Study design: Retrospective cohort study of 13,927 patients, 820 of whom received iliac crest bone graft (ICBG).
Objective: To compare adverse events, length of stay (LOS), and readmission for patients receiving ICBG with those who did not using multivariate analysis to control for potentially confounding factors.
Summary of background data: The use of ICBG in spinal fusion has been associated with increased surgical time, LOS, and donor site morbidity. Development of expensive bone graft substitutes has been predicated on these issues. Data on the effect of bone graft harvest on LOS and readmission rate are sparse, and multivariate analysis has not been used to control for confounding factors.
Methods: Prospectively collected data from the American College of Surgeons National Surgical Quality Improvement Project 2010-2012 database were retrospectively reviewed. This includes demographics, comorbidities, surgical data, and hospital and 30-day follow-up outcomes data including adverse events, LOS, and readmission.
Results: Only 5.9% of spinal fusions use ICBG. Bivariate logistic regression (used for categorical variables) found the ICBG cohort was more likely to have a postoperative blood transfusion (11.6% vs. 5.5%, P < 0.001). Bivariate linear regression (used for continuous variables) found the ICBG cohort to have an extended operative time (+36.0 min, P < 0.001) and extended LOS (+0.6 d, P < 0.001).Multivariate analyses controlling for comorbidities, demographics, and approach-determined postoperative blood transfusion (odds ratio, 1.5), extended operative time (+22.0 min, P < 0.001), and LOS (+0.2 d, P = 0.037) to be significantly associated with ICBG use.No other adverse event was significantly associated with ICBG use. Readmission rates were not significantly different.
Conclusion: This study used a large national database cohort and identified increased postoperative blood transfusion, extended operative time, and increased LOS as short-term outcomes associated with ICBG on multivariate analysis. Other short-term morbidities were not significantly associated with ICBG. Readmission rates were not affected.
Level of evidence: 4.