Main

Necrotizing enterocolitis (NEC), a gastrointestinal emergency, affects ∼5 to 10% of very LBW (VLBW) infants. It is a devastating disease with mortality as high as 30%. Prematurity is the greatest risk factor for development of NEC (1,2). Several studies suggest that the initiation of an intense systemic and local inflammatory cascade leads to intestinal necrosis in response to inciting risk factors (3–8).

Ureaplasma parvum and Ureaplasma urealyticum are commensals of the genital tract of 40–80% childbearing aged women (9,10) and are the most common organisms isolated from infected amniotic fluid and placentas (11). Infertility, chorioamnionitis, preterm delivery, and morbidity such as bronchopulmonary dysplasia (BPD) have all been associated with perinatal Ureaplasma infection (12). The organisms elicit both systemic and local host inflammatory responses in humans (13,14) and in cell (15) and animal models (16–18). The intestinal and respiratory tracts are directly exposed to infected amniotic fluid containing inflammatory mediators, which could enhance the inflammatory response to certain bacteria and their products. In addition to being isolated from the respiratory tract, Ureaplasma has been detected in gastric aspirates by culture (19–21) and molecular methods (22) and in rectal cultures (21). The effects of such synergistic inflammatory interactions could be potentially detrimental to the preterm host leading to a compromised intestinal barrier with development of diseases such as NEC and gastrointestinal-related sepsis.

Although preterm respiratory colonization with Ureaplasma is a known risk factor for neonatal morbidities, its association with NEC has not been previously determined. We hypothesized that preterm infants exposed to Ureaplasma spp. in utero or colonized at birth are at increased risk for NEC. To evaluate the relationship of Ureaplasma colonization with NEC, we examined the incidence and associated clinical and inflammatory variables of NEC in two prospectively recruited cohorts of preterm infants with Ureaplasma colonization status during the first week of life confirmed by culture and PCR who were born at GA <33 wk and birth weight <1501 g.

METHODS

Sample.

Infants born at GA <33 wk and birth weight <1501 g admitted to the NICUs at the University of Maryland Medical Center and Mercy Medical Center (Baltimore, MD) were eligible for study participation. We enrolled patients in two studies designed to characterize the effects of Ureaplasma on preterm infant outcomes from 1999 to 2003 [cohort 1; details of this cohort have been previously reported (23,24)] and from 2007 to 2009 (cohort 2). The objective of the first study was to determine the incidence of invasive disease with U. parvum and U. urealyticum and the relationship with adverse outcomes in VLBW infants. The objective for the current study is to analyze potential single nucleotide polymorphisms in relevant toll-like receptor genes associated with risk for Ureaplasma respiratory tract colonization and BPD. For both studies, infants were excluded if they had confirmed diagnoses of congenital brain/neural tube defects or congenital viral infections. Parental consent was obtained, and the institutional review boards of both institutions approved the study protocols.

NEC assessment.

Cases of stage ≥2 NEC according to the modified Bell criteria (1,25) were confirmed by typical radiological findings (pneumatosis intestinalis, portal venous air, pneumoperitoneum, and/or fixed intestinal loop) and/or pathology, if available, and were classified as medical or surgical NEC. A radiologist blinded to Ureaplasma culture status reviewed all abdominal radiographs of suspected NEC cases. Cases that were confirmed by pathological examination as spontaneous intestinal perforation (n = 2) were excluded. Postnatal age, presence or absence of feeding, and white blood cell and absolute neutrophil counts at birth and onset of NEC were recorded.

Ureaplasma detection.

One or more tracheal or nasopharyngeal aspirates were obtained during the first week of life from enrolled infants. Samples were processed, 10-fold serially diluted in 10 B broth to 10−4 and incubated at 37°C in humidified 5% CO2. Dilutions in which a color change occurred were inoculated on A8 agar and incubated at 37°C in humidified 5% CO2. Cultures were examined daily for 1 wk for color change or colonies typical of Ureaplasma (26). DNA was extracted from original tracheal aspirate or nasopharyngeal samples and culture-positive isolates using QiAmp DNA Blood Mini kits (Qiagen, Valencia, CA) according to the manufacturer's protocol. PCR for cohort 1 was performed as previously described with primers directed against the 5′ region of the multiple-banded antigen (MBA) gene to identify all positive samples and primers targeting urease gene to identify species (24). For cohort 2, DNA samples were analyzed by multiplex real-time PCR to differentiate the two Ureaplasma species simultaneously as previously described using the Roche LightCycler 2.0 (27).

