You are here

Article Text

Article menu

Download PDFPDF

Rehabilitation medicine
Original research
Understanding the clinical profile of patients with frozen shoulder: a longitudinal multicentre observational study
  1. Michel GCAM Mertens1,2,
  2. Mira Meeus1,2,3,
  3. Suzie Noten4,5,
  4. Olivier Verborgt1,6,7,
  5. Erik Fransen8,
  6. Enrique Lluch Girbés2,9,10,
  7. Marta Aguilar Rodríguez9,
  8. Santiago Navarro-Ledesma11,
  9. Manuel Fernandez-Sanchez12,
  10. Alejandro Luque-Suarez13,
  11. Filip Struyf14,
  12. Lirios Dueñas9
  1. 1Research Group MOVANT, Department of Rehabilitation Sciences and Physiotherapy (REVAKI), University of Antwerp Faculty of Medicine and Health Sciences, Wilrijk, Belgium
  2. 2Pain in Motion international Research Group, www.paininmotion.be, Belgium
  3. 3Department of Rehabilitation Sciences and Physiotherapy, Ghent University, Ghent, Belgium
  4. 4Department of Rehabilitation Medicine, Erasmus MC University medical Center, Rotterdam, The Netherlands
  5. 5Rijndam Rehabilitation, Rotterdam, The Netherlands
  6. 6Department of Orthopedic Surgery and Traumatology, AZ Monica, Antwerpen, Belgium
  7. 7Department of Orthopedic Surgery, University Hospital Antwerp, Edegem, Belgium
  8. 8StatUa Center for Statistics, University of Antwerp, Antwerpen, Belgium
  9. 9Physiotherapy in Motion, Multi-Specialty Research Group (PTinMOTION), Department of Physical Therapy, University of Valencia, Valencia, Spain
  10. 10Department of Physiotherapy, Human Physiology and Anatomy (KIMA), Vrije Universiteit Brussel, Brussels, Belgium
  11. 11Department of Physiotherapy, University of Granada, Melilla, Spain
  12. 12Department of Nursing, Physiotherapy and Medicine, University of Almeria, Almeria, Spain
  13. 13Department of Physiotherapy, University of Malaga, Malaga, Spain
  14. 14Rehabilitation Sciences, Universiteit Antwerpen Campus Drie Eiken, Wilrijk, Belgium
  1. Correspondence to Professor Filip Struyf; filip.struyf{at}uantwerpen.be

Abstract

Introduction There is a large diversity in the clinical presentation of frozen shoulder (FS) and the clinical outcome is not always satisfactory. The aim of the current study was to examine to what extent range of motion (ROM) limitation, metabolic factors (diabetes mellitus and thyroid disorders), autonomic symptoms and pain sensitivity may contribute to the prognosis in terms of shoulder pain and disability and quality of life in patients with FS.

Methods Patients with stage 1 or 2 FS were longitudinally followed-up during 9 months after baseline assessment. They completed six questionnaires and underwent quantitative sensory testing (pressure pain thresholds, temporal summation and conditioned pain modulation) and ROM assessment.

Results One hundred and forty-nine patients with FS were initially recruited and 121 completed at least one follow-up measurement. Shoulder pain and disability improved over time and diabetes mellitus was found to be a prognostic factor for final outcome. Several domains of quality of life also improved over time and external rotation ROM, diabetes mellitus, thyroid disorder and autonomic symptoms were found to be prognostic factors for final outcome. These prognostic factors explained 2.5%–6.3% of the final outcome of shoulder pain and disability and quality of life.

Discussion and conclusion In patients with FS, prognostic variables were able to predict different outcomes, indicating that outcomes in this population can be variable-dependent. Other variables not explored in this study might contribute to the prognosis of patients with FS, which should be investigated in future research. In clinical practice, baseline assessment of prognostic factors and focusing on a more holistic approach might be useful to inform healthcare practitioners about progression of patients with FS during a 9-month period.

  • Shoulder
  • Musculoskeletal disorders
  • PRIMARY CARE
  • REHABILITATION MEDICINE
  • RHEUMATOLOGY

Data availability statement

Data are available on reasonable request. Data are available on reasonable request at the corresponding author: filip.struyf@uantwerpen.be.

http://creativecommons.org/licenses/by-nc/4.0/

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/.

https://doi.org/10.1136/bmjopen-2021-056563

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    STRENGTHS AND LIMITATIONS OF THIS STUDY

    • A multicentre design with a large sample at final follow-up was performed.

    • A more holistic assessment, easily applicable and similar to clinical settings was used for patient examination.

    • Diagnosis was based only on recommended diagnostic criteria.

    • Blood glucose levels were not verified with objective tools.

    Introduction

    Frozen shoulder (FS) is a medical enigma, which is difficult to understand and manage. It is characterised by spontaneous onset of shoulder pain and a gradual decrease of both active and passive shoulder range of motion (ROM).1 The prevalence of primary FS in the general population is 2%–5%.1 2 No intervention has demonstrated superior efficacy for FS and disease duration does not appear to be influenced by any treatment.3 The natural history varies between 1 and 3 years, although there is an incomplete recovery in 7%–50% of patients, who maintain a slightly painful and restricted shoulder at long-term follow-ups.4–7 Indeed, 6% of patients with FS still have severe symptoms at more than 7 years.4 An explanation for this large variety in disease duration might be related to an incomplete understanding of both the most effective treatment8–10 and natural history of FS11–13 as well as no proper formulation of treatment success criteria.14 Another possible explanation might be that existing research is more focused on population and intervention averages rather than the experience of individual patients.15 For instance, a slight improvement for the majority of patients after a treatment could be judged as ineffective, while being beneficial for a small group of patients. Contrary, a treatment could be judged effective while it can have adverse effects for example, activating the inflammatory reaction, further decreasing ROM, for a small group of patients.15

    From other pathologies, such as fibromyalgia, low back pain and musculoskeletal pain, it is known that there are subgroups of patients with different characteristics or clinical profiles.16–23 This knowledge about subgroups has changed the treatment approach of several pathologies, resulting in a more effective, patient-tailored approach.24–28 In FS, some studies are already advising the use of interventions based on subgroups as well, thus increasing the likelihood of more efficient treatment approaches.4 29–31 Massive mobility deficits, muscle weakness, presence of comorbidities and unbearable symptoms during the first 6 months after onset predicted a worse prognosis in patients with FS.4 30 31

    Since there is a large diversity in the clinical presentation of an individual patient with FS and the clinical outcome is not always satisfactory, identification of factors that might predict outcome would be beneficial to tailor treatment. In a previously published review, some prognostic factors were proposed.32 In particular, patients with diabetes mellitus (DM) or thyroid disorders have a 5–7 times higher risk of developing FS.33 Indeed, the prevalence of DM and thyroid disorders in patients with FS ranges from 3% to 41%.34–37 Furthermore, the prevalence of FS increases up to 40% in patients with DM36 38 39 and 10.9% in patients with thyroid disorders.40 However, it is unclear what the effect is of DM in the prognosis of treatment outcome in patients with FS, with few studies demonstrating conflicting results.41–43

