Interventional Management for Neck Pain Following Cervical Spinal Fusion: A Scoping Review

Abstract

Purpose

A subset of patients develop chronic neck pain after cervical spinal fusion without identifiable or correctable surgical pathology. Interventional procedures have been proposed, but evidence specific to interventional management of neck pain following cervical fusion has not been comprehensively synthesized. This scoping review aimed to map the evidence on interventional procedures (Concept) for persistent neck pain in adults after cervical spinal fusion (Population) in clinical settings (Context) and identify gaps in the literature.

Methods

A scoping review of MEDLINE, Embase, and CINAHL was conducted to March 2026 per PRISMA-ScR guidelines. The Population, Concept, and Context framework was used to determine inclusion criteria. Risk of bias was assessed using the Cochrane Risk of Bias 2.0 tool, Newcastle–Ottawa Scale and Moga checklist. Findings were synthesized qualitatively and organized by intervention.

Results

Thirteen studies met our inclusion criteria: 12 therapeutic studies (N = 532) across three intervention categories, and one prevalence study. Intervention categories included cervical epidural injection (CEI; one randomized controlled trial (RCT) with no placebo or sham, N = 116), facet-mediated interventions (FMIs; three studies, N = 219), and spinal cord stimulation (SCS; eight studies, N = 197). The CEI RCT reported 53–54% pain reduction, with responder rates at 69% and 71% at 24 months. FMIs showed 53% and 59% responder rates in two studies reporting outcomes in mixed populations. Reported SCS outcomes ranged from 25.7–82.5% across poorly controlled studies. Seven studies enrolled mixed cohorts with either confirmed post-fusion or cervical surgery patients, only providing indirect evidence. GRADE assessment was Moderate for CEI, Very Low for FMIs, and Very Low for SCS.

Conclusion

The evidence was predominantly of very low certainty. An exploratory evidence-informed clinical framework was developed. A lack of evidence was identified for posterior cervical fusion patients and stratified outcome reporting for post-fusion subgroups.

Introduction

Cervical spinal fusion is among the most commonly performed spinal procedures worldwide. There are over 100,000 cases per year in the United States alone.¹ Although procedures are generally effective at treating radiculopathy or myelopathy, patients may develop persistent or new-onset neck pain post-operatively,^2,3 with up to 28% reporting chronic neck pain at 24 months.⁴ With rising procedural volumes, neck pain after cervical fusion is becoming a growing clinical concern for patients, surgeons, and interventionists.^5,6

The etiology of post-operative neck pain differs from that of chronic neck pain and is likely multifactorial and warrants thorough diagnostic workup. Surgical causes must be excluded, which can include adjacent segment disease (eg instability, fracture, disc herniation, or stenosis above or below the fused construct), pseudarthrosis, and hardware failure.^5,6 Other contributors may be considered, including myofascial pain, neuropathic and radiating pain, and facet arthrosis from adjacent segment wear that has not yet progressed to the threshold for revision surgery.^7,8

Comprehensive management of chronic post-surgical neck pain requires a multidisciplinary team and multimodal strategy (eg pharmacotherapy, appropriate physical therapy). Psychological factors must be assessed and addressed, as previous studies suggested even fear of movement may limit balance in chronic non-specific neck pain.⁹ Back school programs and targeted resistance band exercises have also been suggested, with efficacy in non-specific neck pain, and improvements in disability and quality of life.^10–12 When conservative measures fail, and revision surgery is not indicated, interventional options can be considered for pain management. Pain phenotypes become essential to determine which intervention is indicated for specific clinical presentations. Posterior axial pain with occiput or shoulder girdle referral, without radiation or neurological deficit, is linked to zygapophysial (facet) joint pain. Facet-mediated axial neck pain can be diagnostically evaluated using medial branch blocks (MBBs) and treated using facet-joint injections with or without radiofrequency neurotomy (RFN).¹³ Cervical epidural injections (CEIs) can be considered for patients presenting with radicular or nerve root-mediated patterns. The pain may be accompanied by sensory or motor findings, and reflects root compression or irritation, most often from adjacent-segment disc herniation or foraminal stenosis.^7,8,14 Spinal cord stimulation (SCS) can be indicated for patients presenting with refractory neuropathic pain (ie often termed failed neck surgery syndrome (FNSS)), and persistent upper-limb-dominant neuropathic pain with or without an axial component.¹⁵ The evidence supporting these interventions has been established in the context of general chronic pain management, while post-fusion studies focus on post-lumbar or mixed-level spinal fusion, or do not stratify the subgroup of post-cervical spinal fusion patients.^14–18 Anterior and posterior surgical approaches also affect different anatomical structures, with posterior procedures potentially disrupting the cervical extensor musculature and posterior ligamentous attachments.^2,3 These post-operative changes may produce a different combination of pain generators from those encountered in non-surgical chronic neck pain. The evidence in the post-cervical fusion population therefore remains scattered across small, non-specific, and heterogeneous studies.^14–18

This review aims to identify, categorize, and synthesize the available evidence for interventional treatments of neck pain following cervical spinal fusion, to identify potential gaps in the literature, and propose an exploratory evidence-informed framework to assist surgeons and pain physicians in their treatment decisions.

