Abstract
Keywords
Abbreviations
Extremely Low Gestational Age Newborns Necrotizing Enterocolitis Spontaneous Intestinal Perforation Fresh Frozen Plasma Body Mass Index Score for Neonatal Acute Physiology SNAP-Perinatal Extension-II
Introduction
In recent years, gut pathology in Extremely Low Gestational Age Newborns (ELGANs) leading to significant morbidity and mortality has been differentiated into necrotizing enterocolitis (NEC) and spontaneous intestinal perforation (SIP). NEC has been classified using the modified Bell’s classification with each stage having well-defined clinical, radiological and histopathological findings [1]. In contrast to NEC, which is primarily inflammatory, the histopathology of SIP is more consistent with focal necrosis and often has better outcomes than NEC [2–6].
The antecedents of NEC and SIP have often been evaluated in samples of infants defined by birth weight [7]. This poses a problem in light of the evidence that the incidence of NEC is elevated among infants who experienced growth restriction in utero [8] and that samples defined by birth weight have the potential to result in conflicting and inconsistent inferences about the etiology of disorders more clearly related to gestational age [9]. These problems would be reason enough to evaluate the antecedents of NEC and SIP among ELGANs and focusing on gestational age as opposed to birth weight. In addition, NEC appears to have its origins close to birth and influenced by exposure to certain antenatal and early postnatal factors [10–12]. Some of the conflicting findings in the literature might reflect evaluation of postnatal events, exposures, and characteristics when ignoring antenatal contributions [13–15].
The ELGAN Study of extremely low gestational age newborns (delivered before the 28th week of gestation) allowed us to evaluate the postnatal antecedents of NEC and SIP in light of antenatal characteristics in a large sample of high-risk newborns enrolled on the basis of gestational age.
Subjects and methods
The ELGAN study was designed to identify characteristics and exposures that increase the risk of structural and functional neurologic disorders in ELGANs (Extremely Low Gestational Age Newborns) [16]. Between 2002–2004, women delivering before 28 weeks gestation at one of 14 participating institutions were asked to enroll in the study. The enrollment and consent processes were approved by the individual institutional review boards.
Mothers were approached for consent either upon antenatal admission or shortly after delivery, depending on clinical circumstance and institutional preference, and 1249 mothers of 1506 infants consented. Approximately 260 women were either not approached for participation or did not consent to participate.
The sample for the current study consists of the 1320 newborns who survived beyond 28 days or died before then with a diagnosis of Stage II or worse NEC. Placenta histology analysis was limited to the 1230 newborns whose placenta was examined histologically, and for placental bacteriology analyses were restricted to the 1193 newborns whose placenta was sampled under sterile conditions and eligible for culture.
Demographic and pregnancy variables
After delivery, a trained research nurse interviewed each mother in her native language using a structured data collection form while adhering to the study’s procedure manual. The clinical circumstances leading to preterm delivery were operationally defined using both data from the maternal interview and data abstracted from the medical record [17]. Each mother/infant pair was assigned to the category describing the primary reason for the preterm delivery [17].
Maternal body mass index (BMI)
BMIs have been categorized as follows: <18.5 is underweight, 18.5–24.9 is normal, 25.0–29.9 is overweight, 30.0–34.9 is obese, 35.0–39.9 is very obese, and 40+ is extreme obesity [18]. We collapsed BMI into 3 groups: <25, 25–29.9, and 30+ .
Newborn variables
The gestational age estimates were based on a hierarchy of the quality of available information. Most desirable were estimates based on the dates of embryo retrieval or intrauterine insemination or fetal ultrasound before the 14th week (62%). When these were not available, reliance was placed sequentially on a fetal ultrasound at 14 or more weeks (29%), last menstrual period (7%), and gestational age recorded in the log of the neonatal intensive care unit (1%). The birth weight Z-score is the number of standard deviations the infant’s birth weight is or was above or below the median weight of infants at the same gestational age in referent samples not delivered for preeclampsia or fetal indications [19, 20].