    Furthermore, chronic low-grade inflammation seems to play an important role in the pathophysiology of FS. It is suggested that the nervous, endocrine and immune system function interdependently and that a disturbance in one system disturbs another system.44 As chronic low-grade inflammation might disturb the immune system, the nervous system could be disturbed as well. The autonomic nervous system is at least partially involved in the release of proinflammatory and anti-inflammatory cytokines45 and might therefore be involved with chronic inflammation. Proinflammatory cytokines, which are released during an inflammatory state, are able to stimulate the vagal nerve and cholinergic anti-inflammatory pathway, resulting in an increase of hyperalgesia in the central nervous system (CNS).46 Therefore, inflammation would interact with both dysautonomia (a condition in which the autonomic nervous system does not work properly)47 48 and CNS dysfunction48 and these dysfunctions might be present in patients with FS. Dysautonomia may result in altered cardiorespiratory, thermoregulatory, gastrointestinal and bladder function.49 Altered central pain processing is reflected by a hypersensitivity (or amplification response) to sensory input (eg, innocuous, noxious or repeated stimuli) and changes in CNS function (ie, glial activation).50 51 Another factor favouring altered central pain processing is persistent nociceptive input,52 53 which might be the case in patients with FS.

    Identifying prognostic factors might contribute to the reformation of current treatment strategies to improve patients’ outcomes. Consequently, a patient-tailored approach could result in a better prognosis. Therefore, the aim of this study was to determine to what extent local (shoulder ROM), metabolic (DM and thyroid), autonomic and central pain processing (hyperalgesia and endogenous function) factors might predict prognosis in patients with FS.

    In summary, metabolic factors (DM and thyroid disorder), autonomic dysfunction and/or altered central pain processing may play a more dominant role in patients with FS than currently thought, although this hypothesis needs to be tested. Therefore, we hypothesise that patients with smaller shoulder ROM limitations, absence of DM, thyroid disorders and/or self-reported autonomic symptoms and/or normal central pain processing at baseline will have a better prognosis for shoulder pain and disability and quality of life over 9 months follow-up.

    Methods

    Study design

    Patient and public involvement

    Patients or the public were not involved in the design, recruitment, conduct, reporting or dissemination plans of our research. Dissemination of the results to participants was done by sending an email with all general results.

    Participants

    Patients with FS were recruited at the orthopaedic departments of the University Hospitals of Valencia and Malaga (Spain) and AZ Monica campus Deurne (Belgium), and through general practitioner practices in Antwerp (Belgium). The eligibility criteria are presented in table 1.

    View this table:
    Table 1

    Eligibility criteria

    Procedure

    First, patients completed a general demographic survey and four questionnaires: Shoulder Pain and Disability Index (SPADI), Visual Analogue Scale, 36-item Short Form Health Survey (SF-36) and Composite Autonomic Symptom Score-31. These questionnaires have shown to be valid and reliable.54–57

    Second, quantitative sensory testing (QST) including pressure pain threshold (PPT), temporal summation (TS) and conditioned pain modulation (CPM) was performed as a proxy for central pain processing. Finally, ROM of both shoulders was determined. Patients were examined by six physical therapists, all previously trained by two physical therapists with more than 10 years of experience in the examination of shoulder disorders and QST measurements.

    All these measurements were repeated at 3, 6 and 9 months follow-up after baseline measurement. For the current research question only the follow-up scores from the SPADI and SF-36 were used in the analysis. The remaining data will be used to provide information about the natural history of FS in another paper.

    Quantitative sensory testing

    Pressure pain thresholds

    Mechanical hyperalgesia was assessed by determining PPTs. The PPTs were measured on the affected side at the centre of the belly of the anterior deltoid (2 cm below the acromion) and quadriceps muscle (middle point between the anterior superior iliac spine and superior edge patella as described previously58 while the patient was seated on a bench without arm rests. The quadriceps site was chosen to explore hyperalgesia at a remote area (widespread hyperalgesia), which might indicate altered central pain processing.51 Mechanical pressure pain was measured on the above-mentioned sites in an Excel-generated random order using a digital algometer with a rubber tip of 1 cm2 (Wagner Force Dial FDX 50, Wagner Instruments, Greenwich, USA). To determine the PPT, the assessor applied a gradually increasing pressure at a speed of 1 kg/s until the patient experienced the stimulus as annoying and uncomfortable. Two PPT measurements were performed at each site with a rest interval of 30 s and the mean was used for analysis. PPT was found to be reliable.59–63

    Temporal summation

    TS was measured following the procedure described by Cathcart et al.64 A train of 10 repeated pressure stimuli was applied at the quadriceps muscle belly using a digital algometer. The pressure used for TS was the mean PPT previously determined for the quadriceps. Patients rated the pain intensity on a Numeric Pain Rating Scale (NPRS), ranging from 0 (no pain) to 10 (worst imaginable pain) perceived during the 1st, 5th and 10th repetition of the train stimuli. TS was calculated as the difference between the tenth and the first pulse as previously reported.64

    CPM at the unaffected upper arm

    CPM provides information about the efficacy of descending pain modulation pathways and was evaluated in this study by examining the effect of a conditioning stimulus on a test stimulus. The test stimulus consisted of mechanical pressure stimuli elicited as described above in the PPTs section. The conditioning stimulus was provided by means of an ischaemic occlusion applied to the unaffected arm with an inflatable air cuff (Boso Profitest). The cuff was positioned just above the cubital fossa and inflated until the patient experienced the stimulus as annoying and uncomfortable. Thirty seconds after the application of the inflatable cuff, the patient was requested to rate the intensity of perceived pain of the conditioning stimulus on an NPRS. Next, the cuff pressure was adapted (ie, increased or decreased) until the patient experienced a pain intensity of three on the NPRS. Then the PPT was repeated at the deltoid muscle as described above. The cuff was deflated immediately after the second test stimulus. The CPM effect was calculated following the formula: Embedded Image, with negative values implying an anti-nociceptive effect and positive values a pronociceptive effect. This method was found to be reliable and effective in evaluating descending pain modulation pathways.64 65

    Shoulder active ROM

    Shoulder ROM was measured with a gravity referenced analogue plurimeter (Dr Jules Rippstein) into the directions of external rotation (in 0° abduction), flexion and abduction. Both the affected and unaffected sides were measured. The side measured first was allocated based on an Excel-generated random sequence. For measuring shoulder external rotation, the inclinometer was attached to the dorsal side of the forearm and patients, while lying supine, were asked to perform shoulder external rotation while keeping the shoulder at 0° of abduction.

    During the assessment of shoulder flexion and abduction, patients were sitting on the bench with the hips and knees flexed 90° and the feet flat on the floor. The plurimeter was attached to the upper arm just below the insertion of the deltoid muscle. Shoulder flexion and abduction were performed in the sagittal and frontal plane, respectively, with the thumb pointing upwards.

