Defining a High Risk Adult Spinal Deformity Patient: Utilizing Predictive Analytics to Assess Factors Associated with Increased Risk of Medical and Surgical Complications Following Corrective Surgery

Thursday, June 1, 2023

8:30 AM – 6:15 PM

Location: Salons 7-8

Co-Author(s)

PT

Peter S. Tretiakov, BS

United States
OK

Oscar Krol

United States
PD

Pooja R. Dave

Miami, Florida, United States
JM

Jamshaid M. Mir, MD

Research Fellow
NYU Langone Health | NY Spine Institute, United States
JL

Jordan Lebovic, MD, MBA

Resident
NYU Langone Orthopedics
New York, New York, United States
SV

Shaleen Vira, MD

Assistant Professor, Orthopaedic Surgery & Neurosurgery
UT Southwestern Medical Center
Dallas, Texas, United States
MJ

M. Burhan Janjua

United States
AM

Angel Macagno

United States
TR

Timothy Roberts

Orthopaedic Spine Surgeon
Westbury, New York, United States
NP

Nicholas Post

United States
AD

Alexandre De Moura

United States
AS

Andrew J. Schoenfeld, MD, MSc

Associate Professor
Brigham and Women's Hospital, Department of Orthopaedic Surgery
Boston, Massachusetts, United States

Presenting Author (Person doing the Podium Present(s)

PP

Peter G. Passias, M.D.

Spine Surgeon
NYU Langone Orthopedic Hospital
New York, New York, United States

View E-Poster

Submitter(s)

PP

Peter G. Passias, M.D.

Spine Surgeon
NYU Langone Orthopedic Hospital
New York, New York, United States

View E-Poster

Disclosure(s):

Peter G. Passias, MD: No relevant disclosure to display

Background: High risk patients may have increased operative risk and recruit a multi-disciplinary discussion. Both surgical and medical risk factors can lead to the occurrence of adverse events and prolonged recovery course. Little consensus has been reached as to which components of patient profiles and surgical factors predispose pre-operative discussion.

Purpose: To investigate baseline and surgical factors that will identify which ASD patient presents as a high-risk patient.

Methods: Operative ASD patients (scoliosis >20º, SVA>5cm, PT>25º, or TK>60º) with available baseline (BL) and 2-year (2Y) radiographic and HRQL data were included. Determination of risk of surgery was assessed by medical and surgical complications. High medical risk was defined as experiencing a major medical complication within 90 days of surgery. High surgical risk was defined as experiencing a major surgical complication or revision surgery within 90 days of surgery. Conditional inference tree machine learning developed threshold cutoffs for continuous variables. Multivariable logistic regression developed medical risk and surgical risk scores.

Results: 377 ASD patients met inclusion criteria (61.1yrs±14.0, 78%F, BMI: 28.0 kg/m2 ±6.0, CCI: 1.9 ±1.7). Surgically, patients had 11±4.3 levels fused, length of stay was 7.9±4.6, EBL 1476±1434ml, and Op-time 376±133 min, with 63% undergoing an osteotomy, 54% a decompression. In terms of surgical approach, 71% posterior-only and 28% had a combined approach. For High Medical Risk, age >70, BL Frailty Index >5, CCI >3, prior revision, history of heart disease, liver disease, lung disease, BL SVA >15cm, BL C7PL >7cm, and BL PI-LL >25 predicted a poor outcome with an AUC of 97% and accuracy of 98%. Patients with at least 1 of these factors had a greater degree of major (21% vs 5%), mechanical (19% vs 10%), and overall complications by 2 years (64% vs 46%, all p<.05). For High Surgical Risk, age >70, BL BMI >34, Frailty >5, BL SVA >15cm, BL C7PL >7cm, BL PI-LL >25, previous surgical fusion, >16 levels fused, 3CO, instrumentation removal, and >3 interbody fusions predicted poor outcome with an AUC of 93% and accuracy of 94%. Patients with at least 2 of these factors had higher rates of major (20% vs 9%), mechanical (20% vs 10%), and overall complications (65% vs 48%, all p<.05).

Conclusion: Recognition of patient-specific and surgical factors that contribute to a high risk of major medical and surgical complications with poor clinical outcomes will allow surgeons to better profile which patients may require multi-disciplinary collaboration for appropriate perioperative optimization. The present study reveals patients greater than 70 years of age, increased baseline ad deformity, heightened frailty status, and planned 3-column osteotomy are at increased risk of major, mechanical, and overall complications.