Erratum to ‘Cost differences between autologous and nonautologous cranioplasty implants: A propensity score–matched value driven outcomes analysis’ [World Neurosurgery: X (22C) (2024) (100358)]

Abstract

Objective

Despite having higher rates of failure and revision than other implant materials, autologous bone implants are the historical gold standard for restoring cranial defects after decompressive hemicraniectomy. More reliable synthetic implants have been developed, but they are significantly more expensive than autologous bone. The authors sought to compare the initial and long-term costs of custom synthetic implants and autologous cranioplasty grafts.

Methods

The authors reviewed the hospital billing records over a 12-year period at their institution to identify patients who underwent cranioplasty after decompressive hemicraniectomy for trauma or stroke. The costs for imaging, hospital supplies, implants, pharmacy services, facility usage, and laboratory studies were captured for initial surgeries and for subsequent cranioplasty-related revision hospitalizations and surgeries. Clinical characteristics, long-term outcomes, and cost differences of autologous versus custom implants were compared by using univariate and multivariate analyses. These analyses were repeated after propensity-score matching adjusted for factors predictive of cranioplasty failure.

Results

On unmatched analysis, 32 custom and 128 autologous implants were analyzed. Differences in initial cranioplasty failure rates between custom and autologous grafts were insignificant (12.5 % custom vs. 23.4 % autologous, p = 0.18). On univariate analysis, autologous implants cost 46.8 % of custom grafts during the initial hospitalization period (p < 0.01) and 58.7 % of custom grafts once long-term aggregated costs were factored (p < 0.01). Upon multivariate analysis, although custom cranioplasty was independently predictive of higher initial hospitalization costs (standardized β = 0.20, p < 0.01), it was not predictive of long-term total aggregated costs (standardized β = 0.10, p = 0.13). After propensity score matching (29 custom, 29 autologous cases), multivariate analysis of the matched cohorts likewise found custom implants were predictive of greater initial hospitalization costs (standardized β = 0.56, p < 0.01), but were not an independent driver of aggregated long-term costs (standardized β = 0.22, p = 0.11).

Conclusion

Our matched and unmatched multivariate analysis found nonsignificant cost differences between custom and autologous cranioplasties once initial and long-term costs were aggregated, suggesting custom implants could be considered as a primary therapeutic intervention.