    Statistics

    The group with at least one follow-up measurement was compared with the complete sample at baseline by providing mean difference (95% CI). To determine the contribution of 10 candidate prognostic factors (external rotation, flexion and abduction ROM, DM, thyroid disorder, autonomic (dys)function and PPT at the shoulder and quadriceps, TS and CPM) and the change over time, linear mixed models for each individual candidate prognostic factor were fitted (as recommended66) using restricted maximal likelihood. This regression technique accounts for the dependence between observations from the same individual, and can include individuals with incomplete follow-up data.67 In this model, outcome scores at the three follow-up measurements (3, 6 and 9 months follow-up) were included to determine the effect of candidate prognostic factors. To adjust for covariates, the following factors were added to the model: baseline score of the dependent variable (ie, SPADI or SF-36), age, gender, treatment category and demographic region. To account for the dependence between measurements from the same individual, the individual identifier was entered as random effect. Shoulder pain and disability (SPADI) and quality of life (eight domains of SF-36) at 9 months follow-up were entered as dependent variables. When a candidate prognostic factor was found significant, both the marginal and conditional R2 were reported as recommended.68 Significance was set at α<0.05.

    Statistical analysis was performed in R (V.3.6.2, Vienna Austria). Graphs were created with ggplot2.69 Linear mixed models were fitted using add-on packages lme470 and lmerTest.71

    Results

    Subjects

    Figure 1 shows the flow diagram of this study. Initially 149 patients were included, 121 patients completed at least one follow-up measurement and 88 completed all follow-up measurements. Patient characteristics of the full sample at baseline and patients completing at least one follow-up measurement are presented in table 2.

    Figure 1
    Figure 1

    Patient flow diagram during follow-up measurements.

    View this table:
    Table 2

    Patient characteristics at baseline and for the sample with at least one follow-up measurement

    None of the characteristics shows a strong difference between the total sample and the sample analysed, making it quite unlikely that the results presented here have been biased due to attrition bias.

    Development over time

    Figure 2 shows the mean scores and 95% CIs for the SPADI for all follow-up measurements and the mean scores for the eight domains of the SF-36 for all follow-up measurements.

    Figure 2
    Figure 2

    Mean scores for the dependent variables with the 95% CI for shoulder pain and disability.

    Table 3 shows the strength and direction of the analysis over time and indicates a significant improvement for the SPADI (F=32.274, p<0.001) and for the physical functioning (F=3.187, p=0.044), physical problem (F=3.176, p=0.045), emotional problem (F=5.768, p=0.004) and vitality (F=11.126, p<0.001) domains of the SF-36. Post hoc comparison reveals a difference from 3 months follow-up to 6 and 9 months follow-up (p<0.01) for the SPADI and vitality domain and from 6 to 9 months follow-up for the SPADI and emotional problem domain (p<0.03). No post hoc differences were found for the physical functioning and physical problem domains of the SF-36 (p>0.05). Time explained 1.5%–5.9% of the outcome over 9 months follow-up.

    View this table:
    Table 3

    Results of time analysis

    Prognostic factors

    Table 4 shows the strength and direction of the candidate prognostic factor analysis. Active external rotation was found to be a prognostic factor for the physical functioning (F=11.203, p=0.001) and pain (F=4.082, p=0.048) domains of the SF-36 and explained 2.6%–6.3% of the outcome over 9 months follow-up. DM was found to be a prognostic factor for SPADI (F=4.936, p=0.030) and the physical functioning (F=4.156, p=0.046) domain of the SF-36 and explained 2.5%–2.6% of the outcome over 9 months follow-up. Thyroid disorder was found to be a prognostic factor for the emotional problem (F=4.650, p=0.035) domain and explained 2.9% of the outcome over 9 months follow-up. Finally, autonomic symptoms were found to be a prognostic factor for the emotional problem (F=8.228, p=0.006), mental health (F=9.488, p=0.003), vitality (F=6.675, p=0.012), pain (F=4.007, p=0.049) and general health (F=5.378, p=0.024) domains of the SF-36 and explained 2.5%–5.0% of the outcome over 9 months follow-up.

    View this table:
    Table 4

    Results of the prognostic factor analysis for SPADI and the eight domains of quality of life (SF-36)

    Discussion

    Our results showed that shoulder pain and disability decreased, and quality of life (reflected by emotional problem and vitality domains) increased over 9 months follow-up in patients with FS. There were different factors that influenced clinical outcomes over 9 months follow-up adjusted for baseline scores, age, gender, treatment received and geographical region. Only DM was found to be a prognostic factor for shoulder pain and disability, while for the physical functioning domain of the SF-36, external rotation ROM and presence of DM were found to be prognostic factors. Furthermore, for the emotional problem domain presence of thyroid disorder and self-reported autonomic symptoms were found to be prognostic factors and for mental health and vitality domains of the SF-36, self-reported autonomic symptoms were found to be a prognostic factor. For the pain domain, active external rotation and self-reported autonomic symptoms were found to be prognostic factors and for the general health domain, self-reported autonomic symptoms were found to be a prognostic factor.

    In this study, a significant improvement in shoulder pain and disability and quality of life (vitality domain) was found from 3 months to 6 and 9 months. Shoulder pain and disability and quality of life (emotional problem domain) improved from 6 months to 9 months follow-up. There seems to be rather an early increase (from 3 months follow-up to 6 and 9 months follow-up) observed in the vitality domain of quality of life that slowed with time, while the emotional problem domain improved only from 6 to 9 months. This delayed improvement might be a consequence of improved shoulder pain and disability and vitality.

    Prognostic factors

    In this study, no factors were found to be prognostic for all outcomes.

    Higher levels of external rotation ROM at baseline resulted in a worse score on the physical functioning and pain domain of the SF-36. This is an unexpected finding, as we hypothesised that less ROM would result in worse scores. It might be possible that patients with more movement restriction receive more treatment focused on this restriction, however, our data do not contain this detailed information to confirm this hypothesis. Contrary to our results, Yang et al72 found a prognostic value of shoulder external rotation ROM on shoulder function. This difference might be explained by the fact that Yang et al72 provided standardised treatment including mobilisation and stretching techniques and this study was an observational study without standardised treatment.

    Presence of DM was found to be a prognostic factor for worse improvement of shoulder pain and disability and the physical functioning domain of the SF-36. A recent systematic review41 reported conflicting evidence regarding DM as a possible prognostic factor for influencing clinical outcomes in patients with FS. This is confirmed by our results, as it was found to be a prognostic factor in only one domain of quality of life.

    Presence of thyroid disorder was found to be a prognostic factor for improvement in the emotional problem domain. This unexpected finding might be explained because patients with thyroid disorder could have already gone through a period with emotional problems and be able to put the complaints of FS into perspective more easily.

    The presence of more self-reported autonomic symptoms at baseline resulted in a worse prognosis over 9 months follow-up in terms of quality of life (physical functioning, emotional problem and mental health domain). These autonomic symptoms might be the result of the interaction between the nervous, endocrine and immune system. It is suggested that a disturbance in one of these systems will lead to a disturbance in another system.44 Since the pathogenesis of FS is thought to be one of chronic inflammation,48 73 the immune system may be disturbed, resulting in a disturbance of the autonomic nervous system. Whether these autonomic symptoms were already present before the development of FS or developed simultaneously with the FS complaints is unknown. However, autonomic symptoms appear to be present before other disorders such as DM or rheumatoid arthritis develop.74 If this is the case in patients with FS as well, autonomic symptoms may be considered a risk factor for the development of FS.

