Prenatal exposure to intra-amniotic infection with Ureaplasma species increases the prevalence of bronchopulmonary dysplasia

Abstract

Objectives

The present study investigated the relationship between bronchopulmonary dysplasia (BPD) and intra-amniotic infection with Ureaplasma species.

Methods

This was a single-center, retrospective cohort study. Patients with singleton pregnancies who underwent inpatient management at our department for preterm premature rupture of membranes (PPROM), preterm labor, cervical insufficiency, and asymptomatic cervical shortening at 22–33 gestational weeks were included. Amniocentesis was indicated for patients with PPROM or an elevated maternal C-reactive protein level (≥0.58 mg/dL). Patients with an amniotic fluid IL-6 concentration ≥3.0 ng/mL were diagnosed with intra-amniotic inflammation, while those with positive aerobic, anaerobic, M. hominis, and Ureaplasma spp. cultures were diagnosed with microbial invasion of the amniotic cavity (MIAC). Patients who tested positive for both intra-amniotic inflammation and MIAC were considered to have intra-amniotic infection. An umbilical vein blood IL-6 concentration >11.0 pg/mL indicated fetal inflammatory response syndrome (FIRS). The maternal inflammatory response (MIR) and fetal inflammatory response (FIR) were staged using the Amsterdam Placental Workshop Group Consensus Statement.

Results

Intra-amniotic infection with Ureaplasma spp. was diagnosed in 37 patients, intra-amniotic infection without Ureaplasma spp. in 28, intra-amniotic inflammation without MIAC in 58, and preterm birth without MIR/FIR and FIRS in 86 as controls. Following an adjustment for gestational age at birth, the risk of BPD was increased in patients with intra-amniotic infection with Ureaplasma spp. (adjusted odds ratio: 10.5; 95% confidence interval: 1.55–71.2), but not in those with intra-amniotic infection without Ureaplasma spp. or intra-amniotic inflammation without MIAC.

Conclusion

BPD was only associated with intra-amniotic infection with Ureaplasma species.

Keywords:

• Chorioamnionitis
• interleukin-6
• intra-amniotic inflammation
• microbial invasion of the amniotic cavity
• preterm birth
• preterm premature rupture of membranes

Introduction

Bronchopulmonary dysplasia (BPD) is a severe complication in preterm infants; however, its incidence remains high and has not markedly decreased in the past 20 years [1,2]. The difficulty of reducing BPD is attributed to prenatal adverse factors and its multifactorial origin [3]. Histological chorioamnionitis has been the focus of much attention concerning BPD.

However, the relationship between BPD and chorioamnionitis significantly differs between studies [4–7]. The developing fetus in patients with chorioamnionitis is exposed to intra-amniotic inflammation in the absence or presence of various organisms [8]. The type and intensity of this inflammation may affect the subsequent development of BPD [9,10]. Respiratory colonization by Ureaplasma species (spp.) has been implicate in the development of BPD in neonates [11–14], and the relationship between Ureaplasma spp. in amniotic fluid and BPD has been investigated. Previous studies reported that Ureaplasma spp. in amniotic fluid was associated with the development of BPD [15,16], whereas others did not [17,18]. A meta-analysis was performed and showed that Ureaplasma spp. in amniotic fluid was associated with BPD [19]. However, a multivariate analysis has not yet been conducted to account for confounding factors, such as gestational age at birth. The relationship between intra-amniotic inflammation without MIAC and BPD also remains unknown.

A recent study reported that antenatal antibiotic therapy with azithromycin intravenously administered for preterm premature rupture of membranes (PPROM) attenuated BPD [20]. Therefore, it is crucial to identify and treat prenatal adverse factors associated with BPD. The present study investigated the relationships between BPD and intra-amniotic infection with and without Ureaplasma spp. as well as intra-amniotic inflammation without MIAC.

Materials and methods

This was a single-center retrospective cohort study. All procedures were performed in accordance with the ethical standards of the Institutional Review Board of the National Hospital Organization, Saga National Hospital (approval number: R4-8) and the Declaration of Helsinki and its later amendments.

