The effectiveness of cervical medial branch radiofrequency ablation using a three-tined electrode: A real-world cross-sectional cohort study Authors: Hasan Sen, Napatpaphan Kanjanapanang, Amanda N. Cooper, Seth Bires, Audrey Adler, Akbar Nabi, Blake Dickenson, William Tang, Chase Young, Taylor Burnham, Allison Glinka Przybysz, Aaron Conger, Zachary L. McCormick Published: 2025 ABSTRACT BackgroundCervical medial branch radiofrequency ablation (CMBRFA) using conventional electrodes is a proven treatment for chronic cervical zygapophysial (facet) joint pain. More recently, multi-tined RFA technology has been adopted in clinical practice, warranting evaluation of treatment outcomes.ObjectivesEvaluate the effectiveness of CMBRFA using a three-tined RFA electrode for the treatment of cervical facet joint pain in a real-world cohort.MethodsIn this cross-sectional cohort study, electronic medical records of patients who underwent CMBRFA utilizing a three-tined RFA electrode at a single academic medical center between 2022 and 2024 were reviewed. Outcomes were collected at ≥6 months post-CMBRFA via standardized telephone survey, querying numerical rating scale (NRS) pain scores, patient global impression of change (PGIC) scores, and opioid use. The primary outcome was the proportion reporting ≥50 % NRS reduction. Secondary outcomes included ≥2-point NRS reduction, PGIC scores ≥6 (indicating at least “much improved”), and pre-versus post-CMBRFA opioid use. Potential predictors of treatment success were evaluated with multivariate Poisson regression analysis.Results66 patients were included. At mean follow-up of 16.3 ± 7.0 months, 51.5 % (95 %CI: 39.7–63.2) reported ≥50 % NRS reduction, 77.3 % (95 %CI: 65.8–85.7) experienced ≥2-point NRS reduction, and 63.6 % (95 %CI: 51.6–74.2) had PGIC ≥6. Opioid use was decreased significantly by 15.2 % (95 %CI: 5.0–25.3). Follow-up time and repeat CMBRFA were significantly associated with outcomes for PGIC.ConclusionIn this real-world cohort, we observed clinically meaningful improvements in neck pain (defined as ≥50 % NRS reduction from baseline) in 51.5 % of participants at an average of approximately 16 months after treatment with CMBRFA using a three-tined electrode. Introduction Neck pain is one of the most common musculoskeletal complaints in the United States and a significant global health concern, contributing to considerable disability and healthcare burden [1,2]. Cervical zygapophysial (facet) joint dysfunction is a well-recognized source of chronic neck pain, affecting an estimated 50 % of patients who present with neck pain [[3], [4], [5], [6]]. This high prevalence underscores the need for effective treatments for cervical facet joint pain. Cervical medial branch radiofrequency ablation (CMBRFA) is a minimally invasive procedure used to treat chronic cervical facet joint pain by interrupting nociceptive signals in the medial branches of the dorsal rami [7]. By creating a thermal lesion to disrupt pain transmission, CMBRFA has become a widely utilized treatment for facet-mediated neck pain [8,9]. The current technique recommended by the International Pain and Spine Intervention Society (IPSIS) for CMBRFA is a two-stage procedure. It involves oblique and sagittal insertions of a conventional monopolar electrode to achieve parallel alignment with the target nerve(s), which is intended to maximize the area of nerve coagulation and potentially prolong the duration of pain relief [7]. Excellent clinical outcomes have been demonstrated when this technique is used [10]. This approach requires a high level of technical skill to lay the electrode parallel to the third occipital nerve and/or medial branch nerve for effective facet joint denervation. It also requires a posterior approach with traverse of a considerable amount of soft tissue and is significantly more time intensive than a perpendicular approach [10,11]. As such, interest has grown in novel radiofrequency ablation (RFA) technology that allows for a perpendicular approach to the target nerve(s) to potentially increase procedural efficiency and patient comfort. This includes a three-tined electrode that creates a pyramidal-shaped lesion, allowing for a lateral approach to the cervical articular pillar [12] in which the skin-to-nerve distance is significantly reduced compared to a posterior approach in most patients. The pyramidal lesion geometry produced by the multi-tined electrode enlarges the ablation volume around the most aspect of the cannula, increasing the likelihood of encompassing the targeted medial branch, particularly when the cannula is positioned perpendicularly to the articular pillar. This design obviates the need for parallel alignment with the medial branch nerve required with conventional monopolar electrodes. Once the cannula has been placed in its final position, expandable tines are deployed to increase the active treatment area. This single-pass, perpendicular trajectory method using a three-tined electrode has the potential to reduce procedure time, radiation exposure, and patient discomfort. Given these potential benefits, the present study aimed to evaluate the effectiveness of CMBRFA using