Secondary Logo

Cages in ACDF are Associated With a Higher Nonunion Rate Than Allograft

A Stratified Comparative Analysis of 6130 Patients

Pirkle, Sean, BA; Kaskovich, Samuel, BSA; Cook, David J., BA, BEng; Ho, Alisha, BA; Shi, Lewis L., MD; Lee, Michael J., MD

doi: 10.1097/BRS.0000000000002854

Study Design. A retrospective database review.

Objective. The purpose of this study was to analyze the rate of nonunion in patients treated with structural allograft and intervertebral cages in anterior cervical discectomy and fusion (ACDF).

Summary of Background Data. Existing literature consists primarily of single-center studies with inconsistent findings.

Methods. We performed a retrospective analysis of 6130 patients registered in the PearlDiver national database through Humana Insurance from 2007 to 2016. All ACDF patients with anterior plating who were active in the database for at least 1 year were included in the study. Patients with a fracture history within 1 year of intervention, past arthrodesis of hand, foot, or ankle, or a planned posterior approach were excluded from the study. Patients were stratified by number of levels treated, tobacco use, and diabetic condition. Nonunion rates of structural allograft and intervertebral cage groups after 1 year were compared using Chi-squared analyses.

Results. Four thousand sixty-three patients were included in the allograft group, while 2067 were included in the cage group. Overall nonunion rates were significantly higher in the cage group (5.32%) than in allograft group (1.97%) (P < 0.01). When controlling for confounders, increased rates of nonunion were consistently observed in the cage group, achieving statistical significance in 25 of the 26 analyses.

Conclusion. The increased rate of nonunion associated with intervertebral cages may suggest the superiority of allograft over cages in ACDF.

Level of Evidence: 3

The long-term success of anterior cervical discectomy and fusion is associated with solid fusion. In this study, we found that both anterior cervical discectomy and fusion with structural allograft and cage have high fusion rates. However, these data suggest that the use of a cage is associated with a significantly higher rate of nonunion as compared to structural allograft.

Department of Orthopaedic Surgery and Rehabilitation Medicine, University of Chicago Medical Center, Chicago, IL.

Address correspondence and reprint requests to Sean Pirkle, BA, 5120 S Greenwood Ave, Apt #2, Chicago, IL 60615; E-mail:

Received 30 May, 2018

Revised 1 August, 2018

Accepted 9 August, 2018

The device(s)/drug(s) is/are FDA-approved or approved by corresponding national agency for this indication.

No funds were received in support of this work.

Relevant financial activities outside the submitted work: grants.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (

Anterior cervical discectomy and fusion (ACDF) is widely recognized as a highly successful surgical treatment of cervical radiculopathy and myelopathy.1,2 Anterior decompression alone of the neural elements may result in spinal instability and pain, and thus, a concurrent arthrodesis is performed to achieve stability. The technique, initially performed with harvesting of autologous bone graft from the ilium, has since been modified by the utilization of allograft, and more recently, intervertebral cages with bone graft material. The importance of achieving fusion is generally agreed upon3 in that the failure to achieve fusion is associated with a higher likelihood of symptoms and a higher likelihood for revision surgery.

Although allograft does not share the same osteoinductive and osteogenetic properties as autologous iliac crest bone graft, the distinct advantage of the use of allograft is the avoidance of the morbidity of iliac crest bone graft harvest. Allograft, despite its biological inferiority to autograft, does retain osteoconductivity and has been demonstrated in the literature to have high fusion rates.4 Intervertebral cages in the cervical spine are designed to provide structural support between vertebral bodies and to allow bone fusion to occur within and around the cage. However, the cage itself carries no biological properties to promote fusion formation. Furthermore, the cage itself does occupy surface area and intervertebral volume for fusion mass to form that would otherwise have been occupied by a structural graft, auto or allo. This may decrease the likelihood of achieving solid arthrodesis as compared to structural bone graft. Despite these theoretical concerns, there has been a widespread enthusiasm for the use of intervertebral cages in anterior cervical discectomy and arthrodesis. Yoon et al, 3 in an international survey of spine surgeons, reported that the majority (64%) utilize cages as the structural graft component. In North America, 36% of surgeons reported using cages in ACDF.3

