Background: Accurate and timely diagnosis of community-acquired bacterial versus viral infections in children with systemic inflammatory response syndrome (SIRS) remains challenging both for clinician and laboratory. In the quest of new biochemical markers to distinguish bacterial from viral infection, we have explored the possible role of the soluble secreted form of ST2 (sST2).
Methods: This explorative prospective cohort study included children with SIRS who were suspected of having community-acquired infections. Plasma samples for sST2 measurement were obtained from 64 hospitalized children, 41 of whom had SIRS of bacterial etiology and 23 SIRS of viral etiology, and from 20 healthy, age- and sex-matched control children. sST2 measurement was carried out by enzyme-linked immunosorbent assay in parallel with standard measurements of procalcitonin (PCT) and C reactive protein (CRP).
Results: Our findings demonstrate that children with SIRS associated with bacterial infection present significantly increased levels of sST2, when compared with patients with SIRS of viral etiology and healthy children. More important, receiver operating characteristic curve analysis indicated that sST2 has a significant diagnostic performance with respect to early identification of SIRS of bacterial etiology, which was similar to that of PCT and greater than that of CRP. Finally, the combination of sST2 plus PCT and/or CRP, and PCT plus CRP increased their sensitivity and negative predictive value compared with sST2, PCT and CRP alone.
Conclusions: In conclusion, sST2 level may prove useful in predicting bacterial etiology in children with SIRS.
From the *Immunological and Infectious Disease Unit, University Department of Pediatrics, Bambino Gesù Children’s Hospital; †General Pediatrics and Infectious Disease Unit, Department of Pediatrics, Bambino Gesù Children’s Hospital; ‡Department of Pediatric Hematology and Oncology, Bambino Gesù Children’s Hospital; §Department of Internal Medicine, University of Tor Vergata; and ¶Clinical Trial Unit, University Department of Pediatrics, Bambino Gesù Children’s Hospital, Rome, Italy.
Accepted for publication July 24, 2013.
A.F. and P.D. shared seniorship in this study.
The research leading to these results has received funding from the European Union Seventh Framework Programme FP7/2007-2013 under grant agreement no. 261060. The authors have no other funding or conflicts of interest to declare.
Address for correspondence: Francesca Ippolita Calò Carducci, MD, PhD, University Department of Pediatrics, Bambino Gesù Children’s Hospital, Piazza Sant’Onofrio 4, 00165 Rome, Italy. E-mail: email@example.com.
Early diagnosis of bacterial versus viral infection in children with systemic inflammatory response syndrome (SIRS) remains a hard challenge.1 Most children will benefit from antibiotic treatment, and some substantially do, because untreated bacterial infections may cause serious complications. On the contrary, treating viral illnesses with antibiotics is not only ineffective, but also contributes to the development of resistance, adds the risks of toxicity and allergic reactions, affects the quality of care and inflates health care costs.2–4 Results from cultures, antibody titers and tests for viral antigens are often delayed, and rapid immunological or genomic tests require prior suspicion of the infectious agent. Also, blood analysis data (blood count and differential leukocyte count) poorly discriminate children with bacterial infection from those with viral infection. Procalcitonin (PCT) and C reactive protein (CRP) levels are frequently used to aid in the diagnosis of SIRS.5,6 CRP is used to differentiate between viral and bacterial infections. However, CRP is neither highly specific nor sensitive for bacterial infection since it can remain at low concentrations in bacterial infections and can increase significantly in viral infections.7,8 Because of its shorter half life and the fact that elevated concentrations are achieved earlier than with CRP, PCT can offer advantages compared with CRP in the differential diagnosis of SIRS in children. However, the clinical utility of PCT remains surrounded by considerable controversy.9–11 To date, there has been no single biomarker that offers clinicians caring for sick children the absolute diagnostic ability to distinguish bacterial from viral infection. Thus, a great deal of effort is focused on looking for new biomarkers to be used alone or in combination with other biomarkers as part of a panel of tests to assist clinicians in diagnosis.
