Article Text

Download PDFPDF

One-year results from a randomized comparative trial of targeted steroid injection via epidural catheter versus standard transforaminal epidural injection for the treatment of unilateral cervical radicular pain
  1. Aaron Conger1,
  2. Richard W Kendall1,
  3. Beau P Sperry2,
  4. Russell Petersen3,
  5. Fabio Salazar1,
  6. Shellie Cunningham1,
  7. A Michael Henrie1,
  8. Erica F Bisson4,
  9. Masaru Teramoto1 and
  10. Zachary L McCormick1
  1. 1 Division of Physical Medicine and Rehabilitation, Department of Orthopedic Surgery, University of Utah, Salt Lake City, Utah, USA
  2. 2 Los Angeles David Geffen School of Medicine, University of California Los Angeles David Geffen School of Medicine, Los Angeles, California, USA
  3. 3 Alabama College of Osteopathic Medicine, Dothan, Alabama, USA
  4. 4 Department of Neurosurgery, The University of Utah School of Medicine, Salt Lake City, Utah, USA
  1. Correspondence to Dr Aaron Conger, Physical Medicine and Rehab, University of Utah Health, Salt Lake City, Utah, USA; aaron.conger{at}


Objectives The objective of this study was to evaluate the long-term effectiveness of catheter-directed cervical interlaminar epidural steroid injection with triamcinolone compared with cervical transforaminal steroid injection with dexamethasone for the treatment of refractory unilateral radicular pain.

Design Prospective, randomized, comparative trial.

Methods The primary outcome was the proportion of participants with ≥50% Numeric Rating Scale ‘dominant pain’ (the greater of arm vs neck) reduction from baseline. Secondary outcomes included ≥30% Neck Disability Index reduction and Patient Global Impression of Change response indicating ‘much improved’ or ‘very much improved’.

Results Data from 117 participants (55.6% women; 52.3±12.5 years of age; body mass index, 28.2±6.5 kg/m2) were analyzed. The proportion of participants who experienced ≥50% pain reduction at 1 month, 3 months, and 6 months has been previously reported. At 1 year, 61.2% (95% CI, 46.9% to 73.9%) of the catheter group compared with 51.9% (95% CI, 38.4% to 65.2%) of the transforaminal group reported ≥50% ‘dominant’ pain reduction (p=0.35). The proportion of participants who experienced ≥30% improvement in Neck Disability Index score was 60.4% (95% CI, 45.9% to 73.3%) and 47.1% (95% CI, 33.7% to 60.8%) in the catheter and transforaminal groups (p=0.18). Patient Global Impression of Change improvement was similar in both groups: 60.5% (95% CI, 44.2% to 74.8%) and 57.5% (95% CI, 41.7% to 71.9%) of the catheter and transforaminal groups reported being ‘much improved’ or ‘very much improved’, respectively (p=0.79).

Conclusion Both cervical catheter-directed interlaminar epidural injection and cervical transforaminal steroid injection were effective in reducing pain and disability in the majority of participants with refractory unilateral cervical radiculopathy for up to 1 year.

  • neck
  • radiculopathy
  • fluoroscopic
  • pain
  • treatment

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Cervical radiculopathy is a common source of pain and disability.1 Cervical epidural steroid injection (CESI) is often performed for patients who fail to respond to conservative treatment.2–4 Analysis of claims data for nearly 200 000 patients suggests that CESI is associated with reduction in surgical rates.5 Multiple techniques for the delivery of corticosteroid to epidural space exist, but few studies have compared the relative effectiveness of these approaches.

Interlaminar CESI with a catheter (C-CIESI) is performed by first accessing the dorsal epidural space with loss of resistance technique and contralateral oblique fluoroscopic imaging, followed by careful advancement of a soft radiopaque epidural catheter to the targeted level and side of symptoms. This approach allows for the safe use of particulate steroid, but risks include shearing and/or retention of a catheter fragment, dural puncture that has been reported with stiff neuroplasty catheters6 7 and epidural hematoma. This latter complication may occur despite appropriate management of anticoagulation and has been described in a case series of patients receiving C-CIESI.8 Additional risks include transient upper extremity weakness, but a recent study has suggested that low-dose lidocaine and saline produce similar degrees of myotomal weakness when used as dilutants during interlaminar CESI.9 The transforaminal approach to CESI (CTFESI) involves guiding a spinal needle directly into the intervertebral foramen near the cervical nerve root. Unique advantages of this approach may include greater diagnostic specificity,10 but risks include seizure or catastrophic neurologic injury when performed incorrectly or with particulate corticosteroid.11 12