Methods

Study Design and Reporting Framework

This scoping review was conducted in accordance with the Joanna Briggs Institute (JBI) methodology and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR).^19–21 The completed PRISMA-ScR checklist can be found in Supplementary Table 1. A review protocol was not registered for this study. The scoping review methodology was selected in this case owing to high levels of heterogeneity in intervention types, study designs, outcome measures, and patient populations, which preclude accurate quantitative synthesis.²⁰

Search Strategy

A librarian-assisted search of multiple databases, including MEDLINE (Ovid), Embase, and CINAHL was developed using standard protocols with MeSH and keyword terms covering three domains. The databases were searched for terms related to cervical fusion, cervical arthrodesis, anterior cervical discectomy and fusion, posterior cervical fusion, and cervical laminectomy with fusion. The clinical problem included neck pain, persistent pain, failed neck surgery, adjacent segments, and post-surgical pain. Interventions included SCS, MBBs, facet-joint injections (FJI), RFN, CEI, neuromodulation, nerve block, and regional anesthesia. The searches were limited to English-language publications with no date restrictions. The reference lists of the included studies and relevant systematic reviews were screened to identify additional eligible records. The search was conducted up to March 2026. The search strategy used for all databases can be found in Supplementary Table 2.

Eligibility Criteria

The Population, Concept, and Context (PCC) framework was used to determine the eligibility criteria appropriate for this scoping review.²² Population included adult patients (≥18 years) with a cervical spinal fusion from any approach (ie anterior, posterior or combined) who experienced persistent post-operative neck pain or new-onset neck pain. Fusions were included for any degenerative indication (eg myelopathy, radiculopathy, instability). However, non-degenerative populations such as those with traumatic fractures, neoplastic conditions, and cervical deformities were excluded. The concept included interventional procedures geared toward treating neck pain, including SCS (tonic, high-frequency, or burst), MBBs and RFN for facet-mediated pain, and CEIs. Procedures that did not target the cervical spine, including peripheral nerve stimulation, botulinum toxin injections, trigger point injections, acupuncture, and chiropractic manipulation were excluded. Studies were required to include the population or subpopulation with persistent post-operative or new-onset neck pain following cervical spinal fusion. All study designs were eligible for inclusion. Studies on perioperative or acute post-operative analgesia were excluded. Reviews, letters to the editor, abstracts, and other grey literature were excluded. The context included any clinical setting, with no geographic restriction.

Study Selection

All retrieved records were managed using Covidence (Veritas Health Innovation) for systematic screening. Title and abstract reviews were completed by two reviewers independently and in duplicate. A full-text review was also performed in duplicate for all potentially eligible studies. Disagreements at each stage were resolved through a consensus discussion or by a third reviewer. Agreement between reviewers for title/abstract screening was quantified using Cohen’s kappa statistic and was reported.

Data Extraction

Two reviewers extracted data independently and in duplicate using standardized forms, in line with the PRISMA guidelines.²¹ Extracted variables included demographic data, including year, authors, journal, country, and study design. Patient and surgical data were also retrieved, including the sample size, pathology, surgical approach, fusion levels, time from surgery to intervention, and comparators, if present. Outcomes were reported separately, including pain intensity, as reported by the Numerical Rating Scale (NRS) and Visual Analog Scale (VAS). Functional scores were included when present, including the Oswestry Disability Index (ODI) and Pain Disability Index (PDI), along with adverse events and responder rates.

Risk of Bias and Study Quality Assessment

The risk of bias was assessed independently by two reviewers using study-appropriate tools. Disagreements were resolved by consensus among the reviewers or by third-reviewer adjudication. Randomized Controlled Trials (RCTs) were appraised with the Cochrane Risk of Bias 2 tool (RoB 2).²³ Non-randomized comparative and cohort studies were evaluated using the Newcastle–Ottawa Scale (NOS), whereby ≥7/9 was defined as high quality.²⁴ Case series and case reports were appraised with the Moga checklist, a validated tool for single-arm observational studies.²⁵ Each study was assigned a level of evidence using the Oxford Centre for Evidence-Based Medicine (OCEBM) framework (I–V).²⁶ The overall certainty of evidence was evaluated using GRADE (Very Low, Low, Moderate, or High), accounting for risk of bias, imprecision, inconsistency, indirectness, and publication bias.²⁷ RoB 2, NOS, and the Moga checklist appraise the internal validity and reporting quality of individual studies. The OCEBM framework classifies each study according to study design hierarchy, and GRADE evaluates the certainty of evidence for an outcome across a body of studies. Accordingly, certainty of evidence was rated with GRADE at the level of each intervention category for the outcome of pain relief.

Qualitative Synthesis

Studies were grouped by intervention, including CEIs, Facet-Mediated Interventions (FMIs) (ie MBBs, FJIs, and RFN), and SCS (subdivided by modality: tonic, 10 kHz high-frequency, and burst). The findings within each group were summarized with particular attention paid to effect sizes, responder rates, study quality, and consistency across studies. When available, surgical approaches were also reported, with attention to subgroup presence across all included studies. Studies that exclusively included post-cervical fusion patients were classified as direct post-fusion evidence. Studies that included only a subgroup of post-fusion patients and studies reporting outcomes for mixed populations were classified as indirect.

