Predictive ability of neonatal illness severity scores for early death in extremely premature infants

Comparing different scoring systems for predicting mortality risk in preterm infants: a systematic review and network meta-analysis

Background

This study aimed to compare the predictive values of eight scoring systems (Neonatal Critical Illness Score [NCIS], Neonatal Therapeutical Intervention Score System [NTISS], Clinical Risk Index for Babies [CRIB], Clinical Risk Index for Babies II [CRIB-II], Score for Neonatal Acute Physiology Perinatal Extension [SNAPPE], Score for Neonatal Acute Physiology Perinatal Extension II [SNAPPE-II], Score for Neonatal Acute Physiology [SNAP], and Score for Neonatal Acute Physiology II [SNAP-II]) for the mortality risk among preterm infants.

Methods

The Embase, PubMed, Chinese Biomedical Database, Web of Science, and Cochrane Library databases were searched to collect studies that compared different scoring systems in predicting the mortality risk in preterm infants from database inception to March 2023. Literature screening, data extraction, and bias risk assessment were independently conducted by two researchers. Subsequently, the random-effects model was used for the network meta-analysis.

Results

A total of 19 articles were included, comprising 14,377 preterm infants and 8 scoring systems. Compared to CRIB-II, NCIS, NTISS, SNAP-II, and SNAPPE-II, CRIB demonstrated better predictive efficiency for preterm infant mortality risk (P < 0.05). Relative to CRIB, CRIB-II, and SNAPPE, SNAP-II had worse predictive efficiency for preterm infant mortality risk (P < 0.05). The surface under the cumulative ranking curve of the eight scoring systems was as follows: CRIB (0.980) > SNAPPE (0.718) >SNAP (0.534) >CRIB-II (0.525) >NTISS (0.478) >NCIS (0.422) >SNAPPE-II (0.298) >SNAP-II (0.046).

Conclusion

The CRIB scoring system showed the highest accuracy in predicting preterm infant mortality risk and was simple to perform. Therefore, CRIB selection can be prioritized in clinical practice.

Systematic Review Registration

https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=434731, PROSPERO (CRD42023434731).

Comparison of STOPS and SNAPPE-II in Predicting Neonatal Survival at Hospital Discharge: A Prospective, Multicentric, Observational Study

OBJECTIVE

To compare SNAPPE-II and STOPS admission severity scores in neonates admitted to neonatal intensive care unit (NICU) with a gestational age of ≥ 33 wk.

METHODS

In this multicenter, prospective, observational study, the sickness scoring was done on all the neonates at 12 h after admission to the NICUs. The scoring systems were compared by the area under the curve (AUC) on the receiver operating characteristics (ROC) curve.

RESULTS

A total of 669 neonates with gestational age ≥ 33 wk (mortality rate: 2.4%), who were admitted to five participating NICUs within 24 h of birth, were included. Both SNAPPE-II and STOPS had the good discriminatory and predictive ability for mortality with AUCs of 0.965 [95% confidence interval (CI): 0.94-0.98] and 0.92 (95% CI: 0.87-0.99), respectively. The STOPS scoring system with a cutoff score ≥ 4 on the ROC curve had 85% accuracy, whereas the SNAPPE-II cutoff score ≥ 33 on the ROC curve had 94% accuracy in predicting mortality.

CONCLUSION

In infants with the gestational age of ≥ 33 wk, SNAPPE-II and STOPS showed similar predictive ability, but the STOPS score, being a simpler clinical tool, might be more useful in resource-limited settings.

Prediction Models for Bronchopulmonary Dysplasia in Preterm Infants: A Systematic Review and Meta-Analysis

OBJECTIVE

To review systematically and assess the accuracy of prediction models for bronchopulmonary dysplasia (BPD) at 36 weeks of postmenstrual age.

STUDY DESIGN

Searches were conducted in MEDLINE and EMBASE. Studies published between 1990 and 2022 were included if they developed or validated a prediction model for BPD or the combined outcome death/BPD at 36 weeks in the first 14 days of life in infants born preterm. Data were extracted independently by 2 authors following the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (ie, CHARMS) and PRISMA guidelines. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool (ie, PROBAST).

