Elevated Plasma and Cerebrospinal Fluid Interleukin-1 Beta and Tumor Necrosis Factor-Alpha Concentration and Combined Outcome Of

Home , Hypertonia

ORIGINAL ARTICLE Elevated plasma and cerebrospinal ﬂuid interleukin-1 beta and tumor necrosis factor-alpha concentration and combined outcome of death or abnormal neuroimaging in preterm neonates with early-onset clinical sepsis

S Basu1, P Agarwal1, S Anupurba2, R Shukla3 and A Kumar1

OBJECTIVE: Prematurity and sepsis are the major contributors of neonatal mortality and neurodevelopmental sequelae. The present study was conducted to measure the plasma and cerebrospinal fluid (CSF) concentration of interleukin (IL)-1β and tumor necrotic factor (TNF)-α in preterm neonates with early-onset clinical sepsis (EOCS), and to find out their association with combined outcome of death or abnormal neuroimaging. STUDY DESIGN: Thirty-two preterm (⩽34 weeks) neonates with EOCS and 32 gestational age-matched, healthy neonates served as cases and controls, respectively. Samples were collected soon after birth. Neonates were followed up clinically and by serial cranial ultrasonography (CUS) until discharge and subsequently by magnetic resonance imaging (MRI) of brain until 1 year. Developmental screening was done by Denver Developmental Screening test-II. RESULT: In EOCS group, no neonate had any clinical/microbiological evidence of meningitis. Blood culture was positive in 17 (53%). CUS was abnormal in 12 (37%) (intracranial hemorrhage-11, periventricular leukomalacia-1). Ten (31%) neonates expired. Significant elevation of plasma and CSF IL-1β and TNF-α was observed in the EOCS group. On follow-up, seven (22%) neonates showed evidence of white matter damage in MRI, two of them had developmental delay and microcephaly. Plasma and CSF IL-1β and TNF- α concentration were significantly elevated in deceased neonates and those with abnormal neuroimaging. Both biomarkers demonstrated high predictive accuracy for poor outcome in receiver-operating curve analysis. CONCLUSION: Elevation of plasma and CSF IL-1β and TNF-α is associated with an increase in the combined outcome of death or abnormal neuroimaging in preterm neonates with EOCS in the absence of clinical/microbiological evidence of meningitis with high predictive accuracy. Journal of Perinatology (2015) 35, 855–861; doi:10.1038/jp.2015.86; published online 30 July 2015

INTRODUCTION insults that can be picked up as early predictors of neuronal Preterm birth, with an incidence of 7.6 to 12% in developed damage are intracranial hemorrhage and periventricular leuko- countries to 415% in developing countries, is one of the major malacia (PVL). Though cranial ultrasonography (CUS) is a clinically contributors of neonatal mortality and long-term neurodevelop- attractive screening tool to detect neuronal damage because of its mental sequelae.1,2 Chorioamnionitis, the forerunner of early- easy availability and ability to study sick infants at bedside, it has limitations in accurate prediction of parenchymal brain injury in onset neonatal sepsis after birth, is the most common cause of 12 preterm delivery, accounting for 30% of preterm births.3 Recent very preterm infants. Until date, magnetic resonance imaging (MRI) with high contrast resolution remains one of the best studies have shown that perinatal infection and inflammation are diagnostic modality, which allows the identification of subtle the potential risk factors for cerebral white matter injury (WMI) 4–7 changes in brain morphology of preterm infants in chorioamnio- and cerebral palsy. Infection-induced maternal immune activa- nitis mediated damage.13 fl tion leads to fetal in ammatory response syndrome, which has a Previous authors have tried to correlate WMI in MRI with plasma 8 fl key role in fetal neurological damage. Cytokines and in amma- and cerebrospinal fluid (CSF) cytokine concentrations in neonates tory cells are known to be the mediators in the final common with proven or high risk of sepsis14 and histologic chorioamnio- 9 pathway associated with perinatal brain injury. Among the nitis,15 but no significant correlation was documented. Until date, proinflammatory cytokines, tumor necrotic factor-α (TNF-α) and no study has followed up neonates who were born to mothers interleukin-1β (IL-1β) have been documented to be the early and with clinical chorioamnionitis and subsequently developed clinical primary executioners of neuroinflammation, both in animal features suggestive of early-onset neonatal sepsis. In the present models8–10 and human neuroprogenitor cells.11 study, we have measured plasma and CSF TNF-α and IL-1β Early radiological prediction of adverse neurodevelopmental concentration in neonates with early-onset clinical sepsis (EOCS) outcome in premature neonates is a challenging task. Cerebral born to mothers with clinical chorioamnionitis with a hypothesis

