The Journal of Emergency Medicine, Vol. 47, No. 6, pp. 682–688, 2014 Copyright Ó 2014 Elsevier Inc. Printed in the USA. All rights reserved 0736-4679/$ - see front matter
http://dx.doi.org/10.1016/j.jemermed.2014.07.034
Clinical Laboratory in Emergency Medicine
USE OF SERUM PROCALCITONIN IN EVALUATION OF FEBRILE INFANTS: A META-ANALYSIS OF 2317 PATIENTS Jasmin Tamsut England, MD,* Michael T. Del Vecchio, MD,† and Stephen C. Aronoff, MD† *Department of Pediatrics, UCLA Geffen School of Medicine, Los Angeles, California and †Temple University School of Medicine, Philadelphia, Pennsylvania Corresponding Address: Michael T. Del Vecchio, MD, Department of Pediatrics, Temple University School of Medicine, 2nd Floor Kresge West, 3440 N. Broad St., Philadelphia, PA 19140
, Abstract—Background: Serum procalcitonin (PCT) concentrations have been studied as a diagnostic test for serious bacterial infections (SBIs) in children. However, the utility of a single measurement in the evaluation of SBIs in febrile infants younger than 91 days is not clear. Objective: Use a systematic review and meta-analysis to determine: 1) the ability of serum PCT concentrations to identify febrile infants < 91 days of age at high and low risk for SBIs, and 2) to compare its utility with available clinical prediction rules. Methods: The literature search identified studies of febrile infants segregated into risk groups using serum PCT concentrations. Some authors were contacted to provide subgroups < 91 days of age or to provide data with 0.3 ng/mL PCT cutoff values. Data were combined and validated using standard methodologies. Results: Seven studies encompassing 2317 patients were identified; five of seven studies used a PCT discriminating concentration of 0.3 ng/mL. No heterogeneity or publication bias was identified. The overall relative risk (RR) was 3.97 (95% confidence interval [CI] 3.41–4.62) and was consistent by sensitivity analysis. The RR from a systematic review of clinical prediction rules was 30.6 (95% CI 7.0–68.13) and 8.75 (95% CI 2.29–15.2) for infants untreated and treated with antibiotics, respectively. Conclusions: Alone, measurement of serum PCT concentrations, though able to identify a group of young infants at risk for SBIs, is inferior to the available clinical prediction rules for identifying young, febrile infants at risk for SBIs. Serum concentrations # 0.3 ng/mL may be
helpful as an add-on test to current rules for identifying low-risk, febrile infants. Ó 2014 Elsevier Inc. , Keywords—procalcitonin; infants; bacterial; febrile; meta-analysis
INTRODUCTION Procalcitonin (PCT), a precursor of the vitamin D regulatory hormone calcitonin, is produced mainly by the liver, but is also produced by the kidney as well as most other tissues during sepsis and infection (1–5). Induction of PCT synthesis is complex and seems to be dependent on bacterial products, such as lipopolysaccharide, as well as soluble host mediators such as tumor necrosis factora, interleukin (IL)-2, and IL-6 (2,6). The observations that PCT mediates release of proinflammatory cytokines from circulating blood cells and that antibody blockade of this prohormone reduces mortality in a septic hamster model suggest a role for PCT in the initial host response to infection (7,8). PCT appears in the serum prior to other acute inflammatory markers such as peripheral leukocytes and C-reactive protein (CRP) in children with serious bacterial infections (SBIs) (9,10). As a result, serum PCT concentrations have been studied as a diagnostic test for neonatal sepsis, severe infection in febrile, neutropenic children, sepsis in adults, and bone and joint infections
Reprints are not available.
RECEIVED: 30 January 2014; FINAL SUBMISSION RECEIVED: 10 June 2014; ACCEPTED: 1 July 2014 682
Procalcitonin in Febrile Infants
(11–14). These studies reported a variety of results. Yu et al., in a meta-analysis of studies focused on neonates, found PCT to have moderate accuracy, with greater specificity than sensitivity (11). Lin et al. used a metaanalysis to evaluate the use of PCT in febrile, neutropenic, pediatric oncology patients and found PCT to be comparable to CRP, with overall higher specificity but lower sensitivity (12). Simon et al., in a meta-analysis comparing PCT and CRP in all age groups with bacterial infections, found PCT to be more accurate than CRP (13). Finally, Shen et al., in a meta-analysis of bone and joint infections from all age groups, found PCT to be more specific than sensitive (14). The assessment and management of febrile infants < 91 days of age is a vexing problem in pediatrics. These infants, particularly those without focal findings, are at significant risk for SBIs (15,16). Many of these patients are admitted to the hospital, evaluated for serious infections including meningitis, and treated with parenteral antimicrobial therapy (17). In 1985, Dagan and his colleagues in Rochester developed a clinical prediction rule that successfully separated the population of febrile infants < 91 days of age into high-risk and low-risk for SBI groups; this rule was subsequently validated in a large, prospective study (18,19). Based on these and other data, a modified version of this clinical prediction rule was recommended to the pediatric community at large as a guideline for the management of this patient population (20). This systematic review and metaanalysis were undertaken to determine: 1) if serum concentrations of PCT in healthy, febrile infants < 91 days of age could discriminate between those infants with SBIs and those without; and 2) if the measurement of serum concentrations of PCT was a superior discriminator to the currently used clinical prediction rules.
683
2. Documented a temperature by history or on examination of 38 C or greater; 3. Obtained serum concentrations of PCT as part of the initial evaluation and reported the results quantitatively; 4. Used an a priori protocol to define SBIs; and 5. Excluded infants with known risk factors for infection such as prematurity, heart defects, or congenital abnormalities. Data Abstraction A standardized tool was developed to extract data from each article meeting inclusion criteria. The data were extracted by one author and reviewed by two others for completeness and accuracy. Inclusive age range, serum PCT values, PCT cut-off values, number of patients with SBI, and the number of patients with nonbacterial infections were recorded. The primary outcome was the number of patients in each group with SBIs. Statistical Analysis From extracted data, the relative risks (RR) were calculated for each study. Heterogeneity among studies was determined from the c2 distribution of Cochran’s Q statistic; p < 0.05 was considered to be significant (22). I2 was calculated to determine the percentage of heterogeneity across studies (23). In the absence of significant heterogeneity, the fixed-effect method was used to combine study outcomes (24). Sensitivity analysis was performed by reiterative calculations of the composite RR after excluding single studies one at a time. All calculations were performed with Comprehensive Meta-Analysis Software, version 2.2.050, www.Meta-analysis.com.