The ST2 protein is a member of the interleukin-1 receptor family. ST2 is a receptor that is present in 2 main forms, in a membrane-anchored form12 and in a soluble secreted form.13 The membrane-anchored form (ST2) is preferentially expressed by activated Th2 cells and plays an essential role in Th2 effector functions.14,15 The soluble form of ST2 (sST2) is mainly secreted by fibroblasts16 and has been suggested to act as a decoy receptor by binding interleukin-33, thereby inhibiting signaling by ST2.17,18 Expression of sST2 protein can be induced in vitro by proinflammatory stimuli, like lipopolysaccaride, interleukin IL-1β, tumor necrosis factor-α and IL-6 in human and murine inflammatory models.19,20 Elevated levels of sST2 have been found in patients with autoimmune diseases,21 heart disease22 and sepsis.23,24 To our knowledge, plasma levels of sST2 have not been documented in pediatric patients in previous reports. Considering that sST2 contributes to the regulation of the immune response during inflammation, we hypothesized that plasma levels of sST2 would be elevated in children with SIRS and would possibly be helpful in identifying patients suffering from bacterial infection. Therefore, we here performed an explorative study to determine sST2 plasma concentrations in children with SIRS and its diagnostic accuracy in the prediction of SIRS of bacterial etiology.
MATERIALS AND METHODS
This explorative, prospective, observational study was conducted at the Bambino Gesù Children Hospital, between July 2011 and September 2012. The study included all hospitalized children admitted in the Unit of General Pediatrics and Infectious Disease and in the Unit of Immunological and Infectious Disease <12 years who met inclusion criteria. For children above 44 weeks postmenstrual age, we used the criteria of SIRS as defined by Goldstein et al25—the presence of at least 2 of the following 4 criteria, 1 of which must be abnormal temperature or leukocyte count: (1) abnormal temperature defined as a single core temperature >38.5°C or below 36°C; (2) tachycardia defined as mean heart rate > 2 standard deviation (SD) above normal for age in the absence of external stimulus chronic drugs and painful stimuli or otherwise unexplained persistent elevation over a 0.5- to 4-hour period or for children <1 year of age, bradycardia defined as mean heart rate less than tenth percentile in the absence of external vagal stimulus, beta-blockers or congenital heart disease or otherwise unexplained persistent depression over a 0.5-hour period; (3) mean respiratory rate over 2 SD above normal for age or mechanical ventilation for an acute process not related to underlying neuromuscular disease or the receipt of general anesthesia; (4) leukocyte count elevated or depressed for age (not secondary to chemotherapy-induced leucopenia) or >10% immature neutrophils. For children <44 weeks postmenstrual age, we used a list of criteria defined by an Expert Meeting on Neonatal and Pediatric Sepsis on behalf of the European Medical Agency.26 The original article suggest that for enrolling in clinical trial on neonatal sepsis 2 clinical plus 2 laboratory criteria should be present, but for the purpose of this study, we decided to be more sensitive and enrol patient who present at least 1 clinical criteria among (1) hyper- or hypothermia or temperature instability; (2) reduced urinary output or hypotension or mottled skin or impaired peripheral perfusion; (3) apnea or increased oxygen or increased ventilatory support requirement; (4) bradycardia spells or tachycardia or rhythm instability; (5) feeding intolerance or abdominal distension; (6) lethargy or hypotonia or irritability and (7) skin and subcutaneous lesions such as petechial rash or sclerema and at least 1 laboratory criteria among (1) white blood cell count < 4 or > 20 × 109 cells/L; (2) immature to total neutrophil (I/T) ratio > 0.2; (3) platelet count < 100 × 109/L; (4) CRP > 15 mg/L or PCT ≥ 2 ng/mL; (5) glucose intolerance when receiving normal glucose amounts (8–15 g/kg/d) as expressed by blood glucose values > 180 mg/dL or hypoglycemia (< 40 mg/dL) confirmed on at least 2 occasions and (6) acidosis with base excess (BE) < −10 mmol/L or lactate above 2 mmol/L.
The following were considered as exclusion criteria: (1) antibiotic use within the 48 hours before admission to the hospital; (2) vaccination during the previous 2 days which may have caused the febrile syndrome; (3) known immunodeficiencies; (4) any chronic pathology and (5) surgery performed in the 7 days before inclusion in the study. Decisions on diagnostic evaluation were made by the attendant physicians. PCT and CRP were determined at study entry in all patients. Blood samples obtained from 20 age- and sex-matched children attending outpatient clinics for minor surgical procedures were used as controls.