a three-tined electrode with a perpendicular approach in the context of routine clinical practice. Understanding the long-term real-world impact of CMBRFA with a three-tined electrode will help refine clinical decision-making and optimize treatment strategies for patients suffering from chronic cervical facet pain. Methods Data collection This cross-sectional cohort study was conducted at a single tertiary academic medical center. Local Institutional Review Board (IRB 00187326) approval was obtained. The electronic medical records of consecutive patients who underwent CMBRFA using a three-tined electrode from 2022 to 2024 were reviewed. If an individual received multiple CMBRFAs during this time period, only the most recent procedure was included in the analysis. The number of previous CMBRFAs on record was tallied for all participants. All patients had chronic axial neck pain consistent with a facet-mediated pain pattern, typically characterized by reproduction of pain with extension-rotation and facet joint tenderness on examination, without clinical signs or symptoms suggesting a primary radiculopathy. Imaging findings (X-ray or MRI) demonstrating facet arthropathy were considered in the context of full clinical presentation. Both physical examination and imaging findings were considered supportive but not diagnostic in isolation. The diagnosis of for cervical zygapophysial (facet) joint pain was ultimately confirmed or refuted by dual diagnostic medial branch blocks. Patient inclusion criteria were (1) age ≥18 years; (2) previous report of ≥80 % pain relief with dual, concordant medial branch nerve blocks (MBBs) to confirm a diagnosis of cervical facet pain; (3) CMBRFA procedure performed with three-tined RFA electrode technology; (4) cervical spine pain severity score ≥4/10 on Numeric Rating Scale (NRS) documented within 2 months prior to CMBRFA procedure; (5) ≥6 months between CMBRFA and follow-up outcome survey; and (6) willingness to participate in a post-procedural phone call survey. Data extraction was performed by 4 authors (HS, WT, ANC, AVA). Data collected included the following: (1) age; (2) gender; (3) body mass index (BMI); (4) current smoking status; (5) duration of neck pain prior to CMBRFA in years; (6) baseline daily anxiety or depression medication use; (7) baseline opioid use; (8) presence of cervical fusion; (9) highest recorded numeric rating scale (NRS) neck pain score within two months before CMBRFA; (10) percentage of pain improvement with each MBB; (11) date of CMBRFA; (12) CMBRFA level(s) and laterality; and (13) number of previous CMBRFAs. Patients who met the inclusion criteria were contacted via a letter sent on behalf of their treating physician regarding the research study. Those who agreed to participate completed a standardized post-CMBRFA phone call survey that captured current and 7-day average NRS pain scores, self-reported improvement by the Patient Global Impression of Change (PGIC), and whether patients were using daily opioid analgesics to manage their index neck pain. Of the two NRS measures collected, the 7-day average NRS score was designated a priori as the analytic pain measure. The current NRS score was recorded for descriptive context only. Procedures All procedures were performed by Physical Medicine and Rehabilitation physician faculty members with fellowship training in interventional spine procedures. Third occipital nerve and medial branch blocks Medial branch blocks were performed according to the technique described in the IPSIS Technical Manual and Atlas of Interventional Pain and Spine Procedures [13]. Patients were positioned in the lateral recumbent position. The skin was prepped and draped in a sterile fashion. The skin was anesthetized with approximately 1 mL of 1 % lidocaine at each site. Using a lateral fluoroscopic approach, 25-gauge, 1.5–2.5-inch short bevel needles were advanced to the appropriate anatomic landmarks, identifying the third occipital nerve (TON) and C3–C7 medial branch nerves [13]. Correct needle placement was confirmed with an anteroposterior (AP) view on fluoroscopy. Next, approximately 0.1–0.3 mL of contrast medium was injected at each site to ensure proper needle placement and rule out intravascular injection. After correct needle placement was confirmed, 0.3–0.5 mL of either 4 % lidocaine or 0.5 % bupivacaine was injected. Post-procedurally, patients were provided a pain log and instructed to document symptom improvement in 15-min increments for 6 h using the NRS. Positive response to MBBs was defined as ≥80 % pain reduction of concordant duration with the local anesthetic used. Radiofrequency ablation CMBRFA was performed using an 18- gauge multi-tined cannula with a 5-mm active tip (Diros Technology Inc., Markham, Ontario, Canada). Patients were positioned in the lateral decubitus position, and a true-lateral view of the cervical articular pillar was obtained. The RFA cannula was inserted using a trajectory approach to achieve perpendicular placement relative to the target nerve(s). Cannula progression was monitored using lateral and AP fluoroscopic projections (Fig. 1). Once the target was reached, the tines of the