Abstract
INTRODUCTION:
Extremely preterm infants are a population of high risk for morbidity and mortality. NICU's staffing is often lower during nights, weekends and holidays than weekdays, and this fact may contribute to higher morbidities and mortality. Our aim was to analyze the neonatal morbidity and mortality of very preterm infants delivered at our center and admitted to the NICU during the night period, weekends and holidays compared to that registered on weekday admissions.
METHODS:
A retrospective study was conducted at our level III NICU, including data on mother, pregnancy, delivery, and neonatal outcomes of preterm infants with a gestational age below 30 weeks, admitted between January 1st 2005 and December 31st 2017. Statistical analysis was performed using IBM SPSS® Statistics 23.
RESULTS:
220 infants were included in the study; median gestational age 27 weeks (min = 23; max = 29); median birth weight of 922 g (min = 360; max1555); 95 (43.2%) infants were delivered during weekdays and 125 (56.8%) were delivered during weeknights, weekends and holidays. There were no differences on mother's age, pregnancy complications, Apgar scores, birth weights, gestational ages and gender between the two groups. C-sections (p = 0.006), and small for gestational age infants (p = 0.010) were more prevalent in week day births. Chorioamnionitis with chorionic vasculitis (p = 0.028) and cystic periventricular leukomalacia (p = 0.032) were more prevalent in those delivered during the night period, weekends and holidays. In the multivariate analysis, cystic periventricular leukomalacia was not associated to a deliver during weeknights, weekends and holidays (OR = 0.580; 95% CI: 0.19–1.71, p = 0.324).
CONCLUSION:
We did not find any increased morbidity and mortality associated with a birth during nights, weekends and holidays compared to that registered on weekday admissions.
Introduction
Premature birth is an important cause of neonatal morbidity and mortality, and the severity of adverse outcomes is inversely correlated with gestational age [1, 2]. Many current data shows changing trends in mortality and morbidity rates for preterm populations [3–5] and outcomes of very preterm infants also vary considerably among health care facilities [6]. Several risk factors have been associated with increased risk of morbidity and mortality, and studies in adults and pediatric populations have demonstrated that weekend admissions compared with weekday admissions have higher adverse effects on outcomes [7, 8].
Neonatal intensive care units (NICUs) are difficult to staff appropriately due to fluctuations in patient volume and acuity. Health service resources are reduced during night shifts, weekends and holidays in many NICUs and whether this fact may affect preterm neonate outcomes is not known [9–11]. Our hypothesis was that some morbidities as bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH) with or without periventricular infarction (PVI), cystic periventricular leukomalacia (cPVL) and retinopathy of prematurity (ROP) could present an increased prevalence in infants delivered during periods of reduced, and also tired, staff. Among several risk factors, in which gestational age is the most important, these morbidities depend also on expertise management during the first hours of life.
The objective of our study was to compare the neonatal morbidity and mortality of very preterm infants delivered at our center and admitted to NICU during the night period, weekends and holidays with that registered for infants admitted during weekdays.
Methods
A retrospective study was conducted at our level III NICU, including data on preterm infants with a gestational age below 30 weeks (23 weeks +0 days to 29 weeks +6 days inclusive), admitted between January 1st 2005 and December 31st 2017. The study protocol was approved by the ethics committees of our center.
Exclusion criteria included: outborn infants, those affected by a TORCH (Toxoplasmosis, Other, Rubeola, Cytomegalovirus, Herpes) infection, a major congenital anomaly, a chromosomal defect, those with the diagnosis of an inborn error of metabolism performed at prenatal consultation or during the neonatal period, and in the face of diagnostic suspicion of a neuromuscular disease. A pre-structured data form (see supplemental material) including data on demographics, pregnancy, delivery, neonatal morbidity and outcome, and placental histology was completed for each infant.
