What Determines a Positive Outcome of Spinal Manipulation for Persistent Low Back Pain: Stiffness or Pain Sensitivity? A Randomized Trial

1.0 Background 1.1 Treatment of persistent low back pain Low back pain (LBP) is one of the leading causes of years lived with disability. In Denmark, many patients are treated conservatively in private practice via general practitioner, chiropractor and physiotherapist. It is well known that LBP often is recurrent or chronic condition. Patients with LBP often report pain relief following spinal manipulation although clinical and research findings are mixed . There is clear evidence that spinal manipulation is able to give short term pain relief, increase range of motion and the clinical outcomes are comparable to other common interventions for persistent LBP. Unfortunately, there are no reliable means to determine beforehand whether spinal manipulation is likely to provide pain relief for a given patient and it is common practice to simply try a series of treatments to assess the effect.

1.2 Segmental dysfunctions Spinal manipulation is directed at spinal mechanical dysfunctions. However, reliably identifying mechanically dysfunctional segments has proven difficult. In an attempt to objectively quantify such dysfunctions, a mechanical spinal indentation apparatus,, the VerteTrack (VT) has been developed and when compared to the generally unreliable palpation findings of an experienced clinician, no correlation was found.

Without reliable and objective indicators of the presence of spinal dysfunctions, clinical trials of spinal manipulation are bound to include both patients with and without biomechanical dysfunction, thus potentially obscuring any effect of spinal manipulation in the treatment of such dysfunctions. This heterogeneity in the experimental groups could, at least in part, explain the conflicting evidence.

To address this, a preliminary study has compared quantified spinal stiffness measured by VT and the outcome of spinal manipulation which reported a positive correlation between objectively quantified spinal stiffness and clinical improvement following manipulation in a cohort of acute LBP patients. The study suggests that VT may potentially serve to identify a subgroup of patients with LBP, who also has a higher probability of a positive response to manipulative treatment.

1.3 Factors complicating treatment of persistent low-back pain After the onset of low-back pain, the central neural pathways which conduct, modulate and integrate the pain stimulus changes over time as pain persists and becomes more excitable: It is evident that the nociceptive receptors become hyperexcitable, leading to increased transmission of nociceptive sensory input, which in turn causes central hypersensitive and possibly an attenuation of descending pain-inhibiting mechanisms. Overall, this results in enhanced local pain responses (hyperalgesia) and sometimes widespread increased pain sensitivity (generalized hyperalgesia).

There is evidence that such changes occur in the early stages of LBP and that central sensitization develops in step with the progression to persistent pain.

In recent decades, Quantitative Sensory Testing (QST) has been used to examine and quantify the pain experience, including intensity, threshold, tolerance, modulation and more. These findings combined can illustrate generalized hyperalgesia which are indicative of underlying central sensitization and has been reported in persistent LBP.

Central sensitization appears to be reversible, at least in non-spinal conditions when the persistent peripheral actor is resolved. Also, it is predictive of both surgical and non-surgical (analgesic) treatment outcomes. Others however, report that central sensitization is not predictive of the clinical course of spinal pain. A recent systematic review regarding QST as a prognostic tool in LBP concluded that there is a sparsity of studies, which in general are of limited quality, thus the prognostic value of QST in LBP remains unclear.

To our knowledge, no studies have been published examining the prognostic or treatment modifying value of QST in relation to spinal manipulation of persistent LBP.

1.4 Objectives The primary aim of this study is to examine if spinal manipulation is more effective in regards to short term pain relief when directed at spinal segmental stiffness or segmental pain sensitivity in persistent LBP

Secondary aims of this study:

1. Does pain sensitivity and spinal stiffness change in the subgroup of patients who respond to treatment compared to non-responders, i.e. whether pain sensitivity and mechanical stiffness decrease in response to pain relief from spinal manipulation?
2. Does this possibly occur independent of clinical pain relief?
3. Does central sensitization predict short term pain relief of spinal manipulation in persistent LBP?
4. Does segmental spinal stiffness correlates with segmental spinal pain sensitivity in a persistent LBP population?

2.0 Methods 2.1 Design A randomized controlled trial consisting of a maximum of 155 participants with persistent LBP.

2.2 Participants All participants to be included in the cohort will be recruited from a multidisciplinary Spine Center located in Middelfart, Denmark, after being thoroughly examined by a clinician, specialized in spine disorders and diagnosed with persistent Non-specific LBP (NSLBP).

The Spine Center is a hospital department, which handles more than 15.000 new patient cases each year, about 80% of which are referred with LBP and treated, using both conservative as well as surgical interventions.

