Retrospective analysis of leukemoid reactions in extremely preterm infants in a tertiary NICU from 2018–2021

Abstract

Background

Neonatal leukemoid reaction is associated with higher risk of mortality, chronic lung disease and has been associated with chorioamnionitis. Literature on extremely low birth weight infants with leukemoid reaction is limited.

Objectives

The aim of our study was to characterize the maternal and placental factors associated with neonatal leukemoid reaction and to describe outcomes of these ELBW infants. Our objective was to assess if there were maternal factors that would assist the decision-making process regarding the delivery of preterm infants at risk of chorioamnionitis and the sequelae of this inflammatory process.

Methods

This was a retrospective case-control study performed in a single, tertiary Maternity Hospital in Dublin. Two matched controls were identified for each case based on gestation and year of birth and data was collected on both the infants and their mothers.

Results

7 extremely preterm neonates were identified as having a leukemoid reaction, defined as a total white cell count of >50,000 or in the first seven days of life. Baseline characteristics between the groups were similar. The median gestational age in the cases group was 24 + 4 weeks and in the control group was 24 + 1. The mean birthweight was 650 g in the cases group vs. 655 g in the control group. There was a higher percentage of males in the control group, 42.9% vs 28.6% in the cases. The preterm infants with leukemoid reaction had a longer duration of ventilation with a median of 18 days (7.5–23.5 days) compared to 6.5 days (2.8–24.5 days) in the control group. More infants in the leukemoid reaction group required inotropes for hypotension in the first 72 h after delivery (42.9% vs 7.1% in the controls, p value .169). Death or Bronchopulmonary dysplasia (BPD) occurred in 85.7% of the cases identified with a leukemoid reaction vs 71.4% of the controls matched. Median maternal CRP was higher in cases prior to delivery vs the controls (66 vs 18.1 mg/L, p-value = .2151). There was histological evidence of maternal inflammatory response in all cases with fetal inflammatory response in 71% of cases.

Conclusions

Leukemoid reaction in ELBW infants with evidence of maternal and fetal inflammatory response syndrome on placental histology is associated with a longer duration of initial ventilation, increased need for inotropes in the first 72 h after birth, higher rates of death, and BPD. Prospective studies are required to identify potential biomarkers such as proinflammatory cytokines, IL-6, which might aid the decision-making process in delivery.

Keywords:

• Leukemoid reaction
• fetal inflammatory response syndrome
• prematurity

Introduction

Neonatal leukemoid reaction, defined as white blood cell count >50x10⁹cells/L or the absolute neutrophil count >30 × 10⁹ cells/L [1], is associated with a higher risk of mortality, chronic lung disease and has been associated with chorioamnionitis [2]. Literature on extremely low birth weight infants with leukemoid reactions is limited. We know that there is a relationship between leukemoid reaction and chronic lung disease in premature infants [3] as well as with increased mortality. The mothers of these infants are more likely to have chorioamnionitis on placental histology and to have been treated with antibiotics for preterm prolonged rupture of membranes and suspected sepsis [4]. As well as adverse respiratory outcomes, previous studies have also demonstrated the association between histological chorioamnionitis, necrotizing enterocolitis and retinopathy of prematurity [5–8].

The incidence of neonatal leukemoid reaction is variable between 1.3% and 15% based on a small number of studies [1,9]. The association between leucocytosis in preterm neonates and maternal glucocorticoid administration for lung maturation is unclear as data is limited [2].

The aim of our study was to characterize the maternal factors, including placental histology, associated with neonatal leukemoid reaction and to review neonatal outcomes of extremely preterm infants with a leukemoid reaction within the first seven days of life. We aimed to assess if there were maternal factors that would assist the decision-making process regarding the delivery of preterm infants at risk of chorioamnionitis and the sequelae of the inflammatory process underlying leukemoid reaction in VLBW infants.