1Division of Neonatology, Department of Pediatrics, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India; 2Department of Microbiology, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India and 3Department of Radiodiagnosis, Institute of Medical Sciences, Banaras Hindu University, Varanasi, India. Correspondence: Dr S Basu, Divison of Neonatology, Department of Pediatrics, Institute of Medical Sciences, Banaras Hindu University, Varanasi 221005, India. E-mail: [email protected] Received 9 April 2015; revised 30 May 2015; accepted 16 June 2015; published online 30 July 2015 Interleukin-1 beta and tumor necrosis factor-alpha in neonatal sepsis S Basu et al 856 that concentration of these biomarkers may have an association Work-up in controls with combined outcome of death or abnormal neuroimaging. In controls, blood samples for the estimation of IL-1β and TNF-α were collected at the time of routine sampling for thyroid proﬁle (which is the part of our hospital protocol) before discharge. Control CSF samples were METHODS obtained from those neonates who were admitted in the unit for clinically This prospective observational study was conducted in Sir Sunderlal suspected sepsis but had a negative sepsis work-up in blood and CSF and Hospital, Institute of Medical Sciences, Banaras Hindu University, Varanasi, showed spontaneous clinical improvement without any antibiotics. None India over a period of 18 months. It is a regional, tertiary-care teaching of them had any maternal risk factors for chorioamnionitis. Blood and CSF hospital that receives antenatally referred cases from adjoining states. The sepsis work-up was not done in those controls who never had any clinical study was approved by the Institute Ethics Committee. Informed consents feature suggestive of sepsis. CUS was done at the age of 48 to 72 h and in local language were taken from all the parents before inclusion in repeated at 4 weeks in follow-up. MRI was not done in any of the control the study. neonates.

Study participants Laboratory estimation of IL-1β and TNF-α The cases comprised of singleton preterm (⩽34 weeks of gestation) For the estimation of IL-1β and TNF-α, plasma was separated immediately inborn neonates delivered to mothers with clinical chorioamnionitis, from blood by centrifugation. Both plasma and CSF samples were stored at who subsequently developed clinical features suggestive of early-onset − 20 °C until analyzed. IL-1β and TNF-α were estimated by Human IL-1β neonatal sepsis and grouped as EOCS. Clinical chorioamnionitis was and TNF-α ELISA kits as per manufacturer’s instructions. The kits for defined as prolonged rupture of membrane 418 h and intrapartum research GENPRO ELISA kit, CAT#850006096, LOT#101B29, France was used maternal fever (437.8 °C in the period from onset of labor to delivery) for the estimation of IL-1β and GENPRO ELISA kit, CAT#950090096, with two or more of the following: fetal tachycardia, uterine tenderness, LOT#110084, France was used for the estimation of TNF-α. malodorous vaginal discharge or maternal leucocytosis (415 000 leukocytes μl − 1).16 EOCS was diagnosed on the basis of clinical signs and symptoms after excluding other common conditions, which may simulate CUS and MRI examination or co-exist with sepsis, such as respiratory distress syndrome, patent All CUS and MRI images were interpreted by a single qualified radiologist ductus arteriosus, intracranial hemorrhage, metabolic disorders and so on. to avoid any inter-observer variation. The radiologist was blinded to the Infants with antenatal evidence of fetal distress, perinatal asphyxia and infant’s clinical details and group. Infants were swaddled and oral dextrose hypoxic ischemic encephalopathy, intrauterine infections (toxoplasma, was used as a pacifier, if necessary. CUS was done using Siemens and rubella, cytomegalovirus, herpesvirus, syphilis, infective hepatitis, human Xario, Toshiba corporation (Minato-ku, Tokyo, Japan) equipment in quiet immunodeficiency virus), other systemic diseases and congenital mal- neonates without any pressure provocation. A 7.5 MHz curved electronic formations were excluded. Equal number of gestational age-matched array transducer was used to detect early changes. Germinal matrix- healthy neonates without history of clinical chorioamnionitis formed the intraventricular hemorrhage was graded as per Papile, et al.19 control group. For MRI, 1.5 Tesla MRI machine, Magnetom Avanto (Version BV-I7A) Siemens medical system, Erlangen, Germany was used. WMI on MRI was classified as mild, moderate and severe using scoring sheet as used by Work-up in EOCS group Inder, et al.20 Detailed antenatal and perinatal history were recorded and thorough clinical examination was made at birth. Birth weight and gestational age were noted. Enrolled subjects were followed up for the development of Sample size calculation clinical signs of sepsis until 72 h. As there was no similar publication in literature, the present study was Peripheral venous blood samples were collected soon after birth and conducted as a pilot trial with a convenient sample size of 64 (32 cases and sent for blood culture, sepsis screen and estimation of IL-1β and TNF-α. For 32 controls). blood culture, a 10-ml culture bottle containing sheep brain heart infusion ’ broth was inoculated with 1 ml of infant s blood under full aseptic Statistical analysis precautions. Bacterial growth was noted after 48 h and 7 days. Drug sensitivity was determined whenever bacterial growth was noted. Sepsis The statistical program SPSS version 16.0 (SPSS Inc, Chicago, IL, USA) was screen included total leukocyte count, absolute neutrophil count, serum used for data entry and analysis. The data were presented as mean ± s.d. C-reactive protein and micro-erythrocyte sedimentation rate. Cutoff values and median (range) for continuous variables and as frequency and for positive sepsis screen were defined as total leukocyte count percentage for categorical variables. Comparison between two groups was 9 − 1 9 − 1 – ⩽ 5×10 l , absolute neutrophil count ⩽ 1.8 × 10 l , C-reactive protein done using independent samples t-test/Mann Whitney U-test for con- 2 ⩾ 10 mg l − 1 and micro-erythrocyte sedimentation rate ⩾ 10 mm in 1st tinuous variables and χ -test for categorical data. Analysis of variance and hour. Lumbar puncture was done within 24 h of birth and CSF samples post hoc Bonferroni test were used to compare variables among multiple were sent for cell count, estimation of protein and sugar, culture and groups. Pearson correlation coefficient was calculated to detect the estimation of IL-1β and TNF-α concentration. correlation between different parameters. Sensitivity and specificity were Neonates were managed as per our unit protocol. At admission, calculated at different selected cutoff values. Receiver-operating char- intravenous ampicillin and amikacin were started empirically. Antibiotics acteristic curve analyses with measurement of area under the curve were were modified after blood culture reports or on clinical deterioration, if performed to identify the appropriate cutoff values. All the tests were two- necessary. Investigations, such as X-rays, tests for biochemical and tailed. A P-value of o0.05 was considered statistically significant. hematological parameters, arterial blood gas analysis, echocardiography with color Doppler and so on were done as and when required. Respiratory distress syndrome was excluded by chest X-ray, and patent ductus RESULTS arteriosus was excluded by echocardiography. During the study period, a total of 3284 deliveries took place in Progress during hospital stay and outcome were noted. Screening for our University hospital out of which a total of 32 neonates, retinopathy of prematurity (ROP) was carried out initially after 4 weeks of delivered to mothers with clinical chorioamnionitis, who subse- birth and subsequently repeated every 2 weeks or earlier, if indicated. Staging was done according to The International Classification of quently developed EOCS were included as cases. The recruitment Retinopathy of Prematurity.17 Serial CUS was carried out at 48 to 72 h of participants into the study is shown in Figure 1. after delivery, 7 days and at the term equivalent age (post-conceptional age of 40 weeks). Neonates were followed up clinically and neurologically Baseline parameters until the corrected age of 12 months. Developmental screening was done by Denver Developmental Screening test-II in four spheres of develop- Baseline parameters of cases and controls were comparable ment, namely, gross motor, fine motor-adaptive, personal-social and (Table 1). There was no difference in mean birth weight, gesta- language.18 MRI of brain was done at term equivalent age (40 weeks) and tional age, gender, 1 and 5 min Apgar score and receipt of was repeated at the corrected age of 12 months. two doses of antenatal betamethasone between the two groups.

