Role of neutrophil elastase in predicting infection among children with chemotherapy-induced febrile neutropenia
Article information
Abstract
Background
Infection is a significant cause of death following chemotherapy-induced febrile neutropenia (FN). Neutropenia and compromised neutrophil function are the primary reasons for the decreased defense against infections.
Purpose
This study aimed to evaluate the significance of neutrophil elastase (NE) in predicting the outcomes of childhood hematological malignancies with FN.
Methods
The study included 64 patients with FN and 64 healthy children matched for age and sex as controls. The patients were selected from the Hematology and Oncology Unit, Menoufia University, Pediatric Department, and Tanta Oncology Institute. Complete blood counts, C-reactive protein (CRP) levels, NE levels, and blood cultures for bacteria and fungi were performed.
Results
The levels of CRP and NE were increased among children with FN; of them, 21.9% had Gram-negative bacteremia, 17.2% had Gram-positive bacteremia, and 3.1% had Candidemia. NE level was increased in patients with bacterial infections, with a significant positive correlation with duration of FN. NE level had a cutoff of 6.5, with an area under the curve of 0.899, sensitivity of 83.33%, and specificity of 87.50% signifying a higher risk of mortality compared to other variables.
Conclusion
NE levels were elevated in children with FN, suggesting its usefulness for the early detection of infection that could decrease infection-related morbidity and mortality.
Key message
Question: Can neutrophil elastase (NE) levels predict infection— the primary cause of mortality—among children with hematological malignancies and febrile neutropenia (FN)?
Finding: Elevated levels of NE were found in children with chemotherapy-induced FN and a bacterial infection.
Meaning: Increased NE levels and prolonged FN are important factors associated with mortality risk.
Graphical abstract
Introduction
Patients receiving intensive chemotherapy for hematological malignancy face a significant difficulty with febrile neutropenia (FN) [1,2]. Even with antibiotics given during chemotherapy-induced neutropenia and prompt use of broad antimicrobial drugs when FN begins, approximately 35% of individuals with bacterial sepsis do not survive, increasing to 85% for severe sepsis with multiple organ failure [3]. Between 10%–25% of patients experience invasive fungal infections, with mortality rates reaching 30% for Aspergillus infections and up to 96% for cerebral fungal disease [4,5]. If antimicrobial strategies, such as granulocyte transfusions, were started sooner, it could lead to a reduction in mortality rates [6], Scoring systems like the MASCC (multinational association for aupportive care in cancer) or MEWS (modified early warning score) are used to clinically identify high-risk patients [7].
Neutrophil elastase (NE) is a proteolytic enzyme that belongs to the chymotrypsin-like family of serine-proteinases. NE is a protein that consists of 218 amino acids and is found in cytoplasmic azurophilic granules, which have the potential to be discharged during neutrophil degranulation. NE is an inflammatory agent that decreases the rate at which cilia beat, enhances the production of mucus, and causes damage to the airway epithelium [8].
NE breaks down various substances such as elastin, collagen, and fibronectin. NE hydrolysis makes up around 80% of all protease hydrolysis activity in the human body, leading to inflammation, advancing bacterial infections, and excessive mucus secretion. The disease state and severity are reflected by the levels and activity of NE [9]. Despite its connection to multiple illnesses, NE plays a role in cancer [10] and inflammation resistance [11].
Our goal was to assess the importance of NE as a biomarker for detecting outcomes in pediatric hematological cancers, specifically looking at its capability to predict infections in children with chemotherapy-induced FN.
Methods
1. Study design
The study was conducted on 64 FN patients who were selected from the Hematology and Oncology Unit, Menoufia University Hospital, Pediatric Department and Tanta Oncology Institute and 64 healthy children who were matched in terms of age and sex and had no previous medical issues as controls.
2. Ethical approval
The Institutional Review Board of the Menoufia Faculty of Medicine approved the study. Research work was performed in accordance with the Declaration of Helsinki (approval ID: 3/2023 PEDI 17). All patients and their guardians were informed of the benefits and drawbacks of the study before providing their consent.
3. Inclusion criteria
Patients with either one oral temperature exceeding 38.3°C or 2 temperatures more than 38.0°C within 12 hours, each lasting at least 1 hour, and an absolute neutrophil count (ANC) below 1,000/mm3.
Healthy children with matched age and sex who agree to participate with an ANC within normal range.
4. Data gathering
Every patient received a thorough physical and clinical assessment. Tests conducted in the laboratory included a complete blood count, C-reactive protein (CRP), NE, and blood cultures for bacteria and fungi upon admission on the first day of FN. A chest computed tomography (CT) scan was carried out on the fifth day of FN.
5. Study methods
The Sysmex XT 1800, an automated cell counter from Germany, was utilized to analyze complete blood counts. CRP levels were measured. A blood culture was carried out utilizing the BACT/ALERT automated microbial detection system from Biomerieux in France.
