Extracorporeal Membrane Oxygenation for Severe Refractory Respiratory Failure Secondary to 2009 H1N1 Influenza A

Discussion

Throughout 2009, H1N1 influenza infection led to a considerable number of critically ill patients, as many as 30% of whom were refractory to maximal conventional respiratory therapies.^1,4,6–9,12 In our 7 patients with refractory respiratory failure secondary to H1N1, ECMO was successful, and 5 of the 7 patients survived. This finding consistent with published survival rates for ECMO worldwide during the H1N1 pandemic and suggest that ECMO is a viable therapy in both pediatric and adult patients with respiratory failure and refractory ARDS.^12–14

Consistent with prior reports that the H1N1 influenza attack rate was highest in adolescents and young adults,^2,18 6 of our patients were between 15 and 33 years old. Additionally, all but one of those 6 patients had an important underlying comorbidity, including the one patient under the age of 15, who had hypoplastic left-heart syndrome and had undergone the Fontan operation. Given the substantial risk factors in these patients, the fact that their ARDS was progressive and unresponsive to standard ICU therapies was not surprising.

The pre-ECMO therapies in our patients included a lung-protective ventilation strategy, initially with primarily pressure controlled synchronized intermittent mandatory ventilation and a target tidal volume of 5–7 mL/kg. Because of disease progression, all 7 patients were transitioned to high-frequency oscillatory ventilation and then to ECMO, following a brief bridge with inhaled nitric oxide (Table 5). In addition, we used intravenous steroids in 5 of the patients, and oral oseltamivir in all 7 patients. However, patients 2 and 4 did not begin oral oseltamivir within 48 hours of symptom onset. Following ECMO initiation, patients 4 and 7 also received compassionate-use intravenous zanamivir.

Because of the expansion and evolution of our ECMO program during the H1N1 pandemic,¹⁹ these 7 patients were managed with 2 different ECMO systems. Patients 1 through 4 were treated with a 3 Stockert rollerhead pump system (Sorin Group, Arvada, Colorado) and a Quadrox D oxygenator (Maquet, Hirrlingen, Germany). Patients 5 through 7 were treated with a Rotaflow centrifugal pump and Quadrox D oxygenator (both from Maquet, Hirrlingen, Germany). With both systems we used a standardized protocol of continuous heparin infusion titrated based on activated clotting time. Following ECMO initiation and improvement of gas exchange, the level of mechanical ventilatory support was immediately weaned to rest the lungs and minimize ongoing secondary injury. Despite expected variations in mode and settings, all 7 patients, within 24 hours of ECMO cannulation, were on F_IO2 ≤ 0.50, peak inspiratory pressure < 29 cm H₂O, and a target tidal volume of 5–7 mL/kg (see Table 3).

While this series demonstrates successful use of ECMO for refractory respiratory failure from H1N1 influenza in several complex patients, further data are needed to determine under what circumstances ECMO is appropriate. We feel that ECMO deserves broad consideration in critically ill neonatal, pediatric, and adult patients, even if the patient has important comorbidities. The impact of comorbidities is unclear, but ECMO can unquestionably be a life-saving therapy in many circumstances. The possibility of saving the lives of these most severely affected patients with refractory respiratory failure suggests that there are few absolute contraindications to ECMO.

Prospective identification of patients who will benefit from ECMO remains quite difficult, despite decades of successful use in a wide range of clinical settings.^{6,7,12,14,15,20} In the setting of refractory respiratory failure, however, reported ECMO survival is lower when lung injury is associated with comorbidities, especially in the setting of substantial immunocompromise or multiple organ dysfunction.^6,7,21 Our series demonstrates that ECMO can be successful in the presence of considerable underlying comorbidities in pediatric and adult patients, thus potentially expanding the pool of ECMO candidates. Given this potential benefit, it is important to evaluate the patient's ECMO eligibility early, since some data suggest that early application of ECMO may lead to a shorter ECMO course and more rapid lung recovery.¹ The time needed for lung recovery on ECMO, and resource availability are important considerations in the expansion of ECMO utilization, especially during a pandemic.