MATERIALS AND METHODS Study Identification and Abstraction The PRISMA protocol was used (21). In September 2012 and again in March 2013, the Medline database dating back to 1966 was searched using PubMed as the primary search engine. Search terms included ‘‘procalcitonin,’’ ‘‘infant or neonatal or pediatric,’’ ‘‘infection,’’ ‘‘cohort,’’ and ‘‘observational.’’ Filters included ‘‘infant,’’ ‘‘newborn,’’ and ‘‘neonate.’’ The bibliographies of the selected studies and review articles were searched for relevant studies. Study Selection Included studies met the following criteria: 1. Studied a patient population that included subjects < 91 days of age;
RESULTS The results of the literature search and culling process are shown in Figure 1. The initial search of Medline and selected bibliographies yielded 158 titles. Seventeen studies met inclusion criteria after abstracts and titles were scanned. Four of these articles were excluded because SBIs were not a reported study outcome (25–28). Two studies included only subjects at increased risk for infection; two studies failed to report quantitative PCT data; one study did not separate infants < 91 days of age from an older subject population; and one study was a review (9,29–33). The remaining seven studies are summarized in Table 1 (34–40). The authors of the three studies that included patients up to 36 months of age provided summary data for all patients < 91 days of age (34,36,38). Dr. Borja
684
J. T. England et al.
Figure 1. Flow chart of study selection.
Gomez kindly provided summary data using a cut-off PCT concentration of 0.3 ng/mL (35). Two studies used alternate cut-off values for serum PCT concentrations: Olacirgui et al. used 0.5 ng/mL and Maniaci et al. used 0.12 ng/mL (37,40). All of the studies used immunoluminometric assays to determine PCT; three different products were used (Liaison BRAHMS PCT and BRAHMS PCT-Q, both from BRAHMS Diagnostica, Hennigsdorf BEI, Berlin; and quantitative PCT Roche Diagnostics, Manheim, Germany). Similar definitions for SBIs were used. In all of the studies, positive bacterial cultures from typical sterile sites were definitive for infections of blood, cerebral spinal fluid, urine, and joint spaces. Bacterial enteritis was defined by the identification of a known pathogen from stool. In three studies, sepsis was defined by the criteria
determined by Levy et al. (34,36,40,41). Pulmonary infections were defined radiographically in four studies; bone infections were defined by scintigraphy in two studies (34,37,38,40). All of the studies required more than 104 colony-forming units (CFU)/mL of a single organism to define an infection of the urinary tract; in three cases, additional data were used to divide patients into possible or definite groups (35–37). For the purposes of the present review, these groups were combined. The ability of serum PCT concentration to discriminate between febrile infants with and without SBI is shown in Figure 2 and Table 2. All of these studies demonstrated that patients with serum PCT concentrations above the selected cut-off values were more likely to have an SBI than those with serum concentrations below the cut-off value. The RR point estimates ranged from 3.55 to 8.47
Table 1. Summary of Included Studies PCT > Cut-off Study (Reference No.)
Age Range
Andreola et al. 2007 (34)
7 days to 36 months < 91 days of age 1–36 months
Gomez et al., 2012 (35) Luaces-Cubells et al., 2012 (36) Maniaci et al., 2008 (37) Manzano et al., 2010 (38) Olacirgui et al., 2009 (40) Woelker et al., 2012 (39) Pooled data
PCT < Cut-off
No. Evaluable < 91 Days of Age
PCT Cut-off ng/mL
SBI+
107
0.3
20
13
60.6%
8
66
10.8%
1112 325
0.3 0.3
137 8
88 40
60.9% 16.7%
152 12
735 265
17.1% 4.3%
234 37 347 155 2317
0.12 0.3 0.5 0.3
40 6 52 11 274
143 2 34 47 367
21.9% 75.0% 60.5% 19.0% 42.7%
2 4 30 2 210
49 25 231 95 1466
3.9% 13.8% 11.5% 2.1% 12.5%
< 91 days of age 1–36 months < 91 days of age 2–60 days
PCT = procalcitonin; SBI = serious bacterial infection.
SBI
SBI Rate
SBI+
SBI
SBI Rate
Procalcitonin in Febrile Infants
685 Table 3. Sensitivity Analysis of 7 Studies: Procalcitonin and Serious Bacterial Infection in Infants < 91 Days of Age Composite Relative Risk with Study Removed
Figure 2. Forest plot of relative risk from included studies. CI = confidence interval.
among the seven studies. The point estimate for the overall RR was 3.97 (95% confidence interval [CI] 3.41–4.62), favoring patients with SBIs and serum concentrations of PCT above the cut-off values. Despite the variability of cut-off values used, no heterogeneity among the studies was identified (Q = 6.16, p = 0.41, I2 = 2.61). The sensitivity analysis is shown in Table 3. The summary RR varied from 3.76 (95% CI 3.19–4.44) to 5.25 (95% CI 3.95–6.97), with single studies elimination (35,40). Excluding the study with the lowest (0.12 ng/ mL) or the highest (0.5 ng/mL) PCT cut-off concentration did not significantly alter the resulting composite RR: 3.95 (95% CI 3.39–4.60) and 3.76 (95% CI 3.19–4.44), respectively (37,40). DISCUSSION The evaluation and management of the febrile infant < 91 days of age is one of the most common yet complex issues in pediatrics. Roberts and Borzy described 61 febrile infants 8 weeks of age or less; 9 of these infants were bacteremic and 23 had an identified focus of infection (15). In a systematic review, Baraff et al. found that Table 2. Individual and Composite Relative Risks of Serious Bacterial Infection in 7 Studies Using Serum Procalcitonin Concentrations as the Discriminator Study (Reference No.)
Relative Risk
95% CI
Andreola et al. 2007 (34) Gomez et al., 2012 (35) Luaces-Cubells et al., 2012 (36) Maniaci et al., 2008 (37) Manzano et al., 2010 (38) Olacirgui et al., 2009 (40) Woelker et al., 2012 (39) Overall
5.61 3.55 3.85 5.57 5.44 5.26 9.20 3.97
2.76–11.4 2.97–4.25 1.66–8.92 1.39–22.28 2.01–14.69 3.61–7.67 2.11–40.1 3.41–4.62
CI = confidence interval.
Study
Relative Risk
95% CI
All studies Excluding: Andreola et al. 2007 (34) Gomez et al., 2012 (35) Luaces-Cubells et al., 2012 (36) Maniaci et al., 2008 (37) Manzano et al., 2010 (38) Olacirgui et al., 2009 (40) Woelker et al., 2012 (39)
3.97
3.41–4.62
3.9 5.25 3.97 3.95 3.94 3.76 3.94
3.34–4.56 3.95–6.97 3.41–4.63 3.39–4.6 3.38–4.59 3.19–4.44 3.42–4.62
CI = confidence interval.