We collected all the physiology, laboratory and therapy data for the first 12 hours needed to calculate a SNAP-IITM score [21]. Of our sample, 28% had a SNAP-II value of >30 and 32% had a SNAPPE-II>45. We also identified cut-offs for each week of post-menstrual age at birth that defined the top quartile and top decile of SNAP-II and SNAPPE-II (Supplementary material). Documented early bacteremia was defined as recovery of an organism from blood drawn during first week (Supplementary material). Medications administered during the first postnatal week, including surfactant, vitamin A, and steroids (i.e., hydrocortisone and dexamethasone) were recorded, but dosages were not. The infant’s chart was reviewed for receipt of blood products within the first seven days of life.
NEC was classified by the Modified Bell’s staging criteria [1]. A NEC “watch” identified infants who were evaluated for NEC between 48 to 72 hours, but their abdominal radiographs did not demonstrate pneumatosis. Stage 1 included infants with “suspected” NEC and, despite the absence of pneumatosis on abdominal radiographs, was treated with antibiotics and suspension of enteral feedings for at least one week. Stage IIa infants had pneumatosis, but did not experience clinical deterioration or laboratory derangements. Stage IIb included infants with pneumatosis and metabolic (acidosis) and/or hematologic changes (thrombocytopenia). Stage IIIa included infants with Stage IIb criteria plus respiratory or cardiovascular deterioration (e.g., increased need for respiratory support, new vasopressor requirement, oliguria, disseminated intravascular coagulation). Finally, Stage IIIb identified those infants who required surgical intervention, either exploratory laparotomy or placement of Penrose drain. For analysis, we classify NEC as medical (Stages IIa, IIb, and IIIa) (mNEC) or surgical (Stage IIIb) (sNEC). SIP was defined as a gastrointestinal perforation documented on radiograph in the absence of other clinical features of NEC.
Placentas
Placental parenchyma was biopsied under sterile conditions. Eighty-two percent of the samples were obtained within one hour of delivery. The microbiologic procedures and placental inflammation grading were previously described [22, 23] (Supplementary material).
Data analysis
We evaluated the generalized null hypothesis that the risks of mNEC, sNEC, and SIP were not associated with any antenatal or early postnatal characteristic. Initially, we adjusted for gestational age by both week of gestation [23–27], and by collapsing weeks into three groups [23–27]. No differences were noted; therefore, we chose to present data adjusted for gestational age categorically. We carried out bivariate analyses (Tables 1–4 and Supplementary Appendix Tables A-C) to identify candidates for multivariable analyses. Because postnatal phenomena, such as the need for mechanical ventilation on day seven, can be influenced by antenatal and/or delivery phenomena, such as indication and mode of delivery, chorioamnionitis, and extent of growth restriction, we developed multinomial multivariable logistic regression models in which risk factors were ordered in a temporal pattern, so that the earliest occurring predictors/covariates of the outcome (mNEC, sNEC, and SIP) are entered first and are NOT displaced by later occurring covariates. For these time-oriented risk models (TORMs), we categorize sets of antecedents/covariates by the time they occur or are identified (identified as antenatal, delivery, and first postnatal week). Variables that occur earliest (e.g., demographic, pregnancy characteristics; epoch 1) are entered into the regression and non-significant variables are dropped, one by one, until we have only significant variables remaining. Then the previously dropped variables are entered, one at a time, to see if they now become significant. The final variable set at this step is then joined by later occurring variables in the next step. We add the next set of variables (e.g., neonatal; epoch 2). No variable from the epoch 1 set can be displaced. We drop non-significant variables only from the newer set, and then try adding back as before, until we reach the point where all newer-set variables are significant. This is repeated for later occurring sets, say, early postnatal, late postnatal, etc. Only variables from the most recent in time set can be dropped. All variables brought forward from early epochs remain [24–27]. We used a step down procedure within each set seeking a parsimonious solution without interaction terms. The contributions of relevant variables are presented as risk ratios with 95% confidence intervals where the referent group for the risk ratios consists of the infants without the antecedent/covariate.
