A review of chronic lung disease in neonates at Charlotte Maxeke Johannesburg Academic Hospital from 1 January 2013 to 31 December 2014

Abstract

Background. Chronic lung disease (CLD) remains a significant morbidity in preterm babies despite advances in neonatal care. The use of postnatal corticosteroids (PNCS) to treat CLD remains controversial. Objectives. To describe the clinical characteristics of babies with CLD at Charlotte Maxeke Johannesburg academic hospital (CMJAH) and to explore the use of PNCS for the prevention and treatment of CLD. Methods. This was a 2-year retrospective review of neonates admitted to CMJAH. Neonates who were in hospital for ≥28 days were included. Comparisons were made between neonates with evolving CLD and those with no CLD. Results. A total of 485 neonates were analysed, 237 had evolving CLD and 245 did not have CLD. Overall incidence of evolving CLD was 5%. More neonates with CLD needed resuscitation at birth (48.5% v. 39.8%; p=0.02) and had low 5 minutes Apgar scores (17.2% v.10.6%; p=0.001). Neonates with CLD had increased prevalence of patent ductus arteriosus (30.4% v. 7.7%; p=0.001) and late onset sepsis (56.5% v. 23.6%; p=0.001). The mortality rate was also higher in CLD babies (10.2 v. 2.4%; p=0.001). Necrotising enterocolitis (NEC) (29.2% v.8%; p=0.005) and sepsis (83.3% v. 53.8%; p=0.008) were associated with increased mortality. The use of PNCS was associated with less NEC (3.5% v. 17.2%; p=0.001) and improved survival (95.6% v. 81.7%; p=0.001). Conclusions. CLD remains a common morbidity in neonates despite advances in neonatal care. The use of PNCS was shown to have short-term benefits. To get the most out of PNCS use for CLD, further studies need to be conducted to determine the safest type of steroid, the safe doses and the duration of treatment. S Afr J Child Health 2016;106(6):xx-xx. DOI:10.7196/SAJCH.2016v106i4.1060

Background The clinical definition of neonatal chronic lung disease (CLD) also known as bronchopulmonary dysplasia (BPD) has evolved over time. It was first defined by Northway et al. in 1967 as persistent respiratory signs and symptoms along with the need for supplemental oxygen and an abnormal chest X-ray (CXR) at 28 days of age.[1] The definition of CLD was subsequently modified and defines BPD as oxygen dependence at 36 weeks post menstrual age (PMA) with or without the use of respiratory support and with or without the characteristic radiographic changes.[2] These definitions do not consider gestational age (GA) and do not indicate the level of oxygen dependence that can range from needing low-flow oxygen to being ventilator-dependent. To address this issue, the National Institute of Health has developed a consensus severity-based definition. This definition includes all babies born as needing more than 21% supplemental oxygen for at least 28 days. CLD is further classified into mild, moderate and severe, depending on the FiO2 needed and the duration of oxygen therapy for preterm babies.[3] Because of the complexity of the definition, some units just use the need for oxygen on day 28 of life and often refer to this as evolving CLD. The incidence of CLD as defined by the need for oxygen supplementation at 36 weeks PMA is ~30% of premature infants with birth weight (BW) <1 000 g and is uncommon in infants born at >30 weeks of gestation or weight of >1 250 g.[3] CLD has multifactorial aetiology and remains a major cause of morbidity in premature infants[1].Contributing factors include infection, exposure to high oxygen levels with the formation of toxic oxygen free radicals and ventilator-induced lung injury that results in arrested lung development and impaired lung function.[4] Several maternal risk factors including increasing age, hypertension, lack of antenatal steroid usage and chorioamnionitis have been associated with BPD.[4] Over the past few years, the pathophysiology and aetiology of BPD has changed. This pathophysiological and aetiological shift has been bought about by improved survival in extremely premature babies as a result of for instance, antenatal steroids and surfactant therapy. The classic BPD was characterized mainly by lung damage and fibrosis due to oxygen toxicity and mechanical ventilation. The new BPD is characterised by a disorder in lung development with fewer, larger and simplified alveoli. The management of CLD includes several preventive and therapeutic strategies that target several pathways and processes involved in pathogenesis of CLD. Some provide antioxidant protection; others minimise specific aspects of inflammation, reduce elastolytic and proteolytic injury or regulate growth. In addition, supportive pharmacological treatments that target the development of pulmonary oedema, bronchoconstriction and impaired gas exchange are used. The success of these interventions has been variable.[5] Postnatal corticosteroids (PNCS) have been extensively studied and have been found to be effective in weaning infants off mechanical ventilation.[6-8] This effect has been proven for dexamethasone, which is the most widely studied PNCS in randomized controlled trials (RCTs). Despite the short-term benefits, dexamethasone has not been shown to reduce the total days of hospitalisation, duration of supplemental oxygen therapy, or incidence of CLD.[6-8] In the era before PNCS treatment, the long-term neurodevelopmental outcome for survivors with CLD was worse than that in similar infants without CLD.[9] Adverse effects of PNCS that includes hyperglycaemia, gastrointestinal (GIT) perforation, hypertension, infection, steroid-induced cardiomyopathy, long-term neurodevelopmental effects and growth retardation complicate the use of PNCS. The most worrisome long term effect is increased risk for poor neurological outcome including cerebral palsy (CP). Corticosteroids can have direct toxic effects on the developing brain, including neuronal necrosis, interference with healing and inhibition of brain growth.[9,10] A systematic review showed a significantly higher rate of CP after corticosteroid treatment and a non-significant reduction in mortality.[11] A multicentre double blinded RCT testing early postnatal dexamethasone therapy for prevention

