Jiang Y;  School of Medicine, Tongji University, Shanghai, 200065, China.
Shen T; Shi C; Li D; Zhan M; Li G; Qian L; Huang Q; Zheng L;
CheL; Wang L; Shen Y

BMC Cardiovascular Disorders. 26(1):91, 2025 Dec 27.
VI 1

BACKGROUND: Acute coronary syndrome (ACS) is a major global health burden
with a high risk of adverse outcomes. Existing predictive models (e.g.,
GRACE) primarily rely on static indicators and focus on short-term
prognosis, limiting their ability to comprehensively assess patient status
and predict long-term mortality. To address the need for improved
long-term risk prediction in this specific patient subgroup, this study
developed and validated a long-term mortality prediction model for ACS
patients incorporating cardiopulmonary exercise testing (CPET) and other
clinical indicators.

METHODS: This retrospective cohort study included ACS patients treated at
Tongji Hospital from January 1, 2007, to December 31, 2018. Demographic
data, medical histories, CPET indicators, laboratory indicators, and other
baseline data were collected, and all-cause mortality was followed up
until June 30, 2023. All data sets were randomly divided into derivation
and validation cohorts in a ratio of 7/3. Least absolute shrinkage and
selection operator regression and Cox multivariate analysis were used to
identify independent risk factors and a risk prediction model was
established using nomograms.

RESULTS: A total of 299 patients were included in this cohort (211 in the
derivation cohort and 88 in the validation cohort), with an average age of
57.00 years, including 280 males (93.6%). The median follow-up time was
3821 days, and 46 cases (15.4%) reached the study endpoint. The derivation
cohort identified four independent predictive factors: age, blood urea
nitrogen (BUN), ejection fraction (EF), and heart rate reserve (HRR), and
a Nomogram scoring model was constructed based on these factors. The model
demonstrated good discrimination in the derivation cohort (C-index: 0.83)
but this decreased in the validation cohort (C-index: 0.72), suggesting
potential overfitting. Time-dependent calibration analysis showed poor
agreement at 5 years in the validation cohort (R2 = 0.1819), but improved
at 10 years (Slope = 0.8006, R2 = 0.5575) and 15 years (R2 = 0.5638). The
model’s applicability is strictly limited to the studied population: a
predominantly male, lower-risk subset of ACS survivors capable of
completing CPET.

CONCLUSIONS: A model based on four readily available variables-age, BUN,
EF, and the key CPET parameter, HRR-may have utility for predicting
long-term all-cause mortality. This model provides a preliminary tool for
the long-term management of a specific subpopulation of acute coronary
syndrome (ACS) survivors, namely a predominantly male, lower-risk cohort
with the capacity to complete CPET. Further external validation in similar
populations is required before prospective clinical application.