Relative Efficacy of Cervical Total Disc Arthroplasty Devices and Anterior Cervical Discectomy and Fusion for Cervical Pathology: A Network Meta-Analysis

Abstract

Study Design

Systematic Review and Meta Analysis

Objective

This study sought to compare patient-reported outcomes, success, complications, and radiographic outcomes directly and indirectly between different cervical total disc arthroplasty (TDA) devices and anterior cervical discectomy and fusion (ACDF).

Methods

Patients of prospective randomized controlled trials of 1-level cervical TDA with a minimum of 2 years follow up were identified in the literature. A frequentist network meta-analysis model was used to compare each outcome across the different TDA devices included and ACDF using the mixed effect sizes.

Results

15 studies were included for quantitative analysis, reporting the outcomes of 2643 patients with an average follow-up was 67.3 months (range: 24-120 months), 1417 of whom underwent TDA and 1226 of whom underwent ACDF. Nine TDA devices were compared to ACDF, including the Bryan, Discover, Kineflex, M6, Mobi-C, PCM, Prestige ST, ProDisc-C, and Secure-C cervical prostheses. Several devices outperformed ACDF for certain outcomes, including Visual Analog Scale (VAS) Arm, Physical Component Score of the Short-Form Health Survey (SF PCS), neurological success, satisfaction, index-level secondary surgical interventions (SSI), and adjacent level surgeries. Cumulative ranking of each intervention assessed demonstrated the highest performance with the M6 prosthesis (P = .70), followed by Secure-C (P = .67), PCM (P = .57), Prestige ST (P = .57), ProDisc-C (P = .54), Mobi-C (P = .53), Bryan (P = .49), Kineflex (P = .49), Discover (P = .39), and ACDF (P = .14).

Conclusion

Cervical TDA was found to be superior on most outcomes assessed in the literature of high-quality clinical trials. While most devices demonstrated similar outcomes, certain prostheses such as the M6 were found to outperform others across several outcomes assessed. These findings suggest that the restoration of near-normal cervical kinematics may lead to improved outcomes.

Introduction

Cervical total disc arthroplasty (TDA) has proven to be an efficacious treatment option for patients with radiculopathy and/or myelopathy who have failed conservative management.^1,2 Additionally, it has been associated with lower complication rates as compared to anterior cervical discectomy and fusion (ACDF).^3,4 Given its motion-preserving qualities and preferable complication profile, its use has been increasing over several years in patients meeting criteria for arthroplasty.^5,6

As TDA aims to restore physiologic motion and load sharing of the diseased level, its effectiveness can be reliant on both patient and device-specific factors. As each device has a unique design, they offer distinct mechanical properties that may impact clinical outcomes. ⁷ Randomized controlled trials (RCT) of multiple prostheses have demonstrated significant improvement in patient-reported outcome metrics (PROMs) with long-term follow up.^1,8-16 However, inter-device performance comparisons are currently rare in the literature.

Given the increasing incidence of TDA use in patients with cervical pathology, along with the abundance of devices available on the market, performance analysis between devices has become of increased interest and importance. This study aims to assess the impact of prosthesis design on clinical outcomes relative to ACDF, based on the current body of high-quality RCTs. Improved understanding of device design and performance will empower surgeons to navigate the myriad factors that may currently influence surgical indications and device selection.

Methods

Search Strategy

The methodology of this network meta-analysis adhered strictly to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A search of the PubMed and MEDLINE databases was performed in 2020 to identify all published English language studies that pertained to cervical TDA. The terms utilized for the search were “total disc arthroplasty”, “total disc replacement”, “cervical disc arthroplasty”, and “cervical disc replacement”. The reference lists of all relevant studies were also reviewed to identify any additional studies that may not have been captured by the initial query.

Study Eligibility and Data Extraction

Studies were included in the network meta-analysis if they met the following inclusion criteria: (1) the study design was a RCT comparing cervical TDA to another intervention, (2) included patients who underwent index surgery at a single motion segment, (3) included TDA patients who all underwent surgery with a single device, (4) allocated a minimum of 45 patients to each treatment arm, (5) obtained a minimum average follow-up of 2 years, and (6) reported 1 or more outcome of interest to this network meta-analysis determined a priori. Exclusion criteria were: (1) early results of clinical trials that later reported further follow-up, and (2) studies reporting the same outcomes on the same patient cohort also described elsewhere (ie, results from a single institution participating in a multicenter trial). The same inclusion and exclusion criteria were maintained for all studies and for all interventions assessed.

Following the initial search, the titles and abstracts of all identified studies were reviewed independently by 2 authors for relevancy to the subject of interest. The full texts of potential studies were then screened by the same 2 authors and studies for which a consensus on eligibility was achieved were then included for data extraction. If a consensus was not reached, a third author reviewed the full text to make the final decision on inclusion. Once the included studies were identified, the same 2 authors extracted data on predetermined variables of interest.

Outcomes Assessment

The outcomes of interest were selected based on common variables frequently reported in the device-specific Food and Drug Administration (FDA) Investigational Device Exemption (IDE) trials. PROMs included the Neck Disability Index (NDI), Visual Analog Scale (VAS) for neck and arm pain, and Physical Component Score of the Short-Form Health Survey (SF PCS). Additionally, data on overall procedural success, neurological success, patient satisfaction, postoperative dysphagia, device or procedure-related adverse events, index-level secondary surgical interventions (SSI), and adjacent segment surgeries were collected. Radiographic data included were postoperative segmental range of motion (ROM), cervical (C2-C7) ROM, and heterotopic ossification (HO) in TDA or successful fusion following ACDF. The summary of safety and effectiveness data (SSED) reports for the device FDA IDE trials were also searched to supplement data points missing from the manuscripts.

Statistical Analysis

Unlike a standard pairwise meta-analysis, which compares 2 competing treatments directly compared by the pooled studies, the network meta-analysis statistical method allows for the comparison of multiple treatments if there is a common comparator, which in this case was the control group (ACDF). A frequentist approach to network meta-analysis was performed to directly and indirectly compare the performance of individual TDA devices and ACDF with respect to the previously specified outcome measures. All statistical methods for the network meta-analysis were performed utilizing the netmeta package for R Studio version 2021.09.2 build 382 (R Core Team, Vienna Austria). The effect size (ES) and standard error (SE) of each outcome was computed for each of the included studies. For continuous variables, the ES was the mean difference (MD) between the 2 treatments of interest, while the log odds ratio (OR) was used for categorical data. Network meta-analysis was utilized to compute effect size estimates for treatment comparisons by pooling the effect sizes under a random-effects model. A separate analysis was performed for each outcome assessed, and effect sizes were reported with the corresponding 95% confidence interval (95% CI). For both MD and log OR effect size comparisons, a statistically significant difference was established if the corresponding 95% CI did not include 0. The Q statistic was used to assess the total heterogeneity within the network.

Forest plots were constructed to graphically represent the comparison of TDA devices across all the outcomes assessed with ACDF as the reference intervention. The efficacy of the competing treatments in the network was ranked by computing the P-Score, a representative linear score of the certainty that the given intervention is superior to others in the network. ¹⁷ A P-Score was generated for each outcome assessed, then the average P-Score was calculated to create a cumulative rank of the efficacy of each treatment under investigation in the network while weighting each variable equally. When calculating the average P-Score for ACDF, segmental ROM and the presence of bridging bone were excluded from the equation since, unlike TDA, the goal of the procedure is to create a robust fusion mass, thus obviating ROM at the motion segment.