Results
Of the 1320 infants who met entry requirements, 67 (5.1%) developed medical mNEC, 80 (6.1%) developed surgical sNEC and 52 (3.9%) developed SIP. We present our results, utilizing bivariate and multivariate analysis, for distinct categories.
Increased risk of medical NEC
The risk of developing mNEC was higher for infants born to mothers who had singleton gestation (Table 1: 6 v 3%), identified as Black (Table 2: 8 v 2–4%), acknowledged smoking during this pregnancy (Table 3: 8 v 4%), were exposed to second-hand smoke (Table 3: 7 v 4%), had vaginal bleeding after the 12th week of gestation (Table 3: 7 v 4%), had periodontal disease during this pregnancy (Supplementary Appendix Table A: 9 v 5%) or pneumonia (13 v 5%), consumed aspirin(12 v 5%), (less so if they consumed an NSAID(8 v 5%), did not receive any corticosteroid (Supplementary Appendix Table C: 9 v 5%), gave birth prematurely because of placenta abruption (9 v 2–7%).The infant was at increased risk if the PCO2 was in the lowest quartile on two or more of the first 3 postnatal days (8 v 4%), had presumed or definite late bacteremia (8 and 7 v 4%), and during the first postnatal week received packed red blood cells (6 v 3%) or albumin (9 v 5%).
Decreased risk of medical NEC
Reduced risk was associated with no vaginal bleeding in the first 12 weeks of gestation (Table 3: 3 v 6%), histologic documentation of fetal stem vessel thrombosis (Table 4: 1 v 5%), and with Cesarean section delivery (Supplementary Appendix Table C: 4 v 7%).
Increased risk of surgical NEC
The risk of developing sNEC was higher for infants born to mothers who identified as Hispanic (Table 2: 9 v 6%), supported themselves (7 v 4%), were obese prior to this pregnancy (Table 3: 8 v 6 and 4%), had anemia during this pregnancy (Supplementary Appendix Table A: 11 v 6%), the placenta harbored an anaerobe (Supplementary Appendix Table B: 8 v 5%) or two or more organisms (8 v 5%), had histologic evidence of fetal stem vessel thrombosis (Table 4: 9 v 6%), and delivered prematurely because of cervical insufficiency (Supplementary Appendix Table C: 9 v 6 and 5%) or for a fetal indication (12 v 6 and 5%).
The lower the gestational age, the higher the risk of surgical NEC (Table 1: 13 v 6 v 2%). Babies who weighed 750 grams or less were at greater risk than their heavier peers (10 v 4 and 3%). A SNAP-II above 20 was associated with increased risk (9 and 7 v 4%), as was presumed or definite late bacteremia (8 v 4%), mechanical-ventilation on postnatal day 7 (8 v 3%), and receipt any time during the first week of packed red blood cells (7 v 3%), fresh frozen plasma (FFP) (13 v 6%), albumin (9 v 6%), or platelets (12 v 5%).
Decreased risk of surgical NEC
Reduced risk of surgical NEC was associated with mother’s lack of acknowledging a sexually-transmitted disease during this pregnancy (Supplementary Appendix Table A: 0 v 6%), and receipt of a tocolytic agents (2 v 6%).
