Inflammatory markers in cord blood or maternal serum for ... - Nature

Journal of Perinatology (2014) 34, 268–274 & 2014 Nature America, Inc. All rights reserved 0743-8346/14 www.nature.com/jp

ORIGINAL ARTICLE

Inflammatory markers in cord blood or maternal serum for early detection of neonatal sepsis—a systemic review and meta-analysis H Su1, S-S Chang2,3, C-M Han1, K-Y Wu1, M-C Li4, C-Y Huang1, C-L Lee1, J-Y Wu5,6 and C-C Lee6,7 OBJECTIVE: To perform a quantitative review of the evidence on the diagnostic value of inflammatory markers in maternal serum or umbilical cord blood for the diagnosis of early-onset neonatal sepsis (EONS). STUDY DESIGN: We searched multiple databases for studies published through March 2013 that evaluated the diagnostic performance of procalcitonin (PCT), C-reactive protein (CRP) and interleukin-6 (IL-6), and leukocyte count (white blood cell, WBC) in either umbilical cord blood or maternal serum for diagnosis of EONS. We summarized test performance characteristics with the use of forest plots, hierarchical summary receiver operating characteristic curves and bivariate random effects models. RESULT: Our search identified 3874 citations, of which 15 studies evaluating 2178 episodes of suspected neonatal infection were included for analysis. IL-6 in cord blood with a pooled-positive likelihood ratio (LR þ ) of 9.47 (95% confidence interval: 3.86 to 23.3), PCT in cord blood with a LR þ of 5.72 (1.56 to 21.0) and IL-6 in maternal serum with a LR þ of 5.47 (2.10 to 14.2) can be qualified as a valid rule-in test. IL-6 in cord blood with a LR of 0.10 (0.05 to 0.21) and PCT in cord blood with a LR of 0.20 (0.12–0.37) can be qualified as a useful rule-out test. Either CRP or WBC was inadequate for diagnosis of EONS. CONCLUSION: For cord blood sample, IL-6 or PCT can be used as reliable rule-in and rule-out tool. For maternal serum, only IL-6 appeared to be sufficient for rule-in diagnosis. An interventional study may be needed to answer whether the addition of these tests will improve the outcome of patients with EONS. Journal of Perinatology (2014) 34, 268–274; doi:10.1038/jp.2013.186; published online 23 January 2014 Keywords: early-onset neonatal sepsis; procalcitonin; C-reactive protein; interleukin-6

INTRODUCTION Early-onset neonatal sepsis (EONS) is one of the most common causes of morbidity and mortality in neonates.1–3 During pregnancy, bacteria may be transmitted to the developing placenta by ascending infections or through the hematogenous route. Cytotrophoblasts may invade the uterine vasculature and release maternal blood into the intervillous space and lead to EONS. Other than the intrauterine route, neonates may also become infected during delivery. EONS is a systemic infection that occurs in neonates within the first week of life, with the majority appearing within 24 h of birth. Infants of mothers with Group B Streptococcus infection or chorioamnionitis, and who have prolonged rupture of membranes are at higher risk for developing EONS. Early diagnosis remains a major diagnostic problem because clinical signs are nonspecific.4 Routine laboratory tests such as white blood cell (WBC) count, absolute neutrophil count, immature/total neutrophil ratio and C-reactive protein (CRP) levels in the neonatal serum have low sensitivity.5,6 The ‘Gold Standard’, blood cultures are often negative because of low blood volumes drawn, and single cultures as well as the prenatal administration of antibiotics.7 Repeat blood sampling also may

cause deterioration and possible anemia in the premature newborns. In addition, certain biomarkers, such as procalcitonin (PCT) and CRP, show a physiological increase in the first few days of life, which complicates the interpretation of results.6,8,9 Cord blood or maternal serum specimens can obviate the aforementioned problems associated with neonatal blood samples.10,11 However, the accuracy of the inflammatory markers in the cord blood or maternal serum in predicting EONS remains to be validated. The aim of this study is to evaluate the usefulness of the concentrations of the inflammatory markers in umbilical cord blood or maternal serum samples in the detection of intrauterine infections in newborns. Specifically, we seek to quantitatively summarize the current evidence on the measurement of CRP, PCT, interleukin-6 (IL-6), as well as leukocyte count in the cord blood and maternal serum and compare their accuracy for the early diagnosis of EONS.

