A simplified nomogram for predicting deep vein thrombosis: development in NHANES and retrospective single-center...

Area of Science:

• Cardiovascular Medicine
• Epidemiology
• Biostatistics

Background:

• Deep vein thrombosis (DVT) poses a significant cardiovascular risk.
• Current risk assessment tools often require laboratory biomarkers, limiting primary care application.
• A simplified, laboratory-free nomogram was developed to predict DVT in community settings.

Purpose of the Study:

• To develop and validate a simplified, laboratory-free nomogram for predicting deep vein thrombosis (DVT) risk.
• To enhance DVT risk stratification in primary care settings.
• To provide a tool for identifying adults at higher risk of DVT without laboratory tests.

Main Methods:

• The nomogram was developed using a large cohort (NHANES, 39,692 participants) and validated in a separate cohort (3,136 hospitalized Chinese patients).
• Predictors were selected using LASSO regression and fitted with multivariable logistic regression, accounting for complex survey design.
• Model performance was evaluated using Area Under the Curve (AUC), Mean Absolute Error (MAE), and Decision Curve Analysis (DCA).

Main Results:

• A four-variable nomogram including age, smoking, physical activity, and diabetes was developed.
• The nomogram demonstrated good discrimination in development (AUC, 0.816) and external validation (AUC, 0.792).
• Diabetes and ever-smoking were identified as the strongest predictors of DVT risk.

Conclusions:

• The four-variable nomogram can aid in identifying adults at higher risk for DVT in primary care.
• This tool is intended as an exploratory screening aid, not a diagnostic test for acute DVT.
• Prospective validation is necessary prior to widespread clinical implementation.