Inflammatory Prognostic Markers in COPD: Clinical Utility of CAR and MGPS

Background: Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are associated with substantial morbidity and mortality, highlighting the need for simple and reliable biomarkers for early risk stratification. The C-reactive protein to albumin ratio (CAR) and the modified Glasgow Prognostic Score (mGPS) reflect systemic inflammation and nutritional status, but their prognostic value in hospitalized patients with AECOPD remains incompletely defined. This study aimed to evaluate the associations of CAR and mGPS with one-year mortality and indicators of disease severity in hospitalized patients with AECOPD. Methods: In this retrospective single-center cohort study, 1556 adult patients hospitalized with a primary diagnosis of AECOPD between January 2020 and January 2026 were included. Patients with concomitant pneumonia and other major inflammatory conditions were excluded. Demographic, clinical, laboratory, arterial blood gas, and pulmonary function data were retrospectively analyzed. CAR was calculated from admission C-reactive protein and serum albumin levels, whereas mGPS was determined according to established criteria. Independent predictors of one-year mortality were identified using multivariable logistic regression analyses. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the discriminatory performance of CAR. Results: During the one-year follow up, 140 patients (9.0%) died. Compared with survivors, non-survivors had significantly higher CRP levels, higher CAR values, lower serum albumin levels, more severe hypercapnia and acidosis, and higher rates of intensive care unit admission and invasive mechanical ventilation (all p < 0.001). The distribution of mGPS differed significantly according to mortality status, with patients in the mGPS 2 category exhibiting the highest mortality rates (p < 0.001). After adjustment for clinically relevant covariates, age, acidosis, invasive mechanical ventilation, and CAR remained independently associated with one-year mortality. CAR demonstrated excellent discriminatory performance for predicting one-year mortality (AUC 0.907, 95% CI 0.893–0.922; p < 0.001), with an optimal cut-off value of >3.96, yielding 98.6% sensitivity and 84.3% specificity. Increasing mGPS scores were associated with progressively worse clinical outcomes, including higher rates of intensive care unit admission, invasive mechanical ventilation, in-hospital mortality, and one-year mortality. Conclusions: Inflammation and nutrition-based biomarkers are closely associated with disease severity and one-year mortality in hospitalized patients with AECOPD. CAR remained an independent predictor of one-year mortality and demonstrated excellent discriminatory performance, supporting its role as a simple and readily available complementary biomarker for prognostic assessment. Increasing mGPS scores were associated with progressively worse clinical outcomes, suggesting that mGPS may also contribute to risk stratification in hospitalized patients with AECOPD. Prospective multicenter studies with external validation are warranted to confirm these findings and to further define the prognostic value of these biomarkers.