Background and aims: Multistep algorithms for noninvasive identification of patients with metabolic dysfunction-associated steatotic liver disease (MASLD), such as the recent American Association for the Study of Liver Disease and European Association for the Study of the Liver MASLD guidelines aim to identify patients who require hepatology referral. We evaluated the performance of these algorithms in the National Health and Nutrition Examination Survey cohort.

Methods: We analyzed 7768 adult participants from the National Health and Nutrition Examination Survey 2017-2020 with valid liver stiffness measurement (LSM) data. MASLD was defined as hepatic steatosis (controlled attenuation parameter measurement of ≥ 248 dB/m) with ≥1 cardiometabolic risk factor; significant fibrosis and advanced fibrosis were defined as LSM ≥8 kPa and ≥12 kPa respectively (with sensitivity analysis of 0-30kPa and using the FibroScan-aspartate aminotransferase or Agile algorithms).

Results: A total of 44.8% (n = 3479) met metabolic risk criteria, 7.5% (n = 586) met raised Fibrosis-4 Index criteria, and 3.0% (n = 235) met referral criteria. The multistep pathway demonstrated high specificity but limited sensitivity for identifying MASLD with significant fibrosis (sensitivity 21%, 95% confidence interval [CI] 17%-25%) or advanced fibrosis (29%, 95% CI 23%-37%). Sensitivity was particularly low in key subgroups: age < 50 years (6%, 95% CI 3%-11%), women (15%, 95% CI 11%-21%) and patients without type 2 diabetes (16%, 95% CI 12%-20%). Sensitivity improved when targeting more advanced disease: 29% for advanced fibrosis (LSM ≥12 kPa), 54% for FibroScan-aspartate aminotransferase ≥0.67, 35% for Agile 3+ ≥0.68, and 66% for Agile 4 ≥ 0.57.

Conclusion: A multistep screening strategy based on metabolic risk criteria, Fibrosis-4 Index, and LSM demonstrates high specificity but low sensitivity for detecting MASLD with significant fibrosis in a US-based populational cohort.