Comparison of cervical disc arthroplasty versus anterior cervical discectomy and fusion for the treatment of single-segment cervical degenerative disc disease with a minimum of 4-year follow-up: a systematic review and meta-analysis of randomized controlled trials

Abstract

Objective

Our aim was to evaluate long-term effectiveness and safety of cervical disc arthroplasty (CDA) compared to anterior cervical discectomy and fusion (ACDF) in single-segment cervical degenerative disc disease (CDDD).

Methods

Comprehensive literature searches were performed in four databases from their inception until February 6, 2025. Meta-analyses were conducted with use of Stata 17.0.

Results

Seventeen randomized controlled trials were included in our meta-analysis. CDA group exhibited greater blood loss but demonstrated a higher neurological success rate, an increased range of motion, and lower scores on neck disability index as well as visual analog scales. Additionally, reoperation rates and adjacent segment degeneration rates were significantly reduced in comparison to ACDF group. There were no significant differences between two groups in surgical duration or adverse event rates.

Conclusion

In the long term, cervical disc arthroplasty presents favorable safety profile for treating single-segment CDDD with lower reoperation rates while demonstrating superior efficacy over ACDF along with effective reduction in adjacent segment degeneration rates.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-025-06189-x.

Introduction

The incidence of intervertebral disc degeneration is increasing annually, posing significant risks to public health and emerging as one of the major global health challenges [1]. Intervertebral disc degeneration is a progressive chronic condition that frequently results in persistent pain [2]. The symptoms of chronic pain are typically attributed to the physiological compression of spinal nerve roots caused by deformed discs [3]. Diseases related to symptomatic intervertebral disc degeneration consist of herniated discs, radiculopathy, myelopathy, spinal stenosis, and instability; these represent some of the most prevalent disorders diagnosed by spine physicians.

Cervical degenerative disc disease (CDDD) is a prevalent disorder that affects numerous individuals worldwide; it manifests through symptoms ranging from axial neck pain to radiculopathy and myelopathy [4]. The conditions can severely interfere with everyday activities and diminish quality of life—particularly among those engaged in physically demanding occupations or active lifestyles. An operation needs to be performed in cases where conservative therapies have failed to relieve symptoms [5–7].

Anterior cervical discectomy and fusion (ACDF) is widely considered one of the most prevalent and effective surgical interventions for patients suffering from refractory CDDD. This procedure involves the complete excision of the pathological intervertebral disc, followed by bone grafting within the intervertebral space. Subsequently, fixation is achieved using a plate or fusion device, thereby alleviating nerve compression caused by the diseased disc and mitigating associated symptoms. For decades, the conventional cage-plate system has served as the primary method of internal fixation in ACDF for treating symptomatic CDDD. However, numerous reports have highlighted plate-specific complications, including adjacent segment degeneration (ASD), swallowing difficulty and soft tissue injury [8–10].

The quest to minimize biomechanical alterations resulting from fusion surgery, reduce interference with adjacent cervical segments during recovery, and avoid potential complications has prompted exploration into alternative surgical therapies. Drawing inspiration from knee and hip replacement techniques developed in the 1970s, cervical disc arthroplasty (CDA) emerged as one significant non-fusion surgical option [11–13]. CDA entails an anterior approach to discectomy followed by implantation of an artificial prosthesis designed to restrict motion while mimicking natural disc function. This innovative technique aims to preserve cervical spine mobility and mitigate ASD risks [14]. Over recent decades, research has demonstrated that CDA can provide clinical outcomes comparable to or even superior than those achieved with ACDF [15]. However, CDA could lead to complications, including subsidence, migration, and misalignment [16]. In recent years, the necessity for reoperation following cervical disc arthroplasty has gained increasing attention among researchers [17].