Risk of Bias Assessment

Risk of bias was assessed in accordance with the Cochrane Back and Neck Group guidelines for included studies Figure 1. ¹⁸ Overall, the RCTs included for quantitative analysis demonstrated a low risk of bias in most criteria analyzed. High-risk for bias was primarily attributed to the lack of blinding of the care provider or the absence of an intention-to-treat analysis.

Figure 1.

Risk of bias assessment for the included studies.

Results

Search Results and Study Characteristics

A total of 2310 studies were identified, which after filtering for duplicates, resulted in a total of 2307 studies for screening. After the title and abstract screen, 186 studies were identified for full-text review. A total of 15 studies satisfied the inclusion criteria and were selected for data extraction.^{1,8-11,13–16,19-24} The search strategy is summarized by Figure 2

Figure 2.

PRISMA flow chart summarizing the search strategy.

The 15 studies included initially enrolled 3952 patients (2061 TDA, 1891 ACDF), and reported the outcomes of 2643 patients (1417 TDA, 1226 ACDF) who adhered to the per-protocol follow-up. The weighted mean average follow-up was 67.3 months (range: 24-120 months). All studies were two-arm RCTs comparing a single TDA device to ACDF. Nine TDA devices were compared to ACDF, including the Bryan, Discover, Kineflex, M6, Mobi-C, PCM, Prestige ST, ProDisc-C, and Secure-C cervical prostheses. Device specifications are described in Table 1. The studies by Lavelle et al. and Loidolt et al. both reported on 10-year outcomes of the FDA IDE trial for the Bryan disc, however, each study reported unique outcomes, and both were deemed appropriate for inclusion.^14,15 Radcliff et al. reported the 7-year outcomes of the Mobi-C IDE trial. ²⁰ Of note, Kim et al. recently published the 10-year results of the same patient population, however, the authors pooled certain outcomes for both 1 and 2-level TDA, therefore the decision was made to include the earlier trial report. ²⁵ Adequate data was available for all predetermined outcomes of interest, except for cervical (C2-C7) ROM, which was not reported by any study included. Study characteristics are described in Table 2 and study inclusion/exclusion criteria for enrollment as well as definitions of success are described by Table 3.

Table 1.

Summary Data of Studies Included for Quantitative Analysis.

Author	FDA IDE	Device	Patients (n)		Age (SD)		Follow-Up (SD)
			TDA	ACDF	TDA	ACDF
Burkus et al 2014	Y	Prestige ST	212 (of 276)	183 (of 265)	43.3 (r: 25-72)	43.9 (r: 22-73)	84
Coric et al 2018	Y	Kineflex	93 (of 136)	83 (of 133)	43.7 (7.8)	43.9 (7.39)	60
Donk et al 2017	N	Bryan	50	47	44.1 (6.4)	43.1 (7.5)	106.8 (22.8)
Hou et al 2016	N	Mobi-C	51 (of 56)	48 (of 51)	46.3 (7.8)	48.5 (8.3)	61 (1.2)
Janssen et al 2015	Y	Prodisc-C	79 (of 103)	79 (of 106)	42.1 (8.42)	43.5 (7.15)	84
Lavelle et al 2019	Y	Bryan	128 (of 242)	104 (of 221)	44.4 (7.9)	44.7 (8.6)	120
Loidolt et al 2021	Y	Bryan	130 (of 242)	104 (of 221)	44.4 (7.9)	44.7 (8.6)	120
MacDowall et al 2019	N	Discover	67 (of 83)	70	46.9 (6.8)	47 (6.9)	66 (r: 57-77)
Phillips et al 2015	Y	PCM	163 (of 218)	130 (of 185)	45.3 (9.0)	43.7 (8.3)	60
Phillips et al 2021	Y	M6	152 (of 160)	164 (of 189)	43.6 (9.1)	44.7 (7.9)	24
Radcliff et al 2017	Y	Mobi-C	131 (of 164)	54 (of 81)	43.3 (9.2)	44.0 (8.2)	84
Rozankovic et al 2017	N	Discover	51 (of 52)	50 (of 53)	41.32 (8.8)	41.94 (9.36)	24
Sundseth et al 2017	N	Discover	60 (of 68)	60 (of 68)	44.7 (7.2)	43.4 (6.8)	24
Vaccaro et al 2018	Y	Secure-C	124 (of 151)	101 (of 140)	43.4 (7.50)	44.4 (7.86)	84
Zhang et al 2012	N	Bryan	56 (of 60)	53 (of 60)	44.8 (5.6)	45.57 (5.83)	24

FDA, Food and Drug Administration; IDE, Investigational Device Exemption; TDA, Total Disc Arthroplasty; ACDF, Anterior Cervical Discectomy and Fusion; n, Number; SD, Standard Deviation, Y, Yes; N, No; r, Range.

Table 2.

Summary of Inclusion and Exclusion Criteria Upheld by All Included Trials and Definitions of Overall and Neurological Success.