We considered a cycle of antenatal steroids to be full when at least 12 hours had elapsed after the last dose of dexamethasone (four doses of 6 mg of dexamethasone given intramuscularly 12 hours apart) or betamethasone (two doses of 12 mg of betamethasone given intramuscularly 24 hours apart) administered to the mother, in line with the National Institutes of Health (NIH) Consensus Development Panel on the Effect of Corticosteroids for Fetal Maturation on Perinatal Outcomes [12]. Hypertension in pregnancy and preeclampsia were defined according to the American College of Obstetricians and Gynecologists [13]. Clinical chorioamnionitis was defined as maternal fever combined with two or more findings of maternal tachycardia, fetal tachycardia, leukocytosis and/or elevated C-reactive protein, uterine tenderness and/or malodorous amniotic fluid [14]. Histological chorioamnionitis was defined according to Blanc criteria [15]. Placental abruption was defined as a premature separation of the placenta before delivery [16]. Two groups were considered for the time length of rupture of membranes before birth, less than 18 hours and longer than 18 hours.
Gestational age was assessed from date of embryo transfer (for in vitro fertilization), menstrual age (women with regular menstrual cycles), ultrasound examination (when a discrepancy of one or more weeks existed between the age derived from menstrual dating and the age derived sonographically, or in the absence of a menstrual date) [17] or the New Ballard Score (in the absence of obstetrical indexes) [18]. Small for gestational age was defined as a birth weight below the 10th centile of Fenton's fetal growth charts [19].
Delivery room resuscitation practices were used according to the guidelines of the Portuguese Society of Neonatology (available at www.lusoneonatologia.com). We try, whenever possible, the use of early nasal continuous airway positive pressure (early NCPAP) in the first 15 minutes following birth. Respiratory distress syndrome (RDS) was made with a combination of clinical and radiographic features according to the criteria of RDS of the Vermont Oxford Network (VON), once our unit is a collaborative member. The VON criteria of RDS are: (1) PaO2 <50mmHg in room air, central cyanosis in room air, a requirement for supplemental oxygen to maintain PaO2 >50mmHg or a requirement for supplemental oxygen to maintain a pulse oximeter saturation over 85% within the first 24 h of life; and (2) a chest radiograph with reticulogranular appearance to lung fields with or without low lung volumes and air bronchograms, within the first 24 h of life. Noninvasive ventilation is the preferred mode of ventilation in preterm infants with respiratory drive. Conventional mechanical ventilation is used in cases of NCPAP failure or preterm infants without respiratory drive. The most used ventilation mode was synchronized intermittent positive pressure associated with volume guarantee. The strategy of permissive hypercapnia is advocated at our center. Caffeine citrate was routinely used since day one of life, independent of apneas, until 34 weeks of gestational age. Oxygen was administered to maintain saturations within 90–95%, given by pulse oximetry (SpO2). BPD was defined according to the National Institute of Health [20]. In this study, we considered BPD as the dependency on administering oxygen until 36 weeks of corrected age.
Parenteral nutrition is used as soon as there is clinical stability, preferably from day one of life, starting with a volume of 70 ml/kg/day, increasing daily in increments until 150 mg/kg/day by day seven of life. Incubator humidity is usually 70% during the first week. Hemodynamically significant patent ductus arteriosus (PDA) was screened and diagnosed on the basis of the echocardiographic findings. The first evaluation was usually performed between 24 and 72 hours of life, with daily evaluation until closure of the ductus. The standard treatment was ibuprofen. Proven neonatal sepsis was defined as any systemic bacterial or fungal infection documented by a positive blood culture. Early onset sepsis and late onset sepsis were considered in our analysis if diagnosed before and after the first 72 hours of life, respectively. Newborn infants affected by pneumonia were those who presented with signs of respiratory distress plus significant results from laboratory studies (anomalies of white blood cell count, increase in C-reactive protein level or positive blood, urine and/or broncho-tracheal sample cultures) and a radiography pattern compatible with a lung infection. Meningitis was considered in our analysis in the presence of elevated polymorphonuclear counts and biochemical alterations in the cerebro-spinal fluid, with or without isolation of microbiological agent. The modified criteria of Bell were used for the diagnosis and staging of necrotizing enterocolitis (NEC) [21]. Staging of ROP was done according to the International Classification [22, 23]. IVH was classified according to Volpe JJ [24]. PVL was classified according to de Vries L and Rennie JM [25].
