Infectious Central Venous Line Complications During Induction Therapy for Newly Diagnosed Acute Leukemia

Childhood acute leukemia treatment requires central venous lines (CVL) for instillation of chemotherapy and blood products. Ideally, a proper white cell count (WBC) and absolute neutrophil count (ANC) ensure proper healing of CVLs, but this is challenging in children with acute leukemia. We sought to investigate the CVL complication rate in newly diagnosed children with acute leukemia during their induction therapy, and determine if the degree of neutropenia at the time of CVL placement correlated with the number of CVLs lost due to infection, wound dehiscence, or thrombosis.

We conducted a retrospective chart review of children diagnosed with leukemia between January 2007 and December 2009 and recorded leukemia type, WBC and ANC at diagnosis and at the time of CVL placement, the type of CVL placed (external line, subcutaneous port) or placement of peripherally inserted central (PICC) line. We recorded complications, including infection, line malfunction, wound dehiscence, and thrombosis within their induction therapy phase.

Ninety-five children were evaluable, including 68 children with precursor B acute lymphoblastic leukemia (pre B ALL), 19 with acute myelogenous leukemia (AML), and 8 with T-cell acute lymphoblastic leukemia (T cell ALL). Ninety-eight CVLs were placed in 94 children (1 child died of complications of APML before initiation of therapy). There were 77 subcutaneous ports and 21 external lines placed. Eleven patients received PICC lines for various reasons (ex – sedation risk due to large mediastinal mass or altered mental status due to leukocytosis, coagulopathy, refractory thrombocytopenia, previously placed PICC line at outside hospital). ANC at the time of CVL insertion was reviewed: ANC<500 in 39 central lines, 500–1000 in 29 central lines, and >1000 in 30 central lines. Only 1 central line was removed due to wound dehiscence in a child with T cell ALL, and 2 central lines were removed for cellulitis in children with pre B ALL, and all these patients had ANC<500 at the time of line insertion. Two of the 98 central lines developed an associated thrombosis (1 CVL associated extensive arm venous thrombus and 1 external line with small atrial thrombus at tip of catheter), as opposed to 2 of the 11 PICC lines placed (both extensive arm venous thrombi). Seventeen positive blood cultures occurred during the first month of induction (15 from central lines and 2 from PICC lines), and all infections cleared with antibiotics except for 1 patient with PICC-associated venous thrombosis and persistent MRSA bacteremia. One subcutaneous port had to be revised after 3 days due to deep insertion and difficultly accessing; this child had ANCs<500 during each surgery and healed without complications. Three external lines were removed due to malfunction (2 with ANC<500, 1 with ANC 500–1000 at time of insertion).

Nearly 40% of CVLs were placed in times of severe neutropenia (ANC<500), and only 3 were lost due to cellulitis or wound dehiscence. No CVL was lost due to persistent bacteremia compared to 1 PICC line. There was an increased incidence of thrombosis in PICC lines (2 of 11 placed) compared to external lines or ports (2 of 98 lines placed). We failed to see an increased risk of infection due to degree of neutropenia at the time of CVL insertion.

No relevant conflicts of interest to declare.