Some randomized controlled trials (RCTs) compared the surgical treatment outcome of CDA versus ACDF, however, there was no consensus on better surgical intervention for CDDD. Previous meta-analyses compared merits and demerits between ACDF and CDA, but their conclusions were controversial, mainly associated with short-term follow-up (two-year follow-up) [18–20]. Long-term follow-up studies were necessary to confirm effectiveness of CDA as a surgical alternative for treating CDDD. Many researches in recent years evaluated medium- and long-term results of these two interventions, but there was still no consensus. Our aim was to compare postoperative long-term outcomes of CDA and ACDF by systematically evaluating RCTs in treating single-segment CDDD.

Methods

This review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [21].

Search strategy

Literature searches were conducted in Cochrane Library, Web of Science, PubMed and Embase databases from the start of the database until February 6, 2025. Only English language publications were included. References were managed using EndNote 20 and duplicates were automatically removed. A full search strategy was provided in Supplementary Material 1. Discrepancies in search results between the two researchers were resolved through consultation with a third evaluator.

Eligibility

Selection criteria were developed in accordance with PICOS principles.

Inclusion criteria

Participants: patients with a definitive diagnosis of single-segment CDDD who required surgery.

Intervention: CDA.

Comparison: ACDF.

Outcomes: operative time, blood loss, neck disability index (NDI), visual analog score (VAS) of neck pain and arm pain, neurological success, reoperation, adverse events, range of motion (ROM) of operated segment and ASD. Include at least one of the above outcomes.

Study design: RCT.

Exclusion criteria

(1) Studies with incomplete data, reviews, meta-analyses, conference papers and dissertations;

(2) Case reports;

(3) A nimal studies;

(4) D uplicate publications;

(5) A rticles with less than 48 months of follow-up;

(6) N on-English language publications.

Data extraction

Publication screening and data extraction were carried out and cross-validated individually by two investigators, according to the selection criteria. In case of discrepancies, a third researcher was involved in collaborative decision-making. If data were displayed by graphs only, GetData Graph Digitizer software was applied to extract the values. The extraction process included the following information:

1. Study characteristics: author, country, year of publication and study design;

2. Demographic and clinical data: age, sample size, duration of follow-up, prosthesis type;

3. Outcomes of interest.

Quality assessment

For RCTs, the risk of bias was assessed using Cochrane Risk of Bias Tool and the following types of bias were assessed for each study: selection bias (randomized sequence generation and allocation concealment), implementation bias (blinding of subjects and staff), detection bias (blinding of outcome assessment), attrition bias (incomplete data on endpoints), reporting bias (selective reporting) and other bias. Each domain was categorized as having a low, high or unclear risk of bias. The results of the quality evaluation were visualized using RevMan 5.4.1.

Statistical analysis

Meta-analyses were performed using Stata 17.0. Binary outcomes were expressed using odds ratios (OR) and 95% confidence intervals (CI), while continuous variables were analyzed using mean difference (MD) and 95% CI. Heterogeneity was assessed using I² statistic, with fixed-effects models used for low heterogeneity (I² < 50%) and random-effects models used for high heterogeneity (I² ≥ 50%). Sensitivity analysis was performed to evaluate the robustness of results. Egger test was used to assess publication bias.

Results

Study selection

Firstly, 3273 papers were searched from the database and 1552 duplicate publications were removed after importing them into EndNote 20 software. The remaining 1721 papers were further screened. After careful evaluation of titles and abstracts, 1309 were excluded. The remaining 412 papers were carefully read in full text and a total of 395 articles were excluded, including 152 involving CDA only, 112 involving multi-level surgery, 107 non-RCTs, and 24 with follow-up periods of less than four years. Finally, 17 studies were finally identified for inclusion in the meta-analysis [22–38] (Fig. 1).

Fig. 1.

Flow chart of literature search

Study characterization

Seventeen trials were included in this study following a thorough screening process based on predetermined selection criteria. From 2010 to 2024, these trials included 3,303 participants: 1,741 in CDA group and 1,562 in ACDF group. Thirteen of the RCTs were performed in United States, 2 in Netherlands, and 2 in China. Prostheses employed for CDA were the following: ActivC, Bryan, Discover, ProDisc-C, Kineflex|C, Mobi-C, SECURE-C, Prestige and PCM. More information on the selected trials was provided in Table 1.