Increased risk of SIP
The risk of developing SIP was higher for infants born to mothers who were prescribed an anti-hypertensive drug (Supplementary Appendix Table A: 7 v 4%), their placenta harbored a vaginal organism (Supplementary Appendix Table B: 6 v 3%), had increased syncytial knots (Table 4: 6 v 3%), or if they were given a discharge diagnosis of sepsis (Supplementary Appendix Table C: 5 v 1%). The lower the gestational age and birth weight, the higher the SIP risk (Table 1: 8 v 4% and 6 v 3%, respectively). Severely growth restricted newborns were also at increased risk (7 v 4%). The risk of SIP was increased among infants who had acidemia (arterial pH in lowest quartile) on two or more of the first three postnatal days (7 v 3%), definite bacteremia during the first postnatal week (8 v 3 and 4 %), were mechanically ventilated on postnatal day 7 (6 v 1%), or during the first postnatal week received packed red blood cells (5 v 2%), FFP (13 v 3%), or platelets(9 v 3%).
Decreased risk of SIP
Decreased risk of SIP was associated with maternal identification as non-Hispanic (Table 2: 1 v 4%), no history of any conception assistance (Table 3: 2 v 5%), no pneumonia (Supplementary Appendix Table A: 0 v 4%), or no anemia during this pregnancy (1 v 4%).
We began time-oriented multinomial risk models with characteristics and exposures before delivery. (Table 5). In this model, infants were at increased risk of mNEC if the mother identified as Black(RR 3.0; CI 1.8, 5.1), acknowledged smoking during the pregnancy (RR 2.3; CI 1.2, 4.4), had vaginal bleeding after the 12th week (RR 1.9; CI 1.1, 3.3), or consumed aspirin (RR 3.0; CI 1.3, 7.0). Infants were at increased risk of sNEC if the mother did not rely on others for financial support (RR 2.0; CI 1.1, 3.6), or acknowledged anemia (RR 2.1; 1.1, 3.8).
When delivery variables were added to the model, Cesarean section delivery was associated with reduced risk of mNEC (RR 0.6; CI 0.3, 0.96), while low gestational age was associated with increased risk of both sNEC (RR 2.6; CI 1.4, 4.7) and SIP (RR 2.9, CI 1.4, 5.9). Low birth weight provided additional information about increased risk of sNEC (RR 2.1; CI 1.1, 3.8), while increased syncytial knots in the placenta were associated with increased risk of SIP (RR 2.1; CI 1.1, 4.2).
When early postnatal (week 1) variables were added to the Antenatal and Delivery model, receipt of plasma during the first week was the only one associated with both sNEC (RR 2.3, CI 1.1, 4.8) and SIP (RR 3.7, CI 1.8, 7.8). Mechanical ventilation on day 7 was prominently associated with increased risk of SIP alone (RR 4.5, CI 1.7, 12).
Discussion
The multifactorial etiology of NEC and SIP makes early identification of at risk infants challenging and this is further complicated by the timing of infant exposure to the implicated factors – antenatal, intra-partum and post-natal. In our large prospective cohort of ELGANs study, using time oriented multinomial risk models we have identified antecedent factors specifically associated with medical and surgical NEC as well as SIP. The combined incidence of mNEC and sNEC for our cohort was 11% which is similar to the nationally reported data, making our study results generalizable [28].
Infants who developed mNEC were more likely than others to be born to mothers who were Black, smoked during the pregnancy, consumed aspirin during pregnancy and had vaginal bleeding after 12th week of gestation. Not only does pregnancy induced hypertension (PIH) occur in higher proportion of Black as compared to white women but is also associated with higher morbidity and mortality [29]. Maternal smoking appears to alter gene expression and may be associated with placental villus hypoxia, which can influence angiogenesis and apoptosis [30]. The U.S. Preventive Services Task Force recommendations released in 2014 suggest using low-dose aspirin for mothers at high risk for PIH after 12 weeks of gestation [31]. However, aspirin, a known cyclooxygenase inhibitor can disturb fetal prostaglandin synthesis and apparently thereby contribute to premature fetal closure the ductus arteriosus and alteration of the coagulation profiles [32, 33]. Similarly, vaginal bleeding beyond the 12th week of gestation, either due to placental abruption or placenta previa, might be a correlate of impaired placental growth, especially in the cases of chronic placental abruption. Thus, all the factors that we found associated with mNEC, through varied mechanisms have the potential to convey information about placental function as well as hemodynamics, potentially causing compromise to gastrointestinal mucosal integrity, the very first histopathological perturbation in the sequence of events in occurrence of NEC. Our finding that delivery by C-section seemed to confer a protective effect is especially interesting in light of evidence that delivery via a cesarean section does not confer any advantage, and instead may have increased morbidity [34, 35].
