Short-term outcomes of extremely preterm infants at discharge: a multicenter study from Guangdong province during 2008–2017

The collaborative study group composed of 26 NICUs was established before data collection. These NICUs were located in four regions of Guangdong province which were representative of medical units offering neonatal intensive care in the irrespective areas. The Third Affiliated Hospital of Guangzhou Medical University was responsible for coordinating this survey, where all the data were aggregated, stored and analyzed. This study was approved by the Ethics Committees of the Third Affiliated Hospital of Guangzhou Medical University. The same diagnostic criteria were applied to all NICUs.

All EP infants discharged from our collaborative NICUs were studied, whereas live-born infants not admitted to NICUs and still births in the delivery room were not enrolled. The study protocol was fully discussed by all members, and a standardized questionnaire for data collection including maternal and neonatal demographics, major complications, treatments and outcomes was designed. In fact, this study was initiated at the end of 2012 and is still ongoing at present. Therefore, the data from January 1, 2008 to December 31, 2012 were collected retrospectively, and data from January 1, 2013 to December 31, 2017 were collected prospectively. The relevant records of all enrolled infants and their mothers were reviewed and filled in the questionnaire. All sheets were sent to the Third Affiliated Hospital of Guangzhou Medical University and the data from each questionnaire were input into the database. In order to minimize bias among centers and investigators, comprehensive and systematic training was provided to the staffs involved in the survey. Data collected by the researcher at each collaborative NICU was supervised and checked by the NICU director, who was responsible for the quality assurance. Meanwhile, these records were also checked for accuracy and completeness by collaborative centers.

In this study, survivors were defined as neonates who survived to the time of discharge. GA was calculated from the date of the last menstrual period or was determined by fetal ultrasound assessment. RDS was diagnosed in preterm infants with the onset of respiratory distress shortly after birth and a compatible chest radiograph appearance [11]. BPD was defined as oxygen dependency at 36 weeks of corrected age or at discharge [12]. The criteria utilized in our survey for the diagnosis of NEC and for grading the severity of disease were based on Bell’s stage [13]. ROP and the graded standard were defined by the international classification of ROP [14]. IVH and periventricular leukomalacia (PVL) were diagnosed by cranial ultrasonography or magnetic resonance imaging (MRI). The Papile grading system was used to grade IVH [15], and PVL was defined as degeneration of white matter adjacent to the cerebral ventricles following cerebral hypoxia or brain ischemia [16]. Nosocomial sepsis was defined as blood culture-positive sepsis occurring beyond 48 h of life [17].

During the 10-year survey, in total, 2051 EP infants were enrolled. The overall survival rate at discharge was 52.5% (1076 of 2051). The smallest GA and lowest BW in survivors were 22 weeks and 480 g respectively. Totally, the mean GA was 26.68 ± 1.00 weeks, and the distribution ranged from 26 (1.3%) for less than 24 weeks, 105 (5.1%) for 24 weeks, 263 (12.8%) for 25 weeks, 548 (26.7%) for 26 weeks to 1109 (54.1%) for 27 weeks. The mean BW was 935 ± 179 g, and the distribution ranged from 16 (0.8%) for less than 500 g, 36 (1.8%) for 500 - 599 g, 106 (5.2%) for 600 - 699 g, 255 (12.4%) for 700 - 799 g, 398 (19.4%) for 800 - 899 g, 495 (24.1%) for 900 - 999 g to 745 (36.3%) for equal or more than 1000 g. In order to clarify the current treatment and outcome of EP infants, we specifically grouped the EP infants based on survival or not, as it was presented in Table 1. The overall mean GA or BW in the survivor group was greater than in the non-survivor group (p < 0.001). There was no significant difference in the sex ratio between the survivor and non-survivor group. There were fewer infants with Apgar scores ≤ 3 at 1 min and ≤ 7 at 5 min in the survivor group (both p < 0.001). Comparing with the non-survivor group, the survivor group had higher rates of receiving surfactant therapy (p < 0.001) and non-invasive ventilation (p < 0.001), but the rate of requiring mechanical ventilation was lower (p < 0.01). Even more, there were significant differences in the days of mechanical ventilation, non-invasive ventilation, oxygen therapy and hospital stay between the two groups (p < 0.001).

