Objective
The aim of this study was to show and discuss an association between fetal inflammatory response syndrome (FIRS) and an adverse neonatal outcome defined as combined severe neonatal morbidity and mortality in preterm neonates hospitalized in our neonatal intensive care unit.
Study Design
This was an observational study including all preterm neonates hospitalized in our neonatal intensive care unit over a 21 month period. FIRS was defined as cord blood interleukin (IL)-6 greater than 11 pg/mL. Main outcome parameter was an adverse neonatal outcome defined as hospital mortality and/or the presence of any of 5 prespecified morbidities (bronchopulmonary dysplasia, periventricular leukomalacia, intraventricular hemorrhage, and early- or late-onset sepsis).
Results
Fifty-seven of 176 preterm infants hospitalized during the study period (32%) had an adverse neonatal outcome and 62 of these 176 infants (35%) had FIRS with median IL-6 values of 51.8 pg/mL (range, 11.2 to >1000 pg/mL). In a regression analysis, FIRS was significantly associated with adverse neonatal outcome ( P < .001) and with the single outcome parameters, intraventricular hemorrhage and early-onset sepsis ( P = .006 and P = .018, respectively). In the bivariate analysis, FIRS was associated with death and bronchopulmonary dysplasia ( P = .004 and P < .001, respectively). IL-6 correlated with adverse neonatal outcome (r = 0.411, P < .001). When comparing the correlation in neonates less than 32 weeks’ gestational age (r = 0.481, P < .001) with neonates 32 weeks or longer (r = 0.233, P = .019), the difference was nearly significant ( P = .065).
Conclusion
FIRS is a risk factor for adverse neonatal outcome in preterm infants. In particular, the combination of IL-6 greater than 11 pg/mL and low gestational age increased the risk for severe neonatal morbidity or death.
Preterm birth is the leading cause of infant mortality in the industrialized civilization. The rates of both mortality and significant long-term sequelae decrease with the duration of pregnancy, with very immature infants being at the highest risk for an adverse outcome. Major morbidity occurs in 65% of the survivors with a birthweight of 501-750 g and in 11% of the survivors with a birthweight of 1251-1500g.
An expanding body of evidence supports a causal link between intrauterine infection or inflammation and spontaneous preterm birth, with a stronger association the more preterm the delivery. Different authors have shown higher rates of positive amniotic fluid cultures in women with preterm labor but intact membranes at earlier gestational ages, higher interleukin (IL)-6 levels in the amniotic fluid in women with spontaneous labor in pregnancies less than 34 weeks vs 34 weeks or longer, and higher IL-6 levels in cord blood in preterm compared with term neonates born after microbial invasion of the amniotic cavity.
The fetal inflammatory response syndrome (FIRS) is a condition characterized by systemic activation of the fetal immune system. FIRS was originally defined as an elevation of the fetal plasma IL-6 concentration in fetuses of mothers with preterm labor and premature rupture of the membranes (PROM). Affected fetuses had evidence of multiorgan involvement and were more likely to have a subsequent spontaneous preterm delivery in cases of preterm PROM. Some authors have provided evidence that FIRS represents a significant risk factor for neonatal morbidity including bronchopulmonary dysplasia and neurological pathology.
The aim of this study was to determine whether there was an association between FIRS and an adverse neonatal outcome in preterm neonates admitted to the intensive care unit.
Materials and Methods
Study design
The study site was the neonatal intensive care unit of the Division of Neonatology, Pediatric Department, Medical University of Graz, a tertiary care center with approximately 2700 inborn births per year. The main focus of interest was the association of umbilical cord blood IL-6 with combined severe neonatal morbidity (as defined below) and mortality in preterm infants.
From July 1, 2009, to March 31, 2011, IL-6 was determined from umbilical cord blood in all preterm neonates at risk of bacterial infection (pregnancy complicated by PROM, chorioamnionitis, maternal fever during labor, maternal leukocyte count greater than 12,000/μL or C-reactive protein [CRP] greater than 8 mg/L, or preterm spontaneous onset of labor) born at the university hospital obstetrics unit.
Blood was drawn from a clamped umbilical vein and IL-6 levels were determined from plasma using the Endogen Interleukin-6 ELISA (Endogen Inc, Cambridge, MA), an enzyme-linked immunosorbent assay, performed according to the recommended procedure. The upper and lower detection limits were 1 and 1000 pg/mL, and values below or above were estimated by using a maximum likelihood estimation. Within the text, these values are given as less than 1 or greater than 1000 pg/mL. FIRS was defined present at a IL-6 level greater than 11 pg/mL.