of CLD had to be stopped before completion because of concern about serious side effects such as GIT perforation and periventricular leukomalacia (PVL).[12] The American Academy of Pediatrics (AAP) also recommended that alternative corticosteroids undergo studies and that all infants enrolled in RCTs for PNCS receive long-term neurodevelopmental follow up.[13] Due to concern about the safety of PNCS, in 2002 the AAP released a policy statement regarding the use of PNCS for prevention or treatment of CLD stating that the routine use of dexamethasone could not be recommended.[13] The AAP recommended that dexamethasone use be limited to RCTs with long-term follow up. Since the publication of the AAP statement, postnatal use of dexamethasone for CLD has reduced; however the incidence of CLD has not diminished. Some reports have suggested that the incidence and severity of CLD may have actually increased.[14] The data available for PNCS use in CLD are inconclusive and conflicting. As a result, clinicians are advised to use their own clinical judgment to balance potential adverse effects of CLD with the potential adverse effects of PNCS for each individual patient. The incidence of CLD in very low birth weight (VLBW) babies at Charlotte Maxeke Johannesburg academic hospital (CMJAH) is lower than that reported in the Vermont Oxford network (VON).[15] At CMJAH, babies who are on supplemental O2 for >28 days are given oral prednisolone (OP) for prevention/treatment of CLD. Babies who remain ventilator dependent are given dexamethasone. Alternative PNCS include hydrocortisone, nebulised dexamethasone and oral prednisone.[5] One study looking at the effect of short course of oral prednisone (OP) in infants with O2 dependent BPD provided evidence that OP is effective in a select patients with BPD.[16] Characteristics of babies with CLD and the use of PNCS have not been reviewed at CMJAH. This study aims to describe babies with evolving CLD and to explore the use of PNCS for evolving CLD at CMJAH. Methods This study is an institution-based retrospective audit conducted in the neonatal unit at CMJAH in Parktown, Gauteng Province. The objectives were to determine the incidence of CLD at CMJAH, to describe clinical and demographic characteristics and survival to discharge in babies with CLD and to compare these to those of babies without CLD. Evolving CLD was defined as oxygen use at 28 days of life and the need for supplemental oxygen at 36 weeks PMA was considered as definite CLD (VLBW babies only). Following approval by the Committee for Research in Human Subjects at the University of the Witwatersrand (Medical), the CEO and the HOD of Paediatrics at CMJAH, a 2-year (1 January 2013 to 30 December 2014) retrospective review of neonatal medical records was performed. Data from the existing CMJAH neonatal database (Research Electronic Data Capture hosted by the University of the Witwatersrand) was used for analysis. [17] The data are collected prospectively on an ongoing basis for the purpose of clinical audit from clinician completed hospital records. All babies admitted to the CMJAH neonatal unit within 72 hours of life (inborn and outborn), with BWs of ≥500 g, who were in hospital for ≥28 days were included. Babies with irretrievable data were excluded from the study. The group was divided into babies with evolving CLD and those without. The neonatal unit at CMJAH has 84 beds, 35 of which are high-care, 14 in paediatric/neonatal intensive care unit, 20 low-care beds and 15 kangaroo mother care beds. Respiratory support includes early rescue surfactant (SVT), supplemental oxygen via low-flow nasal cannulae (NPO2), nasal continuous positive airway-pressure ventilation (NCPAP), intermittent positive-pressure ventilation (IPPV) and high frequency oscillatory ventilation. Due to limited resources ventilatory support in the form of NCPAP was only offered to babies with BW ≥750 g who showed