Symptom trajectory types during recovery from acute exacerbation of COPD and their association with systemic inflammatory burden and prognosis.

OBJECTIVE: To identify symptom trajectory types during recovery from AECOPD using GBTM, explore the role of admission systemic inflammatory burden (assessed by CRP and PCT) in shaping recovery heterogeneity, and analyze the association between trajectory types and prognosis. METHODS: A total of 216 AECOPD patients hospitalized between January 2023 and December 2024 were enrolled. CAT and mMRC were used to assess symptom changes at 7 time points from admission to 12 weeks post-discharge. Admission systemic inflammatory biomarkers (CRP and PCT) were collected as biomarkers of inflammatory burden. Sepsis was defined according to Sepsis-3 criteria as life-threatening organ dysfunction caused by a dysregulated host response to infection, identified by an acute SOFA score increase of ≥2 points. Available SOFA components included the respiratory sub-score (derived from PaO2/FiO2) and the coagulation sub-score; hepatic, renal, cardiovascular, and neurological sub-scores were not systematically collected. Accordingly, formal complete SOFA-based Sepsis-3 classification was not feasible; CRP and PCT are used herein as surrogate biomarkers of systemic inflammatory burden alongside partial SOFA data. GBTM was applied using the traj procedure in SAS 9.4 to identify symptom recovery trajectory types, and baseline characteristics and prognostic differences were compared among trajectory groups. Cox proportional hazards regression and Fine-Gray competing risk models were used to analyze the association between trajectory types and first re-exacerbation risk. Negative binomial regression was used to analyze the association between trajectory types and re-exacerbation and rehospitalization frequency. Multivariable logistic regression was used to identify independent predictors of non-rapid recovery trajectories. RESULTS: Four symptom recovery trajectory types were identified: rapid recovery group (82 cases, 38.0%), slow recovery group (62 cases, 28.7%), persistent symptoms group (48 cases, 22.2%), and symptom rebound group (24 cases, 11.1%). Patients in non-rapid recovery groups exhibited significantly higher admission inflammatory biomarker levels (CRP and PCT) compared to the rapid recovery group (all P < 0.05). The median time to first re-exacerbation was 248, 156, 78, and 102 days, respectively (P < 0.001). Multivariable Cox regression showed that, using the rapid recovery group as reference, the slow recovery group (aHR = 1.62, 95%CI: 1.08-2.42), persistent symptoms group (aHR = 2.58, 95%CI: 1.68-3.96), and symptom rebound group (aHR = 2.24, 95%CI: 1.32-3.80) all had elevated first re-exacerbation risk (all P < 0.05). Negative binomial regression showed that non-rapid recovery groups had increased re-exacerbation and rehospitalization counts within 12 months (all P < 0.05). Higher admission CAT score (OR = 1.18), lower FEV1%pred (OR = 0.96), ≥2 exacerbations in the previous year (OR = 2.42), higher admission CRP (OR = 1.01, an admission inflammatory biomarker), and current smoking (OR = 2.15) were independent predictors of non-rapid recovery trajectory (all P < 0.05). CONCLUSION: Symptom trajectories during AECOPD recovery are heterogeneous, and elevated admission systemic inflammatory burden is associated with non-rapid recovery. Non-rapid recovery trajectories are associated with increased re-exacerbation risk and worse quality of life. Trajectory-based prognostic stratification combined with early inflammatory burden assessment may facilitate identification of high-risk patients and development of individualized follow-up management strategies.