All neonates were initially assessed by and cared for immediately postnatally by a neonatologist. All infants were admitted to the neonatal intensive care unit unless they were near term and in a good general condition. Inclusion criteria were prematurity (<37 completed weeks postmenstrual age), IL-6 determination from cord blood, and admission to our neonatal intensive care unit. Exclusion criteria were major congenital anomalies. Cases of chromosomal aberrations and genetic diseases without organ dysfunction were not excluded.
Clinical data were entered in the electronic patient filing system (MEDOCS), which included maternal and perinatal information; the postnatal course including all relevant clinical information, ventilation protocols, reports, and relevant images from all ultrasound scans; and all laboratory parameters determined during the hospital stay.
The study was approved by the local ethics committee (number 23-446 ex 10/11).
Outcome parameters
The primary outcome parameter was adverse neonatal outcome defined as hospital mortality or presence of any of 5 morbidities diagnosed during hospitalization: bronchopulmonary dysplasia (BPD), periventricular leukomalacia (PVL), intraventricular hemorrhage (IVH), or early- or late-onset sepsis (EOS and LOS). Furthermore, we searched for an association between any of the aforementioned 5 morbidities and mortality and FIRS.
The diagnosis of BPD was based on the Eunice Kennedy Shriver National Institute of Child Health and Human Development/National Heart, Lung, and Blood Institute/Office of Rare Diseases consensus and was graded according to severity: treatment with oxygen greater than 21% for at least 28 days but breathing room air at 36 weeks’ postmenstruational age or at discharge whatever comes first (mild BPD), need for less than 30% oxygen at 36 weeks/discharge (moderate BPD), and need for 30% or greater oxygen or positive pressure ventilation at 36 weeks/discharge (severe BPD).
Cranial ultrasound (US) scans were obtained routinely in all preterm infants on days 1, 3, and 5 of life and thereafter weekly in cases with abnormal findings, and in all patients, a final scan was performed at discharge. Real-time US scans were performed with a commercially available unit (Advanced Technology Laboratories Inc, Bothell, WA) using a 7.5 or 8.2 MHz transducer).
PVL grades were defined according to de Vries et al : grade I, periventricular flares present for 7 days or more; grade II, presence of small-localized frontoparietal cysts; grade III, extensive periventricular cystic lesions involving occipital and frontoparietal white matter; and grade IV, extensive subcortical cystic lesions.
IVH was classified using the grade 1-4 system according to Papile et al, with grade 4 hemorrhage being periventricular hemorrhage (PVH). The highest grade documented was used for analysis.
EOS and LOS were defined as culture-proven or clinical sepsis with onset at less than 72 hours of life and 72 or more hours of life, respectively. Culture-proven EOS and LOS were defined as positive bacterial culture from umbilical cord blood, peripheral blood, or cerebrospinal fluid. Clinical EOS and LOS were defined as negative culture from umbilical cord blood, peripheral blood, or cerebrospinal fluid but the presence of clinical signs of sepsis in 3 or more categories ([1] respiratory signs; [2] cardiocirculatory signs; [3] neurological signs; [4] poor skin color or prolonged capillary refilling time [more than 2 seconds]; or [5] fever or hypothermia [core temperature >38.5°C or <36.0°C]), with either 1 or more maternal risk factors (PROM [sterile speculum examination demonstrating pooling of fluid with a positive nitrazine test more than 1 hour before onset of labor], clinical chorioamnionitis [uterine tenderness or foul-smelling amniotic fluid, elevated maternal leukocyte count >12,000/μL, and maternal or fetal tachycardia], maternal fever higher than 38.5°C during labor], or ≥2 abnormal laboratory markers [leukocyte count and absolute neutrophil count with age-adapted cutoff values, immature to total neutrophil ratio >0.2, CRP >8 mg/L]). For bivariate and for regression analysis culture proven and clinical sepsis were grouped to one variable.
Statistical analysis
SPSS 17 (SPSS Inc, Chicago, IL) and Minitab 16 (Minitab Inc, State College, PA) were used for statistical analysis. Descriptive statistics (median with range, case number, percentage calculations) were used to characterize the study population.
Bivariate analysis of risk factors for an adverse neonatal outcome and for every single outcome parameter was performed using a Fisher exact test, a Mann-Whitney U test, or a Student t test as appropriate. Odds ratio (OR) with 95% confidence interval (CI) was calculated for categorical variables.
