Low accuracy of FIB-4 test to identify people with diabetes at low risk of advanced fibrosis.

Abstract

Non-invasive tests accurately stratify patients with NAFLD based on their risk of liver-related eventsJournal of HepatologyVol. 76Issue 5PreviewPrevious studies on the prognostic significance of non-invasive liver fibrosis tests in non-alcoholic fatty liver disease (NAFLD) lack direct comparison to liver biopsy. We aimed to evaluate the prognostic accuracy of fibrosis-4 (FIB4) and vibration-controlled transient elastography (VCTE), compared to liver biopsy, for the prediction of liver-related events (LREs) in NAFLD. Full-Text PDF We read with great interest the article by Boursier et al. recently published in Journal of Hepatology.[1]Boursier J. Hagström H. Ekstedt M. Moreau C. Bonacci M. Cure S. et al.Non-invasive tests accurately stratify patients with NAFLD based on their risk of liver-related events.J Hepatol. 2022; 76: 1013-1020Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar In a multicentre cohort of patients with NAFLD, the authors examined the prognostic accuracy of the stepwise Fibrosis-4 (FIB-4) – vibration-controlled transient elastography (VCTE) algorithm for non-invasive fibrosis risk stratification,[2]Tsochatzis E.A. Newsome P.N. Non-alcoholic fatty liver disease and the interface between primary and secondary care.Lancet Gastroenterol Hepatol. 2018; 3: 509-517Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar for the prediction of cirrhosis complications, hepatocellular carcinoma, and death. The algorithm proposes FIB-4 as a first step to identify individuals at low risk (FIB-4 <1.3) of advanced fibrosis who can be managed in primary care. People with indeterminate or high-risk scores (FIB-4 ≥1.3) need additional assessment with VCTE, and those at increased risk of significant fibrosis (VCTE ≥8 kPa) require hepatology referral for further evaluation of possible advanced fibrosis. The authors found that overall, patients with FIB-4 <1.3 had a very low risk of liver-related events, validating use of the pathway to retain patients with a good prognosis in primary care. However, the presence of diabetes (defined as use of antidiabetic medications) was an independent predictor of liver-related events. In the subgroup with FIB-4 <1.3, the prevalence of VCTE ≥8.0 kPa was higher among people with diabetes compared to those without diabetes (p <0.001), and Kaplan-Meier curves demonstrated poorer 8-year prognosis (p = 0.003).We share the concern of Boursier and colleagues regarding the use of FIB-4 <1.3 to classify people with type 2 diabetes as being at ‘low risk’ of advanced fibrosis. In our pooled cohort3Hayward K.L. McKillen B.J. Horsfall L.U. McIvor C. Liew K. Sexton J. et al.Towards collaborative management of nonalcoholic fatty liver disease (TCM-NAFLD): a 'real-world' pathway for fibrosis risk assessment in primary care.Intern Med J. 2021; PubMed Google Scholar, 4Patel P. Hossain F. Horsfall L.U. Banh X. Hayward K.L. Williams S. et al.A pragmatic approach identifies a high rate of nonalcoholic fatty liver disease with advanced fibrosis in diabetes clinics and at-risk populations in primary care.Hepatol Commun. 2018; 2: 893-905Crossref PubMed Scopus (37) Google Scholar, 5Gracen L. Hayward K.L. Aikebuse M. Russell A. O'Beirne J. McPhail S. et al.Implementing the right care in the right place at the right time for non-alcoholic fatty liver disease (NAFLD-RRR study): a study protocol for a community care pathway for people with type 2 diabetes.BMC Health Serv Res. 2022; 22: 487Crossref PubMed Scopus (1) Google Scholar of patients with type 2 diabetes (n = 382) prospectively recruited from primary care and diabetes clinics, VCTE was used as a point-of-care test to assign risk of fibrosis based on liver stiffness measurement (LSM), and the most recent biochemical and haematologic data were used to calculate the FIB-4 score. Cohort 1[3]Hayward K.L. McKillen B.J. Horsfall L.U. McIvor C. Liew K. Sexton J. et al.Towards collaborative management of nonalcoholic fatty liver disease (TCM-NAFLD): a 'real-world' pathway for fibrosis risk assessment in primary care.Intern Med J. 2021; PubMed Google Scholar was recruited in primary care from June 2019 to December 2020, Cohort 2[4]Patel P. Hossain F. Horsfall L.U. Banh X. Hayward K.L. Williams S. et al.A pragmatic approach identifies a high rate of nonalcoholic fatty liver disease with advanced fibrosis in diabetes clinics and at-risk populations in primary care.Hepatol Commun. 