Defining a Surgical Invasiveness Threshold for Increased Risk of a Major Complication Following Adult Spinal Deformity Surgery

Brian J. Neuman, MD; Andrew B. Harris, BS; Eric O. Klineberg, MD; Richard A. Hostin, MD; Themistocles S. Protopsaltis, MD; Peter G. Passias, MD; Jeffrey L. Gum, MD; Robert A. Hart, MD; Michael P. Kelly, MD; Alan H. Daniels, MD; Christopher P. Ames, MD; Christopher I. Shaffrey, MD; Khaled M. Kebaish, MD


Spine. 2021;46(14):931-938. 

In This Article


This study was approved by the institutional review board at each participating institution.

Patient Selection

We retrospectively reviewed a longitudinally collected, multicenter database of patients enrolled in a study evaluating operative versus nonoperative treatment of ASD. Patients were enrolled at 11 sites and treated by 15 surgeons across the United States. The following inclusion criteria were used for patients who underwent operative treatment: age >18 years; more than 5 spinal levels fused; and diagnosis of degenerative or idiopathic scoliosis. In addition, patients had to meet at least one of the following radiographic parameters for inclusion: sagittal vertical axis (SVA) >5 cm, Cobb angle ≥ 20°, pelvic tilt (PT) >25°, or thoracic kyphosis >60°. Patients were excluded if they had spinal deformity secondary to malignancy, neuromuscular causes, or deformity due to infectious causes.

At the time of writing, 1281 patients had been enrolled in the study; however, some of these patients underwent surgery more recently than 2 years ago and were not yet eligible for 2-year clinical follow-up. Of the 760 patients who were eligible for 2-year follow-up based on their date of surgery, 574 (76%) had complete preoperative and 2-year clinical data and were included in our analysis.

Radiographic Evaluation

Radiographic measurements were taken using full-length standing scoliosis lateral radiographs preoperatively and at 6-week follow-up. Measurements were performed using a validated software system (Spine-View, ENSAM Paritech, Paris, France)[18,19] at a central location with a dedicated study team that was not involved in the care of enrolled patients.

HRQoL Measures

Two standardized HRQoL questionnaires, the Oswestry Disability Index (ODI) and the Scoliosis Research Society version 22-revised (SRS-22r), were used to evaluate patients preoperatively and at 2-year follow-up. Both are validated questionnaires that are widely used when studying outcomes in patients with ASD.[20,21]

ASD-SR Index

The ASD-SR index assigns point values to 13 operative and radiographic factors to quantify the degree of invasiveness of surgical ASD procedures. Point values for surgical and radiographic components of the ASD-SR index were determined by using beta coefficients from multivariate regression models to predict estimated blood loss (EBL) and operative time.[11] The surgical components of the ASD-SR index are decompression, fusion, osteotomies performed, the use of interbody spacers, iliac fixation, and revision surgery. The radiographic components of the ASD-SR index are the preoperative to postoperative changes in pelvic incidence minus lumbar lordosis (PI-LL), pelvic tilt (PT), sagittal vertical axis (SVA), and thoracic kyphosis (TK) (Table 1).


Data on complications were collected prospectively, and complications were classified as major or minor. Major complications were defined as those requiring return to the operating room, those necessitating a prolonged hospital stay, or those that were unresolved at the time of data analysis. Minor complications were defined as those requiring only minor changes to routine clinical care.

Statistical Analysis

Descriptive statistics are reported as mean ± standard deviation or number (%) unless stated otherwise. A cutoff point that optimized sensitivity and specificity of predicting major complications was determined using a receiver operating characteristic curve (Youden index).[22,23] Patients with scores above this cutoff comprised the "high invasiveness" group, and those with scores below this cutoff comprised the "low invasiveness" group. These 2 groups were divided in half to create four quartiles. Patient-reported outcomes were compared among quartiles using one-way analysis of variance. χ2 tests were used to compare the prevalence of major, minor, and individual complications above and below the ASD-SR cutoff score. Student t tests were used to compare patient-reported outcome scores between the high- and low-invasiveness groups. Odds of developing a major complication in each invasiveness quartile were analyzed using binary logistic regression, controlling for baseline ASD frailty index values[24] and radiographic deformity measures. Statistical analysis was performed using Stata software, version 15 (StataCorp, LLC, College Station, TX). Significance was considered at P < 0.05.