Patients with singleton pregnancies (between August 2014 and April 2020) who underwent inpatient management at our department for PPROM, preterm labor, cervical insufficiency, and asymptomatic cervical shortening between 22 weeks 0 days and 33 weeks 6 days of gestation were included. Preterm labor was defined as regular uterine contractions (six or more in 60 min). Cervical insufficiency was diagnosed as mid-trimester painless cervical dilatation with membranes visible through the cervical os and intact membranes confirmed by a sterile speculum examination. In our department, pregnant women whose cervical length was <15 mm before 34 gestational weeks were admitted to hospital for management, even if asymptomatic.

Exclusion criteria were as follows: (1) delivery after 34 gestational weeks, (2) MIAC in which Ureaplasma spp. and Mycoplasma hominis were not identified, (3) intrauterine fetal death, (4) multiple congenital malformations, (5) delivery at another hospital, (6) positive MIR/FIR, FIRS, an unclear placental pathology or umbilical vein blood interleukin (IL)-6 concentration in patients in whom intra-amniotic inflammation or MIAC was not detected by amniocentesis or for whom amniocentesis was not indicated.

All patients with PPROM and suspected intra-amniotic inflammation with intact membranes who had preterm labor, cervical insufficiency, and asymptomatic cervical shortening in addition to an elevated maternal concentration of C-reactive protein (≥0.58 mg/dL) and who were being followed up for intra-amniotic inflammation underwent amniocentesis as part of their clinical management. All patients provided written informed consent. Amniocentesis was performed as follows. The abdominal wall was disinfected and a 5-ml sample of amniotic fluid was withdrawn under transabdominal ultrasound using a 25-G percutaneous transhepatic cholangiography needle. One milliliter of amniotic fluid was immediately subjected to measurements of IL-6 concentrations by a chemiluminescent enzyme immunoassay with a detection range of 0.2–1000 pg/mL (Human IL-6 CLEIA Fujirebio, Fujirebio, Tokyo, Japan). The remaining 4 ml was used in routine bacterial aerobic and anaerobic cultures as well as cultures for M. hominis and Ureaplasma spp. with Urea-Arginine LYO2 (BioMerieux, France). A molecular biological analysis was performed to identify M. hominis, Ureaplasma parvum, and U. urealyticum [21].

Maternal corticosteroids were administered to patients diagnosed with PPROM at <34 gestational weeks and to those for whom a premature birth was anticipated at <34 gestational weeks. Figure 1 shows the protocols for the antibiotic treatment of PPROM and preterm labor, cervical insufficiency, and asymptomatic cervical shortening. Ampicillin + Clarithromycin was administered for 7 days after PPROM. If an increase in the amniotic fluid IL-6 concentration or a positive Gram stain was observed, Sulbactam/Ampicillin (SBT/ABPC) + Azithromycin (AZM) was administered and changed according to culture results. In cases with intact membranes, SBT/ABPC + AZM was only administered when the amniotic fluid IL-6 concentration was >11.3 ng/mL, bacterial cultures were positive, or bacteria were confirmed under a microscope and changed according to culture results.

Figure 1. Protocols for the antibiotic treatment of preterm premature rupture of membranes and preterm labor, cervical insufficiency, and asymptomatic cervical shortening.

AMF: amniotic fluid; IL-6: interleukin-6; Regimen 1: ampicillin (ABPC) 2 g IV q6h for 2 days followed by amoxicillin 250 mg PO q6h for 5 days and clarithromycin 200 mg PO q12h for 7 days, Regimen 2: sulbactam/ABPC 1.5 g IV q6h daily and azithromycin 500 mg IV q24h for 3/7 days.

Patients with intact membranes received tocolytic therapy, whereas those with PPROM did not. However, tocolytic agents were discontinued in patients with intact membranes after intra-amniotic infection or markedly elevated amniotic fluid IL-6 concentrations. Indications for delivery depended on the onset of labor, clinical chorioamnionitis, a non-reassuring fetal status, and the exacerbation of intra-amniotic infection. Umbilical vein blood samples were withdrawn via venipuncture of clamped umbilical cords. The placenta, fetal membranes, and umbilical cord were then fixed in 10% neutral buffered formalin.

