A Meta-Analysis of the Diagnostic Performance of Procalcitonin in the Diagnosis of Serious Bacterial Infection in Pediatric Febrile Neutropenia

Background:

Continued advancements and intensification of chemotherapy regimens have enhanced outcomes for children with cancer. With more aggressive therapy, however, infection rates due to immune system suppression have increased1. Although progress in supportive care strategies has improved management of such infections2, infection remains a major cause of morbidity and mortality among pediatric oncology patients3. Serious bacterial infections often present non-specifically with only fever and neutropenia; however, infection represents only one etiology of fever and neutropenia.

Objective:

Identification of a biomarker to more quickly diagnose serious infection in fever and neutropenia episodes would be advantageous. Prior small studies suggest that procalcitonin may be a feasible and reliable marker for bacterial infection in febrile neutropenic pediatric patients. Because the implementation of such a biomarker would be significant to pediatric oncology practice, stronger evidence is required before the assay can be meaningfully, reliably, and safely applied to patient care.

Design/Methods:

We conducted a systematic review and meta-analysis of the literature regarding the use of procalcitonin as a marker of serious infection in neutropenic febrile children. We searched several electronic databases that included records from 1966 to July 2011 using the keyword ¿¿¿¿¿¿¿procalcitonin¿¿¿¿¿¿¿ alone and in combination with several terms related to our topic.

Results:

The ten selected studies included 591 pediatric oncology patients with a total of 1161 febrile episodes. Regression analysis and the random effects model were used to estimate the common effect size for each measure of diagnostic performance for procalcitonin: pooled odds ratio was 39.29, with 95% confidence interval (12.61, 122.4); sensitivity was 0.709, or 70.9%, with p<0.0001 and 95% confidence interval (0.357, 1.06); specificity was 0.826, or 82.6%, with p<0.0001 and 95% confidence interval (0.629, 1.02).

Conclusions:

Procalcitonin is an appropriate candidate biomarker for serious bacterial infection in pediatric oncology fever and neutropenia patients, given that the odds of having an elevated procalcitonin assay and a serious bacterial infection are 39.29 times higher than the odds of having an elevated procalcitonin assay in the absence of serious bacterial infection. Furthermore, this meta-analysis produced evidence of modest to good sensitivity and better specificity with values of 71% and 83%, respectively.

References:

1. Finberg R. and Talcott J., New England Journal of Medicine, 1999. 2. Rubnitz J., Lensing S., Zhou Y., et al, Cancer, 2004. 3. Slats A., Egeler R., Van Der Does-van den Berg A., et al, Leukemia, 2005.