The study included 39 patients (LDH) who presented at the Orthopaedics and Traumatology Department of Büyük Anadolu Hospital between November 2020 and February 2021 with the complaint of low back pain and were diagnosed with lumbar disc herniation. A control group was formed of 38 healthy subjects, selected at random from those invited to participate via social media and announcements, who had not experienced any low back pain in the last year and were not determined with any lumbar problem in physical or radiology examinations. The LDH group included patients aged <65 years, with low back pain ongoing for the last 3 months, who were diagnosed with lumbar hernia on MRI. For patients with suspected root compression, EMG was requested, and after confirmation of no root compression, these patients were included in the study.
Patients were excluded from the study if radiculopathy was determined on MRI or EMG, if they had any rheumatological or infectious disease, any spine or hip deformity, a history of lumbar surgery, or acute pain in any other part of the body.
Demographic data and disease-related information were recorded on a demographic information form in face-to-face interviews. When necessary for the diffierential diagnosis, hemogram, erythrocyte sedimentation rate, full urine analysis, ASO, CRP, RF, salmonella, and brucella tests were requested. The physical examinations of 77 patients were performed by the same orthopaedics and traumatology specialist, experienced in spinal surgery, and the MRIs were analyzed by the same radiology specialist (K.T.) experienced on this subject and blinded to the clinical history. All the MRIs were taken by the same radiology technician.
Approval for the study was granted by the Local Ethics Committee (YDU/2020/83-1160). Informed consent was obtained from all the study participants.
Magnetic Resonance Imaging
Images were obtained using a 1.5 Tesla MRI unit (Sigma Explorer SV25.3 with up-to-date software and 16 channels; General Electric, Milwaukee, WI, USA). With the patient in a supine position, a pillow was placed under the knees and the coil was placed on the spine. The measurements were taken with a routine protocol applied to the lumbar spine to pass through the centre of the disc at the measurement level between L3-S1 (L3-4/L4-5/L5-S1) without leaning to the right or left. Turbo-spin echo T1 and T2-weighted sagittal slices and turbo-spin echo T2 axial slices parallel to the disc spaces were obtained at a thickness of 4mm. Evaluations were made on the T2 axial slices. Fat content was evaluated at all 3 levels of L3-S1 of the left and right erector spina (m. iliocostalis+m. longissimus) and the m. multifidus.
Fat infiltration of the muscles was evaluated semi-quantitatively, using the simplified 3-grade system and the 4-grade system. In the simplified 3-grade system, fat infiltration is graded as Grade 1: normal (fat infiltration of up to 10% of the muscle cross-sectional area [CSA), Grade 2: moderate (10%-50%) and Grade 3: severe (>50%) (Figure 1) . In the simplified 4-grade system, fat infiltration is graded as Grade 1: normal (fat infiltration of up to 10% of the muscle CSA), Grade 2: mild (10%-30%), Grade 3: moderate (30%-50%), and Grade 4: severe (>50%) (Figure 2).
For internal reliability, 10 randomly selected patients were evaluated again by the same radiologist after an interval of 1 month. Intra-observer agreement was examined with Kappa and the Kappa value obtained was 0.934 for the 3-grade system, and 0.921 for the 4-grade system.
Data obtained in the study were analyzed statistically using IBM SPSS vn. 23 software. The risk factors affecting lumbar disc herniation were examined with binary logistic regression analysis, and the results were presented as odds ratio (OR) and 95% confidence interval (CI). In the univariate analysis, each variable was added separately to the model and the effect of each variable alone was examined. In the multivariate analysis, all the variables were placed in the model at the same time and the effect of these together was examined. According to the post hoc power analysis, a minimum of 50 subjects should be included in the study to provide 95% CI (1-α), 95% test power (1-β), and degeneration values of 0.32 and 0.80. Taking the potential loss of patients into consideration, the study was completed with a total of 77 subjects, with test power of 99.5% according to the post hoc power analysis .