Confirmed Triple I was diagnosed when fever was present with biochemical or microbiological amniotic fluid results consistent with microbial invasion of the amniotic cavity [22]. Patients with an amniotic fluid IL-6 concentration ≥3.0 ng/mL were diagnosed with intra-amniotic inflammation [23], while those with positive aerobic, anaerobic, M. hominis, and Ureaplasma spp. cultures were diagnosed with MIAC. Intra-amniotic infection was defined as positivity for both intra-amniotic inflammation and MIAC [24], while an umbilical vein blood IL-6 concentration >11.0 pg/mL indicated fetal inflammatory response syndrome (FIRS) [25]. The maternal inflammatory response (MIR) and fetal inflammatory response (FIR) were staged as described by Redline et al. and verified by the Amsterdam Placental Workshop Group Consensus Statement [26,27]. Patients were classified into the following groups: (1) intra-amniotic infection with Ureaplasma spp., (2) intra-amniotic infection without Ureaplasma spp., (3) intra-amniotic inflammation without MIAC, and (4) preterm birth without MIR/FIR and FIRS.

Neonatal mortality and morbidities were defined as follows. BPD was defined according to the 2001 definition of the National Institute of Child Health and Human Development [28]. Mild BPD was diagnosed in infants treated with supplemental oxygen for at least 28 days after birth, moderate BPD in those receiving <30% oxygen at 36 weeks or discharge, and severe BPD in those treated with ≥30% oxygen or positive pressure ventilation at 36 weeks or discharge. In infants with a gestational age ≥32 weeks, the severity of BPD was assessed 56 days after birth or at discharge from the hospital. Respiratory distress syndrome was diagnosed based on the neonate’s clinical presentation of symptoms and characteristic radiographic signs [29]. Cranial ultrasonography or cranial magnetic resonance imaging was performed to diagnose intraventricular hemorrhage, which was classified using the system proposed by Papile et al. [30]. Cranial ultrasonography and cranial magnetic resonance imaging were employed to diagnose periventricular leukomalacia. Early- and late-onset neonatal sepsis were defined as culture-proven or clinical sepsis with an onset after <72 h of life and ≥72 h of life, respectively. Infant mortality was defined as in-hospital death prior to discharge. Adverse neonatal outcomes were neonatal or infant mortality in a hospital or a diagnosis of BPD, periventricular leukomalacia, intraventricular hemorrhage, or early- or late-onset neonatal sepsis [31].

Primary outcomes in the present study were the relationships between the prevalence of BPD and intra-amniotic infection with or without Ureaplasma spp. as well as intra-amniotic inflammation without MIAC after an adjustment for gestational age at birth. Secondary outcomes were the relationships between the prevalence of adverse neonatal outcomes, intra-amniotic infection with or without Ureaplasma spp., and intra-amniotic inflammation without MIAC after adjustments for potential confounders.

Continuous and categorical variables are shown as medians (interquartile ranges (IQR)) and numbers (percentages), respectively. The non-parametric Kruskal-Wallis test compared the clinical characteristics and short-term neonatal morbidity rates of the four patient groups. Fisher’s exact test compared categorical variables. Post-hoc tests for categorical and continuous variables were the Bonferroni correction and Steel-Dwass test, respectively.

Following adjustments for potential confounders, the relationships between BPD, adverse neonatal outcomes, intra-amniotic infection with and without Ureaplasma spp., and intra-amniotic inflammation without MIAC were investigated by a logistic regression analysis. Gestational age at delivery was included in adjustments for analyses of BPD [32], and gestational age at delivery, severe birth asphyxia, and FIRS in adjustments for analyses of adverse neonatal outcomes from background knowledge [31]. The assumption of 8–10 events per predictor variable was used to select the sample size for a reliable multivariate logistic regression analysis [33].

P values <0.05 indicated a significant difference. EZR software (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R software (The R Foundation for Statistical Computing, Vienna, Austria) [34], was used for statistical analyses.

Results

A total of 209/347 (60%) singleton pregnancies that received inpatient management at our department for PPROM, preterm labor, cervical insufficiency, and asymptomatic cervical shortening were included in the present study (Figure 2). Intra-amniotic infection with Ureaplasma spp. was diagnosed in 37 (18%) patients, intra-amniotic infection without Ureaplasma spp. in 28 (13%), intra-amniotic inflammation without MIAC in 58 (28%), and preterm birth without MIR/FIR and FIRS in 86 (41%).

