Three-Level Cervical Disc Arthroplasty Combined With Fusion Versus Three-Level ACDF: A Systematic Review and Meta-Analysis

Abstract

Study Design

Systematic Review.

Objectives

To compare clinical, radiological, and complication outcomes between three-level hybrid surgery [combining cervical disc arthroplasty (CDA) with anterior cervical discectomy and fusion (ACDF)] and three-level ACDF alone in patients with multilevel cervical degenerative disease.

Methods

A systematic review and meta-analysis were conducted following PRISMA guidelines. Nine retrospective cohort studies (704 patients: 450 HS, 254 ACDF) were included. Pooled mean differences (MD) with 95% confidence intervals (CI) were calculated for clinical, radiological, perioperative, and complication outcomes.

Results

Both HS and ACDF significantly improved disability and myelopathy scores, with no differences in NDI or mJOA. HS preserved better the operated-level ROM (MD + 5.79°; 95% CI 2.38 – 9.21; P = 0.0009), while adjacent segment motion did not differ. Arm pain improvement showed a non-significant trend favouring HS (MD –2.63; P = 0.08), and neck pain was similar. Both procedures improved segmental lordosis; global alignment remained stable. Complications differed: instrumentation failure was more common with ACDF (33.3% vs 5.2%, P = 0.02), while HO occurred in 41.2% of hybrid constructs, reflecting loss of arthroplasty motion.

Conclusion

Three-level HS and ACDF provide comparable clinical outcomes. HS is a promising alternative, providing better motion preservation but at the cost of a higher risk of heterotopic ossification, while ACDF may be associated with a higher rate of instrumentation failure. Long-term prospective studies are still needed to strengthen these conclusions.

Keywords

hybrid surgery cervical disc arthroplasty three-level cervical spine surgery ACDF multilevel cervical degenerative disease

Introduction

Degenerative cervical spine disease is one of the leading causes of neck pain, radiculopathy, and myelopathy, with a considerable impact on quality of life and health-care costs worldwide.^1,2 When non-operative management fails, anterior cervical discectomy and fusion (ACDF) has traditionally been the standard of care, particularly for multilevel pathology.^3,4 ACDF provides reliable decompression, alignment correction, and fusion rates; however, it eliminates motion at the operated levels and has been associated with increased biomechanical stress at adjacent segments, potentially accelerating adjacent segment degeneration (ASD) and raising the risk of secondary surgery.^5–7 These concerns are especially relevant in multilevel fusions, where the loss of mobility is more pronounced and rates of non-union, hardware-related complications, and ASD are higher.^8,9

Cervical disc arthroplasty (CDA) was introduced as a motion-preserving alternative to fusion. Randomised controlled trials and long-term follow-up studies have consistently demonstrated that, in one- and two-level disease, CDA achieves at least equivalent clinical outcomes to ACDF, while preserving range of motion (ROM) and lowering the incidence of adjacent-level reoperations.^10–13 Despite this robust evidence base, CDA has not been widely adopted for three-level disease. Regulatory restrictions, device limitations, and technical challenges limit its availability beyond 2 levels in most regions.¹⁴

For patients with symptomatic three-level pathology, hybrid surgery (HS), a combination of CDA at mobile segments and ACDF at others. Has therefore emerged as a pragmatic middle ground.^15,16 By maintaining mobility at selected segments, HS may reduce compensatory motion at adjacent levels and improve postoperative spinal kinematics while still providing stability where fusion is most appropriate.^17,18 Early single-centre reports and comparative cohort studies suggest that HS offers similar improvements in disability and pain to ACDF, with the added advantage of preserving operated-level ROM.^19,20,21,22 However, findings are not entirely consistent: questions remain regarding the impact of HS on global sagittal alignment, its long-term durability, and whether the benefits of motion preservation translate into fewer adjacent-level complications.^23,24

Existing systematic reviews and meta-analyses have largely focused on one- and two-level disease, leaving a gap in evidence for three-level or more procedures.²⁵ Given the increasing number of younger, active patients presenting with multilevel cervical disease, clarifying the role of HS vs ACDF at three levels has direct clinical relevance.

Evidence comparing three-level hybrid surgery and three-level ACDF is limited to retrospective cohort studies, as no randomised or prospective comparative trials currently exist for this relatively new surgical approach. These retrospective studies, although imperfect, represent the entirety of available comparative data and consistently evaluate similar patient populations, indications, and outcome measures. Given the increasing use of hybrid constructs in clinical practice, a systematic synthesis of these datasets is necessary to provide clinicians with consolidated evidence and to guide surgical decision-making. The aim of this study was therefore to conduct a systematic review and meta-analysis of three-level HS compared with three-level ACDF. We sought to evaluate clinical outcomes, radiological parameters, complication profiles, and differences between hybrid subtypes. By synthesising the available evidence, this study provides a more comprehensive understanding of the relative benefits and limitations of these 2 strategies in the management of complex three-level cervical degenerative disease.

Methodology

Study Design

This study was conducted as a systematic review and meta-analysis in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.²⁶

Data Sources and Search Strategy

We conducted a search across 5 key electronic databases: PubMed, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, and Web of Science from inception to June 2025. Our search strategy utilised both free-text keywords and controlled vocabulary (MeSH terms and specific subject headings for each database). The terms used in the search included: “cervical disc arthroplasty”, “cervical disc replacement”, “artificial disc replacement”, “ACDF”, “Anterior cervical discectomy and fusion”, “three level”, “three-level”, “3 level”, “multilevel”, “multi-level”, “triple level”, “three segment”, “3 segment”, “≥3”, “4 level”, “four level”. Each database was queried using a customised strategy to maximise sensitivity and specificity. No language restrictions were initially applied during the search phase. However, only articles published in English were considered eligible during screening.

