Publications

Abstract

Background: Machine-learning methods are gaining in popularity to predict medical events but their added value to other methods is still to be determined. We compared performances of clinical prediction models for bronchopulmonary dysplasia (BPD) or death in very preterm infants using logistic regression and random forests methods.

Methods: Two population-based cohorts of very preterm infants were used: EPIPAGE-2 (France, 2011) for development and internal validation and EPICE (Europe, 2011) for external validation. Eligible infants were born before 30 weeks’ gestation and admitted in neonatal units. BPD was defined as any respiratory support at 36 weeks postmenstrual age. Candidate predictors were available shortly after birth or at day 3. Logistic regression and random forest models performance was assessed in terms of discrimination (c-statistic) and calibration plots.

Results: Prevalence of BPD/death was 32.1% (668/1923) in EPIPAGE-2 and 41.0% (1368/3335) in EPICE. At both time points, logistic regression and random forest models showed similar performance during internal validation. At birth, external validation in EPICE showed good discrimination (logistic regression model: c-statistics 0.81, 95% CI 0.80-0.83; random forest: 0.80, 95% CI 0.79-0.81) but both models underestimated the probability of BPD/death. Model performances were heterogeneous throughout European regions.

Conclusions: Both modelling methods performed similarly to predict BPD/death shortly after birth in very preterm children

Impact: Whether machine-learning methods predict better short-term respiratory outcomes in very preterm infants than logistic regression models is debated. Random forest-based prediction models did not perform better than logistic regression to predict bronchopulmonary dysplasia or death shortly after birth in very preterm infants. Calibration performances varied among European countries. While offering the same performance, regression models are easier to understand, to disseminate and to apply to different populations.

Abstract

Background: Compromised myocardial function and persistent elevated pulmonary vascular resistance are common among neonates treated with therapeutic hypothermia (TH) for hypoxic-ischemic encephalopathy (HIE). There is a lack of data regarding persistence of cardiac alterations after discharge from the neonatal intensive care unit (NICU).

Methods: We assessed cardiovascular profiles after NICU discharge. Echocardiogram data, including speckle-tracking echocardiography (STE), were extracted from the latest outpatient scan. Data were compared by initial amplitude-integrated encephalography (aEEG) profiles on admission [normal vs. abnormal].

Results: In total, 70 (19%) neonates had a follow-up echocardiogram (22 with initial normal aEEG, 48 with abnormal aEEG). Age at follow-up was similar between the two groups (6.2 vs. 7.7 months, [p = 0.08]). Neonates with an initially abnormal aEEG showed more negative Right Ventricle (RV)-peak global longitudinal strain (-28.2 vs. -26.0%, [p = 0.02]), RV-peak free wall longitudinal strain rate (-1.24 vs. -1.10 [1/second], [p = 0.01]), and RV-peak free wall longitudinal strain rate (-1.50 vs. -1.27 [1/second], [p = 0.001]). These associations remained after multilinear regression analysis, indicating persistent enhanced RV contraction in the abnormal aEEG group.

Conclusion: Neonates with initial abnormal aEEG profiles exhibited increased RV contraction after NICU discharge. Future studies should explore long-term cardiovascular follow-up of neonates with HIE, beyond the perinatal period.

Impact: What is the key message of your article? Cardiac performance in hypoxic ischemic encephalopathy is linked to adverse outcomes. Survivors with an abnormal aEEG at admission showed increased right ventricular contractility at follow-up, possibly related to an adverse adaptation to the initial insult. What does it add to the existing literature? This study offers insights into long-term cardiovascular outcomes in neonates with HIE, focusing on the link between initial aEEG abnormalities and later RV function. What is the impact? The findings underscore the importance of early cardiovascular assessments and monitoring in neonates undergoing TH for HIE, potentially guiding future follow-up protocols.

Abstract

Introduction: Acute respiratory disorders and bronchopulmonary dysplasia (BPD) are frequent and serious complications in very preterm infants (VPI). These conditions are influenced by several factors, including fluid imbalances and abnormal postnatal changes in body water composition (BWC). We aimed to investigate the association between BWC and the severity of respiratory disease in VPI using bioelectrical impedance analysis (BIA).

Material and methods: We conducted an observational study in a third-level neonatal intensive care unit at Reunion Island University Hospital. Infants born before 32 weeks of gestation, with BWC monitoring during hospital stay, were included. The severity of respiratory disease was assessed using mean airway pressure (MAP) recorded at the same time as BIA. Secondary outcomes included the Lung Ultrasound Score (LUS), measured weekly from 28 days of life, and BDP occurrence and severity at 36 weeks of postconceptional age.

Results: We included 34 VPI (BW 1,131 ± 352 g, gestational age 28.0 ± 2.2) and performed 267 BWC analyses (average: 8 BIA/patient). MAP was directly associated with extracellular water (ECW, % of body weight) (p < 0.001), even after adjustment for confounding variables at multivariate analysis (p < 0.02). No significant association was observed between total body water measures and the LUS. ECW was significantly associated with the occurrence of severe BPD (p < 0.003).