The study protocol was approved by the local Ethic Committee. Before enrolment in the study, informed consent was obtained from the parents or legal guardians.
On the basis of their final diagnosis, children were classified into 2 groups corresponding to: bacterial SIRS (B-SIRS) and nonbacterial (non-B-SIRS). The following were considered as B-SIRS: (1) bacteremia recovery of a single bacterial pathogen using standard culture techniques; (2) documented urinary tract infection urine-reactive strip positive, confirmed by urinary sediment (a positive-reactive strip was defined as the presence of leukocytes and/or nitrites and positive urinary sediment when leukocytes or microorganisms on the Gram stain were detected) and growth of a single urinary tract pathogen at ≥105 colony-forming units/mL in urine samples; (3) probable urinary tract infection urine-reactive strip positive, confirmed by urinary sediment but no growth of urinary tract pathogens or growth of urinary tract pathogens at ≤105 colony-forming units/mL; (4) bacterial meningitis-positive cerebrospinal fluid culture; (5) enteritis isolation of a microorganism in a feces sample. The non-B-SIRS group was composed of children with SIRS of proved viral infection (positive antigen or polymerase chain reaction detection, without evidence of bacterial superinfection and with negative bacterial cultures) or probable viral infection (negative cultures and no signs for focal bacterial infection).
Obtaining and Processing the Samples
Samples of 1 mL of blood were obtained within 24 hours from study entry. After centrifugation, plasma was kept at −80°C until assayed. sST2 values were determined in duplicate by enzyme-linked immunosorbent assay using a commercial kit provided by R&D (Minneapolis, MN). The manufacturer’s instructions were followed. Results are given as pg/mL determined by comparison with a standard titration curve included in each assay plate. Quantitative measurements of PCT concentrations were performed using a sandwich immunoluminometric method (LIAISON BRAHMS PCT; Brahms Diagnostica, Henningdorf BEI, Berlin). CRP values were determined employing an immunonephelometric assay (N High Sensitivity CRP, Dade Behring Diagnostics, Deerfield, IL).
All statistical analysis were performed with MedCalc for Windows (version 7.3, MedCalc Software, Mariakerke, Belgium), with the use of contingency tables, χ2 test and Fisher exact test for categorical variables and the Student t test or the Mann-Whitney U test for continuous variables and are presented as mean ± SD or median with 25th and 75th percentile where appropriate. Statistical significance was set at P < 0.05. The diagnostic performances of sST2, PCT and CRP were investigated by receiver operating characteristic (ROC) analysis.27
The demographic and clinical characteristics and the causes of SIRS in 64 children included in the study are summarized in Tables 1 and 2. Among B-SIRS 10 patients (24%) were <44 weeks postmenstrual age, whether in the NB-SIRS group they were 4 (19%). All patients were given antibiotics during the first 24 hours of hospitalization. Among the B-SIRS group, Escherichia coli (n = 8), Enterococcus faecalis (n = 3), Pseudomonas spp. (n = 1) were isolated in urine cultures; Group B Streptococcus (n = 3), Staphylococcus epidermidis (n = 2), Enterobacter cloacae (n = 1), Streptococcus pneumoniae (n = 1), Staphylococcus aureus (n = 3), Klebsiella oxytoca (n = 1) and E. coli (n = 2) were isolated in blood cultures and Streptococcus pyogenes (n = 1), Neisseria meningitidis (n = 1), S. pneumoniae (n = 1) were isolated in cerebrospinal fluid cultures and Salmonella enteritidis (n = 4) in feces.
sST2 in Patients With Non-B-SIRS and B-SIRS
Table 3 summarizes data on sST2 in comparison with PCT and CRP in non-B-SIRS and B-SIRS children. Median sST2, PCT and CRP were significantly higher in children with B-SIRS compared with those with non-B-SIRS. The levels of sST2 in both non-B-SIRS and B-SIRS patients were also considerably increased over healthy controls. Table 3 also shows that among children with B-SIRS, median sST2 and PCT was significantly higher in the 12 patients with bacteremia compared with the 29 patients with other causes of B-SIRS, whereas no significant differences were documented in median levels of CRP. In addition, sST2, PCT and CRP levels were not different between patients with proved and probable viral infection (data not shown).