    Variance explained

    All the prognostic factors investigated in this study explained between 2.5% and 6.3% of clinical outcomes over 9 months follow-up, which would indicate there are more variables that contribute to shoulder pain and disability and to the different domains of quality of life in patients with FS. Other studies have suggested that muscle strength,30 number of comorbidities30 and scapular movement72 could contribute to the prognosis in this population. In addition, there is conflicting evidence regarding the role of gender, age and duration of symptoms and preliminary evidence for onset pain intensity to be prognostic factors for disability in patients with FS.41 Interestingly, the latter factor was also found in a longitudinal study about the long-term outcomes of FS.4 Future studies should include these variables to determine whether all of them in isolation or in combination could provide even better prognostic models for patients with FS than those provided so far.

    Clinical implications and suggestions for further research

    Our results indicate that increased external rotation ROM and presence of DM and autonomic symptoms at baseline may result in a worse prognosis of patients with FS. This information may be useful to tailor treatment in this population. For instance, if more autonomic symptoms are present, it could be beneficial in patients with FS to target the autonomic nervous system (ie, vagal nerve stimulation or yoga). Vagal nerve stimulation and yoga have been found to be effective for improving disease activity and sympathovagal balance in patients with rheumatoid arthritis75 76 and vagal nerve stimulation improved pain and disability in patients with osteoarthritis.75 However, the presence of autonomic symptoms in this study was assessed with a self-reported questionnaire and although this questionnaire is valid and reliable,77 more objective tools to assess the autonomic nervous system, such as heart rate and blood pressure variability,49 should be investigated in patients with FS before implementing specific interventions to treat these symptoms.

    Based on our findings, we suggest considering a more holistic assessment than focusing solely on the shoulder in patients with FS. It seems that DM, the autonomic nervous system and possibly some elements of central pain processing might play a role in the prognosis of patients with FS. This might be important to tailor treatment strategies. However, more evidence of the presence of these prognostic factors is needed and the results of this study need to be replicated in future studies. When these results are confirmed the assessment of these factors becomes more important.

    Strengths

    The greatest strength of this study is the multicentre design with different geographical locations. This provides multiple benefits when compared with a single centre design. One of those strengths is the large sample, even after dropouts, which has been used to determine prognostic factors for FS. Other benefits of this approach are decreased personal bias, stronger statistical data analysis and larger generalisability of the results. Furthermore, the assessment done in this study is convenient to apply in clinical settings and so it is easy to examine the relevant prognostic factors. Some assessment elements are common practice (eg, ROM, pain, disabilities), while others are not routinely performed (eg, self-reported autonomic symptoms). The assessment in this study was about more than simply joints and muscles. With the inclusion of elements of the autonomic and CNS and metabolic factors, we investigated a more holistic assessment approach. Finally, the stage of the FS condition was standardised to patients only in stage 1 or 2 as reflected by our inclusion criteria, minimising the influence of disease duration.

    Limitations

    Besides the strengths, some limitations need to be acknowledged as well. First, the diagnosis FS was based on recommended diagnostic criteria, but the presence of potential other disorders (ie, osteoarthritis) was not ruled out with imaging. Second, the diagnosis of DM was made by an endocrinologist and was used as the dichotomy present/absent. There was no verification of blood glucose levels with objective tools (ie, blood glucose measurement tool), so patients with prediabetic levels might have been given the level ‘absent’ and this might have influenced the results. The analysis of glucose levels in patients with FS is therefore recommended for future studies. Lastly, there were relatively high numbers of patients lost to follow-up. There were 121 patients with at least one follow-up measurement, however, only 88 patients (59%) completed all follow-up measurements. Therefore, results need to be interpreted with caution. Nevertheless, with the current sample, we are able to detect a correlation of 0.22 between candidate prognostic factors and final outcome with a power of 0.80 and a significance level of 0.05.

    Conclusion

    External rotation ROM, and presence of DM, thyroid disorder and autonomic symptoms at baseline emerged to be prognostic factors for shoulder pain, disability and quality of life over 9 months follow-up in patients with FS. These factors explained 2.5%–6.3% of the variance of those outcomes with adjustment for several covariates (ie, baseline score shoulder pain and disability and quality of life respectively, age, gender, treatment category and geographical region). There might be additional prognostic factors, such as muscle strength or number of comorbidities that might be important as suggested by previous studies. All these factors together should be investigated in future research to determine whether these factors in isolation or in combination have prognostic value for clinical outcomes in patients with FS. Finally, further research could investigate whether specific treatments targeting DM and autonomic nervous system symptoms at baseline result in better prognosis. Meanwhile, treatment of patients with FS should be performed according to clinical practice guidelines and assessment should focus on a more holistic approach to signal possible additional treatment goals.

    Data availability statement

    Data are available on reasonable request. Data are available on reasonable request at the corresponding author: filip.struyf@uantwerpen.be.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This multicenter prospective longitudinal study was performed in the research laboratories of the University of Valencia (Spain), University of Malaga (Spain), and the University of Antwerp (Belgium). The study was approved by the Human/Clinical Research Ethics Committees of the University of Valencia (H1432625002427), Hospital Costa del Sol (001_abr17_PI2) and University Hospital Antwerp (B300201422072). Data were collected between November 2014 and October 2020. This study adhered to the Declaration of Helsinki.

    Acknowledgments

    We would like to thank all participants who volunteered to take part in our study.

