Preregistered analytic plan: the gut microbiome and acute kidney injury in sepsis

OverviewWe here share a pre-registered analytic plan for a matched case-control study nested in a retrospective cohort of hospitalized patients with suspected sepsis. We will compare gut microbiota (measured near the time of admission) among patients with sepsis who do and do not develop sepsis-induced acute kidney injury. RationaleSepsis afflicts nearly 50 million people annually, and sepsis-induced acute kidney injury (AKI) is a frequent complication that contributes to morbidity, mortality, and increased healthcare costs. Despite the clinical significance of AKI in sepsis, we still do not understand why some patients with sepsis develop AKI while others do not. The gut microbiome has been implicated in other clinical features and sequelae of sepsis, but to date has not been studied in sepsis-induced AKI. ObjectiveTo determine whether gut microbiota predict AKI in patients with suspected sepsis. HypothesisWe hypothesize that among patients with suspected sepsis, gut bacterial density and identity (at the time of admission) predict the onset and severity of AKI. Study designWe will perform a matched case-control study nested in an observational cohort. The cohort includes patients admitted to the University of Michigan in 2016-2020 with suspected sepsis. We will divide patients into cohorts that did and did not develop AKI. We will derive matched cohorts based on relevant clinical covariates. We will characterize their gut microbiota using 16S rRNA gene amplicon sequencing of rectal swabs obtained within 24 hours of AKI onset. We will compare admission gut microbiota across these matched cohorts to test our primary hypothesis.