Table 1.

Study characteristics

Author	Publication	Country	Study	Sample size		Age(years)		Follow-up	Prosthesis
	Year		Design	CDA	ACDF	CDA	ACDF	(years)	Type
Burkus	2010	USA	RCT	276	265	43.3(25–72)	43.9(22–73)	5	Prestige
Delamarter	2010	USA	RCT	103	106	42.1 ± 8.4	43.5 ± 7.2	4	ProDisc-C
Coric	2013	USA	RCT	41	33	49.5	49.3	4	Bryan
Zhang	2014	China	RCT	55	56	44.8	46.7 (18–68)	4	Mobi-C
Phillips	2015	USA	RCT	163	130	NR	NR	7	PCM
Janssen	2015	USA	RCT	103	106	42.1 ± 8.42	43.5 ± 7.15	7	ProDisc-C
Hisey	2016	USA	RCT	164	81	NR	NR	5	Mobi-C
Sasso	2017	USA	RCT	19	23	NR	NR	10	Bryan
Donk	2017	Netherlands	RCT	50	47	44.1 ± 6.4	43.1 ± 7.5	5	Bryan
Vaccaro	2018	USA	RCT	151	140	NR	NR	7	SECURE-C
Coric	2018	USA	RCT	136	133	NR	NR	6	Kineflex|C
Ghobrial	2018	USA	RCT	130	105	NR	NR	10	Bryan
Lavelle	2019	USA	RCT	242	221	NR	NR	10	Bryan
Liu	2022	China	RCT	36	36	40.3(32–50)	40.8(30–40)	4	Discover
Goedmakers	2023	Netherlands	RCT	35	36	46.5 ± 8.7	47.5 ± 8.0	5	ActivC
Sasso	2024	USA	RCT	19	23	39.8 ± 6.6	43.0 ± 6.6	20	Bryan
Foley	2024	USA	RCT	18	21	40.7	43.6	20	Bryan

CDA, cervical disc arthroplasty; ACDF, anterior cervical discectomy and fusion; RCT, randomized controlled trial; NR, not reported

Quality of studies

The risk of bias in the included researches is presented in Figs. 2 and 3. Two major drawbacks of all RCTs were implementation bias and detection bias. Blinding might not be possible in these studies due to the nature of the research. These trials demonstrated a low risk of bias regarding reporting, attrition and randomized sequence generation and showed a moderate risk of bias for allocation concealment.

Fig. 2.

Risk of bias summary

Fig. 3.

Risk of bias graph

Meta-analysis results

Operative time

Surgical time was analyzed in two studies. Pooled results showed no significant difference in operative time between CDA and ACDF groups [WMD = 29.02, 95% CI (-11.73, 69.78), p = 0.16] (Fig. 4) (Table 2).

Fig. 4.

Forest plot of operative time

Table 2.

Findings of meta-analysis

Outcome	Study	Effect size	95% CI		P-value	Heterogeneity		Effect
	size	WMD/OR	Lower limit	Upper limit		I2 (%)	P-value	model
Operation time	2	29.02	-11.73	69.78	0.16	93.86	0.00	Random
Blood loss	2	15.67	5.96	25.37	0.00	0.00	0.49	Fixed
VAS of neck pain	7	-3.48	-6.59	-0.37	0.03	96.34	0.00	Random
VAS of arm pain	7	-2.32	-4.42	-0.21	0.03	92.81	0.00	Random
Neck disability index	8	-2.80	-5.52	-0.07	0.04	96.40	0.00	Random
Neurological success	7	1.60	1.30	1.98	0.00	41.21	0.12	Fixed
Range of motion	4	10.37	4.69	16.05	0.00	95.93	0.00	Random
Adverse events	9	1.00	0.56	1.78	0.99	81.65	0.00	Random
Reoperation	14	0.39	0.30	0.51	0.00	12.04	0.32	Fixed
ASD	3	0.56	0.36	0.85	0.01	0.00	0.91	Fixed

VAS, visual analogue scale; ASD, adjacent segment degeneration; OR, odds ratios; CI, confidence interval; WMD, weighted mean difference

Blood loss

Data on blood loss was recorded by 2 papers analyzing 251 patients. Meta-analysis showed significantly more blood loss in CDA group, compared to ACDF group [WMD = 15.67, 95% CI (5.96, 25.37), p < 0.01] (Fig. 5).

