Supplementary Material for: Comparison of Various Pharmacologic Agents in the Management of Hemodynamically Significant Patent Ductus Arteriosus in Preterm: A Network Meta-Analysis and Risk-Benefit Analysis

<b><i>Introduction:</i></b> Various pharmacological treatments are available for preterm infants with patent ductus arteriosus (PDA), but their risks and benefits are controversial. This study aimed to identify the best treatment for PDA using network meta-analysis (NMA) and risk-benefit assessment (RBA). <b><i>Methods:</i></b> Relevant randomized controlled trials (RCTs) were identified from MEDLINE, Scopus, and the Cochrane Library. RCTs were eligible if they were studied for preterm or low birth weight infants with presymptomatic PDA and hemodynamically significant PDA (hsPDA). The outcomes were PDA closure for a benefit and the composite risk outcome of adverse effects (AEs) for risk. An NMA was used to estimate the treatment effects of benefit and risk. The RBA helped to incorporate the risk and benefits of multiple treatments. Then, an incremental risk-benefit ratio was calculated by dividing the incremental risk by benefit using data from NMA, and they were jointly simulated using Monte Carlo methods. Finally, net clinical benefit (NCB) probability curves were constructed at varying acceptability thresholds. <b><i>Results:</i></b> Seventy RCTs with hsPDA were eligible considering 13 different interventions, but data on presymptomatic PDA were not enough for pooling. The clustered ranking plot from NMA indicated that 3 interventions (i.e., high-dose oral ibuprofen, standard-dose oral acetaminophen, and standard-dose oral ibuprofen) yielded high PDA closure and low AE. These three treatments and additional commonly used indomethacin were considered in the RBA. Given an acceptable threshold of 25% or having one AE out of four PDA closures, high-dose oral ibuprofen had a 36% chance of having the highest NCB, followed by standard-dose oral acetaminophen (27%), and oral ibuprofen (23.7%). Subgroup analysis indicated that the chances of having the highest NCB of GA ≥28 weeks were similar to that of all available studies. The best for GA &lt;28 weeks, no data for high-dose oral ibuprofen, was standard-dose oral acetaminophen, followed by standard-dose oral ibuprofen. <b><i>Conclusions:</i></b> Trade-off RBA indicated that high-dose oral ibuprofen might be the best treatment for preterm, GA ≥28 weeks, with hsPDA followed by the standard-dose oral acetaminophen and ibuprofen. Preferably, optimal high doses, postnatal age to start treatment, and long-term outcomes are needed to study in the future.

<b><i>引言：</i></b> 针对合并动脉导管未闭（patent ductus arteriosus, PDA）的早产儿，目前已有多种药物治疗方案，但其风险与获益仍存在争议。本研究旨在通过网络Meta分析（network meta-analysis, NMA）与风险获益评估（risk-benefit assessment, RBA），明确动脉导管未闭的最优治疗方案。
<b><i>方法：</i></b> 研究从MEDLINE数据库、Scopus数据库及Cochrane图书馆中检索相关随机对照试验（randomized controlled trials, RCTs）。纳入标准为针对早产儿或低出生体重儿的症状前动脉导管未闭及血流动力学显著性动脉导管未闭（hemodynamically significant PDA, hsPDA）的随机对照试验。本研究的获益结局为动脉导管闭合，风险结局为不良事件（adverse effects, AEs）复合风险结局。采用网络Meta分析估算获益与风险的治疗效应，通过风险获益评估整合多种治疗方案的风险与获益。随后基于网络Meta分析的数据，以增量风险除以增量获益计算增量风险获益比，并通过蒙特卡洛方法进行联合模拟。最终构建不同可接受阈值下的净临床获益（net clinical benefit, NCB）概率曲线。
<b><i>结果：</i></b> 最终纳入13种不同干预措施的70项针对血流动力学显著性动脉导管未闭的随机对照试验，但症状前动脉导管未闭的相关数据不足以进行合并分析。网络Meta分析的聚类排序图显示，3种干预措施——大剂量口服布洛芬、标准剂量口服对乙酰氨基酚及标准剂量口服布洛芬——可实现较高的动脉导管闭合率且不良事件发生率较低。将上述3种治疗方案及临床常用的吲哚美辛纳入风险获益评估。当可接受阈值为25%，或每闭合4例动脉导管未闭仅发生1例不良事件时，大剂量口服布洛芬的净临床获益最高概率为36%，其次为标准剂量口服对乙酰氨基酚（27%）与标准剂量口服布洛芬（23.7%）。亚组分析显示，孕龄（gestational age, GA）≥28周亚组的净临床获益最高概率与全部纳入研究的结果相似；而针对孕龄<28周的患儿（尚无大剂量口服布洛芬的相关数据），最优治疗方案为标准剂量口服对乙酰氨基酚，其次为标准剂量口服布洛芬。
<b><i>结论：</i></b> 权衡式风险获益评估结果显示，对于孕龄≥28周的血流动力学显著性动脉导管未闭早产儿，大剂量口服布洛芬可能为最优治疗方案，其次为标准剂量口服对乙酰氨基酚与标准剂量口服布洛芬。未来仍需开展相关研究，明确最优给药剂量、开始治疗的产后年龄及远期临床结局。