Sepsis is a life-threatening organ dysfunction triggered by severe infection and remains a major global health challenge due to its persistently high mortality (1, 2). Accurate early risk stratification is essential for improving prognosis, and reliable, easily accessible biomarkers play a critical role in this process (3). In sepsis, two key pathophysiological components dominate: the overwhelming inflammatory cascade initiated by pathogen invasion, and profound metabolic dysregulation (4, 5). Conventional inflammatory biomarkers, such as high-sensitivity C-reactive protein (hs-CRP) and procalcitonin (PCT), broadly reflect infection severity but have limited diagnostic specificity (6). In particular, hs-CRP is frequently elevated in chronic non-infectious conditions (e.g., diabetes, autoimmune disorders, chronic liver disease), thereby reducing its ability to reliably distinguish infectious from non-infectious inflammation (7–9). High-density lipoprotein cholesterol (HDL-C), which exerts anti-inflammatory and antioxidant effects and participates in reverse cholesterol transport, is an important protective molecule in lipid metabolism (10, 11). During sepsis, pro-inflammatory cytokines (e.g., TNF-α, IL-6) impair the structure and function of HDL-C and cause a rapid decline in circulating HDL-C levels, which has been linked to poor prognosis in sepsis patients (12, 13). However, HDL-C alone primarily reflects metabolic protection and fails to capture the complex interaction between inflammation and metabolic dysfunction. Despite extensive research on individual inflammatory or lipid biomarkers, the current evidence on composite inflammation-metabolic indices for sepsis prognosis remains limited, and there is no widely accepted composite marker that can robustly capture the interaction between systemic inflammation and metabolic dysregulation in sepsis (14, 15), partly due to small sample sizes, heterogeneity in study populations and outcomes, and limited external validation (14). In recent years, the composite inflammation-metabolic ratio has demonstrated superior prognostic value in cardiovascular diseases (16), but its utility in sepsis remains under-validated.

The hs-CRP/HDL-C ratio (HCHR) constitutes a composite biomarker that integrates both inflammatory injury (elevated hs-CRP) and metabolic protection loss (reduced HDL-C) (17). Across cardiovascular and metabolic disease settings, multiple studies have consistently demonstrated that higher HCHR is associated with worse clinical outcomes, including increased risk of mortality and major adverse events, and in many analyses it shows stronger risk stratification than hs-CRP or HDL-C alone (16, 18, 19). Given that sepsis is fundamentally driven by an interplay between uncontrolled inflammation and metabolic dysregulation, we propose that HCHR may serve as a more accurate prognostic indicator in sepsis patients. Therefore, the present study aims to investigate the association between HCHR and 28-day all-cause mortality in a large retrospective cohort of sepsis patients.

This study investigated the utility of HCHR as an integrated inflammatory–metabolic biomarker for prognostic assessment in sepsis. In this large retrospective cohort, higher HCHR was independently associated with increased 28-day all-cause mortality after adjustment for demographic characteristics, comorbidities, clinical variables, and laboratory parameters (fully adjusted model: OR 6.10, 95% CI 3.48–10.68). Moreover, HCHR demonstrated moderate discrimination (AUC 0.686), outperforming hs-CRP or HDL-C alone, suggesting that combining inflammatory burden with metabolism may provide more informative risk stratification than single biomarkers.

