To our knowledge, we are the first to report on the within-group changes of pain scores from clinical trials of a single session of a manual therapy for chronic neck pain in adults. In their recent review, Bronfort et al18
did not address this issue directly and included only the study by Sloop et al54
. As that study acquired its outcomes three weeks after the intervention, they likely regarded it as a longitudinal study, even though only a single intervention was provided. Bronfort et al18
provided no summary of evidence nor any clinical recommendations regarding single-session studies.
In recent Cochrane reviews, Gross et al36,37,38
included 4 single-session studies, three of which50,54,55
were included in this review. Those authors also included the trial by Howe et al58
; however, this study did not include a group receiving only manipulation, thus we excluded it from our review. Gross et al36,37,38
conducted a pooled analysis of Vernon et al55
and Sloop et al54
, given that they were clinically comparable and not statistically heterogeneous
. Outcome measures were pooled despite the fact that Vernon et al55
obtained only pressure algometry readings, not pain scores, at 5 minutes post-intervention and included a mobilization group, while Sloop et al54
obtained pain outcome only at 3 weeks post-intervention. Despite these considerable problems, Gross et al36,37,38
concluded from their analysis that “when compared to a control (other treatments deemed to be ineffective), there was moderate evidence that single sessions [of manipulation or mobilization] did not result in short-term pain relief [pooled SMD −0.51 (95% CI: −1.10 to 0.07)]” for acute, sub-acute, or chronic neck pain.
Our review differs from these two previous reviews in a number of ways. First, our primary objective was an analysis of the within-group, as opposed to between-group change scores in groups having received a single session of the treatments of interest. Second, our interest was exclusively in studies of a single session. Third, the scope of our review was larger, involving massage, pressure therapies, and manual traction, leading to the inclusion of additional studies such as Yurkiw et al51
for manipulation and Sterling et al56
and Hanten et al49
for mobilization (the more recent studies of Haas et al53
and Martinez-Segura et al52
were not available at the time for either the reviews of Bronfort et al18
or Gross et al36,37,38
). Fourth, more sophisticated analyses such as pooling of effects were not conducted. Lastly, our review remained within the boundaries of studies of chronic neck pain treated with one or more forms of manual therapy.
The subjects included in these studies were relatively homogeneous, not only because of the application of our selection criteria but also because of the similarity of baseline pain severity. In addition, these studies appeared to be similar in regard to the types of treatments employed: manipulations were delivered as single-thrust procedures while mobilizations appeared to be relatively similar in application. Finally, most studies employed the same outcome measure of pain severity on a 100 mm scale (see below re: Farrar et al59
), although a few reported on pressure pain thresholds measured by algometry. Given these clinical and statistical similarities, and given the sufficiently high quality of the studies reviewed, we felt that it was tenable to summarize the results of the 4 manipulation studies that reported pre- and post-intervention VAS scores and then derive a mean change score. This was not possible for the mobilization studies.
The question then arises as to how to interpret our findings. In a previous review60
, we referred to the work of Farrar et al59
in establishing that the minimum clinically important difference in pain severity scores on the VAS as registered by a large variety of patients with chronic pain is 20 mm. This applies to participants in studies of courses of various treatments for their pain, courses extending over days and weeks. In the case of the present review, outcomes were obtained immediately following the interventions. We could not identify any study that provided a similar reference frame for studies of a single-session of treatment for chronic pain patients.
We did identify one study concerning single-session changes
in acute pain patients. Bird and Dickson61
studied emergency-room pain patients and reported that with initial pain VAS scores between 34–66, the minimal clinically important in-session difference in pain score was 17 [95% CI:13, 21] mm; for achieving the outcome of a lot less pain
, the mean score was 33 [24, 42] mm. The baseline pain scores of the 4 pre- and post-manipulation studies in our review50,51,52,53
lie in the same range as those of Bird and Dickson. Their level of minimal clinically important difference (17 mm) in a single emergency-room treatment session is similar to the average level of change (–18.94 [7.42,30.16] mm) that we found in these 4 studies. With Farrar et al59
, Vernon et al60
, and Bird and Dickson61
as references, we advocate that the mean change in pain score of −18.9 [7.42,30.16] mm immediately after a single manipulation in 4 similar studies50,51,52,53
represents a clinically important finding.
The findings with respect to manual pressure techniques and mobilization permit neither summarization nor simple interpretation. We believe that the results of the manual pressure techniques and mobilization studies are best left in the studies' descriptive mode for use by future clinical trial planners. The results provide some indication of the magnitudes of clinical change that can be expected in a single session of the various types of mobilizations included in these studies.
With respect to our approach to within-group analysis, it could be asked if it is appropriate to conduct such analyses from a set of published RCTs. Only one of the manipulation trials54
included a comparison between a form of manual therapy and a placebo control procedure (anamnestic valium only). Only one of the mobilization trials included a sham procedure56
. The remaining majority of trials are more properly seen as randomized comparative trials in which none of the subjects in these trials were blinded as to the form of treatment they received. Two studies compared manipulation to mobilization50,52
whereas two studies compared two different modes of manipulation51,53
We maintain that once the inter-group outcomes of manual therapy trials have been analyzed in standard systematic reviews18,36,37,38
, it then becomes appropriate to assess the magnitudes of change within each treatment group randomized to receive the therapy of interest and, if possible, to summarize the results among studies. In fact, several studies in this review only reported change scores53,54,55
. After hypothesis testing has been conducted, it is reasonable to assess these scores individually for clinical relevance. Our sub-group analysis only extends this exercise to the collective body of trials in this area of interest.
Our study has limitations. The major limitation was the lack of high-quality RCTs that included clearly identifiable patient groups undergoing single-session manual treatments. Consequently, our review included only 4 trials of spinal manipulative therapy, only 4 trials of spinal mobilization and only 1 trial of trigger point therapy. No trials were identified for massage therapy or traction. In terms of quality, only one manipulation trial and one mobilization trial included a placebo control or sham group.
Our quality assessment suggests that weaknesses were associated with concealment of allocation and in blinding of subjects and therapists. In addition, most studies failed to report the presence or absence of adverse reactions. Only Sloop et al54
and Haas et al53
incorporated a double-blind design, while in none of the other studies were either therapists or subjects blinded to the interventions. This is a common problem in clinical trials of manual therapies40
. There is an urgent need for a reliable sham procedure for future trials of these manual therapies, whether with single or multiple intervention sessions.
In addition, not all trials in this review used pain severity as measured on a 100-mm scale as the main outcome. Some used pressure algometry or a mix of outcomes. There were also variations in post-measurement for the main outcome measure. Some outcomes were measured immediately, others at 5 hours or 3 weeks after the single manual therapy intervention. Calculations such as effect sizes and comparisons of manual therapies were made more complicated as a result of differences in the outcomes measured, data reported (continuous versus ordinal; means or median scores; varied measures of variability; change scores versus pre- and post-intervention scores) and the timing of measurements (post-treatment).
Another problem was the differences in baseline pain severity scores for different trials, which made it difficult to pool patient treatment groups for the various manual therapies. In terms of inclusion criteria, chronic neck pain varied from one to three months depending on the trial.
Finally, as discussed previously, the minimal clinically important difference (MCID) in pain severity immediately following a single session of manual therapy has not been defined for manual therapies. For this review, work by Bird and Dickson61
in emergency-room pain patients was used. More studies are needed to identify the MCID for manual therapies after a single session.