Fetal and Neonatal Effects of N-Acetylcysteine When Used for Neuroprotection in Maternal Chorioamnionitis.

OBJECTIVE: To evaluate the clinical safety of antenatal and postnatal N-acetylcysteine (NAC) as a neuroprotective agent in maternal chorioamnionitis in a randomized, controlled, double-blinded trial. STUDY
DESIGN: Twenty-two mothers >24 weeks gestation presenting within 4 hours of diagnosis of clinical chorioamnionitis were randomized with their 24 infants to NAC or saline treatment. Antenatal NAC (100 mg/kg/dose) or saline was given intravenously every 6 hours until delivery. Postnatally, NAC (12.5-25 mg/kg/dose, n = 12) or saline (n = 12) was given every 12 hours for 5 doses. Doppler studies of fetal umbilical and fetal and infant cerebral blood flow, cranial ultrasounds, echocardiograms, cerebral oxygenation, electroencephalograms, and serum cytokines were evaluated before and after treatment, and 12, 24, and 48 hours after birth. Magnetic resonance spectroscopy and diffusion imaging were performed at term age equivalent. Development was followed for cerebral palsy or autism to 4 years of age.
RESULTS: Cardiovascular measures, cerebral blood flow velocity and vascular resistance, and cerebral oxygenation did not differ between treatment groups. Cerebrovascular coupling was disrupted in infants with chorioamnionitis treated with saline but preserved in infants treated with NAC, suggesting improved vascular regulation in the presence of neuroinflammation. Infants treated with NAC had higher serum anti-inflammatory interleukin-1 receptor antagonist and lower proinflammatory vascular endothelial growth factor over time vs controls. No adverse events related to NAC administration were noted.
CONCLUSIONS: In this cohort of newborns exposed to chorioamnionitis, antenatal and postnatal NAC was safe, preserved cerebrovascular regulation, and increased an anti-inflammatory neuroprotective protein. TRIAL REGISTRATION: ClinicalTrials.gov: NCT00724594.

RCT Entities:   Population Interventions Outcomes

Intrauterine infection is associated with significant white and gray matter brain injury in newborns and is particularly important in the pathogenesis of periventricular leukomalacia (PVL) and cerebral palsy.[1,2] Pathogens initiate toll-like receptor signaling in immune cells at the fetal-maternal interface, amplifying production of inflammatory cytokines in the placental membranes and amniotic fluid.[3] Subsequently, fetal inflammatory cells are activated in the cord (funisitis) and in the fetal circulation, and both cytokine storm and leukocyte and endothelial cell activation lead to production of reactive oxygen species, without true infection.[4] Endotoxin exposure, toll-like receptor signaling, leukocyte, and vascular endothelial activation all contribute to a fetal inflammatory state (FIRS) induced by chorioamnionitis. Neuroinflammation rapidly follows FIRS, within 1–4 hours in animal models.[5]

Cytokines and oxidative stress may cause direct central nervous system (CNS) injury.[6,7] Oxidative stress before delivery predisposes the fetus to subsequent hypoxic ischemic (HI) injury, even when the interruption of blood flow during delivery is mild.[8,9] Not surprisingly, the fetus frequently shows poor tolerance to labor with heart rate decelerations and bradycardia, which may further compromise cerebral perfusion and exacerbate white matter injury.

N-acetylcysteine (NAC) directly scavenges oxygen free radicals through its thiol-reducing group and is a promising neuroprotectant in animal models of chorioamnionitis.[10-13] NAC also crosses the blood brain barrier, decreases oxidative stress and cytokine production in the CNS, and replenishes glutathione, a major intracellular antioxidant.[14,15] In an animal model of chorioamnionitis, NAC provides optimal neuroprotection when given antenatally.[11] However, targeting the fetal brain through the maternal circulation requires consideration of placental metabolism and transfer, and pharmacokinetics (PK) in the fetus and mother.

Although NAC has a favorable safety profile in human infants and pregnant mothers with acetaminophen overdose,[16,17] translational studies must carefully consider potential adverse effects in this vulnerable population. In a fetal sheep model of septic shock, antenatal NAC administration increased fetal hypoxemia, raising safety concerns about its application in human chorioamnionitis.[18] To address safety and possible pharmacodynamic (PD) side effects of NAC, we conducted a randomized, controlled pilot trial in pregnant women ≥24 weeks gestation within 4 hours of clinical diagnosis of chorioamnionitis, and their infants postnatally. We have previously reported the PK of NAC administered to this cohort of mothers with chorioamnionitis and in their infants.[19] Intravenous NAC administered to the mother undergoes rapid placental transport, with slightly more than 1:1 transfer to the fetus, ascertained at delivery. In this report, NAC effects on serial cerebral blood flow (CBF) and oximetry, cardiac function, clotting studies, and blood pressure (BP) in mothers and infants exposed to chorioamnionitis, as well as serum cytokines and magnetic resonance spectroscopy (MRS) are reported with long-term outcomes.

