Spine MRI Patterns And Predictors of Disability In Patients With Lower Back Pain. A Prospective Cross-Sectional Study at University of Gondar Comprehensive Specialized Hospital, North West Ethiopia. 2020.

Background: Low back pain (LBP) has become one of the most serious public health problems with substantial socioeconomic implication. Degenerative disc disease one of the commonest cause of LBP. Magnetic resonance imaging (MRI) is routinely utilized in evaluation patients with degenerative changes of the lumbar spine. However there are contradictors reports with regards to association of MRI ndings of lumbar spine and patients’ symptoms. Objective: This study is aimed determine correlation of degree of disability with lumbar spine MRI ndings in patients with LBP at University of Gondar comprehensive Specialized Hospital (UoGCSH), North West Ethiopia, 2020. Methods and Materials. A prospective cross-sectional study was conducted on 72 consecutively enrolled patients with lower back pain who underwent lumbar MRI scan. Degree of disability was measured using Oswestry disability Index (ODI) questionnaire translated to local language. Association between lumbar spine MRI parameters and ODI score and category was tested using Spearman’s rank correlation coecient and Chi square tests. Results The mean age the study subjects was 43.81±1.88 years (range 22-83 years). 59.7% of the study population were Female. In terms of ODI category, most fell under minimal and moderate disability 33 (45.8%) and 25(34.7%) respectively. Disc bulge (81.9%) and foraminal stenosis were the most frequent MRI abnormalities detected. ODI score showed weak correlation with grade of spinal canal stenosis. Foraminal stenosis grade was not correlated ODI. Conclusion: The clinical relevance of MRI ndings in patients with degenerative disc disease is limited and MRI should be sparingly ordered in evaluation of these patients particularly in resource constrained settings.


Introduction
Lower back pain is a major public health problem globally with life time prevalence reaching 11-84% (1).
According to the Global Burden of Disease (GBD) report, the population prevalence of lower back pain for all age groups and gender was estimated to be more than half a billion people per year in 2015 and that gure was 17.3% larger compared to the prevalence in 2005. According to the same report lower back and neck pain were the leading global cause of disability in most countries in 2015. (2) Numerous disease processes could cause lower back pain and disability of which disc disease is the commonest.(3) Disc degeneration can be de ned as an aberrant, cell mediated response to progressive structural failure or simply as a degenerate disc that is painful (4). According to the lumbar disc nomenclature version 2.0 intervertebral disc disease is broadly classi ed in to degeneration and herniation. Degeneration include disc desiccation, disc space narrowing, disc bulge, mucinous degeneration, intradiscal gas and associated bone and marrow changes like endplate sclerosis and Modic changes. Herniation is de ned as focal displacement of disc material involving less than 25% of the disc dimension on axial plane. A more diffuse disc material displacement is referred as disc bulge (5).
Imaging plays an important role in the diagnosis, pre-surgical evaluation and follow up of patients with low back pain. Plain radiography, myelography, computed tomography (CT) and MRI Though there have been previous studies on patterns and prevalence of MRI nding from Africa, their clinical signi cance in terms of level of pain and disability hasn't been explored su ciently. In this study we purposed to describe lumbar MRI patterns and their association with degree of disability measured using ODI index in a third world setting.

Objectives
The objective of this research is to determine Lumbar spine MRI patterns and their association with Oswestry disability index in patients with lower back pain who presented to Radiology department of University of Gondar Comprehensive Specialized Hospital, northwest Ethiopia, 2020

Methods And Materials
Study Area University of Gondar comprehensive specialized Hospital is located in Gondar town, North West part of Ethiopia, 738 km from Addis Ababa. Gondar town is the capital of Central Gondar zone of Amhara Region. The Hospital is a major tertiary teaching Hospital giving service to over ve million population across the region.
The Department of Radiology provides diagnostic and basic interventional services. The department is equipped with two multi-detector CT scans (64 and 4 slice CT scans), one 1.5 tesla MRI, multiple multipurpose ultrasound machines and 2 Digital radiography machines.

Study design
Hospital based prospective cross sectional study design was employed for this study.
Study Period.
The study was conducted between June and September 2020.

Source and study population
Study population All patients with lower back pain referred to the radiology department of UoG CSH for a lumbar spine MRI scan.

Inclusion and exclusion criteria
Inclusion criteria All patients above 20 years of age who had history of lower back pain who were referred for lumbar spine MRI and consented to be part of the study were included.

Exclusion criteria
Patients with Previous lumbar spine operation, patients with non-degenerative cause of lower back pain and patients with contraindication for MRI or unable to complete lumbar spine exam were excluded.
Variables of the study  Modic changes which refers to vertebral end plate signal changes as a result of degeneration was graded as: Modic type 1 low signal intensity on T1WI and high on T2WI, representing brovascular tissue, in ammatory changes and edema.
Modic type 2 high signal intensity on T1WI and iso-intense/high on T2WI, representing bone marrow replacement by fat.
Modic type 3 low signal intensity on both T1WI and T2WI, representing reactive sclerosis.
Disc desiccation: is classi ed as:-Focal disc desiccation involvement of two or less than two discs.
Multifocal disc desiccation involvement of greater than two discs.
Sample size and sampling procedure All eligible patients who presented to the Radiology department with in the study period were enrolled consecutively. A total of 72 patients who ful lled the inclusion criteria were nally included in the study.