Journal of Perinatology (2015), 855 – 861 © 2015 Nature America, Inc. Interleukin-1 beta and tumor necrosis factor-alpha in neonatal sepsis S Basu et al 857 None of the study neonates required delivery room resuscitation. and controls) had any intracranial hemorrhage at the time of ﬁrst Mean duration of prolonged rupture of membrane in the EOCS CUS at 48 to 72 h. The mean time of blood sample collection was group was 37.5 h (range 27 h to 7 days). Fourteen (43.8%) mothers 6.2 ± 3.8 h in EOCS group and 25.8 ± 8.7 h in controls. The mean received at least one dose of intravenous ceftriaxone 4 h prior to time of CSF collection was 8.5 ± 4.6 h in EOCS group and delivery, as a part of our hospital protocol for intrapartum 10.2 ± 3.1 h in controls. antibiotic prophylaxis in prolonged rupture of membrane cases. No neonate received any postnatal steroid, and none (both cases Clinical and laboratory details The most common clinical manifestation of EOCS was respiratory distress (30; 94%), only two neonates presented with lethargy and feeding intolerance. None had any clinical feature suggestive of meningitis. Twelve neonates (37%) in the EOCS group and none in the control group had positive sepsis screen. Microscopic examination of CSF showed a total cell count of ⩾ 30 μl − 1 in six neonates (18.8%). CSF biochemistry showed elevated protein (4100 mg dl − 1) in four neonates (12.5%). Blood culture was positive in 17 (53%) of cases, CSF was sterile in all. The most commonly grown pathogen in blood was Klebsiella pneumoniae (n = 5), followed by Pseudomonas aeruginosa (n = 4). Other pathogens grown were Staphylococcus aureus (n = 3), Citrobacter freundii (n = 2), Acinetobacter baumanii (n =1) Escherichia coli (n = 1) and Candida albicans (n = 1). During the study period, 10 neonates (31%) expired in the EOCS group. Autopsy could not be done in any of them because of lack of consent.