METHODS We adhered to the methods and procedures of the Cochrane Collaboration and the Preferred Reporting Items for Systematic Reviews and Meta Analyses guidelines for reporting systematic reviews.12

1 Department of Obstetrics and Gynecology Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Taoyuan, Taiwan; 2Department of Family Medicine, Chang Gung Memorial Hospital, Taoyuan, and Chang Gung University College of Medicine, Taoyuan, Taiwan; 3Graduate Institute of Clinical Medical Sciences, College of Medicine, Chang Gung University, Taoyuan, Taiwan; 4Institute of Occupational Medicine and Industrial Hygiene, College of Public Health, National Taiwan University, Taipei, Taiwan; 5 Department of Emergency Medicine, Chang Gung Memorial Hospital, Keelung, and Chang Gung University College of Medicine, Taoyuan, Taiwan; 6Department of Emergency Medicine, National Taiwan University Hospital, Taipei, Taiwan and 7Department of Emergency Medicine, National Taiwan University Hospital Yunlin Branch, Douliou, Taiwan. Correspondence: Dr J-Y Wu, Department of Emergency Medicine, Chang Gung Memorial Hospital, Keelung, and Chang Gung University College of Medicine, 3F, No.3, Ln. 325, Sec. 2, Shipai Road, Beitou District, Taoyuan 11267, Taiwan. E-mail: [email protected] or C-C Lee, Department of Emergency Medicine, National Taiwan University Hospital Yunlin Branch, No. 579 Sec 2 Yunlin Road, Douliou, Taiwan. E-mail: [email protected] Received 24 August 2013; revised 18 November 2013; accepted 20 November 2013; published online 23 January 2014

Inflammatory markers for diagnosis of EONS H Su et al

269 Eligible studies

Statistical analysis and data synthesis

Eligible studies were primary studies that assessed the diagnostic accuracy of PCT, CRP, IL-6 and leukocyte count in the cord blood or maternal serum for the diagnosis of neonatal sepsis. Studies have to adhere to the following outcome definition of EONS to be eligible for inclusion.

The sensitivity and specificity and their 95% confidence intervals (CIs) were plotted in forest plots and summary receiver operating characteristic (ROC) curves. We used the bivariate model for diagnostic meta-analysis to obtain weighted overall estimates of the sensitivity and specificity.14 The bivariate approach adjusts for the negative correlation between the sensitivity and specificity of the index test that may arise from different thresholds used in different studies. The bivariate model also allows a DerSimonian-Laird random effects approach for both sensitivity and specificity. The random effects model accounts for both within-study variability (random error) and between-study variability (heterogeneity).15 Diagnostic accuracy studies are expected to show considerable heterogeneity, and the models used are by default random effects models. We also performed diagnostic odds ratio (DOR) meta-analysis. The DOR combines both positive and negative likelihood ratios (LRs) and is a global measure of test performance. To investigate the source of heterogeneity, we calculated the I2 statistics.16 Statistically significant heterogeneity was considered present at I2 450%. Statistical analyses were conducted using STATA 11.0 (Stata Corp, College Station, TX, USA), notably the midas and metandi commands. All statistical tests were two-sided, and statistical significance was defined as a P–valueo0.05.

Definitions A worldwide agreement on the definition of EONS has not been reached. In the included studies, EONS is generally defined as a premature (o37 week of gestation) or term infants (X37 week of gestation) with new-onset systemic infection based on clinical findings (clinically documented infection) or microbiological findings (microbiologically documented infection) within 7 days of birth. Clinical findings usually include at least three of the following clinical manifestations: (1) abnormal skin color (pallor, jaundice and cyanosis); (2) respiratory distress (apnea, tachypnea with respiratory rate 460 per min, grunting, nasal flaring, intercoastal or sternal retractions and need for high ventilator); (3) cardiovascular function abnormality (bradycardia or tachycardia, poor capillary refill and hypotension); (4) neurological findings abnormality (hypotonia, hyporeflexia, irritability, lethargy and seizures); (5) gastrointestinal symptoms (abdominal distension, green or bloody residuals, vomiting and temperature instability), and/or elevated leukocyte count. Microbiological findings include positive blood cultures, or positive culture results from other sterile sites.

Search strategy Electronic databases were searched with the provided search terms. We performed a comprehensive search of the MEDLINE, EMBASE and Cochrane databases for pertinent studies published since inception up to March 2013. We used ‘PCT or CRP or IL-6 or leukocyte count’ as the search terms, without any language restriction. Then, we combined biomarkers with the search keywords representing pregnant women and neonatal populations. The complete search terms used were (PCT or CRP or IL-6 or biomarkers or inflammatory marker) and (neonatal infection or neonatal sepsis or chorioamnionitis or pregnancy or premature rupture of membrane or intrauterine infection). Reference lists of the selected studies, narrative and systematic reviews, and primary studies were also manually checked for other relevant studies. Two authors independently assessed all titles/abstracts to determine that inclusion criteria were satisfied. Full-text articles were retrieved if either of the reviewers considered the abstract potentially suitable. The two reviewing authors then independently assessed the full text of the retrieved studies for their suitability for inclusion. Disputes were resolved by having an additional reviewer (C-CL) to assess the full article, and thus consensus was reached regarding inclusion in the meta-analysis.