Author	Inclusion Criteria	Exclusion Criteria	Overall Success	Neuro Success
Burkus et al 2014	1) Age >18 years	1) Multi- level symptomatic DDD	1) NDI improvement ≥15	Maintenance or improvement in
	2) Single- level symptomatic DDD between C-3 and C-7 and intractable radiculopathy, myelopathy, or both	2) Evidence of cervical instability on dynamic flexion–extension radiographs, sagittal-plane translation of greater than 3.5 mm, or sagittal-plane angulation of greater than 20	2) Neurological status maintenance/improvement	1) Motor function
	3) NDI >30	3) Symptomatic C2-3 or C7–T1 disc disease	3) No loss of disc height due to sub- sidence	2) Sensory function
	4) Neck pain score >20	4) Previous surgery at the involved level	4) No serious implant/procedure related serious adverse event	3) Deep tendon reflexes
	5) Minimum 6-week history of neck and arm pain unresponsive to nonoperative treatments	5) Severe facet joint disease	5) No additional surgical procedure
	6) Surgery was performed within 6 weeks in patients experiencing progressive neurological worsening	6) History of discitis
	7) Radiographic evidence of single-level cervical disc disease and preserved motion at the symptomatic level	7) Osteoporosis
		8) Metastases
		9) Long-term use of steroids or NSAIDs
Coric et al 2018	1) Single-level, cervical disc disease from C-3 to C-7 with radiculopathy	1) Severe facet degeneration	1) 20% improvement in NDI	Maintenance or improvement of neurological status
	2) Failure of 6 months of nonoperative care or progressive symptoms	2) Bridging osteophytes	2) No device failures
	3) Signs of nerve root compression	3) Prior cervical fusion	3) No supplemental index surgery
	4) NDI of 40 or more	4) Severe myelopathy (<3/5 muscle strength)	4) No major device-related adverse event
Donk et al 2017	1) Age between 18 and 55	1) History of cervical spine surgery	NR	NR
	2) Monoradicular signs and/or symptoms due to a herniated cervical intervertebral disk and/or an osteophyte
	3) The radiological findings should be in accordance with the clinical presentation
	4) The involved level should be mobile on dynamic radiographs
Hou et al 2016	1) Age between 21 to 60	1) More than 1 level cervical degeneration	NR	NR
	2) Diagnosis of degenerative cervical spondylosis	2) Other diseases making the patient unsuitable for surgery
	3) Single level cervical disc degeneration	4) Evidence of metal allergy
	4) Failure to respond to at least 3 months of conservative treatment
Janssen et al 2015	1) Age between 18-60	1) More than 1 vertebral level requiring treatment	1) NDI success	Maintenance or improvement in
	2) Cervical disc disease in only 1 vertebral level between C3–C7	2) Marked cervical instability (translation >3 mm and/or more than 11° of rotational difference to that of either adjacent level)	2) Neurological success	1) Motor function
	3) Neck or arm (radicular) pain and/or functional/neurological deficit confirmed by imaging (herniated nucleus pulposus, spondylosis, loss of disc height)	4) Fused level adjacent to the level to be treated	3) Device success	2) Sensory function
	4) Unresponsive to at least 6 weeks nonoperative treatment or progressive symptoms	5) Radiographic confirmation of severe facet joint disease or degeneration	4) Absence of adverse events related to the implant or its implantation	3) Deep tendon reflexes
	5) NDI score ≥15/50 (30%)	6) Known allergy to cobalt, chromium, molybdenum, titanium, or polyethylene
		7) Prior surgery at the level to be treated
		8) Neck or arm pain of unknown etiology
		9) Clinically compromised vertebral bodies at the affected level because of current or past trauma
		10) Active infection—systemic or local
		11) Severe spondylosis at the level to be treated (eg, bridging osteophytes, loss of disc height >50%, absence of motion (<2°)
		12) Metabolic bone disease
		13) Severe diabetes mellitus requiring daily insulin management
		14) Pregnant or interested in becoming pregnant in the next 3 years
		15) Autoimmune disease
		16) Systemic disease including AIDS, HIV, or hepatitis
		17) Osteoporosis
		18) Taking medications or any drug known to potentially interfere with bone/soft-tissue healing (eg, steroids)
		19) Active malignancy
Lavelle et al 2019 and Loidolt et al 2021	1) At least 21-year-old	1) Marked spondylosis	1) ≥ 15-point improvement in their NDI scores	Maintenance or improvement in
	2) Radiculopathy or myelopathy from single-level cervical disc disease secondary to disc herniation	2) Marked reduction or absence of motion or collapse of the intervertebral disc space of greater than 50% of its normal height	2) Maintenance or improvement in their neurologic status	1) Motor function
	3) Unresponsive to at least 6 weeks of nonoperative management, with the exception of cases of myelopathy requiring immediate treatment	3) Facet joint arthrosis	3) No serious adverse events related to the implant or implant/surgical procedure	2) Sensory function
		4) Segmental instability or cervical kyphosis	4) No subsequent surgery or intervention	3) Deep tendon reflexes
		5) Active infection
		6) Metabolic bone disease
		7) Known allergy to titanium, polyurethane, or ethylene oxide residuals
		8) Concomitant conditions requiring steroid treatment
		9) Diabetes mellitus
		10) Extreme obesity
		11) Pregnancy
		12) Inflammatory spondyloarthropathies
		13) Previous cervical spine surgery
MacDowall et al 2019	1) Age 25-60 years	1) Previous cervical spine surgery	NR	NR
	2) Cervical radiculopathy with symptoms of radiating arm pain with duration of at least 3 months	2) Multiple cervical levels requiring treatment
	3) Correlating findings on MRI	3) Severe facet arthropathy
		4) Symptoms or marked radiological signs of myelopathy
		5) Drug abuse, dementia or expected low compliance
		6) Cervical malformation or marked instability
		7) History of severe cervical trauma, pregnancy, rheumatoid arthritis, malignancy, active infection, or other systemic disease
		8) Known allergy to implant material or to NSAIDs
Phillips et al 2015	1) Age 18-65 yrs	1) Prior failed cervical fusion (prior decompressions and adjacent and nonadjacent fusions allowed)	1) at least 20% improvement in NDI	Postoperative maintenance or improvement in 1) muscle strength
	2) Diagnosis of single-level radiculopathy and/or myelopathy	2) Previous cervical trauma	2) Absence of reoperation, revision, or removal	2) Sensory deficit
	3) Symptomatic at only 1 level C3-C4 through C7-T1	3) Marked cervical instability demonstrated by > 3.5 mm translation and/or >11° angular difference to that of either adjacent level	3) Maintenance or improvement in neurological status	3) Reflex functions
	4) Symptoms including 1 or more of the following	4) Congenital canal stenosis resulting in a canal diameter of <10 mm	4) Absence of radiographical or major complications
	- Arm/shoulder pain (>30 mm on 100-mm scale)	5) Radiographically confirmed facet pathology
	- Abnormal motor strength, sensation, and/or reflexes	6) Severe myelopathy to the extent that the patient is wheelchair bound
	- Myelopathy symptoms	7) Autoimmune disorders
	5) Radiographically determined pathology at level to be treated including at least 1 of the following	8) Osteoporosis and other metabolic bone disease
	- Decreased disc height	9) Infection (local or systemic)
	- Degenerative spondylosis	10) Diabetes mellitus
	- Disc herniation	11) Morbid obesity
	6) NDI score of >30/100	12) Malignancy/metastases
	7) Unresponsive to 6 weeks nonoperative treatment or progressive symptoms	13) Known allergies to device materials
Phillips et al 2021	1) Age ≥18 years old and ≤75 years old	1) More than 1 cervical level requiring surgery	1) Absence of supplemental surgical intervention at the index level	Postoperative maintenance or improvement in 1) muscle strength
	2) Diagnosis of degenerative cervical radiculopathy with or without spinal cord compression between C3 to C7 demonstrated by signs and/or symptoms of disc herniation and/or osteophyte formation	2) Previous anterior or posterior cervical spine surgery	2) Absence of serious adverse events classified as device or index level procedure-related	2) Sensory deficit
	2) Inadequate response to 6 weeks conservative medical care	3) Axial neck pain as the solitary symptom	3) NDI improvement of at least 15 points	3) Reflex functions
	3) NDI of ≥30% (raw score of ≥15/50)	4) Advanced cervical anatomical deformity (eg, ankylosing spondylitis, scoliosis) at the operative or adjacent levels	4) Maintenance or improvement in neurological function
	4) Neck or arm pain VAS ≥4 on a scale of 0 to 10	6) Symptomatic facet arthrosis
		7) Less than 4° of motion in flexion/extension at the index level
		8) Instability as evidenced by subluxation > 3 mm at the index or adjacent levels as indicated on flexion/extension x-rays
		9) Advanced spondylosis at the index vertebral level (eg, bridging osteophytes, central disc height <4 mm and/or <50% of the adjacent normal intervertebral disc, or kyphotic deformity >11° on neutral x-rays)
		10) Severe cervical myelopathy (ie, Nurick’s classification >2)
		11) Active systemic or local infection
		12) Co-morbid medical conditions of the spine or upper extremities
		13) Metabolic bone disease
		14) History of an osteoporotic fracture of the spine, hip or wrist
		15) History of an endocrine or metabolic disorder
		16) Taking medications that may interfere with bony/soft tissue healing
		17) Known allergy to titanium, stainless steels, polyurethane, polyethylene, or ethylene oxide residuals
		18) Autoimmune disease or a systemic disorder
		19) Insulin-dependent type 1 or type 2 diabetes
		20) Medical condition that may result in patient death or have an effect on outcomes prior to study completion
		21) Pregnant, or intend to become pregnant, during the course of the study
		22) Severe obesity (body mass index >40)
		23) Physical or mental condition that would interfere with patient self-assessment of function, pain or quality of life
		24) Involved in current or pending spinal litigation where permanent disability benefits are being sought
		25) Incarcerated at the time of study enrollment
		26) Current participation in other investigational study that may impact study outcomes
Radcliff et al 2017	1) Age 18-69 years	1) Multiple vertebral levels requiring treatment	1) ≥ 30-point improvement for patients with baseline NDI ≥60 or 50% improvement for patients with baseline NDI <60	Postoperative maintenance or improvement in 1) muscle strength
	2) Symptomatic cervical degenerative disc disease in 1 level between C3-C7	2) Immobile levels between C1 and C7 from any cause	2) No subsequent surgical intervention	2) Sensory deficit
	3) Myelopathy, myeloradiculopathy, decreased muscle strength, abnormal sensation and/or abnormal reflexes	3) Any prior surgery at the operative level or any prior fusion at any cervical level	3) AEs assessed by the CEC as major complications	3) Reflex functions
	4) Deficit confirmed by CT, MRI, or X-ray	4) Disc height less than 3 mm	4) Maintenance or improvement in neurological function
	5) NDI of ≥30/100	5) T-score less than -1.5	5) Radiographic success
	6) Unresponsive to 6 weeks non-operative treatment or presence of progressive symptoms	6) Metabolic bone disease
	7) No prior surgical procedures at the operative level and no prior fusions at any cervical level	7) Active local or systemic infection
	8) Physically and mentally able and willing to comply with the protocol	8) Active malignancy
	9) Willingness to discontinue all use of NSAIDs from 1 week before surgery until 3 months after surgery	9) Marked instability of the cervical spine
		10) Known allergy to device materials including cobalt, chromium, molybdenum, or polyethylene
		11) Segmental kyphosis of greater than 11° at treatment or adjacent levels
		12) Autoimmune disease
		13) Any diseases or conditions that would preclude accurate clinical evaluation
		14) Current or past high dose steroid use
		15) Morbid obesity (BMI >40)
		16) Use of any other investigational drug or medical device within 30 days prior to surgery
		17) Pending litigation relating to spinal injury (worker’s compensation not included)
		18) Smoking >1 pack of cigarettes per day
		19) Mental illness or belonging to a vulnerable population
Rozankovic et al 2017	1) Primary, single-level cervical disk disease from C3 to C7 with radiculopathy and/or myelopathy	1) Significant anatomic deformity such as severe spondylosis and radiographic signs of instability	NR	NR
	2) Failure of conservative treatment for at least 12 weeks	2) Previous procedures at the operative level
	3) NDI of ≥30%	3) ROM <2°
Sundseth et al 2017	1) Age 25 to 60 years	1) Significant spondylosis involving more than 1 level	NR	NR
	2) Clinical C6 or C7 radiculopathy with corresponding radiological findings	2) Adjacent level ankylosis
	3) NDI ≥30%	3) Intramedullary changes on MRI
	4) Unresponsive to 6 weeks non-operative treatment	Myelopathy
Vaccaro et al 2018	1) Age between 18 and 60	1) More than 1 vertebral level requiring treatment	1) NDI improvement of at least 25%	Postoperative maintenance or improvement in 1) muscle strength
	2) Cervical disease in 1 vertebral level between C3-C7	2) Prior fusion surgery at the adjacent level	2) No device failures requiring revision, removal, reoperation, or supplemental fixation	2) Sensory deficit
	3) Radiculopathy or myelopathy	3) Prior surgery at the level to be treated	3) Absence of major complications defined as major vessel injury, neurological damage, or nerve injury	3) Reflex functions
	4) Radiographic evidence of herniated nucleus pulposus, spondylosis, loss of disc height	4) Clinically compromised vertebral bodies due to current or past trauma	4) For patients who underwent ACDF only, radiographic fusion, as defined by the presence of bridging trabecular bone, without evidence of pseudarthrosis
	5) Failed at least 6 weeks of conservative treatment	5) Radiographical confirmation of facet joint disease
	6) NDI of at least 30	6) Marked cervical instability including translation greater than 3 mm and/or more than 11 degrees of rotational difference from that of either adjacent level
	7) Able to follow postoperative management program	7) Severe spondylosis at the level to be treated as characterized by any of the following
		- Bridging osteophytes
		- Loss of disc height >50%
		- Absence of motion (<2°)
		8) Neck/arm pain of unknown etiology
		9) Metabolic bone disease
		10) Pregnant or interested in becoming pregnant in the next 2 years
		11) Active systemic or local infection
		12) Known allergy to titanium, polyethylenem, cobalt, chromium, or molybdenum
		13) Taking medications or any drug known to potentially interfere with bone/soft tissue healing
		14) Autoimmune disease
		15) Systemic disease including AIDS, HIV, hepatitis
		16) Active malignancy
		17) Neuromuscular disorders
		18) Acute mental illness or substance abuse
		19) Use of bone growth stimulator within past 30d
		20) Participation in other investigational device or drug clinical trials within 30 days of surgery
		21) Prisoners
Zhang et al 2012	1) Symptomatic mild degenerative disc disease at 1 cervical level, including disc herniation with radiculopathy caused by foraminal osteophytes, soft disc herniation, or myelopathy	1) Patients with axial neck pain as a solitary symptom	NR	NR
	2) Unresponsive to at least 6 weeks of conservative treatment