Serum cytokines.

For the first cohort, cord serum samples were analyzed for IL-6 and IL-1β in duplicate samples by standard two antibody ELISA using commercial antibody pairs and recombinant standards (Endogen, Boston, MA) as previously described (23). A curve fitted to standards was generated using a computer program (Softpro: Molecular Devices), and cytokine concentrations from each sample were calculated from the standard curve. Assay sensitivities were 1.5 and 0.78 pg/mL for IL-6 and IL-1β, respectively.

Placental pathology.

Placental studies were performed on 197 of 232 (85%) subjects with confirmed Ureaplasma respiratory status of the first cohort. Sections of umbilical cord, membrane roll, placental disc near the cord insertion site, and the midpoint between cord insertion and the periphery of the placental disc were formalin-fixed, paraffin-embedded, and hematoxylin- and eosin-stained. A pathologist blinded to maternal and infant clinical status reviewed the sections. Histologic chorioamnionitis was separated into maternal and fetal involvement and a stage assigned based on the scheme proposed by Redline et al. (28). Fetal vasculitis was defined as polymorphonuclear infiltration of the chorionic vessels or umbilical cord (28).

Statistical analysis.

The t test and ANOVA was used to compare continuous normally distributed data and Mann-Whitney or Kruskal-Wallis test for nonnormally distributed data. The χ2 or Fisher exact test was used to compare categorical variables. Univariate ORs and 95% CIs were calculated for all variables for NEC outcome. Analyses comparing Ureaplasma-positive and negative infants were stratified by NEC status. Statistical analysis was performed using STATA 7.0 (Stata Corp., College Station, TX). A p < 0.05 was considered significant.

RESULTS

Study cohort characteristics.

For cohort 1, NEC status was confirmed for 308 of 313 subjects, and Ureaplasma respiratory status was available on 232 of 308 (75%). For cohort 2, of 324 infants <33 wks gestation who were eligible for the study, 20 were missed because of the lack of parental contact, 168 declined consent, and parental consent was obtained for the remaining 136 infants. NEC and Ureaplasma respiratory status were available for all cohort 2 enrolled subjects. The incidence of NEC was similar for both cohorts [cohort 1, 15/232 (6.5%); cohort 2, 14/136 (10.3%); p = 0.229] and did not differ from the NEC rate for nonenrolled infants during the study periods. The combined NEC rate was 29 of 368 (7.9%). Ureaplasma respiratory tract colonization rate was also similar for both cohorts [cohort 1, 75/232 (32%); cohort 2, 57/136 (42%); p = 0.064] with an overall colonization rate of 132 of 368 (36%). Ureaplasma parvum was the predominant species (67%) compared with U. urealyticum (27%). Both species were present in 6% specimens. For all subsequent analyses, the cohorts were combined.

Ureaplasma respiratory tract colonization and NEC in VLBW infants.

We first analyzed the relationship of demographic, antenatal, and early neonatal factors with NEC. In the combined cohorts, none of the factors included in analyses were significantly associated with NEC (Table 1). Specific details concerning feeding such as age when feedings were started, composition of feeds, or time to full feeds were not recorded. However, all infants were fed according to an established feeding protocol.

