ADHERE Tree Model for Acute HF Mortality Calculator

Clinical calculation support · Verify results before clinical use

Admission Variables

Enter the patient's admission laboratory values and systolic blood pressure to identify the ADHERE CART risk group.

Unit: mg/dL. ADHERE CART threshold: 43 mg/dL.
Unit: mmHg. ADHERE CART threshold: 115 mmHg.
Unit: mg/dL. The CART checks creatinine ≥2.75 mg/dL only when BUN ≥43 mg/dL and SBP <115 mmHg.

About the ADHERE CART model

The ADHERE Classification and Regression Tree (CART) was developed to stratify in-hospital mortality risk among patients hospitalized with acute decompensated heart failure using admission variables.

The tree first splits on BUN, then on systolic blood pressure; serum creatinine is evaluated in the branch with BUN ≥43 mg/dL and SBP <115 mmHg.

Risk groups in the original derivation

Risk groupCriteriaMortality
Low riskBUN <43; SBP ≥1152.1%
Intermediate risk 3BUN <43; SBP <1154.9%
Intermediate risk 2BUN ≥43; SBP ≥1155.6%
Intermediate risk 1BUN ≥43; SBP <115; creatinine <2.7513.2%
High riskBUN ≥43; SBP <115; creatinine ≥2.7521.9%

Clinical use and limitations

Use the model as an adjunct for bedside prognostic stratification, not as a stand-alone decision rule for treatment limitation, discharge, or escalation. Absolute mortality may differ by case mix, treatment era, and healthcare setting; local calibration may be necessary before using these percentages operationally.

References

Fonarow GC, Adams KF Jr, Abraham WT, Yancy CW, Boscardin WJ; ADHERE Scientific Advisory Committee, Study Group, and Investigators. Risk stratification for in-hospital mortality in acutely decompensated heart failure: classification and regression tree analysis. JAMA. 2005;293(5):572–580. DOI: 10.1001/jama.293.5.572. PMID: 15687312.
Risk stratification of acute decompensated heart failure in-hospital mortality using the ADHERE CART method: is there validation in a Brazilian cohort? External cohort evaluation. The study found that only BUN significantly stratified mortality in its small cohort, underscoring that performance may vary across populations. Read the article.