Table 3.

Mixed Effects Comparisons of Patient-Reported Outcome Metrics Reported as the Effect Size and Corresponding 95% Confidence Interval.

NDI	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		.36 (−11.56 - 12.29)	.82 (−6.73 - 8.36)		5.8 (−6.61 - 18.21)	−.5 (−9.49 - 8.48)		5.7 (−6.86 - 18.26)
Bryan		0	.46 (−13.65 - 13.57)		5.44 (−11.77 - 22.65)	−.86 (−15.79 - 14.06)		5.34 (−11.97 - 22.65)
Discover					4.98 (−9.54 - 19.51)	−1.32 (−13.05 - 10.41)		4.88 (−9.77 - 19.53)
Kineflex
M6						−6.3 (−21.63 - 9.02)		−.1 (−17.76 - 17.56)
Mobi-C								6.2 (−9.23 - 21.64)
PCM
Prestige ST
ProDisc-C
Secure-C
VAS neck	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		.24 (−.91 - 1.39)	.95 (−.22 - 1.33)		1.1 (−.15 - 2.35)	.2 (−1.19 - 1.59)		.63 (−.60 - 1.86)		.61 (−.70 - 1.92)
Bryan			.71 (−1.07 - 1.70)		.86 (−.84 - 2.56)	−.04 (−1.85 - 1.77)		.39 (−1.29 - 2.07)		.37 (−1.37 - 2.11)
Discover					.15 (−.93 - 2.02)	−.75 (−1.95 - 1.24)		−.32 (−1.38 - 1.53)		−.34 (−1.47 - 1.57)
Kineflex
M6						−.90 (−2.77 - .97)		−.47 (−2.23 - 1.29)		−.49 (−2.30 - 1.32)
Mobi-C								.43 (−1.43 - 2.29)		.41 (−1.50 - 2.32)
PCM
Prestige ST									−.02 (−1.82 - 1.78)
ProDisc-C
Secure-C
VAS arm	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		.11 (−1.10 - 1.32)	.57 (−.62 - .96)		1.6 (.29 - 2.91)	.81 (−6.52 - 8.14)		.23 (−1.07 - 1.53)		.55 (−.77 - 1.87)
Bryan			.46 (−1.39 - 1.51)		1.49 (−.30 - 3.28)	.7 (−6.73 - 8.13)		.12 (−1.66 - 1.90)		.44 (−1.35 - 2.23)
Discover					1.03 (−.11 - 2.96)	.24 (−6.74 - 8.01)		−.34 (−1.46 - 1.58)		−.02 (−1.16 - 1.92)
Kineflex
M6						−.79 (−8.24 - 6.66)		−1.37 (−3.22 - .48)		−1.05 (−2.91 - .81)
Mobi-C								−.58 (−8.03 - 6.87)		−.26 (−7.71 - 7.19)
PCM
Prestige ST									.32 (−1.53 - 2.17)
ProDisc-C
Secure-C
SF PCS	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF			.50 (−3.05 - 4.05)		−3.10 (−4.43 to −1.77)	−1.70 (−5.01 - 1.61)		−1.90 (−4.29 - .49)		−1.70 (−4.70 - 1.30)
Bryan
Discover					−3.60 (−7.39 - .19)	−2.20 (−7.05 - 2.65)		−2.40 (−6.68 - 1.88)		−2.20 (−6.85 - 2.45)
Kineflex
M6						1.40 (−2.17 - 4.97)		1.20 (−1.54 - 3.94)		1.40 (−1.88 - 4.68)
Mobi-C								−.20 (−4.29 - 3.89)		0 (−4.47 - 4.47)
PCM
Prestige ST										.20 (−3.64 - 4.04)
ProDisc-C
Secure-C

