Dataset Information

Artificial intelligence model comparison for risk factor analysis of patent ductus arteriosus in nationwide very low birth weight infants cohort

ABSTRACT: Despite the many comorbidities and high mortality rate in preterm infants with patent ductus arteriosus (PDA), therapeutic strategies vary depending on the clinical setting, and most studies of the related risk factors are based on small sample populations. We aimed to compare the performance of artificial intelligence (AI) analysis with that of conventional analysis to identify risk factors associated with symptomatic PDA (sPDA) in very low birth weight infants. This nationwide cohort study included 8369 very low birth weight (VLBW) infants. The participants were divided into an sPDA group and an asymptomatic PDA or spontaneously close PDA (nPDA) group. The sPDA group was further divided into treated and untreated subgroups. A total of 47 perinatal risk factors were collected and analyzed. Multiple logistic regression was used as a standard analytic tool, and five AI algorithms were used to identify the factors associated with sPDA. Combining a large database of risk factors from nationwide registries and AI techniques achieved higher accuracy and better performance of the PDA prediction tasks, and the ensemble methods showed the best performances.

SUBMITTER: Na J

PROVIDER: S-EPMC8595677 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundThe treatment of patent ductus arteriosus (PDA) in very low birth weight (VLBW) infants remains a challenge. The ability to predict which infants will respond to indomethacin could spare some from the risks of unnecessary medications. Our objective was to determine if indicators of acid-base homeostasis could predict response to indomethacin treatment for ductal closure, and thus help guide treatment decisions.MethodsWe performed a retrospective analysis of medical records of VLBW (< 1500 g) neonates with hemodynamically significant PDA born at our institution between January 2009 and December 2012; all infants included in the study were treated with indomethacin for ductal closure within the first 2 weeks of life. We extracted data for a number of clinical variables including gestational age, birth weight, blood chemistries, surfactant use, hematocrit, and blood gas parameters. Our primary outcome measure was successful closure of PDA following the first round of indomethacin. Using variables that were significant on initial testing, we created multivariable regression models to determine the independent association of selected variables with indomethacin response.ResultsOf the 91 infants included in the study, 62 (68%) responded to the first course of indomethacin with successful ductal closure. Multivariable regression modeling revealed that both base excess and hematocrit were independently associated with indomethacin response; odds of PDA closure increased with increasing base excess (OR [odds ratio]: 1.81; 95% confidence interval [CI]: 1.36-2.60) and increasing hematocrit (OR: 1.21; 95% CI: 1.01-1.45). The optimal cutoff value for base excess was - 4.56, with a sensitivity of 96.8% (95% CI: 89-100) and specificity of 79.3% (95% CI: 60-92); optimal cutoff value for hematocrit was 40, with 69.4% sensitivity (95% CI: 56-80) and 65.5% specificity (95% CI: 46-82).ConclusionsBase excess and hematocrit may be independent predictors of indomethacin response in VLBW infants with PDA. Low-cost and readily accessible, acid-base indicators such as base excess could help guide treatment decisions.

Project description:Background: Amino acids are increasingly recognized as bioactive molecules in numerous physiological and pathophysiological pathways. The non-essential amino acid glutamate is vasoactive in the rat ductus arteriosus (DA) and a decrease in its levels within the 1st days of life has been associated with the presence of patent DA (PDA) in extremely preterm infants. However, these findings have not been confirmed in other studies. Objective: To investigate the possible association between amino acid concentrations in the 1st day of life and the presence of PDA in a cohort of 121 newborns with gestational age (GA) below 30 weeks and birth weight (BW) below 1,500 g. Methods: Plasma samples were collected 6-12 h after birth and amino acid concentrations were determined by tandem mass spectrometry. Besides PDA, we analyzed the potential association of amino acid concentrations with infant sex, small for GA (SGA, defined as BW < third percentile), antenatal corticosteroids, chorioamnionitis, and preeclampsia. Group differences were analyzed by ANOVA adjusted for GA and BW. A Bonferroni significance threshold of P < 0.0024 was used to correct for multiple testing. Results: PDA was found in 48 of the 121 infants examined. We observed higher mean levels of glutamate in infants with PDA (147.0 μmol/L, SD 84.0) as compared with those without (106.7 μmol/L, SD 49.1, P = 0.0006). None of the other amino acid concentrations in the PDA group reached the level of statistical significance that was pre-set to correct for multiple comparisons. Glutamate levels were not significantly affected by infant sex, being SGA, or by exposure to antenatal corticosteroids, clinical chorioamnionitis, or preeclampsia. Conclusion: Our study not only does not confirm the previous findings of low glutamate levels in preterm infants with PDA, but we have even found elevated glutamate concentrations associated with PDA. Nevertheless, despite the high statistical significance, the difference in glutamate levels may lack clinical significance or may be an epiphenomenon associated with the particular clinical condition of infants with PDA.