Table 1 Association of clinical variables and NEC
Full size table

Ureaplasma-colonized infants were less mature and experienced a higher rate of preterm premature rupture of the membranes, maternal antibiotic exposure, and longer duration of mechanical ventilation but a lower rate of pregnancy-induced hypertension than noncolonized infants regardless of whether they developed NEC (Table 2). The birth weights of Ureaplasma-positive infants were lower than the birth weights of the Ureaplasma-negative infants in the non-NEC group. However, the incidence of NEC was 2.1-fold higher in Ureaplasma-positive (12.3%) than Ureaplasma-negative infants (5.5%) <33 wk (OR, 2.43; 95% CI, 1.13–5.22; p = 0.023; Table 1) and 3.3-fold higher in Ureaplasma-positive (14.6%) than Ureaplasma-negative (4.4%) infants ≤28 wks (OR, 3.67; 95%CI, 1.36–9.93). When adjusted for GA, the association of Ureaplasma colonization and NEC remained significant (OR, 2.47; 95% CI, 1.13–5.43). Inclusion of other clinical variables in the logistics model did not affect the estimate of the association of Ureaplasma colonization and NEC. There were no differences in NEC rates between the Ureaplasma species. Age of onset, hematologic parameters at onset, NEC severity, and mortality were similar between Ureaplasma-positive and -negative NEC infants (Table 3). All Ureaplasma-positive NEC infants had been fed before onset compared with 92% Ureaplamsa-negative NEC infants, but this difference was not statistically significant.

Table 2 Obstetric and neonatal characteristics of study cohort
Full size table
Table 3 Characteristics of NEC infants with and without Ureaplasma respiratory tract colonization
Full size table

Inflammatory markers associated with NEC in Ureaplasma-colonized VLBW infants.

As shown in Table 4, Ureaplasma-colonized infants had significantly higher admission peripheral white blood cell counts and absolute neutrophil counts regardless of NEC status. Although histologic chorioamnionitis was present in 92% placentas from Ureaplasma non-NEC and 100% Ureaplasma NEC infants, inflammation was detected in 1 of 7 (14%) placentas from Ureaplasma-negative NEC infants (p = 0.052). Similarly, fetal vasculitis was present in 79% placentas from Ureaplasma non-NEC and 67% Ureaplasma NEC infants but absent in all placentas available for review from Ureaplasma-negative NEC infants. When restricted to the subset with placental pathology, histologic chorioamnionitis in the absence of Ureaplasma colonization tended to reduce the risk for NEC (OR, 0.524; 95% CI, 0.178–1.44; p = 0.202). Because Ureaplasma colonization rarely occurred in the absence of histologic chorioamnionitis, it was not possible to distinguish the relative contribution of each variable to NEC.

Table 4 Markers of inflammation in infants with and without NEC
Full size table

Cord serum cytokine measurements were available for 101 of 232 (44%) subjects of cohort 1. Cord serum IL-6 and IL-1β concentrations were similar in NEC and non-NEC groups (Table 1) but were significantly higher in Ureaplasma-positive than in Ureaplasma-negative infants. The highest cytokine concentrations were detected in cord blood samples of Ureaplasma-positive NEC infants (Table 4).

DISCUSSION

The two cohorts experienced similar rates of NEC, suggesting that the rate of the disease has been stable in our NICUs over time. The overall rate of 7.9% is within the range of confirmed NEC rates for VLBW infants reported by the National Institute of Child Health and Human Development Neonatal Research Network (10.1%) (1) and the Vermont Oxford Network (6.9%) (2).

Because NEC is primarily a disease of prematurity, immaturity of gut barrier function and local and systemic immune responses have been implicated in susceptibility to the disease. Recently, more attention has focused on the potential role of the intestinal microbiota in initiating mucosal injury and modulating expression of virulence factors and host immune responses (29). Although many bacterial species and enteric viruses have been reported in association with NEC (29), a causal role for these organisms has not been established. This is the first study to demonstrate an association of Ureaplasma respiratory tract colonization and NEC. Although Ureaplasma was only cultured from respiratory secretions, Ureaplasma spp. are known mucosal organisms that colonize the adult genitourinary tract (10) and have been previously recovered from other mucosal sites such as gastric aspirates and rectum in preterm infants (21,30). The observed higher rate of NEC in Ureaplasma-positive than negative infants ≤28 wk gestation supports the contention that immaturity of intestinal functions increases the susceptibility to NEC in very preterm infants perinatally exposed to Ureaplasma infection/inflammation.

