Abstract
Study Design
Retrospective cohort study.
Objectives
In patients undergoing elective anterior cervical discectomy and fusion (ACDF), we sought to determine the impact of screw length on: (1) radiographic pseudarthrosis, (2) pseudarthrosis requiring reoperation, and (3) patient-reported outcome measures (PROMs).
Methods
A single-institution, retrospective cohort study was undertaken from 2010-21. The primary independent variables were: screw length (mm), screw length divided by the anterior-posterior vertebral body diameter (VB%), and the presence of any screw with VB% < 75% vs all screws with VB% ≥ 75%. Multivariable logistic regression controlled for age, BMI, gender, smoking, American Society of Anesthesiology grade, number of levels fused, and whether a corpectomy was performed.
Results
Of 406 patients undergoing ACDF, levels fused were: 1-level (39.4%), 2-level (42.9%), 3-level (16.7%), and 4-level (1.0%). Mean screw length was 14.3 ± 2.3 mm, and mean VB% was 74.4 ± 11.2. A total of 293 (72.1%) had at least one screw with VB% < 75%, 113 (27.8%) had all screws with VB% ≥ 75%, and 141 (34.7%) patients had radiographic pseudarthrosis at 1-year. Patients who had any screw with VB% < 75% had a higher rate of radiographic pseudarthrosis compared to those had all screws with VB% ≥ 75% (39.6% vs 22.1%, P < .001). Multivariable logistic regression revealed that a higher VB% (OR = .97, 95%CI = .95-.99, P = .035) and having all screws with VB% ≥ 75% (OR = .51, 95%CI = .27-.95, P = .037) significantly decreased the odds of pseudarthrosis at 1-year, with no difference in reoperation or PROMs (all P > .05).
Conclusion
Longer screws taking up ≥75% of the vertebral body protected against radiographic pseudarthrosis at 1-year. Maximizing screw length in ACDF is an easily modifiable factor directly under the surgeon’s control that may mitigate the risk of pseudarthrosis.
Keywords
Introduction
Anterior cervical discectomy and fusion (ACDF) has been performed for over 70 years and is currently the most performed surgery to treat cervical spine pathology. 1 Cervical pseudarthrosis, a vexing complication that can occur in up to 15% of patients after an ACDF, 2 is defined as the failure to achieve the intended cervical fusion. 3 Though improved techniques and implant selection have evolved to reduce the risk of pseudarthrosis after ACDF, pseudarthrosis persists.4-8 While pseudarthrosis is asymptomatic in up to 30% of cases, failure to fuse can significantly impact patient quality of life, leading to increased pain, reoperation, and worsened patient-reported outcome measures (PROMs).9-11 One study found that the risk of pseudarthrosis was highest in years 2-5 after surgery. 12
The etiology of pseudarthrosis is multifactorial, including patient and surgical factors. 3 Several patient-related risk factors have been reported, including smoking, diabetes, osteoporosis, obesity, and malnutrition, among others.9,13-18 Operative risk factors include number of levels fused, graft type, and plate and screw properties may also influence fusion success, though number of levels is likely most important factor.15,17,19,20 The topic of screw length in ACDF remains relatively understudied. While longer screws have demonstrated increased pull-out strength and decreased interspinous motion in two-level ACDF,21-24 whether longer screws can increase the risk of complications, including spinal cord injury, remains unknown. In perhaps the most robust study to date on screw length after ACDF of 85 patients undergoing 2-level ACDF, Lee et al 24 found that patients with any screw length <75% of the vertebral body had a higher rate of pseudarthrosis at 1-year. Nevertheless, the impact of screw length on pseudarthrosis remains understudied in single- and multi-level ACDF.
Given the dearth of research regarding screw length in elective ACDF and the lack of consensus on the optimal screw length in ACDF, we attempted to investigate this matter further. In a cohort of patients undergoing elective, primary ACDF, we sought to determine the impact of screw length on: (1) radiographic pseudarthrosis, (2) pseudarthrosis requiring reoperation, and (3) PROMs.
Methods
Study Design
A single-institution, retrospective cohort study was conducted using data from a prospectively collected spine outcomes registry from 2010-2021. Twelve neurosurgery and orthopedic spine surgeons have contributed patients to this registry since its existence. Study approval was obtained from the Institutional Review Board (IRB #100388). Three full-time employees were responsible for collecting preoperative, 3-month, and 12-month PROMs. Due to the prospective nature of data acquisition at time of collection, a signed consent for participation was collected from all study participants.
