Treatment of patients admitted to the hospital with an upper respiratory infection is often complicated by the lack of diagnostics. Even in cases where a patient has a confirmed case of influenza, there is still the possibility that they may have, or become infected by, a secondary bacterial infection. This leads clinicians to treat patients empirically with antibiotics, which can result in unnecessary antibiotic use among patients without a bacterial infection.
Though antibiotics can be a lifesaving drug, their use is not without risks. Estimates suggest that 20% of patients taking common antibiotics experience some side effect. While most side effects are not life-threatening gastrointestinal effects, other nonnegligible side effects include anaphylactic shock, drug‐induced liver injury, increases in the risk of retinal detachment, serious arrhythmias, and superinfection with resistant bacteria. Antibiotics can also lead to secondary infections, such as Clostridium difficile.
Yet, despite these risks, there has been limited research on the percentage of patients with influenza who actually have a bacterial coinfection. A recent systematic review and meta-analysis conducted by my colleagues at Johns Hopkins University and the Center for Disease Dynamics, Economics & Policy examined the frequency of bacterial coinfection among hospitalized patients with influenza and identified the most common infecting bacterial species.
The findings, published in the journal Influenza and Other Respiratory Viruses , found that in the majority of studies, between 11% and 35% of patients with confirmed influenza had a bacterial coinfection. The most common coinfecting bacteria were found to be Streptococcus pneumoniae and Staphylococcus aureus. Combined, S. pneumoniae and S. aureus accounted for more than 60% of the identified coinfecting bacteria; however, many other bacterial species were found to cause infections as well.
The results suggest that while bacterial infection is common in influenza patients, only about a quarter of patients are likely to be infected. However, the studies were widely heterogeneous, both in patient makeup and results. Analyses of age, setting, enrollment year, study type, study size, and bacterial collection methods did not reveal a source for the heterogeneity in results. Thus, additional factors, such as patient comorbidities or prior antibiotic use, which could not be systematically assessed, may affect the likelihood of coinfection.
Given that the symptoms of influenza and bacterial infection often overlap, correctly diagnosing bacterial coinfection without a laboratory culture can present a challenge. This diagnostic uncertainty leads to significant overuse of antibiotics in patients with influenza alone. Most influenza cases will never result in serious bacterial infections (particularly in nonhospitalized patients), and thus a lot of antibiotic use is unnecessary. As mentioned earlier, unnecessary antibiotic use poses a nonnegligible risk to patients, but it also contributes significantly to rising rates of antibiotic resistance, a major public health issue.
The results from this study highlight that the majority of patients hospitalized with influenza are unlikely to be coinfected with a bacterial pathogen. Thus, it is important for clinicians to appropriately treat patients with antiviral drugs, and ensure that bacterial testing is done when presumptively starting patients on antibiotics. Based on the microbiology results, antibiotics can be stopped if no pathogen is identified or altered to be more appropriate depending on the pathogen found.
Although the findings of this study suggest we need a more thorough analysis of the issue, the results should still aid clinicians by improving their understanding of the likelihood of bacterial coinfection in hospitalized patients with influenza, and thus help them balance the need to minimize patient risks as well as the individual and societal risks of nonessential antibiotic use.
Eili Klein, PhD, is assistant professor in the department of emergency medicine at Johns Hopkins Medicine, Baltimore.