cannula were deployed under direct visualization. Appropriate needle position was confirmed with an AP view. Next, 1–2 mL of 1–2 % lidocaine was injected over each medial branch nerve (MBN). A single lesion was performed at each target nerve by heating the electrode to 80–85 °C for 90–180s after a ramp-up time of 15 or 30s, as determined by the treating physician.Fig. 1Fluoroscopic images of cervical medial branch radiofrequency ablation demonstrating three-tined cannula placement with tines deployed. (A) Lateral view showing proper alignment (true lateral view) of the C5 lateral mass; note that at this fluoroscopic angle, the C4 lateral masses are largely superimposed but a true-lateral of the segment is not present. (B) Anteroposterior view showing cannulae positioned within the waists of the lateral masses of C4 and C5.Fig. 1 Fluoroscopic images of cervical medial branch radiofrequency ablation demonstrating three-tined cannula placement with tines deployed. (A) Lateral view showing proper alignment (true lateral view) of the C5 lateral mass; note that at this fluoroscopic angle, the C4 lateral masses are largely superimposed but a true-lateral of the segment is not present. (B) Anteroposterior view showing cannulae positioned within the waists of the lateral masses of C4 and C5. Outcomes The primary study outcome was the proportion of patients with ≥50 % NRS pain reduction at least six months after CMBRFA. Secondary outcomes included (1) mean NRS pain score reduction from baseline, (2) the proportion of patients who reported ≥2-point NRS reduction from baseline, which is the minimal clinically important change (MCID) for neck pain [14], (3) the proportion of patients who reported being “much improved” or “very much improved” on PGIC with a score of 6 or 7, and (4) the proportion of patients using opioids at follow-up compared to baseline. Statistical analysis Descriptive statistics were calculated for patient demographic, clinical, and procedure-related variables, using means/standard deviations for continuous variables and frequencies/percentages for categorical variables. A 95 % confidence interval (CI) was also calculated for statistics of select outcome variables. Relationships between categorical outcomes for NRS and PGIC and selected covariates were explored using multivariate Poisson regression models with the robust variance estimator [[15], [16], [17]], along with calculations of incidence rate ratios (IRRs) and their 95 % CIs. The included predictor variables were follow-up time period, age, opioid use at baseline, antidepressant/anxiolytic medication use at baseline, and whether each participant's CMBRFA was a first-time or repeat procedure. The regression models for ≥50 % and ≥2-point NRS reduction included an additional covariate of baseline NRS score. All analyses were conducted in Stata 18.0 (StataCorp, LLC, College Station, TX, United States) using a significance level of 0.05. Results Of 129 consecutive patients identified, 66 participants (69.7 % female; mean age 62.0 ± 13.6 years; mean BMI 27.3 ± 5.4 kg/m2) met eligibility criteria and successfully completed the standardized outcomes survey. Participant demographics, clinical characteristics, and procedural variables are summarized in Table 1. The majority (n = 56; 84.8 %) of participants in this cohort underwent first-time procedures, while 10 individuals had previously received at least one CMBRFA prior to the study. The mean follow-up time was 16.3 ± 7.0 months across the full cohort. Follow-up outcome data were also stratified into four categories: 6–12 months (n = 21; 31.8 %), 12–18 months (n = 22; 33.3 %), 18–24 months (n = 12; 18.2 %), and ≥24 months (n = 11; 16.7 %) post-CMBRFA.Table 1Baseline participant demographics, clinical characteristics, and procedural variables (N = 66).Table 1VariableNo. (%)Follow-up 6–12 months21 (31.8) 12–18 months22 (33.3) 18–24 months12 (18.2) ≥ 24 months11 (16.7)Gender Male20 (30.3) Female46 (69.7)Current smoking status No62 (93.9) Yes4 (6.1)History of cervical fusion No53 (80.3) Yes13 (19.7)Number of prior CMBRFAs 056 (84.8) 19 (13.6) ≥ 21 (1.5)Daily anxiety/depression medication use No34 (51.5) Yes32 (48.5)Baseline opioid use No51 (77.3) Yes15 (22.7)Percentage pain relief with MBBs 80–99 %27 (40.9) 100 %39 (59.1)CMBRFA laterality Unilateral50 (75.8) Bilateral16 (24.2)Age, mean ± SD62.0 ± 13.6BMI, mean ± SD27.3 ± 5.4Pain duration in years, mean ± SD6.7 ± 6.7Follow-up time in months, mean ± SD16.3 ± 7.0CMBRFA = cervical medial branch radiofrequency ablation; MBB = medial branch block; SD = standard deviation. Table 1Baseline participant demographics, clinical characteristics, and procedural variables (N = 66).Table 1VariableNo. (%)Follow-up 6–12 months21 (31.8) 12–18 months22 (33.3) 18–24 months12 (18.2) ≥ 24 months11 (16.7)Gender Male20 (30.3) Female46 (69.7)Current smoking status No62 (93.9) Yes4 (6.1)History of cervical fusion No53 (80.3) Yes13 (19.7)Number of prior CMBRFAs 056 (84.8) 19 (13.6) ≥ 21 (1.5)Daily anxiety/depression medication use No34 (51.5) Yes32 (48.5)Baseline opioid use No51 (77.3) Yes15 (22.7)Percentage pain relief with MBBs 80–99 %27 (40.9) 100 %39 (59.1)CMBRFA laterality