Patient Population
The inclusion criteria were: adult patients ≥18-year, undergoing elective, primary, single- or multi-level ACDF for degenerative pathology, and having available imaging at 1-year postoperatively. Patients with less than 1-year follow-up on imaging and patients with a history of prior fusions at the same or adjacent levels were excluded. In addition, patients undergoing ACDF for traumatic fractures or tumor resection were excluded. Cases undergoing a combined antero-posterior fusion were excluded. The sample consisted of patients undergoing primary ACDF without a prior history of pseudarthrosis at that level.
Independent Variables
The primary independent variables were: screw length (mm) and screw length divided by the anterior-posterior vertebral body diameter (VB%). The latter variable was dichotomized into the presence of any screw with VB% < 75% vs all screws with VB% ≥ 75%. Mean screw length of the two screws present at each level was calculated. Mean screw length was subsequently divided by the anterior-posterior vertebral body diameter at the corresponding level. The same methodology was reported by Lee et al. 24
Other independent variables influencing the likelihood of successful fusion were demographic characteristics, comorbidities, American Society of Anesthesiology grade (ASA) grade, smoking, levels fused, presence of corpectomy, and perioperative variables such as estimated blood loss (EBL) and operation duration. Perioperative and postoperative variables included discharge disposition, length of stay (LOS), postoperative complications, unplanned readmissions at 3-month, and last office follow-up visit.
Outcomes
The primary outcomes were: (1) radiographic pseudarthrosis at 1-year postoperatively, (2) pseudarthrosis requiring reoperation, and (3) PROMs. In keeping with prior literature, pseudarthrosis was defined as interspinous motion>1 mm on flexion-extension X-ray.3,25 In the absence of flexion-extension X-ray, pseudarthrosis was defined as the lack of intra/extra-graft bridging bone on the 1-year imaging in the following order: Computed Tomography (CT), Magnetic Resonance Imaging (MRI), or static antero-posterior X-ray.3,25 Figure 1A–D and Figure 2A–D illustrate a successful fusion and pseudarthrosis according to each imaging modality, respectively. MRI and static X-ray were considered the least reliable ways. A-D Successful fusion according to each imaging modality: flexion-extension X-ray (A), Coronal CT (B), sagittal MRI (C), and static X-ray (D). ISM: interspinous motion. A-D Pseudarthrosis diagnosis according to each imaging modality: flexion-extension X-ray (A), Coronal CT (B), sagittal MRI (C), and static X-ray (D). ISM: interspinous motion.
PROMs included Numeric Rating Scale (NRS)-Arm/Neck and Neck Disability Index (NDI). PROMs were analyzed both continuously and based on the minimum clinically important difference (MCID), elaborated on below.
Statistical Analysis
Descriptive statistics were reported to compare patients with any screw with VB% < 75% and all screws with VB%>75%. Mean and standard deviation were reported for continuous variables and frequency for categorical variables. Normal distribution and variance for continuous variables were assessed with the Shapiro-Wilk test and F-test, respectively. Parametric data with equal variance were analyzed with a 2-tailed t test, while nonparametric data were compared with the Wilcoxon signed rank test or Mann-Whitney test. χ2 or Fisher’s exact test was used for nominal data. Pseudarthrosis and pseudarthrosis requiring reoperation were each considered as a dichotomous categorical variable. MCID of PROMs was set at 30% improvement at 1-year postoperatively. MCID was defined at 30% improvement from baseline PROMs. 26 Univariate and multivariable logistic regression controlled for age, body mass index (BMI), gender, smoking, ASA grade, number of levels fused, and whether a corpectomy was performed. The multivariable model was constructed in accordance with prior knowledge of potential confounders of pseudarthrosis following ACDF and based on the discussion with senior authors.1,3,13,14,16,18,22,24 An alpha value <.05 was considered significant. Statistical analysis was performed using SPSS 22 (IBM, Armonk, NY).
Results
Demographics, Perioperative, and Postoperative Variables
Of 1130 patients undergoing elective, primary ACDF, 406 (35.9%) had available imaging at 1-year postoperatively, likely displaying surgeons’ varying practices in postoperative follow-up and imaging. Median time to last office visit was 13.5months. Mean age was 52.3 ± 10.7 years, 196 (48.3%) were males, and levels fused were: 1-level (39.4%), 2-level (42.9%), 3-level (16.7%), and 4-level (1.0%). Mean screw length was 14.3 ± 2.3 mm and mean VB% was 74.4 ± 11.2 across all levels. Of 406 patients, 293 (72.1%) had at least one screw with VB% < 75% and 113 (27.8%) had all screws with VB% ≥ 75% (Figure 3A–B). A-B Illustration of the percentage of the vertebral body diameter occupied by the screw: VB% ≥ 75% (A) and VB% < 75% (B).
Demographics Characteristics of Patients Undergoing ACDF.