Methods

This prospective, double blinded study was approved by the Medical University of South Carolina’s Institutional Review Board. Written informed consent was obtained prior to enrollment from pregnant women ≥24 completed weeks gestational age (GA), who presented with a clinical diagnosis of chorioamnionitis. Clinical chorioamnionitis was defined as maternal fever ≥38°C in the presence of rupture of membranes or 2 clinical findings: uterine tenderness, maternal white blood cell >15 000 cells/mm, fetal tachycardia >160 bpm, or malodorous amniotic fluid. Maternal exclusion criteria included bronchodilator or steroid-dependent asthma, sepsis, seizure disorder, suspected major abnormalities, fetal weight, or bi-parietal diameter less than the 10th % for GA, or immediate delivery.

Pregnant mothers were recruited in 2 cohorts based on GA at diagnosis of chorioamnionitis: term/late preterm (≥33 completed weeks) and preterm (24–32 completed weeks). GA cut-offs were chosen for fetal renal maturity and estimated NAC elimination.[17] After randomized to NAC or saline, NAC (100 mg/kg/dose) or an equivalent volume of saline was given intravenously to mothers over 60 minutes every 6 hours until delivery. NAC/saline (preterm 12.5 mg/kg/dose; term 25 mg/kg/dose) was administered to infants, starting 6 hours after last maternal dose, then every 12 hours for 5 doses.

Primary safety outcomes were the incidence of histaminergic reactions, clinical bleeding, hypotension, seizures, and increase in prothrombin time (PT). Secondary safety outcomes were mean BP (mBP), amount/duration of pressor support, amount of fresh frozen plasma, seizures, and anaphylaxis. Adverse events (AEs) were recorded for 2 days following NAC infusion, and serious AEs (SAEs) for 30 days, verified by an independent clinical research monitor (Table I). Potential AEs expected with NAC treatment are histaminergic reactions or bronchospasm in mothers, longer PT, and lower mBP in mothers and infants. Expected complications from chorioamnionitis are hypotension, cardiac dysfunction, intraventricular hemorrhage (IVH), seizures, disseminated intravascular coagulopathy (DIC), thrombocytopenia, renal failure, pulmonary hypertension, respiratory failure, multisystem organ failure, necrotizing enterocolitis (NEC), PVL, and death in infants.[20]

Table I

Demographic and clinical characteristics

	NAC	Control	Total
Entry strata
Preterm mothers: n	6	6	12
Preterm infants: n	7	7	14
GA at birth (wk)	28.2 ± 1.7	29.4 ± 3.3	NS
Birth weight (g)	1207 ± 235	1339 ± 462	NS
Term mothers: n	5	5	10
Term infants: n	5	5	10
GA at birth (wk)	38.6 ± 2.4	38.4 ± 2.8	NS
Birth weight (g)	3450 ± 415	3150 ± 582	NS
Sex infants
Male	7	6	(54%)
Female	5	6	(46%)
Race of mothers
African American	8	3	(50%)
Caucasian	3	8	(50%)
Labor and delivery
Mean duration rupture of membranes (h, range)	55 ± 83 (0–229)	94 ± 127 (0–322)	NS
Mode of delivery
Vaginal (70%)	8	9	NS
Cesarean (29%)	4	3	NS
Maternal dosing
Mean time first maternal dose to delivery (h, range) (excluding 2 NAC mothers who labored 17 and 32 h)	2.9 ± 2.2 (0–7)	1.8 ± 1.8 (0–5)	NS
Maternal heart rate change pre- and 1 h postdosing (bpm)	−1.8 ± 17	−2.5 ± 18	NS
Maternal mBP change pre and postdosing (mm Hg)	1 ± 12	2 ± 10	NS
Maternal hypotension	1	1
Fetal tachycardia first noted postdosing	1	0
Maternal mean PT (s) predose	13.7 ± 0.5	14.0 ± 0.9	NS
Maternal mean PT (s) at delivery	14.2 ± 0.6	14.1 ± 1.2	NS
Mean change PT (s) after dose	0.51 ± 0.5	0.45 ± 0.2	NS
Histopathology (missing 3)
Maternal inflammatory response, any grade	12/12 (100%)	8/9 (89%)
Fetal inflammatory response, any grade funisitis	9/12 (75%)	5/9 (56%)
Infant
Apgar at 1 min, median	6 ± 2.0	6.3 ± 2.7	NS
Apgar at 5 min, median	7.9 ± 1.8	7.5 ± 2.5	NS
Cord pH, mean	7.27 ± 0.08	7.22 ± 0.16	NS
WBC (Kcells/mm 3) at birth, mean	14.4 ± 5.0	8.1 ± 6.3	NS
Absolute neutrophil count (cells/mm3) at birth, mean	8.9 ± 6.2	10.8 ± 5.3	NS
Culture proven sepsis/pneumonia at birth	1	1
PT (s), mean DOL 1	19.5 ± 2.5	19.0 ± 4.0	NS
SAEs
Infant resuscitation-chest compressions/CV meds	0	1
Death	0	1
PVL/IVH at birth	1	0
IVH from 1–2 d of age	1	1
IVH from 5–7 d of age	2	0
NEC within 30 d	1	1

CV, cardiovascular; NS, not significant; WBC, white blood cells.