Sensitivity and Specificity Measurements for Prediction of B-SIRS
ROC curves were generated for sST2, PCT and CRP at admission with SIRS (Fig. 1). The area under the curve demonstrated good individual discriminatory power for sST2 and PCT (0.9, 95% confidence interval (CI): 0.81–0.96; 0.89, 95% CI: 0.77–0.96, respectively), whereas CRP demonstrated poor discriminatory power (0.65, 95% CI: 0.51–0.81). Pairwise comparison of ROC curves demonstrated a similar discriminatory power for sST2 and PCT (difference between areas, 0.0546, 95% CI: 0.04–0.15, P = 0.278) and significantly greater discriminatory power for both sST2 and PCT vs. CRP (difference between areas: sST2 vs. CRP, 0.245, 95% CI: 0.07–0.42, P = 0.006; PCT vs. CRP, 0.19, 95% CI: 0.02–0.36, P = 0.0267). Table 4 shows the computed specificities, sensitivities, positive and negative predictive values, positive and negative likelihood ratios for sST2, PCT and CRP with regard to early identification of B-SIRS, at cutoff values of 426 pg/mL for sST2, 0.55 ng/dL for PCT and 0.33 mg/dL for CRP. These cutoff values were determined using Youdens Index. Then, we evaluated the sensitivity and specificity of combinations of sST2, PCT and CRP in the distinction between non-B-SIRS and B-SIRS using the cutoffs determined by the ROC analysis. Table 5 shows the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the combinations of the biomarkers.
In the quest of new biochemical markers that may offer pediatricians the diagnostic criteria to distinguish bacterial from viral infection, we have explored the possible role of sST2. Our findings demonstrate that children with SIRS associated with bacterial infection present significantly increased levels of sST2, when compared with patients with SIRS of viral etiology. More important, ROC curve analysis indicated that sST2 has a significant diagnostic performance with respect to early identification of SIRS of bacterial etiology, which is similar to PCT and greater than that of CRP.
Previous studies in adults have demonstrated elevated sST2 levels in serum of patients affected by various disorders, including sepsis.21–24 In particular, Hoogerwerf et al24 have established that sST2 levels correlate with disease severity and mortality in severe sepsis and Dieplinger and coll28 showed the prognostic value of sST2 in patients admitted to an intensive care unit. The patient population studied did not include cases of severe sepsis; therefore, any inference on the prognostic value of sST2 cannot be made. However, it is intriguing that the levels of sST2 and of another laboratory parameter indicative of enhanced inflammation, such as PCT, were significantly increased in children with bacteremia when compared with children with other causes of B-SIRS. This suggests that the extent of release of sST2 in the circulation may correlate with the severity of inflammation. However, further investigation is warranted to elucidate the value of assaying sST2 as a tool to measure SIRS severity in children. In addition, considering the association between sST2 and dampening of the immune response,29 it would be of considerable interest to examine a possible relationship between sST2 levels and the occurrence of secondary nosocomial infection.
Early identification of bacterial infection has a major impact on the clinical course, management and outcome of children with SIRS. Efforts have thus been made to find a reliable marker of bacterial infection. Among the potentially useful markers to differentiate between bacterial and viral infections, PCT and CRP have been proposed to be the most promising. However, the clinical utility of these biomarkers remains controversial.5–11 As we show here, sST2 has a high diagnostic reliability. In particular, sST2 showed a relatively high sensitivity (78.3%) and NPV (75%). Moreover, the dosage of sST2 in serum is quite rapid, easy and requires a low sample volume. However, we only detect sST2 level within 24 hours from hospitalization in a relatively small sample size; thus, this study does not provide information on a possible temporal relationship between sST2 and the course of disease. Larger studies are needed to establish whether sST2 is also a late marker and might be of value in identifying patients in whom antibiotics can be stopped early. Indeed, no single SIRS biomarker is without limitation. The complexity of the host response to infection, host characteristics and type and extent of the infectious involved pathogen may not lend to the identification of a single ideal marker. As such, it may prove more useful to combine various markers. Regarding diagnostic accuracy, the combination of sST2 plus PCT and/or polymerase chain reaction, and PCT plus CRP increased their sensitivity and NPV compared with sST2, PCT and polymerase chain reaction alone. Because prompt and effective antibiotic treatment is crucial in the treatment of patients with bacterial infections, diagnostic markers of infections should have a high sensitivity. Moreover, a high NPV may be useful to exclude a nonlife threatening condition such as SIRS of viral etiology and is desirable to avoid unnecessary hospital admissions and the unnecessary use of antibiotics.