Patient-Reported Outcome Metrics

Neck Disability Index

Eight studies^{1,8,11,19-22,24} comparing 5 cervical TDA devices (Bryan, Discover, M6, Mobi-C, Prestige ST) to ACDF contained sufficient data for analysis (Figure 3). Direct comparison of the effect size estimates (Table 4) demonstrated that no single TDA device significantly outperformed ACDF for reducing neck disability. Indirect comparison between devices found similar results across all paired comparisons.

Figure 3.

Forest plots demonstrating the MD and 95% CI of each device compared to ACDF as the reference for (A) NDI, (B) VAS Neck, (C) VAS Arm, and (D) SF PCS.

Table 4.

Mixed Effects Comparisons of Patient Success and Satisfaction Reported as the Effect Size and Corresponding 95% Confidence Interval.

Overall Success	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−2.19 (−4.56 - .18)		−2.41 (−5.25 - .43)	−1.70 (−3.56 - .16)	−1.23 (−2.46 - .005)		−1.76 (−4.03 - .51)		−2.15 (−4.58 - .28)
Bryan				−.22 (−3.92 - 3.48)	.49 (−2.53 - 3.51)	.96 (−1.71 - 3.63)		.43 (−2.86 - 3.72)		.04 (−3.36 - 3.44)
Discover
Kineflex					.71 (−2.69 - 4.11)	1.18 (−1.76 - 4.28)		.65 (−2.99 - 4.29)		.26 (−3.48 - 4.00)
M6						.47 (−1.76 - 2.70)		−.06 (−3.00 - 2.88)		−.45 (−3.51 - 2.61)
Mobi-C								−.53 (−3.12 - 2.06)		−.92 (−3.65 - 1.81)
PCM
Prestige ST
ProDisc-C
Secure-C
Neuro success	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−.62 (−1.02 - −.20)			−1.79 (−2.14 - −1.44)	−.99 (−1.70 - −.28)	−1.74 (−3.29 - −.19)	−1.89 (−4.05 - .27)	−.89 (−1.50 - −.28)	−2.03 (−3.66 - −.40)
Bryan					−1.18 (−1.72 - −.64)	−.38 (−1.20 - .44)	−1.13 (−2.73 - .47)	−1.28 (−3.47 - .91)	−.28 (−1.01 - .45)	−1.42 (−3.10 - .26)
Discover
Kineflex
M6						.80 (.01 - 1.59)	.05 (−1.54 - 1.64)	−.10 (−2.28 - 2.08)	.90 (.20 - 1.60)	−.24 (−1.90 - 1.42)
Mobi-C							−.75 (−2.45 - .95)	−.90 (−3.17 - 1.37)	.10 (−.83 - 1.03)	−1.04 (−2.81 - .73)
PCM								−.15 (−2.80 - 2.50)	.85 (−.81 - 2.51)	−.29 (−2.54 - 1.96)
Prestige ST									1.00 (−1.24 - 3.24)
ProDisc-C
Secure-C
Satisfaction (Cat)	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF					−2.00 (−4.02 - .02)	−3.20 (−5.65 - −.75)				−2.88 (−4.98 - −.78)
Bryan
Discover
Kineflex
M6						−1.20 (−4.37 - 1.97)				.88 (−3.79 - 2.03)
Mobi-C										.32 (−2.90 - 3.54)
PCM
Prestige ST
ProDisc-C
Secure-C
Satisfaction (VAS)	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF							−8.60 (−14.68 - −2.52)		−4.00 (−13.19 - 5.19)
Bryan
Discover
Kineflex
M6
Mobi-C
PCM									4.60 (−6.42 - 15.62)
Prestige ST
ProDisc-C
Secure-C

Visual Analog Scale Neck and Arm Pain

Eight studies^1,8,19–24 comparing 6 TDA devices (Bryan, Discover, M6, Mobi-C, Prestige ST, Secure-C) to ACDF contained sufficient data for analysis of neck and arm pain. For VAS Neck, direct and indirect comparisons demonstrated similar outcomes between all TDA devices and ACDF. For VAS Arm, the M6 device performed significantly better than ACDF in reducing arm pain (MD: 1.6, 95% CI: .29 - 2.91). No other significant differences were noted in direct or indirect comparisons.

Physical Component Score of the Short-Form Health Survey

Five studies^1,8,20,22,23 comparing 5 TDA devices (Discover, M6, Mobi-C, Prestige ST, Secure-C) to ACDF contained sufficient data for analysis. The M6 device performed significantly better than ACDF in improving physical health status (MD: −3.10, 95% CI: −4.43 to −1.77). No other significant differences were noted in direct or indirect comparisons.

Success and Satisfaction

Overall Success

Six studies^{1,8,9,14,20,23} comparing 6 TDA devices (Bryan, Kineflex, M6, Mobi-C, Prestige ST, Secure-C) to ACDF were included for analysis (Figure 4). Both direct and indirect comparisons demonstrated similar achievement of overall success between all TDA devices and ACDF (Table 5).

Figure 4.

Forest plots demonstrating the MD or log OR and 95% CI of each device compared to ACDF as the reference for (A) overall success, (B) neurological success, (C) categorical satisfaction, and (D) VAS satisfaction.

Table 5.

Mixed Effects Comparisons of Adverse Events and Subsequent Surgeries Reported as the Effect Size and Corresponding 95% Confidence Interval.

