Association between the Endothelial Activation and Stress Index and all-cause mortality in patients with chronic...

Endothelial Activation and Stress Index (EASIX) predicts outcome in diverse disorders; however, its independent association with short- to medium-term mortality in critically ill patients with chronic obstructive pulmonary disease (COPD) has not been examined.

Data were drawn from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database and the Critical care datebase of Zigong Municipal Hospital, Sichuan, China. We identified consecutive COPD patients admitted between 2008 and 2019 in MIMIC-IV and between 2019 and 2020 in the Zigong cohort. EASIX was calculated from platelet count, creatinine, and lactate dehydrogenase measured within 24 h of ICU entry and log₂-transformed. Multivariable Cox regression and restricted cubic splines assessed its relationship with 28-day and 90-day mortality after ICU admission; proportional hazards assumptions were verified. Subgroup analyses and interaction tests evaluated robustness. Discrimination was quantified by area under the receiver-operating-characteristic curve (AUROC).

We included 5,012 critically ill COPD patients from MIMIC-IV and 431 from the Zigong registry. In fully-adjusted models treating log₂(EASIX) as a continuous variable, higher values were associated with increased 28- and 90-day mortality: MIMIC-IV: HR = 1.23 (95% CI: 1.17-1.28, p < 0.001) and HR = 1.18 (95% CI: 1.14-1.22, p < 0.001); Zigong: HR = 1.32 (95% CI: 1.17-1.50, p < 0.001) and HR = 1.32 (95% CI: 1.17-1.48, p < 0.001). After dividing log₂(EASIX) into tertiles, we found that patients in the highest tertile faced significantly greater mortality than those in the lowest: in MIMIC-IV the 28-day hazard ratio was 1.66 (95% CI 1.39-1.98, p < 0.001) and the 90-day HR was 1.55 (1.33-1.80, p < 0.001), while in the Zigong cohort the corresponding estimates were 2.38 (1.41-4.00, p = 0.001) and 2.37 (1.47-3.81, p < 0.001). Log₂(EASIX) showed a non-linear association with both 28- and 90-day mortality. In MIMIC-IV the inflection point lay at approximately 3.49; below this breakpoint the hazard ratio rose modestly (28-day HR = 1.168, 95% CI 1.100-1.239, p < 0.001; 90-day HR = 1.144, 1.088-1.202, p < 0.001), whereas above the breakpoint the risk increased sharply (28-day HR = 1.805, 1.367-2.383, p < 0.001; 90-day HR = 1.597, 1.219-2.092, p < 0.001). In the Zigong cohort the corresponding breakpoint was located between 3.42 and 3.70. Subgroups and interaction testing showed the log₂(EASIX)-mortality link was stable. ROC analysis showed that log₂(EASIX) predicts 28- and 90-day death with good accuracy, yielding AUROCs of 0.80 and 0.79 in MIMIC-IV and 0.83 and 0.83 in the Zigong cohort, respectively.

Elevated log₂(EASIX) independently predicts short- and medium-term mortality in critically ill patients with COPD.