2018; 2: 893-905Crossref PubMed Scopus (37) Google Scholar and Cohort 3[5]Gracen L. Hayward K.L. Aikebuse M. Russell A. O'Beirne J. McPhail S. et al.Implementing the right care in the right place at the right time for non-alcoholic fatty liver disease (NAFLD-RRR study): a study protocol for a community care pathway for people with type 2 diabetes.BMC Health Serv Res. 2022; 22: 487Crossref PubMed Scopus (1) Google Scholar were recruited in tertiary diabetes clinics and primary care from 2015 to 2017 and June 2021 to May 2022, respectively. Mean age was 59.0 ± 10.6 years, 56.3% of the patients were male, 71.7% Caucasian, with median BMI 33.9 (IQR 29.4-38.9 kg/m2) and girth 117 (IQR 105-130 cm). VCTE met quality criteria in 342 (89.5%) individuals. One hundred (29.2%) had LSM ≥8 kPa, 73 (21.3%) had LSM ≥9.5 kPa and 51 (14.9%) had LSM ≥12 kPa (suggestive of clinically significant fibrosis, advanced fibrosis, and cirrhosis, respectively).The LSM results according to low, indeterminate, and high FIB-4 risk stratification scores are illustrated in Fig. 1. FIB-4 <1.3 was present in 233 of 335 (69.6%) patients and had a modest negative predictive value (NPV) (78.5%; 95% CI 73.3-83.8) for excluding clinically significant fibrosis. Composite indeterminate and high scores (FIB-4 ≥1.3) were also poorly predictive of having LSM ≥8 kPa, with FIB-4 sensitivity of 48.5% (95% CI 38.5-58.4) and positive predictive value (PPV) of 46.1% (95% CI 36.4-55.8).The NAFLD fibrosis score (NFS) is another simple test that has been recommended for use in community and primary care settings to exclude advanced fibrosis and to identify people requiring further assessment for high-risk NAFLD.[6]Angulo P. Hui J.M. Marchesini G. Bugianesi E. George J. Farrell G.C. et al.The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD.Hepatology. 2007; 45: 846-854Crossref PubMed Scopus (1881) Google Scholar However the test is not favoured for use in patients from diabetes clinics because the algorithm includes the presence of diabetes, leading to an increase in the score for all patients. In our cohort of patients with type 2 diabetes and NAFLD, low NFS scores (<-1.455) were present in 58 of 341 (17.0%) patients. NFS <-1.455 had an NPV of 72.4% (95% CI 60.9-83.9) for excluding clinically significant fibrosis (Fig. 1). The sensitivity of composite indeterminate and high NFS (≥-1.455) for predicting LSM ≥8 kPa was 83.8% (95% CI 76.6-91.1) with PPV of 29.3% (95% CI 24.0-34.6). Using NFS, only a minority of patients do not require a second-line test to investigate indeterminate or high-risk scores.We believe that for people with type 2 diabetes, the use of simple scores as a first step has low accuracy to identify individuals at low risk of advanced fibrosis. In our patients with diabetes and VCTE ≥8 kPa, FIB-4 was <1.3 in 50 of 97 (51.5%) patients, supporting the findings from Boursier et al. that FIB-4 has low sensitivity with a high number of “false negatives” in this population with a moderate prevalence of advanced fibrosis. We concur with Boursier et al. that “This is problematic, as current guidelines indicate that all patients with FIB-4 <1.30 should not be referred to a liver specialist.” The findings suggest that in people with type 2 diabetes, VCTE or a patented serum test[7]Younossi Z.M. Felix S. Jeffers T. Younossi E. Nader F. Pham H. et al.Performance of the enhanced liver fibrosis test to estimate advanced fibrosis among patients with nonalcoholic fatty liver disease.JAMA Netw Open. 2021; 4e2123923Crossref PubMed Scopus (4) Google Scholar may be a preferable initial step in fibrosis risk stratification.Financial supportLG was supported by the Metro South Health Research Support Scheme Grant (#RSS_2022_012). KLH was supported by a Health Innovation, Investment and Research Office (HIIRO) Clinical Research Fellowship. All funding was external and the funding body had no role in the design of the study, collection, analysis and interpretation of data and in writing the manuscript.Authors’ contributionsEEP: Conceptualised, drafted and designed the work. PV: Designed the work, substantially revised the manuscript. LG: Consented the participants and administered the intervention. Collated data and performed the data analysis. Substantially revised the manuscript. KLH: Substantially revised the manuscript. KMI: Substantially revised the manuscript. All authors read and approved the final manuscript. We read with great interest the article by Boursier et al. recently published in Journal of Hepatology.