Does a Loading Dose Improve Outcomes? Colistin Therapy for Urinary Tract Infections Caused by Multidrug-Resistant Gram-Negative Bacteria

Background: Colistin remains an important therapeutic option for complicated urinary tract infections (cUTIs) caused by multidrug-resistant Gram-negative bacteria. A loading dose is often recommended to rapidly achieve steady-state plasma concentrations, but its clinical benefit specifically for cUTIs, where local urinary drug concentrations markedly exceed plasma levels, remains unclear. Methods: We conducted a retrospective cohort study of adult patients treated with intravenous colistin for cUTIs at Nakornping Hospital, Chiang Mai, Thailand, between 2015 and 2022. Patients were classified as receiving loading-dose (LD) or non-loading-dose (non-LD) colistin. The primary outcome was 30-day all-cause mortality. Secondary outcomes included clinical response, microbiological response, and nephrotoxicity (defined by RIFLE criteria). Inverse probability weighting (IPW) using the propensity score was applied to adjust for baseline covariates, and Cox proportional hazards regression was used to estimate crude and adjusted hazard ratios (HR) using the non-LD group as the reference. Results: Among 123 patients (74 LD, 49 non-LD), unadjusted 30-day all-cause mortality was 35.14% versus 40.82% (p = 0.524), clinical response was 75.68% versus 71.43% (p = 0.599), and microbiological response was 78.38% versus 65.31% (p = 0.109) for LD versus non-LD, respectively. Nephrotoxicity occurred in 48.65% versus 38.78% (p = 0.281). After IPW adjustment, no statistically significant differences were observed between the LD and non-LD groups for 30-day all-cause mortality (aHR 0.78, 95% CI 0.43–1.42, p = 0.419), clinical response (aOR 1.48, 95% CI 0.54–4.08, p = 0.445), or microbiological response (aOR 2.10, 95% CI 0.81–5.48, p = 0.128). For the safety outcome, nephrotoxicity was analyzed using the Fine–Gray competing risk model, treating death as a competing event. The IPW-adjusted subdistribution hazard ratio was (asHR) 2.23 (95% CI 0.83–5.95, p = 0.110), indicating no statistically significant difference in renal safety between the LD and non-LD groups. Conclusions: Loading-dose colistin was not associated with a statistically significant reduction in 30-day all-cause mortality or improvements in clinical response and microbiological eradication in patients with cUTIs. Nephrotoxicity was numerically lower without a loading dose, but the wide confidence intervals in this study of 123 patients precluded a definitive conclusion regarding renal safety. Omitting the loading dose may be a reasonable option for localized urinary infections, particularly in patients at high risk for acute kidney injury, pending confirmation in larger prospective studies.