Background
Very preterm infants are at risk for neurodevelopmental impairment because of postnatal morbidities. This study aims to (1) compare the outcomes of very-low-birth-weight (VLBW) infants in Singapore during two time periods over a decade; 2) compare performances among Singaporean neonatal intensive care units (NICUs); and 3) compare a Singapore national cohort with one from the Australian and New Zealand Neonatal Network (ANZNN).
Methods
Singapore national data on VLBW infants born during two periods, 2007–2008 (SG2007, n = 286) and 2015–2017 (SG2017, n = 905) were extracted from patient medical records. The care practices and clinical outcomes among three Singapore NICUs were compared using SG2017 data. Third, using data from the ANZNN2017 annual report, infants with gestational age (GA) ≤29 weeks in SG2017 were compared with their Oceania counterparts.
Results
SG2017 had 9.9% higher usage of antenatal steroids (p < 0.001), 8% better survival for infants ≤26 weeks (p = 0.174), and used 12.7% lesser nonsteroidal anti-inflammatory drugs for patent ductus arteriosus closure (p < 0.001) than those of SG2007 cohort. Rate of late-onset sepsis (LOS) was almost halved (7.4% vs. 14.0%, p < 0.001), and exclusive human milk feeding after discharge increased threefold (p < 0.001). SG2017, in contrast, had a higher rate of chronic lung disease (CLD) (20.0% vs. 15.1%, p = 0.098). Within SG2017, the rates of LOS, CLD, and human milk feeding varied significantly between the three NICUs. When compared with ANZNN2017, SG2017 had significantly lower rates of LOS for infants ≤25 weeks (p = 0.001), less necrotizing enterocolitis for infants ≤27 weeks (p = 0.002), and less CLD across all GA groups.
Conclusion
Postnatal morbidities and survival rates for VLBW infants in Singapore have improved over a decade. Outcomes for VLBW infants varied among three Singapore NICUs, which provide a rationale for collaboration to improve clinical quality. The outcomes of Singaporean VLBW infants were comparable to those of their ANZNN counterparts.
Key Words
1. Introduction
In Singapore, premature neonates contribute up to 8.4% of the annual birth cohort.
1
Very low birth weight (VLBW) infants are those with birth weights less than 1500 g. VLBW infants consume disproportionately large amounts of healthcare resources. They stay for about two months in the hospital and require long-term follow-up.2
Sequelae, such as cerebral palsy and cognitive delay, seen in VLBW infants are significantly influenced by their postnatal neonatal intensive care unit (NICU) complications.3
Efforts in clinical quality improvement have been ongoing for decades, and have resulted in the establishment of many network databases.2
,4
,5
These are useful tools for benchmarking and improving clinical performance.Singapore's VLBW network was established in 2015. It aims to compile national data on demographics and clinical outcomes of VLBW infants and measure them against those of other networks, such as the Australian and New Zealand Neonatal Network (ANZNN) and the Vermont–Oxford Network (VON).
4
,6
Three public sector level III NICUs. which manage approximately 80% of Singaporean VLBW infants, contributed data to the network. Private sector level II NICUs that look after 20% of VLBW infants were excluded.We report a three-year (2015–2017) national VLBW infant cohort data (cohort SG2017). The comparator, cohort SG2007, is a similar one-year (2007–2008) dataset from a national audit conducted by the Singapore Ministry of Health (MOH). We aimed to review the changes in VLBW infant mortality and morbidity in Singapore over approximately a decade. For SG2017, we analyzed the differences in outcomes among the three public NICUs. In addition, we compared local outcomes against those of international benchmarks ANZNN2017 and VON2014 (latest available published data).
6
,7
2. Methods
A steering committee comprising the directors of participating NICUs (A, B, and C) oversaw the running of the Network. The inclusion criteria were all infants admitted to the NICU weighing <1500 g at birth, regardless of gestational age (GA). Infants who died in delivery rooms were excluded.
Data for SG2017 were extracted from patients’ medical records using standardized definitions and entered by personnel from each participating NICU into an electronic database hosted by the Singapore Clinical Research Institute (SCRI). This was a retrospective, multicenter cohort study. Data collected included VLBW infants born between January 2015 and December 2017.
The extracted variables included birth weight (BW), GA, prenatal care, antenatal steroid use, ventilator requirements, patent ductus arteriosus (PDA) treatment, early- and late-onset sepsis, mortality during initial hospitalization, and length of stay during the first hospital episode. Clinical conditions included chronic lung disease (CLD), intraventricular hemorrhage (IVH), retinopathy of prematurity (ROP), and necrotizing enterocolitis (NEC). Relevant perinatal variables included antenatal steroid use, pregnancy complications, and delivery mode. Definitions of clinical conditions followed those of VON unless otherwise specified.
SG2007 data were obtained from the MOH Clinical Quality Audit conducted from June 2007 to May 2008. Similar to those of SG2017, SG2007 infants who died in the delivery room and level II private sector NICU infants were excluded.
This study was approved by the National Healthcare Group Domain-Specific Review Board. Waiver of informed consent was granted because the study involved collection of de-identified health information.
2.1 Statistical analysis
Continuous variables are presented as mean (±SD) and categorical variables as frequency (percentage). Student's t-test and the chi-square test were used for between-group comparisons of continuous and categorical variables, respectively. Univariate and multivariate logistic regression models were used to compare the odds of infant survival and other major neonatal morbidities among the three NICUs. The multivariate regression model was adjusted for outborn status, mode of delivery, and multiple births separately for the ≤27 weeks and 28–31 weeks GA groups. Selected outcomes for infants <30 weeks were compared with those of the same GA in ANZNN2017, provided that data were available. Statistical analyses were performed using Stata/MP version 17 (College Station, Texas, USA). Statistical significance was set at a two-tailed P-value of <0.05.
3. Results
3.1 Comparison of SG2007 and SG2017 cohorts
SG2007 cohort had 286 infants, compared to 905 infants in SG2017. The baseline characteristics are presented in Table S1. The average GA and BW were similar between SG2007 and SG2017 (GA 28.6 weeks vs. 28.6 weeks; BW 1090 g vs. 1076 g, respectively). The proportion of infants with BW ≤ 750 g or those born at GA ≤26 weeks in the 2 cohorts were also similar. Antenatal steroid use increased significantly in SG2017 (89.3 vs. 79.4%, p < 0.001).
