FLI estimates the probability of ultrasound-detectable hepatic steatosis in adults using routine anthropometry and blood tests. It is mainly used in primary care, metabolic, endocrine and hepatology settings to stratify steatosis risk in adults with obesity, type 2 diabetes or other cardiometabolic risk factors.

Derived from the Dionysos Nutrition & Liver Study, which enrolled 496 Italian adults (216 with and 280 without suspected liver disease) from a general population sample; hepatic steatosis was diagnosed by liver ultrasound and predictors were selected by bootstrapped logistic regression among anthropometric, biochemical and metabolic variables. The final four-variable model (BMI, waist circumference, triglycerides and GGT) had ROC-AUC about 0.84-0.85 for ultrasound-defined fatty liver.
In the original cohort, FLI <30 had sensitivity about 87% and a negative likelihood ratio around 0.2 for ruling out fatty liver on ultrasound, while FLI >=60 had specificity about 86% and a positive likelihood ratio around 4.3 for ruling in steatosis.
A 2021 meta-analysis of 10 population studies (27,221 adults without secondary causes of fatty liver disease) reported imaging-diagnosed NAFLD in roughly 14% of those with FLI <30, 42% with FLI 30-59 and 67% with FLI >=60. Using 30 and 60 as cut-offs, pooled sensitivity and specificity were approximately 81% and 65% for the lower threshold, and approximately 44% and 90% for the higher threshold, supporting FLI as a risk-stratification tool with only moderate ability to definitively exclude or confirm NAFLD.
European and international guidelines list FLI among the best-validated non-invasive steatosis scores, particularly when ultrasound is impractical at scale. Cost-effectiveness modelling in adults with type 2 diabetes or metabolic syndrome has found FLI more cost-effective than several other non-invasive strategies, but guideline groups highlight variable specificity across studies and generally recommend using FLI to select adults for imaging and cardiometabolic risk management rather than as a stand-alone diagnostic test.

FLI has not been validated in children, pregnancy or acute severe illness. It detects steatosis, not steatohepatitis or fibrosis; separate fibrosis assessment is required in patients with abnormal liver tests, long-standing metabolic risk or other chronic liver disease. Marked GGT elevation from cholestasis, alcohol or medications can raise FLI disproportionately to true liver fat content, and FLI may perform less well in advanced cirrhosis or when alcohol or viral hepatitis are the dominant aetiology.

Fatty Liver Index