Data extraction The following information was extracted from the included studies: first author, country, patient characteristics, quantitative data required for construction of a 2 2 table, cutoff levels of biomarker test, quantitative or semi-quantitative nature of the test, study settings and definition of neonatal sepsis. In studies that reported multiple pairs of sensitivity and specificity, we consistently used the data with the highest Youden index (sensitivity þ specificity 1) and performed a sensitivity analysis at a later stage. We contacted the authors for additional data if the studies did not report sufficient data to construct a 2 2 contingency table. We excluded the study if we did not get any response from the contacted authors.

Evaluation of methodological quality Methodological quality was assessed using a standardized tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews (QUADAS).13 Differential verification bias was considered if only selected patients underwent a reference test for outcome verification and the selection process was nonrandom. High risk of incorporation bias was considered if interpretation of the reference test and the index test were not blinded to each other. Spectrum bias was considered if the study patients were not prospectively and consecutively enrolled. Other methodological quality indicators, including adequacy of reference standards, and whether withdrawal or uninterpretable results were reported were evaluated for each included study. & 2014 Nature America, Inc.

RESULTS Our initial searches retrieved 3874 studies. After screening title and abstract, 3829 studies were removed. Additional searches from reference lists retrieved another 11 studies. From these 56 studies, 39 were further excluded for several reasons: either they did not study or report outcomes of interest (n ¼ 32) or insufficient data was collected or provided for 2 2 tables (n ¼ 7) (Figure 1). Table 1 lists the characteristics of the 17 included studies in the review. Studies were performed between 1999 and 2012, and included 2197 episodes of suspected neonatal infection episodes.10,17–31 Of the 2197 episodes, 78 (17.2%) were documented as systemic infection. Most studies were conducted in Europe (n ¼ 14). Eleven studies collected maternal serum specimens, and seven studies used cord blood samples. Twelve studies used appropriate cohort design, whereas five studies used case-control design that is at risk for over-evaluation of the specificity of the test. CRP was evaluated in 13 studies, IL-6 in 8 studies, WBC in 5 studies and PCT in 5 studies. The markers were assessed because of the presence of signs of EONS. The cutoff values for each marker varied among the studies. We used the QUADAS tool for studying quality assessment. Figure 2 provides an overall impression of the methodological quality of the studies. The majority of studies satisfied QUADAS criteria, with the exception of the handling of withdrawals and uninterpretable results. None of the included studies explained withdrawals or reported uninterpretable results. All patients were verified by the same reference standard in all studies. Although none of the tests being analyzed in this work were used to form the diagnostic criteria, none of the studies reported whether physicians were blinded to the results of the index tests when verifying the case with the reference standard; therefore, we could not exclude the possibility of incorporation bias. Diagnostic accuracy indices The minimal number of studies required for meaningful statistical combination is three, so only part of the extracted data can derive a summary estimate. The eligible combinations are cord blood specimens for PCT and CRP, and maternal serum specimens for IL-6, CRP and leukocyte count. Pooled diagnostic accuracy indices are summarized in Table 2. We set a positive LR þ of 5 as the threshold to qualify a rule-in test. PCT in cord blood (LR þ : 5.72, 95% CI: 1.56 to 21.0), IL-6 in cord blood (LR þ : 9.47, 95% CI: 3.86 to 23.3) and IL-6 in maternal blood (LR þ : 5.47, 95% CI: 2.10 to 14.2) can be qualified as valid rule-in tests. We set a negative LR of 0.2 as the threshold to Journal of Perinatology (2014), 268 – 274

270

Summary of the characteristics of the included studies

Author, country, year

Prevalence (N)

Specimens

Study design

Biomarkers tested

Cutoff (PCT, ng ml 1; CRP, mg l 1 WBC, 103/mm3; IL-6, pg ml 1)

Pfeiffer KA, Germany,26

0.23(71)

Maternal serum

Lewis DF, USA,25

0.51(57)

Maternal plasma

CRP, WBC, IL-6 IL-6

0.38(32)

Cord blood

Prospective case control Prospective case control Prospective cohort

PCT, CRP

Janota J Czech,21 Kordek A, Poland,

23

0.17(187)