[1]Boursier J. Hagström H. Ekstedt M. Moreau C. Bonacci M. Cure S. et al.Non-invasive tests accurately stratify patients with NAFLD based on their risk of liver-related events.J Hepatol. 2022; 76: 1013-1020Abstract Full Text Full Text PDF PubMed Scopus (11) Google Scholar In a multicentre cohort of patients with NAFLD, the authors examined the prognostic accuracy of the stepwise Fibrosis-4 (FIB-4) – vibration-controlled transient elastography (VCTE) algorithm for non-invasive fibrosis risk stratification,[2]Tsochatzis E.A. Newsome P.N. Non-alcoholic fatty liver disease and the interface between primary and secondary care.Lancet Gastroenterol Hepatol. 2018; 3: 509-517Abstract Full Text Full Text PDF PubMed Scopus (74) Google Scholar for the prediction of cirrhosis complications, hepatocellular carcinoma, and death. The algorithm proposes FIB-4 as a first step to identify individuals at low risk (FIB-4 <1.3) of advanced fibrosis who can be managed in primary care. People with indeterminate or high-risk scores (FIB-4 ≥1.3) need additional assessment with VCTE, and those at increased risk of significant fibrosis (VCTE ≥8 kPa) require hepatology referral for further evaluation of possible advanced fibrosis. The authors found that overall, patients with FIB-4 <1.3 had a very low risk of liver-related events, validating use of the pathway to retain patients with a good prognosis in primary care. However, the presence of diabetes (defined as use of antidiabetic medications) was an independent predictor of liver-related events. In the subgroup with FIB-4 <1.3, the prevalence of VCTE ≥8.0 kPa was higher among people with diabetes compared to those without diabetes (p <0.001), and Kaplan-Meier curves demonstrated poorer 8-year prognosis (p = 0.003). We share the concern of Boursier and colleagues regarding the use of FIB-4 <1.3 to classify people with type 2 diabetes as being at ‘low risk’ of advanced fibrosis. In our pooled cohort3Hayward K.L. McKillen B.J. Horsfall L.U. McIvor C. Liew K. Sexton J. et al.Towards collaborative management of nonalcoholic fatty liver disease (TCM-NAFLD): a 'real-world' pathway for fibrosis risk assessment in primary care.Intern Med J. 2021; PubMed Google Scholar, 4Patel P. Hossain F. Horsfall L.U. Banh X. Hayward K.L. Williams S. et al.A pragmatic approach identifies a high rate of nonalcoholic fatty liver disease with advanced fibrosis in diabetes clinics and at-risk populations in primary care.Hepatol Commun. 2018; 2: 893-905Crossref PubMed Scopus (37) Google Scholar, 5Gracen L. Hayward K.L. Aikebuse M. Russell A. O'Beirne J. McPhail S. et al.Implementing the right care in the right place at the right time for non-alcoholic fatty liver disease (NAFLD-RRR study): a study protocol for a community care pathway for people with type 2 diabetes.BMC Health Serv Res. 2022; 22: 487Crossref PubMed Scopus (1) Google Scholar of patients with type 2 diabetes (n = 382) prospectively recruited from primary care and diabetes clinics, VCTE was used as a point-of-care test to assign risk of fibrosis based on liver stiffness measurement (LSM), and the most recent biochemical and haematologic data were used to calculate the FIB-4 score. Cohort 1[3]Hayward K.L. McKillen B.J. Horsfall L.U. McIvor C. Liew K. Sexton J. et al.Towards collaborative management of nonalcoholic fatty liver disease (TCM-NAFLD): a 'real-world' pathway for fibrosis risk assessment in primary care.Intern Med J. 2021; PubMed Google Scholar was recruited in primary care from June 2019 to December 2020, Cohort 2[4]Patel P. Hossain F. Horsfall L.U. Banh X. Hayward K.L. Williams S. et al.A pragmatic approach identifies a high rate of nonalcoholic fatty liver disease with advanced fibrosis in diabetes clinics and at-risk populations in primary care.Hepatol Commun. 