# Diabetes and obesity reduce FIB-4 accuracy in MASLD referral pathways

## TL;DR

The study finds that FIB-4, a blood test used to assess liver fibrosis, is less accurate in people with diabetes or obesity, leading to under-triaging for further liver testing.

## Contribution

The study identifies that diabetes and obesity significantly reduce the accuracy of FIB-4 in predicting liver fibrosis in MASLD patients.

## Key findings

- FIB-4 under-triages patients with diabetes or BMI ≥30, increasing the risk of ≥8 kPa liver stiffness even at low FIB-4 scores.
- BMI and diabetes strongly modify FIB-4 performance, more than sex, ALT, or age.
- 2D-SWE thresholds align with VCTE cut-offs, supporting harmonized elastography use across modalities.

## Abstract

Primary care referral pathways often use FIB-4 to triage metabolic dysfunction-associated steatotic liver disease (MASLD), but its accuracy may vary by patient characteristics. We aimed to evaluate FIB-4 performance against elastography, compare it with other non-invasive tests, assess effect modification by BMI, diabetes, sex, ALT, and age, and calibrate 2D shear-wave elastography (2D-SWE) and vibration-controlled transient elastography (VCTE) thresholds.

We analyzed two cohorts with paired serum scores and elastography: Calgary (2D-SWE; N = 8,126) and Edmonton (VCTE; N = 985). We summarized fibrosis-risk distributions and used exceedance probabilities to align 2D-SWE with guideline-recommended VCTE cut-offs.

In the Calgary cohort, median 2D-SWE was 4.6 kPa (IQR 3.9–5.8) and 9% had liver stiffness ≥8 kPa; FIB-4 was <1.30 in 70% and ≥2.67 in 5%. In the Edmonton cohort, median VCTE was 5.3 kPa (4.4–6.6) and 14% had liver stiffness ≥8 kPa; FIB-4 was <1.30 in 77% and ≥2.67 in 3%. In exceedance models, at a FIB-4 of 1.30, the probability of liver stiffness ≥8 kPa was significantly higher in individuals with diabetes or BMI ≥30 kg/m2 than in those without these risk factors. Regression analyses showed that BMI and diabetes materially increased the probability of liver stiffness ≥8 kPa at a given FIB-4, whereas sex, elevated ALT, and age had smaller effects. Using the recommended 2D-SWE thresholds of 9, 13, 17 kPa, we observed an expected stepwise correspondence with VCTE 15, 20, 25 kPa, supporting harmonised elastography cut-offs across modalities.

FIB-4 alone under-triages patients with diabetes and/or obesity. These patients should be fast-tracked to elastography even when FIB-4 is <1.30. We recommend validating these findings in other cohorts, as they may affect triaging practices.

Recent guidelines recommend clinical care pathways for risk stratification of patients with MASLD (metabolic dysfunction-associated steatotic liver disease) using sensitive serum-based markers (e.g. fibrosis 4 index [FIB-4]) as a first step, with a cut-off of <1.30 to rule out advanced fibrosis. In patients with diabetes or obesity, the interpretation of a FIB-4 threshold of <1.30 is different than in patients without these conditions. This finding suggests the need to refine referral pathways in which FIB-4 is used as a first test. New referral models incorporating patient characteristics could improve risk stratification of patients with MASLD who need specialized liver care.

Image 1

•In patients with MASLD, FIB-4 under-triages those with diabetes or obesity.•Diabetes/BMI ≥30 increase the risk of ≥8 kPa stiffness even at FIB-4 <1.30.•BMI and diabetes strongly modify FIB-4 performance; sex, ALT, and age have smaller effects.•2D-SWE thresholds align with VCTE cut-offs, supporting harmonized elastography use.

In patients with MASLD, FIB-4 under-triages those with diabetes or obesity.

Diabetes/BMI ≥30 increase the risk of ≥8 kPa stiffness even at FIB-4 <1.30.

BMI and diabetes strongly modify FIB-4 performance; sex, ALT, and age have smaller effects.

2D-SWE thresholds align with VCTE cut-offs, supporting harmonized elastography use.