Eligibility Criteria

Inclusion criteria were structured using the PICOTS framework.

Population: Adults undergoing surgery for degenerative cervical spine disease involving three cervical levels.

Intervention: Hybrid surgery including both ACDF and arthroplasty at three-levels.

Comparator: Anterior cervical discectomy and fusion (ACDF) performed at three-levels.

Outcome:

◦ Primary clinical outcomes: Neck Disability Index (NDI), modified Japanese Orthopaedic Association (mJOA) score, Visual Analogue Scale (VAS) scores for neck and/or arm pain.

◦ Radiological outcomes: Segmental and global range of motion (ROM), cervical lordosis (C2–C7 and operated segment), sagittal vertical axis (SVA), T1 slope, and radiographic evidence of adjacent segment degeneration (ASD).

◦ Secondary outcomes: Operative duration, intraoperative blood loss, and perioperative complications [including dysphagia, C5 palsy, recurrent laryngeal nerve injury, wound infection, instrumentation failure, transfusion, cardiopulmonary events, readmission, heterotopic ossification (HO), and revisions].

Timing: Any follow-up duration.

Study design: Prospective or retrospective cohort studies, case series or RCTs.

Exclusion Criteria

Studies including trauma, neoplastic, infectious, or inflammatory causes were excluded to ensure a clinically homogeneous population.

Index levels <3 only (studies exclusively of single-level or two-level procedures).

Mixed-level cohorts (eg, 1-3 levels) that do not report separate extractable data for the ≥3-level subgroup.

Case reports, narrative reviews, systematic reviews or meta-analyses, opinion pieces, and surgical technique papers without original data.

Cadaveric, biomechanical, animal, or in-vitro studies.

Conference abstracts.

Study Selection

All search results were imported into a reference manager and de-duplicated. Two independent reviewers screened titles and abstracts for eligibility. Full-text screening was conducted for potentially relevant studies. Disagreements were resolved by consensus with the lead researcher. The reference lists of included studies were manually checked for additional eligible trials.

Data Extraction

Data extraction was performed independently by 2 reviewers. Extracted variables included.

Study characteristics: author, year, country, study design, sample size, and follow-up duration.

Patient characteristics: age and sex.

Surgical details: number and levels of disc replacements/fusions performed.

Outcomes: clinical scores (NDI, VAS neck/arm, mJOA, ROM), operative duration, blood loss, complications, revision rates, and radiological parameters (cervical lordosis, SVA, T1 slope, adjacent segment changes).

Most included studies reported clinical and radiological outcomes only at the final follow-up visit rather than at predefined intervals such as 6 or 12 months. Because follow-up timing was inconsistent across studies and insufficiently detailed to extract outcomes at uniform time points, we prespecified the use of each study’s last available follow-up for all pooled analyses. To assess whether follow-up duration influenced results, we conducted a sensitivity analysis restricted to studies with ≥2 years of follow-up.

Radiological Assessments

Included studies assessed radiographic outcomes using standard lateral radiographs in flexion, extension, and neutral positions. Parameters were measured as follows.

(1) T1 slope: angle between a horizontal line and the superior T1 endplate.

(2) Cervical lordosis (C2–C7): Cobb angle measured between the inferior endplates of C2 and C7.

(3) Sagittal vertical axis (SVA): horizontal distance between a vertical plumb line dropped from the centre of C2 and the posterior-superior corner of C7.

(4) Range of motion (ROM): difference in C2–C7 angles between flexion and extension on dynamic lateral radiographs.

All parameters were recorded at baseline and the latest available follow-up.

Risk of Bias Assessment

The methodological quality of included studies was assessed using the NIH Quality Assessment Tool for observational studies.²⁷ Domains evaluated included study population, selection bias, control of confounding, blinding of outcome assessment, and completeness of follow-up.

Data Analysis

All included studies were retrospective; hence a random-effects meta-analysis was chosen to account for methodological variation and inherent heterogeneity among observational datasets. Continuous outcomes such as NDI, mJOA score, ROM, cervical lordosis, pain VAS, and operative parameters were expressed as mean differences (MD) with 95% confidence intervals (CI), while categorical outcomes were reported as risk ratios (RR) with 95% CI. Heterogeneity was assessed using the I² statistic, with thresholds of 25%, 50%, and 75% representing low, moderate, and high heterogeneity. We prespecified the use of each study’s final available follow-up for all pooled outcomes, as intermediate time-point data were inconsistently reported. Where studies lacked sufficient data for pooling, findings were summarised descriptively. A planned subgroup analysis was conducted within the hybrid cohort, comparing “one disc arthroplasty plus 2 ACDF levels (1CDA + 2ACDF)” with “two-disc arthroplasties plus one ACDF level (2CDA + 1ACDF)” to evaluate whether the proportion of motion-preserving levels influenced outcomes. Although we sought to analyse outcomes based on the position of the disc arthroplasty within the construct (upper, middle, or lower level), the included studies did not consistently report the specific CDA location; therefore, a positional subgroup analysis was not feasible.

A sensitivity analysis was performed to evaluate whether results differed in studies with ≥2 years of follow-up. Studies reporting a mean or minimum follow-up period of less than 24 months were excluded from this secondary analysis, and all primary outcomes were re-analysed using the same random-effects model. Publication bias was not formally assessed given the small number of studies (<10) per outcome, and statistical significance was defined as P < 0.05, Change scores were calculated using the formulae on MetaConverter. The pooled data were analysed using RevMan 5.4.1 version software (Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, 2014).

