Comparison of Prevalence of Degenerative Findings on Lumbar MRI among Sciatica Patients classied using the McKenzie Method.

Purpose To determine if the prevalence of degenerative ndings on MRI differ between sciatica patients with radiating pain below the knee assigned into two groups by using the McKenzie method. Methods A comparative study of one hundred sciatica patients referred to the spine clinic at the hospital for specialist consultation to explore the need for surgery. Patients were classied into centralizers (CEN-group), in which leg pain is relieved in response to repeated specic exercises, and non-centralizers (Non-CEN-group), who have not responded. The latter have been shown to have surgery more often than patients in the CEN-group, hypothesized to have more progressed degenerative ndings on MRI. Multiple lumbar MRI characteristics were scored and compared between the groups. There was a statistically signicant difference in the degenerative ndings between centralizers and non-centralizers, respectively: vertebral end-plate changes were 63% and 43%; mean (SD) Prrmann’s disc degeneration score was 12.8 (±3.5) and 10.6 (±3.8); and severity score of total damage was 12.0 (±3.5) and 10.1 (±3.8). The prevalence of disc prolapses were 76% in the CEN-group and 59% in the Non-CEN-group. Further, 75% in the CEN-group and 74% in the Non-CEN group had nerve root compression due to disc herniation with no statistically discernible difference between the groups. Disc contour on MRI were between both groups as as of nerve root compressions, but more disc changes on MRI were found in centralizers than in non-centralizers. Thus, neither nor overall degenerative ndings in sciatica can be used to determine types of radicular


Introduction
Magnetic resonance imaging (MRI) is considered the mainstay investigation among patients who are candidates for disc surgery suspected of lumbar disc herniation (LDH), and the symptoms have not resolved within 6 weeks as expected with conservative treatment [1.2]. Both imaging and clinical signs and symptoms determine the nal decision for surgery in sciatica patients with radiculopathy.
LDH accounts for less than three percent of all low-back problems but is the most frequent cause of sciatic or radicular pain [3]. LDH may take the form of bulged, protruded, prolapsed, or extruded disc material [4]. Low back and radicular pain may also be produced by structural bony changes such as stenosis, degenerative disc- [5] and 7], however these changes can also be found in the asymptomatic population [8].
Radicular pain is sharp pain that travels along a narrow band and arises as a result of irritation of a spinal nerve or its roots and can be associated with other signs of radiculopathy [9]. In radiculopathy, conduction is blocked in the axons of a spinal nerve or its roots by compression or in ammation [9]; in sensory axons this results in numbness, and in motor axons in weakness [9].
The McKenzie method of mechanical diagnosis and therapy (MDT) has been shown to be a reliable, and validated assessment, classi cation, and therapy method [10.11]. The method has been scored highest of any musculoskeletal therapy classi cation system for low-back related leg pain [12].
The MDT method classi es patients with radicular symptoms with repeated movements and sustained postures into subgroups to determine appropriate management strategies and prognosis into centralizers (CEN-group), and non-centralizers (Non-CEN-group) [10]. In the CEN-group, radiating pain is lastingly abolished with speci c exercises, which has been shown to predict good non-surgical treatment outcomes [13]. Whereas, in the Non-CENgroup patients radiating pain remains unchanged or worsens when assessed with mechanical loading, and   patients in this group have been shown to have signi cantly worse pain, disability and psychosocial outcomes compared to the CEN-group in the short and long-term [13,14]. Furthermore, patients in non-CEN-group are six times more likely to undergo disc surgery compared to patients in the CEN-group [13], and thus might be hypothesized to have more progressed degenerative ndings and disc changes on MRI.
The purpose of this study was to determine in patients with sciatica if degenerative ndings on MRI differ between non-CEN and CEN-groups according to the McKenzie method. We hypothesized that non-centralizers have signi cantly more progressed degenerative ndings and disc contour changes compressing a nerve root on MRI than centralizers.