Methods

This retrospective case-control study was performed in a single, tertiary Maternity Hospital in Dublin. Infants born between January 2018 and December 2021 were eligible for inclusion. Cases of term and preterm babies demonstrating a leukemoid reaction (defined as a WCC > 50x10⁹/L [1] before day seven of life were identified through the laboratory information system. Demographic data, laboratory test results, and information on clinical course and outcome were collected using the electronic healthcare record (Maternal Newborn Clinical Management System). Two controls were identified for each case using the maternity information system, matched for year of birth and gestational age at delivery. This research was approved by the local hospital Research and Ethics Committee (RAAG-2021-024).

Data were collected on maternal and pregnancy-related factors such as maternal age, indication for and mode of delivery, antenatal steroids, preterm and duration of rupture of membranes, maternal pyrexia, antenatal antibiotic use, clinical or histological chorioamnionitis and maternal laboratory investigation results including vaginal, urine, placental and blood cultures. Neonatal demographic information and data on gestational age, birth weight, and sex were obtained. The study also collected data on infant condition at delivery, neonatal intensive care supports provided including ventilatory support and use of inotropes as well as neonatal clinical outcomes (neonatal respiratory pathology, BPD and use of postnatal steroids, NEC, IVH, PVL, and mortality). In addition, infant laboratory results including FBC, CRP, and any cultures taken were noted. Continuous variables were described using mean (SD) and compared using student t-test. Non-parametric data were analyzed using median (range) and Wilcoxon rank test for inter-group comparisons. Categorical data was analyzed using chi-squared test.

Results

7 extremely preterm neonates were identified as having a leukemoid reaction, defined as a total white cell count of >50,000 in the first seven days of life, using the laboratory information management system between January 2018 and December 2021. Two matched controls were identified for each case based on gestation and year of birth and data was collected as described above.

Baseline characteristics between the groups were similar (Table 1). The median gestational age in the cases group was 24 + 4 weeks and in the control group was 24 + 1. The mean birthweight was 650 g in the cases group vs. 655 g in the control group. There was a higher percentage of males in the control group, 42.9% vs 28.6% in the cases.

Table 1. Baseline Characteristics.

Variable	Cases (n = 7)	Controls (n = 14)
Gender n (% male)	2 (28.6%)	6 (42.9%)
Maternal age (mean, SD)	32.1 (6)	32.1 (5.4)
Gestation at delivery (median, IQR)	24.571 (23.643–24.929)	24.285 (23.857–25.142)
Birthweight (mean, SD)	692.857 (98.084)	677.857 (102.673)
Apgar at 5mins <7 n(%)	5 (71.429%)	5 (35.714%)
Delivery n (% LSCS)	3 (42.85%)	9 (64.286%)
PPROM n(%)	4 (57.143%)	10 (71.429%)

Antenatal steroids were completed in all cases (Table 2). Birth was preceded by PPROM in 57% of the cases compared with 71% of the controls. The indication for delivery was suspected chorioamnionitis in 42.9% of the cases compared with 21.4% of the controls.

Table 2. Maternal factors.

Variable	Cases (n = 7)	Controls (n = 14)	P value
Maternal CRP (median, IQR)	66.000 (26.35-104.0)	18.1(6.600–24.40)	.2151
Maternal WCC (median, IQR)	18.900(15.100–20.400)	16.450 (14.525– 19.750)	ns
Maternal Pyrexia n(%)	2 (28.571%)	1 (7.143%)	ns

Median maternal CRP was higher in cases prior to delivery vs the controls (66 vs 18.1 mg/L, p-value = .2151). There was a higher percentage of maternal pyrexia prior to delivery in the cases than in controls (28.6% vs 7.1%).

Histology

Examination of the maternal and fetal sides of the placentas in the cases of neonatal leukemoid reaction identified in VLBW infants revealed histological evidence of maternal inflammatory response in all cases with fetal inflammatory response in the majority of cases (5 out of 7 cases, 71%).

6 of the 7 maternal sides of the placenta had evidence of severe chorioamnionitis, classified as stage 3, grade 2 chorioamnionitis (Table 3). The classification system defines these as necrotizing chorioamnionitis with severe confluent polymorphonuclear leukocytes or subchorionic microabscesses [10]. The other placenta revealed acute subchorionitis (stage 1, grade 2).