Results

Study Selection

From 438 unique records, 13 full-text articles were reviewed, and nine met the inclusion criteria. No additional eligible studies were identified from reference lists. The selection process is summarised in Figure 1.

Figure 1.

Prisma Chart

Study Characteristics

Nine retrospective studies were included, comprising 704 patients. Of these, 450 underwent hybrid three-level surgery (CDA + ACDF) and 254 underwent three-level ACDF. Hybrid procedures were further classified as 1CDA + 2ACDF, n = 217 and 2CDA + 1ACDF, n = 116. No four-level cases were reported.

Patients in the hybrid cohorts were significantly younger than those undergoing ACDF (52 vs 60 years; P < 0.0001), whereas sex distribution, follow-up duration, and number of operated levels did not differ. The most commonly operated segments were C3–C6. Mean follow-up was similar between groups (47.7 vs 50.3 months; P = 0.4). All primary outcomes were extracted at the final reported follow-up, as intermediate time-point data were inconsistently reported across studies.

Study characteristics and demographics are shown in Tables 1 and 2.

Table 1.

Summary of Included Study Characteristics

S/N	Study ID	Study design	Indication for surgery	Hybrid surgery sample size (n)	ACDF sample size (n)	Hybrid surgery mean follow up duration, (SD/range) months	ACDF mean follow up duration, (SD/range) months	Outcomes assessed
1	Zheng 2024¹⁸	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	44	24	138.6 ± 7.6	136.9 ± 7.5	Clinical, radiological and perioperative
2	Chen, 2024²⁴	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	81	42	24.3 ± 15.87	29.1 ± 19.19	Perioperative, and clinical
3	Chen, 2023¹⁶	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	73	33	22.15 ± 12.4	29.1 ± 9.3	Perioperative, radiological, and complications
4	Huang, 2022¹⁷	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	84	24	34.03 ± 19.7	29.13 ± 9.3	Perioperative, clinical, radiological and complications
5	Wang, 2021²²	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	64	37	47.8 ± 20	49.03 ± 14.7	Perioperative, clinical and radiological
6	Xu, 2020²³	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	51	32	93.3 ± 8.5	89.3 ± 9.5	Perioperative, and radiological
7	Hung, 2018¹⁹	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	22	20	28 (21-35)	28 (21-35)	Perioperative, clinical, radiological and complications
8	Jang, 2017²⁰	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	19	30	30.5 ± 18.6	32.3 ± 13.1	Perioperative, clinical, radiological and complications
9	Kang, 2013²¹	Retrospective	Cervical spondylotic myelopathy involving 3 cervical levels	12	12	32.8 ± 7.5	33.2 ± 7.7	Perioperative, clinical, radiological and complications

Table 2.

Summary Demographics of Hybrid Surgery Versus ACDF, and Subgroup Summary of Hybrid Surgery

	Hybrid surgery	ACDF	P value/Mean difference (MD)
Number of patients (n)	450	254	n/a
Age (years)	52.06 (8.78)	60.34 (10.93)	P < 0.0001 MD: −8.28 [−9.7619 to −6.7981]
% female	49.80%	44.49%	P = 0.1773
Mean follow up (SD) months	47.78 (38.75)	50.27 (36.59)	P = 0.4039 MD: −2.49 [−8.3471 to 3.3671]
Cervical levels operated	C3 - C6 30.4% C4 - C7 68.4% Skip 1.1%	C3 - C6 30.3% C4 - C7 68.1% Skip 1.6%	P = 0.8708

	1CDA + 2ACDF	2CDA + 1ACDF	P value
Subgroup summary of hybrid surgery cohort
Number of patients (n)	217	116	n/a
Age (years)	52.56 (8.66)	49.71 (9.13)	P = 0.0053
% female	51.15%	50.00%	P = 0.8412
Mean follow up (SD) months	52.18 (43.98)	47.84 (41.71)	P = 0.3831

Across studies, the primary indication for surgery was three-level cervical spondylotic myelopathy, occasionally accompanied by radiculopathy. No study included non-degenerative pathology, and diagnostic criteria were broadly comparable, thereby reducing clinical heterogeneity between cohorts.

Risk of Bias Assessment

Overall, the included studies carry a moderate risk of bias, primarily due to retrospective design, non-random treatment allocation, and limited adjustment for confounding. Exposure and outcome measurements were generally reliable, using validated clinical scores and standard radiographs, though blinding of outcome assessors was rarely reported. Summary of the risk of bias assessment is shown in Table 3.

Table 3.

NIH Quality Assessment Questionnaire Risk of Bias Summary for Included Studies

S/N	Question	Zheng 2024	Chen 2024	Wang 2023	Chen 2023	Huang 2022	Xu 2020	Hung 2018	Jang 2017	Kang 2013
1	Research question/objective clearly stated?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
2	Study population clearly specified/defined?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
3	Participation rate ≥50%?	NA	NA	NA	NA	NA	NA	NA	NA	NA
4	Same/similar populations; prespecified inclusion/exclusion applied uniformly?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
5	Sample size/power/variance/effect estimates provided?	No	No	No	No	No	No	No	No	No
6	Exposure measured prior to outcome?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
7	Timeframe sufficient to see association?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
8	Different levels of exposure examined?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
9	Exposure measures clear/valid/reliable/consistent?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
10	Exposure assessed more than once over time?	NA	NA	NA	NA	NA	NA	NA	NA	NA
11	Outcome measures clear/valid/reliable/consistent?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
12	Outcome assessors blinded to exposure?	No	CD	CD	CD	No	CD	No	CD	Yes
13	Loss to follow-up ≤20% after baseline?	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes
14	Key confounders measured and adjusted?	No	No	No	No	No	No	No	No	No

Clinical Outcomes

Neck Disability Index (NDI)

Four studies^17,18,19,20 reported NDI outcomes. Pooled analysis demonstrated no significant difference in NDI improvement between HS and ACDF (MD –1.38, 95% CI –3.46 to 0.7; P = 0.19; I² = 34%). Figure 2.