Plasma and CSF IL-1β and TNF-α concentration The plasma and CSF IL-1β and TNF-α concentration were significantly elevated in the EOCS group compared with controls (Table 2). The increase in TNF-α concentration was more marked than IL-1β (eightfold vs 21-fold in plasma and 25-fold vs 38-fold in CSF). In the EOCS group, the relative rise in CSF concentration was more marked in IL-1β than TNF-α (mean CSF-plasma ratio was 3.2 for IL-1β and 1.8 for TNF-α). In controls, plasma and CSF concentration was similar for both the cytokines. Pearson bivariate correlation showed signification correlation between plasma and CSF IL-1β and TNF-α concentration (Pearson coefficient 0.928 for IL-1β and 0.935 for TNF-α; Po0.001). Within the EOCS group, there was no difference in plasma and CSF IL-1β and TNF-α between the preterm neonates with positive sepsis screen Figure 1. Flow chart showing the recruitment of participants into including blood culture and those with clinically suspected sepsis the study. (P40.05).

Table 1. Comparison of the baseline characteristics of the study population

Parameter Early-onset clinical sepsis group (n = 32) Controls (n = 32) P-value

Maternal age (years) (mean ± s.d.) 22.4 ± 2.5 23.0 ± 1.7 0.253a (NS) Antenatal care taken, n (%) 18 (56.3) 20 (62.5) 0.799b (NS) Gravida, median (range) 1 (1–2) 1 (1–2) 1.000b (NS) Parity, median (range) 0 (0–1) 0 (0–1) 1.000b (NS) Receipt of 2 doses of antenatal betamethasone, n (%) 12 (37.5) 13 (40.6) 1.000b (NS)

Mode of delivery SVD, n (%) 21 (65.6) 23 (71.9) 787b (NS) Cesarean section, n (%) 11 (33.4) 9 (29.1)

Presentation Vertex, n (%) 28 (87.5) 29 (90.6) 1.000b (NS) Breech, n (%) 4 (12.5) 3 (9.4)

Birth weight (g) (mean ± s.d.) 1304 ± 182 1337 ± 280 0.573a (NS) Median (IQR) 1295 (1183–1455) 1288 (1128–1591) Gestational age (weeks) (mean ± s.d.) 31.3 ± 1.5 31.6 ± 2 0.418a (NS) Median (IQR) 31 (30–32) 32 (31–33) Male : female 1 : 1 1.2 : 1 0.802b (NS) Apgar score, median (range) 8 (7–9) 9 (7–10) 1.000b (NS) Abbreviations: IQR, interquartile range; NS, not signiﬁcant; SVD, spontaneous vaginal delivery. aIndependent samples t-test. bχ2-test.

Table 2. Comparison of plasma and CSF IL-1β and TNF-α between early-onset clinical sepsis group and controls, mean ± s.d. (IQR)

Parameter Plasma IL-1β (pg ml − 1) Plasma TNF-α (pg ml − 1) CSF IL-1β (pg ml − 1) CSF TNF-α (pg ml − 1) (n = 32) (n = 32) (n = 32) (n =8)

Early-onset clinical sepsis group 32.50 ± 16.04 147.07 ± 107.52 103.41 ± 74.05 268.24 ± 196.67 (16.85–47.25) (76.7–218.65) (48.0–134.40) (115.15–476.50) Controls 4.43 ± 1.41 6.62 ± 1.70 4.14 ± 1.06 6.95 ± 2.10 (3.20–5.55) (4.90–8.00) (3.27–4.60) (5.05–9.35) Mann–Whitney U-test o0.001 o0.001 o0.001 o0.001 (P-value) Abbreviations: CSF, cerebrospinal ﬂuid; IL, interleukin; IQR, interquartile range; TNF, tumor necrotic factor.

Table 3. Comparison of plasma and IL-1β and TNF-α between 8 (80%) had IVH (grade I/II, n = 5 and grade III/IV, n = 3). MRI was neonates with intact survival and those who deceased or had done at 40 weeks in all the 22 neonates who survived. Seven fi abnormal MRI neonates in the early-onset clinical sepsis group (32%) of them showed evidence of PVL, ve had mild PVL characterized by low signal intensity in the perirolandic cortex by Parameter Plasma Mann–Whitney apparent diffusion coefficient maps and hyperintensity in the U-test periventricular white matter and perirolandic cortex and two neonates had moderate PVL characterized by T1 hypointensity Neonates Deceased/ P-value and T2 hyperintensity in the periventricular white matter along with intact abnormal with ex vaccuo dilatation of ventricles in axial T1 and T2-weighted survival (n = 15) MRI (n = 17) MR images. Plasma and CSF IL-1β and TNF-α concentration were − fi IL-1β (pg ml 1) signi cantly higher in the deceased neonates and those who had Mean ± s.d. 18.38 ± 3.8 50.04 ± 24.75 o0.001 delayed development and/or abnormal MRI, compared with those (IQR) (15.6–20.8) (37.25–58.00) with intact survival (Table 3). There was no difference in receipt of antibiotics, caffeine, vasopressors and duration of parenteral − 1 TNF-α (pg ml ) nutrition and mechanical ventilation in those with or without ± ± ± o Mean s.d. 67.64 19.15 228.92 132.73 0.001 PVL in MRI. (IQR) (48.20–84.8) (125.3–303.5) On neurological screening, two neonates with moderate PVL Cerebrospinal fluid had developmental delay, hypertonia and microcephaly. Repeat IL-1β (pg ml − 1) MRI in one showed wavy outline of dilated lateral ventricles along Mean ± s.d. 41.68 ± 7.68 121.56 ± 75.11 o0.001 with bilateral periventricular white matter hyperintensities. The (IQR) (35.45–46.50) (62.40–136.5) other had cystic encephalomalacic changes in bilateral temporal and occipital lobes with features of volume loss in coronal fast TNF-α (pg ml − 1) Mean ± s.d. 85.18 ± 18.44 298.68 ± 182.80 o0.001 Spin Echo-T2-weighted images. Rest of the neonates, including (IQR) (67.45–101.40) (131.40–497.00) those with grade I/II intracranial hemorrhage in CUS and mild PVL in MRI had normal Denver Developmental Screening test-II. All the fl Abbreviations: CSF, cerebrospinal uid; IL, interleukin; IQR, interquartile controls had normal CUS at 4 weeks and normal neurodevelop- range; MRI, magnetic resonance imaging; TNF, tumor necrotic factor. ment in follow-up.