The clinical outcomes are shown in Table 1. The survival rates were high in both cohorts (88.1% and 89.9% in SG2007 and SG2017, respectively). Infants aged ≤26 weeks in SG2017 had an 8% better survival rate (74.7% vs. 66.7%, p = 0.174). More infants in SG2017 were diagnosed with respiratory distress syndrome (RDS) (74.0% vs. 62.9%, p < 0.001) and required surfactant therapy (52.0% vs. 38.8%, p < 0.001). Proportionately more infants in SG2017 developed CLD, defined as a need for oxygen supplementation at 36 weeks’ postmenstrual age (PMA) (20.0% vs. 15.1%, p = 0.098). This definition of CLD was used for the comparison between SG2007 and SG2017 because data on the need for assisted ventilatory support at 36 weeks were not available for SG2007. Postnatal steroid use declined in SG2017 (11.2% vs. 15.4%, p = 0.054). The use of nonsteroidal anti-inflammatory drugs (NSAIDs) for PDA declined in SG2017 (30.1% vs. 42.8%, p < 0.001). The frequency of late-onset sepsis (LOS) was significantly lower in SG2017 (7.4% vs. 14.0%, p < 0.001). This reduction was consistent across all pathogens. At discharge, the proportion of infants exclusively fed human milk increased threefold in SG2017 (27.9% vs. 8.7%, p < 0.001).
Table 1Comparison of clinical outcomes - SG2007 vs. SG2017.
| Parameters | Cohort SG2007 (n = 286) | Cohort SG2017 (n = 905) | p-value |
|---|---|---|---|
| Survival (total) | 88.1% (252/286) | 89.9% (814/905) | 0.387 |
| Survival by GA groups | |||
| ≤26 weeks | 66.7% (50/75) | 74.7% (177/237) | 0.174 |
| 27–29 weeks | 91.8% (90/98) | 93.3% (305/327) | 0.627 |
| ≥30 weeks | 99.1% (112/113) | 97.4% (332/341) | 0.271 |
| Mean Length of stay, days | 69.0 | 69.2 | 0.918 |
| Respiratory outcomes | |||
| Respiratory Distress Syndrome | 62.9% (180/286) | 74.0% (670/905) | <0.001 |
| Surfactant | 38.8% (111/286) | 52.0% (471/905) | <0.001 |
| Pneumothorax | 7.0% (20/286) | 4.0% (36/905) | 0.037 |
| Nasal CPAP | 83.5% (238/285) | 86.6% (784/905) | 0.190 |
| Conventional ventilation | 51.9% (148/285) | 59.9% (542/905) | 0.017 |
| HFV | 13.7% (39/286) | 16.2% (147/905) | 0.310 |
| Inhaled nitric oxide | 0.4% (1/285) | 5.5% (50/905) | <0.001 |
| O2 Day 28 | 40.1% (105/262) | 31.2% (256/820) | 0.008 |
| O2 36 weeks | 15.1% (34/225) | 20.0% (158/791) | 0.098 |
| CPAP at 36 weeks | 12.0% (27/225) | 14.7% (116/790) | 0.305 |
| O2 or CPAP at 36 weeks | no data | 23.6% (187/791) | / |
| Postnatal steroids | 15.4% (44/286) | 11.2% (101/905) | 0.054 |
| Discharge on oxygen | 10.6% (30/284) | 8.9% (71/800) | 0.398 |
| Indomethacin/Ibuprofen | 42.8% (122/285) | 30.1% (272/905) | <0.001 |
| PDA ligation | 10.2% (29/285) | 7.2% (65/905) | 0.103 |
| Severe IVH | 5.4% (15/286) | 4.4% (39/905) | 0.483 |
| PVL | 0.7% (2/283) | 1.3% (12/905) | 0.408 |
| NEC Stage 2,3 | 2.8% (8/286) | 2.2% (20/905) | 0.559 |
| Isolated GI perforation | 0.7% (2/286) | 2.5% (23/905) | 0.062 |
| Enteral feeding on discharge | <0.001 | ||
| Human milk only | 8.7% (22/252) | 27.9% (223/800) | |
| Formula only | 13.1% (33/252) | 12.0% (96/800) | |
| Human milk and formula | 78.2% (197/252) | 60.1% (481/800) | |
| Early bacterial sepsis | 3.1% (9/286) | 1.7% (15/905) | 0.144 |
| Late onset sepsis | 14.0% (40/286) | 7.4% (67/905) | <0.001 |
| Gram negatives | 8.4% (24/286) | 4.6% (42/905) | 0.014 |
| CONS | 4.9% (14/286) | 1.4% (13/905) | <0.001 |
| Other gram positives | no data | 1.3% (12/905) | / |
| Fungal | 7.1% (19/286) | 0.1% (1/905) | <0.001 |
| ROP examination | 85.7% (245/286) | 77.2% (699/905) | 0.002 |
| Severe ROP | 8.6% (21/244) | 8.9% (62/699) | 0.876 |
| ROP surgery | 4.9% (12/245) | 4.3% (30/699) | 0.503 |
| Presence of birth defect | 4.6% (13/286) | 3.8% (34/905) | 0.108 |
p-value from Chi-square test (comparison of proportions) and Student's T-test (comparison of means).
a Types of organism do not add up to 14.0% due to >1 episode of sepsis per hospital stay in some subjects.
3.2 Comparison of three NICUs in SG2017
Table 2 compares the data from the three NICUs. NICU B had a higher rate of vaginal deliveries among infants with GA of 28–31 weeks (p = 0.039). NICU C had a higher proportion of outborn infants with GA ≤ 27 weeks (p = 0.020). More infants in NICU C were from multiple births (p = 0.023).
Table 2Comparing 3 Singapore NICUs – Infant characteristics and neonatal outcomes in SG2017.
