Anterior controllable antedisplacement and fusion: quantitative analysis of a single surgeon's learning experience

BACKGROUND CONTEXT Anterior controllable antedisplacement and fusion (ACAF) is a novel surgical technique for the treatment of ossification of the posterior longitudinal ligament (OPLL) but the surgical learning curve for this technique has not been previously characterized. PURPOSE The aim of this study was to quantify a surgeon's learning curve for ACAF and the effect of surgeon experience on postoperative outcomes. STUDY DESIGN Prospective study of a single institution and single surgeon experience with ACAF surgery. PATIENT SAMPLE A total of 70 consecutive patients with OPLL undergoing ACAF surgery by a single, non-ACAF trained surgeon between 2017 and 2020 were analyzed. OUTCOME MEASURES Intraoperative and postoperative outcomes (blood loss, operative time, errors of surgical procedure, length of hospital stay, Japanese Orthopedic Association (JOA) scoring system, and surgical complications) were assessed. METHODS We prospectively reviewed the first 70 ACAF procedures between 2017 and 2020 performed by a single, non-ACAF trained surgeon. The function relationship between the operative time and case number was fitted using a locally weighted scatterplot smoothing (LOESS) plot. Spearman's correlation analysis was performed to determine factors affecting the operative time. The operative time-related learning curve for ACAF was established and difficulty of each procedure was assessed using a cumulative sum (CUSUM) model. The association between the specific errors of surgical procedures and surgeon experience was further analyzed. A modified CUSUM model was also used to establish the surgical procedure-related learning curve, and thus whether these two learning curves matched with each other was observed. Postoperative outcomes in relation to surgeon experience was compared using a Wilcoxon rank sum test and Chi-squared test. RESULTS Operative time presented a specific pattern of fewer patient-dependent changes as the case number increased. Spearman's correlation analysis showed the operative time was more affected by the case number (r=-0.73) than the complexity of condition and number of levels hoisted. The operative time-related CUSUM model identified the early (first 29 cases) and late phase (late 41 cases) of the learning process, which was also confirmed by a modified CUSUM model based on surgical procedure. The critical point of the CUSUM model for bilateral osteotomies was at case number 29, and time reduction after the early phase was approximately 34 minutes. Length of hospital stay and blood loss were less during the late phase than during the early phase (p<.05). Although no significant difference was observed in postoperative JOA scores between two phases, patients in the late phase obtained higher recovery rates of neurologic function than those in the early phase (p<.01). There was significant difference in the number of specific errors between the two phases (p=.02). There were no significant differences in overall complication rates between two phases, but a higher incidence of certain complications caused by specific errors was observed in the early phase (p=.02), including CSF leakage, C5 palsy and incomplete decompression. CONCLUSIONS We described, for the first time, a detailed learning curve for ACAF surgery. About 29 cases were needed to achieve mastery of ACAF. Once mastered, the surgeon could deal with various OPLL presentations in a universal way regardless of condition complexity and number of surgical levels. Bilateral osteotomies were the most difficult part of ACAF and produced the greatest reduction in time after mastery. We found a close association between specific errors of surgical procedure for ACAF and surgeon experience. Furthermore, certain complications caused by these errors should be on the alert during the early phase of learning ACAF, including CSF leakage, C5 palsy and incomplete decompression.