J Manipulative Physiol Ther 2010 (Feb); 33 (2): 96–101
Gert J. Bergman, PhD, Jan C. Winters, MD, PhD, Klaas H. Groenier, PhD, Betty Meyboom-de Jong, MD, PhD, Klaas Postema, MD, PhD, Geert J. van der Heijden, PhD
Senior Research Manager, Formerly Department of General Practice and Center for Rehabilitation, University Medical Center Groningen, The Netherlands
Objective The purpose of this study was to examine the effect of manipulative therapy on the shoulder girdle, in addition to usual care provided by the general practitioner, on the outcomes of physical examination tests for the treatment of shoulder complaints.
Methods This was a randomized controlled trial in a primary care setting in the Netherlands. A total of 150 participants were recruited from December 2000 until December 2002. All patients received usual care by the general practitioner. Usual care included one or more of the following depending on the needs of the patient: information/advice, oral analgesics or nonsteroidal antiinflammatory drugs, corticosteroid injections, exercises, and massage. In addition to usual care, the intervention group received manipulative therapy, up to 6 treatment sessions in a 12-week period. Twenty-four physical examination tests were done at baseline and after 6, 12, and 26 weeks. Factor analysis was done to reduce the number of outcome measures.
Results The factor analysis resulted in 4 factors: “shoulder pain,” “neck pain,” “shoulder mobility,” and “neck mobility.” At 6 weeks, no significant differences between groups were found. At 12 weeks, the mean changes of all 4 factors favored the intervention group; the factors “shoulder pain” and “neck pain” reached statistical significance (95% confidence interval [CI], 0.1-2.1). At 26 weeks, differences in the factors “shoulder pain” (95% CI, 0.0-2.6), “shoulder mobility” (95% CI, 0.2-1.7), and “mobility neck” (95% CI, 0.2-1.3) statistically favored the intervention group.
Conclusion In this pragmatic study, manipulative therapy, in addition to usual care by the general practitioner, diminished severity of shoulder pain and neck pain and improved shoulder and neck mobility.
From the FULL TEXT Article:
Shoulder complaints are characterized by disability, usually due to pain during shoulder movement and restricted range of motion. Complaints of the neck and/or dysfunction of the joints of the cervical spine, the upper thoracic spine, and the adjacent upper ribs (shoulder girdle) often accompany shoulder complaints and are an important factor in duration and/or recurrence of shoulder complaints.
In clinical practice, a dysfunction of the shoulder girdle can be treated by manipulative therapy, which aim is to restore normal functioning of the shoulder girdle. [1, 2] To date, with only 1 randomized trial favoring manipulative therapy for the shoulder girdle, the evidence for the effectiveness of manipulative treatment in the treatment of shoulder complaints is scarce.  Therefore, we conducted a randomized trial to study the effect of manipulative therapy for the shoulder girdle in addition to usual care by the general practitioner in the treatment of shoulder complaints. The design of this study and the main patient-experienced results are already published. [4, 5] The results indicate that additional manual therapy for the structures of the shoulder girdle accelerates recovery of patient-experienced shoulder symptoms and reduces their severity. In the present article, the results for the physical examination outcome measures are presented.
In the clinical research of musculoskeletal complaints, physical testing of pain and mobility by the physician are important outcomes. [3, 6, 7] However, this concerns mostly multiple physical examination tests and multiple outcome measures. This requires multiple statistical testing. Together with small study sizes (more outcomes than patients), this may lead to spurious significant results from randomized trials affecting the interpretability of the outcome of the trial.
The challenge is to reduce the number of variables in such a way that they are clinically sensible and statistically manageable. To overcome the aforementioned problems with pain and mobility as outcome measures of our randomized trial, we tried to reduce the individual physical examination tests for pain and mobility to relevant components. In this study, we used a physical assessment of pain and mobility of the shoulder and shoulder girdle as outcome measures. They consist primarily of the assessment of active and passive limitations in shoulder movement and pain experienced during these movements and a physical examination of the cervicothoracic spine, consisting of passive movements of the neck and pain experienced in these movements. We used factor analysis to identify relevant components from these variables thereby reducing the number of outcome measures in a clinically meaningful sense and to increase statistical power. The purpose of this study was to examine the effect of manipulative therapy on the shoulder girdle, in addition to usual care provided by the general practitioner, on the outcomes of physical examination tests for the treatment of shoulder complaints.