The enzyme-linked immunosorbent assay (ELISA) technique measured levels of serum human NE in both patients and controls. A double antibody ELISA test was conducted using commercial kits from Sunred Biological Technology Co., Ltd. in Shanghai, China (Catalogue No. 201-12-0891). Serum samples were collected, spun at 3,000 rpm for 10 minutes, then divided and stored at -20°C for later NE examination.
Each standard well was given an extra 50 μL. The test wells were given an additional 40 μL of the sample. After adding 50 μL of prepared streptavidin-horseradish peroxidase solution to every standard and test well, 10 μL of NE-antibody was added to each well of the microtiter ELISA plate. Afterward, the plates were transferred to an incubator set at 37°C. Subsequently, the plates were subjected to 5 washes using wash buffer. 50 μL of chromogen solution B and chromogen solution A were added to each well, followed by incubation at 37°C for 10 minutes in the dark, resulting in the liquid turning blue. Then, each well had 50 μL of stop solution added. The liquid changed from blue to yellow right away. The absorption level of each colored solution was evaluated at a wavelength of 450 nm using a microplate reader. The linear regression equation for the standard curve was calculated based on the concentrations of the standards. The OD values obtained were used in the regression equation to find out the sample concentration.
6. Statistical analysis
The data gathered were organized and analyzed using IBM SPSS Statistics ver. 23.0 (IBM Co., USA). Quantitative data were represented using mean and standard deviation, while qualitative data were presented as numbers and percentages. Chi-square test was utilized to examine the relationship between 2 quantitative variables. The Fisher exact test was utilized to investigate the relationship between the 2 qualitative variables if any of the expected cells is less than five. The Student t test was utilized to examine the relationship between the 2 normally distributed quantitative variables. Mann-Whitney U test was used to examine the relationship between the 2 not normally distributed quantitative variables. Spearman correlation (r) was used to study the correlation between the 2 not normally distributed quantitative variables. Receiver operator characteristics (ROCs) were created to evaluate biomarker performance by identifying points where sensitivity and specificity are maximized for accuracy. A P value less than 0.05 is considered statistically significant.
Results
Based on demographic and laboratory investigations of the studied groups, there were significant difference between patients and controls regarding hemoglobin concentration, total leukocyte count, ANC, platelet count, CRP, and NE as presented in Table 1.
Demographic characteristics and laboratory investigations by study group
Based on the type of malignancy in neutropenic patients, acute lymphoblastic leukemia pre-B cells were the most common. Categorization of FN into moderate (ANC, 500–1,000/mm3, severe (ANC<500 mm3) and profound (ANC<100 mm3) based on ANC as demonstrated in Table 2.
Type of malignancy and degree of neutropenia among patients with febrile neutropenia (N=64)
Regarding septic focus, 41 patients (64.1%) had no septic focus, 21 patients (32.8%) had pneumonia, 4 patients (6.3%) had oral mucositis, 29 patients (45.3%) got gastroenteritis, and 5 patients (7.8%) had typhlitis. In terms of CT chest findings, 39 patients (61%) had normal CT chest findings, 15 cases (23.4%) showed pneumonia, 5 patients exhibited a ground glass appearance, and 5 had pleural effusion (7.8% each). Klebsiella pneumoniae was the most common bacterial infection (12.5%) following by Staphylococcus aureus (9.4%) and Streptococcus pneumoniae (7.8%). Candidemia account 3.1% of cases according to Table 3.
Septic focus, blood culture results, and chest computed tomography (CT) findings in patients with febrile neutropenia
Based on the blood culture findings, 37 patients had negative results (57.8%). K. pneumoniae was detected in 8 patients, 2 survivors (3.8%) and 6 nonsurvivors (54.5%) with a statistically significant difference (P<0.001). Pseudomonas aeruginosa was found in 2 nonsurvivors (18.2%), showing a statistically significant difference (P=0.027). There were 2 nonsurvivors (18.2%) with Candida, showing a statistically significant difference (P=0.273). Gram-positive bacteremia (17.2%) was observed in 9 survivors (17%) and 2 nonsurvivors (18.2%) and Gram-negative bacteremia (21.9%) was observed in 6 survivors (11.3%) and 8 nonsurvivors (72.7%) with a P value of 0.003 as indicated in Table 4.
Blood culture results and bacteremia type survivors versus nonsurvivors
Table 5 displayed a notable increase in NE and CRP levels among patients with bacterial infection compared to those without it, with statistical significant difference at (P<0.001 and P=0.001, respectively).
CRP and NE levels of cases with versus without bacterial infections
Fig. 1 illustrates a significant positive association between NE and the length of FN (r=0.6, P<0.001). Fig. 2 illustrates the analysis of the ROC curve of NE and CRP to predict infections in patients with FN revealed a cutoff point of 6.5, with an area under the curve (AUC) of 0.899, a sensitivity of 83.33%, and a specificity of 87.50%. CRP revealed a cutoff point of 6.1, with an AUC of 0.9, a sensitivity of 81.3%, and a specificity of 99.5%.
Correlation between neutrophil elastase level and duration of febrile neutropenia.
Receiver operating characteristic (ROC) curve analysis of ability of neutrophil elastase and C-reactive protein (CRP) levels to predict bacterial infections among patients with febrile neutropenia.