Dysphagia	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−.82 (−1.76 - .12)			−2.32 (−2.59 - −2.05)			−1.09 (−2.32 - .14)
Bryan					−1.50 (−2.48 - −.52)			−.27 (−1.82 - 1.28)
Discover
Kineflex
M6								1.23 (−.03 - 2.49)
Mobi-C
PCM
Prestige ST
ProDisc-C
Secure-C
Adverse events	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−1.11 (−4.25 - .60)	−1.02 (−4.26 - 2.22)	−1.23 (−4.80 - 2.34)	−1.70 (−5.19 - 1.79)	−.62 (−3.88 - 2.64)	−1.42 (−4.92 - 2.08)	−.47 (−3.75 - 2.81)	−1.10 (−4.55 - 2.35)	−3.37 (−8.44 - 1.70)
Bryan			.09 (−3.25 - 4.85)	−.12 (−3.72 - 4.90)	−.59 (−4.13 - 4.37)	.49 (−2.86 - 5.26)	−.31 (−3.85 - 4.66)	.64 (−2.73 - 5.43)	.01 (−3.49 - 4.94)	−2.26 (−7.17 - 4.07)
Discover				−.21 (−5.03 - 4.61)	−.68 (−5.44 - 4.08)	.4 (−4.20 - 5.00)	−.40 (−5.17 - 4.37)	.55 (−4.06 - 5.16)	−.08 (−4.81 - 4.65)	−2.35 (−8.37 - 3.67)
Kineflex					−.47 (−5.46 - 4.52)	.61 (−4.22 - 5.44)	−.19 (−5.18 - 4.80)	.76 (−4.08 - 5.60)	.13 (−4.83 - 5.09)	−2.14 (−8.34 - 4.06)
M6						1.08 (−3.70 - 5.86)	.28 (−4.66 - 5.22)	1.23 (−3.56 - 6.02)	.6 (−4.30 - 5.50)	−1.67 (−7.82 - 4.48)
Mobi-C							−.80 (−5.58 - 3.98)	.15 (−4.48 - 4.78)	−.48 (−5.23 - 4.27)	−2.75 (−8.78 - 3.28)
PCM								.95 (−3.84 - 5.74)	.32 (−4.59 - 5.23)	−1.95 (−8.11 - 4.21)
Prestige ST									−.63 (−5.39 - 4.13)	−2.90 (−8.94 - 3.14)
ProDisc-C										−2.27 (−8.40 - 3.86)
Secure-C
Index-level SSI	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−1.30 (−3.24 - −.63)	−.37 (−2.29 - .13)	−.93 (−3.16 - 1.30)	−2.38 (−4.93 - .17)	−5.01 (−8.86 - −1.16)	−.32 (−2.42 - 1.78)	−2.72 (−6.28 - .76)	−3.41 (−6.95 - .13)	−4.22 (−8.66 - .22)
Bryan			.93 (−.92 - 2.64)	.37 (−1.58 - 3.58)	−1.08 (−3.31 - 2.42)	−3.71 (−7.14 - .98)	.98 (−.86 - 4.08)	−1.46 (−4.58 - 2.93)	−2.11 (−5.25 - 2.29)	−2.92 (−6.91 - 2.34)
Discover				−.56 (−2.39 - 2.68)	−2.01 (−4.13 - 1.52)	−4.64 (−7.97 - .10)	.05 (−1.67 - 3.18)	−2.39 (−5.41 - 2.04)	−3.04 (−6.08 - 1.41)	−3.85 (−7.75 - 1.46)
Kineflex					−1.45 (−4.83 - 1.93)	−4.08 (−8.52 - .36)	.61 (−2.45 - 3.67)	−1.83 (−6.00 - 2.34)	−2.48 (−6.66 - 1.70)	−3.29 (−8.26 - 1.68)
M6						−2.63 (−7.25 - 1.99)	2.06 (−1.24 - 5.36)	−.38 (−4.73 - 3.97)	−1.03 (−5.39 - 3.33)	−1.84 (−6.96 - 3.28)
Mobi-C							4.69 (.31 - 9.07)	2.25 (−2.97 - 7.47)	1.60 (−3.63 - 6.83)	.79 (−5.08 - 6.66)
PCM								−2.44 (−6.54 - 1.66)	−3.09 (−7.21 - 1.03)	−3.90 (−8.81 - 1.01)
Prestige ST									−.65 (−5.64 - 4.34)	−1.46 (−7.13 - 4.21)
ProDisc-C										−.81 (−6.49 - 4.87)
Secure-C
Adjacent segment surgeries	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−3.12 (−5.25 - −1.68)	−.99 (−3.61 - 1.63)		−.67 (−3.25 - 1.91)	−5.09 (−7.74 - −2.36)	−6.67 (−12.56 - −.78)	−2.31 (−5.73 - 1.11)	−3.98 (−8.27 - .31)	−4.38 (−9.22 - .46)
Bryan			2.13 (−.69 - 5.64)		2.45 (−.34 - 5.93)	−1.97 (−4.81 - 1.64)	−3.55 (−9.36 - 2.95)	.81 (−2.70 - 5.01)	−.86 (−5.16 - 4.13)	−1.26 (−6.08 - 4.25)
Discover					.32 (−3.36 -4.00)	−4.10 (−7.81 - −.31)	−5.68 (−12.13 - .77)	−1.32 (−5.63 - 2.99)	−2.99 (−8.02 - 2.04)	−3.39 (−8.90 - 2.12)
Kineflex
M6						−4.42 (−8.11 - −.65)	−6.00 (−12.43 - .43)	−1.64 (−5.92 - 2.64)	−3.31 (−8.32 - 1.70)	−3.71 (−9.20 - 1.78)
Mobi-C							−1.58 (−8.10 - 4.86)	2.78 (−1.61 - 7.09)	1.11 (−4.00 - 6.14)	.71 (−4.87 - 6.21)
PCM								4.36 (−2.45 - 11.17)	2.69 (−4.60 - 9.98)	2.29 (−5.34 - 9.92)
Prestige ST									−1.67 (−7.16 - 3.82)	−2.07 (−8.00 - 3.86)
ProDisc-C										−.40 (−6.87 - 6.07)
Secure-C

Neurological Success

Seven studies^{1,8,13,14,16,20,23} comparing 7 TDA devices (Bryan, M6, Mobi-C, PCM, Prestige ST, ProDisc-C, Secure-C) to ACDF were included for analysis. Direct comparison demonstrated that all devices except for the Prestige ST performed significantly better than ACDF in achieving neurological success. Additionally, indirect comparison between devices found that the M6 outperformed the Bryan, Mobi-C, and ProDisc-C prostheses.

Satisfaction

Patient satisfaction was compared against ACDF categorically by 3 studies^1,20,23 for 3 devices (M6, Mobi-C, Secure-C) and with a VAS Satisfaction scale by 2 studies^13,16 for 2 devices (PCM, ProDisc-C). For categorical satisfaction, both Mobi-C and Secure-C significantly outperformed ACDF. For VAS satisfaction, only PCM performed significantly better than ACDF. Indirect comparisons between devices found similar results across all paired comparisons for both categorical and VAS satisfaction.

Adverse Events and Subsequent Surgery

Dysphagia

Four studies^1,8,15,22 comparing categorical dysphagia for 3 TDA devices (Bryan, M6, Prestige ST) to ACDF were included for analysis (Figure 5). Direct comparison demonstrated that the M6 device had a significantly lower association with dysphagia when compared to ACDF (Table 6). Additionally, comparison between devices found that the M6 device performed significantly better than the Bryan prosthesis.

Figure 5.

Forest plots demonstrating the log OR and 95% CI of each device compared to ACDF as the reference for (A) dysphagia, (B) adverse events, (C) index level SSI, and (D) adjacent segment surgeries.

Table 6.

Mixed Effects Comparisons of Radiographic Outcomes Reported as the Effect Size and Corresponding 95% Confidence Interval.

