Correction and Maintenance of Cervical Alignment

3-Level ACDF Versus Corpectomy-ACDF “Hybrid” Procedures

Abstract

Study Design: 

Longitudinal observational comparative cohorts.

Objective: 

The objective of this study is to compare 3 and 12-month radiographic sagittal parameters and patient-reported outcomes (PROs) in patients who underwent 3-level ACDF or a hybrid procedure.

Summary of Background Data: 

Anterior Cervical Discectomy and Fusion (ACDF), Anterior Cervical Corpectomy and Fusion (ACCF), and hybrids (combination ACCF-ACDF) are common procedures used to treat symptomatic cervical spondylosis. Although there is a relative abundance of literature comparing 1-level ACCF versus 2-level ACDF and 2-level ACCF versus 3-level ACDF, detailed comparisons of 3-level ACDF versus hybrid procedures have not been extensively addressed.

Methods: 

Patients who underwent a 3-Level ACDF (3L-ACDF, N=47) or 1-Level Corpectomy/1-Level ACDF (Hybrid, N=52) with at least a 12-month post-op data available were identified. Standard demographic, surgical and PROs were collected in addition to preoperative and postoperative radiographic data, including C2 plumb line (C2PL), C2–C7 lordosis (CL), segmental lordosis (SL), and T1 slope (T1S).

Results: 

The 2 cohorts were similar in terms of demographics. At 3 months post-op, CL (9.04° vs. −2.12°, P=0.00) and SL (6.06° vs. −2.26°, P=0.003) were significantly greater in the 3L-ACDF group versus the HYBRID group. This significant difference was maintained at 12 months postoperative for CL [(6.62° vs. −0.60°, P=0.015) but not for SL (2.36° vs. −1.09°, P=0.199)]. There were no differences in PROs between the 2 groups before surgery, at 3 months postoperative or 12 months postoperative. Seven patients required revision surgery in the 1-year study period (1 in the 3L-ACDF, and 6 in the Hybrid P<0.001).

Conclusions: 

Three level ACDF resulted in greater C2–C7 lordosis and segmental lordosis postoperatively, which was maintained at 1 year for cervical lordosis. While PROs were similar between the groups, patients with hybrid instrumentation required significantly more revision surgeries than those treated with 3-level ACDF.

Anterior Cervical Discectomy and Fusion (ACDF) and Anterior Cervical Corpectomy and Fusion (ACCF) are common surgical procedures for the treatment of cervical spine pathology, including Cervical Degenerative Myelopathy and Cervical Spondylotic Myelopathy (CSM), with or without cervical radiculopathy.^1,2 In the treatment of multilevel disease from an anterior approach, one can use either a multilevel ACDF, multilevel ACCF, or a combination of the 2, a so-called “HYBRID” construct (Fig. 1).³

FIGURE 1: Lateral radiograph of the cervical spine showing a Hybrid construct with C4-5 ACDF and C6 corpectomy.

While there is a relative abundance of literature comparing 1-level ACCF versus 2-level ACDF and 2-level ACCF versus 3-level ACDF, detailed comparisons of 3-level ACDF versus HYBRID procedures have not been reported. Hilibrand et al⁴ showed that multilevel ACDF had a higher nonunion rate, while ACCF had a higher rate of graft dislodgment and revision surgery. Several authors have reported shorter OR times, less blood loss, and better restoration of cervical lordosis with ACDF versus ACCF.^3,5,6 Others have shown a correlation between the restoration of the cervical sagittal vertical axis(SVA) with improvement in patient-reported outcomes(PRO).^3,7

The objective of this study was to compare PROs and radiographic parameters in patients who underwent 3-level ACDF versus HYBRID procedures.

METHODS

Our institutional database was queried for patients who underwent ACDF, ACCF, and hybrid procedures from 2015–2020. Inclusion criteria were patients over the age of 18 receiving anterior cervical fusion for 3 motion segments. Exclusion criteria were previous cervical spine surgery, acute trauma and lack of available preoperative or postoperative radiographs sufficient for evaluation of radiographic parameters.