Figure 2. Flow diagram of patient selection in the present study.

PPROM: preterm premature rupture of membranes; MIAC: microbial invasion of the amniotic cavity; MIR: maternal inflammatory response; FIR: fetal inflammatory response; FIRS: fetal inflammatory response syndrome; IL-6: interleukin-6.

Gestational age at amniocentesis was lower in patients with intra-amniotic infection with Ureaplasma spp. than in those with intra-amniotic infection without Ureaplasma spp. (median, 28.0, IQR, 26.4-31.3 vs. median, 32.1, IQR, 30.9-33.1, p = 0.02). The interval from amniocentesis to birth was longer in patients with intra-amniotic infection with Ureaplasma spp. than in those with intra-amniotic infection without Ureaplasma spp. and with intra-amniotic inflammation without MIAC (median, 3.0 days, IQR, 1.0-7.0 vs. median, 1.0 days, IQR, 0.0-2.0, p = 0.03, median, 3.0 days, IQR, 1.0-7.0 vs. median, 1.0 days, IQR, 0.0-3.0, p = 0.01, respectively) (Table 1).

Table 1. Maternal demographic and clinical characteristics of study patients.

Factor	Preterm birth without MIR/FIR and FIRS	Intra-amniotic infection with Ureaplasma spp.	Intra-amniotic infection without Ureaplasma spp.	Intra-amniotic inflammation without MIAC
n	86	37	28	58	p-Value	p-Value
	A	B	C	D		A vs. B	A vs. C	A vs. D	B vs. C	B vs. D	C vs. D
Maternal age at admission (years)	32.0 [28.0, 35.0]	32.0 [30.0, 35.0]	31.5 [29.8, 34.3]	30.0 [25.3, 33.0]	0.02	1.0	1.0	0.23	1.0	0.02	0.67
Primipara (%)	23 (26.7)	20 (54.1)	9 (32.1)	33 (56.9)	0.001	0.03	1.0	0.003	0.39	1.0	0.16
Previous preterm birth (%)	13 (15.1)	3 (8.1)	3 (10.7)	9 (15.5)	0.68
Gestational age at amniocentesis (weeks)	N.A.	27.4 [24.9, 31.0]	29.6 [27.5, 32.6]	28.7 [25.3, 31.7]	0.03	N.A.	N.A.	N.A.	0.02	0.96	0.19
Interval from amniocentesis to birth (days)	N.A.	3.00 [1.00, 7.00]	1.00 [0.00, 2.00]	1.00 [0.00, 3.00]	0.005	N.A.	N.A.	N.A.	0.03	0.01	1.0
Amniotic fluid IL-6 (ng/mL) at diagnosis	N.A.	29.22 [7.51, 74.46]	23.39 [5.34, 68.69]	10.55 [4.23, 29.70]	0.08
C-reactive protein (mg/dL) at amniocentesis	N.A.	0.98 [0.45, 2.59]	0.56 [0.18, 2.87]	0.62 [0.30, 2.02]	0.43
White blood cell (/µL) at amniocentesis	N.A.	13100 [10250, 15250]	12400 [10050, 14000]	12650 [10700, 16925]	0.55
Body temperature (°C) at amniocentesis	N.A.	36.9 [36.7, 37.1]	37.0 [36.8, 37.3]	37.1 [36.6, 37.3]	0.33
Preterm premature rupture of membranes (%)	32 (37.2)	29 (78.4)	21 (75.0)	20 (34.5)	<0.001	<0.001	0.005	1.0	1.0	<0.001	0.003
Placental abruption (%)	4 (4.7)	1 (2.7)	0 (0.0)	1 (1.7)	0.56
Diabetes mellitus (%)	0 (0.0)	3 (8.1)	0 (0.0)	1 (1.7)	0.02	0.15	1	1	1	1	1
Preeclampsia (%)	2 (2.3)	0 (0.0)	0 (0.0)	0 (0.0)	0.41
Maternal antibiotics (%)	57 (66.3)	37 (100.0)	28 (100.0)	54 (93.1)	<0.001	<0.001	<0.001	<0.001	1	0.46	0.60
Maternal corticosteroid (%)	55 (64.0)	35 (94.6)	21 (75.0)	45 (77.6)	0.004	0.002	1.0	0.59	0.19	0.25	1.0
Clinical chorioamnionitis (%)	0 (0.0)	1 (2.7)	3 (10.7)	0 (0.0)	0.002	1.0	0.08	1.0	1.0	1.0	0.19
Cesarean section (%)	38 (44.2)	16 (43.2)	12 (42.9)	25 (43.1)	1.0
Maternal inflammatory response (%)	0 (0.0)	30 (83.3)	22 (78.6)	18 (32.7)	<0.001	<0.001	<0.001	<0.001	1.0	<0.001	<0.001
Fetal inflammatory response (%)	0 (0.0)	24 (64.9)	16 (57.1)	9 (15.8)	<0.001	<0.001	<0.001	0.001	1.0	<0.001	0.002

