Pregnancy risk factors and outcomes in the transition from respiratory distress syndrome (RDS) to RDS and acute respiratory distress syndrome (ARDS) in preterm infants: a retrospective cohort study

Background: Neonatal acute respiratory distress syndrome (ARDS) has been proposed in 2017. Growing evidence suggests that pregnancy risk factors (PRF) contribute to the deterioration of respiratory distress. The present study aims to clarify PRF and outcomes in the transition from respiratory distress syndrome (RDS) to RDS and ARDS in preterm infants. Methods : A retrospective study was conducted at a tertiary neonatal intensive care unit from Jan-1, 2017 to July-30, 2018. Preterm neonates diagnosed with RDS were enrolled at baseline and followed up during hospitalization. The primary outcomes were to identify the relationships between PRF and transition from RDS to RDS and ARDS and the incidences of bronchopulmonary dysplasia (BPD). Results: A total of 269 preterm infants were included. Of them, 168 were diagnosed with both RDS and ARDS, and 101 with RDS. Comparing with preterm neonates with both RDS and ARDS, that with RDS was related to the decreased incidence of BPD (22:79 vs 61:107, 95%CI:0.27-0.85, P =0.010), especially in the subgroup of severe BPD (3:98 vs 17:151, 95%CI:0.08-0.95, P =0.032) and two or more of surfactant (14:87 vs 43:125, 95%CI:0.24-0.91, P =0.023). PRF including intrahepatic cholestasis during pregnancy (ICP), pernicious placenta previa (PPP), hypertensive disorder complicating pregnancy (HDCP) and gestational diabetes mellitus (GDM) were not shown differences between the two groups. Conclusion : Preterm infants with both RDS and ARDS were easier to convert to BPD as compared with that with RDS. Trial registration: Chinese Clinical Trial Registry, ChiCTR1900026980


Introduction
Neonatal respiratory distress related to bronchopulmonary dysplasia (BPD) is a major health problem affecting the increasing premature birth population [1]. It is of great importance to recognize and treat the preterm infants at the early stage of respiratory distress. Neonatal respiratory distress syndrome (RDS) is associated with an increased risk for progression to BPD, and its incidence has been con rmed to be gradually increased with the decreased gestational age [2]. Exogenous surfactant replacement is a key treatment to reduce the incidence of BPD [3]. In 2017, the international neonatal acute respiratory distress syndrome (ARDS) collaborative group provides the rst consensus de nition for neonatal ARDS [4]. To date, no studies have indicated bene cial effects of surfactant on adult and pediatric ARDS [5,6], its action for neonatal ARDS is therefore needed to be further elucidated [7]. Otherwise, the impact of RDS and/or ARDS on BPD is also unclear.
Given RDS is a disease of respiratory distress that commences within 24 h after birth [3], and most ARDS also appears in the rst three days [5]. The optimal active measures should be taken before birth.
Pregnancy risk factors (PRF), including gestational diabetes mellitus (GDM), hypertensive disorder complicating pregnancy (HDCP), and intrahepatic cholestasis during pregnancy (ICP), are related to increase the incidences of preterm birth and RDS [3,8]. The identi cation and subsequent active management for PRF before neonates' birth should be more bene cial strategies for preventing deterioration of RDS. We have found that PRF is related to the deterioration of RDS in a Chinese cohort [9]. However, no studies report the relationships between PRF and the transition from RDS to RDS and ARDS in preterm infants, and it also remains unknown whether stayed in RDS is related to the decreased incidences of BPD, as well as other complications.
The aims of the present study are: 1) to report the effects of PRF on the transition from RDS to RDS and ARDS, and 2) to clarify whether stayed in RDS, as compared with transition to RDS and ARDS, is associated with the decreased incidence of BPD. This was a retrospective study conducted in the neonatal intensive care unit (NICU), Children's Hospital of Chongqing Medical University from Jan-1, 2017 to July-30, 2018.

Methods
Eligibility requirements for neonates were as follows: (1) The gestational age was less than 37 weeks and admitted to NICU in 24 h after birth; (2) These neonates were initially diagnosed with RDS. Exclusion criteria were: (1) major congenital anomalies; (2) chromosomal abnormalities; (3) neuromuscular diseases; (4) upper respiratory tract abnormalities.

