This is a retrospective cohort study of patients who underwent a 6-week treatment protocol of non-surgical spinal decompression via the DRX9000. A HIPAA (Health Insurance Portability and Accountability Act) waiver was obtained through Quorum IRB. This waiver permitted a review of medical records and access to CT scans ordered as part of standard of care.
Clinical Trial Registration Number: NCT00828880
Inclusion and exclusion criteria
Patients and their medical records were eligible for inclusion if the patient was at least 18 years of age, consented for the 6-week treatment protocol, and presented with chronic LBP of at least 3 out of 10 on a verbal rating scale and was due to either discogenic LBP or disc herniation according to a radiological diagnosis using standard medical definitions. Discogenic LBP is most succinctly defined as a loss of lower back function with pain due to disc degeneration. Degenerative disc diseases often emerge when abnormal stresses cause the nucleus gelatinosus to unevenly distribute weight, the annular fibrosis and end plate incur structural damage, and a destructive inflammatory response is triggered to accelerate and perpetuate the degeneration of the disc. A herniated disc (synonymous with a protruding or bulging disc) arises when the intervertebral disc degenerates and is weakened to such an extent that cartilage is pushed into the space containing the spinal cord or a nerve root and causes pain.
All patients were treated at the Upper Valley Interventional Radiology facility (McAllen, Texas). Patient symptoms were evaluated by medical history review, physical examination, and a current CT scan (not older than 2 months prior to the start of treatment) to support a diagnosis of chronic discogenic LBP due to bulging, protruding or herniated intervertebral discs that may have been brought on by degenerative disc disease. Patients were only included if pre- and post-treatment CT scans were performed on the same device, measurements taken by the same investigator (WM), and data recorded on standard collection forms. One height measurement was taken by WM for each of the intervertebral discs under study per CT scan. Accuracy of data was confirmed by a second investigator (JP), but only one measurement was made of each intervertebral disc per CT scan. All CT scans analyzed were performed at least one hour after the subject got out of bed. The first CT scan was performed within two months before the initiation of the treatment, and the second CT scan at least one day after or on the day immediately before the final treatment session.
Exclusion criteria for enrollment in the study were any patients with metastatic cancer; previous spinal fusion or placement of stabilization hardware, instrumentation or artificial discs; neurologic motor deficits; bladder or sexual dysfunction; alcohol or drug abuse; or litigation for a health-related claim (in process or pending for workers' compensation or personal injury). Limitations of the spinal decompression system also led to the exclusion of patients with extremes of height (< 147 cm or > 203 cm) and body weight (> 136 kg).
Patients received treatment with the DRX9000 (Axiom Worldwide, Tampa, FL) as dictated by the intervention's operating guidelines. In short, the protocol typically included 22 sessions of spinal decompression over a 6-week period with 28-minute active treatment sessions. At the start of each session, the patient is fitted with adjustable lower and upper body harnesses and is lowered into the supine position. To initiate active treatment the machine then pulls the patient gently on the lower harness while the upper harness remains stationary, thus distracting the patient's spine. A safety button can be pushed at any time by the patient to release all tension immediately. Daily treatments, Monday through Friday, were performed for the first two weeks of treatment. The latter four weeks consisted of treatments every other day, Monday, Wednesday and Friday.
Initial decompression force was adjusted to patient tolerance, starting at 4.54 kg (10 lbs) less than half their body weight. If a patient described the decompression pull as "strong or painful," this distraction force was decreased by 10%-25%. In subsequent treatment sessions, the distraction force was increased as tolerated to final levels of 4.54 kg to 9.07 kg (10 to 20 lbs) more than half their body weight. Patients continued to use analgesics prescribed by their physicians before enrollment, but were allowed to use additional non-steroidal pain medication should their pain increase temporarily and permitted to discontinue pain medication as needed. During the routine physical examination performed by WM prior to beginning the non-surgical spinal decompression treatment session, at the first and final visits maximal pain was evaluated during a flexion-extension range of motion exam with the question "How strong is your pain on a scale of 0-10 with 0 being no pain and 10 as bad as it could be?"
The first main outcome for this study was the change in pain during a range of motion evaluation measured on an 11-point verbal rating scale (VRS), with 0 being no pain and 10 being pain as excruciating as could be imagined, before and after the 6-week spinal decompression treatment regimen.
The second main outcome was the change in average disc height as measured by CT scan. For each patient, average disc height of L3-L4, L4-L5 and L5-S1 was calculated before the first treatment session and at least one day after or on the day before the last treatment session.
Statistical analysis and sample size estimation
We assumed data to be normally distributed unless exploratory analyses suggested otherwise, in which case a Kolmogorov-Smirnov test was to be applied. Since the treatment effect was defined as the difference between before and after the therapeutic intervention, a paired t-test was applied to test whether there was a reduction in pain and an increase in disc height. For the main hypothesis, the correlation between disc height changes and low back pain, we applied linear regression to quantify the relationship with Pearson's correlation coefficient to determine statistical significance.
Sample size estimations were performed to have sufficient power to test with a two sided type I error of 0.05 and type II error of 0.2 (80% power). Given the sizeable treatment effect reported in the retrospective chart review and also in the prospective pilot study mentioned in the introduction, we expected a reduction in range of motion pain from 6 to 2, with a standard deviation of 2.5. This resulted in a sample size estimation of only 5 patients. To test changes in disc height, we expected a standard disc height of about 8 mm with diseased discs being slightly more compressed, i.e. at about 7.5 mm, and anticipated discs after the decompression treatment to measure at about 8.25 mm. Assuming a standard deviation of 1.0 mm, we estimated a required sample size of 16 patients in order to show a difference. The sample size for the main hypothesis, that the degree of pain reduction is associated with the amount of increase in disc height, was more difficult to estimate since no previous study had determined a correlation coefficient. Therefore, we chose a coefficient of 0.5 for a conservative expectation, resulting in a required sample size of 26 patients. Taking into consideration the possibility of drop-outs, we aimed to collect data from 30 patients.