Segmental ROM	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−8.00 (−14.83 - −1.17)			−7.60 (−14.47 - −.73)	−5.13 (−12.33 - −2.53)	−4.70 (−11.56 - 2.16)		−7.46 (−14.41 - −.51)
Bryan					.40 (−9.28 - 10.08)	2.87 (−7.83 - 8.98)	3.30 (−6.38 - 12.98)		.54 (−9.20 - 10.28)
Discover
Kineflex
M6						2.47 (−8.26 - 8.61)	2.9 (−6.80 - 12.60)		.14 (−9.63 - 9.91)
Mobi-C							.43 (−5.70 - 11.15)		−2.33 (−8.53 - 9.47)
PCM									−2.76 (−12.52 - 7.00)
Prestige ST
ProDisc-C
Secure-C
Bridging bone	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF			6.36 (4.13 - 8.59)	5.64 (4.50 - 6.78)	8.61 (6.45 - 10.77)	5.12 (3.81 - 6.43)	5.58 (4.56 - 6.60)	6.38 (4.91 - 7.85)		6.03 (4.76 - 7.30)
Bryan
Discover				−.72 (−3.23 - 1.79)	2.25 (−.85 - 5.35)	−1.24 (−3.83 - 1.35)	−.78 (−3.24 - 1.68)	.02 (−2.65 - 2.69)		−.33 (−2.90 - 2.24)
Kineflex					2.97 (.53 - 5.41)	−.52 (−2.26 - 1.22)	−.06 (−1.59 - 1.47)	.74 (−1.12 - 2.60)		.39 (−1.32 - 2.10)
M6						−3.49 (−6.01 - −.97)	-3.03 (−5.41 - -.65)	−2.23 (−4.84 - .38)		−2.58 (−5.08 - −.08)
Mobi-C							.46 (−1.20 - 2.12)	1.26 (−.71 - 3.23)		.91 (−.92 - 2.74)
PCM								.8 (−.99 - 2.59)		.45 (−1.18 - 2.08)
Prestige ST										−.35 (−2.30 - 1.60)
ProDisc-C
Secure-C

Adverse Events

Ten studies^{1,8-10,13,15,16,20,21,23} compared device or surgery related adverse event profiles for 9 TDA devices (Bryan, Discover, Kineflex, M6, Mobi-C, PCM, Prestige ST, ProDisc-C, Secure-C) to ACDF. Both direct and indirect comparisons between all TDA devices and ACDF demonstrated similar associations with adverse events.

Index-Level Secondary Surgical Intervention

Fourteen studies^{1,8–11,13,15,16,19-24} compared the incidence of index-level SSI in 9 TDA devices (Bryan, Discover, Kineflex, M6, Mobi-C, PCM, Prestige ST, ProDisc-C, Secure-C) to ACDF. Direct comparison to ACDF demonstrated that the Bryan and Mobi-C devices were associated with significantly fewer surgeries at the index operative level. Additionally, when indirectly comparing between TDA devices, Mobi-C was significantly associated with fewer subsequent index-level surgeries than the PCM disc.

Adjacent Segment Surgery

Thirteen studies^{1,8,10,11,13,15,16,19-24} compared the incidence of adjacent level surgery in 9 TDA devices (Bryan, Discover, M6, Mobi-C, PCM, Prestige ST, ProDisc-C, Secure-C) to ACDF and were included for analysis. When compared to ACDF, a significantly lower association with adjacent segment surgery was seen with the Bryan, Mobi-C, and PCM devices. Additionally, indirect comparison between devices demonstrated that the Mobi-C device was associated with fewer adjacent segment surgeries than both the Discover and M6 prostheses.

Radiographic Outcomes

Segmental Range of Motion

Six studies^{1,11,13,16,20,24} provided sufficient data when comparing segmental ROM in 5 TDA devices (Bryan, M6, Mobi-C, PCM, ProDisc-C) to ACDF (Figure 6). The direct comparison between TDA devices and ACDF was performed as a precursor for indirect analysis between devices. Direct comparison between devices and ACDF demonstrated greater segmental ROM for all devices assessed (Table 7). This difference was significant for all devices except PCM. Indirect comparison between devices found similar results across all paired comparisons.

Figure 6.

Forest plots demonstrating the MD or log OR and 95% CI of each device compared to ACDF as the reference for (A) segmental ROM and (B) bridging bone across the operative segment.

Table 7.

Effect Size Estimate Comparisons of Radiographic Outcomes Reported as the Log Odds Ratio (Bridging Bone) or Mean Difference (Segmental ROM) and Corresponding 95% Confidence Interval.

Segmental ROM a	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF		−8.00 (−14.83 - −1.17)			−7.60 (−14.47 - −.73)	−5.13 (−12.33 - −2.53)	−4.70 (−11.56 - 2.16)		−7.46 (−14.41 - −.51)
Bryan					.40 (−9.28 - 10.08)	2.87 (−7.83 - 8.98)	3.30 (−6.38 - 12.98)		.54 (−9.20 - 10.28)
Discover
Kineflex
M6						2.47 (−8.26 - 8.61)	2.9 (−6.80 - 12.60)		.14 (−9.63 - 9.91)
Mobi-C							.43 (−5.70 - 11.15)		−2.33 (−8.53 - 9.47)
PCM									−2.76 (−12.52 - 7.00)
Prestige ST
ProDisc-C
Secure-C
Bridging bone b	ACDF	Bryan	Discover	Kineflex	M6	Mobi-C	PCM	Prestige ST	ProDisc-C	Secure-C
ACDF			6.36 (4.13 - 8.59)	5.64 (4.50 - 6.78)	8.61 (6.45 - 10.77)	5.12 (3.81 - 6.43)	5.58 (4.56 - 6.60)	6.38 (4.91 - 7.85)		6.03 (4.76 - 7.30)
Bryan
Discover				−.72 (−3.23 - 1.79)	2.25 (−.85 - 5.35)	−1.24 (−3.83 - 1.35)	−.78 (−3.24 - 1.68)	.02 (−2.65 − 2.69)		−.33 (−2.90 - 2.24)
Kineflex					2.97 (.53 - 5.41)	−.52 (−2.26 - 1.22)	−.06 (−1.59 - 1.47)	.74 (−1.12 - 2.60)		.39 (−1.32 - 2.10)
M6						−3.49 (−6.01 - −.97)	-3.03 (−5.41 - −.65)	−2.23 (−4.84 - .38)		−2.58 (−5.08 - −.08)
Mobi-C							.46 (−1.20 - 2.12)	1.26 (−.71 - 3.23)		.91 (−.92 - 2.74)
PCM								.8 (−.99 - 2.59)		.45 (−1.18 - 2.08)
Prestige ST										−.35 (−2.30 - 1.60)
ProDisc-C
Secure-C

^aReported as the Mean Difference (95% Confidence Interval).

^bReported as the Log Odds Ratio (95% Confidence Interval).

Bridging Bone

Seven studies^{1,8,9,16,20,21,23} reported the incidence of bridging bone secondary to HO in 7 TDA devices (Discover, Kineflex, M6, Mobi-C, PCM, Prestige ST, Secure-C) and successful fusion in ACDF. The direct comparison between TDA devices and ACDF was performed as a precursor for indirect analysis between devices. As expected, ACDF was associated with a significantly higher incidence of bridging bone than all TDA devices assessed. Indirect comparison of HO with bridging bone between TDA devices demonstrated a significantly lower association with M6 than the Kineflex, Mobi-C, PCM, and Secure-C devices.

Treatment Ranking

The P-scores for each treatment across all outcomes assessed are reported in Figure 7. Cumulative ranking of the average P-Score weighting each outcome assessed equally for ACDF and each TDA devices showed that the M6 prosthesis earned the highest overall rank (P = .70), followed by Secure-C (P = .67, Figure 8). The Discover device was the lowest ranked TDA device with a P-Score of .39, while ACDF was the lowest overall ranked treatment (P = .14). There was minimal variation in the performance of the remaining 6 devices, which had P-scores ranging from .49 to .57.

Figure 7.

Heat map of computed P-Scores assigned to each treatment modality for all outcomes assessed. Green denotes good performance, while red denotes poor performance.