Cord blood

Prospective cohort

Kayem G, France,

0.19(73) 0.22(109)

Joram N, France,10

0.08(197)

Maternal serum or vaginal secretions Maternal serum Cord blood neonatal blood Cord blood

Prospective cohort

Hatzidaki E, Greece,20

Torbé A, Poland,28

0.35(48)


0.23(65) 0.16(286)

Cord blood

Cancelier AC, Brazil,

0.33(120)

Cord blood

van der Heyden JL, Netherlands,31 Torbé A, Poland,35

0.16(299) 0.28(50)

22

PCT, CRP WBC CRP, IL-6

PCT sensitivity, specificity

CRP sensitivity, specificity

WBC sensitivity, specificity

IL-6 sensitivity, specificity

CRP ¼ 1.2; WBC ¼ 12, IL-6 ¼ 11

NA

IL-6 ¼ 18

NA

56% 76% NA

63% 73% NA

PCT ¼ 2; CRP ¼ 10

IL-6 ¼ 108.5

NA

25% 90% 97% 22% 95% 16% NA

NA

CRP ¼ 2; IL-6 ¼ 100

75% 75% 81% 69% NA

81% 76% 75% 56% NA

87.5% 98.7% 53% 45% NA

50% 97% 52% 76% NA

80.4% 71.6% NA

73.9% 77.9% NA

NA

45% 80% 47%, 76%; 71%, 47% 60% 79% 94% 48% 38% 74% 65.2% 59.1%

55% 70% NA

a

PCT ¼ 1.2 ; CRP ¼ 2.5; WBC ¼ 12

Retrospective cohort

IL-6

Prospective cohort

PCT, CRP

PCT ¼ 0.5; CRP ¼ 5

Maternal serum

Prospective cohort Prospective cohort

PCT, CRP, WBC IL-6

PCT ¼ 1.9; CRP ¼ 10; WBC ¼ 15

Maternal serum

PCT, CRP

PCT ¼ 1.22; CRP ¼ 1

Maternal serum

Retrospective case control Prospective case control Prospective cohort Prospective case control Prospective cohort

CRP

0.08(128)

Maternal serum and vaginal fluid Cord blood

Popowski T, France,

0.04(399)

Maternal serum

Prospective cohort

CRP


0.26(31)

Maternal serum

Prospective cohort

CRP, WBC

CRP ¼ 10; WBC ¼ 15

NA

Gulati S, India,19

0.49(45)

Maternal serum

Prospective cohort

CRP, IL-6

CRP ¼ 10; IL6 ¼ 8

NA

Kordek A, Poland,23 17

Cernada M, Spain,

18 27

Author, country, year

IL-6 CRP, WBC

CRP, IL-6

IL-6 ¼ 40

NA NA

NA

CRP ¼ 2.4, 15

NA

CRP ¼ 1.02; IL-6 ¼ 255.8

NA

CRP ¼ 5

NA

42% 63% NA NA NA 48% 85% NA

NA

NA NA 13% 87% NA

NA 56% 35% 95%, 100%; 90%, 100%; 90%, 97.4% NA NA 27% 90% NA 90% 85% NA NA 90% 87.4% NA NA 82.6% 86.3%

Outcome definition

Outcome diagnostic criteria

Pfeiffer KA, Germany, Lewis DF, USA,25 Janota J Czech,21 Kordek A, Poland,23 Kayem G, France,22 Hatzidaki E, Greece,20 Joram N, France,10

Neonatal infection Neonatal infection Neonatal sepsis Early Neonatal Infection Neonatal infection Neonatal sepsis Early-onset neonatal infection

Torbé A, Poland,30 Torbé A, Poland,28 Kordek A, Poland,24 Cancelier AC, Brazil,17 van der Heyden JL, Netherlands,31 Torbé A, Poland,35 Cernada M, Spain,18 Popowski T, France,27 Torbé A, Poland,36 Gulati S, India,19

Congenital infection Congenital infection Neonatal infection Neonatal sepsis Neonatal sepsis Neonatal infection Neonatal infection Neonatal infection Neonatal infection Neonatal infection

MDI and CDI, positive blood culture, tachypnea, tachycardia, fever, poor feeding, CRP, leukocytosis MDI, positive cultures from either the blood or spinal fluid or congenital pneumonia CDI, clinical signs and laboratory tests CDI, based on clinical criteria alone CDI, clinical signs of neonatal sepsis and routine laboratory results MDI, blood culture alone MDI and CDI, clinical signs of neonatal sepsis, and a positive blood or cerebrospinal fluid bacteriological sample CDI, clinical signs and laboratory tests CDI, clinical signs and laboratory tests CDI, clinical signs of neonatal sepsis and routine laboratory results CDI, clinical signs of neonatal sepsis and routine laboratory results CDI, clinical signs of neonatal sepsis and routine laboratory results CDI, clinical signs of neonatal sepsis and routine laboratory results MDI and CDI, clinical findings plus positive blood culture MDI and CDI, clinical findings plus positive blood culture CDI, based on clinical findings and neonatal laboratory data MDI

26

& 2014 Nature America, Inc.

