Abstract

The aim of this study was to investigate the value of a platelet count (PLT) in the early diagnosis of nosocomial invasive fungal infections in premature infants. Based on clinical diagnosis combined with blood culture results, 72 premature infants of 5354 pediatric patients who were hospitalized in the neonatal ward of our hospital between September 2009 and February 2013 were diagnosed with nosocomial invasive fungal infections (fungal infection group). There were 58 premature infants diagnosed with bacterial infections during the same period (bacterial infection group). The control group included 74 premature infants without nosocomial infections who were hospitalized during the same period. Receiver operating characteristic (ROC) curves were used to analyze the sensitivity, specificity, and diagnostic efficacy of the PLT and white blood cell (WBC) counts and C-reactive protein (CRP) level in the diagnosis of fungal infections in premature infants. The risk factors for invasive fungal infections included birth weight < 2000 g, gestational age < 32 weeks, peripherally inserted central catheter (PICC), oxygen inhalation therapy, intravenous nutrition, and administration of antibiotics (p < 0.05). Compared with the control group, the WBC and PLT counts in the fungal infection group decreased in the early and acute stages of infection (p < 0.01), while the CRP level increased (p < 0.01). The PLT count in the bacterial infection group decreased in the early and acute stages of infection (p < 0.01) and the CRP level increased (p < 0.05). Moreover, the decrease in the PLT count in the fungal infection group was more significant than the bacterial infection group (p < 0.01) and the CRP level increased more in the fungal infection group in the early stage of infection (p < 0.01); however, there were no significant differences in the PLT count and CRP level between the fungal and bacterial infection groups in the acute stage of infection (p > 0.05). ROC curve analysis of the WBC and PLT counts and the CRP level in the early diagnosis of fungal infections showed that the area under the curve of the PLT count was 0.912 (95% confidence interval:0.863–0.961), thus indicating a high accuracy with a cutoff PLT count of 157.0 × 109/L. The corresponding sensitivity and specificity were 77.8% and 94.6%, respectively. We conclude that the PLT count is a convenient, economical, and effective predictor of invasive fungal infections in premature infants and has potential in the early diagnosis of fungal infections.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call