Although there is extensive literature on fusion rates in ACDF, there are few studies comparing fusion rates in ACDF when using a cage versus a structural bone graft.5–9 In addition, as fusion rates are fairly high for ACDF in general, small comparative studies are unlikely to have sufficient power to ascertain whether a difference in fusion rate exists or not. A higher sample size is needed to statistically demonstrate a significant difference or nondifference in fusion rates between the two techniques.

Using the PearlDiver database, we hypothesized that the use of an intervertebral cage in ACDF would be associated with a higher nonunion rate when compared with the fusion rates of ACDF using structural bone graft.

Back to Top | Article Outline


We performed a retrospective review of the Humana subset of the PearlDiver Patient Record Database. This is a commercially available dataset with millions of records from the nationwide Humana health insurance provider. In this dataset, queries can be conducted by Current Procedural Terminology (CPT) coding and International Classification of Diseases (ICD) 9th and 10th revision coding.

We utilized CPT 22551 and queried for patients who had undergone ACDF. Specifically, we selected for patients who had undergone (1) ACDF (22551), (2) structural allograft only (20931) or allograft with cage placement (22851), and (3) anterior instrumentation (22845, 22846) (Figure 1). In this population, we selected only patients who had at least 1-year follow-up after their index ACDF procedure. Any use of autologous iliac crest bone graft in ACDF was excluded from this study.

Figure 1

Figure 1

As our goal was to examine the effect of the intervertebral cage, we excluded patients who had a posterior arthrodesis (22600) within 3 months of their ACDF, as this may have been a part of a planned staged procedure. We also excluded patients who may have had other concurrent orthopedic arthrodesis CPT codes in this perioperative period as well. In addition, we excluded patients with an ICD-9 or ICD-10 coding of fracture (Supplementary Table 1, The reason for these exclusions was to ensure to ensure that a “nonunion” code was in reference to the ACDF and not another source such as fracture healing or an alternate body site of fusion.

After our initial univariate analysis, we then performed a stratified analysis controlling for one significant variable (Table 1), then two (Table 2), and then three significant variables (Table 3). In order to determine differences in nonunion rates between the no cage and cage groups, we conducted Chi-squared tests of independence with an alpha level of 0.05 in all of our analyses.







Back to Top | Article Outline


We identified 6130 patients who underwent ACDF with cage or allograft, anterior plating, and without concurrent or delayed posterior arthrodesis in the Humana subsection of the PearlDiver data registry. Of these, 66% (4063) were treated with allograft and 34% (2067) were treated with a cage device. The distribution of tobacco users, diabetics, and different levels of surgery were not observed to be statistically different between the cage group and allograft group (P > 0.05).

In our univariate analysis of nonunion, we observed that patients with a cage experienced a rate of 5.32% and those with allograft 1.97% (P < 0.0001). We also observed statistically increased rates of nonunion for diabetics, multiple levels, and tobacco use (P < 0.0001). In all subanalyses controlling for confounding variables, we observed a higher rate of nonunion with the use of the cage as compared to the allograft group. When controlling for one additional variable (Table 1), we observed that the use of the cage was significantly associated with a higher rate of nonunion in all subanalyses.

When controlling for two additional variables, we observed the use of the cage to be significantly associated with a higher rate of nonunion in 12 of the 13 possible subanalyses (Table 2). Three subanalyses were not possible (three-level ACDF/Tobacco −; three-level ACDF Diabetes −; Tobacco −/Diabetes −) group because the number of nonunions was less than 11. For patient privacy compliance in the PearlDiver data, when the sample size is less than 11, the true value is rendered as “−1.” In these cases, the number may be anywhere between 1 and 11 patients. In some cases, we were able to use the 11 values as a conservative estimate and still perform our analysis. For example, in the two-level ACDF/Tobacco − group, there were 11 or less patients with nonunion in the allograft group, whereas there were 22 patients with nonunion in the cage group. In this case, we assigned the highest possible nonunion rate (11 patients) in the allograft group and still observed a statistically significant difference between the allograft and cage group.

