Hot Patient? Stay Cool.
Author(s):
Mikayla Moody; Robert Sawyer; Saad Shebrain
Background:
Prolonged fever or leukocytosis in patients treated for infection concerns clinicians because of a possible missed source of infection, a secondary infection, or inadequate antimicrobial therapy.
Hypothesis:
We hypothesized that these patients suffer worse outcomes, including higher mortality and longer antimicrobial treatment.
Methods:
From 2017 to 2022, data were prospectively collected on patients treated for infection in the SICUs of two university-affiliated hospitals. Demographics, microbiological data, severity of illness (APACHE-II score), and outcomes were recorded. Patients with white blood cell count (WBC) ≥15.0×109/L or a maximum temperature (TMAX) ≥38.5° C were divided into quartiles based on days until reduction of WBC to ≤15.0×109/L and days until normalization of temperature to ≤38°C for a full calendar day. Univariate analysis followed by logistic regression (LR) analyses were performed to predict factors associated with prolonged fever or leukocytosis, subsequent/secondary infections, and in-hospital mortality. Model performance was assessed using the Hosmer and Lemeshow test and ROC curve analysis.
Results:
697 patients were identified: 343 with WBC ≥15.0×109/L and 238 with TMAX ≥38.5° C. Mean time to normalizing WBC was 5.5±0.4 days with the highest quartile ≥7 days (prolonged leukocytosis-PL). The mean time to resolution of fever was 4.1±0.3 days with the highest quartile ≥4 days (prolonged fever-PF). By LR analysis, younger age, initial WBC, hospital days until diagnosis, and splenectomy were independently associated with PL; younger age and prior transfusion were associated with PF. Mortality in patients with leukocytosis was 21.9% and in patients with fever, 14.7%. By LR, only increasing age and APACHE-II score (but not PL or PF) were associated with increased mortality. Patients with PL received more days of antibiotics compared to patients without PL (18.5±3.0 vs. 9.5±0.5, p < 0.0001). Neither PL nor PF was associated with the development of a subsequent infection.
Conclusions:
While patients at risk for prolonged fever or leukocytosis could be identified, neither of these events was associated with worse outcomes. This may reflect inter-individual variations in the host inflammatory response rather than a marker of infection severity or inadequate treatment. When faced with a patient with prolonged fevers or leukocytosis, the best course of action is to treat these patients based on established evidence-based guidelines.