Maternal clotting studies were measured before NAC infusion and at delivery. Maternal blood samples for NAC serum concentration and cytokines were drawn immediately before and after NAC/saline infusion at 5, 15, and 30 minutes and 1, 2, 3, 4, and 6 hours, or until delivery. NAC PK were previously reported.[19] Placentas and umbilical cords were graded for chorioamnionitis and funisitis.[21]

All infants were evaluated for sepsis at birth, including complete blood count and blood culture, and treated with antibiotics by standardized clinical protocol. Clotting studies were drawn from the cord and at 48 hours. Serial blood samples for NAC and cytokine levels were drawn from the cord, before and at 15 minutes and 8, 12, 36, and 48 hours after the first infant dose of NAC/saline. Lumbar puncture was performed for cell count, culture, and cytokine levels before 12 hours of age, if the infant was stable without coagulopathy. Replicate samples of serum and cerebrospinal fluid (CSF) were prepared for cytokine analysis as previously described.[22] Electroencephalograms (EEGs) were recorded during the first dose of NAC/saline. Cranial ultrasounds (CUSs) for IVH or other abnormalities were performed after admission and at 48 hours (term infants) or at 7 days (preterm infants), and magnetic resonance imaging (MRI) with MRS and diffusion tensor imaging, at term age equivalent. Infants were followed in neonatal follow-up clinic at 12–24 months of age and at 3–4 years by the infants’ general and developmental pediatricians.

Doppler blood flow was measured before and after dosing in the umbilical and/or fetal middle cerebral artery (MCA) prior to delivery when possible, and postnatally in the anterior cerebral artery (ACA), MCA, and basilar artery (BA) before and after the first dose (0–6 hours of life [HOLs]), and at 12, 24, 36, and 48 hours. Near-infrared spectroscopy probe placed over the forehead recorded regional cerebral oxygenation (RcSO2) during and after NAC/saline dosing, at 0–6, 12, 24, 36, and 48 HOLs. Echocardiograms were performed before the first dose and at 12, 24, and 48 HOLs. NAC assay was previously reported, with maternal and infant NAC plasma concentrations and PK, determined by high performance liquid chromatography.[14,19] Intravenous NAC administration in the mother undergoes rapid placental transfer, and PK of NAC are different in preterm and term infants.[19]

Additional details for the Methods section are available in the Appendix (available at www.jpeds.com).

Statistical Analyses

Univariate analysis of variables between groups was performed using unpaired Student t test or Kruskal-Wallis tests. Repeated measures were analyzed by ANOVA or generalized linear mixed models as appropriate. Spearman or Pearson correlations were performed for analyses with NAC plasma concentrations. For all safety variables, significance was designated as P < .05 to minimize type II error in measures of drug safety, without corrections for multiple comparisons. For cytokines, logarithmic transformations were performed before statistical analysis. In assessment of beneficial effects of NAC within treatment groups, multiple comparisons adjusted with Bonferroni correction are also reported.

Results

Twenty-two mothers (12 preterm, 10 term) and 24 infants, including 2 sets of twins, were enrolled from August 2008 to January 2010. There were no significant differences in demographic data between the group treated with NAC and controls (Table I). No participant met stop criteria for drug infusion. Two withdrawals occurred: one for maternal preference and one infant treated with NAC for IVH and cystic changes on CUS. The infant was withdrawn from study prior to first NAC dose, and CUS findings were consistent with an event occurring more than a week prior to delivery.

Seventeen mothers had epidural anesthesia, all with rupture of membrane and other signs of chorioamnionitis. Histologic examination of placentas demonstrated 13 had both maternal and fetal inflammatory response; 5 had maternal inflammatory response alone; 2 placentas from the second twin in twin pregnancies had no inflammation; and 4 placentas were missing.

Additional details for the Results section are available in the Appendix (available at www.jpeds.com).

Maternal/Fetal AEs

There were no differences in maternal AEs between the group treated with NAC or saline. Hypotension during delivery (1 mother treated with NAC and 1 control), resolved with phenylephrine and volume expansion. One mother treated with NAC had an urticarial rash, but no anaphylactoid reactions were noted. Maternal heart rate, BP, and PTs did not show significant changes pre- and post-NAC dosing or over time between mothers treated with NAC and controls (Table I), and did not correlate with NAC concentrations. A single maternal SAE was reported in a control mother at 24 weeks GA with hypotension, fetal heart rate abnormalities, purulent amniotic fluid, severe necrotizing chorioamnionitis, and trivascular funisitis. Her infant died in the delivery room with severe mixed acidosis (cord pH <6.8, base deficit 24).

Fetal Hemodynamics

Two mothers (1 mother treated with NAC and 1 control) had nonreassuring fetal heart rate tracings, with late decelerations and fetal bradycardia. One mother treated with NAC with 39.4°C had new-onset fetal tachycardia. Fetal metabolic acidosis was not different between groups (Table I).

Umbilical cord blood flow was measured predosing (n = 13), and postdosing (n = 10; NAC 8, saline 2), and remained stable without reversal of flow in diastole. Mean cord pulsatility index was similar pre-NAC (1.16 ± 0.28, 95% CI) and post-NAC (1.12 ± 0.15, 95% CI, n = 8). Fetal MCA blood flow was measured pre- (n = 8) and postdosing (n = 6; NAC 4, saline 2) (Figure 1; available atwww.jpeds.com). Preterm and term fetuses treated with NAC had no evidence of systemic or cardiac compromise in utero. An increase in fetal MCA time average maximum velocity (TAMX) after NAC dosing was observed in the 4 preterm fetuses.