Some limitations of this study deserve consideration. First, the number of patients enrolled was rather low, and only 23 were diagnosed as having SIRS associated to viral infection. Second, sST2 measurement was performed when the diagnosis of SIRS was made. This does not exclude that SIRS was already present before enrolment. Third, we used clinical criteria and microbiological evidence, and it may have been difficult to ascertain the exact cause of SIRS in all patients. In particular, there was a relatively high proportion of children diagnosed with probable viral infection based on the exclusion of other known causes. Moreover, because prompt and effective antibiotic treatment is crucial in the treatment of patients with bacterial infections, all children were given antibiotics at admission. Thus, is possible that some of the children included in this group had indeed a SIRS of bacterial etiology. This may have introduced some misclassification bias. Finally, the diagnostic accuracies of disease markers are highly dependent on the setting in which they are tested and predictive value varies with the pretest probability of disease, thus our results may not be applicable to populations of children with different prevalence of bacterial infections.
In conclusion, sST2 levels increase in children with SIRS and its diagnostic accuracy, with respect to early identification of SIRS of bacterial etiology, is similar to that of PCT and greater than that of CRP. Moreover, the combination of sST2 plus PCT and/or CRP increased their sensitivity and NPV but decreased their PPV compared with sST2 alone. The sensitivity and specificity of a test have limited clinical usefulness as they cannot be used to estimate the probability of disease in an individual patient. Predictive values may be used to estimate this but both NPV and PPV vary according to the prevalence of the condition for which the test is used. Therefore, the practical application of sST2 measurement, as well as that of any other diagnostic test, must be made with great caution. Also, the clinical relevance of sST2 measurement in children with SIRS should be evaluated in the context of large multicentric trials. Further studies are also needed to assess the importance of sST2 measurement during the follow up and in predicting outcome of children with bacterial SIRS.
1. King C. Evaluation and management of febrile infants in the emergency department. Emerg Med Clin North Am. 2003;21:89–99
2. Ciesla G, Leader S, Stoddard J. Antibiotic prescribing rates in the US ambulatory care setting for patients diagnosed with influenza, 1997-2001. Respir Med. 2004;98:1093–1101
3. Gaur AH, Hare ME, Shorr RI. Provider and practice characteristics associated with antibiotic use in children with presumed viral respiratory tract infections. Pediatrics. 2005;115:635–641
4. Nyquist AC, Gonzales R, Steiner JF, et al. Antibiotic prescribing for children with colds, upper respiratory tract infections, and bronchitis. JAMA. 1998;279:875–877
5. Harbarth S, Holeckova K, Froidevaux C, et al. Diagnostic value of procalcitonin, interleukin-6, and interleukin-8 in critically ill patients admitted with suspected sepsis. Am J Respir Crit Care Med. 2001;164:396–402
6. Simon L, Saint-Louis P, Amre DK, et al. Procalcitonin and C-reactive protein as markers of bacterial infection in critically ill children at onset of systemic inflammatory response syndrome. Pediatr Crit Care Med. 2008;9:407–413
7. Jaye DL, Waites KB. Clinical applications of C-reactive protein in pediatrics. Pediatr Infect Dis J. 1997;16:735–46; quiz 746