Fig. 5.

Forest plot of blood loss

Visual analog score

Seven studies recorded separate data on VAS of neck pain and arm pain at the final follow-up, including a total of 1262 cases (CDA group 690, ACDF group 572). At final follow-up, CDA group provided lower neck pain VAS [WMD = -3.48, 95% CI (-6.59, -0.37), p = 0.03] (Fig. 6) and arm pain VAS [WMD = -2.32, 95% CI (-4.42, -0.21), p = 0.03] (Fig. 7), compared to the ACDF group.

Fig. 6.

Forest plot of visual analog score of neck pain

Fig. 7.

Forest plot of visual analog score of arm pain

Neck disability index

Eight studies, including 1,359 patients, reported NDI at the final follow-up. CDA significantly reduced NDI compared to ACDF at final follow-up [WMD = -2.80, 95% CI (-5.52, -0.07), p = 0.04] (Fig. 8).

Fig. 8.

Forest plot of neck disability index

Neurological success

Seven papers presented information regarding neurological success. Neurological success rate at ending follow-up was significantly improved in CDA group compared to ACDF group [OR = 1.60, 95% CI (1.30, 1.98), p < 0.01] (Fig. 9).

Fig. 9.

Forest plot of neurological success

Range of motion in operated segments

Four studies analyzing ROM of surgical segments involved 515 patients. ROM at the surgical level was significantly better in CDA group than in ACDF group at last follow-up [WMD = 10.37, 95% CI (4.69, 16.05), p < 0.01] (Fig. 10).

Fig. 10.

Forest plot of range of motion in operated segments

Adverse events

Nine papers reported adverse events. The pooled results indicated no significant difference in adverse event rates between CDA and ACDF groups at final follow-up [OR = 1.00, 95% CI (0.56, 1.78), p = 0.99] (Fig. 11).

Fig. 11.

Forest plot of adverse events

Reoperation

Fourteen reports analyzed information on reoperations, comprising 2533 patients. At final follow-up, reoperation incidence was significantly lower in CDA group than in ACDF group [OR = 0.39, 95% CI (0.30, 0.51), p < 0.01] (Fig. 12).

Fig. 12.

Forest plot of reoperation

ASD

Three trials recorded ASD rate and included 567 patients (CDA group 311, ACDF group 256). At last follow-up, CDA significantly diminished ASD rates compared to ACDF [OR = 0.56, 95% CI (0.36, 0.85), p = 0.01] (Fig. 13).

Fig. 13.

Forest plot of adjacent segment degeneration rate

Sensitivity analyses

Sensitivity analyses were conducted in this study with the aim of determining whether the findings were stable. Sensitivity analyses were conducted using an item-by-item method of excluding individual studies. The combined effect size was recalculated after excluding one study at one time. At last, we found excluding any of the studies did not significantly change the overall pooled effect sizes, except for surgical time, VAS, NDI, and ASD rate. Therefore, the findings of this meta-analysis for surgical time, VAS, NDI, and ASD rate might not be sufficiently robust and needed to be interpreted with caution (Supplementary Material 2).

Publication bias

The Egger test revealed no significant publication bias (Supplementary Material 3).

