Photo Credit: vkovalcik
The following is a summary of “How likely is it that a virus or bacteria is causing a patient’s symptoms? A new approach to interpret the outcome from multi-pathogen PCR,” published in the January 2025 issue of Infectious Disease by Hulme et al.
The distinction between a pathogen requiring treatment and a commensal ‘passenger’ has been unclear, especially with the broader use of multi-pathogen PCR.
Researchers conducted a retrospective study to propose a new statistical procedure for analyzing and presenting data from case-control studies to clarify the probability of causality.
They conducted a case-control study in US outpatient settings, enrolling 18 to 75-year-old patients with acute lower respiratory tract infections and controls without symptoms. Multi-pathogen PCR testing was performed on patients. The positive etiologic predictive value was calculated to estimate the probability of each pathogen causing the symptoms. Outcomes were shown using a modified forest plot, categorizing pathogens into 5 groups based on the likelihood of causality.
The results showed 618 adult cases, and 497 asymptomatic controls with modified forest plot and causality classification helped facilitate understanding. Influenza A and B, SARS-CoV-2, rhinovirus, and parainfluenza viruses were likely causative pathogens, while Staphylococcus aureus was mostly commensal. Broad confidence intervals for the positive etiologic predictive value hindered conclusions for potential pathogens with low prevalence.
Investigators concluded the statistical approach was likely practical for larger case-control studies or meta-analyses, potentially aiding in the interpretation of multi-pathogen PCR results.
Source: tandfonline.com/doi/full/10.1080/23744235.2025.2456902#abstract
