Colistin is essential for treating multidrug-resistant infections but has a narrow therapeutic window, causing dose-dependent nephrotoxicity in up to 70% of patients. Conventional dosing based on creatinine clearance (CrCl) overestimates kidney function—and risks overdosing—in patients with low muscle mass, in whom serum creatinine is unreliable. The combined creatinine–cystatin C equation (eGFRcr-cys) improves GFR estimation in this group, but whether it improves kidney outcomes when used for dose adjustment is unproven. We therefore examined whether eGFRcr-cys-guided colistin dosing reduces acute kidney injury compared with conventional creatinine clearance–based dosing.

We conducted a single-center, randomized, controlled trial in hospitalized adults receiving intravenous colistin within 24 hours and at risk of low muscle mass, defined by a sarcopenia index <0.8 (serum creatinine/serum cystatin C) and excluded patients on dialysis, kidney transplants, severe AKI, and moribund state. Participants were randomized 1:1 to receive colistin dosing based on either combined creatinine-cystatin C estimated GFR (CKD-EPI 2021 equation; eGFRcr-cys) or Cockcroft-Gault estimated CrCL (CG-CrCl). The primary outcome was AKI incidence, defined by KDIGO 2012 criteria, within 14 days. Secondary outcomes included treatment duration, cumulative colistin exposure, renal replacement therapy use, microbiological clearance, hospital mortality, and major adverse kidney outcomes at day 30 (MAKE30). The trial protocol was published in Thai Clinical Trial Registry (TCTR20240605006).

Of 150 randomized patients with mean sarcopenia index of 0.48 ± 0.16, 75 received eGFRcr-cys dosing and 75 CG-CrCL dosing. Daily colistin exposure was lower in the eGFRcr-cys group compared to control (mean 232.92 ± 65.8 vs 261.6 ± 67.1 mg; mean absolute difference -28.65 mg/day, 95% CI -50.1 to -7.2, p=0.009). However, AKI incidence at day 14 was similar between groups (58.7% vs. 64%; p=0.62). Other clinical outcomes, including RRT use, microbiological clearance, hospital mortality, and MAKE30, did not differ significantly.

This trial, one of the first to test eGFRcr-cys-guided colistin dosing, showed that although eGFRcr-cys-guided dosing reduced daily colistin exposure without increasing serious adverse clinical outcomes, it did not improve kidney outcomes. At present, the evidence remains insufficient to recommend routine colistin dose adjustment using the combined creatinine–cystatin C equation.