The false dichotomy of viral versus bacterial aetiology in upper respiratory tract infections

It's time to move away from the viral versus bacterial dichotomy and embrace an evidence‐based syndromic approach

Antimicrobial resistance is a key emerging threat to human health, driven largely by overuse of antibiotics in humans and animals.1 Upper respiratory tract infections (URTIs) are the most common indication for which antibiotics are prescribed in the community,2 despite the fact that antibiotics are of minimal benefit for most of these syndromes. We define URTI broadly to include acute rhinitis, bronchitis, otitis media, pharyngitis and sinusitis. When faced with a patient presenting with symptoms and signs of a URTI, general practitioners face many challenges to antimicrobial stewardship, including patient expectation, but they often frame the problem as “is this viral or bacterial?”. They then base their decision to prescribe antibiotics or not on this distinction. GPs cite diagnostic uncertainty about viral versus bacterial aetiology as a key factor driving their antibiotic prescribing for URTIs.3 Many educational resources aimed both at prescribers and the community refer to the viral versus bacterial paradigm to inform the choice of whether antibiotics should be prescribed or not.

However, we would argue that this is a false dichotomy, and we should be seeing this question through the lens of evidence‐based medicine, not through theoretical ideas about aetiology. There is extensive high level evidence from randomised controlled trials that antibiotics have minimal or no benefit for URTIs;4,5,6 most of these trials enrolled patients based on clinical syndromes, not microbial aetiology. The normal flora of the healthy human upper respiratory tract includes both viruses (the virome)7 and bacteria (the microbiome).8 Healthy asymptomatic humans often harbour pathogenic viruses transiently in their upper respiratory tracts. For example, in a prospective study in Utah, 108 individuals in 56 households underwent weekly nose and throat swabs for multiplex respiratory virus polymerase chain reaction assay for a year.9 A potentially pathogenic respiratory virus (eg, rhinovirus, parainfluenza virus) was detected in 50% of all time points in children aged less than 5 years, and in 11% of time points in adults. The patient from whom the virus was detected was asymptomatic in 44% of 783 episodes. Sampling from the middle ear10 and bronchial tree11 during acute URTI shows a mixed community of bacteria and viruses. Take bronchitis as an example: there is not a pure growth of a typical pathogen such as Haemophilus influenzae in an otherwise sterile field, rather there is a shift in the usual microbiome and virome so that H. influenzae becomes dominant. Moreover, there are complex interactions between various elements of the microbiome. For example, acute viral infections have been shown to lead to a disturbance of the normal respiratory microbiome, and in turn, an altered host immune response which allows relative overgrowth of pathogenic bacterial species.12

Although antibiotics have little effect on clinical outcomes in URTIs, they do have substantial effects on the microbiome. Even a single course of antibiotics alters the gut microbiome for up to 12 months.13 Emerging data on the importance of the microbiome in diverse aspects of health and disease ranging from metabolism to inflammation underline the need to avoid disturbing the microbiome wherever possible. Antibiotic use increases not only the community risk14 but also the individual risk of a subsequent antibiotic‐resistant infection.15 This evidence of harm needs to be weighed against any potential benefit of antibiotics in a particular patient.

We suggest that we move away from this false dichotomy in consumer and prescriber education and move to evidence‐based syndromic explanations of our therapeutic reasoning. For example, rather than saying “antibiotics are needed for acute otitis media if it's bacterial but not if it's viral”, we could say “21 children with otitis media need to be treated with antibiotics for one less child to have pain after 3 days”. The bottom line is that most URTIs (even those traditionally considered bacterial) resolve whether or not antibiotics are used. As an example, acute sinusitis resolves in 80% of patients with no antibiotics and 90% with antibiotics after 7 days.4 Such messages need caveats — for example, in acute sore throat, antibiotics are indicated if the patient is from a high risk population for acute rheumatic fever. The results of the cited randomised controlled trials do not apply to immunosuppressed patients or those with severe comorbidities such as chronic lung disease, where a lower threshold for using antibiotics is appropriate. The human immune system is usually excellent at responding to and eliminating URTIs, whether they be predominantly bacterial or viral. Even in invasive bacterial infections (such as bacteraemia), antibiotics simply help shift the balance in favour of the immune system, rather than replace it entirely.

Given our increasing understanding of the complexity of the human microbiome, and the imperative to reduce the use of antibiotics, it's time we moved away from the black and white bacterial versus viral dichotomy and embraced a new concept of disease reflecting a shift in the usual spectrum of bacteria and viruses. So, the next time you see a patient with a URTI and are deciding whether to prescribe antibiotics or not, think “syndromically” and consider the published evidence in that clinical syndrome, not the false dichotomy of bacterial versus viral aetiology: it's not bacterial or viral, it's both!