Profile of febrile neutropenia in childhood cancer patients and the clinical utility of procalcitonin and C-reactive protein in identifying severe infections

INTRODUCTION

Chemotherapy-related neutropenia causing infections is a major concern in pediatric malignancies. Febrile neutropenia (FN) is a medical emergency. The mortality rate among febrile neutropenic children is reported as 0.7–3.9% in developed countries.^[1,2] In neutropenic patients, serious infections can manifest with minimal symptoms and fever can be the only sign of infection. In 1970s, Gram-negative agents were more frequently observed followed by Gram-positive organisms in 1990s and recently, there is again an increase in the frequency of Gram-negative organisms being isolated in febrile neutropenic patients.^[3-5]

Mortality and morbidity of FN can be significantly reduced by the initiation of appropriate antibiotics at the early stage and hence identifying the causative microorganisms and their antibiotic susceptibility is important. The clinical outcome of a child with FN depends on child’s underlying disease, age, duration of the neutropenia, severity of neutropenia, foci of infection, presence of indwelling catheter, duration of indwelling catheter, spectrum of antibiotics usage, and antimicrobial resistance. This study was to describe the clinical profile of FN and to assess the diagnostic utility of C-reactive protein (CRP) and procalcitonin (PCT) in identifying infection.

MATERIAL AND METHODS

Children aged <18 years of age with a malignant condition with an absolute neutrophil count of <1000/mm³ or expected to decrease to <500 cells/mm³ during next 48 h admitted between 2017 and 2019 were included in the study. Institutional ethics committee approval was obtained.

Informed consent was obtained from the parents. After relevant history and physical examination, blood samples for complete hemogram, CRP, and serum PCT were collected for everyone. X-ray chest, urine routine, urine culture, and other cultures (e.g., pus and stool) were done when clinically indicated. Blood cultures (two sets of blood culture, 5 mL each one from central line, and the other from peripheral line) were taken and were repeated after 72 h if fever persisted or while stepping up the antibiotics. Computed tomography (CT) scan of paranasal sinuses, CT thorax and abdomen, and serum galactomannan was done for fungal screening in children with prolonged fever after 72 h of appropriate antibiotics or whenever clinically indicated. Invasive fungal infections were classified using the 2002 European Organization for Research and Treatment of Cancer/National Institute of Allergy and Infectious Diseases Mycoses Study Group criteria.

For collecting oropharyngeal swab, the tongue was held down by depressor and with a sweeping motion; the posterior pharyngeal wall and tonsillar pillars were swabbed. For nasopharyngeal swabs, the head was tilted back to 70° and a swab was inserted deep into nostril, rotated, and taken out. Febrile neutropenic patients were managed as per institutional FN guidelines consistent with IDSA recommendations. For all febrile neutropenic children who were hemodynamically stable after relevant investigations first-line antibiotics Inj. Piperacillin tazobactam and amikacin was started. But in case of severe mucositis, or risk for staphylococcal infection or previous port/PICC infection, Inj Teicoplanin was started. In case of children with hemodynamic instability Inj. Meropenem and Teicoplanin was started as first-line drug. Anti-viral and antifungals were added in those who were expected to have prolonged severe neutropenia, those with severe oral ulcers and those who were febrile after 72 h of appropriate iv antibiotics.

RESULTS

The demographic details of the study population are shown in [Table 1].

Only 15 (16.6%) had identifiable foci at the time of admission. Two of them were identified to have dengue fever after laboratory confirmation. Our culture positivity rate was 20%. Among the 18 culture positive cases, 14 (77.7%) of them had central venous catheter. Among the culture positivity in children with central lines, 8 (57.2%) were Gram-negative organisms and 6 (42.9%) were Gram-positive organisms.

Among the total 18 culture positive cases, 11 were gram negative organisms (61%) and 7 were Gram-positive organisms (38.9%). Acinetobacter was the commonest Gram-negative bacilli (4; 22.2%), followed by Klebsiella (3; 16.7 %), Escherichia coli (3; 16.7%), and Pseudomonas (1; 5.6%). Among Gram-positive organisms, Staphylococcus (4; 22.2% Staph aureus-3, CONS-1) and Streptococcus (3; 16.7%; Streptococcus pneumoniae-1, Streptococcus mitis-1, Group a Streptococci-1) were commonly reported.