| Parameters | Gestational weeks | |||||||
|---|---|---|---|---|---|---|---|---|
| ≤27 weeks | 28–31 weeks | |||||||
| NICU A (n = 227) | NICU B (n = 46) | NICU C (n = 61) | P value | NICU A (n = 285) | NICU B (n = 73) | NICU C (n = 59) | P value | |
| CLINICAL INDICATORS OF THE MOTHER | ||||||||
| Maternal race | ||||||||
| Chinese | 131 (57.71%) | 22 (47.83%) | 33 (54.10%) | 0.452 | 147 (51.58%) | 27 (36.99%) | 32 (54.24%) | 0.245 |
| Malay | 54 (23.79%) | 13 (28.26%) | 10 (16.39%) | 78 (27.37%) | 24 (32.88%) | 17 (28.81%) | ||
| Indian | 16 (7.05%) | 5 (10.87%) | 8 (13.11%) | 28 (9.82%) | 13 (17.81%) | 6 (10.17%) | ||
| Others | 26 (11.45%) | 6 (13.04%) | 10 (16.39%) | 32 (11.23%) | 9 (12.33%) | 4 (6.78%) | ||
| Maternal Hypertension | 31 (13.66%) | 8 (17.39%) | 5 (8.20%) | 0.353 | 92 (32.28%) | 18 (24.66%) | 12 (20.34%) | 0.118 |
| Parental Care | 223 (98.24%) | 44 (95.65%) | 58 (95.08%) | 0.304 | 278 (97.54%) | 71 (97.26%) | 59 (100%) | 0.463 |
| Assisted Reproduction | 37 (16.30%) | 10 (21.74%) | 18 (29.51%) | 0.063 | 27 (9.47%) | 12 (16.44%) | 19 (32.20%) | <0.001 |
| Vaginal delivery | 94 (41.41%) | 23 (50.00%) | 27 (44.26%) | 0.551 | 67 (23.51%) | 27 (36.99%) | 12 (20.34%) | 0.039 |
| Location of birth | ||||||||
| Inborn Outborn/BBA | 220 (96.92%) 7 (3.08%) | 42 (91.30%) 4 (8.70%) | 54 (88.52%) 7 (11.48%) | 0.020 | 277 (97.19%) 8 (2.81%) | 69 (94.52%) 4 (5.48%) | 54 (91.53%) 5 (8.47%) | 0.108 |
| Antenatal Steroids | 207 (91.19%) | 37 (80.43%) | 55 (90.16%) | 0.093 | 251 (88.07%) | 64 (87.67%) | 56 (94.92%) | 0.289 |
| Chorioamnionitis | 130 (57.27%) | 11 (23.91%) | 20 (32.79%) | <0.001 | 60 (21.05%) | 6 (8.22%) | 2 (3.39%) | <0.001 |
| CLINICAL INDICATORS OF THE INFANT | ||||||||
| Mean Gestational Age, weeks (SD) | 25.52 (1.22) | 25.83 (1.20) | 25.26 (1.30) | 0.065 | 29.42 (1.07) | 29.23 (1.06) | 29.41 (1.07) | 0.388 |
| Mean Birth Weight, grams (SD) | 809 (194) | 845 (216) | 796 (175) | 0.404 | 1191 (219) | 1202 (195) | 1189 (195) | 0.918 |
| Multiple Gestation | 51 (22.5%) | 8 (17.4%) | 23 (37.7%) | 0.023 | 66 (23.2%) | 25 (34.3%) | 21 (35.6%) | 0.043 |
| Male infant | 131 (57.71%) | 21 (45.65%) | 32 (52.46%) | 0.293 | 147 (51.58%) | 38 (52.05%) | 28 (47.46%) | 0.833 |
| Mean APGAR score 1min (SD) | 4.35 (2.13) | 4.89 (2.65) | 4.79 (2.03) | 0.170 | 6.10 (2.15) | 6.79 (2.41) | 5.93 (1.90) | 0.032 |
| Mean APGAR score 5min (SD) | 6.89 (1.76) | 7.37 (2.08) | 7.39 (1.63) | 0.066 | 8.16 (1.23) | 8.41 (1.45) | 8.20 (1.23) | 0.335 |
| Mean Temperature within 1st hour after admission to NICU (SD) | 36.42 (0.68) | 36.22 (1.11) | 35.55 (0.96) | <0.001 | 36.47 (0.63) | 36.56 (0.65) | 35.77 (0.74) | <0.001 |
| Mean First Base Excess, w/in 1 h admission (SD) | −5.52 (4.43) | −7.76 (5.11) | −4.71 (4.52) | 0.002 | −3.90 (3.78) | −5.60 (4.59) | −3.95 (3.01) | 0.004 |
| CRIB II score | ||||||||
| Level 1: 0–5 | 0 (0%) 86 (39.81%) 115 (53.24%) 15 (6.94%) | 0 (0%) 17 (36.96%) 28 (60.87%) 1 (2.17%) | 0 (0%) 14 (26.92%) 33 (63.46%) 5 (9.62%) | 0.291 | 123 (46.59%) 136 (51.52%) 5 (1.89%) 0 (0%) | 28 (38.89%) 41 (56.94%) 3 (4.17%) 0 (0%) | 23 (43.40%) 29 (54.72%) 1 (1.89%) 0 (0%) | 0.656 |
| Level 2: 6–10 | ||||||||
| Level 3: 11–15 | ||||||||
| Level 4: >15 | ||||||||
| Early bacterial sepsis, ≤3 days | 9 (3.96%) | 0 (0%) | 2 (3.28%) | 0.389 | 3 (1.05%) | 1 (1.37%) | 0 (0%) | 0.695 |
| Late bacterial sepsis, >3 days 1 | 22 (9.69%) | 12 (26.09%) | 11 (18.03%) | 0.006 | 11 (3.86%) | 5 (6.85%) | 1 (1.69%) | 0.313 |
| IVH Grade 3–42 | 19 (8.48%) | 5 (11.11%) | 7 (12.07%) | 0.653 | 6 (2.11%) | 0 (0%) | 1 (1.85%) | 0.464 |
| Cystic periventricular leukomalacia 3 | 2 (0.90%) | 0 (0%) | 2 (3.51%) | 0.202 | 1 (0.35%) | 2 (2.82%) | 2 (3.51%) | 0.056 |
| NEC Stage 2–34 | 9 (3.96%) | 2 (4.35%) | 2 (3.28%) | 0.956 | 4 (1.40%) | 0 (0%) | 2 (3.39%) | 0.266 |
| Any NEC Surgery done | 6 (2.64%) | 0 (0%) | 3 (4.92%) | 0.297 | 1 (0.35%) | 0 (0%) | 2 (3.39%) | 0.031 |
| Other surgery done | 61 (26.87%) | 7 (15.22%) | 6 (9.84%) | 0.008 | 60 (21.05%) | 8 (10.96%) | 0 (0%) | <0.001 |
| ROP examination | 191 (84.14%) | 41 (89.13%) | 47 (77.05%) | 0.226 | 242 (84.91%) | 70 (95.89%) | 43 (72.88%) | 0.001 |
| ROP Surgery done | 15 (6.61%) | 5 (10.87%) | 6 (9.84%) | 0.495 | 1 (0.35%) | 2 (2.74%) | 1 (1.69%) | 0.144 |
| ROP Stage 3 or above 5 | 38 (19.90%) | 7 (17.07%) | 13 (27.66%) | 0.410 | 3 (1.24%) | 1 (1.43%) | 0 (0%) | 0.750 |
| ROP Stage 3 or death | 82 (36.12%) | 14 (30.43%) | 24 (39.34%) | 0.633 | 15 (5.26%) | 4 (5.48%) | 3 (5.08%) | 0.995 |
| Oxygen at Day 28 | 127 (64.8%) | 22 (52.4%) | 45 (88.2%) | 0.001 | 33 (12.00%) | 2 (2.86%) | 22 (40.00%) | <0.001 |