This study is a continuation of a previously published randomized controlled trial (RCT) that reported 1-month, 3-month, and 6-month outcomes comparing the effectiveness of C-CIESI with CTFESI for the treatment of unilateral cervical radicular pain.13 At those time points, higher rates of neck pain reduction were observed at 1 month and 3 months for C-CIESI; other measures of treatment success including arm pain, analgesic medication utilization, function, Patient Global Impression of Change (PGIC), surgery rates, and repeat injection were similar between groups. Here, we report the 1-year outcomes from this RCT and discuss implications for clinical practice and future research.


Study design

This study was a single-center, randomized, non-blinded, comparative trial. The study was registered on (December 4, 2017) and approved by the local institutional review board (January 3, 2018). All participants provided written informed consent prior to enrollment. We described further details of study methodology previously.13

Recruitment, enrollment, and randomization

Study enrollment began in January 30, 2018, and the last participant was recruited on March 18, 2019. Participants were identified from clinic and procedural schedules and were referred for CESI by departments including Orthopedics, Physical Medicine and Rehabilitation, and Neurosurgery. Final determination for study eligibility was made by the study investigators. Participants were randomized using a computer-generated 1:1 allocation ratio into either the C-CIESI or CTFESI groups. Given the inherent differences in procedural technique for C-CIESI or CTFESI, participant and proceduralist blinding was not feasible; however, all other aspects of the treatment experience and follow-up protocol were identical.

Patient population

Adults aged 18–80 years with a clinical diagnosis of unilateral C4–C8 radicular pain supported by pathology on MRI consistent with clinical signs and symptoms were considered for inclusion. Pain score >4/10 pain on the Numerical Rating Scale (NRS) must have been present for at least 6 weeks despite conservative treatment (analgesic medications, chiropractic care, physical therapy, etc.). Exclusion criteria included refusal to participate or follow-up for the duration of the study, active infection, bleeding disorders, allergy to medications used, pregnancy, neurological disease (neuromuscular, demyelinating, or cerebrovascular), prior cervical spine surgery, current glucocorticoid use, or CESI within the prior 6 months.

Study interventions

All injections were performed by three board-certified physiatrists with subspecialty certification in either Sports Medicine or Pain Medicine. Both C-CIESI and CTFESI were performed in accordance with accepted guidelines and current safety recommendations11 12 14; we described these techniques previously.13 A repeat of the index procedure was offered to participants with recurrent pain who initially reported treatment success. Crossover to the alternative group was offered to participants in whom the index procedure was ineffective. Up to three repeat epidural injections were allowed during the study period.

Catheter-directed cervical interlaminar steroid injection

Participants were positioned prone on a fluoroscopy table. The cervical spine region was prepared and draped in usual sterile fashion. After injection of 1–2 mL of 1% lidocaine into the skin and subcutaneous tissues, an 18-gauge 3.5-inch RX Epidural Needle (Epimed International, Dallas, Texas, USA) was placed at the C7–T1 level per practice guidelines and current safety recommendations,11 12 and a loss of resistance technique was used to gain access to the epidural space. Needle position was confirmed in anteroposterior and contralateral oblique views. A radiopaque styleted 21-gauge epidural catheter (C-KATH; Epimed International, Dallas, Texas. USA) was placed into the introducer needle and advanced under live image guidance to the level and side of radicular pathology. Once a satisfactory target position was achieved with the catheter, 0.5–1 mL of contrast (Omnipaque-240; GE Healthcare, Princeton, New Jersey, USA) was injected. On confirmation of a satisfactory epidural contrast pattern, the injectate was delivered: 1 mL of triamcinolone acetonide (40 mg/mL) and 0.5 mL of 1% lidocaine (total volume, 1.5 mL).

Cervical transforaminal epidural steroid injection

Participants were positioned in supine or side-lying position on a fluoroscopy table. The cervical spine region was prepared and draped in usual sterile fashion. After injection of 1–2 mL of 1% lidocaine into the skin and subcutaneous tissues, a 25-gauge spinal needle was advanced to the appropriate target position in the neuroforamen under fluoroscopic guidance according to practice guidelines.11 Once a satisfactory target position was achieved and confirmed in both anteroposterior and oblique views, 0.5–1 mL of contrast (Omnipaque-240; GE Healthcare) was injected under live fluoroscopic observation with and without digital subtraction imaging. On confirmation of a satisfactory epidural contrast pattern without vascular uptake, the injectate was delivered: 1 mL of dexamethasone sodium phosphate (10 mg/mL) and 0.5 mL of 1% preservative-free lidocaine (total volume, 1.5 mL).