Unilateral50 (75.8) Bilateral16 (24.2)Age, mean ± SD62.0 ± 13.6BMI, mean ± SD27.3 ± 5.4Pain duration in years, mean ± SD6.7 ± 6.7Follow-up time in months, mean ± SD16.3 ± 7.0CMBRFA = cervical medial branch radiofrequency ablation; MBB = medial branch block; SD = standard deviation. Baseline participant demographics, clinical characteristics, and procedural variables (N = 66). CMBRFA = cervical medial branch radiofrequency ablation; MBB = medial branch block; SD = standard deviation. At the average follow-up time of 16.3 ± 7.0 months, mean NRS score reduction from baseline was 3.1 ± 2.5 points for the entire cohort (Table 2a), while 51.5 % (n = 34; 95 % CI: 39.7, 63.2) and 77.3 % (n = 62; 95 % CI: 65.8, 85.7) of participants reported ≥50 % and ≥2-point NRS reductions, respectively (Table 2b). The proportion of responders at each follow-up time interval did not differ statistically for either outcome measure (p > 0.05). Similarly, there were no significant differences in responder rates for participants whose CMBRFA was a first-time procedure versus a repeated treatment (p > 0.05; Table 2c).Table 2aContinuous outcomes for pain relief by follow-up period.Table 2aOutcomeMeanSDMinMax95 % CI (means)Baseline NRS score6.41.64106.0, 6.8NRS score reduction from baseline (baseline score minus follow-up score) 6–12 months3.12.5−282.0, 4.3 12–18 months3.12.3−571.9, 4.2 18–24 months3.32.3071.9, 4.8 ≥ 24 months2.92.8−181.0, 4.8 16.3 ± 7.0 months, mean ± SD3.12.5−582.5, 3.7CI = confidence interval; NRS = Numerical Rating Scale; SD = standard deviation.Table 2bCategorical outcomes for pain relief by follow-up period.Table 2bOutcomeYesNop*95 % CI (yes)≥50 % NRS reduction 6–12 months10 (47.6)11 (52.4)0.91828.3, 67.6 12–18 months12 (54.5)10 (45.5)34.7, 73.1 18–24 months7 (58.3)5 (41.7)32.0, 80.7 ≥ 24 months5 (45.5)6 (54.5)21.3, 72.0 16.3 ± 7.0 months, mean ± SD34 (51.5)32 (48.5)39.7, 63.2≥2-point NRS reduction 6–12 months16 (76.2)5 (23.8)0.24654.9, 89.4 12–18 months19 (86.4)3 (13.6)66.7, 95.3 18–24 months10 (83.3)2 (16.7)55.2, 95.3 ≥ 24 months6 (54.5)5 (45.5)28.0, 78.7 16.3 ± 7.0 months, mean ± SD51 (77.3)15 (22.7)65.8, 85.7Note: Values are frequency (%), unless specified otherwise. *From Fisher's exact test.CI = confidence interval; NRS = Numerical Rating Scale; SD = standard deviation.Table 2cCategorical outcomes for pain relief by first-time versus repeat procedures.Table 2cOutcomeYesNop*95 % CI (yes)≥50 % NRS reduction First-time CMBRFA28 (50.0)28 (50.0)0.73437.3, 62.7 Repeat CMBRFA6 (60.0)4 (40.0)31.3, 83.2≥2-point NRS reduction First-time CMBRFA43 (76.8)13 (23.2)1.00064.2, 85.9 Repeat CMBRFA8 (80.0)2 (20.0)49.0, 94.3Note: Values are frequency (%), unless specified otherwise. *From Fisher's exact test.CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; NRS = Numerical Rating Scale. Table 2aContinuous outcomes for pain relief by follow-up period.Table 2aOutcomeMeanSDMinMax95 % CI (means)Baseline NRS score6.41.64106.0, 6.8NRS score reduction from baseline (baseline score minus follow-up score) 6–12 months3.12.5−282.0, 4.3 12–18 months3.12.3−571.9, 4.2 18–24 months3.32.3071.9, 4.8 ≥ 24 months2.92.8−181.0, 4.8 16.3 ± 7.0 months, mean ± SD3.12.5−582.5, 3.7CI = confidence interval; NRS = Numerical Rating Scale; SD = standard deviation. Continuous outcomes for pain relief by follow-up period. CI = confidence interval; NRS = Numerical Rating Scale; SD = standard deviation. Table 2bCategorical outcomes for pain relief by follow-up period.Table 2bOutcomeYesNop*95 % CI (yes)≥50 % NRS reduction 6–12 months10 (47.6)11 (52.4)0.91828.3, 67.6 12–18 months12 (54.5)10 (45.5)34.7, 73.1 18–24 months7 (58.3)5 (41.7)32.0, 80.7 ≥ 24 months5 (45.5)6 (54.5)21.3, 72.0 16.3 ± 7.0 months, mean ± SD34 (51.5)32 (48.5)39.7, 63.2≥2-point NRS reduction 6–12 months16 (76.2)5 (23.8)0.24654.9, 89.4 12–18 months19 (86.4)3 (13.6)66.7, 95.3 18–24 months10 (83.3)2 (16.7)55.2, 95.3 ≥ 24 months6 (54.5)5 (45.5)28.0, 78.7 16.3 ± 7.0 months, mean ± SD51 (77.3)15 (22.7)65.8, 85.7Note: Values are frequency (%), unless specified otherwise. *From Fisher's exact test.CI = confidence interval; NRS = Numerical Rating Scale; SD = standard deviation. Categorical outcomes for pain relief by follow-up period. Note: Values are frequency (%), unless specified otherwise. *From Fisher's exact test. CI = confidence interval; NRS = Numerical Rating Scale; SD = standard deviation. Table 2cCategorical outcomes for pain relief by first-time versus repeat procedures.Table 2cOutcomeYesNop*95 % CI (yes)≥50 % NRS reduction First-time CMBRFA28 (50.0)28 (50.0)0.73437.3, 62.7 Repeat CMBRFA6 (60.0)4 (40.0)31.3, 83.2≥2-point NRS reduction First-time CMBRFA43 (76.8)13 (23.2)1.00064.2, 85.9 Repeat CMBRFA8 (80.0)2 (20.0)49.0, 94.3Note: Values are frequency (%), unless specified otherwise. *From Fisher's exact test.CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; NRS = Numerical Rating Scale. Categorical outcomes for pain relief by first-time versus repeat procedures. Note: Values are frequency (%), unless specified otherwise. *From Fisher's exact test. CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; NRS = Numerical Rating Scale. Overall, 63.6 % (n = 42; 95 % CI: 51.6, 74.2) of