INTRODUCTION

Five decades ago, Eastman and DeLeon ¹ linked maternal fever to cerebral palsy (CP) in full-term infants. Since then, it has been known that chorioamnionitis (CA) can induce preterm birth and potentially cause brain damage in the infant ².

CA is produced by bacterial invasion of the placenta with subsequent infiltration of polymorphonuclears (PMN) ³, and is considered one of the main causes of obstetric morbidity in Lima ⁴. Clinical CA (CCA) affects 0.5% to 10% of all pregnancies, and increases the risk of premature birth by three times ⁵. Meanwhile, histological CA (HCA) is present in 25-40% of all preterm births and up to 80% in preterm births under 30 weeks ⁶. In this regard, several studies have demonstrated the poor correlation between clinical and histological diagnosis ^{2 , 5 , 7}.

In Peru, brain lesions prevalent in premature infants under 34 weeks have been reported. The diagnosis was made using clinical and serial ultrasound examinations; lesions found were more severe as the gestational age decreased ⁸. Likewise, experimental studies and clinical observations have shown that CA can produce brain damage by direct invasion of pathogenic bacteria (sepsis/meningitis) ^{9 - 11} or by the effect of inflammatory products (cytokines) generated during placental infection ^{6 , 12}. In this process, cytokines could pass the blood-brain barrier (BBB), injure the cerebral white matter ^{6 , 11} and sometimes generate intraventricular hemorrhage ².

In the long term, it has been observed that infants exposed to HCA are at increased risk of developing CP and other neurodevelopmental disorders ^{6 , 13}. However, recent studies have questioned these findings and propose the development of new research methodologies to resolve the possible underlying mechanism ^{14 , 15}.

The histopathological study of the placenta and the clinical follow-up offer more evidence to characterize the possible association between HCA and brain damage in the preterm infant. Therefore, the objective of this study was to explore the risk of brain damage in preterm infants under 34 weeks of age exposed to HCA, identified sequentially by age, up to 40 weeks of corrected gestational age. As a secondary objective, due to their relevance in intensive care, maternal and neonatal characteristics were analyzed according to the condition of HCA.

MATERIALS AND METHODS

RESULTS

Patient enrollment

In 2015, 102 premature infants under 34 weeks of age were born. For this study 85 patients were included, 66 of them concluded the follow-up up to 40 weeks of corrected age, 6 abandoned the follow-up (7.1%) and 13 died (15.3%). The study was concluded in March, 2016 at the completion of the last preterm patient follow-up (Figure 1).

Figure 1 Patient inclusion flow chart 

Characteristics associated with histological chorioamnionitis

HCA was distributed as follows: subchorionitis 38.9% (14/36), chorionitis 38.9% (14/36), and chorioamnionitis 22.2% (8/36). All the cases of funisitis 44.4% (16/36) were combined with some type of affectation in the placental layers.

Maternal education was a factor negatively associated with HCA (p = 0.042). PROM and rupture of membranes for over 18 hours were more frequent in preterm infants exposed to HCA (p = 0.004 and p = 0.033, respectively). Prenatal steroid use was less frequent in the HCA group (p = 0.002). Vaginal delivery was more frequent in the HCA-exposed group (p = 0.044), while caesarean section without labor was more frequent in the non-exposed group (p = 0.044). Other characteristics of the mother, according to the status of the HCA, are detailed in Table 2.

Table 2 Maternal characteristics according to the histological chorioamnionitis status. 

^a Student’s T-test, ^b Fisher’ s exact test, ^c Chi-square test

The average head circumference percentile for the group with HCA was 47.6 (standard deviation: 28.3), lower (p = 0.035) than the group without HCA (60.7, standard deviation: 27.6). Of the total of patients, 57.6% (49/85) had probable or confirmed sepsis, which was more frequent in preterm infants exposed to HCA (p = 0.013). Other characteristics of the newborn, according to the status of the HCA, are described in Table 3.

Table 3 Neonatal characteristics of preterm infants according to the histological chorioamnionitis status. 