MIR: maternal inflammatory response; FIR: fetal inflammatory response; FIRS: fetal inflammatory response syndrome; MIAC: microbial invasion of the amniotic cavity.

Neonates with intra-amniotic infection with Ureaplasma spp. had a lower gestational age at birth (p < 0.001), a higher prevalence of BPD, moderate-severe BPD, and more adverse neonatal outcomes than those without MIR/FIR and FIRS (median, 28.0, IQR, 26.4-31.3 vs. median, 32.1, IQR, 30.9-33.1, p < 0.001, 15/37 (40.5%) vs. 3/86 (3.5%), p < 0.001, 10/37 (27.0%) vs. 1/86 (1.2%), p < 0.001, 19/37 (51.4%) vs. 6/86 (7.0%), p < 0.001, respectively) (Table 2). The prevalence of neonatal sepsis was higher in neonates with intra-amniotic infection without Ureaplasma spp. than in those without MIR/FIR and FIRS (6/28 (21.4%) vs. 0/86 (0.0%), p < 0.001, respectively). Neonates with intra-amniotic inflammation without MIAC had a lower gestational age at birth (median, 29.5, IQR, 25.6-31.7 vs. median, 32.1, IQR, 30.9-33.1, p < 0.001), a higher prevalence of BPD, moderate-severe BPD, and more adverse neonatal outcomes than those without MIR/FIR and FIRS (15/58 (25.9%) vs. 3/86 (3.5%), p < 0.001, 12/58 (20.7%) vs. 1/86 (1.2%), p < 0.001, 18/58 (31.0%) vs. 6/86 (7.0%), p < 0.001, respectively). When neonates <26 weeks of gestation at birth and those ≥26 weeks were analyzed separately, BPD developed in 65% of neonates <26 weeks of gestation at birth. In neonates ≥26 weeks of gestation at birth, the prevalence of BPD and moderate-severe BPD was higher in neonates with intra-amniotic infection with Ureaplasma spp. than in those without MIR/FIR and FIRS (9/30 (30.0%) vs. 0/83 (0.0%), p < 0.001, 6/30 (20.0%) vs. 0/83 (0.0%), p = 0.001, respectively).

Table 2. Perinatal data of enrolled patients.