participants reported PGIC scores ≥6 at the mean follow-up time of 16.3 ± 7.0 months (Table 3a). As with the outcomes for pain relief, responder rates for ≥6 on PGIC did not vary significantly between follow-up intervals (p > 0.05). No significant disparities in this outcome measure were identified between participants with first-time versus repeat CMBRFA procedures (p > 0.05; Table 3b).Table 3aOutcomes for Patient Global Impression of Change by follow-up period.Table 3aOutcomeYesNop*95 % CI (yes)≥6 PGIC 6–12 months17 (81.0)4 (19.0)0.07460.0, 92.3 12–18 months10 (45.5)12 (54.5)26.9, 65.3 18–24 months9 (75.0)3 (25.0)46.8, 91.1 ≥ 24 months6 (54.5)5 (45.5)28.0, 78.7 16.3 ± 7.0 months, mean ± SD42 (63.6)24 (36.4)51.6, 74.2Note: Values are frequency (%), unless specified otherwise. PGIC scores ≥6 indicate at least “much improved”. *From Fisher's exact test.CI = confidence interval; PGIC = Patient Global Impression of Change; SD = standard deviation.Table 3bOutcomes for Patient Global Impression of Change by first-time versus repeat procedures.Table 3bOutcomeYesNop*95 % CI (yes)≥6 PGIC First-time CMBRFA34 (60.7)22 (39.3)0.30647.6, 72.4 Repeat CMBRFA8 (80.0)2 (20.0)49.0, 94.3Note: Values are frequency (%), unless specified otherwise. PGIC scores ≥6 indicate at least “much improved”. *From Fisher's exact test.CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; PGIC = Patient Global Impression of Change. Table 3aOutcomes for Patient Global Impression of Change by follow-up period.Table 3aOutcomeYesNop*95 % CI (yes)≥6 PGIC 6–12 months17 (81.0)4 (19.0)0.07460.0, 92.3 12–18 months10 (45.5)12 (54.5)26.9, 65.3 18–24 months9 (75.0)3 (25.0)46.8, 91.1 ≥ 24 months6 (54.5)5 (45.5)28.0, 78.7 16.3 ± 7.0 months, mean ± SD42 (63.6)24 (36.4)51.6, 74.2Note: Values are frequency (%), unless specified otherwise. PGIC scores ≥6 indicate at least “much improved”. *From Fisher's exact test.CI = confidence interval; PGIC = Patient Global Impression of Change; SD = standard deviation. Outcomes for Patient Global Impression of Change by follow-up period. Note: Values are frequency (%), unless specified otherwise. PGIC scores ≥6 indicate at least “much improved”. *From Fisher's exact test. CI = confidence interval; PGIC = Patient Global Impression of Change; SD = standard deviation. Table 3bOutcomes for Patient Global Impression of Change by first-time versus repeat procedures.Table 3bOutcomeYesNop*95 % CI (yes)≥6 PGIC First-time CMBRFA34 (60.7)22 (39.3)0.30647.6, 72.4 Repeat CMBRFA8 (80.0)2 (20.0)49.0, 94.3Note: Values are frequency (%), unless specified otherwise. PGIC scores ≥6 indicate at least “much improved”. *From Fisher's exact test.CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; PGIC = Patient Global Impression of Change. Outcomes for Patient Global Impression of Change by first-time versus repeat procedures. Note: Values are frequency (%), unless specified otherwise. PGIC scores ≥6 indicate at least “much improved”. *From Fisher's exact test. CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; PGIC = Patient Global Impression of Change. At baseline prior to CMBRFA, 22.7 % (n = 15; 95 % CI: 14.3, 34.2) of participants were using opioids, compared to only 7.6 % using opioids at the mean follow-up time of 16.3 ± 7.0 months (Table 4a). The 10 participants who had ceased opioids at follow-up assessment represented a statistically significant decrease in opioid utilization by 15.2 % (95 % CI: 5.0, 25.3) pre-versus post-CMBRFA (p = 0.002; Table 4b). No participants had initiated opioids at follow-up when outcomes were collected.Table 4aOpioid use at baseline and follow-up.Table 4aOpioid useYesNo95 % CI (yes)Baseline15 (22.7)51 (77.3)14.3, 34.2Follow-up5 (7.6)61 (92.4)3.3, 16.5Note: Values are frequency (%), unless specified otherwise.CI = confidence interval.Table 4bChanges in opioid use from baseline to follow-up.Table 4bOpioid use (follow-up)Difference in 9roportions (95 % CI)Opioid use (baseline)YesNop*Yes5 (7.6)10 (15.2)0.00215.2 (5.0, 25.3)No0 (0.0)51 (77.3)Note: Values are frequency (%), unless specified otherwise. Bold font denotes statistical significance at p < 0.05. *From McNemar's exact test.CI = confidence interval. Table 4aOpioid use at baseline and follow-up.Table 4aOpioid useYesNo95 % CI (yes)Baseline15 (22.7)51 (77.3)14.3, 34.2Follow-up5 (7.6)61 (92.4)3.3, 16.5Note: Values are frequency (%), unless specified otherwise.CI = confidence interval. Opioid use at baseline and follow-up. Note: Values are frequency (%), unless specified otherwise. CI = confidence interval. Table 4bChanges in opioid use from baseline to follow-up.Table 4bOpioid use (follow-up)Difference in 9roportions (95 % CI)Opioid use (baseline)YesNop*Yes5 (7.6)10 (15.2)0.00215.2 (5.0, 25.3)No0 (0.0)51 (77.3)Note: Values are frequency (%), unless specified otherwise. Bold font denotes statistical significance at p < 0.05. *From McNemar's exact test.CI = confidence interval. Changes in opioid use from baseline to follow-up. Note: Values are frequency (%), unless specified otherwise. Bold font denotes statistical significance at p < 0.05. *From McNemar's exact test. CI = confidence interval. Poisson regression models for outcomes on NRS and PGIC are