Neonatal Characteristics	Histologic chorioamnionitis		p value
	No (n = 49)	Yes (n = 36)
Gender			0.679 a
Female	24 (49.0)	16 (44.4)
Male	25 (51.0)	20 (55.6)
Gestational age			0.097a
24 to 29	14 (28.6)	12 (33.3)
30 to 31	20 (40.8)	7 (19.4)
32 to 33	15 (30.6)	17 (47.2)
Weight	1,688.7 (500)	1,645.6 (567)	0.711 b
Length, median (IQR)	41.6 (4.6)	40.8 (4.3)	0.215 c
Cranial circumference at birth, mean (SD)	29.1 (2.3)	28.2 (3.3)	0.179 b
Cranial circumference percentile, mean (SD)	60.7 (27.7)	47.6 (28.3)	0.035 b
Weight/Gestational Age Adequacy			0.818 d
Adequate	1 (2.0)	1 (2.8)
Low birth weight	34 (69.4)	22 (61.1)
Very low birth weight	9 (18.4)	7 (19.4)
Extremely low birth weight	5 (10.2)	6 (16.7)
Deceased	8 (16.3)	5 (13.9)	0.758 a
Sepsis			0.013 d
No	25 (51.1)	7 (19.4)
Possible	1 (2.0)	3 (8.3)
Probable	12 (24.5)	16 (44.4)
Confirmed	11 (22.4)	10 (27.8)
Acute renal failure	6 (12.2)	6 (16.7)	0.563 a
Hyaline membrane disease	29 (59.2)	16 (44.4)	0.179 a
Hypoglycemia	15 (30.6)	6 (16.7)	0.141 a
Pneumonia	8 (16.3)	6 (16.7)	0.967 a
Anemia	16 (32.7)	13 (36.1)	0.740 a
Prematurity apnea	12 (24.5)	12 (33.3)	0.371 a
Necrotizing Enterocolitis	4 (8.2)	2 (5.6)	1.000 d
Jaundice	26 (53.1)	23 (63.9)	0.318 a
Neonatal Encephalopathy	5 (10.2)	4 (11.1)	1.000 d
Bilirubin Encephalopathy	0 (0.0)	1 (2.8)	0.424 d
Ductus arteriosus persistence	12 (2.5)	10 (27.8)	0.732 a
Pneumothorax	2 (4.1)	3 (8.3)	0.646 d

SD: standard deviation, IQC: interquartile range

^a Chi-square test, ^b Student’s T-test, ^c Mann-Whitney U-test, ^d Fisher’s exact test

DISCUSSION

Histological chorioamnionitis has a deleterious effect at different ages of the premature infant under 34 weeks, with a risk of producing or contributing to some early and late brain injuries. These effects are known as “distantly injurious” because they emphasize cytokine mediated mechanisms ¹².

The research model we present is innovative, because it allows us to establish the effect of HCA on brain tissue as seen by preterm age. The design, underutilized in the literature, highlights the importance of histological examination of the placenta and its relationship with neurological and ultrasound monitoring of the patient ¹⁶.

The neurological damage related to histological chorioamnionitis in premature infants is a controversial issue that has been analyzed in experimental animal studies. Although the susceptibility of brain tissue has been demonstrated ¹⁹, the results of clinical studies have not been able to conclusively clarify this association ^{15 , 20 , 21}.

CA is an intense inflammatory process, whose main mechanism begins with the premature rupture of the amniotic membranes, which facilitates the entry of microorganisms from the vaginal flora into the intrauterine space ³. Thus, PROM and prolonged rupture of membranes are frequently associated ²². Although the most common germs are from the vaginal flora ^{12 , 20}, resulting in as many as 71.1% positive cultures ²⁰, there is often sterile chorioamnionitis, induced in conditions of stress, damage or cell death, such as placental detachment or additional infections ^{3 , 20 , 23}.

Regardless of the bacteriological risk, a powerful immune response is triggered in both, mother and fetus. In the mother, the production of prostaglandins and cytokines (IL-1β, IL-6, IL-8, TNF-α) that induce premature labor are activated ^{6 , 13}, and in the fetus, systemic inflammatory response syndrome is triggered, which correlates with increases of IL-6 in cord blood ^{5 , 19} and with the increase of MMP-8 in amniotic fluid ^{3 , 20 , 24}. It has already been reported that the use of steroids in the prenatal period could reduce various postnatal effects regarding inflammatory response ^{( 25}. The possible beneficial effect that would tend to lower the frequency of HCA should be handled with caution and be subject to future analysis. However, the effect of CA on brain injury, specifically on the frequency of cerebral palsy ²¹, and the potential deleterious effects on neurodevelopment are under discussion ²⁶.

In brain tissue, cytokines have the capacity to permeability alter the BBB and produce tissue disruption when exposed to hypoxia ⁶, ischemia ^{2 , 7}, sepsis ²⁷ or necrotizing enterocolitis ⁶. During this ischemic and inflammatory process, cells of the microglia ^{12 , 19} and astroglia can be activated, producing free radicals and a greater amount of cytokines (IL-α, ILβ, IL-6 and IL-8) that directly damage the pre-oligodendrocytes ¹⁹ and induce apoptosis through the caspase pathway ⁶. In this regard, Volpe explains that vulnerability of preoligodendrocytes and subplate neurons develops between 24 and 32 weeks of gestation ^{7 , 27}.

