1032

Abstract

Introduction: Procalcitonin (PCT) levels correlate with the extent and type of systemic inflammation associated with bacterial infections and outcomes in both adult and pediatric populations. However, limited data is available on the diagnostic and prognostic value of PCT in critically ill cancer patients with sepsis. Methods: As part of an internal validation study to investigate the role of PCT in the ICU, we prospectively enrolled patients with cancer and suspected sepsis who were admitted to a 20-bed medical-surgical oncologic ICU between March and August 2013. Following informed consent, plasma PCT samples were collected daily with routine blood draws starting within 24 hrs of ICU admission through day 7 or until ICU discharge. PCT results were not utilized for patient care. Demographics, clinical data, severity of illness scores on ICU admission, PCT levels and outcomes of 56 septic patients were analyzed retrospectively. Seven non-septic patients served as controls. Patients were divided into 4 groups: non-septic controls (n=7), SIRS/sepsis (n=10), severe sepsis (n=35), and septic shock (n=11). Mean PCT values on days 1 and 3 and outcomes were compared between survivors and nonsurvivors (ICU and hospital) and correlated with the severity of sepsis. PCT levels >0.5ng/dl were considered abnormal. Student t-test, chi-square test and Mann-Whiney tests were used for statistical analyses. P-value of <0.05 was considered significant. Results: Of the 56 septic patients, 34 (61%) were males; mean age was 61 years and 34 (61%) were medical oncology patients. 30 (54%) patients received recent chemotherapy and 10 (18%) had neutropenia. Mean MPM II and SOFA scores on ICU admission were 40% and 7.6, respectively and lactate was 2.5 mmol/L. 80% had culture documented infections. ICU and hospital mortality rates were 25% and 37.5%, respectively. Nonsurvivors had a mean PCT of 31 ng/dl on day 1 and 28 ng/dl on day 3 vs. 27 ng/dl and 24 ng/dl, respectively for survivors (p=.318 and p=.296). There was an exponential correlation between the mean PCT levels in controls (0.2 ng/dl), patients with SIRS/sepsis (5.4 ng/dl), severe sepsis (18.9 ng/dl) and septic shock (36.8 ng/dl) with a correlation coefficient of 0.99. Conclusions: Our validation study of a cohort of critically-ill patients with cancer shows a correlation between PCT levels and severity of sepsis, similar to the general critically ill population. However, PCT values were not predictive of mortality. Further study is needed to determine the diagnostic utility of PCT in the critically-ill oncologic population.