Factor	Preterm birth without MIR/FIR and FIRS	Intra-amniotic infection with Ureaplasma spp.	Intra-amniotic infection without Ureaplasma spp.	Intra-amniotic inflammation without MIAC
n	86	37	28	58	p-Value	p-Value
	A	B	C	D		A vs. B	A vs. C	A vs. D	B vs. C	B vs. D	C vs. D
Gestational age at birth (weeks)	32.1 [30.9, 33.1]	28.0 [26.4, 31.3]	29.8 [28.1, 32.9]	29.5 [25.6, 31.7]	<0.001	<0.001	0.08	<0.001	0.17	1.0	0.33
Male (%)	55 (64.0)	20 (54.1)	16 (57.1)	34 (58.6)	0.74
Birth weight (g)	1711.0 [1417.5, 1888.5]	1172.0 [848.0, 1530.0]	1453.0 [1058.0, 1856.5]	1288.0 [827.5, 1713.5]	<0.001	<0.001	0.42	<0.001	0.03	1.0	0.57
Small for gestational age (%)	6 (7.0)	5 (13.5)	1 (3.6)	2 (3.4)	0.24
Apgar score at 1 min	8.0 [6.0, 8.0]	6.0 [4.0, 8.0]	6.0 [3.8, 8.0]	5.5 [3.0, 8.0]	0.02	0.15	0.29	0.03	1.0	1.0	1.0
Apgar score at 5 min	9.0 [8.0, 9.0]	8.0 [7.0, 9.0]	8.0 [6.8, 9.0]	8.0 [6.0, 9.0]	<0.001	0.05	0.01	0.002	1.0	1.0	1.0
Umbilical artery pH	7.32 [7.28, 7.36]	7.32 [7.27, 7.37]	7.32 [7.30, 7.35]	7.32 [7.28, 7.36]	0.82
Umbilical artery base excess	−3.0 [-5.0, −2.0]	−3.0 [-5.3, −1.8]	−4.0 [-6.0, −2.0]	−4.0 [-6.0, −2.0]	0.40
Umbilical vein blood IL-6 (pg/mL)	2.5 [1.6, 3.7]	24.3 [4.7, 53.3]	43.2 [5.4, 241.2]	4.2 [2.4, 13.7]	<0.001	<0.001	<0.001	<0.001	1.0	0.006	0.001
Fetal inflammatory response syndrome (%)	0 (0.0)	22 (64.7)	18 (64.3)	16 (30.8)	<0.001	<0.001	<0.001	<0.001	1.0	0.02	0.03
Bronchopulmonary dysplasia (BPD) (%)	3 (3.5)	15 (40.5)	1 (3.6)	15 (25.9)	<0.001	<0.001	1.0	<0.001	0.003	1.0	0.1
<26 weeks gestational age at birth	3/3 (100)	6/7 (85.7)	0/1 (0.0)	13/16 (81.2)	0.16
≥26 weeks gestational age at birth	0/83 (0.0)	9/30 (30.0)	1/27 (3.7)	2/42 (4.8)	<0.001	<0.001	1.0	0.67	0.08	0.04	1.0
BPD (moderate - severe) (%)	1 (1.2)	10 (27.0)	1 (3.6)	12 (20.7)	<0.001	<0.001	1.0	<0.001	0.11	1.0	0.32
<26 weeks gestational age at birth	1/3 (33.3)	4/7 (57.1)	0/1 (0.0)	6/16 (62.5)	0.53
≥26 weeks gestational age at birth	0/83 (0.0)	6/30 (20.0)	1/27 (3.7)	2/42 (4.8)	<0.001	0.001	1.0	0.67	0.63	0.36	1.0
Respiratory distress syndrome (%)	20 (23.3)	6 (16.2)	9 (32.1)	14 (24.1)	0.52
Neonatal sepsis (%)	0 (0.0)	0 (0.0)	6 (21.4)	0 (0.0)	<0.001	1.0	<0.001	1.0	0.03	1.0	0.005
Intraventricular hemorrhage (%)	2 (2.3)	2 (5.4)	3 (10.7)	3 (5.2)	0.34
Periventricular leukomalacia (%)	1 (1.2)	4 (10.8)	0 (0.0)	0 (0.0)	0.003	0.17	1.0	1.0	0.77	0.12	1.0
Neonatal or infant death (%)	0 (0.0)	2 (5.4)	0 (0.0)	2 (3.4)	0.15
Adverse neonatal outcomes (%)	6 (7.0)	19 (51.4)	7 (25.0)	18 (31.0)	<0.001	<0.001	0.09	0.001	0.25	0.33	1.0

MIR: maternal inflammatory response; FIR: fetal inflammatory response; FIRS: fetal inflammatory response syndrome; MIAC: microbial invasion of the amniotic cavity, adverse neonatal outcomes, neonatal or infant mortality in a hospital or the presence of bronchopulmonary dysplasia, neonatal sepsis, periventricular leukomalacia, intraventricular hemorrhage.

Table S1 shows the bacterial species identified in amniotic fluid obtained by amniocentesis in patients with intra-amniotic inflammation. There were 6 co-infection cases. BPD occurred in 2/2 of neonates <26 weeks of gestation at birth and in 1/4 of neonates ≥26 weeks of gestation at birth.