Since CA is based on clinical criteria that can lead to false positives or negatives ^{3 , 5 , 24}, we used an analysis model that relates HCA to neurological damage by age. Surprisingly, 42% of placentas examined presented HCA, a frequency that is at the upper limit of national and international reports ^{3 , 28}. Although most of the affected placentas showed early stage infiltration ^{3 , 20} when the chorion and amnion were affected simultaneously, the risk of neurological damage doubled or tripled both in the first week of life and between 7 and 30 days, with different patterns of associated brain damage, rarely described in the literature.

The brain lesions often were hemorrhagic, ischemic and affected white matter. Except for meningitis, all were identified by brain ultrasound, a method chosen for its great accessibility and efficiency in determining the type and extent of brain injury ²⁰. The noteworthy SLV frequency, a type of white matter lesion related to inflammatory processes, was followed by IVH, as already described in our setting ⁸.

The adjusted analysis by HCA subtypes, excluding confounding factors, and according to age allowed us to determine the association with some brain injuries. Only the concurrent condition of having both, chorion and amnion affection, was related to risk of damage. For IVH, the risk doubled during the first week, possibly due to the association of the inflammatory component with capillary fragility ²⁹, independent from other factors ². In contrast, the risk for PVL and SLV type white substance lesions tripled between 7 and 30 days of age, possibly due to the inflammatory and ischemic mechanisms described above ^{7 , 19}.

However, at 40 weeks of corrected age no association was found despite controlling for the confounding effects of preceding injuries and concurrent systemic diseases. We consider that cerebral hypoplasia and other injuries were the effect of multiple combined processes among which malnutrition, metabolic disorders (hypoglycemia, hyperbilirubinemia), bronchopulmonary dysplasia, pneumonia, late sepsis and necrotizing enterocolitis stand out, among these the inflammatory effect of HCA itself is diminished ^{( 7 , 20 , 26}. Other very small or diffuse brain lesions that would have improved the analysis were not identified due to lack of magnetic resonance imaging ^{29 , 31}.

This study has shown that various inflammatory mechanisms contribute to brain damage. The main one was placental infection that leads to sepsis; since there is no evidence that meningitis plays a fundamental role in brain damage, we consider that the inflammatory factor derived from the sepsis itself would be responsible for the long-term harmful effect ^{26 , 29}. Other infectious and inflammatory processes, such as pneumonia, necrotizing enterocolitis, persistent ductus arteriosus and hyaline membrane disease, were associated with greater neurological damage, especially between 7 and 30 days of life ²⁹, which opens up a series of possibilities for the management and control of cerebral lesions by age, which require further analysis in the future.

However, recent studies have concluded that there is insufficient evidence to justify the association between HCA and brain injury in preterm infants ^{30 , 32}. Only the study by Pappas et al. ²⁹ found a correlation between severe IVH and HCA. None of the studies after 2005 found any association with PVL, the best-known type of lesion. However, the hypothesis remains in question because these clinical studies differ in methodology and conversely there is adequate experimental support. Shi et al ²¹, based on a systematic review and meta-analysis, noted that HCA is associated with risk of cerebral palsy when studies are prospective and start from the presence or absence of chorioamnionitis. Others emphasize analyzing amnion involvement ³². On the long-term effects, Ylijoki et al. showed that HCA was associated with lower cognitive performance, memory and learning skills at 5 years of age ²⁶.

The effects on lower gestational age premature infants was of particular interest. These infants are more frequently and extensively brain damaged, at the expense of white matter; and are likely to have smaller cranium size. These findings have long been described ^{7 , 19} and generally translate into severe disabilities in early childhood ¹⁶. However, there are many related factors, in which the role of HCA has not been clearly defined ^{23 , 26}.

Based on current knowledge, future projects should take into account types (acute and chronic) and grades of CA, germs, umbilical cord clamping, fetal inflammatory response (funisitis and amnion), maternal inflammatory response, circulation of deleterious and protective cytokines, use of prenatal and postnatal steroids, head circumference monitoring, MRI findings and neurodevelopmental examinations ^{3 , 6 , 30}.

One of the strengths of the study is that the anatomopathological examinations of the placenta were available, and neurological diagnoses were performed by neurologists trained in brain ultrasound. Likewise, an analysis model by age was carried out, understanding the importance of the physiopathology of brain injuries in premature infants. Also, a multivariate analysis was performed, which allowed considering confounding factors. As a limitation, it should be mentioned that the time of umbilical cord clamping was not recorded. Late sepsis, which could be a confounding variable on the effect at 40 weeks, was also not differentiated, and no membrane culture or amniotic fluid culture was obtained that could better explain the germ’s virulence factor.

In conclusion, HCA was found in almost half of all preterm births under 34 weeks. Premature rupture of membranes was the main cause of neonatal sepsis, and in turn sepsis was associated with neurological damage. Chorioamnionitis caused brain damage in preterm infants aged 0-7 days and 7-30 days. The most frequent injuries were IVH and damage to white matter (SLV and PLV). At 40 weeks of corrected age, extreme preterm infants with HCA had more extensive brain damage.