Shotgun metagenomics for the diagnosis of infections: A prospective study

Shotgun metagenomics (SMg) promises to significantly improve the microbiological diagnosis of infectious diseases. However, the prospective evaluation of its clinical utility in different infectious syndromes remains poorly documented.

We conducted a prospective study including all patients who underwent SMg as part of their care at the French Henri Mondor Hospital National Reference Laboratory for accredited SMg between February 2018 and January 2020. Patients were categorized as having either a "high likelihood" or "low likelihood" of infection based on their clinical presentation. The contribution of the SMg to the final diagnosis was assessed by a multidisciplinary team of infectious disease specialists.

202 patients were enrolled in the study. Of the 123 patients considered to have a high likelihood of infection, SMg confirmed the cause of infection in 38 cases (30.9%), including 12 cases (9.8%) diagnosed exclusively by this method. In the 79 patients classified as having a "low likelihood" of infection, SMg did not detect any microorganisms compatible with an infectious cause. In particular, patients undergoing immunosuppressive treatment within the latter group showed no deterioration after 6 months.

SMg facilitated microbiological diagnosis in over 30% of complex cases, regardless of sample type or site of infection. Compared with conventional techniques, SMg provided diagnoses in 10% more cases, highlighting its broad utility across different infectious diseases. Our results suggest that SMg is a promising tool for documenting complex infectious diseases alongside traditional microbiology tools. Furthermore, negative SMg results are useful for the management of patients with a low likelihood of infection.