Evidence-Informed Clinical Framework

The synthesized evidence was organized into an evidence-informed clinical framework. The framework is exploratory and is not a validated clinical guideline or treatment algorithm. It includes components from synthesized evidence, clinical expertise and opinion of senior authors, and available literature. For each intervention category, the framework integrated the reported effect sizes, responder rates, study quality (NOS or Moga checklist), OCEBM level of evidence, and GRADE certainty. Considerations were ordered first by pain phenotype, then by increasing invasiveness, and then by the certainty and directness of the supporting evidence. Because most included studies were uncontrolled, reported improvements at or above the minimal clinically important difference (MCID) threshold were treated as within-group change rather than as evidence of comparative effectiveness. The workup for surgical pathology was based on the clinical experience of senior authors practicing as spinal surgeons. Clinically meaningful improvement was interpreted against established thresholds for pain intensity only. A reduction of ≥2 points on an 11-point NRS, or ≥30% from baseline, was taken as the minimal clinically important difference, with ≥50% reduction regarded as substantial improvement.²⁸ MCID estimates for the disability instruments reported across the included studies (Neck Disability Index (NDI), PDI, ODI) are less consistently established, so disability outcomes were summarized descriptively and were not assessed against a fixed threshold.

Results

Search Results and Study Selection

A total of 2657 studies were retrieved across all databases (ie MEDLINE (Ovid), Embase, and CINAHL). After deduplication (390) and screening (2267), 2,200 studies were excluded. A total of 66 full-text reviews were performed, yielding 13 studies for inclusion (Figure 1). The full study demographics are provided in Supplementary Table 3.

Figure 1.

PRISMA study selection flow diagram. Flow of study inclusion per PRISMA-ScR guidelines. Across CINAHL, MEDLINE (Ovid), and EMBASE, 2,657 studies were identified. A total of 390 duplicate studies were removed, 2,267 studies received title/abstract screening, and 2,200 studies were excluded. Of the 67 studies sought for retrieval, one was not obtained, 53 were excluded, and 13 met the inclusion criteria and were included in the qualitative analysis.

Abbreviation: PRISMA-ScR, Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews.

Across all title/abstract records, inter-rater reliability was high: raw agreement was 99% and Cohen’s κ was 0.55, indicating moderate chance-corrected agreement.²⁹ Thirteen studies met the inclusion criteria: 12 therapeutic studies (N = 532, 2007–2022),^30–41 and one prevalence study (Manchikanti et al,⁴² 2008; 45 of 251 patients post-surgical). The 12 therapeutic studies examined three separate intervention categories (Table 1): CEI (1 study, N = 116),³⁰ FMIs, including MBBs and RFN (3 studies, N = 219),^31–33 and SCS (8 studies, N = 197).^34–41 Designs included RCTs, prospective and retrospective cohorts, and case series, with most being observational without control groups (Table 1).

Table 1.

Study Design Characteristics of Included Studies

Author	Year	Study Design	Control/Comparator	Blinding	Post-Fusion Specific?	Approach	Fusion %	Follow-Up Duration	Sample Retention
Cervical Epidural Injections
Manchikanti et al30	2018	RCT	LA alone vs LA + steroid	Double-blind	Subgroup	~90% anterior	90%	24 mo	96/116 (83%)
FMIs: Radiofrequency Neurotomy/Medial Branch Block
Klessinger et al31	2010	Retrospective cohort (RFN)	None	None	Subgroup	Anterior	48%a	Mean 15 mo (1–40)	32/32 (100%)
Klessinger et al32	2010	Retrospective cohort (MBB)	None	None	Subgroup	Anterior	59%a	Mean 20.4 mo	87/104 (84%)
Burnham et al33	2022	Cross-sectional cohort	80–99% vs 100% MBB paradigm	None	Subgroup	NR	7%	≥6 mo	100/195 (51%)
Manchikanti et al42	2008	Retrospective comparative	Non-postsurgical	Single-blind	Subgroup	78% Posterior	18%	NR	251/251 (100%)
Spinal Cord Stimulation
Vallejo et al34	2007	Retrospective case series	None	None	Yes	Anterior	100%	1–9 mo	5/5 (100%)
Chivukula et al35	2013	Retrospective case series	None	None	Yes	Mixed	100%	Mean 2.0 yr	6/6 (100%)
Chivukula et al36	2014	Retrospective cohort	None	None	Subgroup	Mixed	NR	Mean 4.2 yr	100/121 (83%)
Haider et al37	2017	Prospective observational	None	None	Subgroup	NR	NR	12 mo	24/30 (80%)
Hunter et al38	2018	Prospective registry	None	None	Yes	Mixed	100%	12 mo	15/15 (100%)
Verrills et al39	2020	Prospective multicenter	None	None	Subgroup	Anterior	NR	12 mo	31/38 (82%)
Grider et al40	2020	Prospective pilot	None	None	Yes	NR	100%	12 mo	15/23 (65%)
Dahbour et al41	2021	Case report	None	None	Yes	ACDF + posterior	100%	4 days (trial)	1/1 (100%)

Notes: Studies are grouped by intervention categories. The table presents each study’s design, including presence of control groups, blinding, specificity to post-cervical fusion, surgical approach, follow-up duration and sample retention. ^aThese percentages represent the ratio out of the total patients screened. For Klessinger et al³¹ (RFN) it was 48% and 59% for Klessinger et al.³²

Abbreviations: Mo, Months; Yr, Years; ACDF, anterior cervical discectomy and fusion; LA, local anesthetic; FMIs, facet-mediated interventions; MBB, medial branch block; NR, not reported; RCT, randomized controlled trial; RFN, radiofrequency neurotomy; FNSS, Failed neck surgery syndrome.