Other outcomes Out of 22 neonates in the EOCS group who survived, ROP was β Table 4. Receiver-operating-curve analysis for plasma and CSF IL-1 detected in 8 (36.4%). Three neonates had stage I ROP, one had α and TNF- with immediate poor outcome (death or abnormal stage III and two each had stage IV and V. None of the study neuroimaging) neonates developed chronic lung disease or necrotizing enter- Parameter Area under Cutoff Sensitivity Speciﬁcity ocolitis. There was no difference in outcome between two genders the curve value and receipt of antenatal antibiotic prophylaxis (P40.05).

Plasma IL-1β (pg ml − 1) 0.914 ⩾ 31.25 100% 81.8% CSF IL-1β (pg ml − 1) 1.000 ⩾ 98.95 100% 100% Receiver-operating curve analysis Plasma TNF-α (pg ml − 1) 1.000 ⩾ 166.30 100% 100% Receiver-operating curve analysis for plasma and CSF IL-1β and CSF TNF-α (pg ml − 1) 1.000 ⩾ 217.90 100% 100% TNF-α showed high predictive accuracy for poor outcome (death Abbreviations: AUC, area under curve; CSF, cerebrospinal ﬂuid; IL, and/or abnormal neuroimaging) (Table 4 and Figure 2). interleukin; TNF, tumor necrotic factor. DISCUSSION In the present study, we observed a signiﬁcant increase in plasma as well as CSF concentration of IL-1β and TNF-α in preterm, very CUS and MRI brain low birth weight neonates with EOCS, born to mothers with Among 32 cases with EOCS, CUS was abnormal in 12 (37%). Eleven clinical chorioamnionitis. Increase in cytokine concentration was had germinal matrix-intraventricular hemorrhage (grade I/II highly correlated with combined outcome of death or abnormal in eight and grade III/IV in three) and one had increased neuroimaging in the absence of clinical and microbiological echogenicity suggestive of PVL. Out of 10 neonates who expired, evidence of meningitis. The concentration of both the cytokines

Figure 2. Receiver-operating characteristic (ROC) curve for plasma and CSF IL-1β and TNF-α with measurement of area under the curve (AUC).

was significantly higher in CSF than plasma, but, the increase in routes, such as intracervical, intrauterine or maternal LPS CSF concentration was more marked for IL-1β than TNF-α. administration, are characterized by reduction in pre-OLs or The principal cellular target of cerebral WMI in premature myelin markers without macroscopic lesions being evident. Both infants is premyelinating oligodendrocytes (pre-OLs). The major antenatal and neonatal LPS exposures were found to increase the mechanisms of damage to pre-OLs are hypoxia–ischemia and sensitivity of the brain to subsequent hypoxic/ischemic events.24 systemic infection/inflammation. It is difficult to differentiate In our study, the proportional increase in CSF concentration was between hypoxia–ischemia and infection/inflammation, as the more marked in IL-1β than TNF-α. Several authors have reported initiating cerebral insult since the final mechanisms of cellular preferential CNS upregulation of IL-1β after infection.25–27 Girard injury converge into three major events (namely, microglial et al28 have analyzed the IL-1β system using immunohistochem- activation, excitotoxicity and free radical attack) in both the istry to characterize the expression of IL-1 receptors, IL-1R conditions.21 Sometimes, hypoxia–ischemia and infection/inflam- antagonist and induction of downstream effectors in human mation interact with each other to potentiate the cerebral injury brains with white matter damage. The authors have found an even further. In the present study, we have excluded neonates IL-1β overexpression in association with IL-1 receptors immuno- with antenatal fetal distress, perinatal asphyxia and hypoxic reactivity in areas with WMI. Immunolabeling was detected on ischemic encephalopathy, so that we are able to observe the astrocytes and microglia/macrophages. Cai et al. have suggested a extent of neural damage caused by chorioamnionitis and EOCS. differential role of IL-1β and TNF-α in mediating LPS-induced brain A series of animal studies have demonstrated that pre-OLs are injury. Whereas IL-1β has an important direct role in mediating the target cells for lipopolysaccharide (LPS)-induced brain injury, LPS-induced brain injury, TNF-α has a complicated, probably dual, and inflammatory cytokines are the possible mediators.22 Cai effects in LPS-induced brain injury.29 TNF-α has been found to et al.23 have demonstrated an increased expression of IL-1β and elicit autocrine and paracrine effects on surrounding cells,30 TNF-α messenger RNA in the fetal brain along with an increase in leading to further cytokine and chemokine production. Besides glial fibrillary acidic protein-positive astrocytes, injuries to pre-OLs, inflammation, TNF-α treatment can also trigger apoptosis through hypomyelination along with a decrease in myelin basic protein its receptor and intracellular signaling.31,32 and alteration of immunoreactivity of microglia in the brain of In the present study, the mothers belonged to poor socio- offspring after maternal LPS administration. Wang et al.24 have economic status, receipt of antenatal care was poor and suggested that fetal inflammation is the strongest predictor of intrapartum antibiotic prophylaxis was inadequate in most of brain lesions. The authors have shown that cerebral effects them. All these factors probably resulted in the development of following fetal LPS exposure intravenously or via more distant EOCS with neurological involvement and mortality in a significant