Figure 8.

Cumulative ranking of each treatment based on average P-Score.

Discussion

This study quantitatively compared PROMs, patient success and satisfaction, complications, and radiographic outcomes between ACDF and TDA with 1 of 9 different devices, including the Bryan, Discover, Kineflex, M6, Mobi-C, PCM, ProDisc-C, Prestige ST, and Secure-C prostheses. When compared directly to ACDF, several devices were associated with superior outcomes for VAS Arm, SF PCS, neurological success, satisfaction, index-level SSI, and adjacent level surgeries. Few differences were noted in paired comparisons between individual devices. Cumulative ranking of treatments based on paired comparisons demonstrated the highest overall performance with the M6 device, followed by Secure-C. The remaining devices performed at a similar level, except for the Discover prosthesis which had the lowest performance among TDA devices. However, ACDF was the lowest performing treatment among all assessed.

The current body of high-quality literature supports the use of cervical TDA for the treatment of cervical radiculopathy and myelopathy, with most studies demonstrating non-inferiority to ACDF.^{1,8,9,12-14,16,23,25} The primary proposed benefit of the procedure includes the maintenance of motion at the operative segment, thus potentially decreasing abnormal strain at adjacent segments known to accelerate degeneration.^26-29 Further, the current evidence suggests that preservation of cervical kinematics may translate to superior pain and disability outcomes both in the short and longer term.^13,16,30 Findlay et al. conducted a systematic review of RCTs, finding that 6 out of the 8 trials assessed reported significantly less disability 3 months after the index surgery, as assessed by the NDI, while 5 of 7 trials reported lower neck pain. ³¹ The same review identified similar outcomes between TDA and ACDF at 2 years postoperatively among the included trials, however, at longer follow up between 4 to 7 years, more studies favored TDA. ³¹ Our quantitative analysis identified similar outcomes between all treatments in the network for neck disability and pain at an average follow up of nearly 6 years, while only the M6 device demonstrated superior outcomes to ACDF for arm pain and general physical health. The superior ability of the M6 design with 6 degrees of freedom to biomechanically mimic the kinematics of a native cervical motion segment has been described.^32,33 Thus, the better performance of the M6 disc noted in our study may be reflective of more physiologic kinematics.

The current study found similar rates of device or surgery-related adverse events between all TDA devices and ACDF, although subjective dysphagia was less frequently associated with the M6 device compared to both ACDF and TDA with the Bryan prosthesis. This difference in dysphagia between TDA and ACDF may be attributed to the lower profile of the TDA, contrasted to ACDF with an anterior plate, as used in the IDE trials, which is known to contribute to irritation and edema of prevertebral structures.^34-36 Anderson et al. compared the adverse event profiles reported in the Bryan FDA IDE trial at 2 years follow-up, finding a higher incidence of dysphagia and dysphonia among the investigational cohort, which the authors attributed to longer operative times, the use of a special frame, heightened sensitivity among researchers, or a potential true difference. ³⁷ Additionally, it is possible that the milling process required for the proper placement of the Bryan disc may release more proinflammatory cytokines contributing to soft tissue edema and transient dysphagia relative to placement of the M6 device, which relies on small fins for immediate stability until osseointegration is achieved.

Subsequent index-level revision or reoperation following ACDF is most commonly secondary to pseudarthrosis, subsidence/fracture, or instrumentation failure.^38,39 While subsidence and fracture may be seen following TDA, recent literature has identified a wide range of device-related complications that may necessitate further surgical intervention, including recurrent stenosis, device migration, motion loss, and even wear particle-induced osteolysis.^40-42 Index-level SSI were less frequently associated with the Mobi-C and Bryan devices than ACDF, and less frequently with Mobi-C than PCM. Device migration has been associated with the PCM disc, whose press-fit design and less optimal immediate fixation leaves it prone to migration in the early in the postoperative period. ⁴⁰ Further, the PCM disc achieves motion via gliding movement that is limited by the facet joints and soft tissues, rather than intrinsic device core limitations, increasing the strain on the joints which may accelerate the progression of facet arthropathy.^43,44

Although not typically revised surgically, HO with bridging bone results in restricted or lost ROM, in effect leading to a nearly fused or ankylosed motion segment. While the underlying cause remains poorly understood, some have suggested that the development may be related to device-induced biomechanical stress, constitutional factors such as osteophytes and preoperative calcification, or iatrogenic factors including endplate milling and over-distraction of the disc space via placement of an oversized device.^45-47 Conversely, gentle handling of soft-tissue structures, meticulous hemostasis, and the use of non-steroidal anti-inflammatory drugs (NSAIDs) early in the postoperative period may have prophylactic value in attenuating HO formation.^48,49 This analysis identified that the M6 device was less likely to be associated with motion-restricting HO than the Kineflex, Mobi-C, PCM, and Secure-C prostheses. While the shorter follow up reported by Phillips et al. for M6 may have likely contributed, device-specific design elements may also predispose them to HO formation. For example, both the Kineflex and Secure-C discs contain a keel for immediate fixation, potentially promoting HO.^50,51 Additionally, alteration in the segmental center of rotation (COR) has been correlated with a higher rate of HO. ⁵² For most devices, the COR is fixed, thus TDA implantation leads to a change from the physiologic axis, which is typically inferior to the intervertebral disc and posterior to the center of the endplate at C1 and gradually moves more superiorly and anteriorly at more caudal segments.^53-55 Patwardhan et al. compared cervical kinematics before and after implantation with either the M6 or Mobi-C discs, finding that the M6 restored native cervical kinematics, while significant intra-specimen differences were observed following Mobi-C placement. ³³

This network meta-analysis is not without limitations. All studies included in the quantitative analysis compared a single TDA device to ACDF, thus all between-device analyses in the network relied on indirect comparisons. Furthermore, due to the stringent inclusion and exclusion criteria maintained by the systematic review, a small number of studies were included in the final analysis, which can primarily be attributed to the few randomized trials currently in the literature for cervical TDA. Although all trials included in the analysis were of high quality, there were minor differences in the inclusion and exclusion criteria for enrollment and randomization for each trial, meaning that the patient population across all studies was not entirely homogenous. Finally, the studies included in the analysis reported outcomes at follow-ups ranging from 2 to 10 years. While the IDE trials that have spanned 10 years follow up have largely demonstrated maintenance of comparable outcomes from the short to long-term, certain variables assessed by this analysis have a temporal relationship, including HO and subsequent surgeries at both the index and adjacent segments. Despite these limitations, this is the first study to compare outcomes quantitatively and comprehensively across a wide range of cervical TDA devices using a network meta-analysis model. Our findings suggest that TDA may be superior to ACDF in the appropriately indicated patient population. Moreover, restoration of normal cervical kinematics with newer generation devices may improve patient outcomes, although future studies directly comparing device performance are necessary.

Conclusion

Cervical TDA was found to be superior on most outcomes assessed in the literature of high-quality clinical trials. When compared directly to ACDF, several devices were associated with superior outcomes for VAS Arm, SF PCS, neurological success, satisfaction, index-level SSI, and adjacent level surgeries. While most devices demonstrated similar outcomes, newer generation devices were found to outperform others across several outcomes assessed. This may be related to improved restoration of cervical kinematics, although future longitudinal studies directly comparing different TDA designs are necessary.

ORCID iDs

Athan G. Zavras https://orcid.org/0000-0001-9969-5815

Vincent P. Federico https://orcid.org/0000-0003-3749-9282

Michael T. Nolte https://orcid.org/0000-0001-9293-270X