For the purpose of this study, week days births were considered those delivered between eight and 20 h from Monday to Friday. The night period starts at 20 h in most Portuguese centers. We compared morbidity and mortality of infants admitted during the night period, weekends and holidays with that registered for infants admitted during weekdays, for the full 13-year period and for the epochs 2005–2008, 2009–2012, and 2013–2017.
Statistics
Data collection and statistical analysis were performed using IBM SPSS® Statistics 23. Categorical variables were described by their respective absolute and relative frequencies, continuous variables with symmetric distribution by mean (±standard deviation) and continuous variables with asymmetric distribution by median (minimum-maximum). Chi-square or Fisher’s exact test were applied to compare categorical variables, while Independent t-tests and Mann-Whitney U tests were used for normally and non-normally distributed continuous variables, respectively. A multivariate analysis by logistic regression was performed to evaluate the association of any morbidity with the time of birth (week days vs week nights, weekends and holidays). A p-value <0.05 was considered statisticallysignificant.
Results
Out of a total of 256 inborn preterm neonates, 220 (85.9%) were included in the study. There were 116 (52.7%) males and 104 (47.3%) females with a median gestational age of 27 weeks (min = 23; max = 29) and a median birth weight of 922 g (min = 360; max1555); 33 (15%) were small for gestational age. Ninety-five (43.2%) infants were delivered during weekdays and 125 (56.8%) were delivered during weeknights and weekends and holidays, Figs. 1 and 2.
Flow chart of patients.
Distribution of patients along years.
Data on mother, pregnancy, delivery, demographics, neonatal morbidity and mortality according to time of birth are reported in Tables 1 and 2. C-sections were more frequent during weekdays and infants delivered to mothers with chorioamnionitis with chorionic vasculitis born more frequently during weeknights, weekends and holidays (Table 1). Small for gestational age infants were more frequently delivered during week days(Table 2).
Data on mother, pregnancy and delivery, according to time of delivery
aindependent T-test; bchi-square test; cFisher's exact test; HELLP - hemolysis, elevated liver enzyme levels, and low platelet levels; SD standard deviation; bold p – statistically significant.
Data on demographics, transition, neonatal morbidity and mortality according to time of birth
bchi-square test; cFisher's exact test; dMann-Whitney U test; cPVL – cystic periventricular leukomalacia; GA- gestational age; HS-PDA- haemodinamically significant patent ductus arteriosus; IVH- intraventricular hemorrhage; MV – mechanical ventilation; NCPAP – nasal continuous positive pressure; NEC- necrotizing enterocolitis; PVI- periventricular venous infarction; RDS – respiratory distress syndrome; ROP- retinopathy of prematurity; TPN- total parenteral nutrition; bold p – statistically significant.
In the univariate analysis cPVL was the only morbidity that was significantly increased in infants delivered during week nights, weekends and holidays (Table 2). In the multivariate analysis, after adjusting for gestational age, birthweight, small for gestational age, C-section and chorioamnionitis with chorionic vasculitis, cPVL was no longer significantly associated to a deliver during weeknights, weekends and holidays (ORadj = 0.580; 95% CI: 0.19–1.71, p = 0.324).
The morbidity and mortality, according to the time of birth over time (periods 2005–2008; 2009–2012; 1013–2017) are reported in Table 3. In the univariate analysis, cPVL has significantly increased during the night and week-end/holiday period in the epoch 2013–2017. (p = 0.023). In the multivariate analysis, after adjustment for birth weight, gestational age, small for gestational age, C-section, days of mechanical ventilation, days of dependency on oxygen, IVH, sepsis and chorioamnionitis the association was no longer significant (OR = 9.601; 95% CI:1.00–92.17, p = 0.006).
