Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 27, 2026Frontiers in MedicineOpen Access

A seven-variable model predicts 1-year MACE with ~0.83 AUC in hospitalized T2DM and hypertension patients.

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Patients with coexisting T2DM and hypertension face synergistically elevated MACE risk, but evidence for prediction models developed specifically in this population remains limited.

Does an interpretable machine learning model based on routine clinical variables predict 1-year MACE in hospitalized patients with coexisting T2DM and HTN?

Population

1,054 hospitalized patients with coexisting T2DM and HTN

Comparison

Four algorithms (logistic regression, random forest, support vector machine, and XGBoost)

Design

Retrospective study

Follow-up

1-year

Key result

An interpretable logistic regression model based on seven routine clinical variables predicted 1-year major adverse cardiovascular events with an ROC-AUC of 0.828 in hospitalized patients with coexisting type 2 diabetes and hypertension.

Authors

JLJuan LvXWXirui WangZZZ Zhang

Discussion

Loading...

Member takes

Overview

May inform MACE risk stratification in T2DM-HTN; leaves open prospective validation before clinical use.

Key Points

  • To explore and evaluate a machine learning framework for predicting 1-year MACE in patients with T2DM and HTN.
  • Retrospective analysis of 1,054 hospitalized patients with T2DM and HTN, with 249 experiencing MACE during follow-up.
  • Dataset divided into training (60%), validation (20%), and test cohorts (20%).
  • LASSO regression used for feature selection; models included logistic regression and others, evaluated using various performance metrics.
  • Logistic regression model achieved ROC-AUC of 0.828, PR-AUC of 0.656, and Brier score of 0.116 on the test set.
  • LASSO identified six stable predictors: HbA1c, age, hypertension duration, T2DM duration, CysC, and CIMT.
  • Model predictions linked to factors like glycemic burden and subclinical atherosclerosis using SHAP analysis.

Study Design

Type

Cohort (n=1,054)

Multicenter

No

Structured PICO

Does an interpretable machine learning model based on routine clinical variables predict 1-year MACE in hospitalized patients with coexisting T2DM and HTN?

P
Population
1,054 hospitalized adults aged 18 and older with coexisting type 2 diabetes and hypertension, followed for 1 year to develop and validate a machine learning prediction model for major adverse cardiovascular events.
E
Exposure
Interpretable machine learning framework (logistic regression model based on seven routine clinical variables: HbA1c, age, hypertension duration, cystatin C, T2DM duration, carotid intima-media thickness, and sex)
O
Outcome
1-year composite major adverse cardiovascular events (MACE)composite

Main Result

Effect estimate: ROC-AUC 0.828 (95% CI 0.749-0.895)

An interpretable logistic regression model using seven routine clinical variables demonstrated good internal performance for predicting 1-year MACE risk in patients with coexisting T2DM and hypertension.

Limitations

  • Retrospective single-center design
  • Learning curves indicated limited incremental improvement with increasing training sample size
  • Lack of external validation
  • External validation is required before the model can be considered for clinical decision support

Cite This Study

Lv et al. (2026) conducted a cohort in Type 2 diabetes mellitus and hypertension (n=1,054). Logistic regression prediction model was evaluated on 1-year major adverse cardiovascular events (MACE) (ROC-AUC 0.828, 95% CI 0.749-0.895). An interpretable logistic regression model based on seven routine clinical variables predicted 1-year major adverse cardiovascular events with an ROC-AUC of 0.828 in hospitalized patients with coexisting type 2 diabetes and hypertension.

synapsesocial.com/papers/6a6794fcbb2605d1235fba0dhttps://doi.org/10.3389/fmed.2026.1871693
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Development and validation of a machine learning model to predict comorbid hypertension in patients with type 2 diabetes2026 · 2 citations
  2. 2Development and Validation of a Machine Learning–Based Model for Predicting All-Cause Mortality Risk in Patients With Type 2 Diabetes Mellitus Combined With Hypertension: National Cohort Study2026
  3. 3Development and validation of an interpretable machine learning model for predicting 5-year major adverse cardiovascular events in patients with coronary artery disease2026
  4. 4Development and validation of a model to predict cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke in patients with type 2 diabetes mellitus and established atherosclerotic cardiovascular disease2022 · 12 citations
  5. 5Exploration and analysis of risk factors for coronary artery disease with type 2 diabetes based on SHAP explainable machine learning algorithm2025