Piperacillin tazobactam had the highest in vitro efficacy against E. coli, Carbapenem resistance was most prevalent among Klebsiella and only 56% of Klebsiella and 31% of Pseudomonas was sensitive to aminoglycosides. Piperacillin tazobactam and Amikacin were administered as first line antibiotics in 84 (93.3%) patients. Meropenem and Teicoplanin were administered as first-line antibiotics in 6 (6.7%) who presented in shock. Acinetobacter was sensitive to amikacin, piperacillin tazobactam, ciprofloxacin, levofloxacin, cefaperazone with sulbactum and resistant to ampicillin and cephalexin.

Out of 84 children who received piperacillin-tazobactam and amikacin, 54 (64.2%) children required meropenem and teicoplanin as second-line antibiotics (did not respond to first-line antibiotics/or based on culture reports). Among 90 children, 37 (41.1%) required antifungals (Voriconazole/ Amphotericin) and 36 (40%) children required anti-viral (Acyclovir/oseltamivir).

Details of the children who presented with shock are shown in [Table 2].

A total of 85 (94%) had uneventful course and overall mortality rate was 5.5%. [Table 3] shows the details of the children who died.

Among 18 culture positive cases, CRP was positive in 10 cases and PCT was positive in 12 cases. Sensitivity, specificity, positive predictive value, and negative predictive value of CRP was 55.5%, 50%, 21.7%, 81.8%, and of PCT was 66.6%, 55.5%, 27.2%, and 86.9%, respectively. [Tables 4 and 5] depict the correlation between CRP and PCT and culture positivity.

DISCUSSION

In neutropenic patients, inflammatory response is suppressed and serious infections can manifest with minimal symptoms and the origin of fever is not being identified in more than 50% of the cases despite advanced laboratory techniques.

A total of 90 episodes of FN were observed during the study. The hematological malignancies were associated with more FN episodes than the solid tumors (74% vs. 10%). Among 90 FN episodes, 18 (20%) microbiologically documented infections were noted. No foci of infection could be identified in 73 (81.1%) at the time of admission. This correlates with study by Kapoor et al. who has reported that 78% of his study population had no foci of infection.^[6] The most common focus of infection among our study population was lower respiratory tract infection 6 (40%) which is similar to study by Kar et al. who has reported mucositis (33.4%) and pneumonia (24.7%) as the common clinical foci.^[7] Gurlinka et al. have reported 14.4 % culture positivity and only case of identifiable foci during 69 episodes of FN.^[8]

Blood culture positivity rates were 20% in our study and urine culture was positive in 2 cases. Ghosh et al. have reported 30% culture positivity in his study about high risk FN patients.^[9] Das et al. have reported 19.8% culture positivity in his study.^[10] In our study, 17 (94.4 %) had isolated bacterial infection and 1 (5.55%) had polymicrobial sepsis and one had invasive fungal infection.

The mortality rate of our study was 5.5%. Among the 5 children who died, 2 were relapsed ALL and both of them presented in shock, one acute myeloid leukemia had invasive candida infection, three of them had no documented blood culture positivity though CRP and PCT was positive in all the five cases.

Kar et al. have reported 13.2% mortality rate in his study.^[7] Das et al. have reported a mortality rate of 10%.^[10] Paganini et al. concluded that presence of advanced stage of malignancy, severe associated comorbidities and bacteremia at presentation can be predictors of mortality in children with FN.^[11]

Our study proved that in developing country, Gram-negative bacteremia is still more common. The shift toward greater incidence of Gram-negative infections is attributed to increased use of indwelling intravascular catheters and fluoroquinolone and trimethoprim sulfamethoxazole prophylaxis.