10.5% of 306 febrile infants < 3 months of age had a serious bacterial infection (16). Given this high rate of infection and the inability to clinically differentiate infected from noninfected infants, experts advocated for a universal ‘‘hospitalize and treat’’ approach (17). Using a large, prospective cohort of febrile infants < 3 months of age, investigators at the University of Rochester developed a clinical prediction rule to identify a subgroup of infants at low risk for SBI (18). In this study, 1/144 (0.7%) of the infants in the low-risk group had a serious bacterial infection, compared to 22/89 (25%) who did not meet low-risk criteria. In the following years, a large number of studies and commentaries appeared in the literature both supporting and refuting the clinical use of the ‘‘Rochester’’ criteria. In 1993, a practice guideline was published that supported outpatient management of febrile infants aged 29 to 90 days who fulfilled specific clinical, social, and laboratory criteria (20). In 2010, Huppler et al. systematically reviewed all of the studies published to date that prospectively and retrospectively evaluated clinical prediction rules designed to identify a low-risk population of febrile infants < 91 days of age (42). From prospective studies only, pooled rates of serious bacterial infections among low-risk febrile infants were 2.7% and 0.67%, respectively, for those who did and did not receive antimicrobial therapy a priori; among high-risk infants, the SBI rates were 23.8% and 20.6%, respectively. In the prospective studies, the incidence of SBI was 11.1% (497/4481) and 9.4% (174/1858) for all febrile infants who received and did not receive empirical antimicrobial therapy, respectively. Using serum PCT concentrations as the discriminator, pooled data from the present study found that SBIs occurred in 42.4% (274/646) of febrile infants in the high-risk group (PCT concentrations > cut-off value) and in 12.5% (210/ 1676) of febrile infants in the low-risk group (PCT concentrations < cut-off value); the incidence of SBIs in
686
the pooled cohort was 20.9% (484/2317) (Table 1). The difference in incidence of SBIs between the pooled data for the clinical prediction rule studies and the PCT studies, independent of the strength of the tests, may account for the apparent differences in positive and negative predictive values. Using clinical prediction rules as the discriminator and only prospective studies, Huppler et al. calculated overall RR for SBIs in high-risk febrile infants of 30.6 (95% CI 7.0–68.13) and 8.75 (95% CI 2.29–15.2) for nonrecipients and recipients of antimicrobial therapy, respectively (42). In the present study, the RR for SBIs among high-risk febrile infants differentiated by serum PCT concentrations was 3.97 (95% CI 3.41–4.62). These data suggest that using a serum PCT cut-off value of 0.3 ng/mL to identify those infants at risk for SBIs is inferior to the validated clinical prediction rules currently used for the evaluation of febrile infants < 91 days of age. Limitations Selection bias in systematic reviews and meta-analyses may arise from two sources. Failure to recover all of the data during the literature search process is always a concern. Search strategies are never perfect, so completeness is always an open question with these studies. Medline was the only database searched for this study; it is possible relevant publications that appeared in other databases may have been missed. Publication bias contributes to the selection issue because there is risk that wellperformed studies evaluating the utility of PCT serum concentrations as a predictor of SBI in infants failed to pass peer-review muster. The results from the Egger test argue against this problem in the present study. Combinability of studies is another potential source of bias. The included studies employed different assays for measuring serum concentrations of PCT. de Wolf et al. demonstrated that the Roche assay and the BRAHMS Kryptor kit yielded comparable results, whereas Schuch et al. demonstrated comparability between the two BRAHMS PCT assays (43,44). Together these studies suggest that the PCT results are comparable. The lack of heterogeneity among the studies used in this review further support this notion. Variability in the outcome definitions is another source of bias in this study. Overall, the studies were in agreement with regard to definitions of blood, cerebrospinal fluid, and gastrointestinal infections. Whereas there was some variability in the definition of urinary tract infection, the use of 104 CFU/mL was uniform. Conversely, only two studies included bone infections; four of seven studies included pulmonary infections (34,37,38,40). Finally, the conclusions of this study regarding the effectiveness of serum PCT vs. the clinical prediction
J. T. England et al.
rule for the identification of febrile infants with serious bacterial infections are based on the results of two meta-analyses. A head-to-head, prospective comparison study is required to resolve this issue. CONCLUSION The data from this review suggest that serum PCT concentrations # 0.3 ng/mL identify a population of febrile infants < 91 days of age at low risk for SBIs. Febrile infants with a serum concentration in excess of 0.3 ng/mL have, on average, a 3.97-fold increased risk of SBIs. Recent data suggest that infants who fail to meet standardized, low-risk, clinical criteria for SBI have, at best, an average increased risk of 8.75-fold. These differences are largely the result of the broad range of falsenegative results noted in Table 1 (2.1% to 17%). Together these observations suggest that serum PCT concentrations alone are a poorer predictor of SBI. These observations do not support the ‘‘stand alone’’ use of serum PCT concentrations as a screen for SBIs in febrile infants < 91 days of age. Serum PCT concentrations may be used in combination with a clinical prediction rule. Future studies should incorporate serum PCT into existing clinical prediction rules to determine if it adds to the effectiveness of the rules. REFERENCES 1. Bracq S, Machason M. Calcitonin gene expression in normal human liver. FEBS Lett 1993;331:14–8. 2. Assicot M, Gendrel D, Carsin H, Raymond J, Guilbaud J, Bohoun C. High serum procalcitonin concentrations in patients with sepsis and infection. Lancet 1993;341:515–8. 3. Redl H, Schalg G, Togel E, Assicot M, Bohoun C. Procalcitonin release patterns in a baboon model of trauma and sepsis: relationship to cytokines and neopterin. Crit Care Med 2000;28:3659–63. 4. Mesiner M. Pathobiochemistry and clinical use of procalcitoin. Clin Chem Acta 2002;323:17–29. 5. Morganthaler N, Struck J, Chancerelle Y, et al. Production of procalcitonin (PCT) in non-thyroidal tissue after LPS injection. Horm Metab Res 2003;35:290–5. 6. Nijsten M, Olinga P, The T, et al. Procalcitonin behaves as a fast responding acute phase protein in vivo and in vitro. Crit Care Med 2000;28:458–61. 7. Nylen E, Whang K, Snider R, Steinwald P, White J, Becker K. Mortality is increased by procalcitonin and decreased by an antiserum reactive to procalcitonin in experimental sepsis. Crit Care Med 1998;26:1001–6. 8. Liappis A, Gibbs K, Nylen E, et al. Exogenous procalcitonin evokes a pro-inflammatory cytokine response. Inflamm Res 2011; 60:203–7. 9. Galleto-Lacour A, Zamora S, Grervaix A. Bedside procalcitonin and C-reactive protein tests in children with fever without localizing signs of infection seen in a referral center. Pediatrics 2003;112: 1054–60. 10. Hatherill M, Tibby S, Sykes K, Turner C, Murdoch I. Diagnostic markers of infection: comparison of procalcitonin with C-reactive protein and leukocyte count. Arch Dis Child 1999;81:417–21. 11. Yu Z, Liu J, Sun Q, Han S, Guo X. The accuracy of the procalcitonin test for the diagnosis of neonatal sepsis: a meta-analysis. Scand J Infect Dis 2010;42:723–33.