When controlling for three additional variables (Table 3), we observed a statistically higher nonunion rate in the cage group in all subanalyses that were possible. Of the 12 possible permutations, only five subanalyses had sufficient data to allow for analysis. The remaining seven subanalyses had nonunion numbers of less than 11 in both groups. As the true numerator is not known in both groups, we could not estimate a comparison in these subanalyses.

Back to Top | Article Outline


In anterior cervical discectomy and arthrodesis, the achievement of stability is of obvious importance for the maintenance of long-term benefit. Nonunion has been associated with poor clinical outcome and the need for revision surgery,10–12 whether posterior or anterior, presents an additional risk to the patient.

The use of an intervertebral cage in anterior discectomy and arthrodesis has gained popularity.3 Although the nonresorbable nature of the cage may allow for stability as the fusion mass forms, it may also represent a literal mechanical block for fusion formation. Fusion mass cannot form in the space occupied by the synthetic cage. With less endplate surface area and less intervertebral volume available for arthrodesis, we hypothesized that the use of a cage in ACDF would be significantly associated with the development of a nonunion.

Our study using the Humana data within the PearlDiver registry suggests that the use of an intervertebral cage in ACDF is statistically associated with a higher nonunion rate as compared to allograft. In every analysis and subanalyses, we observed a higher rate of nonunion in patients treated with a cage as opposed to allograft. These observations were statistically significant in 25 of the 26 possible permutations of analysis.

This finding is in contrast with prior literature. To date, there have been five studies that compared union rates in ACDF using cages versus bone graft and none of them were able to demonstrate a difference between the two techniques.5–9 However, the sample size in these studies are low. Even when the data from these five studies are pooled, there are only 122 patients in the bone graft ACDF group and 147 patients in the cage group. This low combined sample size precludes sufficient analysis and control for confounding variables, whereas our study allows for a larger stratified analysis.

Our study observed that ACDF with structural allograft results in a significantly higher rate of arthrodesis than ACDF with a cage. Despite this finding, ACDF cages may continue to have a role in cervical spine arthrodesis. In situations where structural allograft may not be readily available, cervical cages represent a reasonable alternative with a well-documented fusion rate, though perhaps not as high as allograft. However, as autologous iliac crest is widely available, a larger structured study with sufficient power, comparing the pros and cons of ACDF with cages and autologous iliac crest would be of great interest.

With any large database, there are weaknesses. The reliability of the reporting and coding is dependent upon multiple sources in an administrative data registry. We were unable to obtain radiographic evidence of nonunion for individual patients and instead relied on the diagnosis codes for nonunion, an important assumption we have made in this study. As this was an observational database study, we were also unable to determine the constitution of each cage placed, whether that be PEEK, titanium, mesh, or porous material. In our analysis, we stratified our initial population to account for the three most likely confounding variables for nonunion. It is entirely possible that other confounding variables exist and this may affect the analysis. Even with this large database, the nonunion patients whittled down to less than 11 patients in some subanalyses. One of the limitations of PearlDiver is when patient population size is less than 11, the true number is not revealed because of the potential for patient identification. We encountered this in some of our subanalyses and this did limit our ability to analyze the data, particularly in Tables 2 and 3 where we attempted to control for multiple confounders.

Future studies utilizing other data sources with sufficient sample size may be of value in further investigation. However, the PearlDiver data have been widely utilized in peer-reviewed publication.13,14 To date, this study is the largest comparative study examining the fusion rates of ACDF using cages and structural bone graft. Our practice, like the majority of spine surgeons in North America,3 is to utilize structural bone graft in ACDF. These data suggest that allograft, when available, may be a superior option than the use of a cage in achieving arthrodesis in the cervical spine.Key PointsIn this study, both structural allograft and intervertebral cage groups experienced high fusion rates.When comparing nonunion rates, these data suggest the superiority of allograft in ACDF.While the use of a cage and nonstructural bone graft material remains an important surgical option, the use of allograft, when donor bone is available, may be preferable in achieving solid arthrodesis.