Using our institution’s electronic medical record system, the charts of patients meeting criteria were reviewed for demographic information, as well as estimated blood loss, history of diabetes and smoking history. Preoperative, postoperative (0–3 mo) and 1 year (9–15 mo) radiographs were analyzed using Picture Archiving and Communication System; to determine the cervical sagittal vertical axis (SVA), C2–C7 lordosis, segmental lordosis, and T1 slope. SVA was recorded as positive when the C2 plumb line fell anterior to the posterior, superior endplate of C7. Cobb measurements for cervical lordosis, segmental lordosis, and T1 slope were recorded as positive for lordotic and negative for kyphotic alignment, respectively.

Patient charts were also evaluated for PRO data, including the Neck Disability Index,^8,9 Numeric rating scales (0–10) for Neck and Arm Pain,¹⁰ Euro-QOL 5D–3-Level (EQ-5D),¹¹ and reoperation of the involved motion segments within the 1 year study period. Indications for return to OR were based on operative indications as detailed in the surgeon's operative report.

All statistical analyses were performed using IBM SPSS v27.0 (Somes, NY). Continuous variables were compared between the 2 cohorts using unpaired independent t tests. Categorical variables were compared using Fisher exact test. A P-value threshold of 0.05 was deemed statistically significant.

RESULTS

Within a single-institution, multi-surgeon (n=12) database, 99 patients met inclusion criteria: 52 patients in the hybrid cohort were compared with 47 patients in the 3-L ACDF cohort. All patients underwent a single-stage anterior exposure for spinal cord decompression and reconstruction with instrumented fusion, including anterior cervical plating.

Of the 47 3-level ACDF procedures, 24 used titanium cages, 15 polyether ether ketone (PEEK), 7 allograft, and 1 titanium coated PEEK. In the HYBRID cohort, all corpectomy levels were reconstructed with static titanium interbody devices, while at the ACDF levels, 20 used PEEK, 12 titanium, 10 titanium coated PEEK, 9 structural allograft, and 1 carbon fiber. The estimated blood loss [(3-L ACDF: 79.74 cc vs. Hybrid: 144.38 cc, P=0.054) and operative time ((3-L ACDF: 154.s vs. Hybrid: 212.43 min, P=0.108)] were similar between the 2 cohorts.

The 2 cohorts were similar in terms of demographic and radiographic measures at baseline. As seen in Table 1, at 3 months postoperative, cervical lordosis [(3-L ACDF: 9.04° vs. Hybrid: −2.12°, P=0.00) and segmental lordosis (3-L ACDF: 6.06° vs. Hybrid: −2.26°, P=0.003)] were significantly greater in the 3L-ACDF group than in the HYBRID group. This was maintained at 12 months postoperative for cervical lordosis (3-L ACDF: 6.62° vs. Hybrid= −0.60°, P=0.015) but not for segmental lordosis (3-L ACDF: 2.36° vs. Hybrid:−1.09°, P=0.199). There were no significant differences, at any time point, between the 2 cohorts in cervical SVA or T1 slope.

TABLE 1 - Preoperativeand Postoperative Radiographic Outcomes

Variable	3-L ACDF	Hybrid	P
Cervical SVA (mm)
Preoperative	2.96 (1.63)	2.93 (1.53)	0.924
3 mo Postoperative	2.82 (1.33)	2.92 (1.42)	0.725
12 mo Postoperative	3.09 (1.56)	3.06 (1.64)	0.934
C2–C7 Lordosis (degree)
Preoperative	2.37 (13.47)	−2.54 (11.59)	0.056
3 mo Postoperative	9.04 (14.67)	−2.12 (13.58)	0
12 mo Postoperative	6.62 (14.13)	−0.6 (14.96)	0.015
Segmental Lordosis (degree)
Preoperative	0.28 (10.39)	−0.75 (9.93)	0.615
3 mo Postoperative	6.06 (14.34)	−2.26 (11.74)	0.003
12 mo Postoperative	2.36 (13.24)	−1.09 (13.3)	0.199
T1 Slope (degree)
Preoperative	31.48 (10.4)	30.29 (13.32)	0.644
3 mo Postoperative	34.17 (8.42)	31.01 (13.77)	0.206
12 mo Postoperative	33.76 (8.63)	32.02 (12.33)	0.444