Of the 13 studies extracted, seven therapeutic studies did not have an exclusive post-cervical fusion population. Four included a subgroup with confirmed fusion and three a subgroup with confirmed prior cervical surgery, and none reported those subgroups’ outcomes separately. Manchikanti et al³⁰ included 105/116 confirmed anterior post-cervical fusion patients, while the remaining patients had either confirmed posterior cervical fusion or laminectomy. Klessinger et al³¹ (RFN) and Klessinger et al³² (MBBs) both only included a subgroup of fusions among the patients screened, 117/242 and 184/312 respectively. Burnham et al³³ reported 7/100 confirmed post-cervical fusion patients, with no stratification of results for post-surgical patients. Chivukula et al³⁶ reported 23/100 confirmed FNSS patients. Haider et al³⁷ and Verrills et al³⁹ had subgroups of 3/24 and 11/31 confirmed post-cervical surgery patients, respectively. In the case of Verrills et al,³⁹ their inclusion criteria confirmed anterior cervical fusion patients were eligible but excluded any patients who had undergone posterior cervical surgery. The three studies were still included because post-fusion patients could have been enrolled, and to highlight the sparsity of available evidence for this population. In each of these seven studies, the reported outcomes were derived from the full mixed cohort rather than from the post-fusion subgroup alone. The prevalence study by Manchikanti et al⁴² had the only mixed cohort that had stratified results for post-surgical patients (45/251 confirmed post-fusion). These seven studies therefore provide indirect rather than direct evidence for the post-cervical fusion population, and their application to the post-cervical fusion population remains limited. It is noteworthy that none of the therapeutic studies had an exclusive posterior cervical fusion patient population (ie either not reported, anterior, or mixed population) (see Table 1).^30–41 Apart from Manchikanti et al³⁰ who used an active control group, and Burnham et al³³ who included two non-randomized groups selected by different diagnostic block paradigms with the same treatment, no form of control or comparator was included in the therapeutic studies.

Cervical Epidural Injections

Manchikanti et al³⁰ (2018) performed a randomized, double-blind controlled trial with 134 patients who were screened for an ultimate group size of 58 (N = 116), comprising 105 patients with anterior cervical fusion (~90%). The rest had confirmed posterior cervical fusion or laminectomy. The patients were randomized to either local anesthetic (LA) alone or LA plus betamethasone, with injections repeated as needed over 24 months.³⁰

Both study groups showed important outcomes: NRS decreased from (Group I) 8.0 ± 0.98 at baseline to 3.7 ± 1.58 and (Group II) 8.0 ± 0.93 at baseline to 3.8 ± 1.57 at 24 months. They reported no significant difference in outcomes between groups. This represents ~53–54% pain reduction across the two groups.³⁰ Responder rates, defined as ≥50% improvement in both pain relief and functional status, were 69% (Group I) and 71% (Group II). NDI scores improved from 30.4 ± 4.4 to 14.7 ± 6.29 (Group I) and 30.0 ± 5.5 to 14.9 ± 6.64 (Group II).³⁰

OCEBM Levels of Evidence graded the study as Level II. Certainty of evidence was rated Moderate using GRADE, with randomized evidence downgraded one level for serious risk of bias arising from concerns about deviations from intended interventions, outcome measurement, and incomplete follow-up.^26,27 It was not further downgraded for indirectness because a vast majority of patients had confirmed cervical fusion (~90%). Because both groups received epidural injections, the assessment does not establish CEI effectiveness compared with sham or no intervention. Patients reporting significant pain relief in each group (Group I: n = 53, Group II: n = 47) received an average of 5.9 ± 2.5 and 5.6 ± 2.4 injections over a 24-month period, respectively.³⁰

Facet-Mediated Interventions (Medial Branch Block and Radiofrequency Neurotomy)

Three studies evaluated therapeutic facet-mediated interventions for chronic neck pain after cervical fusion. Klessinger et al³¹ (2010, RFN) and Klessinger et al³² (2010, MBB) exclusively enrolled patients with persistent neck pain following anterior cervical surgery, not all of whom had undergone fusion. Klessinger et al³¹ (2010, RFN) screened 242 confirmed anterior cervical operations, of which 117 had confirmed cervical fusion. They identified 76/242 patients with persistent pain, and 67/242 had a positive response initially after MBB treatment. Later, 32/67 patients experienced recurrent pain. They were tested again for cervical zygapophysial joint pain using MBBs and then underwent cervical RFN. At 15-month follow-up, approximately 59% had at least 50% pain relief and 25% had complete relief.³¹ Klessinger et al³² (2010, MBB) screened 312 patients, with 184 confirmed post-cervical fusion patients. A total of 104/312 patients reported persistent neck pain. Of the 104 patients with persistent pain, 87 were included in the analysis, representing 87/312 of the original screened cohort. In this case Klessinger et al³² defined a positive treatment response as at least 80% reduction in pain or sufficient patient satisfaction.³² A total of 46/87 patients (~53%) sustained relief from MBBs using triamcinolone and bupivacaine at a mean follow-up of 20.4 months.³² Approximately 32% of patients that were initially treated successfully experienced recurrence of pain, and 13% did not respond to treatment.³² Klessinger et al³² reported single-level cervical surgery in 64.37% of patients and two-level cervical surgery in 35.63%. Burnham et al³³ examined facet-mediated interventions using medial branch radiofrequency ablation in 100 patients with confirmed cervical facet pain, with only 7 confirmed post-cervical spinal fusion patients. Approximately 48–52% of the patient cohort achieved ≥50% NRS reduction and 67–74% reported a Patient Global Impression of Change (PGIC) of ≥6. Burnham et al³³ reported these outcomes over a follow-up period of at least 6 months without presenting outcomes separately for those with confirmed cervical spinal fusion. The outcomes reported in that study therefore reflect the full mixed cohort and constitute indirect evidence in relation to the post-cervical fusion population.