Figure 1

Mean fetal umbilical cord and MCA blood flow velocity (TAMX) before and after NAC dosing by treatment and GA (n = 4 preterm, n = 2−3 term) with 95% CIs. Doppler measures could not be obtained in preterm control fetuses.

Infant AEs

Eight infants had SAEs (Table I). Sepsis-related events were death (1 control), listeria sepsis (1 control), and pneumonia at birth (1 NAC). All CSF cultures were negative. NEC before 30 days of life (DOLs) occurred in 1 NAC and 1 control infant. Late NEC developed in 1 infant treated with NAC with Escherichia coli sepsis and grade IV IVH (DOL 35), PVL (DOL 45), cerebral palsy, and cognitive deficits at 4 years.

PT was not significantly different in infants treated with NAC or saline (Table I). At birth, PT was >18 seconds in 50% preterm infants (2 infants treated with NAC and 2 controls out of 8 preterm infants with clotting studies at birth), and one-quarter of term infants (1 infant treated with NAC with HI encephalopathy [HIE] out of 4 term infants with clotting studies at birth). Overall, 57% infants with clotting studies within 48 hours of birth had PT >18 seconds (3/6 control; 5/8 NAC). Two infants treated with NAC and 1 control received fresh frozen plasma.

In total 4 preterm infants had IVH grades II-III during first week of life, 3 had DIC at birth (2 infants treated with NAC and 1 control). One infant treated with NAC had grade I IVH bilaterally with cystic changes at 4.5 HOL, consistent with an intrauterine event prior to study drug infusion. NAC was undetectable in this infant’s cord blood, and coagulation studies clotted.

Hemodynamic Variables in Infants

Cardiovascular measures (heart rate, systolic, diastolic, mBP) were not significant between treatment groups or before and after study drug dosing (Figure 2). No infant required vasopressors during the period of NAC infusion. Cord, pre-, or postinfusion NAC plasma concentrations showed no PD correlation with BP measurements.

Figure 2

Systolic, diastolic, and mBPs before and after first NAC/saline dose (0–6 HOLs) by GA. Reference line represents normal BP measure for first week of life by GA.[23] P, preterm infants; T, term infants.

Infant CBF Velocity and Vascular Resistance

Mean CBF in cerebral vessels were not significantly different before or after dosing or between treatment groups in either GA cohort (Figure 3). However, CBF velocity increased and resistance decreased significantly very early in the MCA (0 hour to 12–24 hours) in all preterm and term infants exposed to chorioamnionitis, and by 24 hours in the ACA in preterm infants. The timing of the changes were earlier than previous reports in normal[24] and preterm infants exposed to chorioamnionitis (24–40 hours), and similar to preterm infants who developed IVH.[25] Male sex had a significant effect on MCA (P = .014) and BA resistance (P = .032), which varied by GA (Figure 4; available at www.jpeds.com). CBF and cardiac measures were not significantly different in preterm infants with (n = 4) and without IVH (n = 7) (Figure 5; available at www.jpeds.com).

Figure 3

A, Infant CBF velocity (TAMX) and resistive indices (corrected resistive index [CRI]) in the ACA, MCA, and BA before and after initial dose of NAC or saline in preterm (n = 6 NAC, n = 5 control), and B, term (n = 5 NAC, n = 4 control) cohorts (mean, 95% CI). No significant differences in TAMX or CRI between NAC and control infants in any vessel. *P < .016; †P < .007; ‡P < .05, all versus predosing (0 hour).

Figure 4

CRIs in BA and MCA over 48 HOLs by sex for individual preterm and term infants, regardless of treatment (preterm: n = 6 females, n = 5 males; term: n = 5 females, n = 4 males). Preterm males have higher CRI than females, whereas term males have lower CRI than females, as previously reported by Koch et al.[26] CRI, corrected resistive index.

Figure 5

A, CBF and B, cardiac function in preterm infants with IVH in the first week of life vs those with no insult (mean, 95% confidence intervals, CI). A, CBF velocity (TAMX) and resistance (CRI) in ACA, MCA, and BA before and after treatment with NAC or saline (n = 4 IVH, n = 7 control). B, Echocardiographic measurements before and after NAC/saline (n = 4 IVH, n = 7 control). All were not significant. CO, cardiac output index (ml/kg/min); EF, ejection fraction; LVASMV (mL/s); StokeVol, stroke volume index (mL/kg); SVCFlow, superior vena cava flow velocity (mL/kg/min).