Time required on ECMO may be impacted by both underlying comorbidities and pre-ECMO duration of mechanical ventilation. The mean ECMO course in this series was 432 hours, which is longer than the 280 hour mean ECMO duration for ARDS secondary to H1N1 reported by the Extracorporeal Life Support Organization.²¹ Our patients' comorbidities may have contributed to their time on ECMO, but their pre-ECMO duration of mechanical ventilation may also have been a factor.

The largest reported series of H1N1 ECMO patients had a mean of 2 days of mechanical ventilation prior to ECMO initiation.¹² In our series, 3 patients rapidly deteriorated, consistent with rapid progression of respiratory failure seen in many patients with H1N1 infection, and were cannulated for ECMO in less than 48 hours.² Despite these 3 patients with rapid deterioration, the mean duration of mechanical ventilation prior to ECMO initiation was 8.7 days (median 6 d, range 14 h to 25 d). A substantial contributor to the mean duration of mechanical ventilation prior to ECMO was the duration of ventilation in the 2 non-survivors, who were ventilated for 11 days and 25 days prior to transfer to our PICU for ECMO. This point raises the important issue of transfer criteria and timing of transfer to an ICU capable of providing ECMO support.

One non-survivor initially had improved pulmonary compliance and gas exchange but later developed a severe superinfection with Pseudomonas aeruginosa and consequent septic shock, necessitating ECMO. At the time of cannulation, she had been ventilated for 25 days, and her H1N1 viral culture and respiratory viral polymerase-chain-reaction test had become negative, which suggests that her ECMO course was not directly related to the H1N1 infection. Factors that probably contributed to her death were sepsis at the time of ECMO cannulation, active bacterial infection, and prolonged pre-ECMO ventilation duration, all of which are associated with higher ECMO mortality.^21–24

The other non-survivor in this series required ventilatory support for 11 days at a referring institution prior to ECMO. However, the possible benefit of earlier ECMO initiation in this patient is unclear, given that her underlying comorbidities included plastic bronchitis, bronchopulmonary aspergillosis, and a history of left lower lobectomy. This patient's limited pulmonary reserve and fragile baseline respiratory state surely contributed to the severity of her H1N1 infection, the need for ECMO, and death, despite 890 hours on ECMO.

Of note, one patient did survive despite a pre-ECMO ventilation course of 14 days, but for his entire course in our PICU he was treated with a lung-protective strategy that included predominantly high-frequency oscillatory ventilation. Many experts consider a ventilation duration greater than approximately one week a contraindication to ECMO because of the potential for irreversible lung injury.^22–24 However, the “acceptable” duration of mechanical ventilation prior to ECMO initiation is unclear in the current era of lung-protective ventilation. While it is certainly possible that the prolonged pre-ECMO ventilation impacted mortality and duration of ECMO, our experience in this series suggests that patients should not be uniformly excluded from consideration for ECMO based solely on the duration of pre-ECMO ventilation.

Another important consideration in the determination of ECMO candidacy is underlying comorbidity. Consistent with prior reports, our cohort had a high incidence of comorbidities (see Table 2). It is clear from this series that ECMO can be successful in H1N1 ARDS patients with comorbid conditions, including congenital heart disease, renal insufficiency, connective tissue disease, and obesity, which previously would have precluded ECMO for refractory respiratory failure.

The success in this series would argue for a broader consideration of how comorbidities should impact ECMO candidacy in refractory respiratory failure. We recognize that these patients may require longer ECMO courses and that recovery of lung function may depend on the pre-ECMO duration of ventilation. While ECMO certainly involves substantial resources and risks, clinical series in which ECMO successes continue to be reported in previously unthinkable circumstances further complicate the ECMO decision making process.