Using culture techniques, Ureaplasma spp. have been isolated from blood, cerebrospinal fluid, tracheal aspirates, and lung and brain tissue of newborn infants (11,31–33). Epidemiologic studies and experimental infection models support an etiologic role for Ureaplasma infection or resulting inflammation in preterm birth and several neonatal morbidities. Although the association of Ureaplasma respiratory tract colonization with the development of BPD in preterm infants has been debated, a recent meta-analysis of 31 studies supported this association (34). Experimental antenatal infection models in mice (35), immature sheep (36), Rhesus macaque (37), and baboon (17) confirm that in utero exposure to Ureaplasma infection causes fetal/newborn lung inflammation and altered lung development. In our first study of cohort 1, we observed that Ureaplasma species not only colonize the respiratory tract but also invade the bloodstream and cross the immature blood-brain barrier in 23% VLBW infants (24). Detection of Ureaplasma by PCR in serum, but not cerebrospinal fluid, increased the risk for severe intraventricular hemorrhage 2-fold (24). In a mouse model of antenatal Ureaplasma infection, neuronal injury and microgliosis were evident in Ureaplasma antenatally infected pups (35).

There is compelling data from human studies and animal models that Ureaplasma is proinflammatory in multiple compartments (amniotic fluid, placenta, fetal lung, and brain). The stimulatory effect of Ureaplasma on cytokine release has been confirmed in vitro. In cultured human cord blood preterm monocytes, Ureaplasma stimulated release of TNF-α and IL-8, and when coadministered with Gram-negative lipopolysaccharide, Ureaplasma greatly augmented generation of proinflammatory cytokines while blocking expression of the counter regulatory cytokines, IL-6 and IL-10 (15). In the current study, Ureaplasma-positive infants were more likely exposed to chorioamnionitis and to express a systemic inflammatory response (fetal vasculitis, elevated admission white blood cell and absolute neutrophil counts, and cord blood IL-6 and IL-1β), suggesting that inflammation was initiated in utero. This is consistent with recent evidence that in the setting of preterm premature rupture of membranes, intra-amniotic infection with the genital mycoplasmas is associated with a more intense inflammatory response compared with the response to infections with other microorganisms (38). Antenatal exposure to infection/inflammation may predispose the developing intestinal mucosa to subsequent injury or dysregulated inflammatory responses. Previous studies have linked presence of amniotic fluid infection/elevated cytokines (39), cord blood cytokines (40,41), and fetal vasculitis (42) with risk for NEC in preterm infants. In a rat model of NEC, maternal prenatal exposure to lipopolysaccharide led to increased frequency and severity of intestinal injury (43). Taken together, these observations suggest that intestinal injury may be initiated in utero. Hematologic parameters and postnatal age at NEC onset did not differ between Ureaplasma-positive and -negative infants, suggesting that other postnatal factors are necessary for disease progression such as initiation of enteral feeds, prolonged exposure to antibiotics (44) or H2-blockers (45), or change in the intestinal microbiome.

There are several limitations of this study. Ureaplasma respiratory tract colonization in the first week of life was used as a proxy for intestinal mucosal exposure to this organism. Because the primary outcomes of the studies analyzed for this report were BPD and CNS outcomes, cultures were not obtained at the time of NEC onset. We also cannot exclude that infants were exposed antenatally to other microbes that may have altered intestinal permeability or the local immune response (46). Although the duration of ruptured membranes exceeded 1 h in the majority of Ureaplasma-positive infants, indicating vertical transmission likely occurred via an ascending infection, the duration of exposure to the organism before delivery is unknown. Bacterial load of Ureaplasma that correlates with severity of intrauterine inflammation (47,48) may be an important variable that was not measured in the current study.

Whether there is a causal relationship between perinatal Ureaplasma colonization/infection and NEC pathogenesis is currently unknown, but this can be addressed in in vitro and animal NEC models in future studies. Molecular methods may improve the detection of these organisms in relevant specimens such as gastric aspirates, stool, and surgical specimens (22) to confirm this association.

This study identifies Ureaplasma respiratory tract colonization, a marker of in utero infection/inflammation exposure, as a possible risk factor for NEC in VLBW infants. Whether Ureaplasma directly contributes to intestinal mucosal injury or alters the local immune response is unknown. Future experimental cell and animal models may determine how Ureaplasma contributes to NEC pathogenesis.