Association Between Blood Gas Within 1week of Life and Bronchopulmonary Dysplasia in Preterm Infants With Less Than 32 Gestational Weeks

Objective: This study tested wether preterm infants of (cid:0) 32 gestational weeks (GWs) with a blood gas derangement within 7days of life are at increased risk for moderate and severe bronchopulmonary dysplasia (BPD). Method: 236 preterm infants with (cid:0) 32 GWs from January of 2017 to December of 2019 were included in this study. First, univariate analysing determined wether there existed associations between BPD (moderate and severe) and blood gas values, clinical characteristics, interventions, daily given liquid and eneregy within 7days of life. Then multivarariate regession analysis was performed to know wether there were relationships between BPD (moderate and severe) and risk factors between which and BPD (moderate and severe) univariate analysing showed that P value was less than 0.1. Results: From univariate analysing, we found that PaO2, PaCO2 and HCO3 in 7 th day of life, cesarean section (OR=0.508, 95%CI:0.275-0.94), getational age (GA, OR=0.163, 95%CI:0.077-0.344), birth weight (BW, OR=0.122, 95%CI:0.054-0.273), PDA (OR=2.839, 95%CI:1.1463-5.508), early onset infection (OR=3.00, 95%CI:1.483-6.069), and mechanical ventilation (MV, OR=4.562, 95%CI:2.405-8.653) were signicantly associated with moderate and severe BPD. Because there existed close relationship between GA and BW (R=0.642, P=0.000) and BW dispersion was big in this group, we excluded BW in multivariate analysing. From multivariate analysing, besides GA (Exp (B)=0.176, 95%CI:0.08-0.389) , MV (Exp (B)=3.515, 95%CI:1.746-7.076), PaO2 (Exp (B)=0.468, 95%CI:0.226-0.969) in 7 th day of life was the independent risk factor for moderate and severe BPD in the preterm infants of (cid:0) 32 GWs. Conclusion: Preterm infants of (cid:0) 32GWs with blood gas derangements within 7days of life could be at risk of moderate and severe BPD. and 21d of life). Their results show that early anemia was associated with an increased risk of BPD.We cannot nd the relationship between anemia and BPD. It may be caused by that we collected the anemia data within 4weeks of life and did not separate into early anemia (less than 2weeks of life) and late anemia (more than 2weeks of life).The association between BPD and anemia and how anemia works on BPD remain to further study. Rocha reported that there existed a signicant association between SGA and BPD (OR = 5.2 (CI: 1.46-18.58), p = 0.01) by multivariate analysis[32]. The result was different from ours, we can not nd that BPD was associated with SGA. There exist differences between both in the incidence of SGA (up to 18.6% in this study) and in the study population.


Background
Preterm birth is rising around the world, and the incidence of preterm is different ranging from about 5% in several European countries to 18% in some African countries [1] [2].Bronchopulmonary dysplasia (BPD) is the most common chronic lung disease in preterm infants. BPD is a common cause of death in preterm infants occupying 2.8% [3]. During clinical work, we can not change the fact of preterm delivery. We want to know what we can do to reduce and/or alleviate the development of BPD in the preterm babies, especially in rst several days of life. As we know, prolonged exposure to high oxygen concentration is identi ed as an important contributor to retinopathy of prematurity (ROP) [4]. There exists relationships between severe ROP and blood oxygen, carbon dioxide, and pH levels during the rst 3 postnatal days [5] [6]. We deduced that blood gas values (PH, PaCO2, PaO2, HCO3) in preterm infants be very important for mature of premature organs and systems, and changes of blood gas values in the rst several days of life in uence the development of premature lung and have an effect on occurence or developent of BPD in preterm infants. If that, we could control blood gas values to reduce and/or alleviate the development of BPD in the preterm babies. It is reported that BPD is associated with exposure to antenatal and/or postnatal factors including gestational age, birth weight, gender, patent ductus arteriosus (PDA), mechanical ventilation (MV), oxygen toxicity, and infection [7] [8]. In this study, we collected and analyzed such factors as gestational age (GA), birth weight (BW), cesearen section (CS), early onset infection (EOI), anemia, MV (more than 2 days) and blood gas values within 1week of life (1 st , 2 nd , 3 rd , and 7 th day after birth). We didn't collect data of apgar score due to apgar score in very preterm infants (VPI) and extremly preterm infants (EPI) is lower than that in term babies and low apgar score in VPI and EPI is not equal to neonatal asphyxia, and antenatal corticoid therapy because of antenatal corticoid therapy being as a therapy principle of preterm delivery in the obstetrical department of the rst a liated hospital of Nanjing Medical University.