Figure 2.

Forest Plot Showing Changes in NDI Between Baseline and Final Follow up For Both Cohorts

Modified Japanese Orthopaedic Association Scale (mJOA)

Two studies assessed myelopathy outcomes.^17,18 The pooled analysis showed no statistical difference between changes in mJOA in both the HS and ACDF cohorts (MD –0.28, 95% CI –1.0 to 0.44; P = 0.45).

Pain Scores (VAS)

Three studies reported VAS outcomes.^17,19,20 For arm pain, results favoured hybrid but did not reach significance (MD –2.63, 95% CI –5.58 to 0.32; P = 0.08). Neck pain scores were comparable between groups (MD 0.28, 95% CI –0.11 to 0.66; P = 0.16).

Radiological Outcomes

Range of Motion (ROM) at Operated Levels

Five studies evaluated motion at the operated segments. Hybrid surgery preserved significantly greater ROM compared with ACDF (MD 5.79°, 95% CI 2.38-9.21; P = 0.0009; I² = 39%). Figure 3.

Figure 3.

Forest Plot Showing Changes in ROM Between Baseline and Final Follow up for Both Cohorts

Adjacent Segment ROM

Three studies analysed adjacent segment ROM.^18,21,23 No significant differences were seen at the upper (MD –1.71°, 95% CI –5.34 to 1.92; P = 0.36) or lower segments (MD –1.21°, 95% CI –3.99 to 1.56; P = 0.39).

Cervical Lordosis (CL)

Five studies reported global cervical lordosis.^{16,18,19,22,23} Both groups improved significantly: hybrid (MD –3.14°, 95% CI –4.7 to −1.5; P < 0.0002) and ACDF (MD –2.29°, 95% CI –4.5 to −0.06; P = 0.04). At the operated levels, hybrid produced greater lordotic gain (MD –7.5°, 95% CI –10.4 to −4.5; P < 0.0001) than ACDF (MD –5.6°, 95% CI –9 to −2.2; P = 0.001). Table 4.

Table 4.

Radiological outcomes at Baseline, Final Follow up and Mean Difference (MD) for Both Cohorts

	Hybrid at baseline	Hybrid at final follow up	P value, mean difference	ACDF baseline	ACDF final follow up	P value, mean difference
Cervical lordosis (degrees)	9.37 (9.77) n = 254	12.51 (8.79) n = 254	P = 0.0002 MD: −3.14 [−4.7601 to −1.5199]	7.02 (10.76) n = 146	9.31 (8.37) n = 146	P = 0.0433 MD: −2.29 [−4.5105 to −0.0695]
Lordosis (operated segment) (degrees)	5.72 (10.81) n = 95	13.24 (9.96) n = 95	P = 0.0001 MD: −7.52 [−10.4949 to −4.5451]	3.9 (10.57) n = 56	9.53 (7.15) n = 56	P = 0.0013 MD: −5.63 [−9.0095 to −2.2505]
C2-C7 SVA (cm)	3.29 (5.24) n = 190	3.88 (6.14) n = 190	P = 0.3143 MD: −0.59 [−1.7415 to 0.5615]	5.53 (9.24) n = 109	6.31 (9.93) n = 109	P = 0.5489 MD: −0.78 [−3.3407 to 1.7807]
T1 slope (degrees)	22.53 (6.42) n = 168	23.17 (7.28) n = 168	P = 0.3934 MD: −0.64 [−2.1131 to 0.8331]	22.91 (8.04) n = 89	22.44 (8.84) n = 89	P = 0.7110 MD: 0.47 [−2.0297 to 2.9697]

Sagittal Vertical Alignment (SVA) and T1 Slope

No significant changes were observed in either hybrid or ACDF cohorts for SVA (P = 0.31 vs 0.54) or T1 slope (P = 0.39 vs 0.71).^16,18,19,23 Table 4.

Adjacent Segment Degeneration (ASD)

Xu et al²³ found no difference in radiological ASD between hybrid and ACDF, at either the upper (17.6% vs 18.7%; P = 0.8) or lower segments (39.2% vs 33.3%; P = 0.2). Table 5.

Table 5.

Individual Complications Rates in Both Cohorts

	Hybrid	ACDF	P Value
Upper adjacent segment degeneration (1/9)	9/51 (17.6%)	6/32 (18.7%)	0.8989
Lower adjacent segment degeneration (1/9)	20/51 (39.2%)	17/32 (33.3%)	0.2147
Dysphagia (4/9)	33/189 (17.4%)	19/111 (17.1%)	0.57
Instrumentation failure (1/9)	1/19 (5.2%)	10/30 (33.3%)	0.0218
Prosthesis subsidence (2/9)	14/85 (16.5%)	1/45 (2.2%)	0.66
Total revision (3/9)	3/118 (2.5%)	1/56 (1.8%)	0.91
Heterotopic ossification (2/9)	35/85 (41.2%)	0/45 (0%)	0.02
Non union	8/103 (7.7%)	11/54 (20.3%)	0.3

Secondary Outcomes

Operative Time and Blood Loss

Eight studies^{16,17,18,19,20,21,23,24} reported operative duration, showing no significant difference (MD 3.79 min, 95% CI –4.99 to 12.58; P = 0.4). Hybrid procedures were associated with slightly greater intraoperative blood loss however a clinically negligible volume (MD 6.45 mL, 95% CI 2.26-10.6; P = 0.003). Figures 4 and 5.

