The most appropriate next step in management is to perform deep sampling of the lower respiratory tract. Ventilator-associated pneumonia (VAP) is a serious, preventable complication of mechanical ventilation. It is defined as pneumonia with onset at least 48 hours after endotracheal intubation. VAP is difficult to diagnose, but potential clues to its presence include temperature greater than 38.0 °C (100.4 °F), leukocytosis or leukopenia, increased purulent secretion, new or progressive pulmonary infiltrates, and worsening ventilation parameters, particularly after a period of improvement. Patients suspected of having VAP should undergo lower respiratory tract sampling, followed by microscopic analysis and culture of the specimen. Nonbronchoscopic sampling methods are simple suctioning of the endotracheal tube and mini–bronchoalveolar lavage (BAL), which involves use of a telescoping catheter (instead of a bronchoscope) to instill and aspirate physiologic saline for microbiologic analysis. Bronchoscopic methods are standard BAL and protected specimen brush. Deep sampling methods may allow for narrower antibiotic choices and more rapid de-escalation of antibiotics.

While waiting for the microscopic and culture results from the lower respiratory tract sampling, initiating empiric antibiotics is a reasonable option. However, this patient has been in the hospital for 7 days and is at risk for multidrug-resistant organisms, including Pseudomonas species. In this situation, an antipseudomonal cephalosporin or carbapenem would be appropriate. Ceftriaxone and azithromycin would be good coverage for community-acquired pneumonia but are inadequate for this patient.

Chest physiotherapy is useful for assisting with the removal of secretions in patients with COPD, cystic fibrosis, and ciliary dyskinesia. In addition to standard chest percussion and drainage, a large number of mechanical devices are now available that help encourage mobilization of secretions with variable results. Although chest physiotherapy will assist with removal of secretions, it will not provide additional information to clarify the diagnosis in this patient. The most appropriate next step in management is a microbiologic diagnosis to guide further therapy.

Switching from oseltamivir to zanamivir is unlikely to help this patient, who demonstrated an initial clinical response to influenza therapy and then declined. VAP is the most likely diagnosis, not influenza resistance to oseltamivir.