Obesity Subtype Calculator

Clinical Measurements

mmHg
optional
optional
optional
optional
U/L
cm
cm
ratio

Please complete all required fields above. WHR will be calculated automatically.

Note: This calculator is designed for obesity subtyping among individuals with a BMI ≥ 30 kg/m². If your BMI is unknown, please use the BMI calculator below. The model uses multinomial ridge regression trained on UK Biobank data (N=88,877). All input data is processed locally and never uploaded.

BMI Calculator

cm
kg
kg/m²

Batch CSV Prediction

Upload a CSV with ID and seven raw UKB-unit indicators. The browser will generate a two-column result CSV locally.

Input: ID, SBP, HbA1c, TG, HDL-C, Creatinine, ALT, WHR

Output: ID, subtype

No CSV selected.

Methodology

Calculation Steps

  1. 1.Z-score normalization: z = (x − μ) / σ using UKB population parameters
  2. 2.Linear scoring: scoreₖ = β₀ₖ + Σ βᵢₖ · zᵢ
  3. 3.Softmax: P(Cₖ) = exp(scoreₖ) / Σ exp(scoreₗ)
  4. 4.Predicted subtype: argmax P(Cₖ)

Model Information

  • Model: Multinomial Ridge Regression
  • Training: 88,877 UK Biobank participants
  • Features: 7 routine clinical variables
  • Validation: 5-fold CV, Macro-AUC ≈ 0.89
  • Privacy: All computation runs locally in the browser
Ridge Regression Coefficients

This model uses multinomial ridge regression (L2-regularized logistic regression). The L2 penalty shrinks extreme coefficients, improving generalizability. The optimal regularization parameter C was selected via 5-fold cross-validation on the training set (80% of UK Biobank data).

SubtypeInterceptSBPHbA1cTGHDL-CCreat.ALTWHR
LRO57.95-12.03-36.61-29.20-36.22-38.03-36.85-59.00
RDO82.515.62-30.49-14.97-22.6934.37-20.9620.39
HGO-135.881.99130.828.31-13.29-10.917.8018.64
HHO-19.34-1.90-36.16-25.7495.4113.91-1.786.79
DHO14.766.32-27.5761.60-23.210.6651.7813.18

Interpretation: A positive β means the feature increases the log-odds of that subtype. HDL-C's large positive coefficient (95.41) for HHO reflects this subtype's defining characteristic of exceptionally high HDL cholesterol.

License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND). By accessing the material through this website, you are confirming that you agree that your use of the material is subject to these licence terms.

Please note that the use of this webtool is limited to personal, non-commercial use.

You are free to:

  • Share — copy and redistribute the material in any medium or format

The licensor cannot revoke these freedoms as long as you follow the license terms.

Under the following terms:

  • BY = Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
  • NC = NonCommercial — You may not use the material for commercial purposes.
  • ND = NoDerivatives — If you remix, transform, or build upon the material, you may not distribute the modified material.

For the full licence terms, please visit: https://creativecommons.org/licenses/by-nc-nd/4.0/.

This webtool is intended for research use only and is not designed or validated for clinical diagnosis, treatment decisions or individual medical risk assessment. If you have any health concerns or clinical conditions, please consult a qualified healthcare professional and follow medical advice.

Example CSV

Obesity Subtype Definitions

LRO — Low-risk obesity
Features the most favorable metabolic profile among the subtypes. Biomarkers such as HbA1c, lipids, and liver/renal enzymes remain within or close to normal physiological ranges, indicating a "metabolically healthy" obesity phenotype with lower immediate cardiometabolic risk.

RDO — Renal-dominant obesity
Primarily characterized by elevated serum creatinine levels relative to other clusters. This phenotype points towards early renal hemodynamic alterations, suggesting a higher baseline susceptibility to obesity-related nephropathy or chronic kidney disease (CKD).

HGO — Hyperglycaemic obesity
Defined by markedly elevated HbA1c levels, reflecting severe insulin resistance and poor glycemic control. Patients in this cluster have the highest clinical overlap with prediabetes and overt type 2 diabetes mellitus.

HHO — High-HDL obesity
A distinct phenotypic cluster maintaining paradoxically elevated levels of High-Density Lipoprotein Cholesterol (HDL-C) despite an obese state. While typical obesity is associated with low HDL-C, this subtype requires context-specific risk assessment.

DHO — Dyslipidaemic–hepatic obesity
Characterised by elevated triglycerides (TG) and Alanine Transaminase (ALT), strongly suggesting hepatic involvement (such as MASLD/NAFLD) alongside systemic dyslipidaemia. HbA1c and blood pressure tend to be intermediate.