# Development and validation of a nomogram for predicting survival in patients with cardiogenic shock

## TL;DR

This study created a tool called the Cardiogenic Shock Survival Nomogram to predict 30-day survival in patients with cardiogenic shock, showing it works well and better than existing methods.

## Contribution

A new nomogram for predicting 30-day survival in cardiogenic shock patients, validated across different subgroups and outperforming existing scores.

## Key findings

- The CSSN achieved a c-statistic of 0.75 in training and 0.73 in validation for 30-day survival prediction.
- The CSSN outperformed the Cardiogenic Shock Score in multiple metrics, including c-statistic and AUC.
- The model showed robust performance in both AMI-CS and non-AMI-CS subgroups.

## Abstract

There is currently a lack of easy-to-use tools for assessing the severity of cardiogenic shock (CS) patients. This study aims to develop a nomogram for evaluating severity in CS patients regardless of the underlying cause.

The MIMIC-IV database was used to identify 1,923 CS patients admitted to the ICU. A multivariate Cox model was developed in the training cohort (70%) based on LASSO regression results. Factors such as age, systolic blood pressure, arterial oxygen saturation, hemoglobin, serum creatinine, blood glucose, arterial pH, arterial lactate, and norepinephrine use were incorporated into the final model. This model was visualized as a Cardiogenic Shock Survival Nomogram (CSSN) to predict 30-day survival rates. The model's c-statistic was 0.75 (95% CI: 0.73–0.77) in the training cohort and 0.73 (95% CI: 0.70–0.77) in the validation cohort, demonstrating good predictive accuracy. The AUC of the CSSN for 30-day survival probabilities was 0.76 in the training cohort and 0.73 in the validation cohort. Calibration plots showed strong concordance between predicted and actual survival rates, and decision curve analysis (DCA) affirmed the model's clinical utility. The CSSN outperformed the Cardiogenic Shock Score (CSS) in various metrics, including c-statistic, time-dependent ROC, calibration plots, and DCA (c-statistic: 0.75 vs. 0.72; AUC: 0.76 vs. 0.73, P < 0.01 by Delong test). Subgroup analysis confirmed the model's robustness across both AMI-CS and non-AMI-CS subgroups.

The CSSN was developed to predict 30-day survival rates in CS patients irrespective of the underlying cause, showing good performance and potential clinical utility in managing CS.