signs of respiratory 10 failure. Babies with BW ≥900 g who showed signs of respiratory failure on NCPAP or became apnoeic were offered IPPV. Respiratory failure was defined as O2 saturation <88% on 60% supplemental O2, respiratory acidosis (pH <7.25 with PaCO2 >60 mmHg) or markedly increased work of breathing. Definitions Maternal hypertension included both chronic and pregnancy induced hypertension. Chorioamnionitis was defined as premature and/or prolonged rupture of membranes, fever and foul smelling liquor in mothers. Resuscitation at birth was defined as the need for bag mask ventilation, chest compressions, or intubation and ventilation. The Ballard score was used to estimate gestational age (GA). Fenton growth charts (2013) were used to assess weight for gestational age.[18] The whole group of neonates was described and then stratified by BW category namely: •! ≥500-999 g, extreme low birth weight (ELBW) •! ≥1000 to 1499 g, very low birth weight (VLBW) •! ≥1500 to 2499 g, low birth weight (LBW) •! ≥2500 g, normal birth weight (NBW). Babies were considered to be small for GA if the BW was <10th percentile.[18] The 5 minute Apgar scores were categorised into two groups, namely Apgar score ≤5 and Apgar score >5. Babies were divided according to GA into two groups, namely <32 weeks and ≥32 weeks. PNCS was defined as steroids given in an attempt to facilitate weaning of patients off prolonged ventilation or supplemental oxygen. Dexamethasone is given to patients failing to wean off mechanical ventilation. NEC was considered as modified Bell’s stage ≥2.[19] Sepsis was classified as culture-proven bacterial or fungal blood stream sepsis only. Statistical analysis The data was assessed for missing information and erroneous or suspicious entries. These entries were verified as far as possible with the original patient records. Information not available from the database was obtained from hospital files drawn from the hospital record archives. The database was then exported to IBM SPSS Statistics version 23.0 for analysis. Babies with evolving CLD: Babies in different weight categories were compared in terms of therapeutic intervention and outcome. Babies who received PNCS were compared to those who did not, and babies who survived to discharge were compared to those who died. Finally the CLD group was then compared to the no CLD group. The data were normally distributed, so continuous variables were described using means and standard deviations (SD) while frequencies (percentages) were reported for categorical variables. For comparison Chi-square tests were used for categorical variables and independent t-tests for continuous variables. All analyses considered a value of p<0.05 as significant.

Results There were 485 babies hospitalised for more than 28 days; records were not retrievable for 3 patients. Therefore, 482 patients were included, 237 with evolving CLD and 245 without CLD. The overall incidence of evolving CLD was 237/4570 (5.1%). The incidence in the VLB)W babies was 206/1,302 (15.8%) and 31/3,268 (0.94%) in the >1 500 g babies (p<0.0001).The incidence of definite CLD was 98/1,302 (7.5%) in the VLBW babies. Demographic and birth characteristics are shown in Table 1. There was no difference in the birth weight between babies with and without CLD (1017 grams (SD 101) vs 1041 grams (SD 104) p = 0.57). Similarly gestational age was not different between the two groups (CLD 28.2 weeks (SD1.9) vs no CLD 28.3 weeks (SD 1.9) p = 0.9). There were, however, more babies who were SGA in the no CLD group than in those with CLD (26.5% vs 12.9% p=0.03). The percentage of males with CLD was greater (50.2) than in the no CLD group (40.8%). There was no significant difference in maternal obstetric and labour room characteristics between CLD and no CLD babies (Table 1). There were more babies who had 5 minute Apgar scores ≤5 and needed resuscitation at birth in the CLD group.