Figure 4.

Forest Plot Showing the Operative Times (minutes) of Included Study, and Mean Difference Between Both Cohorts

Figure 5.

Forest Plot Showing the Blood Loss (mls) of Included Study, and Mean Difference Between Both Cohorts

Complications

Instrumentation failure was reported more frequently in ACDF in the limited number of studies that captured this outcome (33.3% vs 5.2%; P = 0.02). HO occurred in 41.2% of hybrid constructs, where it represents unintended restriction of arthroplasty motion; as expected, HO was not reported as a complication in ACDF because bone formation contributes to the fusion process. Rates of dysphagia (P = 0.57), prosthesis subsidence (P = 0.66), revision (P = 0.9), and non-union (P = 0.3) were similar between groups. Table 5.

HO was reported only in hybrid constructs, where it represents unintended restriction of arthroplasty motion. Because bone formation in ACDF reflects normal fusion rather than a complication, HO rates were not directly comparable between the 2 procedures.

Subgroup Analysis of Hybrid Types

Demographics

Patients with 2CDA + 1ACDF construct were younger than 1CDA + 2ACDF construct (49.7 vs 52.5 years; P = 0.005). Both subgroups most commonly operated level was C4–C7. Sex distribution and follow-up duration were comparable. Table 2.

Clinical Outcomes

No significant differences were seen in disability (NDI: MD 0.29, 95% CI –3.12 to 3.70; P = 0.87), myelopathy (mJOA: MD 0.46, 95% CI –0.26 to 1.18; P = 0.21), or ROM at operated and adjacent levels (all P > 0.5). The 2CDA +1ACDF cohorts showed greater improvement in neck pain (VAS: MD 0.58, 95% CI 0.06-1.0; P = 0.02), though this did not reach the threshold for clinical importance. Arm pain was similar (MD –0.34, 95% CI –1.01 to 0.33; P = 0.3).

Radiological Outcomes

Upper ASD was more frequent in the 2CDA + 1ACDF cohorts (40% vs 8.3%; P = 0.006), whereas lower ASD rates were similar (40% vs 38.8%; P = 0.9). The 1CDA + 2ACDF cohorts achieved greater global lordosis correction (MD –4.22°, 95% CI –7.0 to −1.4; P = 0.003), while both groups demonstrated significant lordotic improvement at operated segments (Type 1: MD –7.17°, P < 0.0001; Type 2: MD –8.3°, P = 0.005). SVA and T1 slope remained stable across both subgroups.

Secondary Outcomes

Operative time (MD 4.78 min, 95% CI –5.8 to 15.3; P = 0.38), blood loss (MD 2.48 mL, 95% CI –2.91 to 7.87; P = 0.37), and complication rates (dysphagia, prosthesis subsidence, revision, HO, non-fusion; all P > 0.2) did not differ significantly between both cohorts.

Sensitivity Analysis (≥2-Year Follow-Up)

When restricting the analysis to studies with ≥2 years of follow-up, only 2 studies^16,24 of the nine included studies were excluded, and the pooled results were consistent with the primary findings. Improvements in NDI and mJOA remained comparable between hybrid surgery and ACDF. Hybrid constructs continued to demonstrate superior operated-level ROM (MD similar direction), while adjacent-level ROM and cervical lordosis findings were unchanged. Confidence intervals widened due to fewer included studies, but no major shifts in effect estimates were observed. Complication trends, including higher heterotopic ossification in hybrid surgery and greater instrumentation failure in ACDF, also remained directionally consistent.

Discussion

In this systematic review and meta-analysis, we evaluated outcomes of three-level HS compared with three-level ACDF. Across nine retrospective studies and 704 patients, both procedures significantly improved pain, disability, and myelopathy scores. Hybrid constructs, however, stood out for preserving greater motion at the operated levels, while maintaining similar outcomes at adjacent segments. Both HS and ACDF improved cervical lordosis; however, neither approach significantly altered sagittal vertical alignment or T1 slope. Operative times were similar, though HS was associated with slightly higher blood loss which is clinically a negligible volume. Complication patterns differed: heterotopic ossification, which represents unintended loss of motion at the arthroplasty level, occurred only in hybrid constructs, whereas ACDF appeared to have a higher rate of instrumentation failure in the limited available data; however, this observation should be considered hypothesis-generating rather than definitive, given the small number of contributing studies. Within HS, patients with 1CDA+2ACDF and 2CDA+1ACDF performed similarly overall; while the 1CDA + 2ACDF construct offered greater global lordosis correction, while the 2CDA + 1ACDF construct showed a modest advantage in neck pain that did not reach clinical significance. Importantly, the ≥2-year sensitivity analysis demonstrated results that were consistent with the overall pooled findings. This suggests that the relative performance of hybrid surgery and ACDF remains stable over time. Although the smaller sample size reduced statistical precision, no clinically meaningful differences emerged between short-term and longer-term follow-up analyses.