Diagnosis of neonatal RDS
The diagnosis criteria of RDS was based on clinical manifestations and chest X-ray ndings. The clinical signs and symptoms of RDS were respiratory distress, tachypnea, nasal aring, groan, and cyanosis appear within 24 h after birth, as well as bene cial response to pulmonary surfactant and/or lung recruitment strategies. The other criteria include the typical X-ray picture of a grain shadow, air bronchogram or white lung [3].
Diagnosis of neonatal ARDS. of surfactant administration were 6 to 12 hours without more than four doses allowed, and the second and later dose was 100 mg/kg.
Also, these preterm infants received a loading dose of 20 mg/kg caffeine citrate and a maintenance dose of 10 mg/kg/d until 34 weeks of GA. Other care was at the discretion of the attending neonatologist.
De nition of PRF.
Four individual PRF were analyzed in the present study, including ICP, HDCP, GDM, and pernicious placenta previa (PPP). Diagnosis of PRF were consistent with recommendations and guidelines from international consensus les [10] and Chinese Medical Association [11][12][13][14]. The preterm infants with RDS were considered to be with PRF as long as having one or more individual PRF. The diagnosis of BPD is consistent with de nition of BPD [15].
Evaluation of primary and secondary outcomes.
The primary outcomes were: 1) to report the effects of PRF on the transition from RDS to RDS and ARDS in preterm infants, and 2) to clarify whether stayed in RDS, as compared with transition to RDS and ARDS, is associated with the decreased incidence of BPD. And the secondary outcomes is mortality and other complications.

Data Collection
The clinical data of all enrolled neonates were collected in standardized case report forms, including gender, gestational age, birth weight, Apgar score, administration of surfactant, respiratory mode, PRF, death, BPD, retinopathy of prematurity (ROP), necrotizing enterocolitis (NEC), intraventricular hemorrhage (IVH), and so on.
Sample Size Estimation.
The sample size estimation was calculated by PASS software (2008 v8.0.3). According to previous studies [16], the incidence of RDS in newborns from ICP was about 28%. Our previous study have indicated that the incidence of severe RDS in newborns from ICP was about 7% [9], the incidence of ARDS was therefore thought to no less than 7%, if the severe RDS was considered as ARDS. With 80% power and a 2-sided signi cance level of 0.05, 21 neonates would be needed at least in RDS and ARDS conversion group. According to the nature of cohort study, the number in the RDS group was usually not less than RDS and ARDS conversion group, so at least 42 neonates would be needed in preterm infants from ICP. In the present study, four PRF were included, the total number was therefore 168. Actually, during the study period, 269 preterm infants were enrolled. Therefore, the actual sample size was more than theoretical need.
Statistical analysis.
Continuous variables, expressed as mean ± standard deviation, were compared using independent samples t test. Categorical variables were compared using the χ2 test or the Fisher's test. The prede ned BPD subgroups were: mild, moderate and severe BPD. To further assess the effects of surfactant administration on the rate of BPD between the two groups, use of surfactant were divided into three subgroups: zero, one and two of surfactant. Subgroup analyses were conducted for the primary outcome.
All analyses were carried out using computer software (SPSS 16.0 for windows). A p-value less than 0.05 was regarded as statistically signi cant.

Results
From Jan-  1) Baseline characteristics for the included preterm neonates.
All the included neonates were born in other hospitals and transferred to the NICU due to respiratory distress within one week after birth. The baseline characteristics of neonates who stayed in RDS and who converted to RDS and ARDS are presented in table 1. There were also no signi cant differences in the main clinical characteristics of neonates, including gestational age, gender, birth weight, Apgar score, and invasive ventilation between two groups. The primary noninvasive modes were nasal continuous positive airway pressure (NCPAP) and nasal intermittent positive pressure ventilation (NIPPV). Compared with neonates who converted to RDS and ARDS, those who stayed in RDS had more primary mode of NCPAP To further assess the effects of surfactant administration on the rate of BPD between the two groups, the subgroup analysis was also conducted. As far as two or more of surfactant was considered, there was still decreased incidence of BPD in the preterm infants with RDS as compared with that with both RDS