The additional prevalence study presented by Manchikanti et al⁴² enrolled 251 patients, of whom 45 were confirmed post-cervical fusion patients. Of the 45 patients, 35 had a previous posterior fusion and 10 had previous anterior fusion.⁴² The prevalence of facet-mediated pain was 36%, with false-positive rates as high as 50%.⁴²

Certainty of evidence for pain relief across this body of three studies was rated Very Low using GRADE, reflecting uncontrolled retrospective designs, imprecision, and indirectness in all three studies. The OCEBM Levels of Evidence were IV for both Klessinger et al^31,32 studies, and III for the Burnham study.³³

Spinal Cord Stimulation

SCS following cervical fusion was evaluated in eight studies, predominantly in patients with upper-extremity radicular pain.^34–41 Three of these studies (Chivukula et al,³⁶ Haider et al,³⁷ and Verrills et al³⁹) enrolled mixed cohorts in which post-cervical surgery patients formed a subgroup without separately reporting outcomes and therefore cannot be considered direct post-fusion evidence for the investigated procedures. Three stimulation techniques were represented: tonic/conventional (five studies, plus one case report),^34–38,41 10 kHz high-frequency (one study),³⁹ and burst (one study).⁴⁰ A total of 197 patients were analyzed, with individual sample sizes ranging from 1 to 100 patients.^34–41

Five studies examined tonic SCS. Vallejo et al³⁴ enrolled 5 patients after anterior cervical discectomy and fusion (ACDF) and reported 4/5 patients had significant pain relief (>50%, mean 82.5% ± 6%, range 50–90%). Chivukula et al³⁵ (2013) reported VAS reduction from 9.3 ± 0.8 to 4.2 ± 0.8 (55 ± 7.8% pain reduction) in 6 patients over a mean of 2.0 ± 0.8 years. Chivukula et al³⁶ (2014) reported VAS reduction from 8.8 ± 1.0 to 3.8 ± 0.7 (mean 56.6 ± 10.7% reduction) over a mean of 4.2 years in a mixed cohort of 100 patients, of whom 23 were FNSS patients. Haider et al³⁷ reported a 25.7% NRS reduction (NRS 7.0 ± 0.6 to 5.2 ± 1.3) in 24 patients at 12 months, with a mixed cohort with only 13% FNSS patients. Hunter et al³⁸ reported PDI reduction from 51.21 ± 11.03 to 23.70 ± 16.88 at 12 months in 15 post-cervical fusion patients using tonic programming (≤1500 Hz). They also reported 71.9% patient-reported pain relief at 12 months. Verrills et al³⁹ conducted a study on 10 kHz high-frequency SCS which separately reported pain scores for the neck and arm for their mixed cohort. They reported neck VAS reduced from 8.1 ± 0.2 at baseline to 2.2 ± 0.5 (74.2% pain relief) with an 85.2% responder rate at 12 months.³⁹ PDI reduced from 42.6 ± 2.6 at baseline to 21.2 ± 3.4 at 12 months.³⁹ Upper limb VAS decreased from 7.3 ± 0.3 to 2.5 ± 0.5 (61.8% pain relief) with a 76.5% responder rate in 13 available patients at 12 months.³⁹ For burst SCS, Grider et al⁴⁰ reported VAS reduction from 8.13 ± 1.0 to 3.85 ± 1.1 (p<0.05) at 12 months (52.6% reduction) in 15 FNSS patients. Dahbour et al⁴¹ reported an 80% pain reduction in a 48-year-old female in a 4-day trial of tonic stimulation. The patient reported a significant improvement in her quality of life and elimination of breakthrough oxycodone in her medication regimen.

An important methodological consideration is that most SCS studies reported combined pain scores without separately measuring midline neck pain and upper-extremity radicular pain. Chivukula et al³⁶ (2014) noted that pain relief was greater along the extremities than axially. All SCS studies reported degrees of pain relief that exceeded the MCID, except Haider et al³⁷ in their tonic study. Because these were uncontrolled cohorts and case series, such values represent within-group change over time and cannot be attributed to stimulation itself.