On the other hand, infants who developed sNEC were born to mothers who supported themselves, were obese and anemic during the pregnancy. Data about impact of maternal Body Mass Index (BMI) on preterm pregnancy outcomes is very limited. In a large Canadian cohort, it has been demonstrated that elevated BMI resulted in increased incidence of pre-eclampsia, gestational hypertension and preterm birth [36]. The higher incidence of pregnancy induced hypertension (PIH) altering the placental function may explain the higher incidence of sNEC for a pregnancy complicated by obesity. We can only speculate that an anemic mother may have other co-existing nutritional deficiencies that adversely affect in-utero fetal growth, which again might place these infants at higher risk for NEC [28, 37]. They were also the smallest infants with gestational age≤24 weeks, birth weight ≤750gm and more likely to have received FFP during the first postnatal week. The lowest GA group (23–24 weeks), which has an over-representation of infants with low birth weight (≤750 gm), high SNAP-II, hypoxemia, hypercapnea, acidemia, and receipt of a steroid postnatally, were at the greatest risk of developing SIP or NEC, especially surgical NEC. These findings are consistent with the observation that the most important risk factor is extreme immaturity [10, 11]. Although immaturity of bowel epithelium might be the main source of endogenous vulnerability, [38] low gestational age provides information about the immaturity of virtually all systems [39, 40], making lower GA infants most susceptible.
Infants with SIP also tended to be more immature than those without any clinical bowel pathology, had placentas with increased syncytial knots and received plasma during the first postnatal week. Syncytial knots are most prominently increased in the placentas of women who had severe pre-eclampsia and considered a marker of premature aging of the placenta [41]. In the univariable analyses, infants with SIP were more likely than others to be born to a mother with a discharge diagnosis of sepsis and to have harbored an anaerobe and/or vaginal organism in the placenta. Cytokines and cytokine mediators such as Toll like receptors appear to contribute to intestinal inflammation and NEC in animal models [41–45] and have been associated with NEC and SIP in our sample [46]. During normal vaginal term birth the gut is colonized with the maternal vaginal flora, which is a vital transitional phase promoting the utilization of the newborn’s largest immune organ (gut) [47–49]. The majority of ELGANs are delivered via C-section delivery, and thus deprived of being exposed to the maternal vaginal flora. Instead they have increased exposure to pathogenic anaerobic bacteria as shown by the persistence even for the infants delivered by Cesarean section. This altered in-utero bacterial exposure may be an inciting factor in the early cascade of events leading to NEC. Additionally, antibiotic exposure in the NICU may alter the gut milieu promoting colonization by other pathogenic bacteria. Altered colonization has been the basis for trials of probiotic and prebiotic use for prevention of NEC [50].
Both mNEC and sNEC were also associated with an indicator of maternal low socioeconomic. For mNEC it was self-identification as Black, while for sNEC it was self-support during the pregnancy. We are aware of only one previous report that low socioeconomic level is a risk factor for NEC onset [51].
We have identified specific antenatal and early postnatal factors that may alter the infants’ susceptibility to developing NEC. The presence/absence of these factors associated with a given subtype of NEC may allow the clinicians to have greater awareness of an at risk infant in their practice thereby allowing for greater vigilance and early intervention when subtle symptoms appear. This may, in turn, either lower the occurrence of the disease or have lesser morbidity/mortality due to earlier identification and intervention.
