0.4% incidence of return to OR due to screw malposition in a large prospective adolescent idiopathic scoliosis database

摘要

Purpose In contrast to infection and curve progression, return to OR for implant malposition is potentially within the surgeon's control. With increasing surgeon familiarity with freehand/fluoroscopic pedicle screw placement, rates of return to OR due to malposition may have decreased over time. We sought to document the incidence and risk factors for return to OR due to screw malposition in a large cohort of patients with idiopathic scoliosis. Methods AIS patients enrolled in a multicenter prospective registry with minimum 2-year follow-up undergoing PSF between 2003 and 2017 were included. Surgeries with the use of intraoperative CT-guided navigation were excluded. Return to OR due to screw malposition over time was tabulated. Lenke class, age at surgery, year of surgery, height, weight, surgeon and site were evaluated. Results 2435 patients underwent PSF with non-navigated open pedicle screw instrumentation. Mean age was 14.6 years, mean thoracic curve magnitude was 53 degrees, and mean lumbar was 40 degrees. Patients did not routinely undergo intraoperative or postoperative CT evaluation. There were 10 returns to OR for screw malposition at a mean of 0.83 years postoperatively (range 0.1-3.4 years), for an overall incidence of 0.4%. Of the 14 screws, 10 were thoracic, 7 were left-sided. No association was found between screw malposition and curve size, enrolling site, surgeon, BMI, Lenke class, or age. Five patients had radiculopathy which improved after screw revision. Return to OR for screw malposition changed from 2003 to 2017 (1-0.2%) but this did not reach statistical significance. Conclusion Although the incidence of asymptomatic malpositioned screws is unknown, the rate of return to the OR for implant malposition overall in this registry was 1 in 250 patients. Recent data suggest the rate has decreased to 1 in 500 patients. Further work may determine whether enabling technologies can reduce the rate, ideally to a "never event".