Table 1 Challenges and considerations in COVID-19 respiratory microbiome studies

Antibiotic usage

More severe patients are more likely to receive antibiotics, making it difficult to differentiate between association w/disease severity and association w/antibiotic use.

Studies profiling patients prior to treatment avoid this effect (i.e., Bradley et al. [49]). Other studies with both COVID and non-COVID-19 patients receiving antibiotics can identify the effect of specific antibiotics, which can then be separated from the effects of COVID (i.e., Lloréns-Rico [48])

Mechanical ventilation

More severe patients require mechanical ventilation, which is known to impact the respiratory microbiome.

Similar mitigation techniques as used for high antibiotic use can apply to mechanical ventilation.

SARS CoV-2 variant

Some SARS CoV-2 variants may cause more or less severe disease, though the impact of variants on the respiratory microbiome is unknown.

The vast majority of studies here predate the emergence of any variants of concern, but in future studies, it will be important to sequence SARS CoV-2 genomes and report the variants included in microbiome studies to determine if there is an impact on the respiratory microbiome.

Sample storage in viral transport media

Swabs initially collected for SARS CoV-2 testing are often stored in VTM, which contains antibiotics and also nutrients that allow some bacteria to grow out.

Several studies sample the upper respiratory tract microbiome without storing in VTM (e.g., Braun et al., Mostafa et al. Hurst et al. [50,51,52]).

Prior immunity to SARS Cov-2

Patients with prior immunity via vaccination or prior infection are less likely to have severe COVID, but the impact of prior immunity on the microbiome are unknown.

No studies have yet reported the microbiome of COVID-19 patients with prior immunity. The impact can be mitigated by ensuring balanced enrollment of immunized and naïve patients at each level of disease severity in future microbiome studies.

Inconsistent analytical techniques

Each study uses different analysis methods to identify differential taxa in the microbiome. This is known to significantly impact which associations are identified.

Making data publicly available for reanalysis and meta-analysis would allow for standardization across studies. A minority of studies make sufficient data available (e.g., Gupta et al., Kullberg et al. [53, 54]).