Fatal Respiratory Events Caused by Zanamivir Nebulization

TO THE EDITOR—A 25-year-old pregnant woman (26 weeks gestation) was referred from a private hospital in August 2009 because of severe hypoxia secondary to influenza A (H1N1) pneumonia proven by reverse-transcription polymerase chain reaction test. Her symptoms commenced with one week of fever, followed by intense cough and dyspnea requiring mechanical ventilation. Chest radiograph showed bilateral alveolar opacification involving three-fourths of the entire lung fields. Oral oseltamivir, nebulized zanamivir, and intravenous dexamethasone were initiated promptly. The patient was ventilated by Nellcor-Puritan-Bennett-840 ventilator in pressure-controlled mode. At the best positive end–expiratory pressure (PEEP) of 14 cm H2O, the plateau pressure was 27 cm H2O and the arterial oxygen partial pressure–to–fraction of inspired oxygen concentration ratio (PaO2/FiO2) was 64. On the second day, we observed inexplicably low exhaled tidal volumes. Replacement of the ventilator, under the same settings, immediately recovered an additional 100 ml of exhaled tidal volume. The patient was in stable condition afterward with a pulse oxygen saturation (SpO2) of 88%–92% until midnight of the fourth hospital day, when the on-call physician was notified of a sudden ventilator malfunction. The nurse in charge stated that the incident happened while nebulizing zanamivir and was preceded by a few alarms that needed machine reset. While the patient was waiting for a new (third) ventilator, her SpO2 dropped to 65%. This SpO2, however, rose slowly and reached 90% in 24 hours. At this circumstance, blockade in the expiratory filter was suspected. To prevent damage to the reusable filter, we placed a disposable filter (CareStar, DrägerMedical) at the point proximal to the main filter. At midnight of the fifth hospital day, the problem recurred. This time, rapid removal of the disposable filter led to immediate resumption of ventilator function. Examination of the removed filter demonstrated inside-blockade caused by sticky material. Following this second episode of ventilator malfunction, patient became severely hypoxemic (SpO2, 71%–78%) and developed bilateral pneumothoraces. Despite all aggressive measures, her oxygenation failed to improve and the patient expired on the eighth hospital day. Diagnosis and treatment of influenza A (H1N1) pneumonia in this patient was late, which resulted in severe acute respiratory distress syndrome. Because sedatives and muscle relaxants routinely used in managing critical patients might cause impaired gastrointestinal function, the absorption of oral oseltamivir becomes unpredictable. Alternative administration of inhaled zanamivir, although safe and effective [1, 2], is also a practical problem in intubated patients. Endotracheal nebulization of zanamivir at double doses (20 mg in saline) is therefore widely practiced in Thailand during this influenza pandemic. We suspect that this incident of filter blockade was caused by the 20 mg lactose in each blister of zanamivir [3]. A simulation was performed by connecting the ventilator to a test lung and a pressure manometer. From the sixth dose of zanamivir nebulization, we observed significant retardation of expiratory flow with 1.5–2 cm H2O elevation of PEEP and airway pressure. Severe ventilator occlusion occurred transiently during the eighth dose and became persistent at the ninth dose of nebulization—during which the ventilator automatically opened its safety valve and switched to a low respiratory rate, a low airway pressure, and a zero PEEP ventilation. Further ventilator reset failed to recover its previous function. We conclude that expiratory filter obstruction caused by nebulized zanamivir resulted in severe hypoxemia and pneumothoraces that led to fatality in this patient. The fatal respiratory events were immediately reported to GlaxoSmithKline, and steps were taken by GlaxoSmithKline and the United States Food and Drug Administration, with subsequent notification to healthcare professionals and hospital risk managers [4]. Potential conflicts of interest. All authors: no conflicts.