General Session: Adult Spinal Deformity - Hall F

Presented by: A. Lovi

Author(s):

A. Lovi(1), L. Babbi(1), G. Ristori(1), S. Razza(1), A. Luca(1), M. Brayda-Bruno(1), Spine Care Group

(1) IRCCS Orthopaedic Galeazzi Institute, Milan, Italy

Abstract

Background data: Spinal surgery is typically perceived as a risky procedure but the lack of standardized definitions and inconsistent methodology of data reporting have led to a significant variation in adverse events (AEs) recording. Retrospective analysis significantly underestimates the risk of complications, and then prospective studies, focused on AEs, are recommended.The Spine Adverse Events Severity System version 2 (SAVES V2) is a simple validated form to collect and report AEs related to spinal surgery. It has been shown to improve AEs detection, compared with standard patient chart abstraction, and has demonstrated an excellent interobserver reliability.

Objective: To determine the prevalence and severity of perioperative AEs associated with pediatric and adult spine surgery. To estimate the impact of perioperative AEs on length of hospital stay (LOS).

Study Design: Prospective observational, monocenter study.

Methods: The study included 346 consecutive patients, admitted between November 1-2014 and November 31-2016 for elective spinal surgery at a sigle Institution. 294 were adults (age: 48 ± 22 years old), 52 were pediatric patients (age: 14±3 years old). Perioperative AEs were collected and reported on SAVES V2 forms by the surgical team during hospitalization. Based on the recorded AEs, the effect on LOS was estimated. AEs occurrence was analysed in relationship with demographics, diagnosis, ASA score and surgical treatment. Statistical tests were based on a two sided significance level of 0.05.

Results: 21,2% (11) of pediatric patients and 20,7% (61) of adults had at least 1 perioperative AEs. 2 pediatric and 10 adult patients had more than 1. In adults, dural tear (3.1%) and neuropathic pain (4,8%), were the most frequent intraoperative and postoperative AE, respectively. In pediatric patients, neurologic deterioration was the most frequent postoperative AE. AEs occurred more often in patients with an ASA score higher or equal to 3. Deformity correction and thoraco-lumbar surgery were associated with higher risk of AEs. Severe AEs (Grade 3, 4 and 5) occurred more frequently in pediatric patients than in adults. Perioperative AEs required a revision surgery in 4,9% (17) of patients, and a prolonged LOS in 55,5% (40) of cases (mean of 6 additional days of hospital stay).

Conclusions: This study provided an assessment of perioperative AEs associated with adult and pediatric spinal surgery based on a prospective design. The SAVES system was a simple and reliable tool. The overall complication rates during hospitalization were high, but not equally impactful; pediatric patients had the most severe. The results of this study may be useful to improve the safety and cost-effectiveness of patient care.

Abstract

Introduction: Prior to surgery, stratifying patients according to risk profiles for inpatient medical and surgical complications and length of stay will allow surgeons to better pre-operatively counsel patients and evaluate the risk/benefit balance of proposed treatment plans. The aim of this study was to validate a composite risk index which stratifies patients undergoing cervical spine surgery based on predictive factors for medical complications, surgical complications, and length of stay.

Methods: Retrospective review of the Statewide Planning and Research Cooperative System (SPARCS) 2014 and National Inpatient Sample (NIS) 2013 data. Patients aged 18 years and older undergoing any cervical fusion or decompression surgery on the NIS database were included. Five multivariate models (controlling for demographic characteristics, comorbidities, number of levels fused, surgical approach, procedure performed, instrumentation and use of biologics) determined independent clinical factors that increased risk of >1 for medical complications, surgical complications and length of stay. 50 points were distributed among the predictors based on their accumulative odds ratio. Area under the curve (AUC) analysis was used to internally validate each of the prediction models, reported with the 95% confidence interval.

Results: 37,815 patients and 10,790 patients were identified from the NIS and SPARCS databases, respectively (mean age 56.2 and 54.5, 50.6% and 51.6% male). The leading predictor in each category were: weight loss for both prolonged length of stay (OR 16.8) and medical complications (OR 6.35) and revisions for surgical complications (OR 18.03). Weight loss was a significant predictor in the two categories regardless of the presence of malignancy (medical complication: OR 1.64, extended length of stay: OR: 4.21). The ten most significant risk factors were: weight loss (27.0), revision-status (21.5), trauma (19.4), paralysis (14.0), pulmonary comorbidity (12.8), cerebrovascular disease (6.8), and congestive heart failure (6.1). These risk factors were scored 9, 7, 6, 5, 4, 2, and 2 respectively. Three risk thresholds were proposed: mild (0-10), moderate (10-20), and severe (>20/50 points). Medical complications, surgical complications, and length of stay were accurately predicted (AUC=70% [70-69], 64% [64-63], and 67% [68-67], respectively).

Conclusions: This study proposes and validates a novel index to quantify morbidity risk prior to cervical spine surgery. This index incorporates risk attributable to medical comorbidities, those directly surgically related, and is the first such index to incorporate prediction of extended length of stay. This will allow surgeons to help patients, hospitals and insurance companies in determining risk/benefit profiles and predicting patient outcomes.