Taken together, certainty of evidence for pain relief with SCS in this population was rated Very Low, downgraded for risk of bias arising from uncontrolled designs, imprecision related to small samples, and indirectness in the three mixed-cohort studies. All eight SCS studies had an OCEBM Level of Evidence of IV.^34–41

General Observations

Surgical Approach as a Moderating Factor

An important detail is the focus of the evidence on the post-anterior cervical fusion population. Apart from Vallejo et al,³⁴ Klessinger et al³¹ (RFN), Klessinger et al³² (MBB), and Manchikanti et al³⁰ (RCT), who predominantly enrolled an anterior post-fusion population, most therapeutic studies did not report the exact surgical approach their patient cohort underwent. Some studies defined their population as post-cervical surgery rather than confirmed cervical fusion. Chivukula et al,³⁶ Haider et al,³⁷ and Verrills et al³⁹ confirmed prior cervical surgery without specifying fusion. Their studies were retained because prior fusion was not ruled out, though this represents a further source of indirectness.

Statistical Power and Outcome Heterogeneity

Studies used different categorical relief thresholds (ie VAS, NRS, percentage reduction in pain rating (PRPR)) and disability indices (ie NDI, ODI, PDI), in different reporting formats and at different follow-up timepoints.^30–41 This heterogeneity in outcome definitions prevents formal quantitative pooling across intervention types. Most extracted studies were underpowered. The median sample size was 27.5 patients per study (range, 1–116), with SCS studies having particularly small samples (median, n = 15).

Methodological Quality of Included Studies

Randomized Controlled Trials

The single RCT performed by Manchikanti et al³⁰ was assessed using the RoB 2 tool (Table 2). Overall, the study was judged to have some concerns, mainly due to deviations from the intended interventions and measurement of outcomes. There were also concerns about critically missing data, as roughly 17% of participants did not complete the entire two-year follow-up.³⁰

Table 2.

Risk of Bias Assessment for Included Randomized Controlled Trials

Study	D1	D2	D3	D4	D5	Overall	Key Notes
Manchikanti et al30	Low	Some concerns	Some concerns	Some concerns	Low	Some concerns	Active-controlled ~17% missing at 24 mo

Notes: RoB 2 domains: D1 = Randomization process; D2 = Deviations from intended interventions; D3 = Missing outcome data; D4 = Measurement of the outcome; D5 = Selection of the reported result.

Observational Studies

Eight studies were scored using the NOS (Table 3 and Figure 2). Two studies were classified as good quality, reaching seven stars. Burnham et al³³ (7/9) had a formal comparison group using validated instruments such as the NRS scoring system. Manchikanti et al⁴² (prevalence study) (7/9) conducted a comparative study with strict response criteria. Grider et al⁴⁰ and Verrills et al³⁹ both scored six out of nine with adequate study design, compliance with trial registration, and reasonable follow-up retention. The remaining four studies all scored five out of nine, mainly due to the lack of control groups, small cohorts, lack of placebo, and short follow-up intervals.

Table 3.

Newcastle-Ottawa Scale Assessment for Included Observational Studies

Study	Design	Selection (max 4)	Comparability (max 2)	Outcome (max 3)	Total	Quality	Key Limitations
Burnham et al33	Cross-sectional cohort	4★	1★	2★	7/9	Good	Non-randomized groups; only 100/195 eligible patients analyzed; small post-fusion subgroup (7%), not reported separately; external funding.
Manchikanti et al42 (2008)	Retrospective comparative	3★	1★	3★	7/9	Good	No treatment outcomes; single-physician practice; small post-surgical subgroup (45/251); external funding.
Grider et al40	Prospective pilot	3★	0★	3★	6/9	Fair	Single-arm pilot; small sample (15 implanted/23 trialed); industry funding.
Verrills et al39	Prospective multicenter	3★	0★	3★	6/9	Fair	Single-arm; fusion not confirmed (PCS 35.5%); industry funding.
Haider et al37	Prospective observational	3★	0★	2★	5/9	Fair	Single-arm, single-center; only 13% prior cervical surgery, fusion not confirmed; no validated disability scale; industry funding.
Chivukula et al36 (2014)	Retrospective cohort	3★	0★	2★	5/9	Fair	Retrospective, single-arm; PCS subgroup (23/100) not reported separately; funding not reported.
Klessinger et al32 (MBB)	Retrospective cohort	3★	0★	2★	5/9	Fair	Single-arm; categorical success outcome (≥80% relief or patient satisfaction); fusion subgroup not reported separately.
Klessinger et al31 (RFN)	Retrospective cohort	3★	0★	2★	5/9	Fair	Single-arm; small sample (n = 32); categorical ≥50% relief outcome only; fusion subgroup not reported separately.

Notes: Quality rating: Good (≥7 stars), Fair (5–6 stars), Poor (<5 stars). Single-arm studies scored 0 on Selection item 2 (non-exposed cohort) and Comparability. The single case report and three case series were assessed using the Moga checklist.

Abbreviations: PCS, prior cervical surgery; MBB, Medial Branch Block, RFN, Radiofrequency Neurotomy.

Figure 2.

Quality assessment for Randomized Controlled Trials and Observational Studies. Visual synthesis of methodological quality for RCTs and observational studies. (A) RCTs appraised with the Cochrane Risk of Bias 2 (RoB 2) tool (Table 2); (B) Observational studies appraised with the Newcastle–Ottawa Scale (NOS, Table 3). The single case report and three case series were appraised separately with the Moga checklist. NOS ratings were good (≥7 stars), fair (5–6 stars), and poor (<5 stars).

