In many patients rewarmed primarily with CPB, ECMO later becomes necessary because of the inability to wean the patient from CPB due to intractable cardiorespiratory failure [10–13]. Furthermore, ECMO has significant advantages over standard CPB technology in emergency situations (Table 14.1). Thus, in an increasing number of hospitals, venoarterial ECMO has become the method of choice for emergency extracorporeal support in hypothermic patients [4, 5, 13]. In a retrospective study of 59 hypothermic patients using multivariate logistic regression analysis, ECMO resuscitation was associated with improved survival as compared to standard CPB resuscitation [4]. The key factor responsible for improved survival was the routine use of prolonged cardiorespiratory support for 24–48 h in the ECMO group, thus preventing early mortality from respiratory insufficiency which is responsible for 64 % of fatalities after CPB resuscitation [4].

ECMO techniques (Table 14.2) have been used with success over the whole range of underlying pathologies associated with severe accidental hypothermia, including near-drowning [13], avalanche burial [4, 10], urban hypothermia [14], and multisystem trauma [15, 16]. Survival rates in published case series vary over a wide range and depend predominantly on the underlying pathology and preexisting co-morbidities [2, 3]. Urban hypothermia and hypothermia associated with avalanche accidents consistently produce poor survival rates [3, 6, 14], whereas hypothermia after prolonged exposure to cold in healthy individuals suffering from intoxication or wilderness accidents is associated with survival rates of 70–90 % in arrested patients [17]. Venoarterial and venovenous ECMO support have been used after CPB rewarming when patients cannot be weaned from extracorporeal support in the operating room because of intractable respiratory or cardiorespiratory failure [10–13, 18] (Table 14.2). Increasingly, however, venoarterial ECMO support with femorofemoral cannulation is used as the primary therapeutic intervention [4, 5]. Percutaneous femoral cannulation techniques have been used with high success rates and further reduce invasiveness [5]. ECMO is regularly implanted outside the operating theater, and transfer of the ECMO team to an outside hospital to implant an ECMO system in patients with hypothermic cardiorespiratory arrest has been reported [13, 19]. Even initiation of ECMO resuscitation at the scene may be a therapeutic option in the near future [20].

In most centers ECMO is used only in hypothermic patients with cardiorespiratory arrest. Based on their experience with 69 profoundly hypothermic patients, Morita and coworkers suggested that ECMO may improve survival also in non-arrested patients [5].

Multiorgan failure, prolonged stay in the intensive care unit for weeks, and full neurological recovery after several months in a rehabilitation unit are regularly observed in hypothermic arrest victims after initial, successful resuscitation [10, 11, 21, 22]. Therefore, one should always be cautious not to terminate maximum therapy too early after ECMO resuscitation when the clinical course is complicated in a hypothermic patient.

Almost no scientifically valid data are available on how to manage a hypothermic patient on ECMO support. Consequently, perfusion protocols vary over a wide range among different institutions. As an example, the perfusion protocol used at Innsbruck University Hospital is shown in Table 14.3. Many institutions support patients with high flow rates in the range 2.5–3 L/m²/min also during hypothermia to compensate for a preexisting oxygen debt, although the optimal flow rate is in fact unknown. High rewarming rates are regularly used, at least until successful defibrillation of the heart. This enables early restoration of pulsatile flow and improved left ventricular unloading. On the other hand, it has been repeatedly stated that rapid rewarming may put the hypoxic, cold brain at risk for additional injury [2, 3]. Vasopressors to maintain mean arterial pressure above 50 mmHg, sodium bicarbonate to correct severe metabolic acidosis, and catecholamines to support left ventricular unloading and restore pulsatile flow are used, although their actual value has not been proven to date. Acid base management is normally done using an alpha stat regime. A systemic heparin dose lower than for standard CPB is almost always used, and the option of avoiding any systemic anticoagulation by using heparin-coated ECMO systems has been repeatedly mentioned [3, 4]. Nowadays, rewarming is normally discontinued at 32–34 °C in patients with a history of cardiorespiratory arrest and therapeutic hypothermia maintained for additional 12–24 h [23]. This approach is based on the extrapolation of data obtained during normothermic cardiorespiratory arrest. In some institutions prolonged ECMO support after initial resuscitation and rewarming is routine practice [4]. When prolonged femorofemoral ECMO support is used, “upper body hypoxemia” can occur in patients in whom recovery of myocardial function is more pronounced than is recovery of pulmonary function (Table 14.3). This phenomenon is caused by poorly oxygenated blood that is ejected by the heart into the proximal parts of the aorta, whereas distal parts of the aorta receive well-oxygenated blood from the ECMO system. Therefore, oxygenation of the patient on prolonged femoral venoarterial ECMO must always be monitored on the right hand to detect this problem early (Table 14.3). If hypoxemia in the proximal aorta does not respond to alterations in mechanical ventilation, the patient must be switched to a venovenous ECMO system. Leg ischemia is a major complication of femorofemoral ECMO repeatedly reported [24]. Incidence of leg ischemia can be decisively reduced by inserting a separate cannula for leg perfusion. As kinking and dislocation of the cannula are possible in patients with distal leg perfusion, in particular during prolonged ECMO support, additional near-infrared spectroscopy monitoring of foreleg muscle perfusion is advisable. An organized, preemptive approach and standardized treatment protocols (Table 14.3) are used at several institutions [4, 5, 25] and are a good option for improving patient care.