| Ventilatory support | ||||||||
| Nasal CPAP | 203 (89.43%) | 40 (86.96%) | 48 (78.69%) | 0.084 | 271 (95.09%) | 70 (95.89%) | 56 (94.92%) | 0.954 |
| ETT ventilation | 212 (93.39%) | 41 (89.13%) | 58 (95.08%) | 0.464 | 145 (50.88%) | 38 (52.05%) | 28 (47.46%) | 0.859 |
| HFOV | 85 (37.44%) | 14 (30.43%) | 15 (24.59%) | 0.145 | 23 (8.07%) | 3 (4.11%) | 6 (10.17%) | 0.389 |
| Surfactant | 203 (89.43%) | 40 (86.96%) | 39 (63.93%) | <0.001 | 127 (44.56%) | 33 (45.21%) | 17 (28.81%) | 0.073 |
| Inhaled Nitric Oxide | 28 (12.33%) | 1 (2.17%) | 7 (11.48%) | 0.126 | 8 (2.81%) | 0 (0%) | 4 (6.78%) | 0.068 |
| PICC | 202 (88.99%) | 43 (93.48%) | 58 (95.08%) | 0.272 | 230 (80.70%) | 70 (95.89%) | 53 (89.83%) | 0.003 |
| Oxygen at 36 weeks | 77 (43.26%) | 13 (32.50%) | 36 (70.59%) | <0.001 | 16 (6.3%) | 1 (1.4%) | 13 (23.2%) | <0.001 |
| CPAP at 36 weeks | 63 (35.39%) | 10 (25.64%) | 17 (33.33%) | 0.505 | 11 (4.35%) | 9 (12.86%) | 3 (5.36%) | 0.030 |
| Oxygen or CPAP at 36 weeks 6 | 89 (50.00%) | 16 (40.00%) | 36 (70.59%) | 0.008 | 20 (7.91%) | 9 (12.86%) | 13 (23.21%) | 0.004 |
| Postnatal steroids | 62 (27.31%) | 12 (26.09%) | 17 (27.87%) | 0.978 | 6 (2.11%) | 1 (1.37%) | 2 (3.39%) | 0.725 |
| Indomethacin/Ibuprofn for PDA | 130 (57.27%) | 29 (63.04%) | 35 (57.38%) | 0.764 | 44 (15.44%) | 15 (20.55%) | 18 (30.51%) | 0.022 |
| PDA Ligation | 29 (12.78%) | 14 (30.43%) | 15 (24.59%) | 0.004 | 4 (1.40%) | 2 (2.74%) | 0 (0%) | 0.420 |
| Respiratory Distress Syndrome (RDS) | 223 (98.24%) | 41 (89.13%) | 54 (88.52%) | 0.001 | 237 (83.16%) | 41 (56.16%) | 43 (72.88%) | <0.001 |
| Pneumothorax | 12 (5.29%) | 4 (8.70%) | 5 (8.20%) | 0.544 | 10 (3.51%) | 2 (2.74%) | 2 (3.39%) | 0.948 |
| Isolated gastrointestinal perforation 7 | 15 (6.61%) | 1 (2.17%) | 4 (6.56%) | 0.502 | 1 (0.4%) | 1 (1.4%) | 0 (0%) | 0.450 |
| Presence of any of the 7 neonatal morbidities 1 , 2 , 3 , 4 , 5 , 6 , 7 | 115 (50.66%) | 25 (54.35%) | 47 (77.05%) | 0.001 | 34 (11.93%) | 10 (13.7%) | 16 (27.12%) | 0.010 |
| DISCHARGE OUTCOME | ||||||||
| Discharge from hospital (Survival) | 173 (76.21%) | 38 (82.61%) | 48 (78.69%) | 0.620 | 270 (94.74%) | 69 (94.52%) | 55 (93.22%) | 0.898 |
| Enteral feeding at discharge | ||||||||
| Human milk only | 43 (24.86%) | 29 (76.32%) | 4 (8.33%) | <0.001 | 59 (21.85%) | 42 (60.87%) | 1 (1.82%) | <0.001 |
| Formula only | 29 (16.76%) | 4 (10.53%) | 0 (0%) | 36 (13.33%) | 9 (13.04%) | 0 (0%) | ||
| Human milk + Formula | 101 (58.38%) | 5 (13.16%) | 44 (91.67%) | 175 (64.8%) | 18 (26.1%) | 54 (98.2%) | ||
| Oxygen at discharge | 51 (29.48%) | 5 (13.16%) | 6 (12.50%) | 0.012 | 7 (2.59%) | 0 (0%) | 1 (1.82%) | 0.392 |
| Mean Length of Stay, days (SD) | 116.1 (48.8) | 110.5 (46.8) | 105.6 (28.7) | 0.339 | 57.9 (25.8) | 59.3 (21.4) | 60.0 (23.3) | 0.811 |
| Mean Weight at discharge, kg (SD) | 3.54 (1.24) | 2.73 (0.74) | 3.16 (0.81) | <0.001 | 2.50 (0.67) | 2.17 (0.41) | 2.55 (0.68) | <0.001 |
BBA, birth before arrival; CPAP, continuous positive airway pressure; ETT, endotracheal tube; HFOV, high frequency oscillatory ventilation; IVH, intraventricular haemorrhage; NEC, necrotising enterocolitis; PDA, patent ductus arteriosus; PICC, peripherally inserted central catheter; RDS, respiratory distress syndrome; ROP, retinopathy of prematurity; SD, standard deviation.
p-value from Chi-square test (comparison of proportions) and Student's T-test (comparison of means).
a A diagnosis of chorioamnionitis was recorded in the maternal or infant medical record.
There were significant differences in the clinical outcomes. Infants aged ≤27 weeks in NICU B had significantly higher rates of LOS (p = 0.006), and more infants required PDA ligation (p = 0.004). In both GA groups, NICU A had more infants with RDS (p = 0.001), accompanied by higher usage of surfactant for infants aged ≤27 weeks (p < 0.001). NICU C had the lowest use of surfactant (p < 0.001) relative to the diagnosis of RDS. Despite higher rates of RDS in infants of NICU A and NICU B, respiratory outcomes consistent with the diagnosis of CLD were significantly greater in NICU C infants. More infants aged ≤27 weeks in NICU A were discharged with supplemental oxygen (p = 0.012). There was no surgical NEC among infants with GA 28–31 weeks in NICU B (p = 0.031). The mean lengths of stay among the NICUs were similar; however, the mean weights at discharge differed, with infants of NICU B having a lower mean weight at discharge (p < 0.001). Upon discharge, more infants from NICU B were exclusively fed human milk (p < 0.001).