Demographics and morbidity and mortality according to time of birth over time (n = 220)
aindependent T-test; bchi-square test; cFisher's exact test; dMann-Whitney U test; cPVL – cystic periventricular leukomalacia; GA- gestational age; HS-PDA- haemodynamically significant patent ductus arteriosus; IVH- intraventricular haemorrhage; NEC- necrotizing enterocolitis; PVI- periventricular venous infarction; RDS – respiratory distress syndrome; ROP- retinopathy of prematurity; bold p – statistically significant.
NICU’s staffing is more often lower during nights, weekends and holidays than weekdays, and this fact may contribute to higher morbidities and mortality in patients admitted on these periods. In this study, we have assessed and compared the neonatal outcome of very preterm infants, according to timing of birth, weekdays versus nights, weekends and holidays, at our NICU. We did not find any increased morbidity associated with a birth during nights, weekends and holiday shifts. After reviewing the available literature we did not find any study assessing the outcomes of very preterm infants according to timing of birth. Studies in this field are lacking.
Our NICU is a level III and a referral centre for extreme prematurity, congenital heart defects and surgical disorders, with about 400 admissions per year. The unit has nine intensive care and eight special care posts. The ratio nurses: patient is 1:2 and 1:4 in intensive and special care, respectively, and this ratio is independent of week day or night. The numbers of doctors and nurses have changed along the study period due to the age of exemption from the emergency service, retirement, pregnancy and health licenses, and these aspects were not covered by substitutions, mainly due to the economical and restrictive policy. The number of fellow physicians also suffered many variations throughout the study period. The NICU has attempted to keep 2–4 physicians in the morning shifts (8–14 am), two during the afternoon shift (2 to 8 pm) and 2 during the evening (8pm-8am) and weekends and holidays. These doctors provide care to all patients in the unit and support the delivery room. Our delivery room has a number of around 2,500 births per year. To maintain this number of professionals it is necessary that they perform work in extraordinary regime, including several night services per week and several weekend services per month.
Care for the newborn in the health area has undergone important advances in recent decades, through the increasing in the production and diffusion of scientific knowledge combined with technological development. On the other hand, there are worrying issues such as the quality of service and patient safety in this innovative, complex and critical scenario of hospital care. The subject of patient safety gained global prominence following the release of the report To Err is Human: Building a Safer Health Care System [26]. This report has spawned a worldwide patient safety movement by revealing that patients are exposed to risks that may interfere with their safety, causing errors and damage to health. In addition, it demonstrated high rates of mortality caused by the health care of American hospitals, and a large part of the deaths could have beenavoided [27].
Preterm infants may be at greater risk due to particular factors of these patients, such as extreme physiological fragility and organic systems in development, which impair their safety because they require specific care. Patient safety may be affected in some cases by NICU environment issues, when there is inadequate infrastructure, overcrowding, inadequate staffing, work overload, limited resources or poor quality of equipment and materials [27].
In our study, in the univariate analysis, cPVL was the only morbidity that was significantly increased in infants delivered during weeknights, weekends and holidays. Nonetheless, this association was no longer found in the multivariate analysis. We also analyzed three separated epochs (2005–2008; 2009–2012; 2013–2017) assuming that some practices have changed over time. Similar results were found for cPVL in the epoch 2013–2017.
In this study, authors sought to evaluate the neonatal outcome of preterm newborns born in periods of greater risk of error due to the lower number of physicians and the fatigue of professionals. Although they did not find differences in neonatal morbidity and mortality, the authors intend to divulge their experience and emphasize the importance of this type of studies in order to prevent factors such as lack of human resources in units of high risk that can contribute to increased morbidity and mortality, as well as long-term sequelae.