Discussion
In the retrospective study, we aimed to report the effects of PRF on the transition from RDS to RDS combined with ARDS, and to clarify whether preterm infants with RDS, as compared with both RDS and ARDS, is associated with the decreased incidence of BPD. As a result, we found that PRF were not related to the increased conversion from RDS to RDS and ARDS. But, comparing with preterm neonates with both RDS and ARDS, that with RDS was related to the decreased incidence of BPD (22:79  We also showed that the effects of RDS and RDS combined with ARDS on the incidence of BPD were partially associated with the exogenous surfactant replacement. When only one of surfactant was administrated, no difference in the incidence of BPD was found between the two groups. But when more than two of surfactant was needed, there was decreased incidence of BPD in the preterm infants with RDS as compared with that with RDS and ARDS (14:87 vs 43:125, 95%CI:0.24-0.91, P = 0.023). To our knowledge, this is the rst study to assess the effects of surfactant on BPD between preterm infants with RDS and RDS combined with ARDS. The result suggests that preterm infants with both RDS and ARDS is associated with the increased incidence of BPD as compared with that with RDS when two or more of surfactant were needed.
Besides caffeine [17], early use of NCPAP has been proved to be one of the most effective pathways to reduce the incidence of BPD. Supplying with an intermittent peak pressure on NCPAP, NIPPV is considered as a strengthened version of NCPAP with increased ow delivery in the upper airway, increased minute volume and functional residual capacity and recruitment of collapsed alveoli, improved stability of the chest wall and reduced asynchrony of thoraco-abdominal movement [18][19][20][21]. And it should be related to the decreased incidence of BPD. However, several studies have compared the effects of NIPPV and NCPAP on the incidence of BPD, and the results were no advantages of NIPPV over NCPAP. Meneses J et al [22] indicated that NIPPV did not signi cantly reduce the incidence of BPD comparing with NCPAP, of which mean gestational ages were about 30 weeks. Our previous study did not also show differences in the incidence of BPD [23], which was consistent with the meta-analyses [24,25]. The largest multicentered study on the comparison of NIPPV and NCPAP also showed that, in very preterm infant, there was no signi cant difference on the rate of intubation and survival to 36 weeks of post-menstrual age without BPD [26]. And the authors suggested that there might be some subtle but important differences between extremely preterm and preterm infants, but they did not tell us what the differences were.
One important cause to induce the difference may be the diagnosis of RDS and ARDS. the above studies were mainly performed in the pre-neonatal acute respiratory distress syndrome (ARDS) era, in which ARDS was usually considered as severe RDS or severe respiratory failure. Actually, ARDS and RDS are two diseases according to their etiological and pathological characteristics. To further reduce the heterogeneities and exclude the potential effects of RDS on outcomes in neonates with ARDS, the keywords "severe RDS" or "severe respiratory distress" were hypothetically regarded as ARDS. An interesting result was that, as compared with conventional mechanical ventilation, high-frequency oscillatory ventilation reduced the mortality (2:9 vs 3:3, 95%CI 0.09-0.89, P = 0.03) (4:177 vs 13:179, 95%CI 0.10-0.94, P = 0.04) [27,28], and the incidence of BPD at 36 weeks (13:173 vs 28:166, 95%CI 0.24-0.83, P = 0.01) [28], and improved long-term neurodevelopment outcomes [28]. Recently, we also reported a randomized controlled trial, and the result demonstrated that, compared with NCPAP, nasal high frequency oscillatory ventilation signi cantly reduced the need for endotracheal ventilation in the infants with ARDS and/or RDS (10:33 vs 21:15 95%CI 0.08-0.57, P = 0.002), but not in the neonates with RDS and ARDS (2:7 vs 11:6 95%CI 0.02-1.00, P = 0.097) [29]. And the result was consistent with the present study.
Another important cause to induce the difference might be the administration of surfactant. In fact, no studies have indicated bene cial effects of surfactant for adult and pediatric ARDS [5,6], including meconium aspiration syndrome, and which is a subtype of ARDS. Cochrane review showed that, although reducing the severity of respiratory illness and decreased number of infants with progressive respiratory failure requiring support with extracorporeal membrane oxygenation in infants with meconium aspiration syndrome, surfactant administration does not reduce the incidence of BPD [30]. And they were consistent with the present study, in which two or more of surfactant was related to the increased incidence of BPD in the preterm infants with both RDS and ARDS as compared with that with RDS.
In the present study, the prevalence of RDS in preterm neonates was 34.7% (269/775) and the conversion incidence from RDS to RDS and ARDS was 21.7% (168/775). Even so, the rate of ARDS was thought to be under-recognized. An observational study in 2016 including 459 ICUs in fty countries showed that the clinical recognition rates ranged from 51.3% for mild ARDS to 78.5% for severe ARDS [32].
There were some basic limitations due to the observational nature of the present study. 1. there were small neonates with PRF. 2. the severity of ARDS was not measured quantitatively. These factors could induce potential bias, including restricted application scope and size effect. These problems could be overcome in additional prospective studies. A de nitive answer to whether PRF increased the conversion from RDS to RDS and ARDS and incidence of BPD would require a larger sample size in randomized controlled trials.

Conclusions
In summary, preterm infants with both RDS and ARDS were associated with the increased incidence of BPD, especially in the subgroup of severe BPD and two or more of surfactant as compared with preterm neonates with RDS. But, PRF was not related to the increased conversion from RDS to RD S and ARDS.
Due to small neonates with PRF, the relationship between PRF and the conversion from RDS to RDS and ARDS need further assessment.

Availability of data and materials
The datasets generated and/or analysed during the current study are not publicly available due hospital policy but are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The study was approved by the Ethics Committee of Children's Hospital of Chongqing Medical University.   Figure 1 Flowchart of study enrollment: a total of 775 newborn infants were screened and 269 subjects were included in the analysis.