presented in Table 5. The analysis revealed no significant associations between ≥50 % or ≥2-point NRS reduction and the selected predictor variables. However, the regression model for PGIC identified both follow-up time and repeat CMBRFA as significant covariates. Participants who completed the follow-up outcomes survey at 12–18 months post-CMBRFA were 44 % less likely to report PGIC scores ≥6 than those with a follow-up period of 6–12 months (IRR = 0.56; 95 % CI: 0.34, 0.92; p = 0.022). Compared to first-time CMBRFA, participants who returned for repeat CMBRFA were 44 % more likely to have a positive PGIC response at follow-up (IRR = 1.44; 95 % CI: 1.02, 2.02; p = 0.039).Table 5Poisson regression models on categorical study outcomes with select covariates.Table 5OutcomePredictorIRR95 % CIp≥50 % NRS reductionFollow-up time (vs. 6–12 months)12–18 months1.100.63, 1.930.73518–24 months1.440.80, 2.590.219≥24 months1.030.45, 2.400.939Baseline anxiety/depression medication (vs. no)Yes0.810.50, 1.290.375Baseline opioid use (vs. no)Yes0.750.40, 1.410.374Repeat CMBRFA (vs. no)Yes1.270.71, 2.290.417Age0.990.97, 1.000.116Baseline NRS score0.880.76, 1.010.069≥2-point NRS reductionFollow-up time (vs. 6–12 months)12–18 months1.170.88, 1.570.28318–24 months1.210.88, 1.670.244≥24 months0.750.44, 1.310.313Baseline anxiety/depression medication (vs. no)Yes0.900.69, 1.180.445Baseline opioid use (vs. no)Yes0.720.46, 1.130.156Repeat CMBRFA (vs. no)Yes0.980.69, 1.390.923Age1.000.99, 1.010.769Baseline NRS score1.070.99, 1.150.080≥6 PGICFollow-up time (vs. 6–12 months)12–18 months0.560.34, 0.920.02218–24 months0.960.63, 1.460.834≥24 months0.690.40, 1.190.182Baseline anxiety/depression medication (vs. no)Yes0.800.56, 1.150.228Baseline opioid use (vs. no)Yes0.850.54, 1.350.505Repeat CMBRFA (vs. no)Yes1.441.02, 2.020.039Age1.010.99, 1.020.527Note: Bold font denotes statistical significance at p < 0.05.CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; IRR = incidence rate ratio; NRS = Numerical Rating Scale; PGIC = Patient Global Impression of Change. Table 5Poisson regression models on categorical study outcomes with select covariates.Table 5OutcomePredictorIRR95 % CIp≥50 % NRS reductionFollow-up time (vs. 6–12 months)12–18 months1.100.63, 1.930.73518–24 months1.440.80, 2.590.219≥24 months1.030.45, 2.400.939Baseline anxiety/depression medication (vs. no)Yes0.810.50, 1.290.375Baseline opioid use (vs. no)Yes0.750.40, 1.410.374Repeat CMBRFA (vs. no)Yes1.270.71, 2.290.417Age0.990.97, 1.000.116Baseline NRS score0.880.76, 1.010.069≥2-point NRS reductionFollow-up time (vs. 6–12 months)12–18 months1.170.88, 1.570.28318–24 months1.210.88, 1.670.244≥24 months0.750.44, 1.310.313Baseline anxiety/depression medication (vs. no)Yes0.900.69, 1.180.445Baseline opioid use (vs. no)Yes0.720.46, 1.130.156Repeat CMBRFA (vs. no)Yes0.980.69, 1.390.923Age1.000.99, 1.010.769Baseline NRS score1.070.99, 1.150.080≥6 PGICFollow-up time (vs. 6–12 months)12–18 months0.560.34, 0.920.02218–24 months0.960.63, 1.460.834≥24 months0.690.40, 1.190.182Baseline anxiety/depression medication (vs. no)Yes0.800.56, 1.150.228Baseline opioid use (vs. no)Yes0.850.54, 1.350.505Repeat CMBRFA (vs. no)Yes1.441.02, 2.020.039Age1.010.99, 1.020.527Note: Bold font denotes statistical significance at p < 0.05.CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; IRR = incidence rate ratio; NRS = Numerical Rating Scale; PGIC = Patient Global Impression of Change. Poisson regression models on categorical study outcomes with select covariates. Note: Bold font denotes statistical significance at p < 0.05. CI = confidence interval; CMBRFA = cervical medial branch radiofrequency ablation; IRR = incidence rate ratio; NRS = Numerical Rating Scale; PGIC = Patient Global Impression of Change. Lastly, we performed several exploratory post-hoc analyses to aid interpretation of our results with respect to the observed losses to follow-up and responses to dual concordant MBBs. Nearly half (n = 63) of identified patients either declined to participate or could not be contacted about the study. As such, supplemental best- and worst-case analyses were conducted by treating all individuals lost to follow-up as treatment responders and non-responders, respectively, with results summarized in Table S1 and S2. The range of possible responder rates across best- and worst-case scenarios was 26.4–75.2 % for ≥50 % NRS reduction, 39.5–88.4 % for ≥2-point NRS reduction, and 32.6–81.4 % for ≥6 on PGIC. Because participants were required to have ≥80 % pain relief with diagnostic MBBs before proceeding to CMBRFA, outcomes for pain reduction using a higher cutoff threshold of ≥80 % are presented in Table S3. As expected, the responder rate at average follow-up decreased to 24.2 % (95 % CI: 24.2, 75.8) when using a more stringent definition of ≥80 % NRS reduction. Finally, Table S4 shows categorical outcomes stratified by percentage pain relief with diagnostic blocks (80–99 % vs. 100 %), as response to MBBs should predict the magnitude of pain reduction after CMBRFA. Compared to those with a block response of 80–99 % relief, participants with complete pain relief had significantly higher responder rates for ≥80 % NRS