Adapted ECMO criteria for newborns with persistent pulmonary hypertension after inhaled nitric oxide and/or high-frequency oscillatory ventilation

作者：van Berkel Saskia; Binkhorst Mathijs; van Heijst Arno F J; Wijnen Marc H W A; Liem Kian D^*

来源：Intensive Care Medicine, 2013, 39(6): 1113-1120.

DOI：10.1007/s00134-013-2907-y

• High-frequency oscillation ventilation
• Nitric oxide
• Extracorporeal membrane oxygenation
• Persistent pulmonary hypertension of newborn
• Selection criteria
• Newborn

Early prediction of extracorporeal membrane oxygenation (ECMO) requirement in term newborns with persistent pulmonary hypertension (PPHN), partially responding to inhaled nitric oxide (iNO) and/or high-frequency oscillatory ventilation (HFOV), based on oxygenation parameters. %26lt;br%26gt;This was a retrospective cohort study in 53 partial responders from among 133 term newborns with PPHN born between 2002 and 2007. Alveolar-to-arterial oxygen gradient (AaDO(2)) values were determined in these 53 partial responders during the initial 72 h of iNO and/or HFOV treatment and compared between newborns who ultimately did (n = 11) and did not (n = 42) need ECMO. %26lt;br%26gt;Over 72 h, partial responders not requiring ECMO showed a more profound AaDO(2) decrease than those who needed ECMO (median decline 242.5 mmHg, IQR 144 to 353 mmHg, vs. 35 mmHg, IQR -15 to 123 mmHg; p = 0.0007). A decline of %26lt; 123 mmHg over 72 h predicted the need for ECMO (sensitivity 82 %, specificity 79 %). At 72 h, AaDO(2) was significantly lower in partial responders without the need for ECMO than in those who did need ECMO (median 369 mmHg, IQR 258 to 478 mmHg, vs. 570 mmHg IQR 455 to 590 mmHg; p = 0.0008). An AaDO(2) %26gt; 561 mmHg at 72 h predicted the need for ECMO (sensitivity 64 %, specificity 95 %, positive predictive value 78 %). %26lt;br%26gt;In term newborns with PPHN partially responding to iNO and/or HFOV, oxygenation-based prediction of the need for ECMO appears to be possible after 72 h. ECMO centers are encouraged to develop their own prediction model in order to prevent both lung damage and unnecessary ECMO runs.

• 出版日期2013-6