Research Review by Dr. Shawn Thistle©


Jan. 2007

Study Title:

Development of a clinical prediction rule for guiding treatment of a subgroup of patients with neck pain: Use of thoracic spine manipulation, exercise, and patient education


Cleland JA et al.

Publication Information:

Physical Therapy 2007; 87(1): 9-23.


Neck pain is very common, with a lifetime incidence of ~20-70%. Although not as common as lower back pain, neck pain presents a higher risk for developing chronic pain (~30%) than LBP (~10%). Cervical manipulation is a common treatment for mechanical neck pain. Simply stated, the risk of insult to neck vasculature is a concern with this treatment, despite being extremely rare. Further, the evidence for pre-manipulative screening to identify those at risk for such incidents is contradictory and very weak (that is, we have no readily available clinical tools to help us identify those at risk). A study published last year (also led by Josh Cleland) suggested that thoracic manipulation can be effective for reducing neck pain. Widely regarded as a safer intervention (keeping in mind the aforementioned low risks of cervical manipulation), thoracic manipulation could provide an alternative, or complementary treatment to cervical manipulation and other therapies aimed at mechanical neck complaints.

A clinical prediction rule (CPR) consists of combinations of variables from the patient history, physical examination, and in some cases pen and paper outcome measures that identify subgroups of patients who are likely to benefit from a particular treatment. The authors of this study have previously developed and validated a CPR for low back pain. Now with this study, they are beginning to develop a similar tool for mechanical neck pain. As stated in other reviews, I feel this type of research is extremely important for optimizing patient care and safety, while assisting manual therapists to match appropriate treatments with certain patients.

This prospective cohort study evaluated consecutive patients who were referred for physical therapy for neck pain. Each subject underwent a standardized history and physical examination, followed by a series of standardized thoracic spine manipulations. Outcomes were dichotomized as successful or non-successful based on perceived improvement (rather than perceived disability, which is often used in such studies).

Seventy-eight subjects were included in this study, all meeting the following criteria:
  • age 18-65
  • primary complaint of neck pain with or without unilateral upper extremity symptoms
  • baseline Neck Disability Index (NDI) score of at least 10%
Exclusion criteria were as follows:
  • identification of medical “red flags”
  • history of “whiplash” injury within 6 weeks of the examination
  • evidence of CNS involvement
  • signs consistent with nerve root compression
Four physical therapists performed the examinations and thoracic manipulations after a participating in a training/review session. All subjects received a standardized treatment regimen or three different types of spinal manipulation regardless of physical examination findings:
  1. ”Distraction Manipulation” – With the subject seated the therapist placed his upper chest at the targeted segment and grasped the patients elbows. An upward thrust to distract the thoracic spine was then performed.
  2. ”Upper Thoracic Spine Manipulation” – The patient lays supine with arms clasped behind their neck. The therapist uses the manipulative hand to stabilize the vertebra below the targeted segment (between T1-T4) while thrusting down through the patient’s arms (basically an “anterior thoracic” adjustment)
  3. ”Middle Thoracic Spine Manipulation” – Performed as #2 above with targeted vertebra at or below T5
Each manipulation was performed twice, for a total of six thrusts per session, regardless of whether an audible cavitation was heard on the first thrust. All subjects were also instructed in a cervical ROM exercise (10 repetitions, 3-4 times per day), and were advised to maintain their normal daily activities within the limits of pain. Up to three treatment sessions were performed at least 2 days apart.

The Global Rating of Change served as the reference criterion for determining a successful outcome. The GROC is a 15 point rating scale from -7 (“a very great deal worse”) to +7 (“a very great deal better”). It was pre-determined that patients would need to score at least +5 at the second session to be categorized as having a successful outcome. If this was not achieved, a third session was performed after which the GROC was measured again.

Pertinent Results:

  • 42 of 78 subjects met the criteria for a successful outcome, while 36 had a non-successful outcome
  • 23 subjects had a successful outcome after the first treatment (55%), with the rest having a successful outcome after only 2 treatments
  • 10 potential predictors with a significance level less than 0.10 were entered into a logistic regression – with 6 retained for the final regression model
  • the six items were: 1) symptom duration < 30 days 2) no symptoms distal to the shoulder 3) subject reports that looking up does not aggravate symptoms 4) Fear Avoidance Beliefs Questionnaire (FABQ) physical activity scale score of < 12 5) diminished upper thoracic spine kyphosis (T3-T5) and 6) cervical extension of < 30°
  • 14/15 subjects who were positive on at least 4 of the above items and 32/37 who were positive on at least 3 items were in the successful group
  • of the 41 subjects with 2 or fewer positive items, 31 were in the non-successful outcome group
  • the pretest likelihood of a successful outcome was 54% (42/78) – this increased to 93% (likelihood ration 12.0) if 4/6 items were positive, and 86% (likelihood ratio 5.5) if 3/6 items were positive

Conclusions & Practical Application:

This study, although small, provides an important step in the development of a CPR for neck pain. The six factors identified in this study can now be validated in a larger patient sample, and compared with other potential factors.

I feel it is important to highlight that there are some striking similarities between the factors identified in this study for neck pain (albeit preliminary), and those already validated for low back pain. In general, short duration of symptoms with no radiation past the shoulder or knee seem to be helpful in predicting those likely to respond to lumbar and thoracic manipulation for low back and neck pain respectively.

Further, the third and sixth items listed above (no aggravation while looking up and < 30° of cervical extension) seem to agree with one of the results of the study by Tseng published in Manual Therapy last year.

They identified “not feeling worse while extending the neck” as one of the factors associated with a positive outcome immediately after neck manipulation. (Note: I feel these findings in combination begin to cast doubt over the utility of the commonly used “Kemp’s” position and its variants – which many practitioners use to identify joint dysfunction that may be amenable to joint manipulation)

Previous literature has identified a biomechanical link between mobility in the cervicothoracic junction and neck pain. The predictors of decreased cervical extension and reduced thoracic kyphosis in the upper thoracic spine suggests that a meaningful relation may exist, which warrants further study.

Further, it is worthwhile to note that patients who report symptom aggravation while looking up who also have symptoms distal to the shoulder may represent a population of patients with cervical radiculopathy, rather than mechanical neck pain. As such, it stands to reason that they may not obtain as much benefit from thoracic manipulation.

Future studies are required to validate these results, including long-term follow-up and comparison groups to further identify the predictive value of these, and potentially other items.