Outcome measures

Outcomes used to measure treatment success included NRS score, Neck Disability Index, five-item version (NDI-5),15 16 PGIC,17 and Medication Quantification Scale version III (MQS-III)18 at 12 months after injection. The outcomes observed at 1 month, 3 months, and 6 months have been previously reported.13 The primary outcome was the proportion of participants reporting ≥50% NRS reduction. This was further reported as a reduction in neck pain, arm pain, and ‘dominant pain’ (the greater of arm or neck pain). Secondary outcomes included the proportion of participants with ≥30% improvement in NDI-5 score,19 20 MQS-III score improvement of at least 6.8 (equal to 10 daily morphine equivalents),21 and PGIC score 6–7 (indicating ‘much improved’ or ‘very much improved’). These data were collected via a standardized email survey or via telephone interview with a clinical research coordinator.

Power analysis

The necessary sample size was determined based on a review of past literature wherein sample sizes of 20–30 participants showed statistically significant differences.22–24 Sixty participants were enrolled into each group to increase the power to detect between-group differences and given anticipated attrition at long-term follow-up, as previously described in additional detail.13

Statistical analysis

Descriptive statistics were calculated for participants’ demographic and clinical characteristics, as well as baseline neck pain NRS score, arm pain NRS score, ‘dominant’ NRS pain score, NDI-5 score, and MQS-III score. Group comparisons were performed using an independent t-test for continuous variables and Pearson’s χ2 test/Fisher’s exact test for categorical variables. If significant between-group differences in clinical, demographic, or healthcare utilization were identified, those variables were included in binary logistic regression models as covariates to ensure that the treatment effects were not significantly altered by those covariates. Contingency table analysis with calculation of a risk ratio (RR) and its 95% CI was performed for each of the dichotomous outcome variables above. Worst-case scenario and intention-to-treat analyses were applied to categorical outcomes wherein participants who were lost to follow-up, underwent surgery, or crossed over to the alternative treatment group were considered treatment failures.25 26


Figure 1 details the enrollment, randomization, and follow-up of selected participants. Between January 18, 2018, and March 18, 2019, 293 patients were evaluated for study inclusion. Of the 293 screened participants, 173 were excluded and 120 were randomized 1:1 to either C-CIESI or CTFESI. Three participants in the C-CIESI group canceled their scheduled procedure; therefore, outcome data for a total of 117 participants (n=57 for C-CIESI, n=60 for CTFESI) were analyzed. Rates of missing data were similar at all time points between the two groups.

Figure 1

CONSORT diagram showing the full details of enrollment, randomization, and follow-up through 12 months after the index injection. *Participants canceled their procedure before receiving the injection. †Unable to contact for follow-up assessments. ‡Based on intention-to-treat, participants lost to follow-up were considered treatment failures. Participants who elected crossover to the alternative route of epidural injection or underwent cervical spine surgery were also considered treatment failures according to their original allocation. C-CIESI, catheter-targeted cervical interlaminar epidural steroid injection; CONSORT, Consolidated Standards of Reporting Trials; CTFESI, cervical transforaminal epidural steroid injection.

Patients’ demographics, clinical characteristics, and healthcare utilization, along with baseline outcome variables, are summarized in table 1. Pain-related trauma was more common in the CTFESI group compared with the C-CIESI group (28.3% vs 12.3%; χ2=4.62; p=0.03). Consequently, this variable was included as a covariate in the subsequent binary logistic regression models. Baseline NDI-5 scores also differed significantly between the groups (t(115)=2.27; p=0.03). As the main outcome of interest was the score change from baseline to follow-up time point (accounting for the between-group baseline differences), NDI-5 score was ultimately not needed as a covariate in the binary logistic regression models. There were no significant baseline differences in any other demographic or clinical characteristic between the two groups (p>0.05). Overall, 53% of participants reported symptoms for ≤3 months.

