Authors: (including presenting author): :
Leung MKM(1), Lam RPK(1)(2), Ng DSY(1), Tang HT(1), Chow TYJ(1), Yau WK(3), Leung RYY(3), Rainer TH(1)(2), Tsui SH(1), Tsang TC(1)
Affiliation: :
(1)Accident and Emergency Department, Queen Mary Hospital, Hong Kong S.A.R. (2)Department of Emergency Medicine, School of Clinical Medicine, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong S.A.R. (3)Department of Pathology, Queen Mary Hospital, Hong Kong S.A.R.
Keyword 1: :
Neutropenic Fever
Keyword 2: :
Antimicrobial Stewardship
Keyword 3: :
Point-of-Care Testing
Keyword 4: :
Cancer / malignancy
Keyword 5: :
Accident and Emergency
Introduction: :
Neutropenic fever after chemotherapy requires prompt antibiotic administration. To achieve a door-to-antibiotic time of < 1 hour in patients with post-chemotherapy fever, ultrabroad spectrum antibiotics are often administered before laboratory absolute neutrophil counts (ANCs) are available in Accident and Emergency (A&E) Departments under the Hospital Authority. This might cause antibiotic overtreatment of patients who turn out to be non-neutropenic.
Objectives: :
We aimed to evaluate the accuracy of a point-of-care test (POCT), the HemoCue® WBC DIFF system, in identifying neutropenia in post-chemotherapy fever.
Methodology: :
A single-center retrospective cross-sectional study of post-chemotherapy febrile adults was conducted from September 2024 to September 2025 at Queen Mary Hospital A&E. Venous and finger-prick capillary blood samples were collected for laboratory testing and the POCT, respectively. The reference standard was the laboratory ANC. The primary outcome was the accuracy of POCT based on the intraclass correlation (ICC) between the POCT and laboratory ANC. The POCT’s sensitivity, specificity, positive and negative predictive values (PPV and NPV), and positive and negative likelihood ratios (LR+ and LR-) in identifying neutropenia (ANC < 1.0 × 109/L) were calculated.
Result & Outcome: :
Among 173 patients, POCT and laboratory ANC results were highly correlated (ICC 0.863, 95% confidence interval [CI] 0.732 to 0.921). The POCT was accurate, with a sensitivity of 90.3% (95% CI 73.1% to 97.5%), specificity of 97.9% (95% CI 93.5% to 99.5%), PPV of 90.3% (95% CI 73.1% to 97.5%), NPV of 97.9% (95% CI 93.5% to 99.5%), LR+ of 42.8 (95% CI 13.9 to 131.8) and LR- of 0.1 (95% CI 0.03 to 0.3). POCT could have prevented ultrabroad spectrum antibiotic use in up to 139 (80.4%) cases. Based on the results, we have introduced a new workflow integrating the POCT to identify non-neutropenic cases in non-haematological malignancies at Queen Mary Hospital since 12th January 2026. Further evaluation of the impact on patient outcomes and antibiotic stewardship is ongoing.
