Dear Editor,

Children with acute respiratory failure may benefit from the use of high-frequency oscillation ventilation (HFOV) [1]. When HFOV fails to stabilise respiratory function, many children are transported to specialised centres for extracorporeal membrane oxygenation (ECMO) [2–5]. HFOV during patient transport represents not only a logistical challenge (see Electronic Supplementary Material) to the pediatric intensive care unit transport team (PICU-TT), but also a severe physiological challenge to the child receiving it.

For more than 20 years the PICU-TT at the Hôpital Robert Debré (Paris, France) has been the only French PICU-TT to use HFOV during the transport of children for subsequent ECMO support. The transport of patients on ECMO started in the region at the end of 2014. Here, we report our experience with the use of HFOV during the transport of pediatric patients for ECMO support and hypothesize that, when using a proportional hazard survival model, pre-transport physical or transport parameters were not associated with mortality.

Fifty-nine children with respiratory failure, deteriorating HFOV ventilation and/or haemodynamics were provided with HFOV during transport for ECMO support during the study period (January 1994–December 2014). Approval for the retrospective study was given by the Hospital Institutional Review Board. All transfers were by road and occurred in the Paris Region, with 57 children transported to Hôpital Trousseau and two children transported to Hôpital Robert Debré. Patient selection criteria were neither standardised nor defined by the referral or ECMO centres and were independent of the PICU-TT.

The children ranged in age and weight from 0 days to 9 years and from 2.3 to 25 kg, respectively. The principal diagnoses of the transported children were diaphragmatic hernia (17 children), acute respiratory distress syndrome (15), meconium inhalation (11), pneumothoraces (5), sepsis (5) and other diagnoses (6). Twenty-seven children were transported using a SensorMedics oscillatory ventilator (SensorMedics Corp., San Diego, CA) and 32 using a SLE ventilator (SLE Ltd., South Croydon, UK). The transport distances ranged from 5 to 242 km.

Thirty-one children (53 %) survived to discharge, and five children (8.5 %) died during transport, none as a result of equipment malfunction. The median time to death for the 59 children was 14 (range 1–26) days. One child died shortly after arrival and was not connected to ECMO. Of the remaining 53 children, 36 (68 %) were connected to the ECMO circuit. The physical and transport parameters of these children on arrival of the PICU-TT are presented in Table 1. The blood oxygen saturation (SpO2) of the five children who died during transport ranged from 30 to 71 % prior to transport.

Table 1 Comparison of the physical and transport parameters of survivors and children who died from the moment of arrival of the PICU transport team to the end of the primary ICU hospitalisation
Full size table

Boedy et al. [5] have suggested that the ECMO mortality data reported to date have been influenced by survival bias, with deaths before and during transport being ‘hidden’. We also observed a significant early mortality in our study. In contrast, in their study of 112 children transported on ECMO, Clement et al. [2] described two deaths prior to arrival of the PICU-TT and one child ‘beyond salvage’ upon arrival of the team; however, they did not describe any deaths during transport. In both our study and that of Clement et al. [2] the criteria for transfer were not predefined and, consequently, a comparison of death rates is difficult. An important advantage of the use of HFOV during the transport of children for ECMO is that not all the children in our study subsequently required ECMO.

In conclusion, the use of HFOV during the transport of children for ECMO in children is feasible. HFOV during transport may reduce the overall numbers of children being connected to ECMO, but early mortality in severely hypoxaemic children may be higher. Until inclusive risk-adjusted data are available, the use of HFOV during transport is an option, especially where a mobile ECMO team is unavailable.