Seven patients required revision surgery in the 1 year study period (1 in the 3L-ACDF, and 6 Hybrid P<0.001) (Table 2). In the 3L-ACDF patient, the reoperation was at 4 months for pseudarthrosis with pullout of the plate and screws. Three of the HYBRID patients had posterior decompression and fusions performed for pseudarthrosis with loss of fixation and hardware migration. Two of these were within 1 month, and the third at 9 months postoperatively (Fig. 2). The other 3 patients underwent posterior decompression and fusions for pseudarthrosis at 6 months in 1 patient, and 1 year in the other 2. The number of patients requiring subsequent surgery in the 3L-ACDF (n=1) versus the HYBRID (n=6) was statistically significant between cohorts. There were no differences in PROs between the 2 groups before surgery or at any time period postoperatively (Table 3).

TABLE 2 - Revision Surgeries (Reoperations) Within 1 Year

Initial Procedure	Reason For Revision Surgery	Time To Revision Surgery
3-Level ACDF	Pseudarthrosis with loss of fixation	4 mo
Hybrid	Loss of Fixation	1 mo
Hybrid	Loss of Fixation	1 mo
Hybrid	Loss of Fixation	9 mo
Hybrid	Pseudarthrosis	6 mo
Hybrid	Pseudarthrosis	1 y
Hybrid	Pseudarthrosis	1 y

FIGURE 2: A, 65-year-old man with progressive symptoms of CSM. B, Patient underwent a Hybrid procedure and did well postoperatively, with the improvement of preoperative symptoms. C, At 6 months postoperatively, the patient complained of continued axial neck and radicular pain. A CT obtained showed subsidence of the corpectomy cage through the caudal endplate and pseudarthrosis.

TABLE 3 - Patient-Reported Outcomes

VARIABLE	3L-ACDF	Hybrid	P
Neck disability index	Mean (SD)	Mean (SD)
Preoperative	52.59 (20.42)	45.81 (21.19)	0.194
3 mo Postoperative	43.55 (21.3)	35.48 (22.57)	0.127
12 mo Postoperative	39.77 (21.35)	34.87 (26.46)	0.366
Neck Pain
Preoperative	6.26 (2.46)	5.67 (2.45)	0.339
3 mo Postoperative	4.55 (2.71)	3.54 (2.87)	0.136
12 mo Postoperative	4.28 (3)	4.18 (3.32)	0.887
Arm Pain
Preoperative	5.68 (3.08)	5.81 (3.29)	0.874
3 mo Postoperative	2.87 (2.69)	2.33 (2.93)	0.428
12 mo Postoperative	3.84 (2.88)	3.14 (3.21)	0.327
Euro-QOL
Preoperative	0.68 (0.27)	0.66 (0.27)	0.733
3 mo Postoperative	0.71 (0.23)	0.79 (0.24)	0.179
12 mo Postoperative	0.74 (0.26)	0.76 (0.28)	0.708

DISCUSSION

The decision to treat multilevel cervical pathology with multilevel ACDF or corpectomy is one often faced by spine surgeons. As previously discussed, multilevel ACDF has previously been shown to result in greater restoration of lordosis, reduced operative time, blood loss and have a lower incidence of unplanned reoperations.^3,5,6 Despite these advantages, others have shown an increased risk of pseudarthrosis because of the increased number of fusion surfaces in multilevel ACDF.⁴ Several authors have previously compared outcomes of multilevel ACDF and ACCF, but there is a lack of evidence regarding outcomes of 3-level ACDF compared with HYBRID procedures.

In our study comparing 3L-ACDF with HYBRID procedures, we found that multilevel ACDF achieved greater overall cervical lordosis, and segmental lordosis, compared with HYBRID procedures. The superiority in cervical lordosis remained significant at 1 year, while segmental lordosis did not. Throughout the study period, there was no statistical difference in cervical SVA between cohorts.

Despite the differences in lordosis achieved, there were no statistically significant differences in PROs between the groups. Tang et al and Villavicencio et al showed that improvement or maintenance of cervical sagittal parameters was associated with improved PROs.^12,13 The discrepancy between their results and our study can possibly be explained by Sakai et al,¹⁴ who showed specifically, that improved cSVA is directly correlated with improved HR-QOL scores.

