Readmission following complex spine surgery in a prospective cohort of 679 patients – 2-years follow-up using the Spine AdVerse Event Severity (SAVES) system

Abstract

BACKGROUND CONTEXTRecent studies suggest that prospective registration more accurately reflects the true incidence of adverse events (AEs). To our knowledge, no previous study has investigated prospectively registered AEs' influence on hospital readmission following spine surgery. PURPOSETo determine the frequency and type of unplanned readmissions after complex spine surgery, and to investigate if prospectively registered AEs can predict readmissions. DESIGNThis is a prospective, consecutive cohort study. PATIENT SAMPLEWe conducted a single-center study of 679 consecutive patients who underwent complex spine surgery defined as conditions deemed too complicated for surgery at a secondary institute, or patients with severe comorbidities requiring multidisciplinary observation and treatment. OUTCOME MEASURESThe outcomes in this study were (1) readmission to any hospital department within 30 days of discharge and (2) readmission to a surgical spine center at any time in follow-up. METHODSAll patients undergoing complex spine surgery, at our tertiary referral center, were consecutively, and prospectively, included from January 1 to December 31, 2013. Demographics and perioperative AEs were registered using the Spine AdVerse Events Severity (SAVES) system. Patients were followed for a minimum of two years. A competing risk survival model was used to estimate rates of readmissions with death as a competing risk. Patient characteristics, surgical parameters and perioperative AEs were analyzed to identify factors associated with readmission. Analyses of 30-day readmission were performed using logistic regression models. A proportional odds model, with death as competing risk, was used for readmissions to a spine center at any time in follow-up. Results were reported as odds ratios with 95% confidence intervals (95% CI). RESULTSWithin 2 years of index discharge, 443 (65%) were readmitted. Only 20% of readmissions were to a spine center. Cumulative incidence (95% CI) of readmission was estimated to 13% (10%–16%) at 30 days, 26% (23%–30%) at 90 days, 50% (46%–54%) at 1 year, and 59% (55%–63%) at 2 years following discharge. Rates were markedly lower for readmissions to a spine center. Increased odds of 30-day readmission were correlated to intraoperative hypotension (p=.02) and major intraoperative blood loss (p<.01). Readmission to a spine center was associated with the number of instrumented vertebrae (p=.047), major intraoperative AE (p=.01), and intraoperative hypotension (p<.01). CONCLUSIONSTo the best of our knowledge, this is the first study to analyze prospectively registered AEs' association to readmission up to 2 years after complex spine surgery. We found that readmissions were more frequent than previously reported when including readmissions to any department or hospital. Factors related to major intraoperative blood loss were associated to increased odds of readmission. This should be considered during planning of postoperative observation and care.