2.3 Technical description of quantitative measures 2.3.1 Spinal stiffness Spinal stiffness using VT: The participant is placed in the prone position as the device will direct a force perpendicular from the L1 to L5 level starting at a comfortable level below 5 N and progresses to a maximal of 60 N as linear displacement will be measured and both force and displacement will be used to calculate the stiffness. VT has shown to be reliable and may be possible to use in clinical practice.

As standard with VT, all of the lumbar landmarks will be located by using ultrasonography. Research have shown that it provides a more accurately identification than manual palpation and the skill is easy obtainable. The lumbar segments will be clearly marked with a black marker to ensure that the both SM, QST and VT is directed at the indicated lumbar segment.

2.3.2 Quantitative sensory testing Manual pressure algometry(MPA): PPT will be measured the spinal process at L1 to L5 level using a pressure algometer (Somedic pressure algometer with a 1 cm2 probe, model 2, Hørby, Sweden). Pressure will be applied manually with a near-constant velocity of approximately 50 kilopascal(kPa)/s until the participant indicates the pressure as becoming painful by pressing an indicator button connected to the algometer. PPT measurements will be repeated 3 times with approximately 10 second rest intervals. The head of the algometer will be placed such that repeat applications overlapped partly. If no pain has been elicited by 1000 kPa, this will be recorded as the PPT. If the first and second measurements are 1000 kPa, a third will not be performed. MPA is a validated and often-used method to asses pain sensitivity.

Thermal pain sensitivity (ThPS): ThPS will be measured with a Medoc TSA-II thermode stimulator at L1 to L5 level as an average of the final 4 measurements of a series of 5 stimulations (baseline 32 celsius, increase 1 celsius/second), with 10 second intervals.

The participant indicates when the stimulation is perceived as pain by pressing an indicator button, upon which the temperature returns to baseline (decrease 10 celsius/second). When measured at the spine and in the L5-dermatome ThPS have shown to have good to excellent reliability.

Central sensitization: will be evaluated using Computer-controlled cuff algometry (CCA) in combination with MPA and ThPS measured at the upper extremity:

CCA will be used to measure pain threshold (cPPT), pain tolerance (cPTT), temporal pain summation (TPS), pain intensity at time of termination (PTS) and conditioned pain modulation (CPM).

CCA directs the pain intensity by inflation of a tourniquet applied around the lower extremity. This allows for assessment of deep-tissue pain sensitivity as a stimulus-response curve. When compared with the pressure algometry it reduces the measurement bias as it is examiner independent. Likewise, CCA has been tested to be both reliable and sensitive as a manual pressure algometry.

The procedure of CCA includes a 13-cm wide silicone tourniquet cuff (VBM, Sulz, Germany) with an equal-sized proximal and distal chamber wrapped around the non-dominant and dominant gastrocnemius muscle 5 cm inferior to the tuberositas tibia. The cuff pressure is able to increase with a rate of 1 kPa/s in both chambers, the maximum pressure limit is 80 kPa. Furthermore, a 200-liter external air tank is attached and provides air to avoid loud noises from the system during the assessment.

Participants indicate the pain on a computerized electronic Visual analog scale(eVas) ("No pain" = 0 cm to "Worst pain imaginable" = 10 cm) sampled at a frequency of 10 Hz. The participants will be instructed to continuously rate the induced pain intensity from initial pain onset. The pressure at which the stimulus is first described as painful, will be noted as cPPT. Participants are instructed to terminate the test, by pressing an indicator button, when the pain becomes intolerable. This resets the cuff and the pressure is noted as cPTT. The pain intensity at time of termination is the cPTS measure. Next, a series of 4 test impulses of equal pressure will be applied with 10 second intervals to familiarize participants with the procedure. Finally, a series of 10 impulses are run with 1 second intervals, of same pressure intensity in order to measure TPS. All of the above stated will be measured at the patient's dominant lower leg.

CPM will be measured as the difference between the original cPPT and a new measure during a continuing pain impulse simultaneously at the non-dominant leg.

Furthermore CPM will be measured using ThPs at the ventral side of the right forearm, both before and after CPM stimuli from CCA

3.0 Statistics and data analysis 3.1 Power sampling Assuming an alpha-level of 0.05 and a beta-level of 0.8, a power analysis (t-test) reveals, that a minimum n=62 participants per group is needed to reveal an expected NRS difference between groups at 1.0 and with a standard deviation of 1.96. These figures are based on estimates from a previous study on a similar patient group Thus, for two groups, a maximum of 155 participants is needed, when factoring in a dropout rate of around 25%.

3.2 Data analysis: The data is continuous and the primary outcome will be analyzed via regression models. All secondary outcomes will be dichotomized into responders/non-responders depending on the change in NRS.

In a secondary analysis we will look at "time" and "segment treated" as possible factors that may contribute to statistical interaction relative to the primary outcome.