Project description:ObjectivesTo standardize the diagnosis of patent ductus arteriosus (PDA) and report its association with adverse neonatal outcomes in very low birth weight infants (VLBW, birth weight < 1500 g).Study designA multicenter prospective observational study was conducted in Emilia Romagna from March 2018 to October 2019. The association between ultrasound grading of PDA and adverse neonatal outcomes was evaluated after correction for gestational age. A diagnosis of hemodynamically significant PDA (hsPDA) was established when the PDA diameter was ≥ 1.6 mm at the pulmonary end with growing or pulsatile flow pattern, and at least 2 of 3 indexes of pulmonary overcirculation and/or systemic hypoperfusion were present.Results218 VLBW infants were included. Among infants treated for PDA closure in the first postnatal week, up to 40% did not have hsPDA on ultrasound, but experienced clinical worsening. The risk of death was 15 times higher among neonates with non-hemodynamically significant PDA (non-hsPDA) compared to neonates with no PDA. In contrast, the risk of death was similar between neonates with hsPDA and neonates with no PDA. The occurrence of BPD was 6-fold higher among neonates with hsPDA, with no apparent beneficial role of early treatment for PDA closure. The risk of IVH (grade ≥ 3) and ROP (grade ≥ 3) increased by 8.7-fold and 18-fold, respectively, when both systemic hypoperfusion and pulmonary overcirculation were present in hsPDA.ConclusionsThe increased risk of mortality in neonates with non-hsPDA underscores the potential inadequacy of criteria for defining hsPDA within the first 3 postnatal days (as they may be adversely affected by other clinically severe factors, i.e. persistent pulmonary hypertension and mechanical ventilation). Parameters such as length, diameter, and morphology may serve as more suitable ultrasound indicators during this period, to be combined with clinical data for individualized management. Additionally, BPD, IVH (grade ≥ 3) and ROP (grade ≥ 3) are associated with hsPDA. The existence of an optimal timeframe for closing PDA to minimize these adverse neonatal outcomes remains uncertain.

Project description:BACKGROUND:Patent ductus arteriosus (PDA) is a particularly common problem in preterm infants. Although surgical ligation is rarely performed in many contemporary neonatal intensive care units, it remains a necessary treatment option for preterm infants with a large hemodynamically significant PDA under strict clinical criteria, and it can reduce mortality in preterm infants. However, the optimal timing of surgical ligation is still controversial. We conducted this systematic review and meta-analysis to compare the mortality and morbidity of early and late surgical ligation of PDA in preterm or very-low-birth-weight (VLBW) infants. METHODS:This review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42019133686). We searched the databases of PubMed, Embase, the Cochrane Central Register of Controlled Trials, and the World Health Organization International Clinical Trials Registry Platform up to May 2019. RESULTS:This review included 6 retrospective studies involving 397 premature or VLBW infants with PDA. Pooled analysis showed that compared with the late ligation group, the early ligation group had a lower fraction of inspired oxygen (FiO2) at 24 hours postoperatively (mean difference [MD] -6.34, 95% confidence interval [CI] -9.45 to -3.22), fewer intubation days (MD -19.69, 95% CI -29.31 to -10.07), earlier date of full oral feeding (MD -22.98, 95% CI -28.63 to -17.34) and heavier body weight at 36 weeks of conceptional age (MD 232.08, 95% CI 57.28 to 406.88). No significant difference in mortality or other complications was found between the early and late groups. CONCLUSION:Our meta-analysis implies that compared with late surgical ligation, early ligation might have a better respiratory outcome and nutritional status for PDA in preterm or VLBW infants. There was no difference in mortality or postoperative complications between early and late ligation. A randomized prospective clinical trial with a possible large sample size is urgently needed to reinvestigate this conclusion. PROSPERO REGISTRATION NUMBER:CRD42019133686.

Dataset Information

Artificial intelligence model comparison for risk factor analysis of patent ductus arteriosus in nationwide very low birth weight infants cohort

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets