Are deep infections that present before and after 90 days from orthopaedic trauma different? An analysis of the validity of the recent change in CDC criteria for infections.

Background: In 2016, the Centers for Disease Control and Prevention (CDC) changed the time frame for their definition of deep surgical site infection (SSI) from within 1 year to within 90 days of surgery. We hypothesized that a substantial number of infections in patients who have undergone fracture fixation present beyond 90 days and that there are patient or injury factors that can predict who is more likely to present with SSI after 90 days.

Methods: A retrospective review yielded 452 deep SSI after fracture fixation. These patients were divided into two groups-those infected within 90 days of surgery and those infected beyond 90 days . Data were collected on risk factors for infection. Univariate and multiple logistic regression analyses were performed to compare the two groups. A randomly selected control group was used to build infection prediction models for both outcomes. The two outcomes were then modelled against each other to determine whether differences in predictors for early versus late infection exist.

Results: Of the 452 infections, 144 occurred beyond 90 days (32% [95% CI, 28%-36%]). No statistically significant patient factors were found in multivariable analysis between the early and late infection groups. The need for flap coverage was the only injury characteristic that differed significantly between groups, with patients in the late infection group more likely to have needed a flap. When modelled against the control group and directly comparing the two models, predictors for early infection include male sex and fractures of the pelvis, acetabulum, or hip, whereas predictors of late infection include hepatitis C and/or human immunodeficiency virus (HIV) and admission to the intensive care unit (ICU).

Conclusion: Use of the recent CDC definition will underestimate the rate of actual postoperative infections when applied to orthopaedic trauma patients. Hepatitis C and/or HIV and ICU admission are predictors of late infection, whereas male sex and pelvis, acetabulum, or hip fractures are predictors of early infection. Patients who receive flap coverage may be more likely to present with late infection.