Table 3 shows adjusted ORs (aOR) for survival and major morbidities. The odds of CLD in infants of NICU C were greater regardless of gestation, aOR 3.92 [95% CI 1.79–8.62], p = 0.001 for infants aged 28-31 weeks, and aOR 2.57 [95% CI 1.26-5.25], p = 0.010 for infants aged ≤27 weeks. NICU C also had more infants with any one of the seven neonatal morbidities, aOR 3.00 [95% CI 1.5–5.99], p = 0.002, for infants aged 28-31 weeks, and aOR 3.24 [95%CI 1.67-6.29], p = 0.001 for infants aged ≤27 weeks. Rate of LOS for infants with GA ≤27 weeks was highest at NICU B (p = 0.005).
Table 3Comparing 3 Singapore NICUs – Unadjusted and adjusted odds ratios of survival and major neonatal morbidities.
| Parameters | Univariate model | Multivariate model | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ≤27 weeks | 28–31 weeks | ≤27 weeks | 28–31 weeks | |||||||||
| OR | 95% CI | p-value | OR | 95% CI | p-value | AOR† | 95% CI | p-value | AOR | 95% CI | p-value | |
| Survival | ||||||||||||
| NICU A | 1.00 1.48 1.15 | (ref) 0.65–3.37 0.58–2.29 | 0.347 0.685 | 1.00 0.96 0.76 | (ref) 0.31–2.98 0.24–2.39 | 0.941 0.643 | 1.00 1.44 1.28 | (ref) 0.63–3.29 0.63–2.60 | 0.392 0.490 | 1.00 1.03 0.72 | (ref) 0.32–3.31 0.23–2.30 | 0.955 0.582 |
| NICU B | ||||||||||||
| NICU C | ||||||||||||
| CLD (O2/[email protected]) | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 0.67 2.40 | (ref) 0.33–1.34 1.23–4.69 | 0.255 0.010 | 1.00 1.72 3.52 | (ref) 0.75–3.96 1.63–7.61 | 0.204 0.001 | 1.00 0.73 2.57 | (ref) 0.36–1.48 1.26–5.25 | 0.378 0.010 | 1.00 1.79 3.92 | (ref) 0.76–4.19 1.79–8.62 | 0.180 0.001 |
| IVH Grade 3 or 4 | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 1.35 1.48 | (ref) 0.48–3.82 0.59–3.71 | 0.574 0.402 | 1.00 / 0.87 | (ref) / 0.10–7.41 | / 0.902 | 1.00 1.35 1.23 | (ref) 0.47–3.90 0.48–3.18 | 0.578 0.663 | 1.00 / 1.08 | (ref) / 0.13–9.26 | / 0.915 |
| ROP Stage ≥3 | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 0.83 1.54 | (ref) 0.34–2.01 0.74–3.20 | 0.679 0.248 | 1.00 1.15 / | (ref) 0.12–11.28 / | 0.902 / | 1.00 0.87 1.70 | (ref) 0.36–2.14 0.80–3.64 | 0.767 0.169 | 1.00 1.12 / | (ref) 0.10–13.04 / | 0.927 / |
| Late-onset sepsis | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 3.29 2.05 | (ref) 1.49–7.26 0.93–4.50 | 0.003 0.074 | 1.00 1.83 0.43 | (ref) 0.62–5.45 0.05–3.39 | 0.277 0.423 | 1.00 3.14 1.84 | (ref) 1.41–6.98 0.82–4.14 | 0.005 0.139 | 1.00 1.83 0.43 | (ref) 0.60–5.58 0.05–3.42 | 0.287 0.423 |
| NEC Stage 2,3 | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 1.10 0.82 | (ref) 0.23–5.27 0.17–3.90 | 0.904 0.804 | 1.00 / 2.46 | (ref) / 0.44–13.78 | / 0.304 | 1.00 1.11 0.69 | (ref) 0.22–5.45 0.14–3.40 | 0.903 0.648 | 1.00 / 2.78 | (ref) / 0.49–15.72 | / 0.247 |
| Isolated gastrointestinal perforation | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 0.31 0.99 | (ref) 0.04–2.44 0.32–3.10 | 0.268 0.989 | 1.00 3.94 / | (ref) 0.24–63.83 / | 0.334 / | 1.00 0.30 0.86 | (ref) 0.04–2.32 0.26–2.78 | 0.246 0.796 | 1.00 6.00 / | (ref) 0.36–98.94 / | 0.210 |
| PVL | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 / 4.02 | (ref) / 0.55–29.16 | / 0.169 | 1.00 8.17 10.25 | (ref) 0.73–91.45 0.91–115.07 | 0.088 0.059 | 1.00 / 4.49 | (ref) / 0.59–33.83 | / 0.145 | 1.00 13.84 8.80 | (ref) 1.15–167.2 0.70–111.4 | 0.039 0.093 |
| Presented with any of the 7 neonatal morbidities | ||||||||||||
| NICU A (ref) NICU B NICU C | 1.00 1.16 3.27 | (ref) 0.61–2.19 1.71–6.27 | 0.648 <0.001 | 1.00 1.17 2.75 | (ref) 0.55–2.50 1.40–5.40 | 0.682 0.003 | 1.00 1.17 3.24 | (ref) 0.62–2.21 1.67–6.29 | 0.634 0.001 | 1.00 1.28 3.00 | (ref) 0.59–2.76 1.50–5.99 | 0.532 0.002 |
Reference group: NICU A.
a Adjusted ORs were based on multiple logistic regression, adjusted for outborn status, mode of delivery and multiple births.
Figs. S1 and S2 show the mean length of stay and mean weight before hospital discharge for the SG2017 group. The mean length of stay for a 24-week-old infant was more than double that of its 29-week-old counterpart, 148.1 ± 49.2 days vs. 62.6 ± 24.9 days.
3.3 Comparison of SG2017 with ANZNN2017 and VON2014
We compared the data of infants born at <30 weeks of gestation in SG2017 (n = 564) with those of their counterparts in ANZNN2017 (n = 2070) (Table 4). Cesarean birth rates among infants ≤25 weeks were significantly higher in SG2017 (50.32% vs. 40.53%, p = 0.031), which also recorded a higher incidence rate of multiple births (31.85% vs. 21.73%, p = 0.010). There was a significantly higher use of invasive ventilation across all three GA groups in SG2017; however, CPAP use was comparable.