The variables investigated were gestational age, birthweight, sex, multiple gestation, gestational diabetes, preeclampsia, HELLP (hemolysis, elevated liver enzymes, and low platelet count) syndrome, bacterial vaginosis, vaginal colonization with group B streptococci, PROM (as defined in previous text), clinical chorioamnionitis (as defined in previous text), maternal fever higher than 38.5°C during labor, maternal leukocyte count greater than 12,000/μL and/or CRP greater than 8 mg/L, pathological CTG, placental abruption, meconium staining of the amniotic fluid, umbilical artery pH, Apgar score at 1, 5, and 10 minutes, severe (Apgar 1 less than 3) and moderate (Apgar 1: 4-6) birth asphyxia, presence of any grade idiopathic respiratory distress syndrome (IRDS; defined as signs of respiratory distress with compatible radiographic findings on the chest X-ray), severe IRDS (IRDS grades III-IV), patent ductus arteriosus (PDA; confirmed by echocardiography and treated with indomethacin or surgically).
We calculated the OR for an adverse neonatal outcome by IL-6 at different cutoffs for neonates at a gestational age of less than and 32 weeks or longer. ORs between the 2 groups of gestational age were compared as described by Altman and Bland.
We evaluated the alternative cutoff value for IL-6 of 17.5 pg/mL identified by Yoon et al by performing a bivariate analysis of the association of the single outcome parameters with IL-6 of 17.5 pg/mL or greater.
Before performing regression analysis, variables with very low case numbers (<5 cases) were excluded (HELLP syndrome, maternal fever during labor, placental abruption). Next, variables were tested for collinearity by using a χ 2 test and Pearson’s or Spearman’s correlation coefficient as appropriate. Factors with a significant correlation ( P < .05) were either summarized or not included in the analysis: gestational age alone was included instead of both birthweight, and gestational age, bacterial vaginosis, and colonization with group B streptococci (GBS) were summarized as abnormal vaginal colonization, severe (Apgar 1 less than 3) and moderate (Apgar 1: 4-6) birth asphyxia were included instead of Apgar scores, any grade IRDS was included instead of both, any, and severe grade IRDS. FIRS correlated with gestational age, birthweight, Apgar score at 1, 5, and 10 minutes, IRDS and PDA, but as a main parameter of interest, it was not summarized with the other variables.
We performed a regression analysis with adverse neonatal outcome and all single primary outcome parameters separately as independent variables. Because of its low case number, death was not analyzed separately. A regression analysis was performed as follows: we entered factors not considered malleable to intervention with respect to time order (all antenatal factors) and added those variables that correlated to the outcome parameters in the bivariate model with P < .1. We checked again for multicollinearity and regarded a correlation coefficient less than 0.5 as not significant. In case of collinearity, we excluded the parameter less likely to have clinical influence.
We used formal testing to further evaluate prespecified effect modification and used an interaction term to test whether the association between FIRS and outcome differed further as a function of gestational age. Model fit was described by Nagelkerke’s R 2 and percentage of neonates correctly classified. We reported all predictor variables with a significance of P < .1. We calculated predicted risk of outcome for every single-outcome parameter and plotted it against gestational age separately for FIRS-positive and -negative neonates. We added observed outcome for every single-outcome parameter separately for FIRS-positive and -negative neonates.
Receiver-operating characteristics (ROC) curve analysis was performed for cord blood IL-6 predicting an adverse neonatal outcome and the single primary outcome parameters. We calculated the area under the ROC curve with 95% CI with P value and determined the best cutoff value (resulting in the highest sum of sensitivity and specificity) for every single-outcome parameter.
Results
During the study period, 177 neonates had cord blood IL-6 determined, but 1 neonate was excluded because of a complex congenital heart defect; thus, the study population consisted of 176 neonates. The median gestational age was 32 completed weeks (range, 24–36 weeks) and the median birthweight was 1775 g (range, 410–3570 g). Further characteristics of the study patients are presented in Table 1 .