    References

    1. ↵
      1. Zuckerman JD,
      2. Rokito A
      . Frozen shoulder: a consensus definition. J Shoulder Elbow Surg 2011;20:322–5.doi:10.1016/j.jse.2010.07.008pmid:http://www.ncbi.nlm.nih.gov/pubmed/21051244
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Kelley MJ,
      2. Shaffer MA,
      3. Kuhn JE, et al
      . Shoulder pain and mobility deficits: adhesive capsulitis. J Orthop Sports Phys Ther 2013;43:A1–31.doi:10.2519/jospt.2013.0302pmid:http://www.ncbi.nlm.nih.gov/pubmed/23636125
      OpenUrlPubMed
    3. ↵
      1. Neviaser AS,
      2. Hannafin JA
      . Adhesive capsulitis: a review of current treatment. Am J Sports Med 2010;38:2346–56.doi:10.1177/0363546509348048pmid:http://www.ncbi.nlm.nih.gov/pubmed/20110457
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Hand C,
      2. Clipsham K,
      3. Rees JL, et al
      . Long-term outcome of frozen shoulder. J Shoulder Elbow Surg 2008;17:231–6.doi:10.1016/j.jse.2007.05.009pmid:http://www.ncbi.nlm.nih.gov/pubmed/17993282
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Robinson CM,
      2. Seah KTM,
      3. Chee YH, et al
      . Frozen shoulder. J Bone Joint Surg Br 2012;94:1–9.doi:10.1302/0301-620X.94B1.27093pmid:http://www.ncbi.nlm.nih.gov/pubmed/22219239
      OpenUrlCrossRefPubMed
    6. ↵
      1. Kim DH,
      2. Kim YS,
      3. Kim B-S, et al
      . Is frozen shoulder completely resolved at 2 years after the onset of disease? J Orthop Sci 2020;25:224–8.doi:10.1016/j.jos.2019.03.011pmid:http://www.ncbi.nlm.nih.gov/pubmed/30952550
      OpenUrlPubMed
    7. ↵
      1. Shaffer B,
      2. Tibone JE,
      3. Kerlan RK
      . Frozen shoulder. A long-term follow-up. J Bone Joint Surg Am 1992;74:738–46.doi:10.2106/00004623-199274050-00013pmid:http://www.ncbi.nlm.nih.gov/pubmed/1624489
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Cho C-H,
      2. Bae K-C,
      3. Kim D-H
      . Treatment strategy for frozen shoulder. Clin Orthop Surg 2019;11:249–57.doi:10.4055/cios.2019.11.3.249pmid:http://www.ncbi.nlm.nih.gov/pubmed/31475043
      OpenUrlPubMed
    9. ↵
      1. Yip M,
      2. Francis A-M,
      3. Roberts T, et al
      . Treatment of adhesive capsulitis of the shoulder: a critical analysis review. JBJS Rev 2018;6:e5.doi:10.2106/JBJS.RVW.17.00165pmid:http://www.ncbi.nlm.nih.gov/pubmed/29916942
      OpenUrlPubMed
    10. ↵
      1. Georgiannos D,
      2. Markopoulos G,
      3. Devetzi E, et al
      . Adhesive capsulitis of the shoulder. is there consensus regarding the treatment? A comprehensive review. Open Orthop J 2017;11:65–76.doi:10.2174/1874325001711010065pmid:http://www.ncbi.nlm.nih.gov/pubmed/28400876
      OpenUrlPubMed
    11. ↵
      1. Wong CK,
      2. Levine WN,
      3. Deo K, et al
      . Natural history of frozen shoulder: fact or fiction? A systematic review. Physiotherapy 2017;103:40–7.doi:10.1016/j.physio.2016.05.009pmid:http://www.ncbi.nlm.nih.gov/pubmed/27641499
      OpenUrlCrossRefPubMed
    12. ↵
      1. Ryan V,
      2. Brown H,
      3. Minns Lowe CJ, et al
      . The pathophysiology associated with primary (idiopathic) frozen shoulder: a systematic review. BMC Musculoskelet Disord 2016;17:340.doi:10.1186/s12891-016-1190-9pmid:http://www.ncbi.nlm.nih.gov/pubmed/27527912
      OpenUrlCrossRefPubMed
    13. ↵
      1. Eljabu W,
      2. Klinger HM,
      3. von Knoch M
      . Prognostic factors and therapeutic options for treatment of frozen shoulder: a systematic review. Arch Orthop Trauma Surg 2016;136:1–7.doi:10.1007/s00402-015-2341-4pmid:http://www.ncbi.nlm.nih.gov/pubmed/26476720
      OpenUrlCrossRefPubMed
    14. ↵
      1. Kelley MJ,
      2. McClure PW,
      3. Leggin BG
      . Frozen shoulder: evidence and a proposed model guiding rehabilitation. J Orthop Sports Phys Ther 2009;39:135–48.doi:10.2519/jospt.2009.2916pmid:http://www.ncbi.nlm.nih.gov/pubmed/19194024
      OpenUrlCrossRefPubMed
    15. ↵
      1. Andrew Moore R
      . What works for whom? Determining the efficacy and harm of treatments for pain. Pain 2013;154 Suppl 1:S77–86.doi:10.1016/j.pain.2013.03.024pmid:http://www.ncbi.nlm.nih.gov/pubmed/23622761
      OpenUrlPubMed
    16. ↵
      1. Yim Y-R,
      2. Lee K-E,
      3. Park D-J, et al
      . Identifying fibromyalgia subgroups using cluster analysis: relationships with clinical variables. Eur J Pain 2017;21:374–84.doi:10.1002/ejp.935pmid:http://www.ncbi.nlm.nih.gov/pubmed/27633925
      OpenUrlPubMed
    17. ↵
      1. Williams TE,
      2. Chalder T,
      3. Sharpe M, et al
      . Heterogeneity in chronic fatigue syndrome - empirically defined subgroups from the PACE trial. Psychol Med 2017;47:1454–65.doi:10.1017/S0033291716003615pmid:http://www.ncbi.nlm.nih.gov/pubmed/28112075
      OpenUrlPubMed
    18. ↵
      1. Smart KM,
      2. Blake C,
      3. Staines A, et al
      . Self-reported pain severity, quality of life, disability, anxiety and depression in patients classified with 'nociceptive', 'peripheral neuropathic' and 'central sensitisation' pain. The discriminant validity of mechanisms-based classifications of low back (±leg) pain. Man Ther 2012;17:119–25.doi:10.1016/j.math.2011.10.002pmid:http://www.ncbi.nlm.nih.gov/pubmed/22074733
      OpenUrlCrossRefPubMed
    19. ↵
      1. Nijs J,
      2. Van Houdenhove B,
      3. Oostendorp RAB
      . Recognition of central sensitization in patients with musculoskeletal pain: application of pain neurophysiology in manual therapy practice. Man Ther 2010;15:135–41.doi:10.1016/j.math.2009.12.001pmid:http://www.ncbi.nlm.nih.gov/pubmed/20036180
      OpenUrlCrossRefPubMed
    20. ↵
      1. Lentz TA,
      2. Marlow NM,
      3. Beneciuk JM, et al
      . Comorbidity subgroups among medicare beneficiaries seeking health care for musculoskeletal pain. J Gerontol A Biol Sci Med Sci 2019;74:1310–5.doi:10.1093/gerona/gly202pmid:http://www.ncbi.nlm.nih.gov/pubmed/30165549
      OpenUrlPubMed
    21. ↵
      1. Stynes S,
      2. Konstantinou K,
      3. Dunn KM
      . Classification of patients with low back-related leg pain: a systematic review. BMC Musculoskelet Disord 2016;17:226.doi:10.1186/s12891-016-1074-zpmid:http://www.ncbi.nlm.nih.gov/pubmed/27215590