From a clinical standpoint, these results are reassuring: patients undergoing either HS or ACDF for three-level cervical disease can expect similar improvements in function and pain. What distinguishes HS is its ability to preserve motion at the operated segments. This supports the long-standing biomechanical rationale for motion-preserving surgery, reducing stress at adjacent levels and potentially protecting them from accelerated degeneration.²⁸ These findings mirror what has been consistently reported in one- and two-level studies of CDA vs ACDF, where CDA maintains mobility and is linked to fewer adjacent-level reoperations over long-term follow-up, even though patient-reported outcomes converge over time.^29–31

Evidence specifically for three-level procedures is still limited, but the available studies align with our results. Hung and colleagues compared three-level CDA, HS, and ACDF and found that all procedures improved symptoms, but motion-preserving constructs better maintained cervical mobility.¹⁹ Xu et al and Zheng et al reported similar findings, CDA and HS preserved segmental motion but did not clearly outperform ACDF in terms of global sagittal alignment or overall patient outcomes.^18,23 Large meta-analyses and 10-year follow-up studies of CDA at one or 2 levels consistently show reduced adjacent-segment degeneration and reoperations compared with ACDF.³² Taken together, these data suggest that the motion-preserving benefits of CDA likely extend into multilevel surgery, although high-quality prospective studies at three levels are still lacking.

Both HS and ACDF improved global cervical lordosis in our pooled analysis, with HS offering slightly greater correction at the operated levels. This agrees with previous reports showing that hybrid constructs can restore or preserve alignment while maintaining motion.¹⁶ Importantly, global sagittal parameters such as SVA and T1 slope remained unchanged in both groups, reinforcing that the clinical benefit of these small radiographic differences is probably limited in the short to medium term.^16,18 When it comes to adjacent segments, we did not find significant differences in motion between HS and ACDF. This is notable, since reducing stress at adjacent levels is a major reason surgeons consider hybrid or arthroplasty-based strategies.³³ Some prior studies and biomechanical models have suggested that HS reduces adjacent segment motion compared with ACDF, but clinical evidence remains mixed.²⁵ Longer follow-up may be required to capture whether these kinematic differences translate into meaningful reductions in adjacent-segment disease.

The pattern of complications differed between procedures. As expected, HO was more common in HS, which reflects the well-documented biological response to arthroplasty devices.³⁴ While HO does not always cause clinical problems, in severe cases it may limit the very motion preservation that CDA is designed to achieve.³⁵ Careful surgical technique and appropriate prosthesis selection are important for mitigating this risk. On the other hand, ACDF was more often associated with instrumentation failure, an issue that becomes more pronounced as additional levels are fused and stress is concentrated across longer constructs. These trade-offs highlight the importance of tailoring the surgical plan to patient-specific factors such as age, bone quality, alignment goals, and activity level.

Looking within the HS cohort, the 1CDA + 2ACDF and 2CDA + 1ACDF constructs performed similarly overall, with only subtle differences. The 1CDA + 2ACDF construct achieved greater correction of global lordosis, whereas the 2CDA + 1ACDF construct tended to preserve more motion and showed a small advantage for neck pain. However, this advantage did not meet the threshold for clinical significance. Interestingly, the 2CDA + 1ACDF cohorts were associated with a higher incidence of upper adjacent segment degeneration, suggesting that the choice of which levels to replace vs fuse may influence long-term segmental loading. An important consideration is whether the position of the arthroplasty within the construct affects postoperative biomechanics and outcomes. Motion preservation may theoretically be more advantageous at certain levels; however, the included studies did not report outcomes based on CDA location, and we were therefore unable to determine whether upper-, middle-, or lower-level arthroplasty constructs performed differently. Future studies should stratify results by CDA position to clarify its influence on motion, adjacent segment loading, and complication rates.

An important source of potential bias is the statistically and clinically meaningful age difference between the hybrid and ACDF cohorts. Patients selected for hybrid constructs were, on average, younger, which likely reflects surgeon preference to offer arthroplasty to patients with better bone quality, fewer degenerative changes, and greater preserved mobility. This confounding by indication may have influenced both clinical outcomes and complication profiles, including the observed differences in motion preservation, heterotopic ossification, and hardware-related complications. As a result, differences between groups cannot be attributed solely to the surgical construct, and residual selection bias remains an important limitation of the available evidence.

Limitation

This review has several important limitations. First, all included studies were retrospective in design, which carries an inherent risk of selection bias, reporting bias, and confounding that cannot be fully controlled for. The absence of randomised controlled trials on three-level hybrid surgery vs ACDF makes it difficult to establish causal relationships, and the level of evidence therefore remains low. Second, although our pooled analysis included over 700 patients, some outcomes were reported by only a small number of studies. For example, instrumentation failure, which appeared to be more frequent in the ACDF group, was reported by a single study. This means the finding should be interpreted cautiously, as it may not be representative of wider clinical practice. Similarly, several complication rates and radiological outcomes were inconsistently reported, limiting the strength of conclusions in those areas. Third, the included studies varied in terms of surgical technique, implant design, follow-up duration, and baseline patient characteristics, which introduced heterogeneity and may have influenced the results. Fourth, because follow-up intervals were not standardized, we pooled outcomes at each study’s final follow-up, which averaged around four years. This duration may still be too short to detect longer-term issues such as adjacent segment degeneration or device-related complications that often arise after 5-10 years. Additionally, the exact position of the CDA within the hybrid construct was rarely specified, preventing analysis of whether arthroplasty at upper, middle, or lower levels yielded different outcomes. Reoperation data were sparsely reported, and no study distinguished index-level from adjacent-level reoperations, limiting our ability to assess an important long-term outcome. Finally, publication bias cannot be excluded, as the number of studies was too few to allow formal assessment with funnel plots. Taken together, these limitations suggest that while our findings provide useful preliminary insights, they should be viewed as hypothesis-generating. High-quality prospective studies with larger sample sizes and long-term follow-up are essential to validate the comparative effectiveness and safety of three-level HS and ACDF.