Discussion

Background and Findings

CDDD is a prevalent condition that affects human health. Its surgical treatment primarily involves ACDF and CDA. With extended postoperative follow-up periods, certain non-negligible drawbacks of ACDF have emerged, thereby promoting the development and refinement of CDA. Since its introduction, CDA has been anticipated to overcome certain limitations related to ACDF, potentially improving surgical outcomes and postoperative quality of life for patients with CDDD. Several studies have compared the clinical outcomes of these two techniques [39–43]. With the ongoing promotion of CDA, an increasing amount of clinical data regarding its motion preservation following prosthetic replacement has become available, leading to a deeper understanding of its efficacy and safety.

Despite decades of clinical application of CDA, ACDF remains the preferred treatment option for CDDD, largely due to firmly established long-term effectiveness of ACDF contrasted with the still-evolving evidence of CDA. Meta-analyses evaluated the effectiveness and safety of CDA versus ACDF in treating CDDD. Nevertheless, most of these analyses involved research with short-term follow-up periods, limiting the capacity in drawing conclusions regarding long-term comparative outcomes. Therefore, we conducted a meta-analysis with a minimum follow-up duration of four years to evaluate the long-term outcomes of CDA and ACDF in single-segment CDDD.

Our meta-analysis revealed CDA demonstrated greater intraoperative blood loss but also showed superior outcomes in terms of neurological success, range of motion, lower VAS scores, NDI, reoperation rates, and ASD rates compared to ACDF. These findings indicate CDA could be more effective than ACDF for treating CDDD.

Safety of CDA versus ACDF

Surgical safety is typically assessed based on adverse events, reoperation rates, operative time, and intraoperative blood loss. Adverse events encompass complications such as postoperative hematoma, cerebrospinal fluid leakage, surgical site infection, heterotopic ossification, adjacent segment degeneration, prosthesis displacement or subsidence, localized bony residue or bridge formation at the operative level, and allergic reactions to metallic components. Reoperation was described as repeat surgical intervention at the index level for all causes, including prosthesis adjustment, revision, removal, or additional fusion and fixation procedures [44, 45]. Common reasons for reoperation in the CDA group include progressive deterioration of postoperative neurological symptoms, persistent neck and shoulder pain, subsidence or dislocation of the prosthesis resulting in voice impairment, and prosthesis displacement [46, 47]. In the ACDF group, reoperations were primarily attributed to worsening neurological deficits, unresolved postoperative symptoms, endplate prosthesis prolapse or displacement, implant failure to achieve fusion, and ASD [48, 49]. Our analysis found no significant difference in adverse event rates between both groups. Nevertheless, reoperation rate was significantly lower in CDA group compared to ACDF group.

Only two papers reported data on operative duration, and results revealed no statistically significant differences between both groups. Surgical duration is influenced by multiple factors, including patient-specific conditions, intraoperative circumstances, and surgeon experience. CDA exhibited higher blood loss than ACDF, which might be attributed to the increased number of osteotomies required for prosthesis placement. While notable significant differences in blood loss were observed among CDA and ACDF, its clinical relevance appeared limited.

In conclusion, no significant differences were observed between CDA and ACDF concerning overall surgical safety. Nonetheless, CDA demonstrated significantly lower reoperation rate than ACDF.

Efficacy of CDA versus ACDF

These included studies argued for the superiority of CDA versus ACDF by comparing surgical results by analyzing VAS, NDI, and neurological success rates. VAS and NDI provided a comprehensive and objective assessment, and the quantitative results were easy to compare, so the conclusions drawn were more reliable.

Postoperative neck and upper extremity discomfort and pain following cervical spine surgery represent critical factors influencing patients’ quality of life and serve as primary indicators for evaluating surgical efficacy. The alleviation of such symptoms is primarily attributed to decompressing spinal cord and nerve roots and restoring spinal stability. Within our systematic review, separate meta-analyses of NDI and VAS were conducted. At last, our findings revealed a significant difference in NDI among CDA and ACDF. This might be associated with the preservation of motion in operative segments during CDA, resulting in a postoperative cervical spine configuration closer to its physiological state. Furthermore, compared to ACDF, at the final follow-up, VAS in CDA group were significantly lower, which could be linked to increased cervical mobility after surgery and improved function of the cervical musculature. These findings align with those reported by Wu [50], suggesting that CDA, as a relatively new non-fusion technique, can more effectively relieve spinal cord and nerve root compression, significantly improve objective neurological symptoms and signs of cervical spondylosis, and achieve comparable efficacy to the established “gold standard” procedure, ACDF.

