A new iScience study combined machine learning with clinical and preclinical experiments to show lactate is both a powerful predictor and a mechanistic driver of acute kidney injury after cardiopulmonary bypass. In diabetic kidney disease, researchers identified a ceramide-gated Calhm2 calcium influx pathway (driven by Aebp1) injuring podocytes, with the authors reporting a promising urinary biomarker angle. Perioperative glucose management, meanwhile, was addressed in a target trial emulation study: continuing GLP-1 receptor agonists before thyroidectomy in adults with type 2 diabetes was not associated with excess perioperative harm, while certain nominal risks remain hypothesis-generating. The combined message is that kidney injury risk can be mechanistically modeled, biomarker-backed, and managed with clearer perioperative guidance for common diabetes medications.