Table 4Comparing SG2017 with ANZNN 2017 by gestational age groups.
| Parameters | Gestational weeks | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ≤25 weeks | 26–27 weeks | 28–29 weeks | ||||||||
| SG2017 | ANZNN 2017 | p-value | SG2017 | ANZNN 2017 | p-value | SG2017 | ANZNN 2017 | p-value | ||
| n | 157 | 497 | 177 | 654 | 230 | 919 | ||||
| Survived to discharge | 65.61% | 71.83% | 0.137 | 88.14% | 92.66% | 0.053 | 92.17% | 96.95% | 0.001 | |
| Any antenatal steroids | 85.99% | 89.92% | 0.170 | 92.66% | 92.80% | 0.949 | 87.83% | 92.10% | 0.040 | |
| Caesarean birth | 50.32% | 40.53% | 0.031 | 62.71% | 66.97% | 0.285 | 70.87% | 69.84% | 0.751 | |
| Multiples | 31.85% | 21.73% | 0.010 | 18.08% | 24.80% | 0.061 | 23.91% | 27.10% | 0.327 | |
| Late-onset sepsis (>48 or>72 h) | 17.20% | 31.03% | 0.001 | 10.17% | 14.20% | 0.162 | 4.78% | 6.00% | 0.477 | |
| Severe IVH (grade 3 or more) | 17.11% | 18.02% | 0.798 | 2.86% | 4.94% | 0.238 | 2.64% | 2.13% | 0.642 | |
| Severe ROP (stage 3 or more) | 40.35% | 34.15% | 0.157 | 7.27% | 7.04% | 0.916 | 1.88% | 1.27% | 0.480 | |
| Surfactant | 93.63% | 94.57% | 0.907 | 76.27% | 79.10% | 0.421 | 55.65% | 54.50% | 0.744 | |
| IPPV/ETT ventilation | 100.0% | 94.16% | 0.002 | 87.01% | 75.70% | 0.001 | 62.61% | 47.80% | <0.001 | |
| CPAP | 77.07% | 80.08% | 0.413 | 96.1% | 97.10% | 0.497 | 97.4% | 95.80% | 0.261 | |
| HFOV | 48.41% | 53.92% | 0.483 | 21.47% | 22.63% | 0.947 | 10.43% | 8.27% | 0.581 | |
| Inhaled Nitric Oxide | 14.65% | 19.72% | 0.362 | 7.34% | 7.65% | 0.991 | 3.04% | 4.03% | 0.786 | |
| Pulmonary air leak/pneumothorax | 10.19% | 7.85% | 0.368 | 2.82% | 4.40% | 0.339 | 5.22% | 3.60% | 0.263 | |
| CLD (O2 or CPAP at 36 weeks) | 70.64% | 80.17% | 0.012 | 40.00% | 55.79% | <0.001 | 18.27% | 26.04% | 0.015 | |
| Length of stay (median) | 110.00 | 116.80 | 0.075 | 89.00 | 88.56 | 0.560 | 62.00 | 65.32 | 0.032 | |
| ≤27 weeks | 28–31 weeks | |||||||||
| SG2017 | ANZNN 2017 | p-value | SG2017 | ANZNN 2017 | p-value | |||||
| NEC stage 2 or 3 | 3.89% | 9.10% | 0.002 | 1.44% | 1.10% | 0.501 | ||||
p-value from Chi-square test.
CLD, Chronic lung disease; CPAP, continuous positive airway pressure; IPPV/ETT, Intermittent Positive Pressure Ventilation/endotracheal tube; HFOV, high frequency oscillatory ventilation; IVH, intraventricular haemorrhage; NEC, necrotising enterocolitis; ROP, retinopathy of prematurity.
The rates of infants’ survival until discharge generally tended to be higher for ANZNN2017, reaching statistical significance for infants with GA 28–29 weeks (96.95% vs. 92.17%, p = 0.001). For infants born at 28–29 weeks, the length of stay was longer in the ANZNN2017 (65.32 vs. 62.00 days, p = 0.032). SG2017 had a significantly lower rate of LOS among infants ≤25 weeks (17.20% vs. 31.03%, p = 0.001), a lower rate of NEC stage 2 or 3 among infants ≤27 weeks (3.89% vs. 9.10%, p = 0.002), and a lower rate of CLD across all three GA groups.
Table 5 shows comparison between SG2017 and VON2014. SG2017 performed better than the best 25% of VON NICUs with respect to rates of mortality, CLD, LOS, NEC, and severe IVH; however, the rate of severe ROP (8.9%) in SG2017 exceeded the 90th percentile (8.6%) in VON2014.
Table 5Comparing SG2017 with VON2014 with selected variables.
| Parameters | SG2017 | VON 2014 | |
|---|---|---|---|
| (%) | Median (%) | Interquartile range (%) | |
| Mortality | 10.1 | 10.9 | 10.5–11.4 |
| CLD (O2 at 36w) | 20.0 | 28.0 | 24.9–31.6 |
| Late onset sepsis | 7.4 | 9.7 | 8.2–11.2 |
| NEC Stage 2,3 | 2.2 | 4.9 | 4.3–5.7 |
| SIVH | 4.4 | 7.9 | 7.6–8.2 |
| SROP | 8.9 | 6.0 | 5.2–7.1 |
CLD, chronic lung disease; NEC, necrotising enterocolitis; SIVH, severe intraventricular haemorrhage; SROP, severe retinopathy of prematurity.
4. Discussion
4.1 Trend in Singapore over a decade
The overall survival of VLBW infants remained stable across the two periods. However, the survival rate of very high-risk infants (GA ≤26 weeks) improved (74.7% vs. 66.7%, p = 0.174). This trend is consistent with observations reported in other studies.
7
, 8
, 9
In a previous Singaporean study, survival of infants ≤26 weeks was found to have increased between two periods, 1990–1998 and 1999–2007 (61.5% vs. 68.8%).10
Survival was further improved in the SG2017 cohort. Better survival in later periods may be due to optimization of care, resulting in less LOS, despite more peri-viable infants being offered NICU care (during SG2017, infants with GA ≥24 weeks were offered NICU care; for GA 23 weeks, it depended largely on parental choice. However, during the SG2007 era most 23 weeks infants were not offered NICU care).The rates of CLD tended to increase over a decade, even though there was no increase in infants ≤26 weeks, and despite an increase in antenatal steroid and surfactant use. In recent years, higher CLD rates have also been reported in Oceania and the USA.
8
,11
The higher CLD rates in SG2017 may be partly attributed to the improved survival of extremely preterm infants, who have the highest risk of developing CLD, or to the decreased use of postnatal steroids because of concerns about its adverse effects on the developing brain.12
,13
Systemic steroid remains one of the most effective therapeutic options for CLD prevention.14
CLD is the most serious postnatal complication associated with an increased risk for impairment.3
The quest for better preventive and therapeutic options continues.15
Management of PDA has seen a gradual but consistent change toward a decline in NSAID use. Although many trials have shown that NSAIDs are effective, no improvement in morbidities such as CLD has been observed.
