Targeted Screening for Liver Fibrosis: Identifying High-Risk Adults for Transient Elastography Evaluation
Abstract
Introduction and Objective. Non-alcoholic fatty liver disease (NAFLD) is a significant clinical and public health issue, affecting over 30% of the general population. It can lead to liver fibrosis, and ultimately to cirrhosis and physical frailty. Early detection is crucial as early NAFLD-related fibrosis is reversible. Transient elastography (TE), a non-invasive screening method, measures liver stiffness measure (LSM) as a surrogate for hepatic fibrosis, guiding subsequent liver biopsy decisions. However, TE availability and costs are limiting factors. Currently, clinical practice relies on individual indicators (e.g., obesity) for TE testing referrals. We aimed to evaluate the accuracy of individual traditional indicators in identifying adults at risk for abnormal LSM>7pka and determine whether combining readily available indicators in primary care settings could enhance screening accuracy. Methods. Cross-sectional analysis utilized National Health and Nutrition Examination Survey (NHANES) 2017-2020 data, including adults aged ≥ 40 years without history of excessive drinking, cirrhosis, and/or hepatitis B/C. The outcome, liver stiffness (LSM) >7pka by TE, was assessed. Potential predictors included diseases (diabetes [DM], cardiovascular disease, hypertension), examination findings (body mass index [BMI], heart rate [HR]), and lab tests (alanine [ALT] and aspartate aminotransferase [AST], gamma-glutamyl-transferase, platelets, glycated hemoglobin [A1c], C-reactive protein, uric acid, FIB-4 [age, AST, ALT, and platelets], and APRI [AST-to-platelet-ratio index]). Stepwise logistic regression using sampling weights predicted LSM>7pKa based on individual and combined indicators. The area under the receiver operating curve (AUC-ROC) compared predictive accuracy across models. Stata 16 was utilized. Results. Mean age of study population was 59.0±.41. Women were majority (52.3%). Prevalence of diabetes, BMI>30 kg/m2, and LSM>7 pKa were 16.9%, 44.4%, and 18.1%, respectively. AUC-ROC for individual indicators had modest accuracy; e.g., FIB-4 score>3.25: 52% (95% Confidence Interval: 52-53%), metabolic syndrome: 64% (65-68%), and BMI>30kg/m2: 66% (65-68). The final model, incorporating diabetes, BMI>30 kg/m2, HR, A1c, FIB-4>3.25, APRI>0.5, and AST>40 IU/l (p < 0.05 for all), yielded an AUC-ROC of 74% (72-76%). Conclusions-Implications. Combining primary care indicators improved accuracy in identifying at-risk adults for abnormal LSM compared to traditional individual indicators. Further research is needed to identify the optimal indicator combination for identifying individuals at risk of liver fibrosis who may benefit from TE screening.
Keywords
Non-alcoholic fatty liver disease (NAFLD); Transient elastography (TE); Liver stiffness measurement (LSM); Risk Assessment
Presentation Type
Oral Presentation
Targeted Screening for Liver Fibrosis: Identifying High-Risk Adults for Transient Elastography Evaluation
Introduction and Objective. Non-alcoholic fatty liver disease (NAFLD) is a significant clinical and public health issue, affecting over 30% of the general population. It can lead to liver fibrosis, and ultimately to cirrhosis and physical frailty. Early detection is crucial as early NAFLD-related fibrosis is reversible. Transient elastography (TE), a non-invasive screening method, measures liver stiffness measure (LSM) as a surrogate for hepatic fibrosis, guiding subsequent liver biopsy decisions. However, TE availability and costs are limiting factors. Currently, clinical practice relies on individual indicators (e.g., obesity) for TE testing referrals. We aimed to evaluate the accuracy of individual traditional indicators in identifying adults at risk for abnormal LSM>7pka and determine whether combining readily available indicators in primary care settings could enhance screening accuracy. Methods. Cross-sectional analysis utilized National Health and Nutrition Examination Survey (NHANES) 2017-2020 data, including adults aged ≥ 40 years without history of excessive drinking, cirrhosis, and/or hepatitis B/C. The outcome, liver stiffness (LSM) >7pka by TE, was assessed. Potential predictors included diseases (diabetes [DM], cardiovascular disease, hypertension), examination findings (body mass index [BMI], heart rate [HR]), and lab tests (alanine [ALT] and aspartate aminotransferase [AST], gamma-glutamyl-transferase, platelets, glycated hemoglobin [A1c], C-reactive protein, uric acid, FIB-4 [age, AST, ALT, and platelets], and APRI [AST-to-platelet-ratio index]). Stepwise logistic regression using sampling weights predicted LSM>7pKa based on individual and combined indicators. The area under the receiver operating curve (AUC-ROC) compared predictive accuracy across models. Stata 16 was utilized. Results. Mean age of study population was 59.0±.41. Women were majority (52.3%). Prevalence of diabetes, BMI>30 kg/m2, and LSM>7 pKa were 16.9%, 44.4%, and 18.1%, respectively. AUC-ROC for individual indicators had modest accuracy; e.g., FIB-4 score>3.25: 52% (95% Confidence Interval: 52-53%), metabolic syndrome: 64% (65-68%), and BMI>30kg/m2: 66% (65-68). The final model, incorporating diabetes, BMI>30 kg/m2, HR, A1c, FIB-4>3.25, APRI>0.5, and AST>40 IU/l (p < 0.05 for all), yielded an AUC-ROC of 74% (72-76%). Conclusions-Implications. Combining primary care indicators improved accuracy in identifying at-risk adults for abnormal LSM compared to traditional individual indicators. Further research is needed to identify the optimal indicator combination for identifying individuals at risk of liver fibrosis who may benefit from TE screening.