Our study has several strengths. First, we included a large number of infants, making it unlikely that we have missed important associations due to lack of statistical power, or claimed associations that might reflect the instability of small numbers. Second, we selected infants based on gestational age, not birth weight, to minimize confounding due to factors related to fetal growth restriction. Third, we collected all of our data prospectively. On the other hand, a weakness of our study is that the sickest infants were probably treated more aggressively than other infants who were not quite so sick in our study population, making our study prone to confounding by indication. We are also limited by the observational nature of our study.
Conclusion
In conclusion, although many antecedents and correlates of NEC and SIP were identified by bivariate analyses, our time-oriented multivariate analyses identified a smaller set of specific antecedents/correlates. For medical NEC these were maternal Black race, tobacco and aspirin consumption in pregnancy, as well as vaginal bleeding beyond the 12th week of gestation. Infants with surgical NEC and SIP were the most immature (as identified here by birth before the 25th week of gestation, and receipt of, or perhaps need for plasma during the first week). Maternal obesity and anemia were the other antenatal antecedents of surgical NEC, while the only other antenatal antecedent of SIP was the presence of increased syncytial knots in the placenta. Mechanical ventilation on day 7 or the need for such assistance conveyed additional information about an increased risk of SIP. Future studies are needed for predictive modeling for NEC utilizing the identifiable risk factors to identify infants at highest risk for NEC and thereby administer appropriate therapies for prevention of NEC and SIP.
Financial disclosure
The authors do not have any financial or other conflict of interest disclosures to make.
Human research statement
Appropriate IRB approvals as well as informed parental consents were obtained prior to conducting any research related activity.
Other disclosures
This study was supported by the National Institute of Neurological Diseases and Stroke (5U01NS040069-05 and 2R01NS040069 - 06A2) and the National Institute of Child Health and Human Development (5P30HD018655-28).
Footnotes
Acknowledgments
The authors gratefully acknowledge the contributions of their subjects, and their subjects’ families, as well as those of their colleagues.
Participating institutions (site principal investigator and colleagues)
Baystate Medical Center, Springfield MA (Bhavesh Shah, Karen Christianson); Beth Israel Deaconess Medical Center, Boston MA (Camilia R. Martin, Colleen Hallisey, Caitlin Hurley, Miren Creixell); Brigham & Women’s Hospital, Boston MA (Linda J. Van Marter); Children’s Hospital, Boston MA (Alan Leviton, Kathleen Lee, Anne McGovern, Elizabeth Allred, Jill Gambardella, Susan Ursprung, Ruth Blomquist); Massachusetts General Hospital, Boston MA (Robert Insoft, Jennifer G. Wilson, Maureen Pimental); New England Medical Center, Boston MA (Cynthia Cole, John Fiascone, Janet Madden, Ellen Nylen, Anne Furey); U Mass Memorial Health Center, Worcester, MA (Francis Bednarek[deceased],Mary Naples, Beth Powers); Yale-New Haven Hospital, New Haven CT (Richard Ehrenkranz, Joanne Williams, Elaine Romano); Forsyth Hospital, Baptist Medical Center, Winston-Salem NC (T. Michael O’Shea, Debbie Gordon, Teresa Harold, Gail Hounsell, Debbie Hiatt); University Health Systems of Eastern Carolina, Greenville NC (Stephen Engelke, Sherry Moseley, Linda Pare, Donna Smart, Joan Wilson); North Carolina Children’s Hospital, Chapel Hill NC (Carl Bose, Gennie Bose, Janice Wereszczak); DeVos Children’s Hospital, Grand Rapids MI (Mariel Portenga, Dinah Sutton); Sparrow Hospital, Lansing MI (Padmani Karna, Carolyn Solomon); University of Chicago Hospital, Chicago IL (Michael D. Schreiber, Grace Yoon); William Beaumont Hospital, Royal Oak MI (Daniel Batton, Beth Kring).