16
Moreover, NSAIDs can cause serious side effects, including intestinal perforation.17
There is increasing evidence that conservative management of PDA is non-inferior to NSAID treatment.18
,19
The LOS rate was much lower in the infants of the SG2017 group. This has also been observed in other neonatal networks.
7
,8
SG2017 had significantly more infants on human milk feeding. Exclusive immunological factors in human milk protect against LOS. Human milk feeding was also associated with fewer days of parenteral nutrition (PN) and percutaneously inserted central catheters (PICCs), known risk factors of LOS.20
,21
We did not have data on the duration of PN or PICC use for both cohorts. Other factors that may have contributed to the reduced rates of LOS might include a better workforce, infection control practices, antibiotic stewardship, and probiotic use,22
but data on these factors were not obtained. The use of prophylactic fluconazole to prevent invasive Candida infection has become more prevalent following the results of several randomized trials.23
All three NICUs in the network have established a neurodevelopmental follow-up program. However, such data was not collected from the Network. Up to 20% of VLBW infants may have a developmental delay at the age of two years.
24
Teo et al. reviewed the neurological outcomes of infants with BW < 1250 g in a Singapore NICU. They found improvement in IQ scores over a 10-year period.25
Even though most Singaporean VLBW infants had outcomes comparable to their peers born at term, a significant number needed extra support during primary school years.26
The Singapore Network should consider tracking neurodevelopmental follow-up.4.1.1 Comparing data from the three NICUs
Differences in infant characteristics and outcomes were observed among the three NICUs. NICU C had more outborn infants. Outborn infants are more likely to have poorer outcomes. Triaging and earlier in utero transfer of high-risk pregnancies (<32 weeks) to a perinatal NICU may reduce the rates of outborn infants. It may be worthwhile to set up a national high-risk pregnancy in utero transfer coordinating center for this purpose.
27
NICU B had more infants born at <28 weeks with LOS. Quality improvement projects conducted whenever there is a spike in LOS are a staple for NICUs.
22
,28
We understand that NICU B performed a comprehensive root-cause analysis, and among others, one factor that might have caused increased LOS was the use of tap aerators colonized with various bacteria. The LOS sepsis rates in NICU B have since reduced to an average of 4.9% for 2018–2020 (personal communication). The CLD rate was higher in NICU C. This remained despite adjustment for outborn status, mode of delivery, and multiple births. There were also no significant differences in mortality and postnatal steroid use among the NICUs. Despite not having more infants diagnosed with CLD, NICU A had more infants who were discharged with supplemental oxygen. Using our current minimal dataset, we could not speculate reasons to account for the differences in CLD rates and home oxygen use. This suggests that we would need a quality improvement initiative for a more detailed investigation into care practice differences, which may have contributed to this. Of note, our CLD definition did not consider grades of severity, as was recently proposed in a refinement of CLD diagnostic criteria.29
The management of PDA varied. NICU C treated more 28–31 weeks infants with indomethacin/ibuprofen, and NICUs B and C performed more PDA ligations in infants <28 weeks. As noted earlier, PDA management in preterm infants is still controversial, and there is a trend towards a less aggressive approach.
30
NICU B discharged infants at a significantly lower mean weight than NICUs A and C. However, the average lengths of hospital stay among the three NICUs were similar, with all 3 hospitals having identical discharge criteria (PMA >35 weeks, weight >1.9 kg). Comparing other relevant variables, such as length, head circumference, PMA at hospital discharge, and feeding protocols, may provide clues to the observed differences. NICU B had significantly more infants fed human milk only at hospital discharge, and there were no cases of surgical NEC. Appropriately fortified human milk promotes healthy lean weight gain in preterm infants, as opposed to infant formula or donor human milk. This might account for NICU B having infants with lighter body weight at discharge. Human milk is also associated with reduced NEC and improved neurodevelopmental outcomes.31
,32
The SG2017 data on the mean length of stay and mean weight before hospital discharge (Figs. S1 and S2) are helpful for planning resource allocation. Infants with GA <28 weeks required a disproportionately longer hospital stay and were discharged with a heavier body weight, probably because of their continual need for in-hospital treatment for CLD.
33
4.2 Benchmarking against ANZNN
Survival rates tended to be higher for infants of ANZNN2017 than for those of SG2017 (statistically significant only for infants 28–29 weeks). Survival data should be interpreted with caution, as they are affected by resuscitation policies and congenital malformations.
34
Proportionately more SG2017 infants with GA ≤25 weeks were from multiple births, and these infants had higher rates of mortality.35
Another important confounder when comparing survival data is that SG2017 only included infants weighing <1500 g, regardless of GA, whereas ANZNN2017 included all infants from the respective GA groups, regardless of their BW (i.e., larger infants >1500 g were also included). Therefore, a comparison between infants in larger GA groups should be performed with this proviso.SG2017 had a significantly lower LOS in infants ≤25 weeks, less NEC in infants ≤27 weeks, and less CLD in all three GA groups. We did not have CRIB II score data from ANZNN for comparison. CRIB II scores provide information on illness severity on admission. Initial illness severity may affect outcomes.
36
Rates of CLD were consistently lower across all GA groups in SG2017 despite it having a higher proportion of infants on invasive ventilation. This observation seems to contradict the prevailing belief among neonatologists that using non-invasive ventilation (NIV) would be beneficial for CLD. However, recent data suggest that CLD rates have increased despite the increased use of NIV.8
,11
NIV may cause atelectrauma in the lungs of infants with more severe lung disease.37
The largest VLBW network is the VON, with over 1000 participating NICUs. The outcomes for SG2017 were better than those for VON2014 for all variables, except for severe ROP. Even though Singapore's data were not risk-adjusted, NICUs might want to review various practices that may influence ROP rates.
38
,39
4.3 Strengths and limitations
To the best of our knowledge, this is the first analysis of national VLBW data in Singapore. This allows for meaningful risk-adjusted comparisons among Singapore NICUs for the SG2017 cohort. However, because of the minimal dataset, risk-adjustment for VLBW infant outcomes could not be performed when comparing SG2017 with SG2007 and when comparing SG2017 with ANZNN2017 and VON2014. Due to the nature of the data collected, potential biases for outcomes, such as NICU occupancy and medical and nursing workforce data, were not available and could not be accounted for. Comparison with VON2014 was also limited by the slightly different enrolment criteria, as VON2014 included VLBW infants with BW > 1500 g and infants who died in the delivery room.
5. Conclusions
Outcomes for Singaporean VLBW infants have improved over the past decade. Compared to international benchmarks, Singapore NICUs performed well in terms of low rates of mortality, LOS, CLD, severe IVH, and NEC; however, the rates of severe ROP were higher. Subtle differences observed in performance among Singaporean NICUs present a rationale for collaboration in quality improvement. A national coordinating NICU for high-risk pregnancies may help streamline NICU resource allocation for the care of vulnerable VLBW infants.