Abstract

Objective : to use the Delphi method to gain insight into approaches to prenatal diagnosis and management of preterm birth (PTB) in twin pregnancies, including complications such as twin-to-twin transfusion syndrome (TTTS) and a short and/or dilated cervix.

Methods : a three-round Delphi process was conducted among an international panel of experts to assess their approach to prevention, monitoring and management strategies for PTB in twin pregnancies. Experts were selected based on their publication record or membership of related organizations. Response options were multiple-choice answers or a five-point Likert scale. A priori, a cut-off of ≥ 70% agreement was used to define consensus.

Results : a total of 117 experts participated in the first round, of whom 94/117 (80.3%) completed all subsequent rounds. Representatives came from at least 22 countries (across five continents), most commonly the USA (50.4%) and the UK (12.0%). Over 70% of experts performed routine screening of cervical length (CL) using transvaginal ultrasound at 18–23 weeks’ gestation, using CL ≤ 25 mm to diagnose short cervix in twin pregnancies, regardless of a history of PTB. In twin pregnancies with a short non-dilated cervix, most experts offered vaginal progesterone rather than pessary or cervical cerclage, regardless of a history of PTB. In twin pregnancies with asymptomatic dilated cervix, consensus was reached (88.3% agreement) for placement of cervical cerclage, performed up to 24 weeks’ gestation (67.5% agreement; no consensus). Similarly, 96.1% of experts agreed that performing serial transvaginal ultrasound measurements of CL at 16–24 weeks’ gestation was warranted in women with a current singleton pregnancy who had a previous twin pregnancy that required physical examination-indicated cerclage; these patients should be considered high risk for PTB (83.1% agreement). In twin pregnancies with TTTS, laser surgery is offered by most experts, regardless of preoperative CL. In patients with TTTS and short CL, most experts would recommend cervical cerclage (71.9%) or vaginal progesterone (65.6%) rather than pessary or expectant management. However, no consensus was reached on measures to prevent PTB in cases of TTTS with cervical dilation.

Conclusions : This Delphi consensus study highlights practice variations among healthcare providers worldwide in the evaluation and management of PTB in twin pregnancies, which often differ from recommendations given by national and international societies.

Abstract

Background: Bronchopulmonary dysplasia (BPD) in extremely low gestational age neonates (ELGANs) was associated with neurodevelopmental impairment (NDI). However, the best endpoint of BPD assessment to predict subsequent NDI remains unclear.

Methods: We re-analyzed the PREMILOC trial, previously designed to test the effect of prophylactic hydrocortisone on survival without BPD at 36 weeks of postmenstrual age (BPD_W36) in ELGANs, to compare predictive models of NDI considering baseline characteristics, respiratory course up to and BPD status at 36 or 40 weeks of postmenstrual age (BPD_W36/BPD_W40).

Results: Among 404/519 (77.8%) infants enrolled in the trial alive at 2 years of age, all neurocognitive scores were available for 302 (74.8%) patients. Gestational diabetes and sex were identified as the only statistically significant baseline predictors of NDI. Adding BPD_W40 to this baseline model was found to be superior to predict NDI compared to BPD_W36, leading to a mean difference of the developmental quotient of -6.7 points (95% confidence interval: -10.0 to -3.50, P < 0.001). The prophylactic hydrocortisone treatment effect on survival without BPD_W40 was found to be highly significant (OR = 2.08 [95% confidence interval: 1.36 to 3.17], P < 0.001).

Conclusions: These data suggest a better accuracy of BPD_W40 to predict NDI in ELGANs, an important finding for future clinical trials and research in drug development.

Registration numbers: EudraCT number 2007-002041-20, ClinicalTrial.gov number, NCT00623740.

Impact: The best endpoint to assess BPD as a surrogate to predict neurocognitive impairment in infants born extremely preterm remains unclear. This study strongly suggests a better discriminative value of BPD as assessed at 40 weeks of postmenstrual age (instead of 36 weeks) to predict neurocognitive impairments at 2 years of age in children born extremely preterm. This study supports the switch up to 40 weeks of the primary outcome chosen in future clinical trials designed to prevent BPD. Our data also provide evidence of the beneficial effect of HC on preventing BPD at full-term equivalent age

Abstract

Introduction: Recent evidence suggests a restrictive approach toward blood transfusions for management of preterm infants. Objective was to survey blood transfusion practises in preterm neonates <29 weeks’ gestation among 12 population-based neonatal networks participating in the International Network for Evaluating Outcomes in Neonates (iNeo).

Methods: An online survey based on 2023 practices was sent to 608 neonatal intensive care units (NICUs): Australia/New Zealand (30), Brazil (20), Canada (32), Finland (5), France (70), Israel (26), Japan (292), Poland (56), Spain (55), Sweden (9), Switzerland (9), and Tuscany, Italy (4). Transfusion thresholds in 4 different scenarios were surveyed: (a) infants invasively ventilated within first 7 postnatal days, (b) infants invasively ventilated after 7 days, (c) stable infants on noninvasive respiratory support, and (d) stable infants requiring no respiratory support.