reduction (33.3 % vs. 11.1 %; p = 0.045), ≥2-point NRS reduction (87.2 % vs. 63.0 %; p = 0.035), and ≥6 on PGIC (74.4 % vs. 48.1 %; p = 0.039), but not ≥50 % NRS reduction (56.4 % vs. 44.4 %; p = 0.453). Discussion This cross-sectional cohort study aimed to assess clinical outcomes of CMBRFA utilizing a three-tined RFA electrode with a perpendicular approach to the third occipital nerve(s) and/or medial branch nerve(s) in a real-world clinical population. At a mean follow-up time of approximately 16 months after CMBRFA, 52 % and 77 % of participants reported ≥50 % and ≥2-point NRS pain score reductions from baseline, respectively. This responder rate was similar to our findings when evaluating the use of an internally-cooled RFA electrode, a technology that also allows for a perpendicular approach to the cervical articular pillar, in a real-world cohort of patients who underwent CMBRFA [18]. In that study, Burnham et al. observed that 54 % of participants experienced ≥50 % improvement in neck pain at a mean follow-up of 17 months. Notably, the present study also demonstrated reduced opioid usage from baseline to follow-up. Within the study cohort a 15 % decrease in the proportion of participants consuming opioid pain medication was observed from baseline to follow-up. This observed reduction in opioid use is consistent with the findings of studies that have evaluated opioid use in patients who underwent lumbar medial branch RFA for lumbosacral facet joint pain [19,20]. Regression modeling to explore potential predictors of treatment success revealed no associations between outcomes for NRS and the chosen demographic, clinical, and procedural covariates. However, the model for ≥6 on PGIC did indicate significant effects of follow-up time and repeat CMBRFA. The likelihood of reporting PGIC scores ≥6 was decreased among participants with an intermediate follow-up time interval of 12–18 months. This pattern likely reflects the natural decline in CMBRFA effectiveness over time [21]. Patients at 12–18 months post-procedure may be transitioning out of the peak benefit phase, leading to lower PGIC scores as symptom relief begins to fade. In contrast, those at 18–24 and ≥ 24 months post-procedure may have adjusted their expectations or adapted to their baseline, resulting in higher perceived benefit despite the time elapsed. Additionally, participants with repeat CMBRFA procedures were significantly more likely to report PGIC scores ≥6 than those who underwent first-time CMBRFA. Multi-society consensus guidelines published for CMBRFA recommend offering repeat procedures to patients whose initial CMBRFA provided clinically significant pain relief lasting a minimum of 3 months [22]. Intuitively, participants who opt to repeat CMBRFA after return of index symptoms were probably satisfied with results of the original CMBRFA. In the present study, initial treatment responders may have anticipated that subsequent CMBRFA procedures would confer similar benefits, prompting this subset of patients to assign higher ratings for overall improvement on PGIC based on previous experience. There are limited studies on CMBRFA using the multi-tined Trident cannula. Künzle et al. reported that 58 % of patients achieved a ≥30 % reduction in pain, with 64.7 % of patients reporting a PGIC score of 6–7, at 2 months post-procedure in an observational study [12]. Filiatrault et al. conducted a double-blind randomized controlled trial comparing a multi-tined cannula (MTC) to a conventional cannula (CC) for CMBRFA [11]. While no significant difference in pain reduction was observed at 3 months, the CC group demonstrated significantly higher rates of ≥50 % NRS reduction at 6 months (61.1 % vs. 25.0 %) and 12 months (70.6 % vs. 17.4 %) compared to the MTC group. The multi-tined Trident cannula, alternatively, offered shorter procedure times and reduced radiation exposure. Procedural pain did not differ significantly between groups. However, three patients in the CC group requested premature termination of the procedure due to severe discomfort and were subsequently excluded from analysis, highlighting a potential bias. Additionally, the MTC group was older and included a patient with prior cervical spine surgery, while the CC group had none, which may have confounded the results. Finally, different lesioning protocols were used with each cannula type, with the CC group receiving 4 lesions per medial branch nerve (6 for the third occipital nerve [TON]) compared to 2 lesions per nerve (3 for the TON) in the MTC group. These methodological limitations, along with the small sample size, limit the generalizability and interpretation of the findings. Interestingly, real-world prospective cohort studies have reported 3-month response rates of up to 65.4 % in patients treated with the multi-tined RFA electrode technique [23,24]. While the aforementioned studies provide valuable insights, there remains a need for further investigation to better understand the long-term effectiveness and potential complications of CMBRFA using the multi-tined RFA electrodes. While not reported in the literature to our knowledge, there is possibility of trespass of the