Declaration of competing interest
The authors have no conflicts of interest to declare.
Acknowledgement
We thank Dr. Mary Grace Tan for her support in data collection for the SGH.
This study was partially funded by the Perinatal Society of Singapore and AbbVie Pte Ltd.
Appendix A. Supplementary data
The following are the Supplementary data to this article.
- Multimedia component 1
- Figure S1
Mean length of stay (in days) in hospital by completed gestational weeks at birth in cohort SG2017 (error bar indicates standard deviations).
- Figure S2
Mean weight (in grams) before discharge home by completed gestational weeks at birth in cohort SG2017 (error bar indicates standard deviations).
References
- Registry of births and deaths: report on registration on births and deaths 2018.Immigration and Checkpoints Authority Singapore, Singapore2018
- Mortality and length of stay of very low birth weight and very preterm infants: a EuroHOPE study.PLoS One. 2015; 10e0131685
- Cognitive outcomes of children born extremely or very preterm since the 1990s and associated risk factors: a meta-analysis and meta-regression.JAMA Pediatr. 2018; 172: 361-367
- Trends in mortality and morbidity for very low birth weight infants, 1991–1999.Pediatrics. 2002; 110: 143-151
- Outcomes of very-low-birthweight infants at 3 years of age born in 2003-2004 in Japan.Pediatr Int. 2011; 53: 1051-1058
- Report of the Australian and New Zealand Neonatal Network 2017.ANZNN, Sydney2019
- Variation in performance of neonatal intensive care units in the United States.JAMA Pediatr. 2017; 171e164396
- Trends in care practices, morbidity, and mortality of extremely preterm neonates, 1993-2012.JAMA. 2015; 314: 1039-1051
- Association between year of birth and 1-year survival among extremely preterm infants in Sweden during 2004–2007 and 2014–2016.JAMA. 2019; 321: 1188-1199
- Borderline viability–neonatal outcomes of infants in Singapore over a period of 18 years (1990–2007).Ann Acad Med Singap. 2013; 42: 328-337
- Ventilation in extremely preterm infants and respiratory function at 8 years.N Engl J Med. 2017; 377: 329-337
- Early dexamethasone therapy in preterm infants: a follow-up study.Pediatrics. 1998; 101: E7
- Hazards of systemic steroids for ventilator-dependent preterm infants: what would a parent want?.CMAJ. 2001; 165: 33-34
- Controlled trial of dexamethasone in respirator-dependent infants with bronchopulmonary dysplasia.Pediatrics. 1985; 75: 106-111
- Present and future of bronchopulmonary dysplasia.J Clin Med. 2020; 9: 1539
- PDA: to treat or not to treat.Congenit Heart Dis. 2019; 14: 46-51
- Spontaneous intestinal perforation in extremely low birth weight infants: association with indometacin therapy and effects on neurodevelopmental outcomes at 18-22 months corrected age.Arch Dis Child Fetal Neonatal Ed. 2013; 98: F127-F132
- Early treatment versus expectant management of hemodynamically significant patent ductus arteriosus for preterm infants.Cochrane Database Syst Rev. 2020; 12: CD013278
- Prolonged ductal patency in preterm infants: does it matter?.J Pediatr. 2021; 229: 12-14
- Impact of early human milk on sepsis and health-care costs in very low birth weight infants.J Perinatol. 2013; 33: 514-519
- Risk factors for late-onset sepsis in preterm infants: a multicenter case-control study.Neonatology. 2019; 116: 42-51
- Late-onset sepsis in preterm infants: update on strategies for therapy and prevention.Expert Rev Anti Infect Ther. 2015; 13: 487-504
- A multicenter, randomized trial of prophylactic fluconazole in preterm neonates.N Engl J Med. 2007; 356: 2483-2495
- Factors affecting neurodevelopmental outcome at 2 years in very preterm infants below 1250 grams: a prospective study.J Perinatol. 2018; 38: 1093-1100
- Long-term neurodevelopmental outcomes of premature infants in Singapore.Ann Acad Med Singap. 2018; 47: 63-70
- When not all is well.in: Outcomes of Singaporean Chinese very low birth weight children in mainstream primary schools. Thesis, University of Leicester, England2002
- All the right moves: why in utero transfer is both important for the baby and difficult to achieve and new strategies for change.F1000Res. 2020; 9 (F1000 Faculty Rev 979)
- Nosocomial infection reduction in VLBW infants with a statewide quality-improvement model.Pediatrics. 2011; 127: 419-426
- The diagnosis of bronchopulmonary dysplasia in very preterm infants. An evidence-based approach.Am J Respir Crit Care Med. 2019; 200: 751-759
- Changing patterns of patent ductus arteriosus surgical ligation in the United States.Semin Perinatol. 2018; 42: 253-261
- Persistent beneficial effects of breast milk ingested in the neonatal intensive care unit on outcomes of extremely low birth weight infants at 30 months of age.Pediatrics. 2007; 120: e953-e959
- Dose-response relationship between donor human milk, mother's own milk, preterm formula, and neonatal growth outcomes.J Pediatr Gastroenterol Nutr. 2018; 67: 90-96
- Prevalence of chronic respiratory morbidity, length of stay, inpatient readmissions, and costs among extremely preterm infants with bronchopulmonary dysplasia.Expert Rev Pharmacoecon Outcomes Res. 2021; 21: 1117-1125
- Outcome comparison of very preterm infants cared for in the neonatal intensive care units in Australia and New Zealand and in Canada.J Paediatr Child Health. 2015; 51: 881-888
- Methods of mortality risk adjustment in the NICU: a 20-year review.Pediatrics. 2013; 131: S68-S74
- CRIB II: an update of the clinical risk index for babies score.Lancet. 2003; 361: 1789-1791
- Lung inflammation in preterm infants with respiratory distress syndrome: effects of ventilation with different tidal volumes.Pediatr Pulmonol. 2006; 41: 357-363
- Predictors for the development of referral-warranted retinopathy of prematurity in the telemedicine approaches to evaluating acute-phase retinopathy of prematurity (e-ROP) study.JAMA Ophthalmol. 2015; 133: 304-311
- Low energy intake during the first 4 weeks of life increases the risk for severe retinopathy of prematurity in extremely preterm infants.Arch Dis Child Fetal Neonatal Ed. 2016; 101: F108-F113
Article info
Publication history
Published online: March 15, 2023
Accepted:
December 20,
2022
Received in revised form:
November 28,
2022
Received:
June 6,
2022
Publication stage
In Press Corrected ProofIdentification
Copyright
© 2023 Taiwan Pediatric Association. Published by Elsevier Taiwan LLC.
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
