METHODS: In a 16-center multinational cohort, single-term fractional-polynomial logistic regression models derived CLA3+ and CLA4 probabilities. Prognostic performance was assessed using time-dependent area under the receiver operating characteristic curve (AUROC), Brier score, integrated discrimination improvement (IDI), and serial-change analyses.
RESULTS: The derivation cohort included 2,811 biopsied patients; validation included 11,196 patients with baseline VCTE, of whom 8,046 had serial assessments. Derivation AUROCs were 0.84 for CLA3+ and 0.89 for CLA4. In validation, both achieved the highest integrated AUROC (0.83) and lowest integrated Brier scores (0.0078-0.0081) among non-invasive tests, with good calibration. CLA3+ classified 83.1%/10.8%/6.1% as low/intermediate/high risk (5-year LRE 0.4%/1.5%/11.8%; sensitivity 0.72), whereas CLA4 classified 92.0%/3.6%/4.4% (0.5%/4.7%/13.0%; specificity 0.93), with the smallest intermediate-risk groups. IDI favored CLA3+ over Agile3+, whereas CLA4 performed similarly to Agile4. Serial risk categories remained stable in 86.7% for CLA3+ and 93.2% for CLA4, while ≥10-30% decreases were associated with lower LRE risk. Two-step testing reduced intermediate classifications to 3.9% with FIB-4→CLA3+ and 1.4% with FIB-4→CLA4, with a 5-year LRE rate of 8.3% in the rule-in group.
CONCLUSIONS: CLA3+ and CLA4 provide well-calibrated estimates for advanced fibrosis and cirrhosis using controlled attenuation parameter, liver stiffness, and alanine aminotransferase, with small intermediate-risk groups and prognostically informative serial changes. They may serve as second-line tests within Fibrosis-4 index-based two-step pathways.
IMPACT AND IMPLICATIONS: Current MASLD care pathways use sequential non-invasive testing to identify patients at risk of advanced fibrosis and liver-related events, but existing composite scores often require multiple clinical inputs and leave substantial intermediate-risk groups. CLA 3+ and CLA 4 use CAP, liver stiffness, and ALT to achieve prognostic performance comparable to Agile scores, while yielding small intermediate-risk groups and informative serial changes. These findings support their potential use as simple second-line tools after FIB-4 and for longitudinal risk assessment in VCTE-based pathways, with relevance to clinical triage and future trial design.
METHODS: We analyzed data from the VCTE-Prognosis cohort, an international multicenter study with serial liver stiffness measurements. Outcomes included liver stiffness progression (≥20% increase and Baveno category upshift), liver stiffness regression (≥20% decrease and Baveno category downshift), and incident LREs. We assessed associations between baseline antihypertensive therapy and liver outcomes using Cox proportional hazards models with inverse probability of treatment weighting, comparing RASi users vs nonusers and CCB users vs nonusers.
RESULTS: We followed 4302 individuals with MASLD and hypertension for a median of 4.6 years, of whom 41.2% received RASis and 32.0% received CCBs. RASi use was associated with a higher likelihood of liver stiffness regression (hazard ratio [HR], 1.220; 95% confidence interval [CI], 1.010-1.474; P = .039) and associated with a lower risk of long-term LREs (HR, 0.431; 95% CI 0.261-0.714; P = .001), with no significant effect on liver stiffness progression. Conversely, CCB use was associated with a lower risk of liver stiffness progression (HR, 0.681; 95% CI, 0.521-0.890; P = .005) and LREs (HR, 0.559; 95% CI, 0.326-0.958; P = .034), but not with liver stiffness regression. Findings were consistent across subgroup and sensitivity analyses.
CONCLUSIONS: RASi and CCB use were associated with favorable liver stiffness trajectories and lower risk of long-term LREs in individuals with MASLD and hypertension.