After adjustments for gestational age at birth and potential confounders, intra-amniotic infection with Ureaplasma spp. increased the risk of BPD and adverse neonatal outcomes (adjusted odds ratio (aOR): 10.5; 95% confidence interval (CI):1.55–71.2, aOR: 6.12; 95% CI: 1.02–36.7, respectively) (Table 3, Table S2). However, intra-amniotic infection without Ureaplasma spp. and intra-amniotic inflammation without MIAC were not associated with BPD or adverse neonatal outcomes.

Table 3. Multivariate logistic regression analysis of the presence of bronchopulmonary dysplasia.

	Reference	Adjusted Odds ratio	95% CI	p-Value
Intra-amniotic infection with Ureaplasma spp.	Preterm birth without MIR/FIR and FIRS	10.5	1.55 − 71.2	0.02
Intra-amniotic infection without Ureaplasma spp.		0.7	0.05 − 10.2	0.79
Intra-amniotic inflammation without MIAC		2.39	0.37 − 15.4	0.36
Gestational age at birth	1 week	0.42	0.31 − 0.57	<0.001

MIR: maternal inflammatory response; FIR: fetal inflammatory response; FIRS: fetal inflammatory response syndrome; MIAC: microbial invasion of the amniotic cavity.

Discussion

We enrolled patients with preterm birth without intra-amniotic infection and intra-amniotic inflammation as controls and demonstrated that intra-amniotic infection with Ureaplasma spp. was associated with BPD, whereas that without Ureaplasma spp. and intra-amniotic inflammation without MIAC were not.

The relationship between Ureaplasma spp. in amniotic fluid collected by amniocentesis and the subsequent development of BPD in infants remains unclear. A previous study examined 44 patients and showed that the incidence of BPD did not significantly differ between patients with Ureaplasma-positive amniotic fluid (n = 9) and those with Ureaplasma-negative amniotic fluid (n = 29) (38% vs. 11%) or between those with positive cultures for other bacterial species (n = 6) (33% vs. 11%) [17]. Another study analyzed 38 patients and found that the incidence of severe BPD did not significantly differ among patients with M. hominis- or Ureaplasma-positive cultures, those with positive cultures for other bacterial species, and those with negative cultures (1/6 (16.7%) vs. 2/8 (25.0%) vs. 7/24 (29.2%)) [18]. Although these studies did not find a relationship between Ureaplasma positivity and BPD, a type 2 error due to the small sample size may have affected the findings obtained. In the analysis of intra-amniotic 16S rRNA gene patterns in preterm infants just before birth, Enterococcus (Firmicutes) and Ureaplasma (Tenericutes) were present at a higher relative abundance in the mild BPD group than in the no BPD group, while the relative abundance of Proteobacteria within the Escherichia-Shigella group was significantly higher in infants with moderate/severe BPD than in those with mild BPD but did not significantly differ from that in infants without BPD [35]. Another two studies from the same institution reported a relationship between Ureaplasma infection and BPD in infants by performing PCR and culturing amniotic fluid and placenta/fetal membranes collected during cesarean section [15,16]. In the study from 2011 on 257 preterm births at <34 gestational weeks, the incidence of BPD was significantly higher in Ureaplasma-positive cases than in Ureaplasma-negative cases (14/85 (16.5%) vs. 6/172 (3.5%), p < 0.001) [16]. Furthermore, the odds ratio of positive Ureaplasma infection for the incidence of BPD was significant (aOR: 11.4; 95% CI: 3.4-38.4) following adjustments for birth weight and the use of artificial ventilation in the first week as confounders, which is consistent with the present results.