Recommendation for Future Researchers

Future research on three-level hybrid surgery should prioritise prospective, adequately powered comparative studies with clearly defined inclusion criteria, robust adjustment for baseline confounders, and transparent reporting of follow-up rates. Standardising outcome assessment at clinically meaningful time points (eg, 6, 12, and 24 months, plus long-term follow-up) would improve comparability across studies and help clarify the temporal patterns of recovery and alignment changes. Detailed reporting of hybrid construct design is also essential. Future studies should specify not only whether 1 or 2 arthroplasty levels were used but also which cervical levels received CDA, as variation in upper, middle, or lower placement may influence biomechanics, motion preservation, and adjacent segment behaviour.

In addition, future investigations should systematically report reoperation rates at both the index and adjacent levels, as this is a key rationale for motion-preserving strategies yet remains poorly documented. Radiological assessment would benefit from standardised definitions for adjacent segment degeneration, heterotopic ossification, and sagittal alignment metrics, alongside validated patient-reported outcomes. Long-term follow-up (≥5-10 years) is needed to determine the durability of hybrid constructs and to evaluate whether their theoretical benefits translate into sustained clinical advantages and reduced adjacent segment disease. Collectively, these improvements will generate higher-quality evidence to guide optimal construct selection in multilevel cervical surgery.

Conclusion

From the retrospective evidence currently available, hybrid surgery and three-level ACDF seem to deliver similar short-to mid-term improvements in pain, disability, and myelopathy. However, these findings should be viewed with caution. Many of the included studies were limited by small sample sizes, incomplete reporting of follow-up, and a lack of adjustment for important differences between patients. As a result, our confidence in the pooled results is necessarily limited. Stronger prospective studies with clearer reporting and better control of confounding factors are needed to more reliably determine how hybrid constructs compare with multilevel ACDF.

Footnotes

ORCID iDs

Chinedu Egu

Elie Najjar

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Fejer

Hartvigsen

. The prevalence of neck pain in the world population: a systematic critical review of the literature. Eur Spine J. 2006;15(6):834-848. doi:10.1007/s00586-004-0864-4

Hoy

Protani

Buchbinder

. The epidemiology of neck pain. Best Pract Res Clin Rheumatol. 2010;24(6):783-792. doi:10.1016/j.berh.2011.01.019

Smith

Robinson

. The treatment of certain cervical-spine disorders by anterior removal of the intervertebral disc and interbody fusion. J Bone Joint Surg Am. 1958;40(3):607-624.

Bernhardt

. Cervical spondylotic myelopathy: a review of surgical indications and decision making. Yale J Biol Med. 1993;66(3):165-177.

Hilibrand

Carlson

Palumbo

Jones

Bohlman

. Radiculopathy and myelopathy at segments adjacent to the site of a previous anterior cervical arthrodesis. J Bone Joint Surg Am. 1999;81(4):519-528. doi:10.2106/00004623-199904000-00009

Davies

Mowforth

Smith

Kotter

. Degenerative cervical myelopathy. Br Med J. 2018;360:186. doi:10.1136/bmj.k186

Eck

Humphreys

Lim

, et al. Biomechanical study on the effect of cervical spine fusion on adjacent-level intradiscal pressure and segmental motion. Spine. 2002;27(22):2431-2434. doi:10.1097/00007632-200211150-00003

Emery

Bohlman

Bolesta

Jones

. Anterior cervical decompression and arthrodesis for the treatment of cervical spondylotic myelopathy. Two to seventeen-year follow-up. J Bone Joint Surg Am. 1998;80(7):941-951. doi:10.2106/00004623-199807000-00002

Wang

McDonough

Kanim

Endow

Delamarter

. Increased fusion rates with cervical plating for three-level anterior cervical discectomy and fusion. Spine. 2001;26(6):643-646. doi:10.1097/00007632-200103150-00015

10.

Mummaneni

Burjus

Haid

Zdeblick

. Clinical and radiographic analysis of cervical disc arthroplasty compared with allograft fusion: a randomized controlled clinical tria. J Neurosurg Spine. 2007;6(3):198-209. doi:10.3171/spi.2007.6.3.198

11.

Radcliff

Coric

Albert

. Five-year clinical results of cervical total disc replacement compared with anterior discectomy and fusion for treatment of 2-level symptomatic degenerative disc disease: a prospective, randomized, controlled, multicenter investigational device exemption clinical trial. J Neurosurg Spine. 2016;25(2):213-224. doi:10.3171/2015.12.SPINE15824

12.

Gornet

Lanman

Burjus

, et al. Two-level cervical disc arthroplasty versus anterior cervical discectomy and fusion: 10-Year outcomes of a prospective, randomized investigational device exemption clinical trial. J Neurosurg Spine. 2019;31(4):508-518. doi:10.3171/2019.4.SPINE19157

13.

Skovrlj

Lee

Caridi

Cho

. Reoperation rates after anterior cervical discectomy and fusion versus cervical disc replacement: a meta-analysis of prospective randomized controlled trials. Spine. 2014;39(9):769-776. doi:10.1007/s11999-016-4707-5

14.

Coric

Nunley

Guyer

, et al. Prospective, randomized, multicenter study of cervical arthroplasty: 269 patients from the kineflex|C artificial disc investigational device exemption study with a minimum 2-year follow-up: clinical article. J Neurosurg Spine. 2011;15(4):348-358. doi:10.3171/2011.5.SPINE10769

15.