Procalcitonin in Febrile Infants 12. Lin S, Hou T, Huang D. Role of procalcitonin in the diagnosis of severe infection in pediatric patients with fever and neutropenia. Pediatr Infect Dis J 2012;31:e182–8. 13. Simon L, Gauvin F, Amre D, Saint-Louis P, Lacroix J. Serum procalcitonin and C-reactive protein levels as markers of bacterial infection: a systematic review and meta-analysis. Clin Infect Dis 2005;40:1386–8. 14. Shen C, Wu M, Lin K, et al. The use of procalcitonin in the diagnosis of bone and joint infection: a systematic review and metaanalysis. Eur J Clin Microbiol Infect Dis 2013;32:807–14. 15. Roberts K, Borzy M. Fever in the first eight weeks of life. Johns Hopkins Med J 1977;141:9–13. 16. Baraff L, Oslund S, Schriger D, Stephen M. Probability of bacterial infections in febrile infants less than three months of age: a metaanalysis. Pediatr Infect Dis J 1992;11:257–64. 17. Roberts K. Management of young, febrile infants. Primum non nocere revisited. Am J Dis Child 1983;137:1143–4. 18. Dagan R, Powell K, Hall C, Menegus M. Identification of infants unlikely to have serious bacterial infection although hospitalized for suspected sepsis. J Pediatr 1985;107:855–60. 19. Jaskiewicz J, McCarthy C, Richardson A, et al. Febrile infants at low risk for serious bacterial infection-an appraisal of the Rochester criteria and implications for management. Febrile infant collaborative study group. Pediatrics 1994;94:390–6. 20. Baraff L, Bass J, Fleisher G, et al. Practice guideline for the managment of infants and children 0 to 36 months of age with fever without source. Pediatrics 1993;92:1–12. 21. PRISMA: Transparent reporting of systematic reviews and meta-analyses. The PRISMA Checklist. 2009. Available at: http://www. prisma-statement.org/statement.htm. Accessed November 27, 2013. 22. Cochrane W. The combination of estimates from different experiments. Biometrics 1954;10:101–29. 23. Higgins J, Thompson S, Deeks J, Altman D. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. 24. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: 719–48. 25. Ibrahim K, Abdel-Wahab A, Ibrahim A. Diagnostic value of serum procalcitonin levels in children with meningitis: comparison with blood leukocyte count and C-reative protein. J Pak Med Assoc 2011;61:346–51. 26. Paize F, Carrol W, Downey C, et al. Diagnostic efficacy of activated partial thromboplastin time waveform and procalcitonin analysis in pediatric meningococcal sepsis. Pediatr Crit Care Med 2011;12: e322–9. 27. Leroy S, Romanello C, Galletto-Lacour A, et al. Procalcitonin is a predictor for high-grade vesicoureteral reflux in children: metaanalysis of individual patient data. J Pediatr 2011;159:644–51. 28. Lacroix J, Manzano S, Galetto A, Gervaix A. Procalcitonin measurement for detection of serious bacterial infection in febrile children: comaprison between two automated immunoassays. Clin Biochem 2012;45:593–5. 29. Unal S, Arslankoylu A, Kuyucu N, Aslan G, Erdogan S. Procalcitonin is more useful than C-reactive protein in differentiatin of fever in patients with sickle cell disease. J Pediatr Hematol Oncol 2012;34: 85–9.
687 30. Bohnhorst B, Lange M, Bartels D, Bejo L, Hoy L, Peter C. Procalcitoniin and valuable clinical symptoms in the early detection of neonatal late-onset bacterial infection. Acta Paediatr 2012;101: 19–25. 31. Bressan S, Gomez B, Mintegi S, et al. Diagnsotic performance of the lab-score in predicting severe and invasive bacterial infections in well-appearing young febrile infants. Pediatr Infect Dis J 2012; 31:139–44. 32. Fernandez-Lopez A, Luaces-Cubells C, Garcia-Garcia J-J, Fernandez Pou J, Spanish Society of Pediatric Emergencies. Procalcitonin in pediatric emergency departments for the early diagnosis of invasive bacterial infections in febrile infants: results of a multicenter study and utility of a rapid qualitative test for this marker. Pediatr Infect Dis J 2003;22:895–903. 33. Srinivasan L, Harris M. New technologies for the rapid diagnosis of neonatal sepsis. Curr Opin Pediatr 2012;24:165–71. 34. Andreola B, Bressan S, Callegaro S, Liverani A, Plebani M, Da Dalt L. Procalcitonin and C-reactive protein as diagnostic markers of severe bacterial infections in febrile infants and children in the emergency department. Pediatr Infect Dis J 2007;26: 672–7. 35. Gomez B, Bressan S, Mintegi S, et al. Diagnostic value of procalcitonin in well-appearing young febrile infants. Pediatrics 2012;130: 815–22. 36. Luaces-Cubells C, Mintegi S, Garcia-Garcia J-J, et al. Procalcitonin to detect invasive bacterial infection in non-toxic-appearing infants with fever without apparent source in the emergency room. Pediatr Infect Dis J 2012;31:645–7. 37. Maniaci V, Dauber A, Weiss S, Nylen E, Becker K, Bachur R. Procalcitonin in young febrile infants for the detection of serious bacterial infections. Pediatrics 2008;122:701–10. 38. Manzano S, Bailey B, Girodias J-B, Galetto-Lacour A, Cousineau J, Delvin E. Impact of procalcitonin on the mangement of children aged 1 to 36 months presenting with fever without source: a randomized controlled trial. Am J Emerg Med 2010; 28:647–53. 39. Woelker J, Sinha M, Christopher N, Powell K. Serum procalcitonin concentration in the evaluation of febrile infants 2 to 60 days of age. Pediatr Emerg Care 2012;28:410–5. 40. Olacirgui I, Hernandez U, Munoz J, Emparanza J, Landa J. Markers that predict serious bacterial infection in infants under 3 months of age presenting with fever of unknown origin. Arch Dis Child 2009; 94:501–5. 41. Levy MM, Fink MP, Marshall JC, et al. International Sepsis Definitions Conference. Crit Care Med 2003;31:1250–6. 42. Huppler A, Eickoff J, Wald E. Performance of low-risk criteria in the evalaution of young infants with fever: review of the literature. Pediatrics 2010;125:228–33. 43. de Wolf HK, Gunnewiek JK, Berk Y, van den Ouweland J, de Metz M. Comparison of a new procalictonini assay from Roche with the established method of the Brahms Kryptor. Clin Chem 2009;55:1043–4. 44. Schuch G, Duc-Marchand C, Venet C, Mann H, Tixler A, Bionda C. Use of procalcitonin in intensive care units: comparison of semi quantitative PCT-Q Brahms assay with autmated PCT-Kryptor assay. Anal Biol Clin (Paris) 2011;69:663–70.