Factors that increase the risk of mortality in cases with FN included NE (P=0.002), duration of FN (P=0.003), bacterial growth (P=0.01) and CRP (P=0.025) as shown in Table 6.
Risk factors associated with mortality in patients with febrile neutropenia
Discussion
Infections pose a significant risk of morbidity and mortality for children with cancer who are undergoing chemotherapy or receiving a transplant. Neutropenia is a serious risk of cancer therapy and increases the chance of bacterial infections which can be life-threatening [12].
The current study demonstrated higher levels of CRP and NE in children with cancer who developed FN due to chemotherapy. During neutropenic episodes, the acute myeloid leukemia patients experienced increased levels of CRP and NE as reported by Van de Geer et al. [13].
The most frequent bacterial infection was K. pneumoniae (12.5%), with S. aureus (9.4%) and S. pneumoniae (7.8%). Gram-negative bacteremia occurred in 21.9% of cases, while Gram-positive bacteremia occurred in 17.2%. Candidemia occurred in 3.1% of cases. Antari et al. [14] reported that 19% of cases of FN were caused by Gram-negative bacteria, with Gram-positive bacteria and Candidemia each accounting for 3% of cases. Gram-negative bacteria that were found were P. aeruginosa, E. coli, and Klebsiella. Van de Geer et al. [13] discovered that bacterial infection occurred in 34.6% of neutropenic episodes, while fungal infection occurred in 7.7%.
Levels of NE were elevated in patients with bacterial infection with a highly significant positive correlation with the duration of FN. NE had a cutoff of 6.5, with an AUC of 0.899, a sensitivity of 83.33%, and a specificity of 87.50% indicating higher mortality risk compared to other variables. CRP revealed a cutoff point of 6.1, with an AUC of 0.9, a sensitivity of 81.3%, and a specificity of 99.5%. NE is more sensitive than CRP indicating that it is increased in true positive cases.
An increase in NE levels was linked to neutrophil regeneration prior to the detection of circulating neutrophils. Comparing the hazard ratio for NE levels above versus below the median concentration, the value was 6.3 (95% confidence interval, 2.6–15.2; P<0.001). This discovery is beneficial for patients with refractory infections and prolonged neutropenia. When deciding on whether intensifying treatment is necessary during prolonged neutropenia, including the use of granulocyte transfusions, it is reassuring to have the ability to anticipate autologous neutrophil recovery for fighting severe infections [6].
CRP seems to provide less useful information. This could be because every neutropenic patient shows increased CRP levels, which may be caused by the inflammatory impact of their primary illness, previous chemotherapy, and/or therapy-related mucosal injuries [15-18].
Mean CRP peak level was reached at 36–48 hours in children with cancer after the start of infection. CRP showed a slower kinetics, this slow kinetic may lead to false negative results [19].
The measurement of human NE-like activity was conducted in cell lysates from isolated peripheral polymorphonuclear neutrophils, demonstrating repeatable and reliable results for up to 26 hours. Consequently, its highest level could persist for as long as 26 hours [20]. So, the NE kinetics may be faster and favorable than that of CRP.
Few studies have examined the kinetics of CRP, procalcitonin (PCT), and interleukin-6 (IL-6) in pediatric patients with fever and neutropenia. De Lucio Delgado et al. [19] illustrated that IL-6 demonstrated the fastest kinetics, then PCT. The highest level of IL-6 was achieved in the initial blood sample after the onset of the fever episode. PCT and IL-6 proved to be more reliable indicators than CRP for predicting bacteremia in individuals with FN [21].
De Lucio Delgado et al. [19] illustrated the ability of plasma interleukin-8 (IL-8) levels measured 24–48 hours prior to diagnosis to distinguish between patients with FN due to bloodstream infections and those with FN resulting from other causes. Many studies have shown that IL-8 concentrations are elevated in patients with bacterial infections during neutropenia, whereas CRP appears to be less effective [15,22].
Also, patients suffering from bacterial pneumonia show elevated NE levels in bronchoalveolar lavage fluid, leading to excessive proteolytic damage and worse clinical outcomes [23,24].
The limitation of this study was small number of patients. The strength of this study was uniquely evaluated NE as a promising biomarker to identify bacterial exacerbation in children with chemotherapy-induced FN and it is the first study investigating NE in many types of childhood malignancy with chemotherapy-induced FN and its role in prediction of FN.
In conclusion, NE levels were elevated in children with FN promising for early detection of infection to decrease infection-related morbidity and mortality.
Notes
Conflicts of interest
No potential conflict of interest relevant to this article was reported.
Funding
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Acknowledgments
The authors are thankful to both the individuals involved in the study and the team who collected the data.
Author Contribution
Conceptualization: ME, DE, MA, EA, MK, AM; Formal analysis, ME, DE, MA, EA, MK, AM; Funding acquisition, ME, DE, MA, EA, MK, AM; Investigation: ME, DE, MA, EA, MK, AM; Methodology: ME, DE, MA, EA, MK, AM; Writing, review, and editing: ME.