"Low back pain (LBP) is a common threat to medicine and a reasonable threat to all national health care systems.... Reducing ineffective treatments is necessary to decrease the LBP associated costs." Left out of von Heymann, et al.'s bold opening to their double-blind, placebo-controlled study, "Spinal High-Velocity Low-Amplitude Manipulation in Acute Nonspecific Low Back Pain," is that appropriate patient care also demands the replacement of ineffective treatments with better alternatives. Fortunately, the study itself provides fuel for that premise, finding spinal manipulation more effective than a nonsteroidal anti-inflammatory drug and placebo in patients with acute LBP. The authors' findings are particularly impressive because the study consisted of a double-blind, randomized clinical trial in which patients receiving treatment and providers registering the results of said treatment were unaware of which intervention the patient was receiving.
From February 2003 to September 2008, patients with acute (less than 48 hours' duration) LBP were recruited for the study on an ongoing basis in five outpatient orthopedic or general practices in Germany. Patients initially were randomized to one of three intervention arms: fixed-dose diclofenac therapy, spinal HVLA manipulation or placebo. The study authors noted the "relative risk of adverse effects of NSAIDs," choosing diclofenac to be acceptable for the study using a fixed dose (50 mg tablet, three times a day) and for a short time frame (less than 10 days).
However, an interim analysis in 2006 showed active treatment to be clinically superior to placebo, and the trial was continued with the two active arms only: HVLA manipulation vs. diclofenac. (Ten of 22 subjects in the initial placebo group dropped out of the study due to treatment failure / ongoing pain, as might be expected.) Sixty-nine patients participated in phase 1 and 32 during phase 2, with 93 patients ultimately proving to be evaluable and comprising the study group: 35 who received spinal manipulation, 36 who received diclofenac therapy, and 22 who underwent placebo treatment.
To ensure patient blinding, subjects in the manipulation group received placebo tablets (representing diclofenac), while patients in the diclofenac group received sham manipulation. To accomplish the latter, treating clinicians delivered HVLA manipulations "designed to treat the [nondysfunctional sacroiliac joint] by traction on the leg combined with a cephalad impulse on the sacrum, which then remains neutral regarding the lumbar spine. ... In addition, this technique [was] applied on the opposite side of the identified segmental dysfunction." Because of this, LBP patients randomized to receive sham manipulation (diclofenac arm) were excluded from the study if they presented with dysfunction of the sacroiliac joint as a characteristic of their back pain, which accounted for the relatively small number of total patients treated during the five-year study period.
Other exclusion criteria included known tolerance to NSAIDs or paracetamol (a "rescue medication" available to all patients during the study, not to exceed six 500-mg tablets per day); occurrence of low back pain and/or receipt of spinal manipulation for any cause within the previous three months; presence of any metabolic, malignant or serious organic or neurological disease; and any structural disturbances of the spine (osteoporosis, scoliosis, disc herniation, spondylolisthesis, hip dysplasia, etc.).
While the treating clinician obviously could not be blinded to the type of therapy being provided, blinding was ensured on a provider level by having a second physician, unaware of the randomization, register treatment results between baseline and following treatment. Specifically, "Before treatment and between 7 and 9 days ... after treatment, the subjects filled out questions and items in [their] personal diary and noted intake of rescue medication. Three days after entrance into the trial, patients were seen again by a physician who treated them initially to undergo another spinal manipulation (according to treatment allocation) if necessary. Immediately thereafter, and at another time between days 7 and 9 after randomization, the patient was seen by [the blinded investigator]. ... Subjects who experienced [persistent] or intolerable aggravation of LBP were free to visit, at any time, either the treating physician (up to day 3) or the blinded investigator (on days 4-7) to decide on continuation or termination of the trial to allow open therapy according to clinical standard."
The second clinician did not apply any treatment whatsoever during their interaction with patients. The primary outcome variable registered by the second clinician was the difference in Roland-Morris Disability Scale score before vs. following treatment. A 100 mm VAS for self-assessment of pain and SF-12 inventory (quality-of-life questionnaire) served as secondary outcome variables. All three items were components of each patient's personal diary, which they completed before and in the days following treatment as specified above.
- "There was a clear difference between the treatment groups: the subjects [receiving] spinal manipulation showed a faster and quantitatively more distinct reduction in the RMS" (compared to subjects receiving diclofenac therapy).
- "Subjects [also] noticed a faster and quantitatively more distinct reduction in [their] subjective estimation of pain after manipulation. ... A similar observation was made when comparing the somatic part of the SF-12 inventory ... indicating that the subjects experienced better quality of life after the spinal manipulation compared to diclofenac."
- "The rescue medication was calculated both for the mean cumulative dose (numbers of 500 mg paracetamol tablets) and for the number of days on which rescue medication was taken. ... In the diclofenac arm, the patients on average took almost 3 times as many tablets and the number of days [taking the tablets] was almost twice as high" compared to patients in the manipulation arm. While the authors note that these results were not significant due to large between-individual variations (meaning a few patients could have taken many tablets, throwing off the overall totals), it still suggests that value of spinal manipulation vs. drug therapy (because even if both patient groups had taken the same amount of rescue medication for the same number of days, it wouldn't discount the fact that patients in the manipulation group showed significant improvement on outcome variables compared to patients in the diclofenac group).
The HVLA Manipulation
According to the study authors, the high-velocity, low-amplitude manipulation performed on patients randomized to the manipulation intervention was "the most commonly used HVLA technique for this indication in many countries," performed as follows:
- "The physician palpates between the spinous processes of the dysfunctional segment and flexes the patient's upper leg at knee and hip until this segment opens in a neutral position of flexion." Patient is in side-lying position.
- "While getting into a deep contact with 2 fingers of the caudad hand to the table-faced side of the upper spinous process of the identified dysfunctional segment, the physician places the cephalad hand in the antecubital fossa of the patient's arm while resting the forearm gently on the patient's upper lateral thorax directly below the shoulder."
- "The patient's shoulder and pelvis ... are axially rotated in opposite directions. The patients inhales and exhales, and during exhalation, further rotational 'slack' is taken up as a diagnostic probation mobilization to exclude contraindications against an impulse."
- "With the patient relaxed and exhaling, the physician applies out of the rotational slack a HVLA thrust simultaneously moving with his forearms the pelvis and sacrum towards him and the shoulder girdle into the opposite direction, while pulling the upper spinous process of the dysfunctional segment upwards."
- von Heymann WJ, Schloemer P, Timm J, Muehlbauer B. Spinal high-velocity low-amplitude manipulation in acute nonspecific low back pain: a double-blinded, randomized controlled trial in comparison with diclofenac and placebo. Spine, April 2013;38(7):540-48.