NAC Maintains Normal Cerebral Arterial Relationships despite Chorioamnionitis

Resistive indices and blood flow velocities are normally tightly correlated in the major cerebral vessels of an individual.[24] Chorioamnionitis disrupts this tight correlation in term infants,[26] and blood flow in different regions of the brain may be affected by neuroinflammatory factors and increased metabolic demand. NAC can restore cerebral autoregulation and nitric oxide (NO) responsiveness.[27,28] We investigated the pattern of CBF correlation in ACA, MCA, and BA in individual infants, for beneficial effects of NAC on neuroinflammation-induced disruption of CBF. CBF velocities in the MCA were strongly associated with those in ACA (P = .0003) and BA (P = .003) in the group treated with NAC from 0–48 hours of the study, but not in the control group. Resistance in the ACA was also significantly associated with MCA (P = .001) and BA (P = .0004) in the group treated with NAC, but marginally associated with MCA (P = .018) in the control group over time. With Bonferroni correction for 4 mixed model comparisons within treatment groups (MCA and ACA, TAMX and corrected resistive index, P < .0125), all NAC associations remained significant.

Cardiac Function and Cerebral Oxygenation

Mean echocardiographic measurements were not significantly different between treatment groups before and after the first dose of NAC or at 12, 24, and 48 hours in either GA cohort (Figure 6; available at www.jpeds.com) and did not correlate with NAC serum concentrations. NAC administration had no main effect on left ventricular output myocardial performance index or velocity, ejection fraction, or stroke volume.[29] Importantly, male sex had a significant negative effect on ejection fraction in these infants exposed to chorioamnionitis (P = .0027).

Figure 6

Echocardiographic measures before and after dosing of NAC or saline in preterm and term infants. Cardiac output (mL/kg/min), EF, LVASMV (mL/s,) and LVMPI, stroke volume (mL/kg), and SVC flow (mL/kg/min) were not significantly different before or after dosing of NAC or saline (mean, 95% CI; preterm: n = 6 NAC, n = 5 control; term: n = 4 NAC, n = 4 control).

Near-infrared spectroscopy data were adequate in 7 term and 10 preterm infants, and there were no differences in oxygenation before, during, or after NAC dosing or over time compared with controls. Mean RcSO2 was marginally associated with cardiac output (r = 0.24, P = .05), but not with other indices of cardiac function. In preterm infants exposed to chorioamnionitis, higher RcSO2 in the frontal lobes in the first 6 hours was associated with lower resistance in the ACA (P = .031), adjusting for HOL and GA. The duration of maternal fever prior to birth also correlated with higher RcSO2 in their infants for the first 12 hours, regardless of GA (r ≥ 0.67, P < .007, n = 15–16). Taken together, our findings suggest that neuroinflammation may determine local vasodilation and RcSO2.[30]

EEG

No electrographic or clinical seizures were recorded in either group. Nine infants treated with NAC and 7 controls had adequate serial EEG tracings for dosing effects. Three infants treated with NAC had abnormal background activity (low amplitude) before NAC dose, which normalized in 2 infants during the NAC infusion. One infant treated with NAC showed persistent low amplitude.

MRI

MRI at term age equivalent had low apparent diffusion coefficients in the posterior limb of the internal capsule in 1 preterm infant treated with NAC and 1 control. Both had normal outcomes. The term infant with HIE treated with NAC had abnormal signal intensities in basal ganglia (BG) bilaterally, but normal Bayley-III scores at 12 months and normal development at 2 years of age.

MRS of myoinositol (mI)/N-acetylaspartate (NAA) ratio in the BG showed a weak trend, with control infants having higher mI/NAA ratios than infants treated with NAC, controlling for GA at birth (P = .08, n = 16). In the preterm group, BG mean mI/NAA ratios were higher in control (1.4 ± 0.23; n = 6) than preterm infants treated with NAC (0.92 ± 0.30; n = 4) exposed to chorioamnionitis (P = .026). Higher mI ratios are associated with astrogliosis, and lower NAA ratios are associated with neuronal injury.[31] There were no significant differences in fractional anisotropy in the corpus callosum or internal capsule in infants scanned at term age equivalent, controlling for GA at birth and scan.

Cytokines

Median and IQRs for cytokines from maternal serum at delivery, infant cord serum, and infant CSF samples are reported by treatment group (Table II; available at www.jpeds.com). Serum fibroblast growth factor 2 (FGF2) concentrations were significantly lower in mothers treated with NAC compared with the saline group over time (Figure 7, A), and interleukin (IL)-17 decreased after dosing only in the group treated with NAC (P = .01). FGF2 is produced by activated endothelial cells, and potentiates leukocyte recruitment and extravasation at sites of inflammation.[32] IL-17 is secreted by a subset of T helper cells-17, which are responsive to bacterial infection and inflammation.[33]

Table II

Cytokine concentrations by treatment group (median, IQR, pg/mL)

