Revisiting the association of preoperative steroids, severe neutropenia, and prophylactic antibiotics on pediatric port infections

The impact of preoperative absolute neutrophil count (ANC), steroids, and prophylactic antibiotics on port infection risk remains controversial. As infections represent a significant complication of pediatric port placement, we sought to identify specific risk factors associated with port infection at our institution.

Ports placed from 2013 to 2024 at our children's hospital were retrospectively reviewed. Exclusion criteria was absence of 72 h preoperative ANC data. Severe neutropenia was defined as ANC<500 cells/mm³. Multivariable logistic regression identified risk factors associated with port infection and removal within 30 days of placement.

Of 1566 ports meeting inclusion criteria, indications were liquid malignancy (44.3%), solid tumors (42.2%), and benign diagnoses (13.5%); 16.4% were placed in context of severe neutropenia. The overall 30-day port infection rate was 5.6%. Of infected ports, 31.0% required removal. Children with severe neutropenia were significantly more likely to develop port-associated infections (aOR 1.832 [1.010-3.325], p = 0.0464). This risk persisted in those who did not receive prophylactic antibiotics (aOR 2.126 [1.064-4.249], p = 0.0328) but became non-significant in the subgroup receiving antibiotic prophylaxis (p = 0.7866). Steroid use was associated with increased odds of port infection (aOR 1.705 [1.064-2.734], p = 0.0266) and removal (aOR 3.592 [1.570-8.216], p = 0.0025).

In a large-scale cohort of pediatric port placement, steroids and severe neutropenia increased the odds of port infection while prophylactic antibiotics helped mitigate this risk. These findings demonstrate a potential protective benefit of prophylactic antibiotic use and support consideration of delayed port placement in severely neutropenic children. Prospective studies are warranted to clarify these relationships and inform clinical standard practice recommendations in children.