Abstract
Background: Serum albumin and prealbumin are increasingly recognized as biomarkers of inflammation rather than nutritional status. Because prealbumin responds more rapidly to acute inflammatory and catabolic stress than albumin, the albumin-to-prealbumin ratio (APR) may better reflect inflammatory burden. However, its prognostic significance in elderly patients with sepsis remains unclear. Methods: This single-center retrospective observational study included 84 patients aged ≥75 years hospitalized with sepsis between April 2023 and March 2024. APR was calculated using serum albumin and prealbumin levels at admission, and patients were categorized into low- and high-APR groups based on the median value. The primary outcome was 28-day mortality. Multivariable logistic regression analysis was performed to evaluate the independent association between APR and 28-day mortality after adjustment for the Sequential Organ Failure Assessment score, maximum serum lactate level, and Mini Nutritional Assessment-Short Form. Subgroup analyses were conducted according to median serum albumin level (<2.9 vs. ≥2.9 g/dL). Results: Twenty-eight-day mortality was significantly higher in the high-APR group than in the low-APR group (40.5% vs. 7.1%, p = 0.001), with significantly poorer survival (log-rank p = 0.003). A high APR remained independently associated with increased 28-day mortality (odds ratio 10.2, 95% confidence interval 2.3-45.1). In subgroup analyses, APR was associated with mortality only among patients with relatively preserved albumin levels. Conclusions: APR is a useful prognostic marker of 28-day outcomes in elderly patients with sepsis and may provide prognostic information beyond conventional severity indices.