ASA = American Society of Anesthesiology grade, BMI = Body mass index, COPD = Chronic Obstructive Pulmonary Disease.
Operative and Postoperative Variables of Patients Undergoing ACDF.
Pseudarthrosis and Reoperation Due to Pseudarthrosis
A total of 141 (34.7%) patients had radiographic pseudarthrosis at 1-year postoperatively, and 15 (3.7%) underwent reoperation due to pseudarthrosis. No difference was found between the number of levels fused and pseudarthrosis rate (1-level = 34.4% vs 2-level = 33.3% vs 3-levels = 39.7% vs 4-levels = 25.0%, P = .785) or pseudarthrosis requiring reoperation (1-level = 5.0% vs 2-level = 2.9% vs 3-levels = 2.9% vs 4-levels = 0%, P = .715). Similarly, no difference was found between patients with vs without a corpectomy in pseudarthrosis rate (19.2% vs 36.0%, P = .092) and pseudarthrosis requiring reoperation (4.0% vs 0%, P = .618).
A higher rate of pseudarthrosis was found in patients with any screw with VB% < 75% compared to all screws with VB% ≥ 75% (39.6% vs 22.1%, P < .001). Similarly, patients who had all screws with VB% < 75% had a higher rate of radiographic pseudarthrosis compared to those who had any screw with VB% ≥ 75% (44.1% vs 29.1%, P = .002). The imaging used to assess for pseudarthrosis at 1-year were flexion-extension X-ray (56.4%), CT scan (3.4%), MRI (3.2%), and static X-ray (36.9%), with no observed difference in the imaging used between the two groups (P = .147). Although a higher reoperation rate due to pseudarthrosis was found in patients with any screw with VB% < 75%, it was not statistically significant (4.1% vs 2.7%, P = .769) (Table 2).
Univariate and Multivariable Logistic Regression of the Impact of Screw Length and VB% on Pseudarthrosis and MCID PROMs, Controlling for Age, BMI, Gender, Smoking, ASA Grade, Number of Levels Fused, and Whether a Corpectomy was Performed.
ASA = American Society of Anesthesiology grade. NRS = Numeric rating scale. NDI = Neck disability index. MCID = Minimal clinically important differences. *Statistical Significance.
Patient-Reported Outcome Measures
PROMs of Patients Undergoing ACDF.
NRS = numeric rating scale. NDI = neck disability index. MCID = minimal clinically important differences.
Discussion
The present study evaluated how screw length, as defined by the percentage of the vertebral body diameter occupied by the screw, affected radiographic pseudarthrosis, pseudarthrosis requiring reoperation, and 1-year PROMs in patients undergoing elective, primary ACDF. Preoperatively, comparable demographics and comorbidities were found between patients with different screw length. Perioperatively, no difference was found in overall complication rate and 3-month readmissions. Postoperatively, having any screw with VB% < 75% was independently associated with increased odds of radiographic pseudarthrosis at 1 year, with no significant difference in pseudarthrosis requiring reoperation. The fusion was evaluated using flexion-extension x-rays and interspinous motion >1 mm in most patients, with no significant differences in the imaging modalities used to assess fusion between patients with any screw with VB% < 75% or all screws VB% ≥ 75%. Similarly, no difference was found in PROMs at 1-year between the two groups. Though no difference was found regarding screw length in pseudarthrosis requiring reoperation, these findings emphasized that maximizing screw length in single and multi-level ACDF may mitigate the risk of radiographic pseudarthrosis.
Patients with any screw with VB% < 75% were at significantly increased risk of pseudarthrosis when controlling for age, BMI, gender, smoking, ASA grade, number of levels fused, and whether a corpectomy was performed. In a similar methodology, a retrospective study of prospectively collected data of 85 patients undergoing 2-level ACDF by Lee et al 24 evaluated pseudarthrosis in relation to screw length utilizing follow-up with flexion-extension radiographs at 6-week, 6-month, and 1-year. The authors found that having any screw with VB% < 75% was significantly associated with pseudarthrosis at every time point. Comparatively, our study examined a greater number of patients, included both single- and multi-level ACDF, incorporated more imaging modalities, and found that having any screw with VB% < 75% was independently associated with pseudarthrosis at 1-year, with no differences in pseudarthrosis requiring reoperation. Our findings are in line with Lee et al 24 and other biomechanical studies,21-23 which demonstrated an increased screw pull-out strength due to stronger bone purchase. A plausible explanation of our findings may be that shorter screws tend to result in a slightly less stable construct when compared to the use of longer screws. Longer screws, due to their greater purchase within the vertebral body, may offer an enhanced biomechanical support, in turn leading to a decreased rate of radiographic pseudarthrosis.