A strong relationship was reported between intra-amniotic infection and bacteria in gastric fluid at birth; intra-amniotic infection with Ureaplasma spp. was associated with Ureaplasma spp. in gastric fluid at birth and the subsequent colonization of airway secretions. By analyzing the airway secretions of neonates, the Ureaplasma colonization of airway secretions at birth was shown to be associated with an enhanced inflammatory response in blood and increased inflammatory cytokine levels in secretions [12], and Ureaplasma positivity often persisted [36], suggesting progression to chronic lung infection [37,38]. Emphysema-like changes have also been detected on chest X-rays relatively early [12]. These findings indicate that a relationship exists between Ureaplasma colonization and the onset of BPD [11,13,14,37]. The incidence of persistent colonization is markedly lower for other bacterial species [36], and a relationship with BPD has not been reported [13]. In autopsies on extremely low birth weight infants, infants who died of respiratory disease and had Ureaplasma spp. in airway secretions or lung tissue showed more severe fibrosis and higher numbers of alveolar macrophages and TNF-α-and TGF β1-immunoreactive cells in lung tissue than those who died of causes other than respiratory disease or those with pneumonia caused by other bacterial species [39].

In experiments on sheep, colonization by Ureaplasma species induced inflammation in the immature lung, but not severe lung injury or the disruption of further lung development [40]. On the other hand, studies on non-human primates, specifically rhesus monkeys, which have similar characteristics to humans, demonstrated that an intra-amniotic injection of U. parvum and Mycoplasma caused a persistent infection of the fetal lungs and resulted in elevated levels of IL-8, tumor necrosis factor-α, and IL-6 in lung tissue and alveolar fluid [41,42]. In a preterm birth model of baboons artificially ventilated for the first 14 days after birth, an intra-amniotic injection of U. parvum induced more extensive fibrosis of lung tissue and more intense α-smooth muscle actin and transforming growth factor-β immunostaining than in non-injected controls [43].

The present results and previous findings on human neonates and animals suggest that intra-amniotic infection with Ureaplasma spp. causes persistent inflammation and fibrosis of the lungs from the fetal to neonatal stages, leading to the onset of BPD.

Strengths and limitations

We are the first to investigate whether relationships exist between BPD and intra-amniotic infection with and without Ureaplasma spp. and intra-amniotic inflammation without MIAC following an adjustment for gestational age at birth. The simultaneous assessment of the degree of intra-amniotic inflammation with amniotic fluid cultures allowed us to investigate MIAC with intra-amniotic inflammation. Furthermore, we analyzed the most extensive and sufficient number of cases among studies that examined amniotic fluid collected by amniocentesis.

However, we only performed amniotic fluid cultures and did not include PCR testing, and in the case of follow-up amniotic fluid testing, amniotic fluid cultures were conducted after antibiotic treatment. Therefore, cases of intra-amniotic infection may have been included among intra-amniotic inflammation cases without MIAC. Nevertheless, this is unlikely to have significantly affected the results obtained because the incidence of BPD in neonates ≥26 gestational weeks at birth was as low as 2/42 intra-amniotic inflammation cases without MIAC. Other limitations are that 1) this was a retrospective analysis from a single center, 2) we only performed a univariate analysis of moderate-to-severe BPD due to the small number of cases, and 3) the relationships between BPD and the severity and duration of intra-amniotic inflammation and intra-amniotic infection were not examined. 4) Our analysis does only apply to populations of pregnant women with clinical signs of preterm delivery and elevated CRP, but not the total populations.

Conclusions

Intra-amniotic infection with Ureaplasma spp. increased the risk of BPD after an adjustment for gestational age at birth. These results indicate the importance of diagnosing intra-amniotic infection with Ureaplasma spp. The efficacy of intravenous macrolides for mothers with intra-amniotic infection with Ureaplasma spp. warrants further study in a randomized controlled trial.

Authors’ contributions

All authors accept responsibility for the contents of this manuscript and approved its submission.

Research ethics

Research involving human subjects complied with all relevant national regulations and institutional policies, was performed according to the tenets of the Helsinki Declaration (as revised in 2013), and received approval from the Institutional Review Boards of National Hospital Organization Saga National Hospital (approval number: R4-8).

Informed consent

All patients provided informed consent for the present study.

Acknowledgments

We would like to thank Dr. Eiji Sadashima (Saga Medical Centre Koseikan) for his support with data analyses.

Disclosure statement

The authors declare no conflicts of interest.

Data availability statement

Raw data are available upon reasonable request from the corresponding author.

Additional information

Funding

This work was supported by JSPS KAKENHI (grant JP20H03564; for IY) and AMED, Japan (grant JP23fk0108677; for MN and IY).