From Host Response to Microbiological Risk: Association of the RIN-IM Score with Culture-Positive Sepsis and Antimicrobial Resistance in Critically Ill Adults

Background/Objectives: Early identification of microbiologically confirmed sepsis and antimicrobial-resistant infection remains challenging in critically ill patients. The renal–inflammatory–nutritional–metabolic (RIN-IM) score integrates serum creatinine, the systemic immune-inflammation index, pan-immune inflammation value, serum albumin, and triglyceride–glucose index. This study investigated the association of the RIN-IM score with culture-positive sepsis and antimicrobial resistance phenotypes in critically ill adults. Methods: This single-center retrospective cohort study included 318 adults with sepsis admitted to the intensive care unit between 1 January 2023 and 31 March 2026. The RIN-IM score was calculated from laboratory measurements obtained within 24 h and categorized as low (0–3), intermediate (4–6), or high risk (7–10). The primary outcome was culture-positive sepsis. Secondary analyses, restricted to patients with culture-positive sepsis, examined associations between RIN-IM and multidrug-resistant (MDR), extensively drug-resistant (XDR), and pandrug-resistant phenotypes. Results: Culture-positive sepsis was identified in 154 patients (48.4%). The median RIN-IM score was higher in culture-positive than in culture-negative patients (6 [interquartile range, 4–8] vs. 4 [2,3,4,5]; p < 0.001). Culture positivity increased from 29.1% in the low-risk group to 50.4% in the intermediate-risk group and 74.6% in the high-risk group. Each one-point increase in RIN-IM was independently associated with culture-positive sepsis (adjusted odds ratio [aOR], 1.35; 95% confidence interval [CI], 1.21–1.51), while high- versus low-risk classification was associated with an aOR of 4.65 (95% CI, 2.25–9.62). Among culture-positive patients, 67 (43.5%) had MDR, and 20 (13.0%) had XDR infections. Each one-point score increase was independently associated with MDR infection (aOR, 1.29; 95% CI, 1.12–1.48). The exploratory XDR analysis was based on 20 events and used Firth’s bias-reduced logistic regression; a one-point score increase was associated with XDR infection (aOR, 1.29; 95% CI, 1.09–1.52), but the categorical estimates were imprecise. The score demonstrated moderate discrimination for culture-positive sepsis (AUC, 0.752; 95% CI, 0.699–0.805) and MDR infection (AUC, 0.724; 95% CI, 0.643–0.805); discrimination for XDR infection was exploratory (AUC, 0.742; 95% CI, 0.635–0.849). Twenty-eight-day mortality increased from 13.6% to 27.7% and 47.9% across the three risk categories (p < 0.001). Conclusions: Higher admission RIN-IM scores were associated with culture-positive sepsis. Among patients with culture-positive sepsis, higher scores were also associated with MDR infection; this subgroup analysis was conditioned on microbiological confirmation and should not be interpreted as prediction of resistant infection in the overall sepsis population. The association with XDR infection was exploratory, based on only 20 events, and requires confirmation in larger independent cohorts. Because RIN-IM assessment and microbiological sampling were not fully time-synchronized, these findings should not be interpreted as real-time microbiological prediction or evidence of clinical utility. RIN-IM should not replace culture-based diagnosis, susceptibility testing, individual risk-factor assessment, local antibiograms, or clinical judgment. The observed associations do not establish a direct biological relationship between RIN-IM and microbial burden or antimicrobial resistance.