Joaquim

Riew

. Multilevel cervical arthroplasty: current evidence. A systematic review. Neurosurg Focus. 2017;42(2):E4. doi:10.3171/2016.10.FOCUS16354

16.

Chen

Deng

Liu

, et al. Cervical sagittal balance after consecutive three-level hybrid surgery versus anterior cervical discectomy and fusion: radiological results from a single-center experience. J Orthop Surg Res. 2023;18(1):345. doi:10.1186/s13018-023-03819-

17.

Huang

Wang

Liu

, et al. Classification of three-level hybrid surgery for the treatment of cervical degenerative disc disease: a retrospective study of 108 patients. BMC Surg. 2022;22(1):179. doi:10.1186/s12893-022-01627-7

18.

Zheng

, et al. Sagittal sequence and clinical efficacy of cervical disc replacement and hybrid surgery in the treatment of cervical spondylotic myelopathy: a retrospective study. Front Surg. 2024;10:1265349. doi:10.3389/fsurg.2023.1265349

19.

Hung

Kao

. Comparison of sagittal parameters for anterior cervical discectomy and fusion, hybrid surgery, and total disc replacement for three levels of cervical spondylosis. Clin Neurol Neurosurg. 2018;168:140-146. doi:10.1016/j.clineuro.2018.03.003

20.

Jang

Lee

Cho

. A comparison of anterior cervical discectomy and fusion versus fusion combined with artificial disc replacement for treating 3-Level cervical spondylotic disease. J Korean Neurosurg Soc. 2017;60(6):676-683. doi:10.3340/jkns.2016.1010.013

21.

Kang

Lin

Ding

Liang

Lian

. Artificial disk replacement combined with midlevel ACDF versus multilevel fusion for cervical disk disease involving 3 levels. Orthopedics. 2013;36(1):e88-e94. doi:10.3928/01477447-20121217-24

22.

Wang

Huang

Meng

Wang

Ding

Hong

. Is cervical disc replacement valuable in 3-Level hybrid surgery compared with 3-Level anterior cervical discectomy and fusion? World Neurosurg. 2021;146:e151-e160. doi:10.1016/j.wneu.2020.10.057

23.

Liang

Wang

Zhu

Liu

. Cervical spine balance of multilevel total disc replacement, hybrid surgery, and anterior cervical discectomy and fusion with A long-term Follow-up. Spine. 2020;45(16):E989-E998. doi:10.1097/BRS.0000000000003474

24.

Chen

Liu

, et al. Overloaded vertebral body following consecutive three-level hybrid surgery comparing with anterior cervical discectomy and fusion. Orthop Surg. 2024;16(12):3036-3046. doi:10.1111/os.14242

25.

Yang

Liang

Sun

, et al. Comparison of hybrid surgery and two-level ACDF in treating consecutive cervical degenerative disc disease: a systematic review and meta-analysis. Glob Spine J. 2025;15:21925682251343524-21925682251343968. doi:10.1177/21925682251343524

26.

Moher

Liberati

Tetzlaff

Altman

PRISMA Group . Preferred reporting items for systematic reviews and Meta analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097. doi:10.1371/journal.pmed.1000097

27.

National Institutes of Health . Quality assessment tool for observational cohort and cross-sectional studies. 2014. https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort

28.

Meng

Rong

Hong

Ding

Chen

Liu

. Biomechanics following skip-level cervical disc arthroplasty versus skip-level cervical discectomy and fusion: a finite element-based study. BMC Muscoskelet Disord. 2019;20(1):49. doi:10.1186/s12891-019-2425-3

29.

Kim

Hoffman

Bae

, et al. Ten-year outcomes of 1- and 2-Level cervical disc arthroplasty from the Mobi-C investigational device exemption clinical trial. Neurosurgery. 2021;88(3):497-505. doi:10.1093/neuros/nyaa459

30.

Nunley

Hisey

Smith

Stone

. Cervical disc arthroplasty vs anterior cervical discectomy and fusion at 10 years: results from a prospective, randomized clinical trial at 3 sites. Int J Spine Surg. 2023;17(2):230-240. doi:10.14444/8431

31.

Radcliff

Davis

Hisey

, et al. Long-term evaluation of cervical disc arthroplasty with the Mobi-C© cervical disc: a randomized, prospective, multicenter clinical trial with seven-year Follow-up. Int J Spine Surg. 2017;11(4):31. doi:10.14444/4031

32.

Dong

Chen

, et al. The change of adjacent segment after cervical disc arthroplasty compared with anterior cervical discectomy and fusion: a meta-analysis of randomized controlled trials. Spine J. 2017;17(10):1549-1558. doi:10.1016/j.spinee.2017.06.010

33.

Bhatt

Orosz

Schuler

, et al. Anterior cervical hybrid constructs reduce superior adjacent segment burden compared to multilevel anterior cervical discectomy and fusion. J Spine Surg. 2024;10(2):165-176. doi:10.21037/jss-23-135

34.

Yang

Zhao

Chai

, et al. A comparison study between hybrid surgery and anterior cervical discectomy and fusion for the treatment of multilevel cervical spondylosis. Bone Jt J. 2020;102(8):981-996. doi:10.1302/0301-620X.102B8

35.

Turkov

Job

Iturriaga

Verma

. Current concepts of cervical disc arthroplasty. Int J Spine Surg. 2022;15(6):1174-1183. doi:10.14444/8149