© 2015 Nature America, Inc. Journal of Perinatology (2015), 855 – 861 Interleukin-1 beta and tumor necrosis factor-alpha in neonatal sepsis S Basu et al 860 number. We recorded a high percentage of cases developing 4 Leviton A, Dammann O. Coagulation, inflammation, and the risk of neonatal white germinal matrix-intraventricular hemorrhage and PVL, opinions matter damage. Pediatr Res 2004; 55: 541–545. differ regarding occurrence of PVL and germinal matrix-intraven- 5 Stoll BJ, Hansen NI, Adams-Chapman I, Fanaroff AA, Hintz SR, Vohr B et al. National tricular hemorrhage in chorioamnionitis and early-onset neonatal institute of child health and human development neonatal research network sepsis. Although majority are of the opinion that WMI and PVL neurodevelopmental and growth impairment among extremely low-birth weight infants with neonatal infection. JAMA 2004; 292: 2357–2365. are commonly associated with chorioamnionitis in preterm 33–35 6 Adams-Chapman I, Stoll BJ. Neonatal infection and long-term neurodevelop- infants, Chau et al. have opined that histopathological mental outcome in the preterm infant. Curr Opin Infect Dis 2006; 19:290–297. chorioamnionitis is not associated with an increased risk of WMI 7 Dammann O, Kuban KC, Leviton A. Perinatal infection, fetal inflammatory on MRI. In contrast, postnatal infections and hypotension are response, white matter damage, and cognitive limitations in children born pre- associated with an increased risk of WMI in the premature term. Ment Retard Dev Disabil Res Rev 2002; 8:46–50. newborn.36 Several studies have found an association between 8 Burd I, Bentz AI, Chai J, Gonzalez J, Monnerie H, Le Roux PD et al. Inflammation- 37–39 induced preterm birth alters neuronal morphology in the mouse fetal brain. J chorioamnionitis and neonatal IVH, though others were 88 – fi 35,40 Neurosci Res 2010; : 1872 1881. unable to con rm the association. Some studies even reported 9 McAdams RM, Juul SE. The role of cytokines and inflammatory cells in perinatal the occurrence of both PVL and intraventricular hemorrhage in brain injury. Neurol Res Int 2012; 2012: 561494. 41,42 preterm infants with chorioamnionitis. 10 Sheng WS, Hu S, Ni HT, Rowen TN, Lokensgard JR, Peterson PK. TNF-alpha- We have taken samples soon after birth to document early induced chemokine production and apoptosis in human neural precursor cells. J changes in the concentrations of the biomarkers and to avoid Leukoc Biol 2005; 78: 1233–1241. changes related to management and advancement of postnatal 11 Pugazhenthi S, Zhang Y, Bouchard R, Mahaffey G. Induction of an inflammatory age. Thirty out of 32 of our neonates were receiving supplemental loop by interleukin-1β and tumor necrosis factor-α involves NF-kB and STAT-1 in 8 oxygen (25 to 40%) by head box or nasal prongs at the time of differentiated human neuroprogenitor Cells. PLoS One 2013; : e69585. 12 van Wezel-Meijler G, Steggerda SJ, Leijser LM. Cranial ultrasonography in neo- blood sampling. A recent study has postulated the role of non- 34 – fl nates: role and limitations. Semin Perinatol 2010; :28 38. infectious causes of systemic in ammation, such as, supplemental 13 Hatfield T, Wing DA, Buss C, Head K, Muftuler LT, Davis EP. Magnetic resonance oxygen exposure in preterm infants, which may contribute to imaging demonstrates long-term changes in brain structure in children born preterm 43 organ injury. and exposed to chorioamnionitis. Am J Obstet Gynecol 2011; 205: 384 e1–384 e8. The major limitation of our study was small sample size. 14 Ellison VJ, Mocatta TJ, Winterbourn CC, Darlow BA, Volpe JJ, Inder TE. The rela- Number of control CSF was only eight, as doing CSF in healthy tionship of CSF and plasma cytokine levels to cerebral white matter injury in the neonates only for research purpose would have been unethical. premature newborn. Pediatr Res 2005; 57:282–286. 15 Reiman M, Kujari H, Maunu J, Parkkola R, Rikalainen H, Lapinleimu H et al. PIPARI We could not document histopathological evidence of chorioam- fl fi Study Group. Does placental in ammation relate to brain lesions and volume in nionitis. Moreover, because of nancial constraints we could not preterm infants? J Pediatr 2008; 152: 642–647. repeat the estimation of biomarkers to document any serial 16 Alexander JM, Gilstrap LC, Cox SM, McIntire DM, Leveno KJ. Clinical chor- change in their concentration over time, which could have been ioamnionitis and the prognosis for very low birth weight infants. Obstet Gynecol more informative. At last, long-term developmental follow-up until 1998; 91: 725–729. the age of 18 to 24 months was not possible. 