Four- and five-level anterior cervical discectomy and fusion with iliac crest autologous bone graft and plate–screw fixation: Radiographic and clinical outcomes with minimum 2-year follow-up Authors: Tony Tannoury, Hayley Denwood, Sung Jun Son, Megan Berube, Rehan R. Khan, Daman P. Dhunna, Jeongeun Kim, Yifei Wang, Chadi Tannoury Published: 2026 ABSTRACT ABSTRACTBackground:In multilevel disease, the risks and benefits of performing anterior versus posterior surgical approaches to the cervical spine remain debated. We sought to evaluate the radiographic and clinical outcomes of patients with multilevel cervical spondylotic myelopathy (CSM) treated with four- and five-level anterior cervical discectomy and fusion (ACDF), using tricortical iliac crest autograft and anterior plate fixation.Materials and Methods:Between 2009 and 2021, 1428 patients who underwent cervical fusion were reviewed. Inclusion required four- or five-level ACDF fusion with a 2-year clinical follow-up. Exclusions required fusions <4 levels; corpectomies; additional posterior fixation; and surgeries for infection, trauma, or malignancy. Demographics, preoperative and postoperative Nurick grades, Visual Analog Scale (VAS) pain scores, radiographic alignment (C2–C7 Cobb angle, T1 slope [T1S]), and fusion status were collected. Postoperative complications were categorized as early (0–3 months), intermediate (3–12 months), and late (13–24 months).Results:Twenty-seven patients (63% of males), age 59.7 ± 8.6 years, were included. Twenty-two patients underwent four-level and 5 underwent five-level ACDF. Significant improvements were observed in Nurick grade (2.78–1.15, P < 0.001), VAS score for neck and arm pain (both P < 0.001), Cobb angle (8.45°–13.39°, P = 0.001), and T1S (28.81°–32.36°, P < 0.001). Dysphagia was the most common early complication (29.6%), decreasing to 7.4% at late follow-up. Three patients (11.1%) developed adjacent segment disease. All patients achieved radiographic fusion.Conclusion:Four- and five-level ACDF with tricortical iliac crest autograft yields excellent clinical and radiographic outcomes in patients with multilevel CSM, with high fusion rates and low complication profile over a 2-year follow-up.Level of evidence:IV (retrospective case series). INTRODUCTION Cervical spondylotic myelopathy (CSM) is characterized by progressive sensorimotor deterioration and commonly occurs secondary to degenerative and age-related kyphogenic pathologies.[123] With an aging global population, U.S. hospitalizations for CSM have increased more than twofold, while the number of patients receiving surgical treatment each year has increased up to sevenfold.[4] When nonoperative management fails, anterior cervical discectomy and fusion (ACDF) with plate fixation, a common surgical treatment for CSM arising from anterior compressive pathologies, is preferred in patients with kyphotic cervical malalignment.[567] ACDF with plate fixation has been associated with significant improvement in associated symptoms and health-related quality-of-life outcome metrics.[89101112] Since Robinson and Smith first described the ACDF technique in 1958, extensive literature has documented outcomes for single-, two-, and three-level ACDF procedures.[13] However, the associated risk of pseudarthrosis appears to directly correlate with the number of fusion levels, altered biomechanics at several segments, and increased micromotion across a longer fusion construct.[141516] Accordingly, Veeravagu et al. reported that increasing the number of levels fused with ACDF correlated with increased rates of complications and revision surgery.[17] Currently, there is a paucity of studies specifically investigating four- and five-level ACDF, but the recent retrospective studies suggest favorable outcomes with acceptable postoperative complication rates.[181920] De la Garza-Ramos et al. reported 86% fusion rate in four-level ACDF, which was not significantly different than the three-level ACDF cohort.[21] Furthermore, Chang et al. reported a 92.6% fusion rate in four-level ACDF, with two asymptomatic pseudarthrosis cases requiring no treatment, while Koller et al. found 100% fusion in four- or five-level ACDF.