688
J. T. England et al.
ARTICLE SUMMARY 1. Why is this topic important? The evaluation of the febrile infant is a common and vexing problem in the care of pediatric patients. 2. What does this study attempt to show? The use of procalcitonin in the evaluation of febrile infants is being used more and more commonly. We have attempted to assess the usefulness of this test in this clinical arena. 3. What are the key findings? The use of procalcitonin for the evaluation of the febrile infant, though of some use, must be used with caution and is inferior to clinical prediction rules already in use. 4. How is patient care impacted? Clinicians have a clearer understanding of the usefulness of procalcitonin in the assessment of the febrile infant. With this understanding they can more accurately apply this test to the clinical situation and enhance their clinical management of this population. Additionally, the present study may be used to guide further research.
http://dx.doi.org/10.1016/j.jemermed.2014.07.034
Clinical Laboratory in Emergency Medicine
USE OF SERUM PROCALCITONIN IN EVALUATION OF FEBRILE INFANTS: A META-ANALYSIS OF 2317 PATIENTS Jasmin Tamsut England, MD,* Michael T. Del Vecchio, MD,† and Stephen C. Aronoff, MD† *Department of Pediatrics, UCLA Geffen School of Medicine, Los Angeles, California and †Temple University School of Medicine, Philadelphia, Pennsylvania Corresponding Address: Michael T. Del Vecchio, MD, Department of Pediatrics, Temple University School of Medicine, 2nd Floor Kresge West, 3440 N. Broad St., Philadelphia, PA 19140
, Abstract—Background: Serum procalcitonin (PCT) concentrations have been studied as a diagnostic test for serious bacterial infections (SBIs) in children. However, the utility of a single measurement in the evaluation of SBIs in febrile infants younger than 91 days is not clear. Objective: Use a systematic review and meta-analysis to determine: 1) the ability of serum PCT concentrations to identify febrile infants < 91 days of age at high and low risk for SBIs, and 2) to compare its utility with available clinical prediction rules. Methods: The literature search identified studies of febrile infants segregated into risk groups using serum PCT concentrations. Some authors were contacted to provide subgroups < 91 days of age or to provide data with 0.3 ng/mL PCT cutoff values. Data were combined and validated using standard methodologies. Results: Seven studies encompassing 2317 patients were identified; five of seven studies used a PCT discriminating concentration of 0.3 ng/mL. No heterogeneity or publication bias was identified. The overall relative risk (RR) was 3.97 (95% confidence interval [CI] 3.41–4.62) and was consistent by sensitivity analysis. The RR from a systematic review of clinical prediction rules was 30.6 (95% CI 7.0–68.13) and 8.75 (95% CI 2.29–15.2) for infants untreated and treated with antibiotics, respectively. Conclusions: Alone, measurement of serum PCT concentrations, though able to identify a group of young infants at risk for SBIs, is inferior to the available clinical prediction rules for identifying young, febrile infants at risk for SBIs. Serum concentrations # 0.3 ng/mL may be
helpful as an add-on test to current rules for identifying low-risk, febrile infants. Ó 2014 Elsevier Inc. , Keywords—procalcitonin; infants; bacterial; febrile; meta-analysis
INTRODUCTION Procalcitonin (PCT), a precursor of the vitamin D regulatory hormone calcitonin, is produced mainly by the liver, but is also produced by the kidney as well as most other tissues during sepsis and infection (1–5). Induction of PCT synthesis is complex and seems to be dependent on bacterial products, such as lipopolysaccharide, as well as soluble host mediators such as tumor necrosis factora, interleukin (IL)-2, and IL-6 (2,6). The observations that PCT mediates release of proinflammatory cytokines from circulating blood cells and that antibody blockade of this prohormone reduces mortality in a septic hamster model suggest a role for PCT in the initial host response to infection (7,8). PCT appears in the serum prior to other acute inflammatory markers such as peripheral leukocytes and C-reactive protein (CRP) in children with serious bacterial infections (SBIs) (9,10). As a result, serum PCT concentrations have been studied as a diagnostic test for neonatal sepsis, severe infection in febrile, neutropenic children, sepsis in adults, and bone and joint infections
Reprints are not available.
RECEIVED: 30 January 2014; FINAL SUBMISSION RECEIVED: 10 June 2014; ACCEPTED: 1 July 2014 682
Procalcitonin in Febrile Infants
(11–14). These studies reported a variety of results. Yu et al., in a meta-analysis of studies focused on neonates, found PCT to have moderate accuracy, with greater specificity than sensitivity (11). Lin et al. used a metaanalysis to evaluate the use of PCT in febrile, neutropenic, pediatric oncology patients and found PCT to be comparable to CRP, with overall higher specificity but lower sensitivity (12). Simon et al., in a meta-analysis comparing PCT and CRP in all age groups with bacterial infections, found PCT to be more accurate than CRP (13). Finally, Shen et al., in a meta-analysis of bone and joint infections from all age groups, found PCT to be more specific than sensitive (14). The assessment and management of febrile infants < 91 days of age is a vexing problem in pediatrics. These infants, particularly those without focal findings, are at significant risk for SBIs (15,16). Many of these patients are admitted to the hospital, evaluated for serious infections including meningitis, and treated with parenteral antimicrobial therapy (17). In 1985, Dagan and his colleagues in Rochester developed a clinical prediction rule that successfully separated the population of febrile infants < 91 days of age into high-risk and low-risk for SBI groups; this rule was subsequently validated in a large, prospective study (18,19). Based on these and other data, a modified version of this clinical prediction rule was recommended to the pediatric community at large as a guideline for the management of this patient population (20). This systematic review and metaanalysis were undertaken to determine: 1) if serum concentrations of PCT in healthy, febrile infants < 91 days of age could discriminate between those infants with SBIs and those without; and 2) if the measurement of serum concentrations of PCT was a superior discriminator to the currently used clinical prediction rules.
683
2. Documented a temperature by history or on examination of 38 C or greater; 3. Obtained serum concentrations of PCT as part of the initial evaluation and reported the results quantitatively; 4. Used an a priori protocol to define SBIs; and 5. Excluded infants with known risk factors for infection such as prematurity, heart defects, or congenital abnormalities. Data Abstraction A standardized tool was developed to extract data from each article meeting inclusion criteria. The data were extracted by one author and reviewed by two others for completeness and accuracy. Inclusive age range, serum PCT values, PCT cut-off values, number of patients with SBI, and the number of patients with nonbacterial infections were recorded. The primary outcome was the number of patients in each group with SBIs. Statistical Analysis From extracted data, the relative risks (RR) were calculated for each study. Heterogeneity among studies was determined from the c2 distribution of Cochran’s Q statistic; p < 0.05 was considered to be significant (22). I2 was calculated to determine the percentage of heterogeneity across studies (23). In the absence of significant heterogeneity, the fixed-effect method was used to combine study outcomes (24). Sensitivity analysis was performed by reiterative calculations of the composite RR after excluding single studies one at a time. All calculations were performed with Comprehensive Meta-Analysis Software, version 2.2.050, www.Meta-analysis.com.