Table 3. Histology with outcomes.

Cases	Maternal Risk Factors	Maternal CRP (mg/L)	Histology Maternal Side	Histology Fetal Side	Neonatal Outcome
1	GBS on HVS	<1	SI, GII	No	No BPD
2	PPROM	49	SIII, GII	SII, GII	Died day 3
3	PPROM, maternal pyrexia	19	SIII	SII	Died day 28
4	PPROM	Not available	SIII, GII	SIII, GII	NEC, BPD, ROP, PDA treated, discharged on home 02
5	PPROM	83	SIII, GII	SIII, GI	Died day 19
6	Maternal Pyrexia	111	SIII, GII	No	Died day 12
7	PPROM	114	SIII, GII	SII, GII	BPD, required HFOV and iNo, PVL

Fetal inflammatory response (FIR) is reported when there is evidence of mural inflammation in the umbilical vessels or chorionic plate vessels. Examination of the fetal side of the placentas revealed evidence of FIR in the majority of cases (5 out of 7). The other two showed no evidence in the sampled cord vessels and chorionic plate vessels, however, it is possible that this was a sampling issue and that if other samples had been taken inflammation may have been seen. Of those that had FIR identified on the placenta, all had at least stage II disease with involvement of the umbilical vein and one of more umbilical arteries or stage III defined as necrotizing funisitis.

Neonatal outcomes

All infants with a leukemoid reaction required resuscitation including intubation at delivery and one infant required CPR. Only 64.3% of the matched controls required intubation at delivery. Apgar scores at 1 and 5 min were lower in the cases vs the controls (Table 4).

Table 4. Neonatal outcomes.

Variable	Cases (n = 7)	Controls (n = 14)
NEC n(%)	1 (14.286%)	2 (14.286%)
Death or BPD at 36/40 n(%)	6 (85.714%)	10 (71.429%)
Inotropes in first 72hrs n(%)	3(42.857%)	1 (7.143%)
Death before discharge n(%)	4 (57.143%)	5 (35.714%)
IVH n(%)	5 (83.333%)	9 (64.286%)
PVL n(%)	1 (14.286%)	0 (0%)
Hydrocortisone for hypotension in first 72hrs	2 (28.571%)	0 (0%)
Need for iNo n(%)	4 (57.143%)	4 (28.571%)
HFOV n(%)	4 (57.143%)	6 (42.857%)
RCC transfusion first 72hrs n(%)	5 (71.429%)	4 (28.571%)
No. of days ventilated (median, IQR)	18 (7.500–23.500)	6.5(2.750–24.500)
Pneumothorax n(%)	3 (42.857%)	5 (35.714%)
Doses of Surfactant 1 2 3	4 2 1	10 2 2
Need for postnatal steroids (DART) n (%)	3 (42.857%)	5 (35.714%)

More infants in the leukemoid reaction group required inotropes for hypotension in the first 72 hours after delivery (42.9% vs 7.1% in the controls, p value .169). The preterm infants with leukemoid reaction had a longer duration of ventilation with a median of 18 days (7.5–23.5 days) compared to 6.5 days (2.8–24.5 days) in the control group.

There were higher rates of severe IVH (Grade 3/4) and PVL in the group with leukemoid reactions in the first week of life. There were similar rates of NEC in both groups (Table 3).

Death or BPD, defined as 0₂ requirement at 36 weeks CGA [11] occurred in 85.7% of the cases identified with a leukemoid reaction vs 71.4% of the controls matched.

There was more use of postnatal steroids for the management of evolving lung disease (DART protocol, [12]) in the patients with leukemoid reaction compared to their matched controls (42.9% vs. 35.7%). This is likely related to the longer duration of initial ventilation in the leukemoid group compared to the control group. Dexamethasone in our unit is used over a ten-day course with a cumulative dose of 712 micrograms/kg adapted from the DART study.