[2223] MATERIALS AND METHODS This study was approved by the institutional review board at our institution. A retrospective investigation of clinical and operative notes was performed to identify all patients who underwent four- or five-level ACDF with tricortical iliac crest autograft and plate fixation at a single academic institution by two orthopedic spine surgeons (C.T. and T.T.) between January 1, 2009, and December 31, 2021. Symptomatic patients were evaluated with static and dynamic cervical spine radiographs and cervical spine magnetic resonance imaging to confirm the cervical pathology. The patients’ medical records and radiographic studies were reviewed, including demographics, preoperative status, diagnosis, and clinical outcomes. Preoperative and postoperative Nurick grade (range: 0–5) and Visual Analog Scale (VAS) (range: 0–10) for neck pain and arm pain were collected to quantify functional outcomes.[24] A total of 1428 patients underwent anterior cervical fusions, and 357 patients who underwent four- or five-level ACDF were identified for this study. Patients were excluded if they did not have follow-up beyond 2 years after surgery, underwent corpectomies (partial or total), had additional posterior fixation, or if their ACDF was performed for the treatment of spinal infections, trauma, or malignancies. Twenty-seven patients met the inclusion criteria and were considered for analysis, and all eligible patients were included. This study was designed to descriptively evaluate clinical and radiographic trends within a rare surgical cohort, and therefore, no predictive or multivariable modeling and no assessment of comparative outcomes were performed. Surgical indications included cervical nerve roots impingement with or without spinal cord compression(s), secondary to multilevel disc herniations and/or disc osteophytes resistant to nonoperative management. All studied patients received tricortical iliac crest autograft with anterior plate–screw fixation. Fusion was determined from upright static and dynamic cervical spine radiographs in all patients, using established dynamic radiographic criteria.[2526] The imaging was reviewed by two independent radiology-trained researchers. Fusion was based on (1) presence of bridging bony trabeculae and/or osseous union at the bone/graft interface on lateral radiographs, (2) interspinous motion <1 mm on dynamic flexion–extension radiographs, (3) no signs of anterior plating system loosening or failure, and (4) absence of radiolucency or halo formation around the grafts. Radiographic fusion was defined as the simultaneous presence of all four criteria described above, consistent with previously published dynamic radiographic fusion standards. Pre- and postoperative cervical lordosis was measured using the Cobb C2–C7 angle between the lower endplates of C2 and C7, while the T1 slope (T1S) was also recorded. Relevant data were collected for early (inpatient to 3 months), intermediate (3–12 months), and late (13–24 months) complications. Early postoperative complications included implant-related complications, dysphagia, dysphonia, reintubation, delayed extubation, and other medical complications. Intermediate and late postoperative complications included implant-related complications, persistent dysphagia, persistent dysphonia, pseudarthrosis, and adjacent segment disease (ASD). Postoperative dysphagia severity was assessed clinically and independently by speech therapists and was graded using the Bazaz dysphagia scale. All statistical analyses in this study were conducted with the intent to describe clinical trends rather than model definitive outcomes. Preoperative and final outcome at most recent follow-up were compared using the Wilcoxon signed-rank test. P <0.05 was considered statistically significant. Operative details Every patient received the standard Smith–Robinson approach for ACDF: a left-sided anterolateral skin incision, along the medial border of the sternocleidomastoid muscle, and the choice for either transverse or longitudinal skin incision was based on the local individual anatomy.