MATERIALS AND METHODS Study Identification and Abstraction The PRISMA protocol was used (21). In September 2012 and again in March 2013, the Medline database dating back to 1966 was searched using PubMed as the primary search engine. Search terms included ‘‘procalcitonin,’’ ‘‘infant or neonatal or pediatric,’’ ‘‘infection,’’ ‘‘cohort,’’ and ‘‘observational.’’ Filters included ‘‘infant,’’ ‘‘newborn,’’ and ‘‘neonate.’’ The bibliographies of the selected studies and review articles were searched for relevant studies. Study Selection Included studies met the following criteria: 1. Studied a patient population that included subjects < 91 days of age;
RESULTS The results of the literature search and culling process are shown in Figure 1. The initial search of Medline and selected bibliographies yielded 158 titles. Seventeen studies met inclusion criteria after abstracts and titles were scanned. Four of these articles were excluded because SBIs were not a reported study outcome (25–28). Two studies included only subjects at increased risk for infection; two studies failed to report quantitative PCT data; one study did not separate infants < 91 days of age from an older subject population; and one study was a review (9,29–33). The remaining seven studies are summarized in Table 1 (34–40). The authors of the three studies that included patients up to 36 months of age provided summary data for all patients < 91 days of age (34,36,38). Dr. Borja
684
J. T. England et al.
Figure 1. Flow chart of study selection.
Gomez kindly provided summary data using a cut-off PCT concentration of 0.3 ng/mL (35). Two studies used alternate cut-off values for serum PCT concentrations: Olacirgui et al. used 0.5 ng/mL and Maniaci et al. used 0.12 ng/mL (37,40). All of the studies used immunoluminometric assays to determine PCT; three different products were used (Liaison BRAHMS PCT and BRAHMS PCT-Q, both from BRAHMS Diagnostica, Hennigsdorf BEI, Berlin; and quantitative PCT Roche Diagnostics, Manheim, Germany). Similar definitions for SBIs were used. In all of the studies, positive bacterial cultures from typical sterile sites were definitive for infections of blood, cerebral spinal fluid, urine, and joint spaces. Bacterial enteritis was defined by the identification of a known pathogen from stool. In three studies, sepsis was defined by the criteria
determined by Levy et al. (34,36,40,41). Pulmonary infections were defined radiographically in four studies; bone infections were defined by scintigraphy in two studies (34,37,38,40). All of the studies required more than 104 colony-forming units (CFU)/mL of a single organism to define an infection of the urinary tract; in three cases, additional data were used to divide patients into possible or definite groups (35–37). For the purposes of the present review, these groups were combined. The ability of serum PCT concentration to discriminate between febrile infants with and without SBI is shown in Figure 2 and Table 2. All of these studies demonstrated that patients with serum PCT concentrations above the selected cut-off values were more likely to have an SBI than those with serum concentrations below the cut-off value. The RR point estimates ranged from 3.55 to 8.47
Table 1. Summary of Included Studies PCT > Cut-off Study (Reference No.)
Age Range
Andreola et al. 2007 (34)
7 days to 36 months < 91 days of age 1–36 months
Gomez et al., 2012 (35) Luaces-Cubells et al., 2012 (36) Maniaci et al., 2008 (37) Manzano et al., 2010 (38) Olacirgui et al., 2009 (40) Woelker et al., 2012 (39) Pooled data
PCT < Cut-off
No. Evaluable < 91 Days of Age
PCT Cut-off ng/mL
SBI+
107
0.3
20
13
60.6%
8
66
10.8%
1112 325
0.3 0.3
137 8
88 40
60.9% 16.7%
152 12
735 265
17.1% 4.3%
234 37 347 155 2317
0.12 0.3 0.5 0.3
40 6 52 11 274
143 2 34 47 367
21.9% 75.0% 60.5% 19.0% 42.7%
2 4 30 2 210
49 25 231 95 1466
3.9% 13.8% 11.5% 2.1% 12.5%
< 91 days of age 1–36 months < 91 days of age 2–60 days
PCT = procalcitonin; SBI = serious bacterial infection.
SBI
SBI Rate
SBI+
SBI
SBI Rate
Procalcitonin in Febrile Infants
685 Table 3. Sensitivity Analysis of 7 Studies: Procalcitonin and Serious Bacterial Infection in Infants < 91 Days of Age Composite Relative Risk with Study Removed
Figure 2. Forest plot of relative risk from included studies. CI = confidence interval.
among the seven studies. The point estimate for the overall RR was 3.97 (95% confidence interval [CI] 3.41–4.62), favoring patients with SBIs and serum concentrations of PCT above the cut-off values. Despite the variability of cut-off values used, no heterogeneity among the studies was identified (Q = 6.16, p = 0.41, I2 = 2.61). The sensitivity analysis is shown in Table 3. The summary RR varied from 3.76 (95% CI 3.19–4.44) to 5.25 (95% CI 3.95–6.97), with single studies elimination (35,40). Excluding the study with the lowest (0.12 ng/ mL) or the highest (0.5 ng/mL) PCT cut-off concentration did not significantly alter the resulting composite RR: 3.95 (95% CI 3.39–4.60) and 3.76 (95% CI 3.19–4.44), respectively (37,40). DISCUSSION The evaluation and management of the febrile infant < 91 days of age is one of the most common yet complex issues in pediatrics. Roberts and Borzy described 61 febrile infants 8 weeks of age or less; 9 of these infants were bacteremic and 23 had an identified focus of infection (15). In a systematic review, Baraff et al. found that Table 2. Individual and Composite Relative Risks of Serious Bacterial Infection in 7 Studies Using Serum Procalcitonin Concentrations as the Discriminator Study (Reference No.)
Relative Risk
95% CI
Andreola et al. 2007 (34) Gomez et al., 2012 (35) Luaces-Cubells et al., 2012 (36) Maniaci et al., 2008 (37) Manzano et al., 2010 (38) Olacirgui et al., 2009 (40) Woelker et al., 2012 (39) Overall
5.61 3.55 3.85 5.57 5.44 5.26 9.20 3.97
2.76–11.4 2.97–4.25 1.66–8.92 1.39–22.28 2.01–14.69 3.61–7.67 2.11–40.1 3.41–4.62
CI = confidence interval.
Study
Relative Risk
95% CI
All studies Excluding: Andreola et al. 2007 (34) Gomez et al., 2012 (35) Luaces-Cubells et al., 2012 (36) Maniaci et al., 2008 (37) Manzano et al., 2010 (38) Olacirgui et al., 2009 (40) Woelker et al., 2012 (39)
3.97
3.41–4.62
3.9 5.25 3.97 3.95 3.94 3.76 3.94
3.34–4.56 3.95–6.97 3.41–4.63 3.39–4.6 3.38–4.59 3.19–4.44 3.42–4.62
CI = confidence interval.