Materials And Methods
This retrospective study was conducted at neonatal department of the rst a liated hospital of Nanjing Medical University. Admitting criteria for the study: The baby was admitted to the neonatal department from January of 2017 to December of 2019, The preterm infant was 32 gestatioal weeks (GWs) , The preterm baby had no serious congenital anomalies. Exclusion criteria: Infants were with congenital anomalies such as congenital lung anomaly, congenital diaphragmatic hernia, severe congenital heart diseases and primary persistent pulmonary hypertension, Recorded information was imcomplete, An infant was dead and/or refused treatment by parents within 2weeks of life. Excluding of 17cases, we collected 236 cases of preterm infants, who were from 24.9 GWs to 32 GWs (median: 30.4weeks), whose BW was from 800g to 3250g (median:1400g). The study was approved by the ethics review committee of the rst a liated hospital of Nanjing Medical University (2014SR038).

Statistics
Categorical variables were compared using χ2-or Fisher's exact tests. Univariate analysis was used to determine relative risk and 95% con dence intervals. Logistic regression analysis was used to determine association between BPD and arterial blood gas values on postnatal days 1, 2, 3,and 7 of the preterm infants who were less than 32GWs. All analyses were performed using Stata software version 19.0 (StataCorp LP, College Station, TX, USA). Signi cance was set at P 0.05. Clinical de nitions BPD was de ned as required additional oxygen and/or respiratory support should be equal to or more than continuous 28 days before assessment at 36 weeks of postmenstrual age in preterm infants and classi ed as mild, moderate and severe based on the required fraction of inspired oxygen at 36 weeks' PMA: mild, no supplemental oxygen; moderate, supplemental oxygen ≤30%; and severe, respiratory support and/or oxygen >30% [9]. Here, oxygen therapy or respiratory support indicated the infant's usual therapy and for oxygen therapy, the treating time was equal to or more than 12 hours of a day in the assessing days, not a transient or acute event; target oxygen saturation is 90-95% [09] [10]. GA was established based on last menstrual period con rmed by an ultrasonographic examination prior to 20 weeks [11]. Small for gestaional age (SGA) was de ned as infant birth weight less than 10th percentile of BW for the same GA [12]. Early onset infection (EOI) was de ned as con rmed infection (including neonatal sepsis, urine tract infection, respiratory infection, etc.) within 7days after birth [13]. Anemia was de ned as Hct <39% [14]. Patent ductus arteriosus (PDA) was diagnosed by two-dimensional color Doppler examination in 1 st week of life.

Incidence of BPD in the preterm infants
We collected 253cases with less than 32 weeks of GA in total. After exclusion of 17 cases including incomplete recorded information, dead, refused treatment by parents and severe congenital anomalies, the left 236 preterm infants were included to be analyzed. There existed 54 cases with moderate and severe BPD. The incidence of moderate and severe BPD was 22.88% in this study.

Classi cation
We choosed the extremely abnormal blood gas values each day (1 st , 2 nd , 3 rd ,and 7 th day of life). When the baby's condition was stable, bood gas could just be done one time per day, and the blood gas value was included. We separated each day blood gas values, daily given liquid and energy into a low group and a high group according to each median value(see table2, table3 and table4).

Associations between BPD and clinical characteristics, interventions, each day blood gas values, daily given liquid and eneregy by univariate analysing
From the univariate analysing, we found that such clinical characters as GA, BW, CS, PDA, EOI and MV were signi cantly associated with BPD . The less GA and lower BW, the more moderate and severe BPD. The preterm infants with PDA, EOI, MV and non CS were at high risk of moderate and severe BPD (see gure1).
By univariate analysing relationships between moderate and severe BPD and daily given liquid and energy within 1 st week of life, the daily given energy in the 7 th day of life of preterm infants with BPD was lower than that preterm infants with no BPD, no signi cant difference of daily given liquid in the 1 st week of life between preterm infants with BPD and those with no BPD (see gure2 and gure3).
From the univariate analysing, we found that the blood gas values in the 7 th day of life such as PCO2, PO2 and HCO3 were signi cantly associated with BPD. Speci cally, higher PCO2, lower PO2 and lower HCO3 menifasted in preterm infants with BPD in 7 th day of life than those in preterm infants with no BPD (see gure4).

Relationships between moderate and severe BPD and risk factors by multivariate analysing
Although univariate analysing relationship between BW and BPD showed that P value was less than 0.01, there existed close relationship between GA and BW (R=0.642, P=0.000) and BW's dispersion was big (800-3250g), we excluded BW in multivariate analysing from the rst step. From multivariate analysing relationships between BPD and the above risk factors between which and BPD univariate analysing showed that P value was less than 0.1, GA, MV and PO2 in 7 th day of life were the independent risk factors of BPD in the preterm infants of less than 32GWs (see table5).