Abbreviations: RoB 2, Cochrane Risk of Bias 2; NOS, Newcastle-Ottawa Scale.

Case Report and Case Series

The single case report conducted by Dahbour et al⁴¹ met 9/20 items in the Moga checklist criteria.²⁵ No clear inclusion criteria were mentioned, and the reported outcome measure was non-definitive. The follow-up period was extended to only four days, which also affected the overall quality score. Vallejo et al,³⁴ Chivukula et al,³⁵ and Hunter et al³⁸ conducted case series and met 10/20, 12/20, and 13/20 items on the Moga checklist criteria respectively.²⁵

Proposed Framework

An exploratory evidence-informed clinical framework was created based on pain phenotypes, senior author clinical expertise, the quality of the synthesized studies, and reported outcomes (Figure 3).

Figure 3.

Exploratory evidence-informed clinical framework proposed by the authors. The evidence-informed clinical framework proposed by the authors labels each step by its source: senior authors’ clinical opinion, external literature, or synthesized evidence. Step 1: Rule out surgically correctable pathology (CT, MRI, flexion-extension radiographs, and clinical assessment); this step reflects senior authors’ clinical opinions and established principles of post-surgical spine care. Step 2: Diagnosis of the pain phenotype based on clinical presentation. Step 3: Conservative management trial, guided by clinical presentation and available evidence from the literature. Step 4: Diagnostic facet evaluation with controlled MBBs could be used to rule out false positives, as suggested in the Manchikanti et al⁴² prevalence study. Step 5: If radicular pain with nerve root-mediated symptoms is suspected, CEIs were evaluated in a single active-controlled RCT.³⁰ For axial pain confirmed with MBBs, cervical-facet-joint RFN was evaluated in one retrospective study after anterior cervical surgery.³¹ Step 6: SCS for refractory upper-extremity radicular pain (with or without axial neck pain) is consistent with FNSS. No included study evaluated SCS for isolated axial neck pain, although one study reported neck pain outcomes separately.³⁹ Most of the evidence originates from anterior surgical approach studies. Manchikanti et al,³⁰ Klessinger et al,^31,32 Burnham et al,³³ Chivukula et al,³⁶ Haider et al,³⁷ and Verrills et al³⁹ did not report separate outcomes for post-cervical fusion patients in their mixed cohorts. The framework is intended to support, not replace, individualized clinical judgments.

Abbreviations: ASD, Adjacent Segment Disease; RCD, Rotator Cuff Disease; GHA, Glenohumeral Arthritis; CBT, Cognitive Behavioral Therapy; PF, Post Fusion; MBB, medial branch block; FNSS, failed neck surgery syndrome; CS, Cervical surgery; RR, Responder rate; HF-SCS, High Frequency Spinal Cord Stimulation.

Discussion

Patients who develop chronic neck pain after cervical spinal fusion represent a particularly challenging clinical population to treat. Having already undergone major surgery, they are left with disabling pain and few remaining treatment options, particularly when revision surgery is not indicated. This scoping review maps and synthesizes the available evidence for spine-directed interventional treatments in this population and, where reported, characterizes it by surgical approach. A principal finding is that the available evidence does not currently permit stratification by surgical approach. Across the 13 included studies and 532 patients in 12 therapeutic studies, three intervention categories were identified: CEIs, FMIs, and SCS.^30–41 Seven studies did not report post-cervical fusion outcomes separately (ie Chivukula et al,³⁶ Verrills et al,³⁹ and Haider et al³⁷ only confirmed post-cervical surgery patients), meaning that the total number of patients represents a mixed population as well. Across the included studies, these interventions were used to treat distinct pain phenotypes, including facet-mediated axial neck pain, radicular neuropathic pain, and mixed axial–radicular pain syndromes. Despite potentially encouraging effect sizes, the evidence base was heterogeneous in study design, follow-up duration, outcome measurement, and methodological quality, with predominantly Very Low certainty by GRADE assessment.

The facet-mediated studies reported the proportion of participants reaching a predefined response threshold, whereas most SCS studies reported only mean group change in VAS or NRS, while only Manchikanti et al,³⁰ Vallejo et al,³⁴ and Verrills et al³⁹ reported both. Because a mean group reduction exceeding the MCID can be produced by a minority of large responders, mean change and responder proportion are not interchangeable and were not pooled or compared directly in this synthesis.²⁸ Most of the included studies were uncontrolled. Therefore, a within-group change of this kind does not by itself establish intervention effectiveness. Natural history, regression to the mean, co-interventions, selection effects, and placebo or contextual effects cannot be excluded. CEI was supported by a single randomized, double-blind trial with an active comparator rather than a placebo or sham control. However, the study demonstrated some concerns regarding the risk of bias assessment, particularly related to deviations from the intended interventions and incomplete outcome data. The Moderate GRADE assessment concerned the comparative effect of adding corticosteroid to local anesthetic on pain intensity at 24 months. Both groups received epidural injections. Therefore, this assessment does not establish the effectiveness of CEIs compared with sham or no intervention. FMI studies reported consistent responder rates and used confirmatory diagnostic blocks. SCS studies comprised the largest body of literature but were mostly limited by small samples, absence of blinding or sham controls, and a primary focus on upper-extremity radicular pain rather than isolated axial neck pain. In addition, seven of the included therapeutic studies enrolled mixed populations, three of which only confirmed post-cervical surgery, with none of the studies reporting separate outcomes for the post-fusion subgroup. Therefore, their findings constitute indirect evidence and cannot be attributed specifically to patients with persistent pain following cervical fusion. The majority of evidence was derived from post-anterior cervical fusion populations, and the generalizability to post-posterior fusion patients remains uncertain. Overall, the literature is scattered, and our best impression is that these interventions should be trialed based on patient presentation in combination with conservative care as part of a multidisciplinary approach after correctable surgical pathology has been excluded.

