Setting and participants
This is a retrospective cohort study. The study protocol was approved by the research institute’s committee of human research in the Second Affiliated Hospital of Shantou University Medical College (NO.2018-23) and abided by the standards of the Declaration of Helsinki. The data were anonymized in this study, so we did not use the consent to participate.
We collected data of pregnant women who gave birth at 34 (0/7) to 36 (6/7) weeks gestation from January 2014 to June 2019. Data were excluded from this study if the pregnant women or preterm infants met the following criteria: (1) pregnant women had serious liver, kidney, lung or heart disease before or during pregnancy, (2) pregnant women received ACS treatment before 34 gestational weeks, (3) preterm infants had a congenital malformation or needed surgery.
Data collection
Variables included age of mother, gestational diabetes mellitus, pregnancy hypertension, method of delivery, premature rupture of membranes (PROM), condition of the placenta, meconium-stained amniotic fluid, multiple gestation, gestational age, birth weight of preterm infants, asphyxia, pulmonary surfactant treatment for preterm infants, and mechanical ventilation for preterm infants. The short-term outcomes of preterm infants in this study included the length of neonatal hospital stay, hospitalization expenses for preterm infants, NRDS, neonatal pneumonia, neonatal hypoglycemia, neonatal sepsis, necrotizing enterocolitis of newborn, neonatal intracranial hemorrhage, and hypoxic-ischemic encephalopathy.
Exposure factor in this cohort is ACS treatment
The patient cohort was divided into two groups: the ACS group and the without-ACS group. Antenatal corticosteroid treatment consisted of four doses of dexamethasone (6 mg intramuscular) 12 hours apart [13]. Whether to use antenatal corticosteroid after 34 weeks of preterm delivery or not was based on the choice of the pregnant women.
Assessment short-term outcomes of preterm infants
We compared nine short-term outcomes of preterm infants between the ACS group and without-ACS group. Complications of premature infants were diagnosed by doctors in the Neonatal Department, who did not know we were going to conduct this retrospective study when they made diagnosis. All doctors in the Neonatal Department made the diagnosis of complications according to standard diagnostic criteria. Assessment methods of these outcomes is described below. Neonatal hypoglycemia was defined as a glucose level less than 2.2 mmol per liter [14].
A diagnosis of NRDS was based upon the findings of respiratory difficulty (cyanosis, grunting, nasal flaring, or tachypnea) that necessitated mechanical ventilation support, and was furthermore consistent with typical radiological findings of the lung (such as frosted glass-like changes, air bronchogram, and white lung). Laboratory findings were characterized initially by hypoxemia and later by progressive hypoxemia, hypercapnia, and variable metabolic acidosis. The clinical course, chest x-ray findings, and blood gas and acid-base values helped to establish the clinical diagnosis of NRDS [15].
Diagnosis of neonatal sepsis was based on symptoms and laboratory evidence. Initial symptoms might be few, and included apnea, tachypnoea, or tachycardia. Late complications of neonatal sepsis might include respiratory failure, pulmonary hypertension, cardiac failure, shock, renal failure, liver dysfunction, cerebral edema or thrombosis, adrenal hemorrhage or insufficiency, bone marrow dysfunction, and disseminated intravascular coagulation. Neonatal sepsis can be diagnosed when blood or other sterile body site culture produces positive pathogenic bacteria or opportunistic pathogens. The inflammatory response was measured by blood count, C-reactive protein, procalcitonin, interleukin 6, interleukin 8, and tumor necrosis factor. [16].
Gastrointestinal signs of necrotizing enterocolitis included feeding difficulty, gastric retention, abdominal distension, bilious vomiting, and stool with blood [17]. Necrotizing enterocolitis of newborn was defined according to Bell’s criteria ≥ stage 2A. Diagnosis was made based on plain abdominal radiographs. A finding of pneumatosis in the intestinal wall confirmed the clinical suspicion and diagnosis of necrotizing enterocolitis of newborn.
Neonatal pneumonia was inflammation of the lung caused by infection. Diagnosis of neonatal pneumonia was made according to the risk factors, such as PROM, chorioamnionitis in the mother, and low birthweight, which predispose to pneumonia. Infants with respiratory distress usually required investigation to identify infection. A chest X-ray, showing increased bronchovascular shadows with small patchy and macular shadows helped to diagnose neonatal pneumonia [18].
Intracranial hemorrhage was suspected on the basis of the history, clinical manifestations, and knowledge of the birthweight-specific risks for intravascular hemolysis. Ultrasonography was the preferred imaging technique for screening. All at-risk infants underwent cranial ultrasonography within the first 3-7 days of age [19].
A diagnosis of hypoxic-ischemic encephalopathy was made according to the pH value of the fetal umbilical artery, the Apgar score at 5 and 10 minutes, and multi-system organ failure, including combined kidney damage, liver damage, blood abnormalities, heart dysfunction, metabolic disorders, and gastrointestinal tract injury [20]. Magnetic resonance imaging was used to evaluate extensive periventricular injury. All at-risk preterm infants underwent an MRI within the first week of age.
Because this is a single center study, we compared the neonatal hospitalization expenses directly. Hospitalization expenses included the examination, drug, and nursing care costs. Length of hospital stay was counted from the day of admission to the day of discharge.
Statistical analysis
We used the Shapiro-Wilk test to determine whether continuous variables were normally distributed, and the Wilcoxon-Mann-Whitney U-test was conducted for skewed distributions (presented as the median and the min-max range). Descriptive statistics for categorical variables are shown as frequency (percentage). The Pearson chi-square test or Fisher's exact test were used to compare categorical variables, as appropriate. Collinearity among all covariates was assessed using the Spearman correlation test [21].
For binary variables, logistic regression was used to analyze the risk factors of neonatal short-term outcomes, and independent variables were chosen based on clinical knowledge. For continuous dependent variables, linear regression was used to analyze the relationship between independent variables. Regression analysis was performed by a forward stepwise method to identify the risk factors. Estimated slope and 95% confidence intervals (CI) were obtained. Statistical analyses were performed using SPSS 24.0 (SPSS, Chicago, IL). P-values of less than 0.05 were considered to be statistically significant.