There are two main challenges in the interventional occlusion of large patent ductus arteriosus (PDA): one is the insufficient size of the devices for the large PDA, and another is pulmonary hypertension (PH) after closure. Some manufacturers have produced large-size occluders, which can significantly meet clinical needs. Some clinical centers have adopted a trial occlusion strategy for the occlusion of large PDA. However, the indications of when to release the devices and the factor that affect persistent postoperative PH still need to be clarified, and the previous studies have yet to elucidate fully.

One hundred thirty-nine consecutive patients were retrospectively analyzed in this study. All of them were diagnosed with a PDA diameter≥10 mm or the ratio of PDA and aortic > 0.5, and all underwent trial occlusion before deciding whether to release the device. We collected the follow-up clinical data to explore the factors affecting follow-up PH.

The closure of large PDA with severe PH is feasible with 96.4% success (5 out of 139 patients failed). Five procedures were deemed to have failed because their mPAP did not decrease, and the patients experienced uncomfortable symptoms after the trial occlusion. The average age of the successful 134 patients was 35.04 ± 10.23 years, and the mean pulmonary artery pressure (mPAP) was all > 50 mmHg. The average follow-up period was 17.56 months. The average size of the occluder (pulmonary end) was almost twice the PDA diameter (22.23 ± 4.77 mm vs. 11.66 ± 2.16 mm). Left ventricular end-diastolic dimension (LVEDD), mPAP, and left ventricular ejection fraction (LVEF) were significantly reduced after occlusion and recovered during follow-up. 42 of the 78 patients with total pulmonary resistance> 4 Wood Units still had PH during the follow-up. Meanwhile, ten of the 42 follow-up PH patients were accompanied by heart failure, and four were hospitalized again because of PH. All the available essential factors, including the mPAP, ratio of systolic pulmonary pressure and systemic pressure, and LVEDD, were analyzed in the logistic regression. The analysis revealed that the postoperative mPAP was an independent risk factor in the multivariable logistic regression model (OR = 1.069, 95% CI: 1.003 to 1.140, p = 0.040) with a receiver operating characteristic curve (ROC) cut-off value of 35.5 mmHg (area under the curve [AUC] 74.4%; p < 0.001). The baseline LVEDD was the only detected risk factor (OR = 0.921; 95% CI: 0.869 to 0.977; p = 0.006) if the regression model only included the preoperative factors, and the cut-off value was 56.5 mm with an AUC of 67.6%. The five failed patients had a higher postoperative mPAP, a lower mPAP decrease, a lower Qp/Qs, a smaller LV diameter, and a higher LVEF. The logistic regression model found that the mPAP after the trial occlusion was the only risk factor for closure failure (OR = 1.115; 95% CI: 1.020 to 1.217; p = 0.016). The cut-off value of the mPAP was 49.5 mmHg, and the AUC was 88.5% (p = 0.004).

mPAP over 35.5 mmHg after closure predicts follow-up PH and clinical outcomes that are likely worse. These patients should be treated with targeted therapy. As for the patients with mPAP less than 35.5 mmHg, it is not necessary to conduct the targeted medical therapy, which could significantly reduce the cost of the patients in developing countries and low-income populations. However, RHC still needs to be performed optimally with pulmonary vascular resistance (PVR) assessment several months after PDA closure. Patients with smaller LV before closure(less than 56.5 mm LVEDD) should be observed more carefully. Unfortunately, the importance of PVR assessment before the closure and evaluation of an upper value of PVR safe for PDA closure cannot be derived from this study.