17 Prepared by an International Committee. An international classification of reti- A high incidence of ROP (36.4%) was observed in the EOCS nopathy of prematurity. Br J Ophthalmol 1984; 68:690–697. group. Previous authors have also demonstrated a higher rate of 18 Frankenburg WK, Dodds J, Archer P, Shapiro H, Bresnick B. The Denver II: a major revision and restandardization of the Denver Developmental Screening Test. ROP in infants born to mothers with histological and clinical 89 – 44,45 fl Pediatrics 1992; :91 97. chorioamnionitis. Fetal in ammatory response syndrome 19 Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of sub- induced by chorioamnionitis might be the factor responsible to ependymal and intraventricular hemorrhage: a study of infants with birth weights increase the risk of ROP by directly sensitizing the developing less than 1,500 gm. J Pediatr 1978; 92: 529–534. retina to oxygen-induced changes in vascular endothelial growth 20 Inder TE, Wells SJ, Mogridge N, Spencer C, Volpe JJ. Defining the nature of the factor availability and subsequent vascular development.46 cerebral abnormalities in the premature infant: a qualitative magnetic resonance To conclude, elevation of plasma and CSF IL-1β and TNF-α imaging study. J Pediatr 2003; 143: 171–179. concentration are associated with increase in the combined 21 Volpe JJ, Kinney HC, Jensen FE, Rosenberg PA. The developing oligodendrocyte: key cellular target in brain injury in the premature infant. Int J Dev Neurosci 2011; outcome of death or abnormal neuroimaging in preterm neonates 29:423–440. with EOCS even in the absence of clinical and microbiological 22 Pang Y, Cai Z, Rhodes PG. Disturbance of oligodendrocyte development, hypo- evidence of meningitis. Both IL-1β and TNF-α may be utilized as myelination and white matter injury in the neonatal rat brain after intracerebral early predictors of death or abnormal neuroimaging before the injection of lipopolysaccharide. Brain Res Dev Brain Res 2003; 140:205–214. appearance of CUS and MRI abnormalities. As the present study 23 Cai Z, Pan ZL, Pang Y, Evans OB, Rhodes PG. Cytokine induction in fetal rat brains was limited by small sample size, larger studies may be carried out and brain injury in neonatal rats after maternal lipopolysaccharide administration. to validate these findings. Pediatr Res 2000; 47:64–72. 24 Wang X, Rousset CI, Hagberg H, Mallard C. Lipopolysaccharide-induced inflammation and perinatal brain injury. Semin Fetal Neonatal Med 2006; 11:343–353. CONFLICT OF INTEREST 25 Scripter JL, Ko J, Kow K, Arimura A, Ide CF. Regulation by interleukin-1beta of formation of a line of delimiting astrocytes following prenatal trauma to the brain fl The authors declare no con ict of interest. of the mouse. Exp Neurol 1997; 145:329–341. 26 Chauvet N, Palin K, Verrier D, Poole S, Dantzer R, Lestage J. Rat microglial cells secrete predominantly the precursor of interleukin-1beta in response to lipopo- REFERENCES lysaccharide. Eur J Neurosci 2001; 14: 609–617. 1 Blencowe H, Cousens S, Oestergaard MZ, Chou D, Moller AB, Narwal R et al. 27 Girard S, Kadhim H, Larouche A, Roy M, Gobeil F, Sébire G. Pro-inflammatory National, regional, and worldwide estimates of preterm birth rates in the year disequilibrium of the IL-1 beta/IL-1ra ratio in an experimental model of perinatal 2010 with time trends since 1990 for selected countries: a systematic analysis and brain damages induced by lipopolysaccharide and hypoxia-ischemia. Cytokine implications. Lancet 2012; 379:2162–2172. 2008; 43:54–62. 2 Chang HH, Larson J, Blencowe H, Spong CY, Howson CP, Cairns-Smith S et al. 28 Girard S, Sébire G, Kadhim H. Proinflammatory orientation of the interleukin 1 Preventing preterm births: analysis of trends and potential reductions with system and downstream induction of matrix metalloproteinase 9 in the patho- interventions in 39 countries with very high human development index. Lancet physiology of human perinatal white matter damage. Neuropathol Exp Neurol 2013; 381: 223–234. 2010; 69: 1116–1129. 3 Elovitz MA, Brown AG, Breen K, Anton L, Maubert M, Burd I. Intrauterine inflam- 29 Cai Z, Pang Y, Lin S, Rhodes PG. Differential roles of tumor necrosis factor-alpha mation, insufficient to induce parturition, still evokes fetal and neonatal and interleukin-1 beta in lipopolysaccharide-induced brain injury in the brain injury. Int J Dev Neurosci 2011; 29:663–671. neonatal rat. Brain Res 2003; 975:37–47.

