This study was approved by the institutional review board of Asan Medical Center (protocol number: 2023-0035). The requirement for informed consent was waived because of the retrospective nature of the study. A retrospective review was conducted on the medical records of patients who underwent CIESIs at a pain clinic (Asan Medical Center, Seoul, Korea) between March 2021 and June 2022. The inclusion criteria were as follows: (1) cervical radicular pain, (2) presence of cervical spinal stenosis or cervical disc degeneration on magnetic resonance imaging (MRI), (3) failure to respond to conservative treatment for > 3 months, (4) MRI performed within 1 year preceding the injection. The exclusion criteria were as follows: (1) insufficient medical records with missing numerical rating scale (NRS) scores; (2) patients lost to follow-up within three months after CIESI; (3) axial neck pain without cervical radicular pain; (4) cervical myelopathy; (5) no or insufficient radiologic tests; (6) prior cervical spinal surgery; (7) cervical epidural neuroplasty performed instead of a single injection; and (8) history of CIESI within the last 3 months.

CIESIs were performed as previously reported [9]. All procedures were performed under fluoroscopic guidance. The needle insertion level was determined at C6–C7 or C7–T1, and the paramedian approach was adopted in all procedures. After sterilization, the skin was infiltrated with 1% lidocaine, and a 22G Tuohy needle (Green Medical Supply) was inserted at the target cervical level while guided by an anteroposterior (AP) view. When the needle was advanced until the pedicle level on the vertebral body just inferior to the target interlaminar space on an AP image, fluoroscopy was rotated obliquely in the contralateral side at an angle of 50° ± 5° to visualize the ventral interlaminar line (VILL) between the laminae. Then, the needle was advanced further without loss of resistance (LOR) immediately before the VILL in the contralateral oblique (CLO) view. Subsequently, the needle was cautiously advanced using an LOR-to-air technique to reach the epidural space. When LOR was obtained, 0.5 mL of contrast medium (Omnipaque 300; GE Healthcare) was injected to confirm the epidural space using the CLO and AP views. If the epidural space was confirmed, a 3-mL mixture of 5 mg dexamethasone and 1% lidocaine was administered.

All CIESIs were performed by three attending pain physicians (≥ 4 years of procedural experience), nine fellows (0 to 8 months of experience), and 16 residents (0 to 3 months of experience) according to identical protocols. The trainees performed all the procedures under the supervision of the attending pain physicians.

The pain intensity was evaluated at baseline and 3 months after CIESI using NRSs which ranged from 0 (no pain) to 10 (worst imaginable pain). A responder was defined as a patient with an NRS score decrease of ≥ 50% [10]. The patients were divided into non-responder and responder groups depending on the outcome.

Demographic data, including age, sex, and body mass index (BMI), were collected. Comorbidities (such as diabetes and hypertension), pain duration, last previous procedure, response to last previous procedure, and pre-procedural symptoms (radicular pain with or without neck pain, weakness) were also collected. The cervical curve type (lordosis, straight, or sigmoidal or kyphosis) was assessed using pre-procedural lateral cervical spine X-rays [11].

Based on the pre-procedural MRI, each severity of central stenosis, foraminal stenosis, and disc degeneration at the closest level in correlation with the patient’s symptoms were assessed. For central stenosis, each grade was defined as follows: grade 0, normal; grade 1, obliteration of > 50% of the subarachnoid space without any signs of cord deformity; grade 2, central canal stenosis with spinal cord deformity, or cord deformation with no signal changes in the spinal cord; and grade 3, increased signal intensity of the spinal cord near the compressed level on T2-weighted images [12]. For foraminal stenosis, each grade was defined as follows: grade 0, normal or absence of neural foraminal stenosis with the narrowest neural foraminal width longer than that of the extraforaminal nerve root; grade 1, non-severe cervical neural foraminal stenosis, with the narrowest neural foraminal width the same as or less than (but > 50% of) the extraforaminal nerve root width; and grade 2, severe cervical neural foraminal stenosis, with the narrowest neural foraminal width the same as or less than 50% of the extraforaminal nerve root width [13]. For disc degeneration, each grade was defined as follows: grade 1, low-intensity or structural changes of the nucleus pulposus; grade 2, disc bulge or herniation with annulus fibrosus degeneration; and grade 3, further degeneration with a disc height decrease > 25% [14]. The severity of central stenosis, foraminal stenosis, and disc degeneration in each case was determined by consensus among the three investigators. Additionally, cervical sarcopenia was also assessed using a grading system on fatty infiltration. Fatty infiltration in the bilateral multifidus muscles were evaluated at the C5–C6 level using the Goutallier classification [15,16]. The C5–C6 level is the most commonly involved cervical spine level with degenerative changes, due to its relatively high load-bearing role and flexibility [17,18]. Thus, this level has been frequently selected as a representative for assessing fatty infiltration in the cervical paraspinal muscles. Axial cuts of T2-weighted MRI sequences were utilized by two blinded independent reviewers (Fig. 1). The Goutallier classification is defined as follows: Goutallier 0, no visible fat streaks; Goutallier 1, minimal fatty streaks; Goutallier 2, more muscle present than fat; Goutallier 3, fat and muscle present in equal quantities; Goutallier 4, more fat present than muscle [19]. After the inter-rater reliability between the reviewers was tested, the grades of the two reviewers were averaged and further classified as follows: mild (Goutallier 0–1), moderate (Goutallier 1.5–2), or severe (Goutallier 2.5–4.0) sarcopenia, based on the Fuchs modification of the Goutallier grading system: no fat to minimal fat as mild, more muscles than fat as moderate, and equal/greater fat than muscle as severe [15,16]. All the reviewers were pain physicians who have more than 5 years of experience in diagnosing spine pathologies with spine MRI. Before the study, the reviewers were trained to evaluate fatty infiltration using the Goutallier classification with 30 cases each.

The pre- and post-procedural pain intensities were classified into one of three categories (mild, NRS 1‒3; moderate, NRS 4‒6; and severe, NRS 7‒10).

Categorical variables were presented as absolute numbers with percentages. Continuous variables were compared between non-responders and responders using the Student’s t-test or Mann–Whitney U-test, and were presented as medians with interquartile ranges. Categorical variables were compared using the chi-squared test or Fisher’s exact test. To determine the factors associated with a successful response to CIESI, univariate and multivariate logistic regression analyses were performed. Variables with P values < 0.1 on univariate logistic regression analysis were included in the multivariate logistic regression analyses. P values < 0.05 were considered significant. The odds ratios (ORs) and 95% confidence intervals (CIs) for successful response in the presence of independent predictors of CIESI were calculated by logistic regression analyses. All statistical analyses were performed using IBM SPSS (version 22; IBM Corp.).