10.5% of 306 febrile infants < 3 months of age had a serious bacterial infection (16). Given this high rate of infection and the inability to clinically differentiate infected from noninfected infants, experts advocated for a universal ‘‘hospitalize and treat’’ approach (17). Using a large, prospective cohort of febrile infants < 3 months of age, investigators at the University of Rochester developed a clinical prediction rule to identify a subgroup of infants at low risk for SBI (18). In this study, 1/144 (0.7%) of the infants in the low-risk group had a serious bacterial infection, compared to 22/89 (25%) who did not meet low-risk criteria. In the following years, a large number of studies and commentaries appeared in the literature both supporting and refuting the clinical use of the ‘‘Rochester’’ criteria. In 1993, a practice guideline was published that supported outpatient management of febrile infants aged 29 to 90 days who fulfilled specific clinical, social, and laboratory criteria (20). In 2010, Huppler et al. systematically reviewed all of the studies published to date that prospectively and retrospectively evaluated clinical prediction rules designed to identify a low-risk population of febrile infants < 91 days of age (42). From prospective studies only, pooled rates of serious bacterial infections among low-risk febrile infants were 2.7% and 0.67%, respectively, for those who did and did not receive antimicrobial therapy a priori; among high-risk infants, the SBI rates were 23.8% and 20.6%, respectively. In the prospective studies, the incidence of SBI was 11.1% (497/4481) and 9.4% (174/1858) for all febrile infants who received and did not receive empirical antimicrobial therapy, respectively. Using serum PCT concentrations as the discriminator, pooled data from the present study found that SBIs occurred in 42.4% (274/646) of febrile infants in the high-risk group (PCT concentrations > cut-off value) and in 12.5% (210/ 1676) of febrile infants in the low-risk group (PCT concentrations < cut-off value); the incidence of SBIs in
686
the pooled cohort was 20.9% (484/2317) (Table 1). The difference in incidence of SBIs between the pooled data for the clinical prediction rule studies and the PCT studies, independent of the strength of the tests, may account for the apparent differences in positive and negative predictive values. Using clinical prediction rules as the discriminator and only prospective studies, Huppler et al. calculated overall RR for SBIs in high-risk febrile infants of 30.6 (95% CI 7.0–68.13) and 8.75 (95% CI 2.29–15.2) for nonrecipients and recipients of antimicrobial therapy, respectively (42). In the present study, the RR for SBIs among high-risk febrile infants differentiated by serum PCT concentrations was 3.97 (95% CI 3.41–4.62). These data suggest that using a serum PCT cut-off value of 0.3 ng/mL to identify those infants at risk for SBIs is inferior to the validated clinical prediction rules currently used for the evaluation of febrile infants < 91 days of age. Limitations Selection bias in systematic reviews and meta-analyses may arise from two sources. Failure to recover all of the data during the literature search process is always a concern. Search strategies are never perfect, so completeness is always an open question with these studies. Medline was the only database searched for this study; it is possible relevant publications that appeared in other databases may have been missed. Publication bias contributes to the selection issue because there is risk that wellperformed studies evaluating the utility of PCT serum concentrations as a predictor of SBI in infants failed to pass peer-review muster. The results from the Egger test argue against this problem in the present study. Combinability of studies is another potential source of bias. The included studies employed different assays for measuring serum concentrations of PCT. de Wolf et al. demonstrated that the Roche assay and the BRAHMS Kryptor kit yielded comparable results, whereas Schuch et al. demonstrated comparability between the two BRAHMS PCT assays (43,44). Together these studies suggest that the PCT results are comparable. The lack of heterogeneity among the studies used in this review further support this notion. Variability in the outcome definitions is another source of bias in this study. Overall, the studies were in agreement with regard to definitions of blood, cerebrospinal fluid, and gastrointestinal infections. Whereas there was some variability in the definition of urinary tract infection, the use of 104 CFU/mL was uniform. Conversely, only two studies included bone infections; four of seven studies included pulmonary infections (34,37,38,40). Finally, the conclusions of this study regarding the effectiveness of serum PCT vs. the clinical prediction
J. T. England et al.
rule for the identification of febrile infants with serious bacterial infections are based on the results of two meta-analyses. A head-to-head, prospective comparison study is required to resolve this issue. CONCLUSION The data from this review suggest that serum PCT concentrations # 0.3 ng/mL identify a population of febrile infants < 91 days of age at low risk for SBIs. Febrile infants with a serum concentration in excess of 0.3 ng/mL have, on average, a 3.97-fold increased risk of SBIs. Recent data suggest that infants who fail to meet standardized, low-risk, clinical criteria for SBI have, at best, an average increased risk of 8.75-fold. These differences are largely the result of the broad range of falsenegative results noted in Table 1 (2.1% to 17%). Together these observations suggest that serum PCT concentrations alone are a poorer predictor of SBI. These observations do not support the ‘‘stand alone’’ use of serum PCT concentrations as a screen for SBIs in febrile infants < 91 days of age. Serum PCT concentrations may be used in combination with a clinical prediction rule. Future studies should incorporate serum PCT into existing clinical prediction rules to determine if it adds to the effectiveness of the rules. REFERENCES 1. Bracq S, Machason M. Calcitonin gene expression in normal human liver. FEBS Lett 1993;331:14–8. 2. Assicot M, Gendrel D, Carsin H, Raymond J, Guilbaud J, Bohoun C. High serum procalcitonin concentrations in patients with sepsis and infection. Lancet 1993;341:515–8. 3. Redl H, Schalg G, Togel E, Assicot M, Bohoun C. Procalcitonin release patterns in a baboon model of trauma and sepsis: relationship to cytokines and neopterin. Crit Care Med 2000;28:3659–63. 4. Mesiner M. Pathobiochemistry and clinical use of procalcitoin. Clin Chem Acta 2002;323:17–29. 5. Morganthaler N, Struck J, Chancerelle Y, et al. Production of procalcitonin (PCT) in non-thyroidal tissue after LPS injection. Horm Metab Res 2003;35:290–5. 6. Nijsten M, Olinga P, The T, et al. Procalcitonin behaves as a fast responding acute phase protein in vivo and in vitro. Crit Care Med 2000;28:458–61. 7. Nylen E, Whang K, Snider R, Steinwald P, White J, Becker K. Mortality is increased by procalcitonin and decreased by an antiserum reactive to procalcitonin in experimental sepsis. Crit Care Med 1998;26:1001–6. 8. Liappis A, Gibbs K, Nylen E, et al. Exogenous procalcitonin evokes a pro-inflammatory cytokine response. Inflamm Res 2011; 60:203–7. 9. Galleto-Lacour A, Zamora S, Grervaix A. Bedside procalcitonin and C-reactive protein tests in children with fever without localizing signs of infection seen in a referral center. Pediatrics 2003;112: 1054–60. 10. Hatherill M, Tibby S, Sykes K, Turner C, Murdoch I. Diagnostic markers of infection: comparison of procalcitonin with C-reactive protein and leukocyte count. Arch Dis Child 1999;81:417–21. 11. Yu Z, Liu J, Sun Q, Han S, Guo X. The accuracy of the procalcitonin test for the diagnosis of neonatal sepsis: a meta-analysis. Scand J Infect Dis 2010;42:723–33.