Discussion
The United States National Institutes of Health (NIH) reported an increase of BPD in the past 20 years, more than 40% preterm infants of 22-28 GWs with BPD [15]. The incidence of BPD in preterm infants of 32GWs was reported to be 30% in the United States[16] [17]. The incidence of moderate/severe bronchopulmonary dysplasia was 22% in less than 1500g of birth weight in one study [18].As to this study, the incidence of morderate and severe BPD was 22.9% in preterm infants with less than 32 GWs. It was similar to the previous study. Infants born at 22-25 GWs have a greater than 50% risk of BPD, and infants born at 26-27 GWs still have greater than 30% chance of developing BPD [19]. The incidence of BPD in infants with BW<1000g, 1000-1249g and 1250-1000g was 62.3%, 25.9% and 17.3% respectively [20] .In this research, not only from univariate analyzing but also from multivariate analyzing, we found that BPD was inversely associatd with GA (P=0.016, OR=0.342, 95%CI: 0.143-0.819), namely, the less GA/BW in preterm infants, the higher incidence of BPD.
Mechanical ventilation is an important risk factor for the development of BPD[21] [22]. Because invasive mechanical ventilation could damage premature pulmonary, there is a high morbidity of BPD when preterm infants received long periods of invasive mechanical ventilation [23].Watterberg reported that reducing mechanical ventilation may improve pulmonary outcomes in ELGAN infants [24]. A metaanalysis compared LISA with mechanical ventilation, LISA had a lower odds of BPD(OR, 0.53; 95% CI, 0.27-0.96) [25]. It was con rmed by this research that mechanical ventilation is associated with BPD in preterm infants of less than 32GWs (OR, 3.515, 95% CI, 1.746-7.076). PDA has also been recognized as a risk factor of BPD [26]. PDA has been linked to the development of BPD [27]. By univariate analyzing, PDA was associated with BPD in this group preterm infants(OR=2.839, 95%CI: 1.1463-5.508). However, in multivariate regession analyzing, PDA was not associated with BPD. PDA is often present during the period that preterm infants are very susceptible to lung injury leading to BPD. There is overlaping time of PDA existing and lung injury. To day, it remains unclear wether PDA is a true causative risk factor for BPD or just a physiological marker associated with the development of BPD [27].
Duan reported that multivarariate logistic regression analysis was performed to determine the association between anemia and BPD [31]. Among 243 preterm infants, the incidence of anemia was higher in BPD patients than that in non-BPD patients (p<0.001). Mean Hct in BPD patients was lower than that in non-BPD patients at different time points( 1d, 7d, 14d, and 21d of life). Their results show that early anemia was associated with an increased risk of BPD.We cannot nd the relationship between anemia and BPD. It may be caused by that we collected the anemia data within 4weeks of life and did not separate into early anemia (less than 2weeks of life) and late anemia (more than 2weeks of life).The association between BPD and anemia and how anemia works on BPD remain to further study. Rocha reported that there existed a signi cant association between SGA and BPD (OR = 5.2 (CI: 1.46-18.58), p = 0.01) by multivariate analysis [32]. The result was different from ours, we can not nd that BPD was associated with SGA. There exist differences between both in the incidence of SGA (up to 18.6% in this study) and in the study population.
Between days 7 and 27 after birth, every 10 kcal/kg/d increase in energy intake was associated with a 9% reduced risk of BPD (p =0.029) in multivariable models [33]. We found that there existed inversely association between BPD and the 7 th day given energy by univariate analysing(OR=0.515, 95%CI, 0.275-0.962), but there existed no association between both by multivariate analysing. We will continue to collect clinical data to expand the population in order to explicit the relationship between both.
The internal enviroment within 1 st week of life such as blood gas values ( PH PCO2 PO2 HCO3) in preterm infants may be very important for the development of premature organs and systems. For infants of less than 32 GWs, the infants are during saccular and alveolar stages of lung. If the critical lung development is interrupted by low PH, low PO2, high PCO2, and low HCO3, there would cause ineffective gas exchange in the preterm lung and the baby would need respiratory support and could suffer from BPD. This study showed that BPD was associated with higher PCO2, lower PO2, and lower HCO3 in 7 th day of life by univariate analyzing, and associated with lower PO2 in 7 th day of life through multivariate recession analyzing, namely low PO2 in 7 th day of life was the independently risk factor of BPD.It suggested that the gas exchang membrane could not recover untill 1week of life and that continued low PaO2 till 1 week after birth furth hurt the premature lung and might cause BPD. The speci c mechanism of low PaO2 in 7 th day of life on BPD need to furth study. As we known, it is rst reported that there existed association between blood gas values in the rst 7 days of life and BPD.
From the above, low PaO2 in 7 th day of life, besides MV, GA/BW, was the independent risk factor of moderate and severe BPD in preterm infants of less than 32GWs. Namely, preterm infants of 32GWs with blood gas derangements within 7days of life could be at risk of moderate and severe BPD.

Declarations
Ethics approval: The study was approved by the ethics review committee of the rst a liated hospital of Nanjing Medical University (2014SR038)Consent for publication: The manuscript is approved by all authors for publicationAvailability of data and material: The data used or analysed during the study are included in this published article. More detailed data are available from the corresponding author if need.Code availability: N/A  Relationship between BPD and daily given liquid by univariate analysing Figure 3 Relationship between moderate and severe BPD and daily given energy by univariate analysing