[27] After decompressive discectomies, tricortical autograft (s) harvested separately from the anterior iliac crest [Figures 1–5] were inserted, followed by placement of anterior lordotic plate–screw fixation spanning either four (C3–C7 or C4–T1) or five (C3–T1) disc levels. All patients were immediately mobilized with supervision by physical/occupational therapists (PT/OT). In addition, all patients were evaluated by speech-language pathology (SLP) therapists to determine the safety of food and pill swallowing. Postoperative cervical orthoses were used as needed for comfort and for a 2–3-week duration following surgery approximately. All patients included in this study followed up in the clinic at 2 weeks, 3 months, 6 months, 9 months, 12 months, 18 months, and 24 months after surgery. The incision healing process was followed, and examples of incisions after 1 year and 21 months postoperative can be seen in Figure 6a and b. Postoperative anterior–posterior and lateral X-rays were obtained for all patients [Figures 7 and 8]. The representative postoperative radiographs shown in Figures 7 and 8 were obtained at late postoperative follow-up (>12 months postoperatively), consistent with routine radiographic surveillance for long-segment ACDF constructs. To obtain the tricortical iliac crest autograft, a small skin incision is made over the iliac crest area The iliac crest is subperiosteally exposed A high-speed oscillating saw is used to harvest the tricortical grafts Four or five tricortical grafts, averaging approximately 6–8 mm in thickness, are harvested After ensuring hemostasis, the wound is closed sequentially (a and b) Well-healed anterior cervical incisions at 1-year (a: four-level anterior cervical discectomy and fusion [ACDF]) and 21-month (b: five-level ACDF) follow-up (a and b) anteroposterior (a) and lateral (b) X-ray of a four-level anterior cervical discectomy and fusion 2 weeks postsurgery (a and b) Anteroposterior (a) and lateral (b) X-ray of a five-level anterior cervical discectomy and fusion 4 weeks postsurgery RESULTS Patient demographics A total of 27 patients with a mean age of 59.67 ± 8.58 years were included. Seventeen (63.0%) patients were male [Table 1]. Fifteen (55.6%) patients were American Society of Anesthesiologists (ASA) Grade 2, and 12 (44.4%) were ASA Grade 3. There were 9 (33.3%) active and 10 (37.0%) former smokers in the cohort. Six (22.2%) patients had type II diabetes, 3 (11.1%) patients had chronic kidney disease, and 1 (3.7%) patient had chronic steroid intake secondary to chronic obstructive pulmonary disease. The mean body mass index was 27.78 ± 5.30 kg/m2. The mean clinical and radiographic follow-up was 43.88 and 22.41 months, respectively. Table 1Characteristics and Radiographic Measurements of Patients Undergoing 4- and 5-Level ACDFCharacteristicsTotal Cohort (n=27) P Age (years)59.678.58BMI27.785.30ASA Class2.440.51Estimated Blood Loss (mL)71.1193.0Hospital Stay (days)3.151.70Sex Female10 (37.0%) Male17 (63.0%)Diabetes None21 (77.8%) T2DM6 (22.2%)Chronic Kidney Disease Yes3 (11.1%) No24 (88.9%)Chronic Steroid Use Yes1 (3.7%) No26 (96.3%)Tobacco Use Current9 (33.3%) Former10 (37.0%) Never8 (29.7%)Number of Levels Fused 4-Level Fusion22 (81.5%) 5-Level Fusion5 (18.5%)Segments Fused C3-C717 (63.0%) C4-T15 (18.5%) C3-T15 (18.5%)Preoperative Nurick Scale2.78±0.89P<0.001Postoperative Nurick Scale1.15±0.82Preoperative VAS Neck Score7.44±1.42P<0.001Postoperative VAS Neck Score3.50±1.89Preoperative VAS Arm Score6.28±2.02P<0.001Postoperative VAS Arm Score2.33±1.64Preoperative C2-C78.45°±6.49°P=0.001Postoperative C2-C713.39°±6.34°Preoperative T1 Slope28.81°±7.63°P<0.001Postoperative T1 Slope32.36°±9.52BMI: Body Mass Index; T2DM: Type II Diabetes Mellitus; VAS: Visual Analog Scale Characteristics and Radiographic Measurements of Patients Undergoing 4- and 5-Level ACDF BMI: Body Mass Index; T2DM: Type II Diabetes Mellitus; VAS: Visual Analog Scale Functional measures All patients included in this cohort demonstrated clinical myelopathy preoperatively, as reflected by a mean baseline Nurick grade of 2.78 ± 0.89. The mean preoperative and postoperative Nurick grading was 2.78 ± 0.89 and 1.15 ± 0.82, respectively [Table 1]. There was statistically significant improvement in the Nurick grading scale postoperatively (P < 0.001). Similarly, the VAS neck pain score improved from 7.44 ± 1.42 preoperatively to 3.50 ± 1.89 postoperatively (P < 0.001). VAS arm pain score also improved from 6.28 ± 2.02 preoperatively to 2.33 ± 1.64 postoperatively (P < 0.001). Perioperative characteristics There were 22 (81.5%) patients who underwent four-level ACDF and 5 (18.5%) who underwent five-level ACDF [Table 1]. Out of the 22 four-level ACDF, 17 (63.0%) involved C3–C7 levels, and 5 (18.5%) involved C4–T1 levels, whereas the five-level ACDF were all at the C3–T1 levels. The mean estimated blood loss was 71.11 ± 93.0 mL. The mean length of hospital stay was 3.15 ± 1.70 days [Table 1]. Delayed discharge was due to a delay in SLP evaluation and clearance in 5 (18.5%) patients and a delay in PT/OT clearance in 9 (33.3%) patients. Clinical and radiographic outcomes There were no intraoperative complications [Table 2]. Nineteen out of 27 (70.3%) patients had an early complication of transient dysphagia. Of note, one of them developed dysphagia secondary to a self-resolving prevertebral hematoma formation in the inpatient setting. The patient was a former smoker with no significant medical comorbidities and exhibited no signs of airway compromise. One out of 27 (3.7%) had transient dysphonia that resolved by week 6 postsurgery. Five (18.5%) patients had moderate–severe donor-site pain; of those, 1 (3.7%) patient had a donor-site seroma, and 1 (3.7%) patient had a donor-site infection which both required donor-site surgical exploration. Table 2Postoperative Outcomes Following 4- and 5-Level ACDFComplicationsPercentageIntraoperative Complications Yes0 (0.0%) No27 (100.0%)Early Postoperative Complications Dysphagia9 (33.3%) Dysphonia1 (3.7%) Donor-Site Pain5 (18.5%) Donor-Site Infection1 (3.7%) Donor-Site Seroma1 (3.7%)Intermediate Postoperative Complications Persistent Dysphagia2 (7.4%) Residual Upper Extremity Numbness and Weakness2 (7.4%)Late Postoperative Complications Persistent Dysphagia2 (7.4%) Residual Upper Extremity Numbness and Weakness1 (3.7%) New Upper Extremity Numbness and Weakness3 (11.1%) Adjacent Segment Disease3 (11.1%) C2-C31 (3.7%) C3-C41 (3.7% C7-T11 (3.7%)Reoperations Required Donor Site Debridement2 (7.4%) Postoperative Outcomes Following 4- and 5-Level ACDF During the intermediate postoperative period, 2 (7.4%) patients had residual upper extremity numbness and weakness, resolving in one patient by the end of the study period. Two (7.4%) patients continued to have persistent dysphagia into the intermediate and late postoperative periods. Both patients reported these symptoms to be mild. Three (11.1%) patients developed radiographic ASD with new mild upper extremity numbness and/or subtle subjective weakness [Table 2]. The areas of ASD included C2–C3, C3–C4, and C7–T1. None underwent secondary cervical operations during the study period. The comparison of radiographic cervical parameters pre- and postoperatively is shown in Table 1. The C2–C7 Cobb angle was corrected from a preoperative mean of 8.45° ±6.49° to postoperative mean of 13.39° ±6.34°. There was a statistically significant difference (P = 0.001) between preoperative and postoperative C2–C7 lordosis. The T1S also significantly increased from a preoperative mean of 28.81° ±7.63° to postoperative mean of 32.36° ±9.52 (P < 0.001). While ASD was noted in 3 patients (11.1%), all studied subjects exhibited successful radiographic fusion with no evidence of hardware complications or pseudarthrosis at an average radiographic follow-up of 22.41 ± 10.65 months (range: 12–48 months). DISCUSSION ACDF is the standard surgical treatment for patients with cervical spondylosis refractory to conservative management.[1328] Basques et al. demonstrated that multilevel ACDF restored lordosis more effectively than single-level fusions, with no significant difference in postoperative outcomes or risk of ASD between the two.[29] At present, there is an ongoing debate on the surgical approach (anterior vs. posterior) that achieves superior postoperative neural function and healthcare cost in patients with multilevel CSM.[3031] When comparing ACDF with other procedures, a meta-analysis reported that postoperative outcomes were more favorable in ACDF compared to posterior laminoplasty for multilevel CSM.[28] Similar to our findings, these results suggest that four- and five-level ACDF are safe and effective for treating multilevel CSM and are associated with excellent clinical and radiographic outcomes. In addition, as discussed by Crawford et al., procedures with relatively lower frequency (i.e, four- and five-level ACDF) necessitate retrospective analysis of cases to set a foundation for future work. The rate of transient dysphagia in our four- or five-level ACDF cohort was 19/27 (70.3%), which is within the range (17.6%–72.4%) of what has been reported in literature.