Skip to main content

The Value of Using Viral Diagnostic Tests in Reducing Antibiotic Prescribing in Children with Pneumonia

There is substantial evidence that the risk of serious bacterial infection is low in children with laboratory-confirmed viral infection. However, the diffuse lower respiratory tract inflammation induced by viral respiratory tract infections predisposes to bacterial super infection, making it difficult to exclude concurrent bacterial pneumonia with certainty in children with laboratory-confirmed viral infections. Viral and bacterial co-infections were detected in 23% of children with pneumonia evaluated at a tertiary-care children’s hospital.
Randomized clinical trials and prospective studies of rapid influenza testing demonstrate significant reductions in ancillary testing and antibiotic use among children evaluated in the emergency department during influenza season.
 Bonner et al enrolled 391 patients (aged 2 months to 21 years) with fever and influenza like illness. Rapid influenza tests were performed on naso pharyngeal specimens for all patients, with 52% positive for influenza. Patients were then randomized so that the treating physician was either provided or not provided with the results of influenza testing. Antibiotics were prescribed to 7.3% of patients for whom the physician was aware of a positive influenza test result, compared with 24.5% of patients for whom the physician was not aware of the results. Similar reductions were noted in the performance of chest radiography and other ancillary tests. No patient had lobar pneumonia.
 Esposito et al randomized 957 children who presented to their clinic with influenza like illness to rapid influenza testing or no testing, with 43 (8.9%) children testing positive. Antibiotics were prescribed to 32.6% of influenza-positive compared with 64.8% of influenza-negative patients; of those who were randomized to no testing, 61.8% were given antibiotics. No significant difference was noted in the performance of chest radiographs between groups.
 In a retrospective cohort study of hospitalized adults with laboratory-confirmed influenza infection, a positive rapid influenza test result was associated with 6-fold higher odds to discontinue or withhold antibiotic therapy compared with influenza-positive patients whose diagnosis was delayed because positive PCR results were not readily available.
Doan et al conducted an open-label randomized controlled trial in which children 3–36 months of age were randomized to receive a multiviral rapid diagnostic test by direct immunofluorescence assay (IFA) (n = 90) or routine care (n = 110). At least one virus was detected in 66% of patients randomized to viral testing. Differences in antibiotic prescribing or in the performance of chest radiography or other ancillary tests between virus-positive and virus-negative or untested patients were not statistically significant. However, patients undergoing viral testing were less likely to receive antibiotics when subsequently seeing their primary care physician for the same illness within 1 week of discharge from the emergency department.
In a retrospective review, Byington and colleagues documented a significant decrease in antibiotic prescribing, with respect to inpatient intravenous therapy and oral antibiotic therapy at discharge, for hospitalized children who tested positive for RSV, parainfluenza 1, 2, 3, or adenovirus, compared with those who tested negative.
Although positive tests for viral pathogens are helpful, the sensitivity and specificity of rapid viral tests are not 100%, and false-negative and false-positive tests occur. For influenza, the sensitivity of each type of test varies by both method and sampling technique, and for the rapid tests, may also vary by the strains of influenza circulating in any given year. For example, the sensitivity of rapid influenza tests during 2009 pandemic H1N1 was poor compared with the performance of tests for seasonal influenza. For children with influenza like illness in a community with documented influenza virus circulation, a negative rapid influenza virus test in a child with CAP and symptoms compatible with influenza may reflect inaccuracies of the test, rather than reliably excluding influenza virus as a pathogen. For children with influenza, particularly those who require mechanical ventilation, initial results of nasopharyngeal testing for influenza may be negative, even with reverse-transcriptase PCR techniques, because of many factors, including poor-quality specimens, sampling of the upper rather than lower respiratory tract, and prolonged duration from illness onset to specimen collection. Multiple specimens on multiple days may be required for diagnosis. Because early influenza antiviral therapy provides the greatest benefit to the child, a clinician should not wait to start empiric antiviral therapy until after obtaining respiratory tract samples for diagnosis.
Some children with viral LRTI may also have an associated bacterial LRTI. In a study of 23 seriously ill, ventilated infants with documented RSV CAP, Levin et al found that 39% had specimens suggestive of concomitant bacterial infection based on tracheal aspirate cultures. They concluded from their patients and a literature review that evidence of bacterial pneumonia in otherwise low-risk infants with RSV presenting with respiratory failure is present in ≥20%, and the use of empiric antibiotics for 24–48 hours pending culture results may be justified until concomitant bacterial infection is excluded. However, for infants who do not have respiratory failure or any other findings that suggest bacterial coinfection, care process models have the potential to decrease inappropriate antibiotic use when they discourage such use in children who are documented to have a positive rapid test for a respiratory virus.
Clinical Infectious Diseases, Volume 53, Issue 7, 1 October 2011, Pages e25–e76,