Despite an increased risk of radiographic pseudarthrosis in patients with any screw with VB% < 75%, no significant difference was found in pseudarthrosis requiring reoperation. Lee et al 24 did not assess screw length in relation to pseudarthrosis requiring reoperation or PROMs. A retrospective study conducted by Alonso et al 27 have found that operative pseudarthrosis occurred in 7 of 211 (3.3%) patients undergoing 1- to 2 level ACDF. In another retrospective study of 2078 patients undergoing ACDF, pseudarthrosis requiring reoperation occurred in .58% of patients and was level-dependent, such as more reoperations were observed in ≥3-level ACDF compared to ≤2-level fusions. 28 Related to the current data, the lack of statistical significance may have been due to the overall low rate of pseudarthrosis requiring reoperation, with 4.1% in patients with any screw with VB% <75% vs 2.7% in patients with all screws VB% > 75%. It is possible with longer term follow-up, and larger samples, these differences may widen and become statistically and clinically significant.
The current study demonstrated that having all screws with VB% ≥ 75% did not increase the risk of perioperative complications. While screw length may help reduce rates of pseudarthrosis, whether longer screws increase intraoperative complications, including dural tears or spinal cord injury, remained relatively unexplored in prior literature. Of course, we caution heavily against using excessively long screws, and all screws should be measured preoperatively and intraoperatively based on caspar pin length. Moreover, once the discectomy is performed, it is our practice to place a screw into the disc space until it touches the dura and add 2 mm on to that given the angling up and down of the cranial and caudal screws, respectively. Evidently, prior studies have demonstrated that intraoperative use of neuronavigation may guide screw insertion and decrease the risk of complications.29,30
The present study has several limitations warranting discussion. First, A significant percentage of patients who received ACDF did not receive imaging to assess for pseudarthrosis 1-year postoperatively (64.1%). This lack of follow-up indicates that regular 1-year postoperative follow-up was not satisfactory and suggests that instances of pseudarthrosis may have been underreported, which may significantly influence our long-term outcomes. Second, MRI and static x-rays may not be ideal in depicting pseudarthrosis, but they were used whenever the patient did not receive postoperative flexion-extension X-ray or CT scan. MRI and x-rays have been shown in previous studies to be adjuncts to detect bone fusion. 31 Additionally, the rates of using MRI and static x-rays were similar in both cohorts. Third, as a retrospective, single-institution, multi-surgeon study, these findings may have limited generalizability. Fourth, pseudarthrosis is a multi-factorial process, and accounting for all cofactors affecting pseudarthrosis was nearly impossible. While we controlled for a significant number of covariates, we did not control for other important factors such as the presence of rheumatoid arthritis, graft type, and surgical technique, among others. 32 Interbody material is certainly an important factor, and we were unfortunately unable to account for this. Fifth, the angular trajectory of screws was not reported, and whether it impacts fusion success remains unknown. Screw angulation should be an important factor to consider in future research. Sixth, although all screws inserted in 4-level fusion were VB% < 75%, no difference was found in pseudarthrosis rate between the different fusion levels. Seventh, we recognize that the fundamental differences between corpectomy procedures and single-level ACDF may still introduce some level of confounding. 32 We tried to mitigate this confounder by including corpectomies as a covariate in our multivariable analysis to control for their influence on pseudarthrosis. Furthermore, no difference was found in pseudarthrosis rate between patients with vs without a corpectomy. Eighth, while pseudarthrosis was radiographically assessed, pseudarthrosis requiring reoperation is surgeon-related and may be an unreliable assessment of outcomes. Finally, the present study was limited by a follow-up time of only 1-year, which may underestimate the rate of pseudarthrosis, reoperations due to pseudarthrosis, and long-term PROMs. Future prospective studies should validate the clinical impact of screw length on pseudarthrosis requiring revisions and long-term PROMs.
Conclusion
In a single-center study of patients undergoing elective single- and multi-level ACDF, screw length taking up to <75% of the vertebral body was significantly associated with increased rates of radiographic pseudarthrosis at 1-year postoperatively. No significant association was found between screw length and pseudarthrosis requiring reoperation or PROMs. These results can help surgeons mitigate the risk of radiographic pseudarthrosis following elective, primary ACDF and create a better understanding of the importance of maximizing screw length without increasing the risk of complications.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Dr Zuckerman reports being an unaffiliated neurotrauma consultant for the National Football League. Dr Stephens is a consultant for Nuvasive and Carbofix and receives institutional research support from Nuvasive and Stryker Spine. Dr Abtahi received an institutional research support from Stryker Spine. No other perceived conflict of interest by any of the listed authors.