For the mothers, there was a similar incidence in the history of pregnancy problems between the survivor and non-survivor groups. Comparing with the mothers in the non-survivor group, the mothers in the survivor group had a higher proportion of age ≥ 35 years (p < 0.05) and antenatal steroids therapy (p < 0.001), while a lower rate of twin/multiple pregnancies (p < 0.05). Interestingly, the mothers in the survivor group had a higher incidence of premature rupture of membranes (p < 0.001) and a lower incidence of cervical incompetence (p < 0.01). However, there was no significant difference in the rate of cesarean section, or in the incidences of infection in the middle trimester of pregnancy, gestational diabetes mellitus, pregnancy-induced hypertension syndrome, placental abruption/placenta previa or thyroid dysfunction.

The number of EP infants discharged from NICUs increased sharply from 58 cases in 2008 to 408 cases in 2017, as it was shown in Table 2. The survival rate improved annually from 36.2% in 2008 to 59.3% in 2017 (p < 0.001). The ratio of EP infants to all preterm infants rose annually from 0.70% in 2008 to 2.29% in 2017 (p < 0.001). Similarly, the ratio of EP infants to all discharged infants increased annually from 0.17% in 2008 to 0.67% in 2017 (p < 0.001).

For the entire cohort, with the increase in GA or BW, the number of EP infants increased from 26 (GA < 24 weeks) or 52 (BW < 600 g) to 1109 (GA = 27 weeks) or 745 (BW ≥ 1000 g), as it was displayed in Table 3. As GA increased from less than 24 weeks to 27 weeks, the survival rate improved from 11.5 to 62.4% (p < 0.001). As BW increased from less than 600 g to above 1000 g, the survival rate rose dramatically from 23.1 to 64.2% (p < 0.001). There was only one infant who survived with BW less than 500 g.However, there were no significant differences in GA or BW between 2008–2012 and 2013–2017 (p = 0.073 or p = 0.500). The mean GA or BW for EP infants during 2008–2012 were 26.73 ± 0.93 weeks or 935 ± 172 g respectively, while during 2013–2017 were 26.66 ± 1.02 weeks or 935 ± 181 g respectively.

According to the level of economic development where the hospitals were located, our collaborative NICUs could be divided into three categories, specifically involving Guangzhou and Shenzhen (including 11 hospitals) with high level, the cities in the Pearl Delta (including 10 hospitals) with medium level and the cities outside the Pearl Delta (including 5 hospitals) with low level. As it was shown in Table 4, from 2008 to 2017, the overall survival rates were positively correlated with the level of regional economic development, and there were significant differences among these regions (p < 0.001). From the low to high level, the mean GA of EP infants was 26.74 ± 0.90 weeks, 26.79 ± 0.92 weeks and 26.62 ± 1.04 weeks, which showed a significant difference (p = 0.002); while the mean BW was 920 ± 174 g, 947 ± 170 g and 932 ± 183 g, which showed no significant difference (p = 0.104).

Among the 26 hospitals, 19 were general hospitals and the others were specialist hospitals (maternal and children’s hospitals), and the overall survival rate of EP infants discharged from general hospitals was lower than specialist hospitals (OR = 0.530, 95% CI: 0.445–0.632, p < 0.001). Nevertheless, the mean GA and BW of EP infants from general hospitals (26.75 ± 0.94 weeks and 944 ± 177 g respectively) were both greater than those from specialist hospitals (26.60 ± 1.05 weeks and 924 ± 180 g respectively), and there were significant differences (p = 0.001 and p = 0.013).