We discovered neurological success rate was higher in CDA than in ACDF. It is hypothesized this outcome may stem from the ability of CDA to restore normal intervertebral height and preserve spinal functional units, thereby more effectively relieving pressure on adjacent segments and intervertebral spaces, ultimately leading to improved neurological status.

Artificial discs were developed later than artificial joints and have not yet achieved the same level of clinical adoption due to various factors. However, CDA and joint replacement share conceptual similarities, as both aim to preserve the function of local anatomical structures. By implanting an artificial disc into the intervertebral space, CDA avoids segmental fusion and preserves maximal mobility at the operated level. Therefore, cervical disc replacements appear to support the biomechanics of neck. It is expected that they will lower ASD incidence. Nevertheless, studies by Jawahar et al. indicated the risk of ASD following CDA was not necessarily lower than that after ACDF and might even be elevated in patients with concomitant osteoporosis or degenerative disc disease [51, 52]. Yin also reported that while CDA preserved ROM at operated segments, there were no significant differences regarding ROM at adjacent segments [53]. In our study, we observed that at the final follow-up, ROM was significantly greater in CDA than in ACDF, and ASD incidence was significantly lower during CDA. However, there remains insufficient evidence to determine whether ASD following ACDF differs from its natural degenerative progression, whether fusion exacerbates this process, or whether preserving ROM after CDA reduces the rate of ASD.

Limitations

Our study has several limitations. Firstly, implant types used for CDA and ACDF varied across the included RCTs, potentially affecting the validity of the pooled results. Secondly, cost-effectiveness could not be assessed because access to individual patient-level data was unavailable. Thirdly, the results were also affected by heterogeneity. For instance, the results of operative duration, VAS, NDI and ROM, and AE showed significant heterogeneity. This might be because measurement errors were unavoidable during the measurement process. Clinical heterogeneity could result from the characteristics of various prostheses, different surgical indications, and operative techniques employed across various medical centers. Moreover, sensitivity analysis showed that the results of the meta-analysis on procedure duration, VAS, NDI, and ASD incidence were not robust. Consequently, our findings should be interpreted with caution. Fourthly, only English-language publications were included, which might introduce language bias. Lastly, some methodological limitations were identified among the included studies, including unclear allocation concealment methods and inadequate blinding procedures. These shortcomings may compromise the overall credibility of the synthesized results. Therefore, further independent, high-quality RCTs with long-term follow-up are warranted to enhance the strength of evidence and provide complementary insights.

Conclusion

In summary, cervical disc arthroplasty demonstrates a favorable long-term safety profile in treating one-segment cervical disc degenerative disease, characterized by lower reoperation rate compared to ACDF. Furthermore, CDA exhibits superior efficacy in comparison to ACDF and is linked to a significantly reduced incidence of adjacent segment degeneration. However, these conclusions require further validation through additional high-quality RCTs with extended follow-up periods.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CDDD

Cervical degenerative disc disease

CDA

Cervical disc arthroplasty

ACDF

Anterior cervical discectomy and fusion

VAS

Visual analog score

NDI

Neck disability index

RCT

Randomized controlled trial

ASD

Adjacent segment degeneration

ROM

Range of motion

OR

Odds ratios

CI

Confidence interval

WMD

Weighted mean difference

Author contributions

YZ and JJ collaborated on the design, literature retrieval, and data extraction, and performed data analysis and quality assessment. YZ and JW made the final assessment and interpretation of the results. YZ drafted the manuscript. JJ and JW contributed to the critical revision of the manuscript. All authors read and approved the final version of the manuscript.

Funding

This research received no external funding.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable in this section.

Consent for publication

Not applicable in this section.

Competing interests

The authors declare no competing interests.