Before any interventional treatment is considered, it is the opinion of the senior authors that a thorough surgical workup must be completed to exclude correctable surgical pathologies. This should include computed tomography (CT) to assess bony fusion and hardware integrity, magnetic resonance imaging (MRI) to evaluate adjacent segment disease or cord compression, and flexion–extension radiographs to assess instability or pseudarthrosis. Clinical correlation with symptom pattern, neurological examination, and timeline of pain onset relative to the index surgery is essential. Revision surgery, when indicated, should take precedence. Rotator cuff disease, adhesive capsulitis, glenohumeral arthritis, and scapular dysfunction should also be ruled out, as these conditions can mimic cervical radiculopathy.^43–47 This recommendation reflects clinical experience and established principles of post-surgical spine care rather than the evidence synthesized in this review.

Based on these findings, we propose an exploratory, evidence-informed, phenotype-oriented clinical framework presented in Figure 3 that summarizes the current literature and may assist clinicians in considering interventional options for this population. Once surgical pathology has been excluded as described above, conservative measures should be trialed before escalation to procedural interventions.^9–12 If facet-mediated pain is suspected, controlled diagnostic blocks should precede therapeutic interventions.⁴² In the synthesized evidence, cervical-facet-joint RFN was used for facetogenic axial pain, diagnostically confirmed using MBBs. It showed a ~59% responder rate in diagnostically confirmed patients.³¹ The Klessinger et al³² MBB study reported 53% relief at 20.4 months. If radicular or nerve root-mediated pain syndromes are suspected, CEIs may be considered, with the single active-controlled RCT reporting ~53–54% pain reduction and 69% (Group I) and 71% (Group II) responder rates sustained at 24 months. However, the absence of a placebo or sham comparator limits what can be inferred about effectiveness. SCS has been studied mainly in patients with refractory neuropathic and/or upper-extremity radicular pain with or without axial neck pain. No included study evaluated SCS for isolated axial neck pain without a radicular or neuropathic component, so its role in that presentation is unsupported by the evidence reviewed here rather than shown to be ineffective. Notably, the majority of evidence supporting this framework is derived from post-anterior fusion populations.

Several critical gaps in the literature have been identified. Among the 12 therapeutic studies, none employed a placebo or sham control. Only Manchikanti et al,³⁰ who used an active comparator, and Burnham et al,³³ who compared two non-randomized groups selected by different diagnostic block paradigms, included a comparator. The remaining studies were single-arm and uncontrolled, limiting the interpretability of their findings. Additionally, Burnham et al,³³ the prevalence study by Manchikanti et al,⁴² and several SCS studies^37–40 have received external funding, which has been suggested to be a potential source of reporting bias.⁴⁸ The current literature shows that patients with neck pain following posterior fusion are substantially underrepresented in chronic pain interventional studies. Eligibility was restricted to English-language publications, and conference abstracts, letters, and other grey literature were excluded. Therefore, potentially relevant evidence published in other languages or outside the peer-reviewed literature may have been missed. In the future, RCT design should become a priority for FMI and SCS studies. Future studies should not only investigate patients after posterior cervical fusion but also report outcomes for post-surgical patients separately and ensure precise stratification of reported outcomes based on the surgical approach.

Conclusion

This study aimed to synthesize the available evidence for interventional treatment in post-cervical fusion patients, identify potential gaps in the stratification of outcomes by post-cervical fusion vs non-surgical population, and present an exploratory evidence-informed clinical framework to support clinical decision making. Thirteen studies met our inclusion criteria, including one prevalence study. A total of 532 patients were analyzed across 12 therapeutic studies (2007–2022).^30–41 Certainty of evidence was Moderate for CEI (one active-controlled RCT without a placebo or sham control), Very Low for FMIs, and Very Low for SCS. The underlying literature was predominantly observational, heterogeneous, and of very low certainty. Current evidence derives predominantly from post-anterior fusion populations and may not generalize to posterior fusion. Seven studies contributed only indirect evidence because their outcomes were reported for mixed cohorts. Therefore, the evidence synthesized for these interventions remains preliminary. Studies focusing on post-cervical fusion populations with adequately powered prospective and comparative designs, with sham or placebo controls, are urgently needed, particularly for SCS and FMIs. This review identified key intervention-specific signals that must be further investigated through formal evaluation and consistent stratification of outcomes by surgical approach. The lack of posterior surgical approach representation and stratification of outcomes by surgical approach is a critical gap that must be resolved to establish a stronger evidence base.

Funding Statement

This research received no specific grants from any funding agency in the public, commercial, or not-for-profit sectors.

Disclosure

Dr Robert Ravinsky reports consulting fees from SI-BONE, outside the submitted work. The other authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.