RESULTS

Sixty-five studies were reviewed, including 158 development and 108 externally validated models. Median c-statistic of 0.84 (range 0.43-1.00) was reported at model development, and 0.77 (range 0.41-0.97) at external validation. All models were rated at high risk of bias, due to limitations in the analysis part. Meta-analysis of the validated models revealed increased c-statistics after the first week of life for both the BPD and death/BPD outcome.

CONCLUSIONS

Although BPD prediction models perform satisfactorily, they were all at high risk of bias. Methodologic improvement and complete reporting are needed before they can be considered for use in clinical practice. Future research should aim to validate and update existing models.

Updates in Neonatal Extracorporeal Membrane Oxygenation and the Artificial Placenta

Extracorporeal life support, initially performed in neonates, is now commonly used for both pediatric and adult patients requiring pulmonary and/or cardiac support. Data suggests the clinical feasibility of Extracorporeal Membrane Oxygenation for premature infants (29-33 weeks estimated gestational age [EGA]). For extremely premature infants less than 28 weeks EGA, an artificial placenta has been developed to recreate the fetal environment. This approach is investigational but clinical translation is promising. In this article, we discuss the current state and advances in neonatal and "preemie Extracorporeal Membrane Oxygenation" and the development of an artificial placenta and its potential use in extremely premature infants.

Neurodevelopmental Outcomes of Infants at <29 Weeks of Gestation Born in Canada Between 2009 and 2016

OBJECTIVE

To evaluate changes in mortality or significant neurodevelopmental impairment (NDI) in children born at <29 weeks of gestation in association with national quality improvement initiatives.

STUDY DESIGN

This longitudinal cohort study included children born at 22^0/7 to 28^6/7 weeks of gestation who were admitted to Canadian neonatal intensive care units between 2009 and 2016. The primary outcome was a composite rate of death or significant NDI (Bayley Scales of Infant and Toddler Development, Third Edition score <70, severe cerebral palsy, blindness, or deafness requiring amplification) at 18-24 months corrected age. To evaluate temporal changes, outcomes were compared between epoch 1 (2009-2012) and epoch 2 (2013-2016). aORs were calculated for differences between the 2 epochs accounting for differences in patient characteristics.

RESULTS

The 4426 children included 1895 (43%) born in epoch 1 and 2531 (57%) born in epoch 2. Compared with epoch 1, in epoch 2 more mothers received magnesium sulfate (56% vs 28%), antibiotics (69% vs 65%), and delayed cord clamping (37% vs 31%) and fewer infants had a Score for Neonatal Acute Physiology, version II >20 (31% vs 35%) and late-onset sepsis (23% vs 27%). Death or significant NDI occurred in 30% of children in epoch 2 versus 32% of children in epoch 1 (aOR, 0.86; 95% CI, 0.75-0.99). In epoch 2, there were reductions in the need for hearing aids or cochlear implants (1.4% vs 2.6%; aOR, 0.50; 95% CI, 0.31-0.82) and in blindness (0.6% vs.1.4%; aOR, 0.38; 95% CI, 0.18-0.80).

CONCLUSIONS

Among preterm infants born at <29 weeks of gestation, composite rates of death or significant NDI and rates of visual and hearing impairment were significantly lower in 2013-2016 compared with 2009-2012.

Comparison of CRIB-II with SNAPPE-II for predicting survival and morbidities before hospital discharge in neonates with gestation ≤ 32 weeks: a prospective multicentric observational study

Various studies validated and compared Score for Neonatal Acute Physiology with Perinatal extension-II (SNAPPE-II) and Clinical Risk Index for Babies-II (CRIB-II) admission sickness severity scores for predicting survival, but very few studies compared them for predicting the morbidities in preterm infants. In this multicenter prospective observational study, SNAPPE-II and CRIB-II newborn illness severity scores were compared for predicting mortality and morbidities in infants with gestational age of ≤ 32 weeks. Major morbidities were classified as bronchopulmonary dysplasia, abnormal cranial ultrasound (presence of intraventricular hemorrhage grade III or more or periventricular leukomalacia grade II to IV), and retinopathy of prematurity requiring treatment. Combined adverse outcome was defined as death or any major morbidity. Comparison of the scoring systems was done by area under the curve (AUC) on receiver operating characteristics curve (ROC curve) analysis. A total of 419 neonates who were admitted to 5 participating NICUs were studied. The mortality rate in the study population was 8.8%. Both CRIB-II (AUC: 0.795) and SNAPPE-II (AUC: 0.78) had good predictive ability for in-hospital mortality. For predicting any one of the major morbidities and combined adverse outcome, CRIB-II had better predictive ability than SNAPPE-II with AUC of 0.83 vs. 0.70 and 0.85 vs. 0.74, respectively.