We had 61% Gram-negative isolates and the most common Gram-negative isolates of our study were Acinetobacter (4; 22.2%), followed by Klebsiella (3; 16.7%), E. coli (3; 16.7%), and pseudomonas (1; 5.6%). Reddy et al. had prevalence as high as 92% for Gram-negative isolates with Klebsiella (48%) being the most common organism.^[12] Ghosh et al. had a prevalence of Gram-negative isolates at 56% with pseudomonas (12%) being most common.^[9] Lakshmaiah et al. had a 64% prevalence of Gram-negative isolates with E. coli (32%) being the most common.^[13]

In our study, carbapenems, cephalosporins, cefoperazone sulbactam, fosfomycin, and piperacillin tazobactam showed the highest in vitro activity against E. coli, whereas in cases with Klebsiella pneumoniae, carbapenem, and fosfomycin were effective compared to cephalosporins and aminoglycosides. All pseudomonas species were sensitive to beta lactam/beta lactamase inhibitor which decreased to 66% in Acinetobacter and 50% in E. coli and 31% in Klebsiella. There was high degree of resistance to aminoglycosides with only 56% of Klebsiella and 31% of Pseudomonas sensitive to aminoglycosides. Among Gram-positive organisms, 85.7% were sensitive to Cefotaxime, and 71.4 % were sensitive to ampicillin. Among our 18 culture positive cases, 17 (94.4%) were sensitive to first-line antibiotics (piperacillintazobactam and amikacin) and only 1 (5.6%) was resistant to piperacillin-tazobactam.

In a study by Lakshmaiah et al., the antibiotic sensitivity among GNB was highest for imipenem (100%), followed by piperacillin-tazobactam (86.95%), most of the E. coli isolates were resistant to cefoperazone-sulbactam, piperacillintazobactam. Gram-positive isolates including MRSA were uniformly sensitive to levofloxacin, teicoplanin, linezolid and vancomycin.^[14] Similar antibiotic sensitivity pattern of high sensitivity of GNB to cefoperazone/sulbactam, carbapenems, piperacillin/tazobactam, and resistance to third generation cephalosporins has been reported in other studies.^[15,16]

PCT has been proposed as a biomarker for discriminating bacterial and viral infections due to the significantly higher increase in its levels in the presence of bacterial stimulus. PCT levels have been shown to be significantly higher in FN especially in Gram-negative infections, but not in to viral illness or elevated in inflammatory conditions or mucositis.^[17] Elevated PCT values at presentation correlated with increased risk of severe infections as per Hemming et al. and but PCT values did not vary between sepsis or less severe infections at admission but was significant on day 2 as per Santolaya et al.^[18,19] CRP, an acute phase reactant synthesized during infection and acute inflammation, too has been widely used as in indicator of infection. CRP has been documented to be higher in clinically or microbiologically documented infections when compared to fever of unknown origin. CRP was reported to be more sensitive but not more specific than PCT. As PCT levels are reported to rise within 3–4 h in response to infection as compared to 24–48 h required for CRP, initial levels of CRP at admission might not reflect the true picture and hence needs serial monitoring.

In our study among 18 culture positive cases, CRP was positive in 10 cases and 12 cases had PCT positivity. Sensitivity, specificity of CRP and PCT was 55.5%, 50%, and 66.6%, 55.5%, respectively. Although CRP and PCT were not very useful in prediction of culture positivity, both CRP and PCT were positive in patients with hemodynamic instability and in those who succumbed to death. In our study, the positive predictive value, negative predictive value of CRP was 21.7% and 81.8% and of PCT was 27.2% and 86.9%, respectively.

In a study by Reitman et al., serum PCT was higher in positive blood cultures than with sterile blood cultures. In ruling out bacteremia, PCT showed 76% specificity and 93% negative predictive value.^[20] In a study by Kim et al., PCT was found to have better diagnostic value than CRP to detect bacteremia in FN.^[21]

Asturias et al. had reported a direct association between elevated CRP, the duration of FN, bacteremia, and mortality.^[22] Avabratha et al. have quoted that antibiotic response in pediatric FN can be assessed by serial monitoring of CRP.^[23] Oberoi et al. have proved that CRP is a strong predictor of complications in pediatric FN.^[24]

CONCLUSION

Gram-negative organisms are still the predominant causative agents reported during FN, as noticed in other studies from similar background. Raised CRP and PCT are indicators of severe adverse events during FN.