PS1273 PREDICTIVE FACTORS FOR LINE‐ASSOCIATED BLOODSTREAM INFECTIONS IN PATIENTS WITH HAEMATOLOGIC MALIGNANCIES: COMPARISON OF TUNNELLED CATHETERS AND PERIPHERALLY INSERTED CENTRAL CATHETERS

Background: Patients with haematological malignancies often require central venous access for administration of chemotherapy and blood products. Infection remains one of the commonest complications of these devices and is associated with significant morbidity and sometimes mortality. Aims: The purpose of our study is to identify clinical and laboratory factors which may predict a higher risk of line‐associated bloodstream infections in patients with haematological malignancies. Methods: Patients treated for haematological malignancies at our centre between January 2016 and December 2018 were retrospectively identified. Only patients who had either a tunnelled line or peripherally‐inserted central catheter (PICC) were included in the study. Line‐associated bloodstream infections were recorded and the predictive value of a number of baseline clinical and laboratory parameters for infections was assessed. Results: One hundred and fifty‐three patients were identified during the study period. The haematological diagnosis was as follows: acute leukaemia in 63 (41.2%) patients, lymphoma in 81 (52.9%) and other diagnoses in nine (5.9%). One hundred and twenty‐five patients (81.7%) had a tunnelled line while 28 patients (18.3%) had a PICC. The total duration of follow‐up was 10,416 line‐days with a median of 55 days. The median in‐situ duration for tunnelled lines was 38 days (range 4–150 days) while that for PICCs was 17 days (range 2–145 days) ( P  = 0.02). Bloodstream infections were identified in 58 (37.9%) patients, with 28 patients (18.3%) satisfying the criteria for a catheter‐related blood‐stream infection (CRBSI). The overall infection rate was 5.5 infections per 1,000 line‐days. There was no significant difference in the rates of infection between tunnelled lines and PICC lines ( P  = 0.25). On multivariate analysis for the whole group, line‐associated infections were more likely in patients with acute leukaemia ( P  < 0.001), those receiving corticosteroids as part of the chemotherapy regimen ( P  = 0.01) and those with neutropenia (absolute neutrophil count <1.0 x 10 9 /L) at the time of line placement ( P  = 0.03). On univariate analysis, predictive factors for line infections in tunnelled catheters included acute leukaemia ( P  < 0.001), steroid use ( P  = 0.01), younger age ( P  = 0.04) and total number of line‐days with neutropenia ( P  = 0.04). On multivariate analysis only acute leukaemia ( P  < 0.001) and steroid use ( P  = 0.06) retained independent prognostic significance. For PICCs only neutropenia at the time of placement was predictive for an increased risk of infection ( P  = 0.01). Summary/Conclusion: In conclusion this study identified acute leukaemia, corticosteroid use and neutropenia at the time of catheter placement as clinical predictors of infection in patients with haematological malignancy.