EUVIMEDEUVIMED entdecken
KohortenstudieOpen Access · CC BY-NC 3.0

Development and Assessment of a Risk Scoring System for Cardiovascular Adverse Events in Patients on Maintenance Hemodialysis: A Retrospective Cohort Study

Li Yin, Dongdong Xiang

Risk Management and Healthcare Policy · 2026-11 · Volume 19 · 1-11

DOI: 10.2147/rmhp.s653112 ↗

Kurz gefasst

Kohortenstudie in der Zeitschrift Risk Management and Healthcare Policy, 2026-11. Verfasst von Li Yin und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Nephrology, Ningguo People's Hospital, Ningguo, Anhui, 242300, People's Republic of China. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 3.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Cardiovascular adverse events (CVAEs) are a major clinical concern in patients receiving maintenance hemodialysis (MHD). Their prediction using routine measurements requires evaluation in a clearly defined outcome window.

Objective

To develop and internally evaluate a preliminary four-variable model and simplified score for 12-month CVAEs in patients receiving MHD.

Methods

This single-center retrospective cohort included 251 patients receiving MHD. Baseline variables were compared between patients with (n = 71) and without (n = 180) a qualifying CVAE within 12 months of the index assessment. The previously reported model incorporating dichotomized age, NT-proBNP, potassium, and hemoglobin was refitted. Apparent discrimination and probabilistic accuracy were assessed. Internal validation used 1000 bootstrap resamples, retaining the predictors and cutoffs while re-estimating the intercept and coefficients.

Results

The 12-month CVAE proportion was 28.3%. NT-proBNP > 5000 pg/mL (OR = 2.997, 95% CI: 1.423– 6.312), potassium > 5.5 mmol/L (OR = 8.055, 95% CI: 3.797– 17.088), and hemoglobin < 100 g/L (OR = 2.979, 95% CI: 1.266– 7.011) were associated with CVAEs in the four-variable model; the age association was not statistically significant. The simplified score ranged from 0 to 237. The regression model and score both achieved an apparent AUC of 0.8599. Bootstrap evaluation of the regression model yielded an optimism-corrected AUC of 0.8504, Brier score of 0.1234, and calibration slope of 0.9381.

Conclusion

Elevated NT-proBNP, hyperkalemia, and low hemoglobin were associated with 12-month CVAEs in this MHD cohort. The preliminary four-variable model showed discrimination in the development cohort and underwent internal evaluation conditional on its fixed specification. Independent external validation and assessment of added clinical value are required before implementation. Keywords: maintenance hemodialysis, cardiovascular adverse events, logistic regression, risk scoring model, NT-proBNP, serum potassium, hemoglobin

Schlagwörter

Stichwörter der Autor:innen

NT-proBNPcardiovascular adverse eventshemoglobinlogistic regressionmaintenance hemodialysisrisk scoring modelserum potassium

Autor:innen und Institutionen

  • Department of Nephrology, Ningguo People's Hospital, Ningguo, Anhui, 242300, People's Republic of China.

Zitieren

APA

Yin, L., & Xiang, D. (2026). Development and Assessment of a Risk Scoring System for Cardiovascular Adverse Events in Patients on Maintenance Hemodialysis: A Retrospective Cohort Study. Risk Management and Healthcare Policy, Volume 19, 1-11. https://doi.org/10.2147/rmhp.s653112

Vancouver

Yin L, Xiang D. Development and Assessment of a Risk Scoring System for Cardiovascular Adverse Events in Patients on Maintenance Hemodialysis: A Retrospective Cohort Study. Risk Management and Healthcare Policy. 2026;Volume 19:1-11. doi:10.2147/rmhp.s653112

BibTeX
@article{Yin2026,
  author = {Yin, Li and Xiang, Dongdong},
  title = {Development and Assessment of a Risk Scoring System for Cardiovascular Adverse Events in Patients on Maintenance Hemodialysis: A Retrospective Cohort Study},
  journal = {Risk Management and Healthcare Policy},
  year = {2026},
  volume = {Volume 19},
  pages = {1--11},
  doi = {10.2147/rmhp.s653112}
}
Kostenlos starten →