carotid artery or jugular vein in patients with aberrant anatomy in which these vessels traverse more posteriorly and superficial to the lateral aspect of the cervical articular pillar, reinforcing the need to review advanced axial imaging for procedural planning. Additionally, patient selection criteria, such as specific pain characteristics and underlying pathology, could play a role in determining the success of this technique. It is possible that certain subgroups may benefit more from this approach compared to others, which could explain the variability in outcomes seen across studies. Future additional randomized comparative studies of conventional and other CMBRFA techniques will be useful in guiding treatment decisions for patients with chronic cervical facet joint pain [25]. Moreover, exploring the role of adjunct therapies, such as physical therapy or pharmacological interventions, in combination with CMBRFA may provide a more comprehensive understanding of the optimal management strategies for patients with chronic neck pain. Limitations This study has several limitations that must be acknowledged. One limitation is the cross-sectional rather than prospective design. Additionally, the lack of a control or comparison group restricts our ability to directly compare the outcomes of CMBRFA with other RFA techniques. Variability in patient characteristics, including comorbidities, and the inability to systematically capture all concurrent or subsequent treatments before and after CMBRFA may introduce confounding factors that could influence the results. Additionally, our definition of baseline pain as the highest NRS score recorded within two months prior to the procedure reflects how pain scores are documented in our electronic medical record but may not represent patients’ typical pre-procedure pain levels, potentially introducing bias. We also experienced a substantial loss to follow-up, with approximately half of treated patients not completing outcome surveys, introducing the possibility of responder bias as patients with favorable outcomes may be more likely to participate. Furthermore, variability in follow-up timing and the absence of serial covariate collection limit the ability to draw causal inferences from our regression analyses, and these findings should be considered exploratory. Despite these limitations, our study offers valuable insights into the potential effectiveness of CMBRFA with the three-tined RFA electrode for the treatment of neck pain in patients with confirmed chronic cervical facet joint pain based on dual-concordant medial branch blocks. Conclusion In individuals with chronic cervical facet-mediated neck pain, clinically meaningful improvements in pain were observed at an average follow-up of approximately 16 months after CMBRFA with a three-tined RFA electrode. At a mean follow-up time of 16.3 ± 7.0 months, 51.5 % and 77.3 % of study participants reported ≥50 % and ≥2-point NRS reductions, respectively, while 63.6 % reported PGIC scores consistent with “much improved” or “very much improved”. Additionally, a significant reduction in opioid analgesic use was observed at follow-up compared to baseline. This study contributes to the evidence base suggesting that CMBRFA performed with more recently developed multi-tined electrodes benefits carefully selected patients with cervical facet joint pain. Funding This work was supported by investigator-initiated grant funding from 10.13039/100032805Avanos Medical. Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Zachary L. McCormick reports financial support was provided by Avanos Medical Inc. Zachary L. McCormick reports a relationship with International Pain & Spine Intervention Society that includes: board membership. Zachary L. McCormick reports a relationship with Avanos Medical Inc that includes: funding grants. Zachary L. McCormick reports a relationship with Boston Scientific Corporation that includes: funding grants. Zachary L. McCormick reports a relationship with Presidio Medical that includes: funding grants. Zachary L. McCormick reports a relationship with Saol Therapeutics that includes: funding grants. Zachary L. McCormick reports a relationship with Spine Biopharma that includes: funding grants. Zachary L. McCormick reports a relationship with SPR Therapeutics Inc that includes: funding grants. Zachary L. McCormick reports a relationship with Stratus Medical that includes: funding grants. Zachary L. McCormick reports a relationship with Avanos Medical that includes: consulting or advisory. Zachary L. McCormick reports a relationship with Saol Therapeutics that includes: consulting or advisory. Zachary L. McCormick reports a relationship with Stryker that includes: consulting or advisory. Zachary L. McCormick reports a relationship with OrthoSon that includes: consulting or advisory. Taylor Burnham reports a relationship with Diros Technology Inc that includes: funding grants. Taylor Burnham reports a relationship with Avanos Medical Inc that includes: consulting or advisory. Aaron Conger reports a relationship with Stratus Medical that includes: funding grants. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.