The lack of difference in PROs in our study may be because of the finding that cervical SVA and T1 slope were not significantly different between groups, or that factors other than cervical alignment are driving PROs. Another possibility is that the average alignment parameters of both cohorts fall within the normal parameters of asymptomatic individuals as described by Hardacker et al^15. Perhaps statistically and clinically significant differences in PROs may be difficult to demonstrate, or altogether nonexistent, between groups with sagittal parameters within the normal range.

In a meta-analysis comparison of ACDF versus Hybrid, Shamji et al¹⁶ showed no difference between ACDF and Hybrid groups in JOA or Neck Disability Index outcomes, or sagittal radiographic parameters. In their retrospective review, Liu et al¹⁷ similarly showed no differences in PROs, sagittal alignment, or postoperative complications.

A major difference between 3L-ACDF and HYBRIDS shown in our study was the significantly higher rate of reoperations in the HYBRID group. While 50% of these were for hardware and graft-related complications within 1 year, all reoperations were associated with pseudarthrosis. It appears in our patients that early reoperation was associated with loss of hardware fixation, while later reoperations were for treatment of symptomatic pseudarthrosis. Though higher revision surgery in corpectomies was previously described by Hilibrand et al,⁴ they also described a lower rate of pseudarthrosis in corpectomies. We did not separately assess fusion rate in the cohorts and instead showed that pseudarthrosis in HYBRID procedures is more symptomatic and prone to failure than in multilevel ACDF.

In their biomechanical study comparing the stability of a 2-level ACDF to 1-level ACCF in cervical models, Ouyang et al.¹⁸ found that 2-level ACDF showed greater stress on the anterior plate and screw fixation while decreasing the stress on the inferior endplate of the cephalad level. Interestingly, there was only 1 (2%) case of hardware-related complications or reoperations in the 3L-ACDF cohort, compared with 3(6%) in the HYBRID cohort. While not measured in our study, the HYBRID patients with reoperation-associated hardware-related complications were often associated with subsidence of the corpectomy cage.

Both ACDF and HYBRID procedures are effective treatments of cervical radiculopathy and CSM. Our study has shown comparable outcomes, both radiographically and in patient-reported outcomes between the 2 procedures. Though a corpectomy can provide greater access to retrovertebral pathology, we have shown that one should keep in mind the potentially greater correction of lordosis and decreased rate of reoperation with 3-level ACDF.

There are several strengths to this study. First, the large sample size with multi-surgeon cohorts makes the study more representative of a larger population because of the range of surgical techniques and patient demographics. One year radiographic follow-up also allows for assessment of the construct durability and PROs over the expected fusion time period.

The main limitation of this study is the inherent patient selection bias between the cohorts. A cervical corpectomy allows for the treatment of retrovertebral compression that may otherwise be difficult to address through a standard ACDF. This difference in pathology could also be associated with differences in spinal cord compression and even the muscular stability of the cervical spine.

A second limitation of the study is the large number of patients excluded because of the lack of adequate postoperative follow-up radiographs. This is a consequence of differences in surgeon preference for radiographic follow-up in asymptomatic patients. This lack of follow-up of asymptomatic patients may overestimate the reoperation rate in our study. Another limitation of our study is the lack of definitive determination of fusion, as previously discussed. Lastly, we did not assess differences in neurological improvement between the cohorts.

3L-ACDF is able to achieve and maintain greater lordosis of the operative levels than HYBRID procedures while maintaining equivalent outcomes. Given the higher revision and return to OR rate among HYBRID procedures in our study, we recommend evaluating patient bone quality, construct stability, and pseudarthrosis risk factors. If there is a concern for the loss of fixation or a high likelihood of pseudarthrosis, a posterior procedure can be considered as a primary or secondary procedure in patients with retrovertebral compression.

Areas of future interest may be definitively assessing fusion rates of the respective constructs, and evaluating differences in patient and implant characteristics in cases requiring revision surgery. Longer follow-up may also show differences in PROs that were not detected in the 1-year time period.