Back to Top | Article Outline


The authors would like to thank Jeff France from the PearlDiver team his assistance throughout this project, as well as all anonymous reviewers for their feedback.

Back to Top | Article Outline


1. Bohlman HH, Emery SE, Goodfellow DB, Jones PK. Robinson anterior cervical discectomy and arthrodesis for cervical radiculopathy. Long-term follow-up of one hundred and twenty-two patients. J Bone Joint Surg Am 1993; 75:1298–1307.
2. Emery SE, Bohlman HH, Bolesta MJ, Jones PK. Anterior cervical decompression and arthrodesis for the treatment of cervical spondylotic myelopathy. Two to seventeen-year follow-up. J Bone Joint Surg Am 1998; 80:941–951.
3. Yoon ST, Konopka JA, Wang JC, et al. ACDF graft selection by surgeons: survey of AOSpine members. Global Spine J 2017; 7:410–416.
4. Samartzis D, Shen FH, Matthews DK, et al. Comparison of allograft to autograft in multilevel anterior cervical discectomy and fusion with rigid plate fixation. Spine J 2003; 3:451–459.
5. Bhadra AK, Raman AS, Casey AT, Crawford RJ. Single-level cervical radiculopathy: clinical outcome and cost-effectiveness of four techniques of anterior cervical discectomy and fusion and disc arthroplasty. Eur Spine J 2009; 18:232–237.
6. Kim MK, Kim SM, Jeon KM, Kim TS. Radiographic comparison of four anterior fusion methods in two level cervical disc diseases: autograft plate fixation versus cage plate fixation versus stand-alone cage fusion versus corpectomy and plate fixation. J Korean Neurosurg Soc 2012; 51:135–140.
7. Kim S-H, Lee JK, Jang JW, et al. Polyetheretherketone cage with demineralized bone matrix can replace iliac crest autografts for anterior cervical discectomy and fusion in subaxial cervical spine injuries. J Korean Neurosurg Soc 2017; 60:211–219.
8. Lee C-H, Hyun SJ, Kim MJ, et al. Comparative analysis of 3 different construct systems for single-level anterior cervical discectomy and fusion: stand-alone cage, iliac graft plus plate augmentation, and cage plus plating. J Spinal Disord Tech 2013; 26:112–118.
9. Liu J-M, Xiong X, Peng AF, et al. A comparison of local bone graft with PEEK cage versus iliac bone graft used in anterior cervical discectomy and fusion. Clin Neurol Neurosurg 2017; 155:30–35.
10. Phillips FM, Carlson G, Emery SE, et al. Anterior cervical pseudarthrosis. Natural history and treatment. Spine (Phila Pa 1976) 1997; 22:1585–1589.
11. Lowery GL, Swank ML, McDonough RF. Surgical revision for failed anterior cervical fusions. Articular pillar plating or anterior revision? Spine (Phila Pa 1976) 1995; 20:2436–2441.
12. Leven D, Cho SK. Pseudarthrosis of the cervical spine: risk factors, diagnosis and management. Asian Spine J 2016; 10:776–786.
13. Myhre SL, Buser Z, Meisel HJ, et al. Trends and cost of posterior cervical fusions with and without recombinant human bone morphogenetic protein-2 in the US Medicare population. Global Spine J 2017; 7:334–342.
14. Virk SS, Diwan A, Phillips FM, et al. What is the rate of revision discectomies after primary discectomy on a national scale? Clin Orthop Relat Res 2017; 475:2752–2762.

ACDF; allograft; anterior cervical discectomy and fusion; cervical spine; fusion rate; intervertebral cage; nonunion; PEEK cage

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.