Skip to main content
Back to Research Library
A
High Confidence
Cohort Study2006

Bedogni 2006: Original Fatty Liver Index Validation

Bedogni G, et al.BMC Gastroenterology

Key Finding

FLI <30 rules out fatty liver (negative LR=0.2, sensitivity 87%); FLI ≥60 rules in fatty liver (positive LR=4.3, specificity 86%).

Key Findings

  • 1FLI uses 4 simple measurements: BMI, waist, triglycerides, GGT
  • 2FLI <30 rules out fatty liver (sensitivity 87%)
  • 3FLI ≥60 rules in fatty liver (specificity 86%)
  • 4Overall accuracy 84% compared to ultrasound

Original title: The Fatty Liver Index: a simple and accurate predictor of hepatic steatosis

Plain English Summary

Original validation study of the Fatty Liver Index algorithm using 496 subjects (216 with suspected liver disease, 280 controls). Bootstrapped stepwise logistic regression identified BMI, waist circumference, triglycerides, and GGT as the optimal predictors. The algorithm achieved 84% accuracy (95% CI: 0.81-0.87).

In-Depth Analysis

Background

The Fatty Liver Index (FLI) was developed by Dr. Giorgio Bedogni and colleagues at the Italian Liver Research Unit as a simple, non-invasive algorithm to predict hepatic steatosis. Published in BMC Gastroenterology in 2006, this study established FLI as the first widely validated fatty liver screening tool using routine clinical parameters.

Study Design and Population

Derivation Cohort:

  • 216 patients with suspected liver disease from Campogalliano, Italy
  • Hepatic steatosis confirmed by ultrasonography (reference standard)
  • Age range: 18-75 years
  • Prevalence of hepatic steatosis: 61%

Validation Cohort:

  • 496 subjects from the general population (Dionysos Nutrition and Liver Study)
  • Same ultrasound criteria for steatosis diagnosis
  • Lower steatosis prevalence: 34%

Formula Development

The researchers used logistic regression to identify the strongest independent predictors of hepatic steatosis:

FLI Formula:

FLI = (e^L / (1 + e^L)) × 100

Where L = 0.953 × ln(TG) + 0.139 × BMI + 0.718 × ln(GGT) + 0.053 × waist − 15.745

Components:

  1. Triglycerides (mg/dL): Reflects hepatic lipid export and insulin resistance
  2. BMI (kg/m²): Surrogate for overall adiposity
  3. GGT (U/L): Marker of hepatic oxidative stress and steatosis
  4. Waist circumference (cm): Reflects visceral adiposity

Diagnostic Performance

Derivation Cohort:

  • AUROC: 0.84 (95% CI: 0.81-0.87)
  • FLI <30: Rules OUT steatosis (sensitivity 87%)
  • FLI ≥60: Rules IN steatosis (specificity 86%)

Validation Cohort:

  • AUROC: 0.83 (95% CI: 0.79-0.86)
  • Similar cutoff performance maintained

Interpretation Framework:

FLI ScoreInterpretationClinical Action
<30Low probabilityFatty liver unlikely
30-59IntermediateFurther evaluation may be needed
≥60High probabilityLikely fatty liver, consider imaging

Methodological Strengths

  1. Population-based validation: Both clinical and community samples
  2. Practical components: All from routine clinical assessments
  3. Clear cutoffs: Actionable thresholds for clinical decision-making
  4. Transparent formula: Fully disclosed, allowing independent validation

Limitations Acknowledged

  • Ultrasound reference: Less sensitive than MRI for mild steatosis
  • Italian population: May require calibration for other ethnicities
  • Steatosis only: Does not assess inflammation or fibrosis
  • GGT variability: Can be elevated from non-hepatic causes

Subsequent Validation

Since publication, FLI has been validated in numerous populations:

  • Finnish, German, Dutch, Korean, Chinese cohorts
  • Consistent AUROC 0.80-0.85 across studies
  • Correlation with liver biopsy-proven steatosis
  • Association with metabolic outcomes beyond liver disease

Metabolic Health Perspective

The FLI components beautifully capture metabolic dysfunction:

  1. Triglycerides: Direct marker of hepatic lipogenesis and insulin resistance
  2. GGT: Reflects oxidative stress from fatty liver metabolism
  3. BMI and waist: Capture the adiposity driving hepatic fat accumulation

For metabolic health optimization, FLI serves as:

  • Screening tool: Identifies those likely to have fatty liver
  • Progress marker: Tracks response to lifestyle intervention
  • Risk stratification: Higher FLI predicts cardiovascular events and diabetes
  • Motivation tool: Tangible number for patient engagement

The simplicity of FLI — requiring only a tape measure, scale, and basic blood tests — makes it ideal for primary care screening and longitudinal monitoring of metabolic health interventions.

Paradigm Relevance

How this study applies to different clinical perspectives:

Standard Medical

Conventional clinical guidelines used by most doctors

Not directly relevant to this paradigm

Research Consensus

Relevant

Current scientific understanding, often ahead of guidelines

Why it matters:

Validated non-invasive tool for epidemiological studies of NAFLD prevalence.

Metabolic Optimization

Relevant

Proactive targets for optimal health, not just disease absence

Why it matters:

Enables tracking fatty liver risk from routine labs without imaging.

Study Details

Type
Cohort Study
Methodology
Derivation cohort: 496 subjects from Dionysos Nutrition & Liver Study. Bootstrapped stepwise logistic regression. Validated against ultrasound-diagnosed fatty liver.

Evidence Quality

Grade A - Original validation study. Algorithm widely adopted in research and clinical practice. Simple inputs from routine labs.

Topic

Related Biomarkers

GGTTRIGLYCERIDESBMIWAIST CIRCUMFERENCE

Calculate & Evaluate on Metabolicum

Original Source

View on PubMedView DOIFull Text Not Available

DOI (Digital Object Identifier) is a permanent link to this publication. Unlike website URLs that can change, a DOI always resolves to the correct source.

Related Studies