| Variable | Value |
|---|---|
| Total | 176 (100) |
| Sex (male:female) | 101:75 (57:43) |
| Gestational age, wks | 32 (24–36) |
| Birthweight, g | 1775 (410–3570) |
| Singletons:twins:triplets | 124:49:3 (70:28:2) |
| Chromosomal aberrations and genetic diseases | 2 (1 trisomy 21 without major organ involvement, 1 Cri-du-Chat-syndrome with multiple small ventriculoseptal defects without hemodynamic compromise) |
| Apgar 1 min | 8 (0–9) |
| Apgar 5 min | 9 (2–10) |
| Apgar 10 min | 9 (3–10) |
| IRDS | 87 (49) |
| EOS | 32 (18) |
| LOS | 8 (5) |
| NEC | 1 (0.5) |
| TRDN/wet lung | 22 (12) |
| Mechanical ventilation | 157 (89) |
| Days | 1 (1–114) |
| Surfactant application | 54 (31) |
| BPD all stages (among survivors longer than 28 d) | 16 (9) |
| Mild BPD | 14 (8) |
| Moderate BPD | 0 (0) |
| Severe BPD | 2 (1) |
| PVL | 13 (7) |
| Grade I | 5 (3) |
| Grade II | 5 (3) |
| Grade III | 3 (2) |
| IVH | 21 (12) |
| Grade I | 13 (7) |
| Grade II | 2 (1) |
| Grade III | 6 (3) |
| Death | 7 (4) |
| Adverse neonatal outcome (EOS, LOS, IVH, PVL, BPD, and/or death) | 57 (32) |
| FIRS | 62 (35) |
Sixty-two of 176 neonates (35%) had FIRS. Neonates with FIRS had lower gestational age and lower birthweight compared with neonates without FIRS (median, 30.7 vs 33.0 weeks, P < .001, and 1419 vs 1878 g, P < .001).
IL-6 and combined severe morbidity and mortality
Fifty-seven of 176 neonates (32%) had an adverse neonatal outcome (presence of BPD, PVL, IVH, EOS, LOS, and/or death). Neonates with an adverse neonatal outcome had higher cord blood IL-6 concentration and lower gestational age compared with infants without any primary outcome parameter (median, IL-6, 29.2 vs 5.0 pg/mL, P < .001, and median gestational age, 30.1 vs 33.4 weeks, P < .001). The OR for an adverse neonatal outcome increased from 7.0 (95% CI, 3.5–14.0) for IL-6 greater than 11 pg/mL to 9.4 (range, 3.8–22.9) for IL-6 greater than 50 pg/mL and to 34.9 (range, 4.4–274.4) for IL-6 greater than 500 pg/mL ( P < .001 for all).
Gestational age had no significant interaction effect on the relationship between IL-6 above the different cutoffs and adverse neonatal outcome, but it was an independent risk factor for an adverse neonatal outcome ( P < .001 for all cutoffs). Outcome by IL-6 concentration and gestational age is presented in Table 2 and Figure 1 . Adverse neonatal outcome was significantly associated with FIRS in the bivariate and in the regression analysis ( P < .001 for both; Tables 3 and 4 ). Figure 2 shows the predicted risk of outcome for FIRS-positive and -negative neonates obtained in a regression analysis plotted against gestational age.
| IL-6, pg/mL | GA <32 wks | P value | GA ≥32 wks | P value | Difference between ORs |
|---|---|---|---|---|---|
| OR (95% CI) | OR (95% CI) | P value | |||
| >5 | 5.1 (1.8–14.3) | .001 | 1.1 (0.4–3.3) | .877 | .0421 |
| >11 | 11.3 (3.7–34.1) | < .001 | 3.1 (1–9.6) | .045 | .1097 |
| >50 | 12.8 (2.7–60.6) | < .001 | 4.8 (1.2–19.9) | .018 | .3598 |
| >500 | All had an adverse outcome a | .001 | 13.1 (1.1–154.7) | .010 | Not calculable a |
a OR not calculable because of zero patients in the group of neonates without adverse neonatal outcome; therefore, also P value for the difference between the ORs is not calculable.
