Co-detection of Pandemic (H1N1) 2009 Virus and Other Respiratory Pathogens

From May through October 2009, a total of 10,624 clinical samples from 23 US states were screened for multiple respiratory pathogen gene targets. Of 3,110 (29.3%) samples positive for pandemic (H1N1) 2009 virus, 28% contained >1 other pathogen, most commonly Staphylococcus aureus (14.7%), Streptococcus pneumoniae (10.2%), and Haemophilus influenzae (3.5%).

Of the respiratory specimens shipped by overnight mail from the 23 states, >95% were nasopharyngeal swabs in transfer buffer. High-throughput nucleic acid extraction was performed automatically by using KingFisher 96 instrumentation (Thermo Scientifi c, Hudson, NH, USA) and MagnetX chemistry (Scigenix, Marietta, GA, USA) according to manufacturers' specifi cations. Multiplex PCR amplifi cation and Luminex (Austin, TX, USA) liquid suspension detection methods were based on internally validated protocols. Reactions were amplifi ed by using ABI 9700 thermocyclers (Applied BioSystems, Singapore), and the resulting PCR products were detected by using the LiquiChip 200 Workstation (Luminex) according to previously described protocols (3,6

Conclusions
The main fi nding of this large-scale clinical study was the co-detection of multiple pathogens with the pandemic infl uenza virus strain. In 44% of samples, no pathogens were detected, which may represent infection with common pathogens not detected by the assay. For example, bocavirus and all coronavirus groups not detected by the assay account for ≈12% and 5%-10%, respectively (8,9), of respiratory infections. An expanded test menu may improve the detection rate for such pathogens.
This study raises 2 questions. First, does co-detection equal co-infection? Second, and more practical, does codetection change the clinical outcome? We chose the word co-detection rather than co-infection or co-colonization because co-infection means all identifi ed microorganisms contributed to the pathogenic effect, and co-colonization may not indicate the causative agent. Co-detection indicates that >1 other pathogen was detected in a sample. The differences among the defi nitions have etiologic meaning, but the data presented here cannot be used directly to address etiology.
Most samples in this study were nasal swabs rather than upper or lower respiratory tract samples. Nasal swab samples have greater potential for contamination with normal fl ora, particularly S. aureus. No data on asymptomatic carriers were available because these persons rarely seek healthcare. However, these fi ndings raise questions about the effectiveness of the single-agent etiology approach toward infectious diseases. Pandemic (H1N1) 2009 virus and multiple other pathogens are often detected during autopsy (1,2), indicating that co-infection may play a major role in the disease process. In addition, detection of multiple pathogens is associated with increased critical illness in children (7).
The Centers for Disease Control and Prevention identifi ed "the need for early recognition of bacterial pneumonia in persons with infl uenza" (2). However, no suggestions were provided for meeting this need. Furthermore, the Centers "underscore the importance of managing patients with infl uenza who also might have bacterial pneumonia with both empiric antibacterial therapy and antiviral medications" (2) without identifying measures that would make this task tangible. Current practices of clinical diagnosis based on signs and symptoms inherently lack this type of information. The true value of a multiplex molecular method of screening for infectious respiratory agents depends on the clinical relevance. Among the samples with >1 positive results, 53% had positive results for viral pathogens without co-detection of bacterial pathogens. For these patients, prescription of antimicrobial drugs on the basis of clinical fi ndings alone could serve to spread drug resistance through selective pressure on normal fl ora. Furthermore, limited secondary treatment resources, such as oseltamivir administration during a pandemic, could be prioritized on the basis of screening results. Of the 10,624 samples studied, 70.7% were negative for the pandemic (H1N1) 2009 virus strain.
Our fi ndings suggest that multiplex screening for respiratory pathogens is useful for providing rapid surveillance information to inform physicians who would otherwise base decisions on clinical signs and symptoms alone. Electronic reporting of empirical laboratory respiratory pathogen detection provided by a Clinical Laboratory Improvement Amendments-approved laboratory can greatly enhance surveillance data collection (10). Because most states have the authority to collect data of public health relevance (10), the screening service provided by the Diatherix Laboratories could facilitate reporting of notifi able diseases.