives support from FS and MMee with respect to topic-selection and study design. TS controls the statistical model and will support statistical data analysis. MMer and PB contributed to the protocol through critical review and intellectual content. DG and MP support the feasibility of the study and access to clinical data through the physiotherapy and orthopaedic clinic at Canton Hospital Winterthur. DG and MP also support data security by providing REDCap software. All authors gave final approval for publication of the present protocol version.
Funding This work is supported by the Swiss physiotherapy association physioswiss through a financial research prize to the first author. Canton Hospital Winterthur supported AS with study fee payment.
Competing interests None declared.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.