Abbreviations: CDI, clinically documented infection; CRP, C-reactive protein; IL-6, interleukin-6; MDI, microbiologically documented infection; NA, not available; PCT, procalcitonin; WBC, white blood cell. a Multiple cutoff values.


Journal of Perinatology (2014), 268 – 274

Table 1.


271 3874 citations identified from Medline, EMBASE, and Cochrane library

3829 citations exclude based on screening titles and abstracts

11 citations added based on search from reference lists of relevant articles

56 potentially relevant articles identified for further review

39 citations excluded after full-text review 32 did not study or report outcomes of interest 7 did not have sufficient data to reconstruct 2x2 tables

17 articles included in meta-analyses

Figure 1.

Flow chart of study identification and inclusion.

Figure 2. Results of quality assessment by QUADAs. The figure provides an overall impression of the methodological quality of the included studies.

quality a rule-out test. IL-6 and PCT in cord blood can be qualified as a valid rule-out test, and can be used to reduce the pretest probability to an acceptable level. Leukocyte count in maternal & 2014 Nature America, Inc.

serum has poor LR þ (1.97, 95% CI: 1.32 to 2.90) and LR (0.70, 95% CI: 0.53 to 0.93), and is not an adequate tool for either rule-in or rule-out diagnosis. Journal of Perinatology (2014), 268 – 274



0.454 0.056 58.3 (0.00–88.1) 80.7 (54.8–91.8) 117 (21.9–633) 16.6 (3.67–75.4) 0.96 (0.67–0.99) 0.88 (0.85–0.91) Abbreviations: CRP, C-reactive protein; PCT, procalcitonin; WBC, white blood cell; IL-6, interleukin-6.

0.10 (0.05–0.21) 0.28 (0.13–0.64) 9.47 (3.86–23.3) 5.47 (2.10–14.2) 0.90 (0.86–0.93) 0.86 (0.69–0.95) 3 5 IL6 Cord blood7,8,10 Maternal serum9,10,16,17,19

0.91 (0.82–0.96) 0.76 (0.53–0.90)

0.848 0.00 (0.0–84.69) 3.17 (1.91–5.25) 0.70 (0.66–0.74) 0.70 (0.53–0.93) 1.97 (1.33–2.90) 0.76 (0.66–0.84) 4 WBC Maternal serum17,19,22,23

0.47 (0.30–0.64)

0.734 0.408 29.5 (0.0–72.7) 0.00 (0.0–67.6) 10.1 (5.95–17.3) 3.49 (2.31–5.26) 0.82 (0.73–0.90) 0.72 (0.68–0.75) 0.49 (0.28–0.83) 0.61 (0.46–0.81) 3.40 (1.21–9.49) 2.22 (1.74–2.82) 0.71 (0.68–0.74) 0.76 (0.65–0.84) 5 8 CRP Cord blood8,11,12,14,15 Maternal serum9,13,17,18,19,21–23

0.71 (0.62–0.79) 0.53 (0.37–0.70)

79.7 (46.2–92.4) 21.5 (5.53–83.8) 0.85 (0.82–0.88) 0.20 (0.12–0.37) 5.72 (1.56–21.0) 0.86 (0.58–0.96) 0.82 (0.72–0.89) 4 PCT Cord blood11,12,14,15

I2 (95% CI) Diagnostic OR (95% CI) AUROC (95% CI) Likelihood ratio, Likelihood ratio, þ Specificity (95% CI) Sensitivity (95% CI) Studies (n) Variable

Diagnostic accuracy of PCT, CRP, WBC count and IL-6 in cord blood sample or maternal serum

Table 2.

