The establishment of a novel upper cervical complex fracture classification system.

Background context: Upper cervical complex fractures are associated with high rates of neurological damage and mortality. The Dickman's classification is widely used in the diagnosis of upper cervical complex fractures. However, it falls short of covering the full spectrum of complex fractures. This limitation hinders effective diagnosis and treatment of these injuries.

Purpose: To address the diagnostic gap in upper cervical complex fractures, the study introduces a novel classification system for these injuries, assessing its reliability and usability.

Study design: Proposal of a new classification system for upper cervical complex fractures.

Patient sample: The study comprised the clinical data of 242 patients with upper cervical complex fractures, including 32 patients treated at our hospital, along with an additional 210 cases from the literature.

Outcome measures: The interobserver and intra-observer reliability (kappa coefficient, κ) of this classification system were investigated by 3 spine surgeons. The 3 researchers independently reevaluated the upper cervical complex fracture classification system 3 months later.

Methods: The proposed classification categorizes upper cervical complex fractures into 3 main types: Type I combines odontoid and Hangman's fractures into 2 subtypes; Type II merges C1 with odontoid/Hangman's fractures into 3 subtypes; and Type III encompasses a combination of C1, odontoid, and Hangman's fractures, divided into 2 subtypes. Meanwhile, a questionnaire was administered in 15 assessors to evaluate the system's ease of use and clinical applicability.

Results: A total of 45 cases (18.6%) unclassifiable by Dickman's classification were successfully categorized using our system. The mean κ value of inter-observer reliability was 0.783, indicating substantial reliability. The mean κ value of intraobserver reliability was 0.862, indicating almost perfect reliability. Meanwhile, thirteen assessors (87.7%) stated that the classification system is easy to remember, easy to apply, and they expressed intentions to apply it in clinical practice in the future.

Conclusions: This system not only offers high confidence and reproducibility but also serves as a precise guide for clinicians in formulating treatment plans. Future prospective applications are warranted to further evaluate this classification system.