The pathophysiological basis of HCHR in sepsis lies in its capture of overwhelming systemic inflammatory and lipid metabolic defense mechanisms. First, hs-CRP is a downstream effector largely driven by IL-6 signaling and reflects the magnitude of systemic inflammatory activation (20). This inflammatory surge contributes to endothelial dysfunction, microcirculatory derangements, and organ injury (20), which are central to sepsis-related mortality. However, because hs-CRP can also increase in various chronic inflammatory conditions, its specificity for assessing infection alone is limited (7–9). Second, decreased HDL-C is not merely a marker of dyslipidemia but reflects the collapse of key homeostatic systems. HDL-C exerts pleiotropic protective effects relevant to sepsis, including endotoxin neutralization, anti-inflammatory and anti-oxidative actions, modulation of innate immune responses, and preservation of endothelial barrier integrity and microvascular function (21, 22). During sepsis, inflammatory cytokines suppress HDL-C synthetase activity and disrupt HDL-C structure, leading to decreased levels and functional inactivation (23). Reduced HDL-C further impairs the body’s compensatory capacity against inflammation, forming a vicious cycle of exacerbated inflammation and weakened protection (24). HCHR quantifies the degree of inflammatory injury via elevated hs-CRP and the loss of metabolic protective capacity via reduced HDL-C. This ratio serves as a composite indicator of the injury-protection imbalance in sepsis. A heightened ratio reflects greater inflammatory severity and inadequate metabolic compensation, both of which are associated with an unfavorable prognosis. This plausibly explains why HCHR provides more discriminative prognostic information than either component alone. Previous studies have supported this: Gao et al. demonstrated in a large national cohort that HCHR exhibited a significantly higher AUC for predicting cardiovascular events in middle-aged and elderly individuals compared with hs-CRP or lipid markers alone (16). Sun et al., using data from NHANES 2015–2023, further reported that HCHR was positively associated with the risk of prediabetes and diabetes, suggesting its utility in early metabolic risk stratification (19). In addition, Chen et al. found that elevated preoperative HCHR increased the likelihood of postoperative SIRS in elderly patients (18). Our study also validated the clinical utility of HCHR in a large-sample sepsis population.

Our results are also consistent with existing clinical literature on the individual components of HCHR in sepsis and critical illness. Multiple studies have demonstrated that elevated CRP or hs-CRP independently predicts mortality in septic or critically ill populations (25, 26), whereas reduced HDL-C levels correlate with heightened cytokine release, increased vasopressor demand, and poor survival outcomes (23, 27). More recently, composite biomarkers combining inflammatory and lipid-related parameters have been shown to outperform single indices in risk stratification, further corroborating the clinical relevance of integrative markers such as HCHR (16, 18, 19). Notably, in cardiovascular disease and related metabolic conditions, HCHR has been repeatedly associated with higher risks of major adverse events and mortality, and may provide incremental prognostic value beyond isolated inflammatory or lipid markers, which is consistent with our study findings. The consistency across disease domains supports that HCHR reflects the comprehensive role of systemic inflammation and lipid metabolism (16, 28), which is relevant not only in atherosclerotic cardiovascular disease but also in sepsis, where systemic inflammation, endothelial injury, and metabolic reprogramming coexist (29, 30).

From a clinical perspective, our RCS analysis revealed a linear dose–response relationship between HCHR and 28-day mortality without an apparent threshold, indicating that mortality risk increases progressively with rising HCHR values. Given its moderate discrimination, HCHR should be positioned as an adjunctive tool to support early risk stratification rather than a standalone decision-making biomarker. In practice, several clinical application scenarios merit prospective evaluation. First, HCHR could be incorporated into early triage workflows within the first 24 h of ICU admission to identify patients who may benefit from closer monitoring or intensified reassessment. Second, regarding integration into clinical scoring systems, our findings support considering HCHR as a candidate variable to be added to established prognostic models (e.g., SOFA- or APACHE-based frameworks) to test whether it provides incremental predictive value. But this still needs multicenter prospective verification. Third, dynamic monitoring of HCHR trajectories may offer additional prognostic information beyond a single early measurement, potentially reflecting treatment response and evolving immunometabolic status during hospitalization.

Nevertheless, several limitations should be acknowledged. First, its single-center retrospective design may introduce selection bias. Causal inference cannot be established and the observed associations should not be interpreted as causal effects. Multicenter prospective studies are needed. Second, HCHR was derived from a single early time-point measurement (worst values within 24 h of ICU admission), and the prognostic significance of longitudinal changes and trajectories of HCHR remains unclear. Third, medication use and treatments that may influence lipid profiles and inflammatory markers (e.g., statins, corticosteroids, and lipid-modifying therapies) were not fully captured or adjusted for, which may have affected HDL-C, hs-CRP, and thus HCHR, contributing to residual confounding. Finally, the biological mechanisms by which HCHR affects sepsis prognosis require further experimental exploration.

In summary, HCHR is independently associated with 28-day mortality in sepsis and demonstrates better prognostic discrimination than hs-CRP or HDL-C alone. Given its accessibility and low cost, HCHR may serve as a practical adjunct biomarker for early risk stratification and as a candidate component for future enhanced prognostic models, pending prospective validation.