DISCUSSION EONS is one of the major causes of neonatal morbidity and mortality, accounting for over one million neonatal deaths worldwide each year.1–3 Early diagnosis and treatment improves outcome. Owing to the immature immune system, signs of infection within 72 h of age are usually subtle and nonspecific. A serum marker may help clinicians reach early diagnosis of EONS, but several markers of infection have a physiological elevation in the first few days of life, making the interpretation of these markers complicated.11 Sampling umbilical cord blood for measurement of PCT before the physiological increase may make interpretation of PCT concentration easier and can raise its diagnostic value in this special population. We showed that PCT and IL-6 in cord blood can be used as a valid rule-in and rule-out test, whereas IL-6 in maternal blood can also be used as a useful rule-in test. We found that both CRP and leukocyte count in maternal serum are not adequate for either rule-in or rule-out diagnosis. Previous meta-analyses addressing the value of neonatal serum IL-6 showed that IL-6 has moderate value (AUC: 0.89, 95% CI: 0.84 to 0.94; LR þ : 4.55; 95% CI: 3.27 to 6.32; LR : 0.26; 95% CI: 0.18 to 0.36) in diagnosing EONS.32 Our meta-analysis showed comparable accuracy of IL-6 in maternal serum specimen but higher accuracy of IL-6 in cord blood specimen. Therefore, cord blood may be a more appropriate sample for measurement of IL-6. Previous meta-analyses addressing the value of PCT in both neonatal and cord blood specimens showed a summary sensitivity of 0.72 (95% CI: 0.63 to 0.81) and a summary specificity of 0.77 (95% CI: 0.72 to 0.81), neither sensitive nor specific enough for diagnosis of EONS.33 They did not provide accuracy estimate for PCT in cord blood sample. We find cord blood PCT has a superior sensitivity and specificity (sensitivity 0.82, 95% CI: 0.72 to 0.89; specificity 0.86, 95% CI: 0.58 to 0.96), and can provide important diagnostic information for clinical decision making. To provide more relevant information for clinicians, we calculated post-test probability for each biomarker-specimen pair.34 In a virtual population with a 20% prevalence (pretest probability) of EONS, a LR þ of 5.72 (cord blood PCT), 5.47 (maternal blood IL-6) and 9.47 (cord blood IL-6) translates into a positive predictive value (post-test probability) of 56, 56 and 70%. Similarly, a LR of 0.20 (cord blood PCT) and 0.10 (cord blood IL-6) translates into 5 and 2% post-test probability. Therefore, if local prevalence of EONS is low (that is, o10%), non-infective causes of respiratory distress or hypotension should be aggressively sought if either cord blood PCT or IL-6 test is negative. In addition, although EONS is a disorder with a high mortality rate if not treated promptly, a negative test with a post-test probability as low as o5% may justify to shorten the duration of empiric antibiotics in well-appearing neonates, given the potential risk of prolonged antibiotics treatment, such as direct antibiotic toxicity, necrotizing enterocolitis, late-onset sepsis and death in preterm neonates.33,35,36 Conversely, positive results for cord blood PCT or IL-6 test could be used for confirmation of bacterial infection to allow systemic antibiotics treatment and close hemodynamic monitoring.

Publication bias (Egger’s test p)

We also calculated two measures, area under the ROC (AUROC) curve and DOR for comparison of global accuracy between different biomarkers. IL-6 in cord blood has the highest AUROC curve (0.96, 95% CI: 0.67 to 0.99), followed by IL-6 in maternal serum (AUROC: 0.88, 95% CI: 0.85 to 0.91), PCT in cord blood (AUROC: 0.85, 95% CI: 0.82 to 0.88), CRP in maternal serum (AUROC: 0.72, 95% CI: 0.68 to 0.75) and leukocyte count in maternal serum (AUROC: 0.70, 95% CI: 0.66 to 0.74). The DOR, interpreted as the number of times of increased odds for EONS after a positive test, also followed a similar trend (Table 2). No significant evidence of potential publication bias was noted using the Egger’s test for funnel plot asymmetry (Table 2).

0.294

272



273 Limitations Despite the promising results, the results have to be interpreted in the context of high between-study heterogeneity. The marked differences in PCT or IL-6 measurement methods, timing of blood sampling, definition of neonatal sepsis and the use of appropriate controls have been reported as likely causes of heterogeneity, and all were present in the studies included. Of note, a wide range of cutoff values for each marker was used. To cope with this problem, we constructed the hierarchical summary ROC curve, and the AUROC curve can be interpreted independently of the threshold effect. The lack of standardization of current evidence may form major obstacles for cord blood PCT or IL-6 test being considered as part of diagnostic criteria of EONS. Future studies that validate measurement of PCT or IL-6 in cord blood for early diagnosis of EONS clearly need to distinguish between use of the test to screen for the presence or absence of EONS in high-risk patients in the absence of signs of illness and its use to confirm or exclude the disease when it becomes manifest. Another major limitations of this work is the limited number of studies in each biomarkerspecimen category, which prevents us from performing sensitivity analysis to verify the stability of the accuracy estimates across various EONS definitions and study designs. In addition, results of this study only show the usefulness of biomarkers in the diagnosis of EONS, but whether they have equal strength in prognosis, risk prediction and therapeutic guidance remains unanswered. This work also does not assess the diagnostic value of umbilical cord measurement of inflammatory biomarkers in detection of lateonset maternofetal infection related to intra-partum or post-natal infection. The work also does not provide answer to whether the biomarker information in maternal serum or umbilical cord provides additional information to that already provided by what is already known about the neonates, which has to be answered in a well-designed clinical trial evaluating relevant clinical outcomes.