[2223] In 10 patients, the dysphagia was mild, and in nine moderate–severe using the Bazaz scale. Twelve of these patients had transient dysphagia that resolved within 3 months. Our cohort’s observed rate of postoperative dysphagia falls in the upper range of the reported rate, which may be related to the systematic implementation of perioperative professional SLP consultation, and an increased sensitivity in dysphagia diagnosis. Moreover, no other soft-tissue injuries (i.e., extensive bleeding, esophageal injury, or odynophagia) were observed in our study, consistent with their relatively low prevalence in the literature. We did not observe graft or hardware-related complications. Similarly, a systematic review by Yee et al. reported a low overall graft or hardware failure rate of 2.1% in ACDF patients. Furthermore, none of our patients required revision surgery, while reported revision rates in the literature range from 8% to 35%.[1418] ASD has been well documented in literature as one of the most anticipated complications following ACDF. A formative study by Hilibrand et al. demonstrated the incidence of ASD to be 2.9% per year during the first 10 years postoperatively. In addition, Bydon et al. reported a 12.2% incidence of ASD in a retrospective series of 888 patients. The incidence of ASD has also been positively correlated with the number of fusion levels, attributed to altered biomechanics in multilevel ACDF that results in greater adjacent segment motion compared to single-level ACDF.[17] Increased rates of pseudoarthrosis following three- and four-level ACDF have been extensively documented throughout the literature, with rates as high as 42% and 56%, respectively.[323334] Conversely, there is a sparsity of literature describing the incidence of pseudoarthrosis following five-level ACDF. To accurately determine the presence of pseudoarthrosis, we used the lateral dynamic radiographic criteria, which have been described as having a higher sensitivity, specificity, and predictive values compared to other methods.[2526] In this study, there were no cases of pseudoarthrosis in either the four- or five-level groups. This may be due to our choice of using an iliac crest autograft, which has been shown to be associated with decreased pseudoarthrosis rates.[32] There are several limitations to this study, including its retrospective design and small sample size. Furthermore, as all patients received tricortical iliac crest autograft, comparison of graft choice and its association with clinical and radiographic outcomes was not possible. Moreover, the small five-level ACDF cohort (n = 5) prevented adequate statistical comparison to the four-level ACDF cohort. Patients with asymptomatic pseudarthrosis or delayed bony union may have been missed. In addition, the mean radiographic follow-up of 22.41 months may underestimate the ASD incidence. Finally, the generalizability of this study may be limited as all patients within this study were operated on at a single institution by two fellowship-trained and experienced senior surgeons who use a similar surgical technique. Five-level ACDFs are extremely infrequent on a national level, so a baseline study such as the present is critical. To our knowledge, this study is one of the first and largest series to date that evaluated four- and five-level ACDF with tricortical iliac crest autograft. Outcomes after four-level ACDF have been reported in literature, whether as a case series or in comparison with other levels of fusion.[9182135] However, five-level ACDF is often not included in the studies, perhaps attributing to the rarity of the procedure.[23] CONCLUSION Excellent clinical, functional, and radiographic outcomes can be achieved with four- and five-level ACDF with tricortical iliac crest autograft. These surgical procedures are safe and effective for treating CSM refractory to conservative management and are associated with high rates of fusion and low complication rate. Conflicts of interest There are no conflicts of interest.