Journal of Perinatology (2015), 855 – 861 © 2015 Nature America, Inc. Interleukin-1 beta and tumor necrosis factor-alpha in neonatal sepsis S Basu et al 861 30 Hanisch UK. Microglia as a source and target of cytokines. Glia 2002; 40:140–155. hemorrhage in infants born before 32 weeks' gestation. Am J Perinatol 1995; 12: 31 Muppidi JR, Tschopp J, Siegel RM. Life and death decisions: secondary complexes 429–436. and lipid rafts in TNF receptor family signal transduction. Immunity 2004; 21: 40 Kaukola T, Herva R, Perhomaa M, Pääkkö E, Kingsmore S, Vainionpää L et al. 461–465. Population cohort associating chorioamnionitis, cord inflammatory cytokines and 32 Varfolomeev EE, Ashkenazi A. Tumor necrosis factor: an apoptosis JuNKie? Cell neurologic outcome in very preterm, extremely low birth weight infants. Pediatr 2004; 116: 491–497. Res 2006; 59:478–483. 33 Perlman JM, Risser R, Broyles RS. Bilateral cystic periventricular leukomalacia in 41 Sarkar S, Kaplan C, Wiswell TE, Spitzer AR. Histological chorioamnionitis and the the premature infant: associated risk factors. Pediatrics 1996; 97:822–827. risk of early intraventricular hemorrhage in infants bornoor = 28 weeks gesta- 34 Wu YW, Colford JM Jr. Chorioamnionitis as a risk factor for cerebral palsy: a tion. J Perinatol 2005; 25: 749–752. meta-analysis. JAMA 2000; 284: 1417–1424. 42 Verma U, Tejani N, Klein S, Reale MR, Beneck D, Figueroa R et al. Obstetric 35 Graham EM, Holcroft CJ, Rai KK, Donohue PK, Allen MC. Neonatal cerebral white antecedents of intraventricular hemorrhage and periventricular leukomalacia in matter injury in preterm infants is associated with culture positive infections and the low-birth-weight neonate. Am J Obstet Gynecol 1997; 176: 275–281. only rarely with metabolic acidosis. Am J Obstet Gynecol 2004; 191: 1305–1310. 43 Chang BA, Huang Q, Quan J, Ladd M, Kwan E, McFadden DE et al. Early inflam- 36 Rocha G, Proença E, Quintas C, Rodrigues T, Guimarães H. Chorioamnionitis and mation in the absence of overt infection in preterm neonates exposed to brain damage in the preterm newborn. J Matern Fetal Neonatal Med 2007; 20: intensive care. Cytokine 2011; 56:621–626. 745–749. 44 Moscuzza F, Belcari F, Nardini V, Bartoli A, Domenici C, Cuttano A et al. Correlation 37 Chau V, Poskitt KJ, McFadden DE, Bowen-Roberts T, Synnes A, Brant R et al. Effect between placental histopathology and fetal/neonatal outcome: chorioamnionitis of chorioamnionitis on brain development and injury in premature newborns. and funisitis are associated to intraventricular haemorrage and retinopathy of Ann Neurol 2009; 66:155–164. prematurity in preterm newborns. Gynecol Endocrinol 2011; 27: 319–323. 38 Vergani P, Patanè L, Doria P, Borroni C, Cappellini A, Pezzullo JC et al. Risk factors 45 Polam S, Koons A, Anwar M, Shen-Schwarz S, Hegyi T. Effect of chorioamnionitis for neonatal intraventricular haemorrhage in spontaneous prematurity at on neurodevelopmental outcome in preterm infants. Arch Pediatr Adolesc Med 32 weeks gestation or less. Placenta 2000; 21:402–407. 2005; 159: 1032–1035. 39 Salafia CM, Minior VK, Rosenkrantz TS, Pezzullo JC, Popek E, Cusick W et al. Maternal, 46 Lee J, Dammann O. Perinatal infection, inflammation, and retinopathy of pre- placental, and neonatal associations with early germinal matrix intraventricular maturity. Semin Fetal Neonatal Med 2012; 17:26–29.