Results: A total of 382 NICUs (63%) responded. Transfusion practices varied within networks and between countries. For invasively ventilated infants, the transfusion threshold during first 7 days after birth was a hematocrit <underline>≤</underline>35% in 79% of NICUs, and at an age ≥8 days, the transfusion threshold was a hematocrit <underline>≤</underline>30% in 68% of NICUs. For stable infants on noninvasive ventilation, the transfusion threshold was a hematocrit <underline>≤</underline>30% in 80%, and in those without respiratory support, the transfusion threshold was a hematocrit of <underline>≤</underline>25% in 68% of NICUs.

ABSTRACT

Background : population-based data are needed to reliably assess the impact of SARS-CoV-2 infection during pregnancy.

Objectives : to estimate the population-based incidence of SARS-CoV-2 infection and its severe forms in the obstetric population, identify risk factors of severe SARS-CoV-2 infection (severe COVID-19) and describe delivery, maternal and neonatal outcomes by disease severity, using a definition of severity based on organ dysfunction.

Methods : a prospective population-based study conducted over the three first pandemic waves between March 2020 and April 2021 in 281 maternity hospitals in six French regions included all women with SARS-CoV-2 infection during pregnancy or within 7 days post-partum, whether symptomatic or not, hospitalised or not. Severe COVID-19 forms were defined a priori using clinical, biological and management criteria of organ dysfunction. We calculated infection and severe infection rates and studied associations between sociodemographic, medical and pregnancy characteristics and severe COVID-19 by univariate and multivariate modified Poisson regression modelling.

Results : from a population of 385,214 deliveries in the participating regions, 6015 women with SARS-CoV-2 infection were identified, including 337 severe cases. The rates of severe COVID-19 were 1.1, 0.9 and 3.6 per 1000 deliveries during the first, second and third pandemic waves, respectively, and the proportions of severe COVID-19 were 8.6%, 3.4% and 9.3%, respectively. On multivariate analysis, the risk of severe COVID-19 was associated with younger and older age, migrant status, living with > 4 people, overweight or obesity, chronic hypertension or diabetes and infection ≥ 22 weeks of gestation rather than earlier in pregnancy. Neonatal morbidity occurred mostly with severe maternal infection.

Conclusion : Using an organ-based definition of severity and population-based data, rates of severe COVID-19 appeared lower than in previous studies. A permanent perinatal surveillance system is needed to assess efficiently and rapidly the impact of future pandemics.

Abstract

Necrotizing enterocolitis (NEC) is probably the most painful intestinal disease affecting infants born preterm. NEC is known to cause highly severe and prolonged pain that has been associated with adverse short- and long-term effects. However, research on pain management in infants with NEC is scarce. This is likely due to its low incidence and very acute occurrence. As a result, the optimal pain management for these vulnerable infants remains unknown, and analgesic therapy practices are highly variable. Therefore, we aimed to establish expert-based consensus recommendations on pain management for NEC. Experts of the European Society for Paediatric Research (ESPR) Special Interest Groups on Neonatal pain and NEC were invited to participate in two consensus meetings. Prior to the first hybrid consensus meeting, an online survey provided input for potential recommendations. During the consensus meetings, experts shared clinical expertise and voted on recommendations. An expert consensus statement, comprising nine recommendations on optimal pain assessment and pain treatment in infants with NEC, was developed. Expert recommendations included regular pain assessments with a neonatal pain scale with additional assessments on indication and pre-emptive administration of analgesic therapy (e.g., paracetamol and an opioid) in infants with NEC stage ≥ II.

Conclusion: This expert consensus statement provides clinical recommendations essential for any healthcare professional caring for premature infants with NEC. The recommended guidance this statement provides on pain management strategies is key to preventing and reducing pain in this vulnerable population.

What is known: • Necrotizing enterocolitis (NEC) is a very painful disease, making effective pain management essential. • Current pain management practices for infants with NEC are highly variable.

What is new: • This expert consensus statement provides recommendations on optimal pain assessment and pain treatment in infants with NEC. • These clinical recommendations may help better prevent pain in these vulnerable infants.

Abstract

Results: in both cohorts, treatment with inhaled nitric oxide was not associated with IQ at age 5–6 years. Analysis identified maternal educational level, gestational age at discharge from hospital, intraventricular haemorrhage and maternal country of birth as important factors associated with IQ scores.

Aim : To investigate whether treatment with inhaled nitric oxide is associated with cognitive performance at age 5–6 years in preterm-born children.

Methods : we analysed preterm children from two large European cohort studies, the German Neonatal Network (GNN) (N = 3606) and the French EPIPAGE-2 cohort (N = 2579) admitted to neonatal care and followed up at age 5–6 years. Both cohorts had recorded data on iNO treatment. General cognitive ability was tested with IQ tests. Classification and Regression trees analysis was used to identify prenatal, perinatal and neonatal, clinical and social-environmental predictors of IQ.

Conclusion : Treatment with inhaled nitric oxide was neither negatively nor positively associated with IQ at age 5–6 years. Neonatal and brain health, as well as socioeconomic factors are important for cognitive performance in early childhood.