What is the role of dynamic cervical spine radiographs in predicting pseudarthrosis revision following anterior cervical discectomy and fusion?

Abstract

Background context: Postoperative dynamic radiographs are used to assess fusion status after anterior cervical discectomy and fusion (ACDF) with comparable accuracy to computed tomography (CT) scans.

Purpose: To (1) determine if dynamic radiographs accurately predict pseudarthrosis revision in a cohort of largely asymptomatic patients who underwent ACDF, (2) determine how adjacent segment motion is affected by fusion status, and (3) analyze how clinical outcomes differ between patients with symptomatic and asymptomatic pseudarthrosis.

Study design: Retrospective cohort study.

Patient sample: Patients ≥ 18 years who underwent primary one- to four-level ACDF at a single institution over a 10-year period.

Outcome measures: Interspinous motion on preoperative and postoperative flexion-extension radiographs and preoperative and postoperative Visual Analogue Scale for Neck Pain (VAS Neck) and Arm Pain (VAS Arm), Neck Disability Index (NDI), Modified Japanese Orthopaedic Association scale (mJOA), Mental and Physical Component Scores of the Short-Form 12 (SF-12) Health Survey (MCS-12 and PCS-12) METHODS: The difference in spinous process motion between flexion and extension radiographs was used to determine motion at each level of the ACDF construct. Pseudarthrosis was defined as ≥ 1 mm spinous process motion on dynamic radiographs. A receiver operating characteristic (ROC) curve was generated to predict the probability of surgical revision for pseudarthrosis based on millimeters of interspinous motion at each instrumented level. Patient reported outcome measures (PROMs) were used to assess the effect of pseudarthrosis on clinical outcomes. Alpha was set at p<.05.

Results: A total of 597 patients met inclusion criteria including 1,203 ACDF levels. Of those, 215 patients (36.0%) were diagnosed with a pseudarthrosis on dynamic radiographs with 29 patients (4.9%) requiring pseudarthrosis revision. ROC analysis identified a "cutoff" value of 1.00 mm of interspinous process motion for generating an optimal area under the curve (AUC). The negative predictive value (NPV) was 99.6%, whereas the positive predictive value (PPV) was 13.7%. When analyzing adjacent segment motion, the Δ supra-adjacent interspinous process motion (ISM) was significantly lower for patients with a superior construct pseudarthrosis (-1.06 mm vs. 1.80 mm, p<.001), whereas the Δ infra-adjacent level ISM was significantly lower for patients with an inferior construct pseudarthrosis (-1.21 mm vs. 2.15 mm, p<.001). Patients with a pseudarthrosis not requiring revision had worse postoperative NDI (29.3 vs. 23.4, p=.027), VAS Neck (3.40 vs. 2.63, p=.012), and VAS Arm (3.09 vs. 1.85, p=.001) scores at 3 months, but not 1-year, compared with patients who were fused. Patients requiring pseudarthrosis revision had higher 1-year postoperative NDI (38.0 vs. 23.7, p=.047) and lower 1-year postoperative Δ VAS Arm (-0.22 vs. -2.97, p=.016) scores.

Conclusions: One-year postoperative dynamic radiographs have a greater than 99% negative predictive value for identifying patients requiring pseudarthrosis revision, but they have a low positive predictive value. Most patients with a pseudarthrosis remain asymptomatic with similar 1-year postoperative patient-reported outcomes compared with patients without a pseudarthrosis.

Keywords: Adjacent segment motion; Anterior cervical discectomy and fusion; Dynamic radiographs; Patient reported outcome measures; Pseudarthrosis; Revision.