	NAC	Control
Maternal cytokines at delivery
CKine	446 (212–912)	181 (77–317)
Eotaxin	46 (20–81)	37 (31–44)
FGF2	7.6 (7.6–26)	26 (9.2–93)
Fractalkine	23 (23–23)	41 (25–69)
GCSF	1345 (357–3932)	729 (218–2050)
GMCSF	7.8 (7.5–9.3)	11 (8.4–25)
GRO	1574 (1520–2083)	944 (787–1566)
IL-1a	9.4 (9.4–9.4)	9.4 (9.4–17)
IL-1Ra	8.3 (8.3–11)	8.3 (8.3–8.3)
IL-6	81 (53–407)	35 (17–157)
IL-7	4.4 (1.4–12)	4.5 (2.5–8.8)
IL-8	13 (6.3–35)	38 (17–65)
IL-10	48.89 (30–352)	132 (38–198)
IL-17	0.7 (0.7–3.0)	20 (9.9–161)
IL-23	37 (32–51)	45 (32–305)
IP10	386 (226–564)	245 (192–506)
LIF	6.0 (5.8–7.2)	6.8 (6.3–17)
MCP1	626 (371–1640)	1018 (333–1114)
MCP2	34 (30–41)	21 (16–34)
MCP4	51 (37–84)	21 (14–58)
MDC	923 (790–953)	699 (500–864)
MIP1a	3.4 (2.9–5.5)	3.8 (2.9–7.2)
MIP1b	43 (33–64)	48 (25–103)
MIP1d	5835 (4266–7965)	6074 (3062–7351)
sCD40L	18 703 (9597–29 332)	5000 (2630–11 719)
sDF1ab	1950 (1141–2643)	831 (173–1724)
sIL-2RA	11 (11–11)	11 (11–36)
TPO	201 (178–446)	116 (77–730)
VEGF	40 (26–129)	209 (26–468)
Cord serum cytokines
CKine	1562 (991–1948)	689 (100–846)
Eotaxin	41 (19–80)	57 (42–71)
FGF2	26 (14–77)	22 (15–76)
Fractalkine	42 (24–60)	28 (24–40)
GCSF	3995 (821–14 423)	827 (176–3618)
GMCSF	7.5 (7.5–12)	7.5 (7.5–7.9)
GRO	2092 (1253–2521)	1151 (966–1570)
IL-1a	9.4 (9.4–22)	9.4 (9.4–9.4)
IL-1Ra	745 (251–1302)	8.3 (8.3–448)
IL-6	125 (53–212)	71 (9.9–195)
IL-7	1.4 (1.4–2.1)	1.4 (1.4–6.4)
IL-8	114 (40–392)	96 (11–139)
IL-10	27 (4.0–107)	4.7 (2.2–32)
IL-17	0.7 (0.7–0.7)	0.7 (0.7–3.4)
IL-23	32 (32–32)	32 (32–32)
IP10	294 (176–589)	255 (194–320)
LIF	8.4 (5.8–14)	15 (7.8–19)
MCP1	887 (725–2882)	1034 (728–1124)
MCP2	39 (33–48)	47 (44–69)
MCP4	85 (49–108)	52 (46–86)
MDC	1517 (946–1988)	1223 (1022–2327)
MIP1a	30 (9.7–88)	11 (9.6–25)
MIP1b	105 (86–143)	88 (85–98)
MIP1d	11 234 (9796–12 754)	11 693 (9968–13 114)
sCD40L	30 619 (20 426–56 028)	18 430 (15 865–40 696)
sDF1ab	2247 (1034–2546)	2582 (383–2981)
sIL2RA	187 (57–230)	11 (11–193)
TPO	269 (89–420)	224 (120–593)
VEGF	69 (26–233)	241 (234–293)
Infant CSF cytokines
Cathepsin D	34 577 (29 711–38 967)	33 890 (28 619–44 022)
Complement C4	1419 (1295–1448)	1618 (1162–1981)
CRP	54 (16–121)	48 (12–179)
FGF2	7.6 (7.6–9.1)	11 (7.9–14)
Flt3L	36 (30–43)	38 (30–41)
Fractalkine	49 (45–59)	56 (51–67)
GCSF	36 (24–224)	46 (35–199)
GRO	36 (32–105)	67 (38–137)
IL-1b	0.8 (0.8–0.8)	0.8 (0.8–0.8)
IL-1Ra	8.7 (8.3–26)	8.3 (8.3–8.9)
IL-2	1.0 (1.0–1.0)	1.1 (1.0–1.2)
IL-6	6.2 (0.9–18)	3.5 (1.5–4.5)
IL-7	3.7 (2.3–5.9)	4.9 (3.9–6.1)
IL-8	403 (243–2239)	876 (408–1735)
IL-10	4.9 (3.0–5.2)	6.2 (2.9–7.7)
IL-17	0.7 (0.7–0.7)	0.7 (0.7–0.7)
IP10	262 (219–357)	333 (266–1535)
MCP1	4359 (2962–5688)	5360 (4504–6168)
MDC	3.6 (3.6–5.5)	3.6 (3.6–3.6)
MIP1a	6.8 (5.5–12)	4.4 (3.5–13)
MIP1b	29 (18–31)	25 (16–41)
MIP4	0.2 (0.1–0.2)	0.1 (0.1–0.2)
NCAM	199 685 (17 8413–269 305)	171 580 (148 688–240 162)
PAI1 total	5221 (2009–16 333)	4053 (2998–4883)
PDGF AA	65 (37–82)	53 (48–104)
PEDF	845 (754–931)	796 (752–955)
S100b	471 (427–652)	453 (395–808)
sCD40L	5.1 (5.1–38)	5.1 (5.1–5.1)
sICAM1	1257 (1071–2443)	1130 (1029–2088)
sIL-2Ra	11 (11–11)	11 (11–11)
sVCAM1	358 857 (261 649–428 756)	353 097 (276 063–373 702)
VEGF	26 (26–26)	26 (26–26)