Discussion

This small case series describes a phenomenon in preterm neonates of fetal inflammatory response syndrome associated with leukemoid reaction and subsequent poorer outcomes for the preterm neonate affected compared with matched controls.

The lack of infectious involvement in infants with neonatal leukemoid reaction indicates the potential for early discontinuation of antibiotics when there is associated FIRS. The use of corticosteroids as a treatment in preterm neonates with leukemoid reactions has not previously been studied. Many of these infants will require hydrocortisone for the management of hypotension in the early postnatal period and later treatment with dexamethasone for bronchopulmonary dysplasia.

PPROM complicates only 2% of pregnancies but is responsible for 40% of preterm deliveries [13]. The three causes of neonatal death associated with PPROM are prematurity, sepsis, and pulmonary hypoplasia. In addition, there are maternal risks associated with chorioamnionitis. The biggest dilemma in relation to PPROM is the timing of delivery. The main issue is balancing the risks and benefits of prolonging the pregnancy to reduce complications of prematurity with the increased risk of intrauterine infection [14]. One-third of pregnancies with PPROM have positive amniotic fluid cultures [15]. The management of PPROM consists of antibiotic prophylaxis and corticosteroids with close monitoring for signs of maternal or neonatal infection or compromise. If the woman does not go into spontaneous labor, delivery is indicated if there is evidence of maternal or fetal compromise, signs of infection (chorioamnionitis), or if a gestational age of 34–37 weeks is reached.

Clinical evidence of chorioamnionitis includes maternal pyrexia, uterine tenderness, offensive vaginal discharge, and fetal tachycardia. Although maternal blood parameters such as C-reactive protein and a full blood count are easy to measure, their specificity in indicating ongoing fetal inflammation or chorioamnionitis is limited [16–18]. The presence of leucocytosis may also be useful clinically in cases where there is doubt about the diagnosis of chorioamnionitis. However, clinical markers from the uterine cavity such as amniotic fluid have shown greater prediction of chorioamnionitis [19,20].

Previous studies reported the association between a systemic fetal inflammatory response and increased rates of bronchopulmonary dysplasia and neonatal morbidity. FIRS is associated with elevated fetal plasma interleukin-6 value [21]. The use of Interleukin 6 has been described as an adjunct to predict the onset of FIRS and neonatal leukemoid reaction [22].

The measurement of IL-6 (interleukin 6) in the vaginal fluid has been shown to accurately predict chorioamnionitis in some cases up to 2 days prior to other clinical signs of chorioamnionitis being apparent [23].

The possibility of measuring IL-6 levels in vaginal fluid in the monitoring phase of women with PPROM could potentially trigger earlier intervention for preterm neonates at risk of FIRS and leukemoid reaction, rather than waiting for a rise in maternal CRP or overt maternal signs of infection.

Our study adds to the literature a well-documented phenomenon of leukemoid reaction, however, there is limited data on extremely preterm infants affected by leukemoid reaction and how to prognosticate outcomes. With the improved survival of our extremely preterm infants, we continue to search for markers of prognostication both physiologically and biochemically in order to aid decision-making processes with these vulnerable infants and their families. We hope that having more information close to the time of delivery about the presence of FIRS would enhance communication with parents about prognostication for preterm infants with leukemoid reaction.

Conclusion

Leukemoid reaction in ELBW infants with evidence of maternal and fetal inflammatory response syndrome on placental histology is associated with a longer duration of initial ventilation support, increased need for inotropes in the first 72 h after birth, and higher rates of death and BPD.

The findings from our study, despite the small sample size, indicate that higher maternal CRP values have an association with neonatal leukemoid reaction and subsequently increases in neonatal morbidity and mortality. Current obstetric management of suspected chorioamnionitis does not indicate an absolute CRP value for intervention and more robust clinical and biochemical markers are required to guide the timing of delivery. Prospective studies are required to identify potential biomarkers such as proinflammatory cytokines, IL-6, which might aid the decision-making process in delivery.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.