Trial procedure
Patients diagnosed with sciatica (N= 132) in primary or occupational health care were referred for further investigations to the spine clinic of a tertiary hospital. Inclusion criteria were age 18 -to-65-years, and lumbar radicular pain that had lasted at least six weeks. Exclusion criteria were pregnancy, previous low back surgery, serious diseases or "red-ags" including cauda equina syndrome (CES) or lower limb palsy that hindered normal functioning, previous high or moderate energy trauma, or osteoporotic fractures for the elderly.
If the medical examination at the spine clinic suggested that the patient may have radiculopathy and need for surgery, they were referred for a lumbar MRI. While waiting for their MRI scan, patients underwent a clinical McKenzie-based assessment by one of two physiotherapists specially trained to perform the examination, and both with several years of clinical experience. Finally, one hundred patients (76%) met the criteria, and were classi ed into the two groups. After an oral and written explanation, patients signed informed consent for their participation in the study. All protocols adhered to relevant ethical guidelines and regulations. This study was the anatomical location of dominant pain; and neural examination of nerve tensions, key muscle strengths and light touch sensitivity tests. This was followed by a standardized loading strategy testing with single and repeated end-range movements and / or sustained end-range positions in exion, extension, lateral side-gliding /bending and rotation in loaded and in unloaded positions. The full clinical assessment lasted 90 minutes.
Before repeated end-range movement testing patients stood with their spine erect and recorded on a pain drawing the exact location of the present pain. After the repeated end-range movement testing, patients completed a second pain drawing blinded to the rst drawing. The assessor compared drawings, and the exact site and any change in the location or intensity of low back and radicular pain was recorded.
The patients were assigned into the CEN-group (n=51), if pain was decreased or abolished distally, but remained more centrally, and remained so after getting up from the treatment couch while walking around for at least three minutes. The patients were assigned into the Non-CEN-group (n=49), if there was no change in location or intensity of the leg pain or symptoms only peripheralized more distally, or the most distal pain increased in intensity.

MRI ndings
Images were rst assessed by a resident orthopedic surgeon, who was speci cally trained to classify degenerative changes on MRI and was familiar with reading MR images of the spine (JR). Multiple MRI characteristics of the degenerated ndings were scored from spinal levels L1 to S1, using both increased signal on T2-weighted/ uidsensitive sequences and decreased signal on T1-weighted images. Classi cation of lumbar disc was based on lumbar disc nomenclature: version 2.0 (4) with bulging, protrusion or prolapse of the lumbar disc. The location of nerve root compressions due to prolapsed discs was assessed by a physiatrist (JY) with extensive experience in clinical practice and reading MR images.
Disc degeneration was classi ed according to P rrmann's grading system [5]. Classi cation is scored from 1 point referring to no degenerative changes to 5 points referring to the disc as inhomogeneous with a hypointense black signal intensity, no differentiation between the nucleus and annulus, and the disc space is collapsed. Degenerative spondylosis was classi ed using the Kellgren classi cation [17]. The Kellgren classi cation is scored from 0 points referring to no degenerative changes to 4 points for severe narrowing of the disc with sclerosis and large osteophytes [17]. The presence of Modic changes [5,18] was assessed and classi ed in three subgroups: Modic type 1 changes, result from bone marrow edema and in ammation, Modic type 2 changes associated with fatty degeneration of the bone marrow and Modic type 3 changes with subchondral bone sclerosis, representing different stages of the same pathological process [19]. Furthermore, degeneration of end plate, a cartilaginous layer between the intervertebral disc and the vertebra, was assessed using the total end-plate score (TEPS) [7]. The TEPS was assigned to each disc separately using a 6-point-scale from 1 (no changes) to 6 (complete end plate damage, irregularity, and sclerosis of the end plate). In spinal stenosis, a portion of the spinal canal narrows to the point at which it exerts compression on the spinal nerves. Spinal stenosis was considered present when the thecal sac measured less than 100mm 2 in surface area or in case of obliteration of perineural fat and compression of lateral recess or foramen. Nerve root canal stenosis was considered present if there was narrowing of the nerve root canal due to osteoarthritis. Spondylolysis was assessed from the MRIs and was considered as a break in the pars interarticularis. Spondylolisthesis was assessed by measuring the anterior and posterior walls of the vertebrae in relation of adjacent vertebrae.

Sociodemographic characteristics and clinical outcomes
Data were collected on sociodemographic characteristics: age, gender, body mass index and smoking. Duration of the spinal disorders were measured in years, and duration of the present symptoms in weeks. Low back and leg pain intensities during last week were assessed with a visual analogue scale (VAS) [15]. Disability was captured using the Finnish version of the Oswestry Disability Index (ODI) [16].

Data analysis
The descriptive statistics are presented as means with standard deviation (SD), as medians with interquartile range (IQR) or counts with percentages. Statistical comparisons between the groups were made using the t-test,