| Outcome parameter and significant risk factors | OR (95% CI) or median in neonates with vs without the indicated outcome parameter | P value |
|---|---|---|
| Adverse neonatal outcome | ||
| GA, wks | 30.6 vs 33.1 | < .001 |
| Birthweight, g | 1410 vs 1925 | < .001 |
| Apgar 1 min | 7 vs 8 | < .001 |
| Apgar 5 min | 8 vs 9 | < .001 |
| Apgar 10 min | 9 vs 9 | < .001 |
| Severe birth asphyxia | 5.5 (1.6–18.8) | .003 |
| FIRS | 7 (3.5–14) | < .001 |
| IRDS | 6.9 (3.3–14.4) | < .001 |
| PDA | 5.2 (2.6–10.3) | < .001 |
| Death | ||
| GA, wks | 27.6 vs 32.6 | .022 |
| Birthweight, g | 1009 vs 1810 | .020 |
| Maternal fever during labor | 28 (1.6–503.3) | .001 |
| Placental abruption | 13.9 (1.1–175.6) | .009 |
| Apgar 1 min | 7 vs 8 | .009 |
| Apgar 5 min | 8 vs 9 | .001 |
| Apgar 10 min | 9 vs 9 | .021 |
| FIRS | 12.1 (1.4–103) | .004 |
| IRDS | All had IRDS; OR not calculable a | .006 |
| BPD | ||
| GA, wks | 26.1 vs 33.0 | < .001 |
| Birthweight, g | 820 vs 1886 | < .001 |
| Apgar 1 min | 7 vs 8 | .028 |
| Apgar 10 min | 9 vs 9 | .034 |
| FIRS | 6.9 (2.1–22.4) | < .001 |
| IRDS | All had IRDS; OR not calculable a | < .001 |
| PDA | 40.1 (5.1–313) | < .001 |
| IVH | ||
| GA, wks | 30.1 vs 32.9 | .002 |
| Birthweight, g | 1315 vs 1870 | < .001 |
| HELLP syndrome | 16.2 (1.4–187.4) | .003 |
| Apgar 1 min | 7 vs 8 | .004 |
| Apgar 5 min | 8 vs 9 | .002 |
| Apgar 10 min | 9 vs 9 | .014 |
| Severe birth asphyxia | 3.8 (1.1–13.6) | .030 |
| FIRS | 7.6 (2.6–21.9) | < .001 |
| IRDS | 5.2 (1.7–16) | .002 |
| PDA | 2.5 (1–6.4) | .044 |
| PVL | ||
| GA, wks | 30.3 vs 32.9 | .008 |
| Birthweight, g | 1250 vs 1820 | .016 |
| EOS | ||
| GA, wks | 28.8 vs 33.0 | < .001 |
| Birthweight, g | 1095 vs 1878 | < .001 |
| Apgar 1 min | 7 vs 8 | .001 |
| Apgar 5 min | 8 vs 9 | < .001 |
| Apgar 10 min | 9 vs 9 | < .001 |
| Severe birth asphyxia | 3.1 (1.0–10.3) | .049 |
| FIRS | 13 (5–34.1) | < .001 |
| IRDS | 10.1 (3.4–30.3) | < .001 |
| PDA | 6.6 (2.9–15.2) | < .001 |
| LOS | ||
| GA, wks | 26.5 vs 32.9 | < .001 |
| Birthweight, g | 740 vs 1830 | < .001 |
| Multiple birth | 4.3 (1–18.7) | .037 |
| IRDS | All had IRDS; OR not calculable a | .003 |
| PDA | 16.1 (1.9–133.9) | .001 |
a OR is not calculable because of zero patients in the group of neonates without the according outcome parameter.
| Outcome parameter | R 2 | Predicted correctly | Predictors | OR (95% CI) | P value | P value for interaction term a |
|---|---|---|---|---|---|---|
| Adverse neonatal outcome | .390 | 79.9% | GA | 1.2 (1.0–1.4) | .027 | |
| Severe birth asphyxia | 3.8 (0.9–16.9) | .076 | ||||
| FIRS | 5.8 (2.5–13.8) | < .001 | .839 | |||
| BPD | .892 | 97.6% | GA | 11.1 (2.0–62.6) | .006 | .976 |
| IVH | .419 | 92.5% | GA | 1.3 (1.0–1.6) | .091 | |
| Multiple birth | 4.3 (0.8–22.4) | .082 | ||||
| Severe birth asphyxia | 5.0 (0.8–32.3) | .090 | ||||
| FIRS | 7.7 (1.8–32.9) | .006 | .158 | |||
| PVL | .119 | 92.6% | GA | 1.3 (1.0–1.5) | .017 | .392 |
| EOS | .532 | 90.2% | GA | 1.31 (1.08–1.59) | .007 | |
| IRDS | 4.81 (1.13–20.51) | .034 | ||||
| FIRS | 10.26 (2.95–35.68) | < .001 | .414 | |||
| LOS | .473 | 96.0% | GA | 2.0 (1.2–3.3) | .006 | .695 |
| Multiple birth | 6.7 (0.8–56.8) | .081 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
