Better Shoulder Outcomes with Spinal Manipulative Therapy

Abstract & Commentary

Synopsis: Adding early manual manipulation of the cervicothoracic spine for shoulder pain or dysfunction disorders, in addition to regular care of information, advice and medications, produced much better recovery at both 12- and 52-weeks post treatments.

Source: Bergman, GJ et al. Ann Intern Med. 2004;141: 432-439.

This prospective, randomized controlled trial from The Netherlands sought to test the contribution of spinal "manipulative therapy" to the standardized treatments currently employed by the Dutch College of General Practitioners. The annual incidence of shoulder symptoms in that country is estimated at 10-25 per 1000 enrolled patients.

Current Dutch treatment guidelines for "pain between the neck and the elbow at rest or during movement of the upper arm" recommend 2 weeks of information and advice about home exercises and limited daily use, along with oral analgesics and NSAIDs. This is followed when needed by up to 3 corticosteroid shoulder injections at 2 week intervals with Triamcinolone 40mg and Lidocaine 10 mg. After 6 weeks of persistent symptoms, physical therapy referral is made for exercises, massage and physical applications.

In this study, all 150 participants received the usual care noted above, and half were randomized to concurrently receive 6 spinal treatment sessions over 12 weeks by physiotherapists registered to provide orthopedic manipulative therapies. Cases of chronic pain, severe trauma and dislocations were excluded. Manipulative techniques were standardized and involved generally low-amplitude, high-velocity thrust techniques for all, and high-amplitude, low-velocity thrusts when indicated at specific locations to decrease any restrictions in movement in the spine and adjacent ribs. Direct shoulder therapy and massage/exercises were discouraged as being a protocol deviation.

Outcomes were measured primarily by patient surveys and secondarily by analyzing pain severity scores and functional disability. Improved scores at 6 weeks favored the manipulation group, but did not reach statistical significance for full recovery until 12 weeks (43% vs 21%).

At the end of one year after therapy began, 52% of manipulated patients considered themselves recovered, vs 35% of the control group who received usual care.

Comment by Mary Elina Ferris, MD

It’s hard to interpret human studies that involve physical interventions, since the treatment group cannot be blinded and obviously know that they are specially selected for a desired improvement. Relying on the patient’s perception as the outcome measure is open to bias, and this study does not have any objective measurement of improvement other than the patient’s impression. The intervention group had traditional physical therapy excluded (possibly to control the number of variables), so it is impossible to determine if there was anything unique to spinal manipulation compared to early institution of other physical treatments.

Furthermore, numbers are small, definitions of specifically which shoulder disorders are being treated are lacking, and manipulative therapies varied depending on the nature of the complaint and the therapist involved. Recovery rates among the individual therapists ranged from 14% to 83%, and there still were almost half of the manipulated group that did not consider themselves completely recovered at the end of the study (compared to 65% of the control group).

Nonetheless, more patients did feel better and considered themselves fully recovered in the intervention group, and isn’t that what we are aiming for? Traditional allopathic medicine’s disdain for unproven spinal manipulative techniques may change after the recent publication of 4 years of HMO claim data involving 1 million back pain patients, showing reduction in costs and radiology, surgery and hospitalization with access to chiropractic care.1 An accompanying editorial, however, points out that lack of a cause-effect relationship, especially since those with chiropractic insurance tended to be younger and healthier.2

My conclusion from both the study reviewed above and the new beneficial chiropractic economic analyses1 is that there is an important role for "hands-on" treatments in musculoskeletal complaints, but for which treatments and for which conditions is far from clear. We should continue to work towards better definitions of these troublesome and pervasive problems, hopefully in collaboration with our physically inclined colleagues.

Dr. Ferris, Clinical Associate Professor, University of Southern California, is Associate Editor of Internal Medicine Alert.


1. Legorreta AP et al. Arch Intern Med. 2004;164: 1985-1992.

2. Ness J et al. Arch Intern Med. 2004;164:1953-54.