This is a prospective study of 50 consecutive episodes of febrile neutropenia (FN) in children suffering from malignancies and undergoing therapy in a pediatric oncology unit of a teaching hospital. Children between 1-15 years of age with marrow/biopsy evidence of malignancy and clinical features of FN were included. Those with liver disease were excluded from the study.

The results of CRP were compared and associated with clinical findings and other laboratory results. SPSS version 11.5 was used for statistical analysis.

Fifty episodes of FN occurred in 33 children, 17 of whom were males, with a mean age of 6.9 years. Thirty episodes of FN occurred in the intensive and 20 in the maintenance phases of chemotherapy. The commonest diagnosis was acute lymphoblastic leu-kemia (34 episodes in 24 patients).

Out of the 50 episodes of FN, 19 (38%) occurred in children with documented infection with clinical and/or radiological evidence of infection and culture positivity (Group I). In these, respiratory tract infections were the commonest (31.6%) followed by urinary tract infections (26.3%) and sepsis (21%). Cellulitis purely on clinical evidence was also included in this group. Nine episodes (18%) occurred in children with probable infection (Group II), in which there was an identifiable site of infection without a positive culture. The rest 22 (44%), were labeled as having fever of unknown origin (Group III).

ANC at presentation was <200/cmm in 19 epi-sodes (38%), between 200-500/cmm in 17 episodes (34%) and >500/cmm in 14 episodes (28%). A majority of the documented and probable infection, occurred with ANC <200/cmm (73.7% and 55.6% respectively) while only a few episodes of documented and probable infection, occurred in children with ANC >500/cmm (10.5% and 22.2% respectively). Fifteen episodes grew organisms from tissue fluid cultures (S. aureus: 4; P. aeuruginosa: 3; Citrobacter: 1; E.coli: 4; S. pneumoniae: 2; and other: 1). CRP was positive in 11 of these episodes (Table I).

TABLE I



Culture Results in Febrile Neutropenia and CRP Positivity

Initial CRP positivity in the study groups revealed that while Group I (D1:15/19; D7:4/19) and II (D1:7/9; D7:2/9) had higher number of CRP positivity, Group I values, when considered in isolation, were significantly higher (P<0.001). In Group III, CRP was positive on D1 in 5/22 episodes and in 2/22 episodes on day 7. When CRP positivity in the various groups were analyzed on day 7 and at the end of antibiotic therapy (if later), it was found that in all three groups, CRP decreased as treatment progressed and the child responded to therapy.

Our findings corroborate the results of previous studies(3,4). However, Bodey, et al.(5) demonstrated 53% documented infections in their series, in contrast to the present study. This could be due to low culture yields, early initiation of empirical antibiotic therapy or better nursing care in our patients.

Of the 50 episodes of FN, true positives (i.e. CRP positive with documented infections) accounted for 15 episodes, false positive (i.e. CRP positive but cultures negative ) in 12, false negative (i.e. CRP negative and culture positive) in 4 and true negative (i.e. CRP and culture negative) in remaining 19 episodes. The sensitivity was 55% and specificity 82% with a positive predictive value of 78%. Putto, et al.(6) have reported a sensitivity of 100% and specificity of 75% for CRP in infections while Peltola and Jaakkola(7) reported a sensitivity of 89% and specificity of 77% with a positive predictive value of 79%.

In Group I, CRP remained high in 4 children out of 15 on day 7 and on the last day of antibiotic therapy. Of these 2 died and 2 required augmentation of therapy. Other than these, in all 3 groups, CRP positivity uniformly decreased as the children improved. Thus, CRP can indicate persistent infection in children with FN. This correlates well with other studies(8,9). Though there are reports of procalcitonin(PCT) being a better marker of infection in FN, it is expensive and not readily available(10). In a pediatric study(11) involving 60 febrile episodes, periodic measurement of PCT was found to be more useful than CRP. However, both PCT and CRP levels were significantly higher in FN than in controls.

We conclude that, CRP is an easily available and significantly sensitive test to diagnose infection and monitor response to therapy in FN.

Contributors: ATK conceived the idea, provided the study design and stands as guarantor. Data was collected and analyzed by VP along with ATK and KSA. KSA, ATK and AR wrote the paper. The final manuscript was approved by all authors.

Funding: None.

Competing interests: None stated.