Reducing Infection Rates in Implant-Based Breast Reconstruction: Impact of an Evidence-based Protocol.

Background: Infection is the most significant complication in implant-based breast reconstruction, potentially leading to reconstructive failure. We hypothesized that implementation of an evidence-based protocol marked by preoperative decolonization and sterility optimization would result in a decline in postoperative infection rates.

Methods: Informed by a literature review, we developed an evidence-based, perioperative infection prevention protocol implemented in 2015. Surgical outcomes were compared between patients who had undergone implant-based breast reconstruction before and after protocol implementation. A Fisher exact test was used to compare infection rates before and after protocol implementation. A logistic regression analysis was modeled to evaluate the impact of the protocol on infection rate while controlling for nonmodifiable risk factors.

Results: Three hundred fifty-eight breasts underwent reconstruction before protocol implementation and 135 afterward. Patients were similar in terms of demographics and surgical characteristics. There was a significantly reduced incidence of clinically relevant infection after protocol implementation (9.5%-2.9%, P = 0.013). Logistic regression analysis confirmed that the protocol was independently associated with a decrease in infection risk (odds ratio, 0.244; P = 0.021). After protocol implementation, no gram-positive bacteria were isolated among cultures obtained from infected periprosthetic fluid. Radiation and drain duration greater than 21 days were independently associated with greater risk for infection.

Conclusions: Our evidence-based protocol was associated with a significant decline in infection rates among implant-based breast reconstruction patients and was particularly effective for gram-positive infections. We will continue to use this protocol in our practice and will consider future directions for addressing gram-negative infections as well.