Table 1

Demographic and clinical characteristics of participants at baseline (N=117)

Table 2 shows the contingency table analysis on the primary and secondary outcomes at 12 months. There were no significant between-group differences in any outcome. Binary logistic regression models on each of the dichotomous variables above, with pain-related trauma as a covariate, did not significantly change the treatment effects, as treatment (C-CIESI vs CTFESI) was not a significant predictor in any model (p>0.05).

Table 2

Proportions of participants with improvement at 12 months

The results of the contingency table analysis, as well as those of the binary logistic regression models, remained unchanged in a worst-case scenario analysis in which patients who were lost to follow-up were considered treatment failures. Specifically, the success rates (≥50% reduction) for ‘dominant’ NRS score at 12-month follow-up in the C-CIESI and CTFESI groups were 52.6% (95% CI, 39.7% to 65.3%) and 45.0% (95% CI, 32.9% to 57.8%), respectively. Similarly, at 12 months, the success rates (≥50% reduction) for neck pain reduction for C-CIESI and CTFESI were 40.4% (95% CI, 28.4% to 53.6%) and 33.3% (95% CI, 22.5% to 46.2%), respectively. At 12 months, arm pain was also similar in both groups, with 47.4% (95% CI, 34.7% to 60.3%) of participants in the C-CIESI group compared with 41.7% (95% CI, 29.8% to 54.5%) in the CTFESI group reporting ≥50% arm pain reduction. The percentages of participants reporting ≥30% reduction in NDI scores at 12 months were 50.9% (95% CI, 38.0% to 63.6%) and 40.0% (95% CI, 28.3% to 52.9%) for the C-CIESI and CTFESI groups, respectively. A separate analysis examining outcomes at 12 months in those who were initial non-responders at 1 month, 3 months, and 6 months showed that only one to three participants reported treatment success at that time point, depending on the outcome considered.

Group differences in healthcare utilization are presented in table 3. There were no significant between-group differences in the rates of repeat injection, imaging, or surgery. Overall, 13.6% of participants had cervical spine surgery, 31.6% had at least one repeat injection, and 29% had some type of repeat cervical imaging in the 12 months following the procedures.

Table 3

Healthcare utilization (N=117)


This report highlights the 12-month outcomes observed after both C-CIESI and CTFESI in patients with unilateral cervical radicular pain. Our prior report noted greater reductions in ‘dominant’ pain and neck pain between 1 month and 3 months in those treated with C-CIESI, without other significant between-group changes for other outcomes at those time points.13 Similar to the results seen at 6 months, at 1 year, there were no significant between-group differences observed in the primary analysis for any outcome including ‘dominant’ pain, neck pain, arm pain, NDI-5 score, MQS-III score, PGIC, and measures of healthcare utilization including repeat injection, imaging, or surgery (tables 2 and 3). Although catheter fracture/shearing and epidural hematoma have been reported after C-CIESI,6–8 no such events occurred during this study. Similarly, no serious adverse effects occurred in participants treated with CTFESI.

Improvements in pain and function observed at short-term follow-up were maintained in both groups at 12 months. For example, in the CTFESI group, 53% and 51% reported ≥50% reduction in arm pain at 1 month and 12 months. At 12 months, treatment success rates for arm pain, NDI-5, and MQS-III were relatively unchanged compared with those observed at 1 month; success rates seen at 1 month in both groups varied by less than 5% at all subsequent time points throughout the study. The higher rates of ‘dominant pain’ and neck pain reduction seen at short-term follow-up which favored C-CIESI were no longer apparent at 6 months and 12 months. This equilibration was driven by a marginal decline in neck and ‘dominant’ pain relief in the C-CIESI group, while pain relief appeared stable or slightly improved from 1 month to 12 months in the CTFESI group. Considering both those who received CTFESI and C-CIESI together, the overall percentage of those who reported ≥50% neck pain reduction was 46.3% (95% CI, 37.0% to 55.8%) at 1 month and 43.4% (95% CI, 33.9% to 53.5%) at 12 months. For arm pain, the success rate was 55.4% (95% CI, 45.5% to 64.9%) at 1 month and 55.9% (95% CI, 45.6% to 65.8%) at 12 months.13 Similarly, significant functional improvement was noted in 55.0% (95% CI, 45.5% to 64.2%) of all participants at 1 month and again in 53.5% (95% CI, 43.6% to 63.2%) at 12 months. These results are similar to those seen in prior studies of CESI, although few trials have reported 1-year results.3 27 28 In both groups, the vast majority of participants who were considered non-responders at 1 month, 3 months, and 6 months continued to report pain and functional impairment at 12 months.

