Are we overlooking a hidden organ in sepsis-associated AKI?

Sepsis-associated acute kidney injury (SA-AKI) is among the most common and deadly complications in the intensive care unit. Its pathogenesis is complex, with hypoperfusion, inflammation, and nephrotoxin exposure layered upon substantial host heterogeneity. Antibiotic selection adds yet another variable. Anti-anaerobic antibiotics are widely prescribed in sepsis, and observational studies have linked these agents to increased risks of mortality and AKI.1,2 At a cursory glance, the ACORN pragmatic trial seemed to put this issue to rest: among infected patients randomized to cefepime or piperacillin-tazobactam, AKI rates were similar.3 However, on closer inspection, nearly half of the cefepime arm received metronidazole; thus, ACORN was not designed to answer the question of anti-anaerobic antibiotics and AKI risk. Mounting evidence suggests that the gut microbiome plays a critical role in the pathogenesis of critical illness.4 Sepsis disrupts the integrity of the intestinal barrier, allowing translocation of bacterial products into circulation, amplifying systemic inflammation, and worsening end-organ injury.4,5 Conversely, AKI can alter gut microbiome composition through direct effects of uremia and intestinal perfusion.6 Anti-anaerobic antibiotics have been shown to deplete protective commensals, reduce diversity, and promote Enterobactericiae overgrowth.7 However, despite a plausible mechanism, evidence directly linking the full sequence of events—anti-anaerobic antibiotic exposure, microbiome perturbation, and SA-AKI—has been limited and largely speculative, until now. In this issue of the Journal, Winner et al. implement a series of complementary analyses to determine whether anti-anaerobic antibiotics play a causal role in SA-AKI, and if so, how (https://doi.org/10.1164/rccm.202411-2281OC).8 First, in a retrospective cohort, administration of anti-anaerobic antibiotics was independently associated with a 61% increased risk of SA-AKI, consistent with prior studies.1,2 To strengthen causal inference, the authors performed an instrumental variable analysis, leveraging a piperacillin-tazobactam shortage to approximate randomization. Both piperacillin-tazobactam and cefepime/metronidazole were linked with a greater risk of AKI onset and decreased likelihood of AKI recovery compared to cefepime alone. Looking under the hood, the authors next performed a case-control analysis of rectal swabs from matched septic patients with and without SA-AKI, identifying compositional differences in the gut microbiome, including increased abundance of Enterobactericiae and Lachnospiraceae in those with SA-AKI. Finally, in murine sepsis models, genetically identical mice harboring different microbiomes exhibited different AKI severity following the same sepsis insult; Lachnospiraceae again was implicated. These studies elegantly paint a picture of anti-anaerobic antibiotics leading to microbiome shifts and ultimately SA-AKI (Figure 1). At the same time, each component analysis has limitations that temper interpretation. The retrospective cohort analysis shares the limitations of preceding observational studies, namely confounding by indication. In the instrumental variable analysis, the stratification of cefepime by metronidazole co-administration, while compelling, was not subject to the same natural experiment and therefore shares the limitations of retrospective studies. In the case-control microbiome analysis, the timing of rectal swab collection and AKI identification was essentially coincident, making directionality difficult to establish. Finally, while the murine experiments offered rigorous control of host factors, they cannot definitively establish causality, and it is possible that microbiota differences influenced immune development well before the sepsis challenge. Taken together, the complementary strengths of these analyses point toward a persuasive narrative, but definitive causality and mechanism remain to be proven. Conceptual model linking anti-anaerobic antibiotics, gut microbiome disruption, and sepsis-associated acute kidney injury. Figure created in BioRender. One of the study’s more intriguing and perplexing findings was that Lachnospiraceae were enriched in both septic patients with SA-AKI and in the mice with higher creatinine. This taxon was recently found by the same research team to explain temperature variability in human and murine sepsis, underscoring its key role.9 However, Lachnospiraceae are obligate anaerobes, which should be depleted by the anti-anaerobic antibiotics that are posited to cause SA-AKI, and they produce short-chain fatty acids, which have been shown to be renoprotective.10 This apparent paradox illustrates the complexity of microbiome research, where the community structure, interspecies interaction, and metabolic activity likely matter more than simple abundance. Looking forward, functional analyses, including microbial metabolomic profiling, may help illuminate not only what microorganisms are present, but what they are doing and how those pathways intersect with AKI. Even if the microbiome acts as an intermediary between anti-anaerobic antibiotics and SA-AKI, as the data suggest, it likely explains only part of the story. In the controlled, antibiotic-free murine experiment, creatinine responses to sepsis varied ­widely, with Lachnospiraceae abundance only accounting for 18% of the variance. Host heterogeneity, including differences in innate immune responses, neurohormonal tone, and renal reserve, likely contributes substantially to SA-AKI, though these features are difficult to measure.11 In addition, there are plausible connections between antibiotics and SA-AKI that are independent of microbiome effects. Some antibiotic regimens, irrespective of anti-anaerobic activity, may simply be more effective at treating the underlying infection, either because local antibiograms favor them or because they more readily achieve pharmacokinetic targets such as time-over-MIC, and better sepsis control translates to better renal outcomes.12,13 In addition, antibiotics can directly affect creatinine. While piperacillin-tazobactam alone does not appear to be particularly nephrotoxic, multiple studies have reported excess nephrotoxicity with the combination of vancomycin/piperacillin-tazobactam, though this is controversial.14 Pseudo-AKI, or altered creatinine secretion without impaired glomerular filtration, has also been documented with several antibiotics, including piperacillin-tazobactam, and further complicates interpretation in studies like this one that define AKI by creatinine-based criteria.14 This important study has strengthened the link between anti-anaerobic antibiotics, gut ecology, and AKI, yet the mechanistic pathway remains unresolved. Our next challenge is to learn precisely how the microbiome contributes to SA-AKI and how to manipulate this axis to improve patient care. Doing so will require a multi-modal approach: preclinical models with deliberate microbiome manipulation, prospective human studies with longitudinal sampling and microbial metabolomic profiling, and interventional trials that truly compare anti-anaerobic versus anaerobe-sparing antibiotics and test microbiome-targeted therapeutics such as purified microbial metabolites, live biotherapeutics, and GI-targeted antibiotic adsorbents.10,15,16 Clinicians must weigh numerous factors when prescribing empiric antibiotics, including prior microbiology, local resistance patterns, institutional guidelines, allergies, illness severity, and uncertainty about infectious source. However, where true equipoise exists in antibiotic selection, perhaps favoring anaerobe-sparing regimens could help protect what Winner et al. remind us is a criti­cal “hidden organ” in SA-AKI. Supplementary material is available at American Journal of Respiratory and Critical Care Medicine online. Please see the ICMJE disclosure forms, which have been provided as supplementary material. No artificial intelligence tools were used in writing this manuscript.