Circulating stem/progenitor cell biomarkers and prognosis in critically ill patients with gastrointestinal bleeding

Abstract

BACKGROUND

Gastrointestinal bleeding (GIB) is a common critical illness, especially in patients in the intensive care unit (ICU), and is often complicated by shock, multiple organ dysfunction, and serious infections, resulting in high mortality. At present, clinical practice mainly relies on the Glasgow-Blatchford, albumin, international normalized ratio, mental status, systolic blood pressure, age ≥ 65 years (AIMS65), and Rockall scoring systems, combined with lactic acid and organ function indicators, to evaluate the condition and prognosis. However, the predictive value of these traditional indicators for short-term mortality in critically ill patients with GIB remains limited, and additional biomarkers are needed to complement, rather than replace, established clinical assessments.

AIM

To explore associations between circulating stem/progenitor cell biomarkers and prognosis in critically ill GIB patients and develop a prognostic model.

METHODS

Overall, 140 GIB patients admitted to the ICU of Jiashan First People’s Hospital (January 2020 and December 2025) were retrospectively included. Flow cytometry analyzed peripheral blood samples within 6 hours of ICU admission, including CD34+ cells, endothelial progenitor cell (EPC) (CD34+KDR+), and EPC (CD34+ CD133+KDR+) subsets. The primary outcome was 28-day all-cause death, and secondary outcomes included 7 days rebleeding, 90-day death and ICU stay. Multivariate logistic regression analyzed factors associated with 28-day all-cause death, and receiver operating characteristic curve analysis evaluated model discrimination.

RESULTS

Twenty-nine (20.71%) patients died within 28 days and 29 (20.71%) experienced rebleeding within 7 days. Patients in the death group were more critically ill, manifested as higher Sequential Organ Failure Assessment score (10.21 ± 2.15 vs 7.96 ± 3.00, P = 0.001), AIMS65 score [4.00 (3.00, 4.00) vs 3.00 (2.00, 3.00), P = 0.008], proportion of cirrhosis [18 (62.07%) vs 27 (24.32%), P < 0.001], and lactic acid levels [4.20 (3.05, 5.35) mmol/L vs 2.95 (2.28, 3.55) mmol/L, P = 0.003]. Among the circulating progenitor cells, lnEPC (CD34+KDR+) levels were significantly lower in the death group [0.85 (0.65, 1.15) vs 1.35 (1.08, 1.60), P < 0.001]. After adjusting for Sequential Organ Failure Assessment score, lactic acid level, AIMS65 score, and cirrhosis, lnEPCs (CD34+KDR+) remained associated with 28-day mortality (odds ratio = 0.557, 95% confidence interval: 0.353-0.879, P = 0.012). The combined model showed moderate discrimination (area under the curve = 0.727); at the optimal predicted-probability cutoff of 0.360, the sensitivity and specificity were 0.483 and 0.937, respectively, with acceptable calibration on the Hosmer-Lemeshow test (P = 0.622).

CONCLUSION

A lower early EPC (CD34+KDR+) level was associated with increased short-term mortality in critically ill patients with GIB. The combined EPC-clinical model may complement organ dysfunction assessment and aid supplementary risk stratification.

Keywords: Critical illness; Gastrointestinal bleeding; Circulating stem/progenitor cells; Endothelial progenitor cells; Flow cytometry; Prognostic model

Core Tip: In this retrospective of intensive care unit patients with gastrointestinal bleeding, early circulating endothelial progenitor cell (CD34+KDR+) levels were lower in patients who died within 28 days and remained associated with mortality after adjusting for clinical variables. The endothelial progenitor cell-clinical model showed moderate discrimination and high specificity, but limited sensitivity; therefore, it should not be interpreted as a standalone early warning tool for high-risk screening. Its current value is mainly hypothesis-generating and may complement organ dysfunction scores and help identify patients with relatively low predicted risk, pending external validation and standardized flow cytometry workflows.