The 4-factor solution of the factor analysis indicated that separate tests concerning pain and mobility of the glenohumeral joint and neck actually can be clustered in clinically recognizable and statistically relevant factors, indicating glenohumeral pain, neck pain, glenohumeral mobility, and neck mobility.
For none of the 4 identified factors, our trial shows a clinically relevant or statistically significant effect of additional manipulative therapy at 6 weeks after randomization. At 12 weeks after randomization, patients who received manipulative therapy in addition to usual care had a small but statistically significant lower score on the factors “shoulder pain” and “neck pain” (factor 2) than patients who received usual care by the general practitioner only. At 26 weeks after randomization, the difference between groups in scores on the factors “shoulder pain” (factor 1), “shoulder mobility” (factor 2), and “neck mobility” (factor 4) favored manipulative therapy, and all were statistically significant. These results are mainly congruent with other main outcome measures of the trial, such as patient-perceived recovery, severity of the main complaint, and shoulder disability, which are reported elsewhere.  In short, during treatment (6 weeks), no significant differences were found. After completion of treatment (12 weeks), 43% of the patients receiving additional manipulative therapy reported full recovery and 21% in the control group. At 26 weeks, recovery rates were about 40% in both groups. There was a significant difference in favor of manual therapy between the intervention groups concerning the main complaint and a borderline significant difference was found in shoulder disability. 
The factors “shoulder pain” and “neck pain” were significantly different at week 12, whereas the other factors “neck mobility” and “shoulder mobility” were only significantly different at week 26 might raise questions whether these outcomes are primarily the effect of the manual therapy. However, all 4 factors favored additional manual therapy in all 3 follow-up measurements. We did a strict randomized study with limited loss of follow-up. Patients with missing values were evenly distributed between treatment groups. Still, missing values and imputation according to the last-observation-carried-forward may have influenced our results. Unfortunately, we cannot compare our results with other studies regarding the physical examination outcomes of manual therapy of the shoulder girdle in shoulder complaints.
There were 150 subjects included in this study in the Netherlands; therefore, it is possible that findings in this study may not necessarily be generalizable to other patient populations. Because the treatment protocols were not standardized for each patient, it is not clear what part of the treatment may have been causing clinical improvement. In this study, we used a set of physical tests, based on treatment guidelines and tests commonly used in clinical practice, which is a selection of all available tests to evaluate pain and mobility of the shoulder. To date, it is unclear, however, which physical test(s) should be used in scientific research or clinical practice to evaluate changes. The results show that the factors derived from our trial data are quite robust with respect to their content and statistical efficiency. Although, it remains to be shown whether these factors will sustain in other studies.
In clinical trials concerning treatment of shoulder complaints, factor analysis is useful for the reduction of multiple outcomes of physical examination data and therefore increases statistical power. On the basis of the factors derived from physical examination tests of the shoulder and the cervicothoracic spine, we conclude that manipulative therapy, in addition to usual care by the general practitioner, diminishes the severity of the pain in the shoulder and neck and improves the mobility of the shoulder and the cervicothoracic spine. Results were most prominent at 26 weeks after initiation of treatment.
Funding Sources and Potential Conflicts of Interest
This study was funded by the Netherlands Organization for Scientific Research (NWO grant no. 904-65-901). No conflicts of interest were reported for this study.
Factor analysis may be useful for the reduction of multiple outcomes of physical examination data and therefore increases statistical power.
Manipulative therapy, in addition to usual care by the general practitioner, diminishes the severity of the pain in the shoulder and neck and improves the mobility of the shoulder and the cervicothoracic spine.