      OpenUrlPubMed
    22. ↵
      1. Asih S,
      2. Mayer TG,
      3. Williams M, et al
      . Does classification of chronic musculoskeletal disorder patients into psychosocial subgroups predict differential treatment responsiveness and 1-year outcomes after a functional restoration program? Clin J Pain 2015;31:1036–45.doi:10.1097/AJP.0000000000000216pmid:http://www.ncbi.nlm.nih.gov/pubmed/25621427
      OpenUrlPubMed
    23. ↵
      1. Drew BT,
      2. Conaghan PG,
      3. Smith TO, et al
      . Toward the development of data-driven diagnostic subgroups for people with Patellofemoral pain using modifiable clinical, biomechanical, and imaging features. J Orthop Sports Phys Ther 2019;49:536–47.doi:10.2519/jospt.2019.8607pmid:http://www.ncbi.nlm.nih.gov/pubmed/31213159
      OpenUrlPubMed
    24. ↵
      1. Malfliet A,
      2. Kregel J,
      3. Coppieters I, et al
      . Effect of pain neuroscience education combined with cognition-targeted motor control training on chronic spinal pain: a randomized clinical trial. JAMA Neurol 2018;75:808–17.doi:10.1001/jamaneurol.2018.0492pmid:http://www.ncbi.nlm.nih.gov/pubmed/29710099
      OpenUrlPubMed
    25. ↵
      1. Lera S,
      2. Gelman SM,
      3. López MJ, et al
      . Multidisciplinary treatment of fibromyalgia: does cognitive behavior therapy increase the response to treatment? J Psychosom Res 2009;67:433–41.doi:10.1016/j.jpsychores.2009.01.012pmid:http://www.ncbi.nlm.nih.gov/pubmed/19837206
      OpenUrlCrossRefPubMed
    26. ↵
      1. Nyström B,
      2. Svensson E,
      3. Larsson S, et al
      . A small group Whiplash-Associated-Disorders (WAD) patients with central neck pain and movement induced stabbing pain, the painful segment determined by mechanical provocation: fusion surgery was superior to multimodal rehabilitation in a randomized trial. Scand J Pain 2016;12:33–42.doi:10.1016/j.sjpain.2016.03.003pmid:http://www.ncbi.nlm.nih.gov/pubmed/28850488
      OpenUrlPubMed
    27. ↵
      1. Yosmaoğlu HB,
      2. Selfe J,
      3. Sonmezer E, et al
      . Targeted treatment protocol in Patellofemoral pain: does treatment designed according to subgroups improve clinical outcomes in patients unresponsive to multimodal treatment? Sports Health 2020;12:170–80.doi:10.1177/1941738119883272pmid:http://www.ncbi.nlm.nih.gov/pubmed/31750786
      OpenUrlPubMed
    28. ↵
      1. Karpinski K,
      2. Müller-Rath R,
      3. Niemeyer P, et al
      . Subgroups of patients with osteoarthritis and medial meniscus tear or crystal arthropathy benefit from arthroscopic treatment. Knee Surg Sports Traumatol Arthrosc 2019;27:782–96.doi:10.1007/s00167-018-5086-0pmid:http://www.ncbi.nlm.nih.gov/pubmed/30128683
      OpenUrlPubMed
    29. ↵
      1. Yang J-lan,
      2. Jan M-H,
      3. Chang C-wei, et al
      . Effectiveness of the end-range mobilization and scapular mobilization approach in a subgroup of subjects with frozen shoulder syndrome: a randomized control trial. Man Ther 2012;17:47–52.doi:10.1016/j.math.2011.08.006pmid:http://www.ncbi.nlm.nih.gov/pubmed/21944980
      OpenUrlPubMed
    30. ↵
      1. Pease B,
      2. Ross M
      . Defining subgroups of patients with a stiff and painful shoulder: an analytical model using cluster analysis. Disabil Rehabil 2021;43:1–8.doi:10.1080/09638288.2019.1631891pmid:http://www.ncbi.nlm.nih.gov/pubmed/31268741
      OpenUrlPubMed
    31. ↵
      1. Yang J-lan,
      2. Chang C-wei,
      3. Chen S-yee, et al
      . Mobilization techniques in subjects with frozen shoulder syndrome: randomized multiple-treatment trial. Phys Ther 2007;87:1307–15.doi:10.2522/ptj.20060295pmid:http://www.ncbi.nlm.nih.gov/pubmed/17684085
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Struyf F,
      2. Meeus M
      . Current evidence on physical therapy in patients with adhesive capsulitis: what are we missing? Clin Rheumatol 2014;33:593–600.doi:10.1007/s10067-013-2464-3pmid:http://www.ncbi.nlm.nih.gov/pubmed/24374758
      OpenUrlPubMed
    33. ↵
      1. Milgrom C,
      2. Novack V,
      3. Weil Y, et al
      . Risk factors for idiopathic frozen shoulder. Isr Med Assoc J 2008;10:361–4.pmid:http://www.ncbi.nlm.nih.gov/pubmed/18605360
      OpenUrlPubMed
    34. ↵
      1. Tzeng C-Y,
      2. Chiang H-Y,
      3. Huang C-C, et al
      . The impact of pre-existing shoulder diseases and traumatic injuries of the shoulder on adhesive capsulitis in adult population: a population-based nested case-control study. Medicine 2019;98:e17204.doi:10.1097/MD.0000000000017204pmid:http://www.ncbi.nlm.nih.gov/pubmed/31574829
      OpenUrlPubMed
    35. ↵
      1. Schiefer M,
      2. Teixeira PFS,
      3. Fontenelle C, et al
      . Prevalence of hypothyroidism in patients with frozen shoulder. J Shoulder Elbow Surg 2017;26:49–55.doi:10.1016/j.jse.2016.04.026pmid:http://www.ncbi.nlm.nih.gov/pubmed/27424251
      OpenUrlPubMed
    36. ↵
      1. Zreik NH,
      2. Malik RA,
      3. Charalambous CP
      . Adhesive capsulitis of the shoulder and diabetes: a meta-analysis of prevalence. Muscles Ligaments Tendons J 2016;6:26–34.doi:10.11138/mltj/2016.6.1.026pmid:http://www.ncbi.nlm.nih.gov/pubmed/27331029
      OpenUrlCrossRefPubMed
    37. ↵
      1. Inayat F,
      2. Ali NS,
      3. Shahid H, et al
      . Prevalence and determinants of frozen shoulder in patients with diabetes: a single center experience from Pakistan. Cureus 2017;9:e1544.doi:10.7759/cureus.1544pmid:http://www.ncbi.nlm.nih.gov/pubmed/29018641
      OpenUrlPubMed
    38. ↵
      1. Prodromidis AD,
      2. Charalambous CP
      . Is there a genetic predisposition to frozen shoulder?: a systematic review and meta-analysis. JBJS Rev 2016;4. doi:doi:10.2106/JBJS.RVW.O.00007. [Epub ahead of print: 23 02 2016].pmid:http://www.ncbi.nlm.nih.gov/pubmed/27490134
    39. ↵
      1. Juel NG,
      2. Brox JI,
      3. Brunborg C, et al
      . Very high prevalence of frozen shoulder in patients with type 1 diabetes of ≥45 years' duration: the dialong shoulder study. Arch Phys Med Rehabil 2017;98:1551–9.doi:10.1016/j.apmr.2017.01.020pmid:http://www.ncbi.nlm.nih.gov/pubmed/28219686
      OpenUrlPubMed
    40. ↵
      1. Cakir M,
      2. Samanci N,