CONCLUSION

In infants with gestational age of ≤ 32 weeks, both CRIB-II and SNAPPE-II are good scoring systems for predicting mortality. CRIB-II, being a simpler scoring system and having better predictive ability for major morbidities and combined adverse outcome, is preferable over SNAPPE-II.

WHAT IS KNOWN

• SNAPPE-II and CRIB-II scores have good predictive ability on in-hospital mortality in preterm neonates.

WHAT IS NEW

• SNAPPE-II and CRIB-II both have good predictive ability for mortality, but CRIB-II has better ability for short-term morbidities related to the prematurity.

Machine Learning Algorithms to Predict Mortality of Neonates on Mechanical Intubation for Respiratory Failure

BACKGROUND

Early identification of critically ill neonates with poor outcomes can optimize therapeutic strategies. We aimed to examine whether machine learning (ML) methods can improve mortality prediction for neonatal intensive care unit (NICU) patients on intubation for respiratory failure.

METHODS

A total of 1734 neonates with respiratory failure were randomly divided into training (70%, n = 1214) and test (30%, n = 520) sets. The primary outcome was the probability of NICU mortality. The areas under the receiver operating characteristic curves (AUCs) of several ML algorithms were compared with those of the conventional neonatal illness severity scoring systems including the NTISS and SNAPPE-II.

RESULTS

For NICU mortality, the random forest (RF) model showed the highest AUC (0.939 (0.921-0.958)) for the prediction of neonates with respiratory failure, and the bagged classification and regression tree model demonstrated the next best results (0.915 (0.891-0.939)). The AUCs of both models were significantly better than the traditional NTISS (0.836 (0.800-0.871)) and SNAPPE-II scores (0.805 (0.766-0.843)). The superior performances were confirmed by higher accuracy and F1 score and better calibration, and the superior and net benefit was confirmed by decision curve analysis. In addition, Shapley additive explanation (SHAP) values were utilized to explain the RF prediction model.

CONCLUSIONS

Machine learning algorithms increase the accuracy and predictive ability for mortality of neonates with respiratory failure compared with conventional neonatal illness severity scores. The RF model is suitable for clinical use in the NICU, and clinicians can gain insights and have better communication with families in advance.

Development of an artificial placenta for support of premature infants: narrative review of the history, recent milestones, and future innovation

Over 50 years ago, visionary researchers began work on an extracorporeal artificial placenta to support premature infants. Despite rudimentary technology and incomplete understanding of fetal physiology, these pioneering scientists laid the foundation for future work. The research was episodic, as medical advances improved outcomes of premature infants and extracorporeal life support (ECLS) was introduced for the treatment of term and near-term infants with respiratory or cardiac failure. Despite ongoing medical advances, extremely premature infants continue to suffer a disproportionate burden of mortality and morbidity due to organ immaturity and unintended iatrogenic consequences of medical treatment. With advancing technology and innovative approaches, there has been a resurgence of interest in developing an artificial placenta to further diminish the mortality and morbidity of prematurity. Two related but distinct platforms have emerged to support premature infants by recreating fetal physiology: a system based on arteriovenous (AV) ECLS and one based on veno-venous (VV) ECLS. The AV-ECLS approach utilizes only the umbilical vessels for cannulation. It requires immediate transition of the infant at the time of birth to a fluid-filled artificial womb to prevent umbilical vessel spasm and avoid gas ventilation. In contradistinction, the VV-ECLS approach utilizes the umbilical vein and the internal jugular vein. It would be applied after birth to infants failing maximal medical therapy or preemptively if risk stratified for high mortality and morbidity. Animal studies are promising, demonstrating prolonged support and ongoing organ development in both systems. The milestones for clinical translation are currently being evaluated.