Few participants advanced to cervical spine surgery. During the study period, only 13.6% of participants underwent surgery, with about 80% being performed before 6 months. The surgical rates in this study mirror closely those recently reported by Kleimeyer et al in a large retrospective analysis of national claims data (n=192 777) wherein 11.2% of patients ultimately underwent surgery by 6 months after CESI, increasing to 14.5% by 1 year.5 In this study, about 70% of participants received only one injection during the 12-month study period, a number similar to Kleimeyer et al.5

Historically, use of particulate corticosteroid has been thought to result in long-lasting pain relief due to a depot effect. While this study was not specifically designed to test the effectiveness of particulate versus non-particulate steroid for CESI, prior study of CTFESI with particulate versus non-particulate preparations has shown equivalent results.29–31 Given multiple reports of catastrophic neurologic injury and death, multisociety guidelines now recommend the exclusive use of non-particulate steroid for CTFESI.12 Despite these guidelines and equivalent results seen in studies to date, close to 10% of providers continue to use particulate steroid during CTFESI.4 For those who wish to use particulate steroid during CESI, the results of this study suggest that a catheter-directed interlaminar approach is likely equally effective and may be preferable for patients experiencing a strong component of neck pain.

Anatomical variables may impact the choice of route/approach for CESI. Studies have suggested that up to one in five individuals may have non-classic vertebral artery anatomy, including the presence of accessory vessels and loops which extend into the posterior neuroforamen.32 At higher cervical levels, CT-based imaging studies have shown periforaminal vessels in up to 40% of patients.33 In cases where a targeted injection at upper cervical levels is desired but CTFESI cannot be safely performed due potential vascular trespass, C-CIESI may be considered.

Limitations of this study must be recognized. Given the inherent approach differences with C-CIESI and CTFESI, blinding of proceduralists and participants was not feasible. Despite this, we observed low rates of crossover at all time points and no significant differences in rates of repeat injection or surgery between groups. Although not an inherent limitation, readers should be aware that about 50% of the total study population had pain for ≤3 months. The results of this study may not be fully generalizable to patients with more chronic symptoms. This study did not examine responses based on the etiology of radicular pain, as prior investigation into the radiographic predictors of treatment success for CESI has been decidedly mixed.34–40 Furthermore, recent studies have questioned the accuracy of conventional MRI for detecting the difference between disc and osteophyte.41 The strengths of this RCT include the wide range of outcomes captured, duration of follow-up, large size (n=117), and performance of interventions according to clinical practice guidelines.11


C-CIESI with triamcinolone and CTFESI with dexamethasone appeared similarly effective in reducing pain and disability for up to 12 months in the majority of patients with refractory unilateral cervical radicular pain. There were no significant between-group differences in the primary analysis for ‘dominant’ pain, neck pain, arm pain, function, medication use, patient impression of change, number of repeat injections, or surgery rates at 12 months following the index injection. This study provides further evidence that CESI results in substantial improvement in pain and function regardless of the approach used.

Ethics statements


The authors wish to thank Marilou Oh, Eric Leach, Alycia Hill, and Sara McEvoy for assistance with patient treatment and clinical follow-up



  • Correction notice This article has been corrected since it published Online First. Table 1 and the methods section have been updated.

  • Contributors RK, ZMcC, EFB, and AM were involved in conception and design of the study. AC, BPS, RP, FWS, SC, RK, ZMcC, EFB, and AMH were involved in acquisition of data. AC, ZMcC, and MT were involved in analysis of data. AC, ZMcC, RK, and MT were involved in interpretation of data. AC, BPS, and ZMcC were involved in drafting of the manuscript. AC, BPS, RP, FW, SC, RK, ZMcC, EFB, AMH, and MT were involved in critical revision of the work for important intellectual content. AC, BPS, RP, FWS, SC, RK, ZMcC, EFB, AMH, and MT provided approval of the version of the manuscript to be published. AC, BPS, RP, FWS, SC, RK, ZMcC, EFB, AMH, and MT provided agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.


  • Funding This was an investigator-initiated study supported by a grant from Epimed International. All funds were paid directly to the study investigators’ institution. No investigator salary support was provided by this grant.

  • Competing interests ZMcC serves on the Board of Directors of the Spine Intervention Society. There are no other potential conflicts of interest to disclose on the part of any of the other authors.

  • Provenance and peer review Not commissioned; externally peer reviewed.