      3. Balci N, et al
      . Musculoskeletal manifestations in patients with thyroid disease. Clin Endocrinol 2003;59:162–7.doi:10.1046/j.1365-2265.2003.01786.xpmid:http://www.ncbi.nlm.nih.gov/pubmed/12864792
      OpenUrlCrossRefPubMed
    41. ↵
      1. MGCAM M,
      2. Struyf F,
      3. Meert L
      . Factors influencing treatment outcome of physical therapy in frozen shoulder patients: a systematic review. Eur J Physiother 2022;24:174–90.doi:10.1080/21679169.2020.1827029
      OpenUrl
    42. ↵
      1. Pushpa C,
      2. Saradakutty G,
      3. Das S, et al
      . Long-term outcome in adhesive capsulitis associated with type 2 diabetes. J Evol Med Dent Sci 2020;9:3783–6.doi:10.14260/jemds/2020/830
      OpenUrl
    43. ↵
      1. Dyer BP,
      2. Burton C,
      3. Rathod-Mistry T, et al
      . Diabetes as a prognostic factor in frozen shoulder: a systematic review. Arch Rehabil Res Clin Transl 2021;3:100141.doi:10.1016/j.arrct.2021.100141pmid:http://www.ncbi.nlm.nih.gov/pubmed/34589691
      OpenUrlPubMed
    44. ↵
      1. Chapman CR,
      2. Tuckett RP,
      3. Song CW
      . Pain and stress in a systems perspective: reciprocal neural, endocrine, and immune interactions. J Pain 2008;9:122–45.doi:10.1016/j.jpain.2007.09.006pmid:http://www.ncbi.nlm.nih.gov/pubmed/18088561
      OpenUrlCrossRefPubMedWeb of Science
    45. ↵
      1. Zila I,
      2. Mokra D,
      3. Kopincova J, et al
      . Vagal-immune interactions involved in cholinergic anti-inflammatory pathway. Physiol Res 2017;66:S139–45.doi:10.33549/physiolres.933671pmid:http://www.ncbi.nlm.nih.gov/pubmed/28937230
      OpenUrlPubMed
    46. ↵
      1. Watkins LR,
      2. Maier SF
      . The pain of being sick: implications of immune-to-brain communication for understanding pain. Annu Rev Psychol 2000;51:29–57.doi:10.1146/annurev.psych.51.1.29pmid:http://www.ncbi.nlm.nih.gov/pubmed/10751964
      OpenUrlCrossRefPubMedWeb of Science
    47. ↵
      1. Mani R,
      2. Cooper C,
      3. Kidd BL, et al
      . Use of laser Doppler flowmetry and transcutaneous oxygen tension electrodes to assess local autonomic dysfunction in patients with frozen shoulder. J R Soc Med 1989;82:536–8.doi:10.1177/014107688908200910pmid:http://www.ncbi.nlm.nih.gov/pubmed/2677369
      OpenUrlCrossRefPubMedWeb of Science
    48. ↵
      1. Pietrzak M
      . Adhesive capsulitis: an age related symptom of metabolic syndrome and chronic low-grade inflammation? Med Hypotheses 2016;88:12–17.doi:10.1016/j.mehy.2016.01.002pmid:http://www.ncbi.nlm.nih.gov/pubmed/26880627
      OpenUrlPubMed
    49. ↵
      1. Wehrwein EA,
      2. Orer HS,
      3. Barman SM
      . Overview of the anatomy, physiology, and pharmacology of the autonomic nervous system. Compr Physiol 2016;6:1239–78.doi:10.1002/cphy.c150037pmid:http://www.ncbi.nlm.nih.gov/pubmed/27347892
      OpenUrlPubMed
    50. ↵
      1. van den Broeke EN
      . Central sensitization and pain hypersensitivity: some critical considerations. F1000Res 2018;7:1325.doi:10.12688/f1000research.15956.2pmid:http://www.ncbi.nlm.nih.gov/pubmed/30416715
      OpenUrlPubMed
    51. ↵
      1. den Boer C,
      2. Dries L,
      3. Terluin B, et al
      . Central sensitization in chronic pain and medically unexplained symptom research: a systematic review of definitions, operationalizations and measurement instruments. J Psychosom Res 2019;117:32–40.doi:10.1016/j.jpsychores.2018.12.010pmid:http://www.ncbi.nlm.nih.gov/pubmed/30665594
      OpenUrlPubMed
    52. ↵
      1. Gracely RH,
      2. Schweinhardt P
      . Programmed symptoms: disparate effects United by purpose. Curr Rheumatol Rev 2015;11:116–30.doi:10.2174/1573397111666150619095125pmid:http://www.ncbi.nlm.nih.gov/pubmed/26088212
      OpenUrlPubMed
    53. ↵
      1. Harte SE,
      2. Harris RE,
      3. Clauw DJ
      . The neurobiology of central sensitization. J Appl Biobehav Res 2018;23:e12137.doi:10.1111/jabr.12137
    54. ↵
      1. Roy J-S,
      2. MacDermid JC,
      3. Woodhouse LJ
      . Measuring shoulder function: a systematic review of four questionnaires. Arthritis Rheum 2009;61:623–32.doi:10.1002/art.24396pmid:http://www.ncbi.nlm.nih.gov/pubmed/19405008
      OpenUrlCrossRefPubMedWeb of Science
    55. ↵
      1. Karcioglu O,
      2. Topacoglu H,
      3. Dikme O, et al
      . A systematic review of the pain scales in adults: which to use? Am J Emerg Med 2018;36:707–14.doi:10.1016/j.ajem.2018.01.008pmid:http://www.ncbi.nlm.nih.gov/pubmed/29321111
      OpenUrlPubMed
    56. ↵
      1. VanderZee KI,
      2. Sanderman R,
      3. Heyink JW, et al
      . Psychometric qualities of the Rand 36-Item health survey 1.0: a multidimensional measure of general health status. Int J Behav Med 1996;3:104–22.doi:10.1207/s15327558ijbm0302_2pmid:http://www.ncbi.nlm.nih.gov/pubmed/16250758
      OpenUrlCrossRefPubMedWeb of Science
    57. ↵
      1. Treister R,
      2. O'Neil K,
      3. Downs HM, et al
      . Validation of the composite autonomic symptom scale 31 (COMPASS-31) in patients with and without small fiber polyneuropathy. Eur J Neurol 2015;22:1124–30.doi:10.1111/ene.12717pmid:http://www.ncbi.nlm.nih.gov/pubmed/25907824
      OpenUrlPubMed
    58. ↵
      1. Kuppens K,
      2. Struyf F,
      3. Nijs J, et al
      . Exercise- and stress-induced hypoalgesia in musicians with and without shoulder pain: a randomized controlled crossover study. Pain Physician 2016;19:59–68.pmid:http://www.ncbi.nlm.nih.gov/pubmed/26815250
      OpenUrlPubMed
    59. ↵
      1. Lewis GN,
      2. Heales L,
      3. Rice DA, et al
      . Reliability of the conditioned pain modulation paradigm to assess endogenous inhibitory pain pathways. Pain Res Manag 2012;17:98–102.doi:10.1155/2012/610561pmid:http://www.ncbi.nlm.nih.gov/pubmed/22518372
      OpenUrlCrossRefPubMedWeb of Science
    60. ↵
      1. Walton DM,
      2. Macdermid JC,
      3. Nielson W, et al
      . Reliability, standard error, and minimum detectable change of clinical pressure pain threshold testing in people with and without acute neck pain. J Orthop Sports Phys Ther 2011;41:644–50.doi:10.2519/jospt.2011.3666pmid:http://www.ncbi.nlm.nih.gov/pubmed/21885906