Data collection procedure
All the lumbar spine MRI scans were taken using 1.5Tesla Philips Achieva MRI machine by experienced radiographic technicians. Axial and sagittal T1 & T2 weighted and sagittal STIR sequences were taken for the vertebral segments L1 to S1. No contrast was used for the MRI studies. Scanned spine MRI studies were interpreted by a senior general radiologist.
Data on degree of disability was obtained by interviewing patients using an ODI questionnaire translated to local language. A semi-structured questionnaire was used to document all the other relevant socio demographic and imaging variables.

Data processing and analysis
Data was checked for completeness and cleaned before analysis. No missing data was identi ed in the dataset. Descriptive statistics is presented in the form of frequencies and percentages for categorical variables and summary statistics for continuous variables. The dependent variable (ODI) was tested for normality using the Shapiro-Wilk Test and showed a skewed distribution. Spearman's non-parametric correlation was applied to determine the relationship between MRI parameters and OSI score. Sciences, University of Gondar. Written informed consent was obtained from the study participants before the study. Participants were informed about the purpose of the research and that they have full right to refuse, withdraw or completely reject part or all of their participation in the study. Participants were also assured that their treatment and relation with the hospital and/or other organizations were not be in uenced by their withdrawal from the study. Con dentiality were ensured using anonymous checklist and questionnaire. The study followed the Strengthening of Reporting of Observational Studies in Epidemiology (STROBE) guideline.

Sociodemographic & Clinical Characteristics
Out of the 72 participants 43 (59.7%) were female. Mean age of patients in this study was 43.8 ± 1.9 years (range 22-83years). The median duration of lower back pain among participants was 2years (IQR 0.5-4years). Figure 1 MRI characteristics Fifty ve percent at L5/S1 and L3/L4 involvement was seen in twenty 29.2%.
Modic change was present in 14 (19.4%) of the studies. Type II Modic change was the predominant type accounting for 86% (12 out of the fourteen). L3-L4 (11.1%) segment was the most frequently involved followed by L2/3 (4/14) & L4/5 (4/14). Summary of the MRI patterns and ODI categories is presented in Table 1 Table  Discussion Disc bulge and foraminal stenosis were the two most prevalent MRI ndings among our study population, followed by disc desiccation, disc herniation and spinal canal stenosis. Most of the patients with foraminal or spinal canal stenosis had mild degree of stenosis and those pathologies tended to concentrate at L4/5 and L5/S1 levels. The prevalence of disc bulge (82%) and foraminal stenosis (65.3%) in the current study is relatively higher than previously reported. Brinjikji et al in their meta-analysis estimated the prevalence of disc bulge to be 43.2%. Whereas a hospital based study from Ethiopia puts the prevalence much lower at 18.5%. On the other hand the distribution of spinal segment involvement in our study was consistent with ndings from previous researches (11,12).
Our study revealed that MRI grades of spinal canal stenosis is weakly correlated with ODI score (r = 0.3).
(a spearman rank correlation between 0 and 0.4 is considered a weak correlation in most grading systems. (21)) Similar nding is reported in a study conducted in 1990 by Hurri et al. who reported association between spinal stenosis and ODI index (22). However plain radiography instead of MRI was used to estimate spinal canal stenosis in that study and their nding couldn't be substantiated by subsequent studies which tried to explore association between spinal canal area or anteroposterior diameter with that of ODI score. (13,14,23). In addition, Categorical ODI groups. Which are more clinically meaningful than absolute scores, failed to correlate with grade of canal stenosis in the current study.
Further study is required to conclusively determine association between canal stenosis and degree of disability.
Grade of foraminal stenosis, disc desiccation, disc bulge or herniation, Modic change and disc desiccation didn't predict disability in the current study. This is in agreement with results from previous studies. ( Our study is not without limitations. Small sample size and relatively few cases in the category of severe canal stenosis could have limited the power of the study. The qualitative grading system used for staging spinal canal and foraminal stenosis could have introduced bias and misclassi cation. We recommend that our study is interpreted in light of these limitations.

Conclusion
This study tried to investigate the pattern of MRI abnormalities in adult patients with lower back pain and determine their association with Oswestry disability score (ODI). Disc bulge and foraminal stenosis are the two commonest MRI patterns and their prevalence was relatively higher in the current study than previously reported. Spinal canal stenosis grade is weakly correlated with ODI score. Foraminal stenosis and other discal MRI parameters are not correlated with degree of disability.

Consent for publication
The authors have consented for the publication of this study.
Availability of data and material