# Nomogram for Risk Prediction of Mortality for Patients with Critical Cardiovascular Disease Treated by Continuous Renal Replacement Therapy in Coronary Care Unit

## TL;DR

This study creates a risk prediction tool to estimate mortality in patients with severe heart disease undergoing kidney therapy in a coronary care unit.

## Contribution

A novel nomogram model is developed using nine clinical predictors to assess mortality risk in CRRT-treated cardiovascular patients.

## Key findings

- Nine clinical factors were identified as significant predictors of mortality in patients undergoing CRRT in the CCU.
- The nomogram demonstrated strong predictive accuracy with a C-index of 0.902 and good calibration.
- The model provides a reliable and practical tool for risk stratification in critical cardiovascular patients.

## Abstract

To establish a nomogram-scoring model for evaluating the risk of
death in patients with critical cardiovascular disease after continuous renal
replacement therapy (CRRT) in a coronary care unit (CCU).

This
retrospective cohort study included data collected on 172 patients, in whom CRRT
was initiated in the CCU between January 2017 and June 2021. Predictors of
mortality were selected using an adaptive least absolute shrinkage and selection
operator logistic model and used to construct a nomogram. The nomogram was
evaluated using the concordance index (C-index) and Hosmer–Lemeshow test.

The number of patients who died in-hospital after CRRT was 91
(52.9%). The results of the multivariate logistic regression analyses clarified
that age, history of hypertension and/or coronary artery bypass grafting, a
diagnosis of unstable angina pectoris or acute myocardial infarction, ejection
fraction, systolic blood pressure, creatinine, neutrophil, and platelet counts
before CRRT initiation were significant predictors of early mortality in patients
treated with CRRT. The nomogram constructed on these predictors demonstrated
significant discriminative power with an unadjusted C-index of 0.902 (95% CI:
0.858–0.945) and a bootstrap-corrected C-index of 0.875. Visual inspection
showed a good agreement between actual and predicted probabilities
(Hosmer–Lemeshow χ2 = 5.032, p-value = 0.754).

Our nomogram based on nine readily available predictors is
a reliable and convenient tool for identifying critical patients undergoing CRRT
at risk of mortality in the CCU.