8. Peltola H, Jaakkola M. C-reactive protein in early detection of bacteremic versus viral infections in immunocompetent and compromised children. J Pediatr. 1988;113:641–646
9. Tang BM, Eslick GD, Craig JC, et al. Accuracy of procalcitonin for sepsis diagnosis in critically ill patients: systematic review and meta-analysis. Lancet Infect Dis. 2007;7:210–217
10. Müller B, Christ-Crain M, Schuetz P. Meta-analysis of procalcitonin for sepsis detection. Lancet Infect Dis. 2007;7:498–499; author reply 502
11. Reinhart K, Brunkhorst FM. Meta-analysis of procalcitonin for sepsis detection. Lancet Infect Dis. 2007;7:500–502; author reply 502
12. Yanagisawa K, Takagi T, Tsukamoto T, et al. Presence of a novel primary response gene ST2L, encoding a product highly similar to the interleukin 1 receptor type 1. FEBS Lett. 1993;318:83–87
13. Tominaga S. A putative protein of a growth specific cDNA from BALB/c-3T3 cells is highly similar to the extracellular portion of mouse interleukin 1 receptor. FEBS Lett. 1989;258:301–304
14. Yanagisawa K, Naito Y, Kuroiwa K, et al. The expression of ST2 gene in helper T cells and the binding of ST2 protein to myeloma-derived RPMI8226 cells. J Biochem. 1997;121:95–103
15. Xu D, Chan WL, Leung BP, et al. Selective expression of a stable cell surface molecule on type 2 but not type 1 helper T cells. J Exp Med. 1998;187:787–794
16. Bergers G, Reikerstorfer A, Braselmann S, et al. Alternative promoter usage of the Fos-responsive gene Fit-1 generates mRNA isoforms coding for either secreted or membrane-bound proteins related to the IL-1 receptor. EMBO J. 1994;13:1176–1188
17. Hayakawa H, Hayakawa M, Kume A, et al. Soluble ST2 blocks interleukin-33 signaling in allergic airway inflammation. J Biol Chem. 2007;282:26369–26380
18. Sanada S, Hakuno D, Higgins LJ, et al. IL-33 and ST2 comprise a critical biomechanically induced and cardioprotective signaling system. J Clin Invest. 2007;117:1538–1549
19. Kumar S, Tzimas MN, Griswold DE, et al. Expression of ST2, an interleukin-1 receptor homologue, is induced by proinflammatory stimuli. Biochem Biophys Res Commun. 1997;235:474–478
20. Tajima S, Oshikawa K, Tominaga S, et al. The increase in serum soluble ST2 protein upon acute exacerbation of idiopathic pulmonary fibrosis. Chest. 2003;124:1206–1214
21. Kuroiwa K, Arai T, Okazaki H, et al. Identification of human ST2 protein in the sera of patients with autoimmune diseases. Biochem Biophys Res Commun. 2001;284:1104–1108
22. Weinberg EO, Shimpo M, Hurwitz S, et al. Identification of serum soluble ST2 receptor as a novel heart failure biomarker. Circulation. 2003;107:721–726
23. Brunner M, Krenn C, Roth G, et al. Increased levels of soluble ST2 protein and IgG1 production in patients with sepsis and trauma. Intensive Care Med. 2004;30:1468–1473
24. Hoogerwerf JJ, Tanck MW, van Zoelen MA, et al. Soluble ST2 plasma concentrations predict mortality in severe sepsis. Intensive Care Med. 2010;36:630–637
25. Goldstein B, Giroir B, Randolph AInternational Consensus Conference on Pediatric Sepsis. . International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med. 2005;6:2–8
27. Campbell G. Advances in statistical methodology for the evaluation of diagnostic and laboratory tests. Stat Med. 1994;13:499–508
28. Dieplinger B, Egger M, Koehler W, et al. Prognostic value of soluble ST2 in an unselected cohort of patients admitted to an intensive care unit - The Linz Intensive Care Unit (LICU) study. Clin Chim Acta. 2012;413:587–593
29. Liew FY, Xu D, Brint EK, et al. Negative regulation of toll-like receptor-mediated immune responses. Nat Rev Immunol. 2005;5:446–458
marker; pediatric; systemic inflammatory response syndrome; diagnosis© 2014 by Lippincott Williams & Wilkins, Inc.