Noninvasive scoring systems have modest predictive ability to identify the future development of severe liver disease in patients in the general population, according to a study published in Gastroenterology.
The aim of this retrospective study was to analyze 5 noninvasive scoring systems on their ability to predict incident severe liver disease during a long-term follow-up. Data were gathered from the Swedish Apolipoprotein Mortality Risk (AMORIS) cohort. Baseline laboratory testing and basic health evaluations were completed between 1985-1996 and included about 35% of the population of Stockholm county. For this study, patients were included in different noninvasive tests based on the number of available biomarkers and necessary variables for the test to be completed.
The 5 noninvasive tests were Aspartate aminotransferase to Platelet Ratio Index (APRI), BARD, Fibrosis-4 (FIB-4), Forns, and nonalcoholic fatty liver disease fibrosis score. The patients were grouped according to low, intermediate, or high risk for advanced fibrosis based on specific cutoffs for each individual test. Outcomes were assessed using the nationwide Swedish registers to evaluate hospitalizations, specialized outpatient care, incident cancers, and causes of deaths through 2011. Severe liver disease, such as cirrhosis, liver failure, hepatocellular carcinoma, liver transplantation, or decompensated liver disease, was evaluated at a 5-year follow-up and a 10-year follow-up.
In total 127,302 patients were included in the APRI test, 75,303 patients were included in the BARD test, 126,941 patients were included in the FIB-4 test, 122,419 patients were included in the Forns test, and 13,160 patients were included in the nonalcoholic fatty liver disease fibrosis test. Patients who were grouped into the high-risk category for each test were predominantly older men with type 2 diabetes. There was a strong association between baseline risk category and the occurrence of severe liver disease.
The hazard ratios were lower in the low risk category and higher in the high risk category. APRI had an adjusted hazard ratio (aHR) of 45.25 in the high-risk category, BARD had an aHR of 1.71 in the high-risk category, Fibrosis-4 had an aHR of 17.01 in the high-risk category, Forns had an aHR of 29.29 in the high-risk category, and nonalcoholic fatty liver disease fibrosis test did not have enough patients in the high risk category to calculate a hazard ratio. The predictability of the tests reduced over time and was more accurate for men when compared with women. The APRI test had the highest predictability with a 96.0% accuracy in the intermediate-risk group and 99.1% in the high risk group. All tests were more accurate when used in patients with type 2 diabetes, who were obese, and who had an elevated alanine aminotransferase score.
Limitations of this study include the retrospective nature of the study, the potential misclassification in the codes in the national Swedish registers, a higher proportion of women included in the study, and the sample population being mostly white with a low prevalence of hepatitis B and C who make generalizability of this study low.
The researchers concluded, “[h]igher scores from noninvasive scoring systems to evaluate fibrosis are associated with an increased risk [for] cirrhosis in a general population, but their predictive ability is modest.”
Disclosure: This clinical trial was supported by Astra Zeneca. Several study authors declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.
Hagström H, Talbäck M, Andreasson A, Walldius G, Hammar N. Ability of noninvasive scoring systems to identify individuals in the population at risk for severe liver disease [published online September 26, 2019]. Gastroenterology. doi: 10.1053/j.gastro.2019.09.008