RDS was the most common complication in EP infants, who were diagnosed with RDS accounted for 88.0% (1804 of 2051). Therefore, the proportion of EP infants receiving surfactant therapy and mechanical ventilation was 78.9% (1618 of 2051) and 78.2% (1603 of 2051) respectively, as shown in Table 1.There was no significant difference in RDS incidence among GA groups as shown in Fig. 1. BPD was another serious complication in EP infants. In this survey, 68.0% (788 of 1158) of EP infants were diagnosed with oxygen dependency on postnatal day 28, but only 32.3% (374 of 1158) still required oxygen therapy at 36 weeks of corrected age or at discharge. The incidence of BPD decreased significantly with the increase of GA (p <0.001). 1117 EP infants underwent eye examination, 504 (45.1%) of whom were diagnosed with ROP at different stages, and the proportion of stage ≥3 was 14.6% (163 of 1117). 88 EP infants received further treatments, 40 of whom underwent intravitreal injection with anti-VEGF. The incidence of ROP declined with the increase of GA (p <0.001). NEC was evaluated in EP infants who survived for more than 48 h. 10.1% (160 of 1588) of EP infants met the diagnostic criteria at any stage of NEC, 27 of whom required surgery. The incidence of NEC appeared to be positive with GA (p <0.05). 1431 EP infants were examined by head ultrasound or MRI, of whom 37.4% (535 of 1431) were diagnosed with IVH, 13.3% (190 of 1431) of EP infants were diagnosed with severe IVH at grade ≥ 3. The incidence of IVH was inversely related to gestational age (p <0.001). The incidence of PVL was 6.2% (89 of 1431), and there was no significant difference among GA < 25 weeks, GA = 25 weeks, GA = 26 weeks and GA = 27 weeks. Among the 1588 infants survived more than 48 h, 250 infants (15.7%) were diagnosed with blood culture-positive nosocomial sepsis. The complications for each subgroup of GA are shown in Fig. 1. Additionally, comparing with the period of 2008–2012, the incidence of RDS and BPD increased sharply during the period of 2013–2017(OR = 2.076, 95% CI: 1.574–2.738, p < 0.001; OR = 1.645, 95% CI: 1.152–2.351, p < 0.01). On the contrary, the incidence of PVL decreased dramatically (OR = 0.252, 95% CI: 0.163–0.392, p < 0.001). Moreover, the incidence of ROP, NEC and IVH did not change significantly. The results were shown in Fig. 2.

Among the 975 non-survivors, 893 (91.6%) of EP infants died during the neonatal period (≤ 28 days), the other 82 (8.4%) died after the neonatal period (> 28 days) as it was shown in Fig. 3. Specifically, 324 (33.2%) of EP infants died on the first day, 139 (14.3%) on the second day, 88 (9.0%) on the third day, 129 (13.2%) during the fourth to seven day, 97 (9.9%) in the second week, 71 (7.3%) in the third week and 45 (4.6%) in the fourth week. The survival days of non-survivors under active treatment or medical care withdrawal were shown in Fig. 3. The chi-square test showed that there was a significant difference in the distribution of survival days between the two groups (p < 0.001). No significant difference in the distribution of survival days of non-survivors was found among the different GA (GA < 25 weeks, GA = 25 weeks, GA = 26 weeks or GA = 27 weeks) by chi-square test.

In our survey, 654 EP infants died mainly as a result of medical care withdrawal. The main reasons for the decision to withdraw care were summarized, as they were shown in Fig. 4. The most important reason was the economic burden and fearing of poor or uncertain outcomes, followed by only fearing of poor or uncertain outcomes, only concerning about economic burden, unknown reason and other factors, which was accounted for 34.4, 33.0, 19.0, 10.7 and 2.7% respectively.

As it was shown in Table 5, the multivariate logistic regression analysis indicated that improved survival of EP infants was associated with discharged from specialist hospital, hospitals located in high-level economic development region, increasing gestational age, increasing birth weight, antenatal steroids use and a history of premature rupture of membranes. On the contrary, twins or multiple births, Apgar ≤7 at 5 min, cervical incompetence and decision to withdraw care were associated with decreased survival.