      OpenUrlCrossRefPubMed
    61. ↵
      1. Walton DM,
      2. Levesque L,
      3. Payne M, et al
      . Clinical pressure pain threshold testing in neck pain: comparing protocols, responsiveness, and association with psychological variables. Phys Ther 2014;94:827–37.doi:10.2522/ptj.20130369pmid:http://www.ncbi.nlm.nih.gov/pubmed/24557645
      OpenUrlAbstract/FREE Full Text
    62. ↵
      1. Vanderweeën L,
      2. Oostendorp RAB,
      3. Vaes P, et al
      . Pressure algometry in manual therapy. Man Ther 1996;1:258–65.doi:10.1054/math.1996.0276pmid:http://www.ncbi.nlm.nih.gov/pubmed/11440515
      OpenUrlCrossRefPubMed
    63. ↵
      1. Farasyn A,
      2. Meeusen R
      . Pressure pain thresholds in healthy subjects: influence of physical activity, history of lower back pain factors and the use of endermology as a placebo-like treatment. J Bodyw Mov Ther 2003;7:53–61.doi:10.1016/S1360-8592(02)00050-5
      OpenUrlCrossRef
    64. ↵
      1. Cathcart S,
      2. Winefield AH,
      3. Rolan P, et al
      . Reliability of temporal summation and diffuse noxious inhibitory control. Pain Res Manag 2009;14:433–8.doi:10.1155/2009/523098pmid:http://www.ncbi.nlm.nih.gov/pubmed/20011713
      OpenUrlCrossRefPubMed
    65. ↵
      1. Mertens MG,
      2. Hermans L,
      3. Crombez G, et al
      . Comparison of five conditioned pain modulation paradigms and influencing personal factors in healthy adults. Eur J Pain 2021;25:243–56.doi:10.1002/ejp.1665pmid:http://www.ncbi.nlm.nih.gov/pubmed/32965727
      OpenUrlPubMed
    66. ↵
      1. Steyerberg EW,
      2. Moons KGM,
      3. van der Windt DA, et al
      . Prognosis research strategy (progress) 3: prognostic model research. PLoS Med 2013;10:e1001381.doi:10.1371/journal.pmed.1001381pmid:http://www.ncbi.nlm.nih.gov/pubmed/23393430
      OpenUrlCrossRefPubMed
    67. ↵
      1. Fitzmaurice GM,
      2. Laird NM,
      3. Ware JH
      . Applied longitudinal analysis. 2nd ed. Wiley, 2012.
    68. ↵
      1. Nakagawa S,
      2. Schielzeth H
      . A general and simple method for obtaining R 2 from generalized linear mixed-effects models. Methods Ecol Evol 2013;4:133–42.doi:10.1111/j.2041-210x.2012.00261.x
      OpenUrlCrossRefPubMed
    69. ↵
      1. Wickham H
      . ggplot2: elegant graphics for data analysis. Springer International Publishing, 2016.
    70. ↵
      1. Bates D,
      2. Mächler M,
      3. Bolker B, et al
      . Fitting linear mixed-effects models using lme4. J Stat Softw 2015;67:48.doi:10.18637/jss.v067.i01
    71. ↵
      1. Kuznetsova A,
      2. Brockhoff PB,
      3. Christensen RHB
      . lmerTest package: tests in linear mixed effects models. J Stat Softw 2017;82:26.doi:10.18637/jss.v082.i13
    72. ↵
      1. Yang J-lan,
      2. Chang C-wei,
      3. Chen S-yee, et al
      . Shoulder kinematic features using arm elevation and rotation tests for classifying patients with frozen shoulder syndrome who respond to physical therapy. Man Ther 2008;13:544–51.doi:10.1016/j.math.2007.07.006pmid:http://www.ncbi.nlm.nih.gov/pubmed/17913564
      OpenUrlPubMed
    73. ↵
      1. Hand GCR,
      2. Athanasou NA,
      3. Matthews T, et al
      . The pathology of frozen shoulder. J Bone Joint Surg Br 2007;89:928–32.doi:10.1302/0301-620X.89B7.19097pmid:http://www.ncbi.nlm.nih.gov/pubmed/17673588
      OpenUrlCrossRefPubMed
    74. ↵
      1. Koopman FA,
      2. van Maanen MA,
      3. Vervoordeldonk MJ, et al
      . Balancing the autonomic nervous system to reduce inflammation in rheumatoid arthritis. J Intern Med 2017;282:64–75.doi:10.1111/joim.12626pmid:http://www.ncbi.nlm.nih.gov/pubmed/28547815
      OpenUrlCrossRefPubMed
    75. ↵
      1. Courties A,
      2. Berenbaum F,
      3. Sellam J
      . Vagus nerve stimulation in musculoskeletal diseases. Joint Bone Spine 2021;88:105149.doi:10.1016/j.jbspin.2021.105149pmid:http://www.ncbi.nlm.nih.gov/pubmed/33548494
      OpenUrlPubMed
    76. ↵
      1. Ganesan S,
      2. Gaur GS,
      3. Negi VS, et al
      . Effect of yoga therapy on disease activity, inflammatory markers, and heart rate variability in patients with rheumatoid arthritis. J Altern Complement Med 2020;26:501–7.doi:10.1089/acm.2019.0228pmid:http://www.ncbi.nlm.nih.gov/pubmed/32326727
      OpenUrlPubMed
    77. ↵
      1. Sletten DM,
      2. Suarez GA,
      3. Low PA, et al
      . COMPASS 31: a refined and abbreviated composite autonomic symptom score. Mayo Clin Proc 2012;87:1196–201.doi:10.1016/j.mayocp.2012.10.013pmid:http://www.ncbi.nlm.nih.gov/pubmed/23218087
      OpenUrlCrossRefPubMed
      1. Hannafin JA,
      2. Chiaia TA
      . Adhesive capsulitis. A treatment approach. Clin Orthop Relat Res 2000;372:95–109.pmid:http://www.ncbi.nlm.nih.gov/pubmed/10738419
      OpenUrlCrossRefPubMed

    Footnotes

    • Twitter @michelmertens4, @FilipStruyf

    • FS and LD contributed equally.

    • Contributors MicM: data acquisition (including recruitment), data curation, formal statistical analysis, interpretation, writing—original draft and editing, responsible for overall content; MirM: conceptualisation, methodology, interpretation, writing—review and editing, supervision; SN: conceptualisation, data acquisition (including recruitment), writing—review and editing; OV: data acquisition (including recruitment), writing—review and editing; EF: formal statistical analysis, writing—review and editing; ELG: conceptualisation, methodology, data acquisition (including recruitment), writing—review and editing; MAR: data acquisition (including recruitment), writing—review and editing; SN-L: data acquisition (including recruitment), data curation, writing—review and editing; MF-S: data acquisition (including recruitment), writing—review and editing; AL-S: data acquisition (including recruitment), writing—review and editing, supervision. FS: conceptualisation, methodology, interpretation, writing—review and editing, supervision; LD: data acquisition (including recruitment), conceptualisation, methodology, data curation, writing—review and editing, supervision. All authors approved the final version and agreed to be accountable for the work published.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.