Procalcitonin in Febrile Infants 12. Lin S, Hou T, Huang D. Role of procalcitonin in the diagnosis of severe infection in pediatric patients with fever and neutropenia. Pediatr Infect Dis J 2012;31:e182–8. 13. Simon L, Gauvin F, Amre D, Saint-Louis P, Lacroix J. Serum procalcitonin and C-reactive protein levels as markers of bacterial infection: a systematic review and meta-analysis. Clin Infect Dis 2005;40:1386–8. 14. Shen C, Wu M, Lin K, et al. The use of procalcitonin in the diagnosis of bone and joint infection: a systematic review and metaanalysis. Eur J Clin Microbiol Infect Dis 2013;32:807–14. 15. Roberts K, Borzy M. Fever in the first eight weeks of life. Johns Hopkins Med J 1977;141:9–13. 16. Baraff L, Oslund S, Schriger D, Stephen M. Probability of bacterial infections in febrile infants less than three months of age: a metaanalysis. Pediatr Infect Dis J 1992;11:257–64. 17. Roberts K. Management of young, febrile infants. Primum non nocere revisited. Am J Dis Child 1983;137:1143–4. 18. Dagan R, Powell K, Hall C, Menegus M. Identification of infants unlikely to have serious bacterial infection although hospitalized for suspected sepsis. J Pediatr 1985;107:855–60. 19. Jaskiewicz J, McCarthy C, Richardson A, et al. Febrile infants at low risk for serious bacterial infection-an appraisal of the Rochester criteria and implications for management. Febrile infant collaborative study group. Pediatrics 1994;94:390–6. 20. Baraff L, Bass J, Fleisher G, et al. Practice guideline for the managment of infants and children 0 to 36 months of age with fever without source. Pediatrics 1993;92:1–12. 21. PRISMA: Transparent reporting of systematic reviews and meta-analyses. The PRISMA Checklist. 2009. Available at: http://www. prisma-statement.org/statement.htm. Accessed November 27, 2013. 22. Cochrane W. The combination of estimates from different experiments. Biometrics 1954;10:101–29. 23. Higgins J, Thompson S, Deeks J, Altman D. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. 24. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: 719–48. 25. Ibrahim K, Abdel-Wahab A, Ibrahim A. Diagnostic value of serum procalcitonin levels in children with meningitis: comparison with blood leukocyte count and C-reative protein. J Pak Med Assoc 2011;61:346–51. 26. Paize F, Carrol W, Downey C, et al. Diagnostic efficacy of activated partial thromboplastin time waveform and procalcitonin analysis in pediatric meningococcal sepsis. Pediatr Crit Care Med 2011;12: e322–9. 27. Leroy S, Romanello C, Galletto-Lacour A, et al. Procalcitonin is a predictor for high-grade vesicoureteral reflux in children: metaanalysis of individual patient data. J Pediatr 2011;159:644–51. 28. Lacroix J, Manzano S, Galetto A, Gervaix A. Procalcitonin measurement for detection of serious bacterial infection in febrile children: comaprison between two automated immunoassays. Clin Biochem 2012;45:593–5. 29. Unal S, Arslankoylu A, Kuyucu N, Aslan G, Erdogan S. Procalcitonin is more useful than C-reactive protein in differentiatin of fever in patients with sickle cell disease. J Pediatr Hematol Oncol 2012;34: 85–9.
687 30. Bohnhorst B, Lange M, Bartels D, Bejo L, Hoy L, Peter C. Procalcitoniin and valuable clinical symptoms in the early detection of neonatal late-onset bacterial infection. Acta Paediatr 2012;101: 19–25. 31. Bressan S, Gomez B, Mintegi S, et al. Diagnsotic performance of the lab-score in predicting severe and invasive bacterial infections in well-appearing young febrile infants. Pediatr Infect Dis J 2012; 31:139–44. 32. Fernandez-Lopez A, Luaces-Cubells C, Garcia-Garcia J-J, Fernandez Pou J, Spanish Society of Pediatric Emergencies. Procalcitonin in pediatric emergency departments for the early diagnosis of invasive bacterial infections in febrile infants: results of a multicenter study and utility of a rapid qualitative test for this marker. Pediatr Infect Dis J 2003;22:895–903. 33. Srinivasan L, Harris M. New technologies for the rapid diagnosis of neonatal sepsis. Curr Opin Pediatr 2012;24:165–71. 34. Andreola B, Bressan S, Callegaro S, Liverani A, Plebani M, Da Dalt L. Procalcitonin and C-reactive protein as diagnostic markers of severe bacterial infections in febrile infants and children in the emergency department. Pediatr Infect Dis J 2007;26: 672–7. 35. Gomez B, Bressan S, Mintegi S, et al. Diagnostic value of procalcitonin in well-appearing young febrile infants. Pediatrics 2012;130: 815–22. 36. Luaces-Cubells C, Mintegi S, Garcia-Garcia J-J, et al. Procalcitonin to detect invasive bacterial infection in non-toxic-appearing infants with fever without apparent source in the emergency room. Pediatr Infect Dis J 2012;31:645–7. 37. Maniaci V, Dauber A, Weiss S, Nylen E, Becker K, Bachur R. Procalcitonin in young febrile infants for the detection of serious bacterial infections. Pediatrics 2008;122:701–10. 38. Manzano S, Bailey B, Girodias J-B, Galetto-Lacour A, Cousineau J, Delvin E. Impact of procalcitonin on the mangement of children aged 1 to 36 months presenting with fever without source: a randomized controlled trial. Am J Emerg Med 2010; 28:647–53. 39. Woelker J, Sinha M, Christopher N, Powell K. Serum procalcitonin concentration in the evaluation of febrile infants 2 to 60 days of age. Pediatr Emerg Care 2012;28:410–5. 40. Olacirgui I, Hernandez U, Munoz J, Emparanza J, Landa J. Markers that predict serious bacterial infection in infants under 3 months of age presenting with fever of unknown origin. Arch Dis Child 2009; 94:501–5. 41. Levy MM, Fink MP, Marshall JC, et al. International Sepsis Definitions Conference. Crit Care Med 2003;31:1250–6. 42. Huppler A, Eickoff J, Wald E. Performance of low-risk criteria in the evalaution of young infants with fever: review of the literature. Pediatrics 2010;125:228–33. 43. de Wolf HK, Gunnewiek JK, Berk Y, van den Ouweland J, de Metz M. Comparison of a new procalictonini assay from Roche with the established method of the Brahms Kryptor. Clin Chem 2009;55:1043–4. 44. Schuch G, Duc-Marchand C, Venet C, Mann H, Tixler A, Bionda C. Use of procalcitonin in intensive care units: comparison of semi quantitative PCT-Q Brahms assay with autmated PCT-Kryptor assay. Anal Biol Clin (Paris) 2011;69:663–70.
688
J. T. England et al.
ARTICLE SUMMARY 1. Why is this topic important? The evaluation of the febrile infant is a common and vexing problem in the care of pediatric patients. 2. What does this study attempt to show? The use of procalcitonin in the evaluation of febrile infants is being used more and more commonly. We have attempted to assess the usefulness of this test in this clinical arena. 3. What are the key findings? The use of procalcitonin for the evaluation of the febrile infant, though of some use, must be used with caution and is inferior to clinical prediction rules already in use. 4. How is patient care impacted? Clinicians have a clearer understanding of the usefulness of procalcitonin in the assessment of the febrile infant. With this understanding they can more accurately apply this test to the clinical situation and enhance their clinical management of this population. Additionally, the present study may be used to guide further research.