CKine, secondary lymphoid-tissue chemokine; FGF2, fibroblast growth factor 2; GCSF, granulocyte colony stimulating factor; GM-CSF, granulocyte monocyte colony stimulating factor; GROa, (CXCL1) growth regulated protein alpha; IP10, interferon gamma inducible protein 10; LIF, leukemia inhibitory factor; MCP, monocyte chemotactic factor; MIP, macrophage inflammatory protein; sCD, soluble cluster of differentiation antige; sDF, stromal cell-derived factor; TPO, thyroid peroxidase; CRP, C-reactive protein; FLT3L, FMS-like tyrosine kinase 3 ligand; MCD, macrophage-derived chemokine; NCAM, neural cell adhesion molecule; sICAM1, soluble intercellular adhesion molecule; sVCAM, soluble vascular cell adhesion molecule; PAI1, plasminogen activator inhibitor 1; PDGF AA, Platelet-derived growth factor -AA; PEDF, pigment epithelium-derived factor; S100b, S100 calcium binding protein B.

Figure 7

Median, IQR of serum cytokine concentrations (pg/mL): A, FGF-2 before and after NAC/saline dosing in mothers (NAC n = 11; saline n = 7, P = .02); B, VEGF and IL-1Ra in infants over 0–48 hours after delivery with GA and time in mixed model (n = 12 NAC, n = 9 saline; P ≤ .014). For IL-6, there was a significant NAC treatment*time interaction effect (P = .014), from 36–48 HOL. FGF-2, fibroblast growth factor 2.

Infants treated with NAC had significantly lower serum proinflammatory vascular endothelial growth factor (VEGF) and higher anti-inflammatory IL-1 receptor antagonist (IL-1Ra) over time compared with controls (Figure 7, B). For IL-1Ra and IL-6, treatment-time interactions were evident, and IL-6 was lower at 36 and 48 hours in infants treated with NAC vs controls. Our infant serum IL-6 concentrations were similar to published values in preterm infants exposed to chorioamnionitis with fetal inflammatory response and early sepsis (>40 pg/mL)[34] and in preterm infants with white matter injury.[35] VEGF contributes to early brain injury as a permeability factor associated with blood brain barrier disruption, hemorrhage, and ischemia.[36] Anti-inflammatory IL-1Ra provides significant neuroprotection in animal models of chorioamnionitis and HI injury, by blocking IL-1b activity, leukocyte infiltration, and microglial activation.[37] Consistent with these data, lower serum IL-1Ra was associated with higher mI/creatinine in the white matter in our infants (n = 13, r = −0.73; P = .01). There were no significant differences between treatment groups in CSF cytokines obtained within 12 hours of birth, controlling for GA (n = 16).

Serum IL-6, IL-8, VEGF, and monocyte chemotactic protein-1 were also significantly related to MCA CBF velocity over the 48 hours of the study in a mixed model with HOL, treatment, and GA (all P ≤ .03, n = 20 subjects, 91 paired CBF/cytokine observations from 0–48 hours). These data are consistent with the roles of these cytokines as biomarkers of endothelial activation and injury.

Developmental Outcomes

Twenty-one infants exposed to chorioamnionitis were available for follow-up to 3–4 years of age (9 controls and 12 infants treated with NAC). Two out of 9 control infants had developmental delay; 1 preterm infant with IVH had speech delay at 3 years, and 1 preterm infant with NEC had fine motor delay at 4 years. Two of the 12 infants treated with NAC had delays; 1 preterm infant with late NEC, grade 4 IVH, and PVL, had spastic quadriplegia and cognitive impairment, and 1 preterm infant had fine motor delay and autism spectrum disorder at 4 years.

Discussion

In this pilot study of neuroprotection in preterm and term infants exposed to chorioamnionitis, NAC administration antenatally in the mother and postnatally to their infants, resulted in no significant adverse effects on CBF, cerebral oxygenation, cardiac function, clotting measurements, or BP. Infants treated with NAC did show beneficial effects, as NAC restored normal cerebrovascular coupling between major cerebral vessels, decreased proinflammatory VEGF, and increased anti-inflammatory IL-1Ra compared with infants who received saline. Mothers treated with NAC also had lower cytokines associated with endothelial activation and leukocyte recruitment during inflammation. Fetal systemic inflammatory response with elevated circulating IL-6 and clotting abnormalities occurred in the majority of infants, strongly suggesting the presence of fetal endothelial activation and neuroinflammation.