Results
Demographic data In the present study the duration of radicular pain median was 16 weeks (IQR 12-24) in the CEN-group, and 20 weeks (IQR 12-28) in the Non-CEN-group. Current low back pain intensity was signi cantly stronger in patients in the Non-CEN-group and the patients were signi cantly more disabled than patients in the CEN-group, p=0.008,and p<0.001 respectively (Table 1).
Thirty-eight patients (75%) in the CEN-group and thirty-six (74%) in the Non-CEN-group had nerve root compression due to intervertebral disc herniation (p=0.44, Table 2). Among other degenerative ndings, 63% in the CEN-group and 43% in the Non-CEN-group had Modic changes (p=0.05). Modic type 1 changes, resulting from bone marrow edema and in ammation was found among 10 patients in each of the two groups. Modic type 2 changes, entailing fatty degeneration of the bone marrow, and Modic type 3 changes with subchondral bone sclerosis was found in 22 and 11 patients in the CEN-and the Non-CEN-groups, respectively. Mean P rrmann's score was 12.8 (SD ±3.5) in the CEN-group and 10.6 (±3.8) in the Non-CEN-group (p=0.003), and mean TEP score was 12.0 (±3.5) and 10.1 (±3.8), respectively (p=0.01), which were most signi cant at L4-L5 level. More detailed ndings are in Table 3. Spondylosis was a more common nding in the CEN-group than in the Non-CEN-group (0.03). There were no other statistical differences between the groups with regard to spondylolysis, spondylolisthesis, and prevalence of central, foraminal, and lateral recess stenosis (Table 2).   Table 3 Comparison of disc degenerative changes on MRI between sciatica patients classi ed by the McKenzie method in the CEN-and the Non-CEN-groups (N=100).

Discussion
In this present study the MRI discogenic ndings were high and the amount of disc contour changes were similar between the groups, as were nerve root compressions, central and foraminal stenosis, lateral recess and spondylolisthesis ndings despite different clinical ndings. Patients in the CEN-group had signi cantly higher disc degeneration scores (P rrmann's), more Modic changes and endplate damages, and more spondylosis on MRI than those in Non-CEN-group.
Diagnostic imaging and surgery may be indicated in patients with severe lumbar radicular symptoms who fail to respond to conservative care in 6-8 weeks [4]. In this present study, the mean duration of radicular pain was over three months before they were referred to tertiary hospital for further medical investigations. These patients had statistically similar sociodemographic data and neurological signs and symptoms, with two exceptions: the non-centralizers had signi cantly stronger back pain and they were signi cantly more disabled than centralizers. There was no statistical difference in lower leg pain.
It has been shown that a single imaging nding such as Modic change or disc degeneration has only weak association with the presence of pain [20]. Modic changes at L5-S1 and Modic type I lesions are more likely to be associated with lower back pain and sciatica pain symptoms than other types of Modic changes or changes located at other lumbar levels [21]. In our analysis, the signi cant amounts of herniated discs, lateral recess ndings, nerve root compressions, degeneration changes, Modic changes and total endplate damages were found mostly at L4-5 levels and secondly at L5-S1 levels. These different types of disc contours changes, other types of degenerative image ndings and Modic changes on MRI are also highly prevalent in asymptomatic populations and increase with advanced age [8].
Albert et al. [22] found that among their sciatica patients the amounts of centralizers was as high as 85%. They concluded that no matter what type of disc lesion is found on MRI and despite the severity of leg pain and neurological ndings; it was possible for the majority of this group of patients to demonstrate centralization [22]. In our analysis the CEN-group prevalence was much lower. The reason for which might be the difference in de nition of centralization. In contrast to Albert et al. we did not include those patients in the CEN-group, if the centralization was unstable i.e., the pain was reduced or abolished during the repeated movements testing or positioning, but after resuming a weight bearing position for at least three minutes, the site and pain intensity level returned to the pre-testing location and intensity. In our study these patients were assigned into the Non-CENgroup.
To the best of our knowledge, our study may be the rst to investigate the difference of the features of spinal discs and other degenerative changes on MRI between sciatica patients classi ed by repeated movements and sustained postures into centralizers and non-centralizers. The MDT system has acceptable inter-examiner reliability for classifying patients with back pain into subgroups when applied by therapists trained into the method [23]. In this study the physiotherapists who classi ed the patients were certi ed and experienced in assessing and classifying sciatica patients into different mechanical subgroups. However, the assessment and classi cation were based only on the rst clinical examination at study entry, thus this may have affected the distribution of patients. Werneke et al. [24] found, that it may take up to seven therapy visits to de ne the exact group among chronic patients such as ours.
Another limitation might have been that the MRI morphology was not subjected to inter-examiner reliability assessment. Lurie et al. [25] concluded that classi cation of disc morphology showed substantial intra-and interreader agreement, whereas thecal sac and nerve root compression showed moderate reliability. In another study excellent agreement was found on the affected disc levels and nerve root that most likely caused the sciatica symptoms [1]. In this present study doctors who evaluated MR images were experienced in reading MR images and were blinded to the results of the clinical examination and classi cations of patients.
In conclusion, this study provides a relationship of sciatica patients' pain behavior classi ed by repeated movements and sustained postures with structural MRI ndings. More progressed degenerative changes on MRI were found in the centralizers than in the non-centralizers. Disc contour changes on MRI were similar between patients in both groups as well as prevalence of disc herniation compressing a nerve root. Thus, neither disc nor overall degenerative ndings in sciatica patients can be used to determine types of radicular pain.