This post-hoc study, utilizing the PLASE database with a sufficient sample size (N = 692) of early preterm infants in Japan, provided a prediction model for future need for PDA surgery using simple clinical and echocardiographic variables at 3 days of age, and cross-validation was performed. The model demonstrated high reliability, with ROC-AUC and ICC values of >0.8 and >0.9, respectively, and a low Brier score of 0.075 in the validation cohort. These results suggest that this model is currently the best available for predicting the future need for PDA surgery in early preterm infants at 3 days of age.

Previously PDA severity scores have been reported, but these models were based on limited sample sizes and lacked cross-validation. For instance, El-Khuffash et al. provided a PDA severity score using echocardiography data (PDAd, maximum PDA velocity, LV output, and late diastolic peak velocity of the lateral mitral annulus (a’) using tissue Doppler imaging) at 2 days of age.5 They reported the usefulness of the model, including the use of variables such as gestational age to predict chronic lung disease or death before discharge.5 Fink et al. and El-Khuffash et al. used the same cohort5 and compared the Shaare Zedek medical center (SZMC) score with the El-Khuffash model.5,6 The SZMC score uses the PDAd, LA/Ao, abdominal aorta, and PDA flow patterns,6 and is believed to be a hemodynamic index showing how PDA is symptomatic like our PDA PLASE score. The SZMC score was well correlated with the El-Khuffash score,5,6 and the El-Khuffash score5 appeared to have considerable overlap in the three groups classified by severity according to the SZMC score.6 El-Khuffash et al.5 and SZMC6 used the same cohort consisting of 141 preterm infants with a mean gestational age of 26.8 ± 1.4 weeks, including only 10 infants who underwent PDA surgery. McNamara and Sehgal model, which incorporates PDAd, transductal flow, LA/Ao, the ratio of peak mitral flow velocities during early and late diastole or isovolumetric relaxation time, and decreased or absent diastolic flow in superior mesenteric artery, middle cerebral artery, or renal artery7 provides information on PDA significance and has been used in multiple studies.8 However, we could not compare the PLASE score with the scores developed by El-Khuffash,5 SZMC,6 or McNamara and Sehgal7 because our model uses only three simple echocardiographic indices (PDAd, LPAedv, and LA/Ao). Our study did not capture tissue Doppler, LV output, transmitral flow, superior mesenteric/ middle cerebral/ renal arterial flow, or abdominal aortic flow which are necessary for calculating those other scores.5,6,7 To our knowledge, no previous studies including these three studies,5,6,7 have developed a PDA severity score using echocardiography based on a sufficiently large sample of infants and validated through appropriate cross-validation.

The previous studies2,3 have not been able to quantitatively predict future need for PDA surgery by a model that integrates various clinical and echocardiographic information. This study advanced to provide a simple but reliable prediction model to predict future need for PDA surgery using GA and three easily measurable echocardiographic variables at 3 days of age with high discrimination and calibration ability. Such prediction of the future need of PDA surgical closure will help neonatologist’s decision-making process on the use of cyclooxygenase inhibitors, respiratory management, or blood transfusion.3 This model easily yields the probability using a spreadsheet. To help understand the nature of this prediction model, we determined the relationship between the PLASE score and PDAd in the four representative settings with LA/Ao ratio of 1.2 and 2.0, and with LPAedv of 10 and 20 cm/s in infants with gestational ages of 23 and 25 weeks (Fig. 3). The results indicated that the PDAd at 3 days of age plays a major role to predict future need for PDA surgery. This aligns with a previous report that gestational age and PDAd were the most important determinants of PDA treatment.9 Although PDAd may be expected to be the most important determinant to predict the need for PDA surgery in symptomatic infants with large PDAd, it is not the case. PDA surgical closure was performed at a mean age of 20–21 days in the included patients, and they were treated by various management strategies after 3 days of age. The PLASE score can reasonably provide information on the PDA status at 3 days of age with respect to the probability of future need for PDA surgical closure in the Japanese real-world practice and may contribute to optimize neonatal management.

The relationship between the PLASE score (probability of future surgical PDA closure, vertical axis) and PDA diameter (horizontal axis) in the four representative settings with LA/Ao ratio of 1.2 and 2.0, and with LPAedv of 10 and 20 cm/s in infants with gestational ages of 23 and 25 weeks. PLASE, the Patent ductus arteriosus and Left Atrial Size Evaluation study in preterm infants; PDA, patent ductus arteriosus; LA/Ao, left atrial to aortic diameter; LPAedv, left pulmonary artery end-diastolic velocity.

This study has a potential limitation inherent in the heterogeneity of the 34 institutes and multiple echocardiographers ( > 200).2 Nonetheless, the findings reflect the real-world practice in Japan. All echocardiographers, mostly neonatologists, had been evaluated and met the predefined quality control criteria for accurate echocardiographic measurements before data acquisition.4 Environmental differences for PDA surgical closure (five institutes [15%] needed to transfer patients, and six [18%] needed to invite surgeons for PDA surgical closure) may have affected the management and decision-making for the surgical closure of PDA.2 The indication criteria for PDA surgical closure were not standardized; instead, the decision for PDA surgical closure was clinically based on careful assessments of respiratory, renal, intestinal, and nutritional conditions as well as the overall tendency for the general condition to deteriorate.2 The current analysis set PDA surgical closure as the primary outcome and did not assess whether the PLASE score is useful for predicting death, severe bronchopulmonary dysplasia, or a composite of these two. These issues need to be investigated in future studies. Finally, international debates and differences in PDA therapies10,11,12 and surgical closure exist. Although the rate of PDA surgery of approximately 11% in the present cohort was comparable to that in other countries,13 the validity of the PLASE score outside Japan remains to be elucidated.