In human infants and animal models of chorioamnionitis, FIRS and neuroinflammation cause injury to the CNS and other organs by direct toxicity of cytokine mediators, NO dysregulation, and alterations in perfusion and cerebral autoregulation, which occur very early after infection. Vascular endothelial injury results in dysregulation of fetal and neonatal CBF,[8,38] as we saw with decreased cerebrovascular resistance and increased CBF within 12–24 hours, particularly in males. Furthermore, the immediate loss of correlation of CBF and resistance between the 3 major cerebral vessels after birth is consistent with intrauterine onset of neuroinflammation and vascular dysfunction.[26,39] Although we did not include a control group of preterm and term infants without chorioamnionitis exposure, another report has found uncoupling of CBF and resistance in term neonates exposed to chorioamnionitis, compared with term infants not exposed to chorioamnionitis.[26] Also, abnormal CBF has been documented in early onset sepsis and in preterm infants who develop IVH.[25,30] Fetal inflammatory cascades also predispose term and preterm infants to significant morbidities of IVH, HIE, white matter injury, and NEC, in which vascular insufficiency may be a contributing factor.[25,30]

Antenatal NAC can counteract fetal neuroinflammation associated with chorioamnionitis by several mechanisms, including scavenging oxygen free radicals, restoring intracellular glutathione levels, and decreasing inflammatory cytokine production.[10-13] NAC stabilizes CBF, enhances autoregulation, and re-establishes normal vascular reactivity, which depends on endothelial synthesis of NO.[27,40] NAC inhibits conversion of NO to peroxynitrite, thereby preserving the bioavailability of NO for normal vascular responsiveness under oxidative stress.[27,41]

NAC also had no untoward effects on cerebral or systemic perfusion in the fetus or infant when started within 4 hours of clinical diagnosis of chorioamnionitis and administered for the first 48 hours after birth. In a similar time frame in adult patients with endotoxic shock, NAC increased cardiac output index, oxygen delivery, and systemic vascular resistance.[42] However, NAC effects are more variable when started later than 24 hours in sepsis.[43] Furthermore, high dose NAC has been associated with reduced left ventricular stroke work in a small number of adult patients, when given more than 24 hours from onset of septic shock.[44] If present in excess, NAC sulfhydryl groups may react with NO to form S-nitrosothiols, a stable, stored form of NO that can cause vasodilation.[45]

Although treatment was instituted within 4 hours of chorioamnionitis in our study, we considered that NAC could interact with specific vulnerabilities of fetal and neonatal physiology to produce hypotension. With extensive physiologic monitoring before and after NAC dosing, we found no adverse hemodynamic changes with 100 mg/kg in mothers or 12.5–25 mg/kg/dose NAC administered to their infants compared with saline treatment. Fetuses had significant NAC plasma concentrations, but umbilical and CBF velocities were stable before and after maternal NAC dosing.[19]

We found no differences in infant BP or RcSO2 over 2 days of dosing between NAC and saline-treated infants, contrary to a report in fetal sheep.[18] In this model of septic shock, daily lipopolysaccharide (LPS) administration was followed by a 5-hour infusion of 50, 100, or 200 mg/kg NAC for 5 days (n = 2 per dose),[18] which causes significant hypoxemia and hypotension for 3–24 hours in this model.[46] When comparing effect of NAC, most of the significant hemodynamic differences were between sham and LPS animals in the mixed model. In post hoc analyses, NAC-LPS animals showed small changes of questionable clinical significance in partial pressure of oxygen in arterial blood, pH, lactate, and mBP compared with LPS-saline animals.[18]

NAC plasma concentrations and cerebral perfusion were not measured in the fetal sheep, which inhibits a PD comparison with our human study, but NAC dose did not correlate with the changes.[18] In our infants, NAC did not adversely affect systemic perfusion, cardiac function, or CBF before and after NAC dosing, or over 48 hours compared with control infants exposed to chorioamnionitis. None of the PD measurements of hemodynamic, cerebral perfusion, or cardiac variables correlated with concurrent NAC plasma concentrations. This indicates that these PD measures are not influenced by NAC at the doses used in infants exposed to chorioamnionitis.

The small sample size in this intensive safety study is a potential limitation. However, other studies have documented no AEs in fetuses of mothers who received NAC for acetaminophen overdose, or in very preterm infants who received NAC continuously for the first 6 DOLs to prevent bronchopulmonary dysplasia.[16,17] Our incidence of NEC and IVH after chorioamnionitis is similar to other reports, but our sample size is not large enough to evaluate effects of NAC treatment on these complications. Infants treated with NAC had no greater adverse outcomes related to early events than control infants.

More than one-half of our infants with clotting studies had significant DIC within the first 24 hours after birth, independent of treatment. Coagulopathy has been noted in other reports in 25%-30% of infants exposed to chorioamnionitis, and provides laboratory evidence of systemic inflammatory response with endothelial activation.[47,48] In addition, these findings suggest that clotting abnormalities and early increases in CBF may contribute to greater incidence of IVH and neurodevelopmental delays in infants exposed to chorioamnionitis.[25]

Although our sample size is limited, we did not see threshold or concentration-dependent adverse effects of NAC on CBF, cardiac function, or cerebral oxygenation. Neuroprotective compounds that can be used in this population are rare, and these data support further evaluation of NAC for antenatal neuroprotection.

For antenatal drug administration to be effective in future treatment trials for chorioamnionitis, consideration should be given to length of the consent process in laboring mothers and the short duration of labor in term mothers with chorioamnionitis. However, our PK data show that even a short infusion of NAC rapidly crosses the placenta and can be measured in the cord blood of the neonate, making antenatal administration of NAC feasible for fetal neuroprotection in chorioamnionitis.[19]