7 8

9

10

11

12 13

14

15 16 17

18

19

CONCLUSION In conclusion, our study showed that PCT and IL-6 in the umbilical cord blood may provide important diagnostic information to help early identification of EONS. CRP and leukocyte count do not provide equally useful information for inclusion or exclusion diagnosis. For maternal serum, only IL-6 appeared to be sufficient for rule-in diagnosis. Clinicians should interpret the results based on their LR and pretest probability. The next step would be to assess whether provision of information from these biomarkers tests can modify the clinical outcomes.

20

21

22

23

CONFLICT OF INTEREST The authors declare no conflict of interest. 24

REFERENCES 1 Koenig JM, Keenan WJ. Group B streptococcus and early-onset sepsis in the era of maternal prophylaxis. Pediatr Clin North Am 2009; 56(3): 689–708. 2 Stoll BJ, Hansen NI, Sanchez PJ, Faix RG, Poindexter BB, Van Meurs KP et al. Early onset neonatal sepsis: the burden of group B Streptococcal and E. coli disease continues. Pediatrics 2011; 127(5): 817–826. 3 Lukacs SL, Schrag SJ. Clinical sepsis in neonates and young infants, United States, 1988-2006. J Pediatr 2012; 160(6): 960–965 e961. 4 Mishra UK, Jacobs SE, Doyle LW, Garland SM. Newer approaches to the diagnosis of early onset neonatal sepsis. Arch Dis Child Fetal Neonatal Ed 2006; 91(3): F208–F212. 5 Christensen RD, Baer VL, Gordon PV, Henry E, Whitaker C, Andres RL et al. Reference ranges for lymphocyte counts of neonates: associations between abnormal counts and outcomes. Pediatrics 2012; 129(5): e1165–e1172. 6 Chiesa C, Natale F, Pascone R, Osborn JF, Pacifico L, Bonci E et al. C reactive protein and procalcitonin: reference intervals for preterm and term new-


25

26

27

28

borns during the early neonatal period. Clin Chim Acta 2011; 412(11-12): 1053–1059. Buttery JP. Blood cultures in newborns and children: optimising an everyday test. Arch Dis Child Fetal Neonatal Ed 2002; 87(1): F25–F28. Chiesa C, Pellegrini G, Panero A, Osborn JF, Signore F, Assumma M et al. C-reactive protein, interleukin-6, and procalcitonin in the immediate postnatal period: influence of illness severity, risk status, antenatal and perinatal complications, and infection. Clin Chem 2003; 49(1): 60–68. Chiesa C, Osborn JF, Pacifico L, Natale F, De Curtis M. Gestational- and age-specific CRP reference intervals in the newborn. Clin Chim Acta 2011; 412(19-20): 1889–1890. Joram N, Boscher C, Denizot S, Loubersac V, Winer N, Roze JC et al. Umbilical cord blood procalcitonin and C reactive protein concentrations as markers for early diagnosis of very early onset neonatal infection. Arch Dis Child Fetal Neonatal Ed 2006; 91(1): F65–F66. Howman RA, Charles AK, Jacques A, Doherty DA, Simmer K, Strunk T et al. Inflammatory and haematological markers in the maternal, umbilical cord and infant circulation in histological chorioamnionitis. PloS ONE 2012; 7(12): e51836. Leeflang MM, Deeks JJ, Gatsonis C, Bossuyt PM. Systematic reviews of diagnostic test accuracy. Ann Int Med 2008; 149(12): 889–897. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003; 3: 25. Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005; 58(10): 982–990. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7(3): 177–188. Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat Med 2002; 21(11): 1525–1537. Cancelier AC, Petronilho F, Reinke A, Constantino L, Machado R, Ritter C et al. Inflammatory and oxidative parameters in cord blood as diagnostic of earlyonset neonatal sepsis: a case-control study. Pediatr Crit Care Med 2009; 10(4): 467–471. Cernada M, Badia N, Modesto V, Alonso R, Mejıás A, Golombek S et al. Cord blood interleukin-6 as a predictor of early-onset neonatal sepsis. Acta Paediatr 2012; 101(5): e203–e207. Gulati S, Bhatnagar S, Raghunandan C, Bhattacharjee J. Interleukin-6 as a predictor of subclinical chorioamnionitis in preterm premature rupture of membranes. Am J Reprod Immunol 2012; 67(3): 235–240. Hatzidaki E, Gourgiotis D, Manoura A, Korakaki E, Bossios A, Galanakis E et al. Interleukin-6 in preterm premature rupture of membranes as an indicator of neonatal outcome. Acta Obstet Gynecol Scand 2005; 84(7): 632–638. Janota J, Stranak Z, Belohlavkova S, Mudra K, Simak J. Postnatal increase of procalcitonin in premature newborns is enhanced by chorioamnionitis and neonatal sepsis. Eur J Clin Invest 2001; 31(11): 978–983. Kayem G, Goffinet F, Batteux F, Jarreau PH, Weill B, Cabrol D. Detection of interleukin-6 in vaginal secretions of women with preterm premature rupture of membranes and its association with neonatal infection: a rapid immunochromatographic test. Am J Obstet Gynecol 2005; 192(1): 140–145. Kordek A, Giedrys-Kalemba S, Pawlus B, Podraza W, Czajka R. Umbilical cord blood serum procalcitonin concentration in the diagnosis of early neonatal infection. J Perinatol 2003; 23(2): 148–153. Kordek A, Halasa M, Podraza W. Early detection of an early onset infection in the neonate based on measurements of procalcitonin and C-reactive protein concentrations in cord blood. Clin Chem 2008; 46(8): 1143–1148. Lewis DF, Barrilleaux PS, Wang Y, Adair CD, Baier J, Kruger T. Detection of interleukin-6 in maternal plasma predicts neonatal and infectious complications in preterm premature rupture of membranes. Am J Perinatol 2001; 18(7): 387–391. Pfeiffer KA, Reinsberg J, Rahmun A, Schmolling J, Krebs D. Clinical application of maternal serum cytokine determination in premature rupture of membranes-interleukin-6, an early predictor of neonatal infection? Acta Obstet Gynecol Scand 1999; 78(9): 774–778. Popowski T, Goffinet F, Maillard F, Schmitz T, Leroy S, Kayem G. Maternal markers for detecting early-onset neonatal infection and chorioamnionitis in cases of premature rupture of membranes at or after 34 weeks of gestation: a two-center prospective study. BMC Preg Childbirth 2011; 11: 26. Torbe A. Maternal plasma procalcitonin concentrations in pregnancy complicated by preterm premature rupture of membranes. Mediators Inflamm 2007; 2007: 35782.



274 29 Torbe A, Czajka R. Are vaginal fluid procalcitonin levels useful for the prediction of subclinial infection in patients with preterm premature rupture of membranes? J Obstet Gynaecol Res 2005; 31(5): 464–470. 30 Torbe A, Czajka R, Kordek A, Rzepka R, Kwiatkowski S, Rudnicki J. Value of vaginal fluid proinflammatory cytokines for the prediction of early-onset neonatal infection in preterm premature rupture of the membranes. J Interferon Cytokine Res 2007; 27(5): 393–398. 31 van der Heyden JL, van Teeffelen SS, Coolen AC, Halbertsma FJ, Aardenburg R, Mertens HJ et al. Is it useful to measure C-reactive protein and leukocytes in patients with prelabor rupture of membranes? Am J Perinatol 2010; 27(7): 543–547. 32 Shahkar L, Keshtkar A, Mirfazeli A, Ahani A, Roshandel G. The role of IL-6 for predicting neonatal sepsis: a systematic review and meta-analysis. Iran J Pediatr 2011; 21(4): 411–417.


33 Cotten CM, Taylor S, Stoll B, Goldberg RN, Hansen NI, Sańchez PJ et al. Prolonged duration of initial empirical antibiotic treatment is associated with increased rates of necrotizing enterocolitis and death for extremely low birth weight infants. Pediatrics 2009; 123(1): 58–66. 34 Halkin A, Reichman J, Schwaber M, Paltiel O, Brezis M. Likelihood ratios: getting diagnostic testing into perspective. QJM 1998; 91(4): 247–258. 35 Cordero L, Ayers LW. Duration of empiric antibiotics for suspected early-onset sepsis in extremely low birth weight infants. Infect Control Hosp Epidemiol 2003; 24(9): 662–666. 36 Kuppala VS, Meinzen-Derr J, Morrow AL, Schibler KR. Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants. J Pediatr 2011; 159(5): 720–725.