2018; 2: 893-905Crossref PubMed Scopus (37) Google Scholar and Cohort 3[5]Gracen L. Hayward K.L. Aikebuse M. Russell A. O'Beirne J. McPhail S. et al.Implementing the right care in the right place at the right time for non-alcoholic fatty liver disease (NAFLD-RRR study): a study protocol for a community care pathway for people with type 2 diabetes.BMC Health Serv Res. 2022; 22: 487Crossref PubMed Scopus (1) Google Scholar were recruited in tertiary diabetes clinics and primary care from 2015 to 2017 and June 2021 to May 2022, respectively. Mean age was 59.0 ± 10.6 years, 56.3% of the patients were male, 71.7% Caucasian, with median BMI 33.9 (IQR 29.4-38.9 kg/m2) and girth 117 (IQR 105-130 cm). VCTE met quality criteria in 342 (89.5%) individuals. One hundred (29.2%) had LSM ≥8 kPa, 73 (21.3%) had LSM ≥9.5 kPa and 51 (14.9%) had LSM ≥12 kPa (suggestive of clinically significant fibrosis, advanced fibrosis, and cirrhosis, respectively). The LSM results according to low, indeterminate, and high FIB-4 risk stratification scores are illustrated in Fig. 1. FIB-4 <1.3 was present in 233 of 335 (69.6%) patients and had a modest negative predictive value (NPV) (78.5%; 95% CI 73.3-83.8) for excluding clinically significant fibrosis. Composite indeterminate and high scores (FIB-4 ≥1.3) were also poorly predictive of having LSM ≥8 kPa, with FIB-4 sensitivity of 48.5% (95% CI 38.5-58.4) and positive predictive value (PPV) of 46.1% (95% CI 36.4-55.8). The NAFLD fibrosis score (NFS) is another simple test that has been recommended for use in community and primary care settings to exclude advanced fibrosis and to identify people requiring further assessment for high-risk NAFLD.[6]Angulo P. Hui J.M. Marchesini G. Bugianesi E. George J. Farrell G.C. et al.The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD.Hepatology. 2007; 45: 846-854Crossref PubMed Scopus (1881) Google Scholar However the test is not favoured for use in patients from diabetes clinics because the algorithm includes the presence of diabetes, leading to an increase in the score for all patients. In our cohort of patients with type 2 diabetes and NAFLD, low NFS scores (<-1.455) were present in 58 of 341 (17.0%) patients. NFS <-1.455 had an NPV of 72.4% (95% CI 60.9-83.9) for excluding clinically significant fibrosis (Fig. 1). The sensitivity of composite indeterminate and high NFS (≥-1.455) for predicting LSM ≥8 kPa was 83.8% (95% CI 76.6-91.1) with PPV of 29.3% (95% CI 24.0-34.6). Using NFS, only a minority of patients do not require a second-line test to investigate indeterminate or high-risk scores. We believe that for people with type 2 diabetes, the use of simple scores as a first step has low accuracy to identify individuals at low risk of advanced fibrosis. In our patients with diabetes and VCTE ≥8 kPa, FIB-4 was <1.3 in 50 of 97 (51.5%) patients, supporting the findings from Boursier et al. that FIB-4 has low sensitivity with a high number of “false negatives” in this population with a moderate prevalence of advanced fibrosis. We concur with Boursier et al. that “This is problematic, as current guidelines indicate that all patients with FIB-4 <1.30 should not be referred to a liver specialist.” The findings suggest that in people with type 2 diabetes, VCTE or a patented serum test[7]Younossi Z.M. Felix S. Jeffers T. Younossi E. Nader F. Pham H. et al.Performance of the enhanced liver fibrosis test to estimate advanced fibrosis among patients with nonalcoholic fatty liver disease.JAMA Netw Open. 2021; 4e2123923Crossref PubMed Scopus (4) Google Scholar may be a preferable initial step in fibrosis risk stratification. Financial supportLG was supported by the Metro South Health Research Support Scheme Grant (#RSS_2022_012). KLH was supported by a Health Innovation, Investment and Research Office (HIIRO) Clinical Research Fellowship. All funding was external and the funding body had no role in the design of the study, collection, analysis and interpretation of data and in writing the manuscript. LG was supported by the Metro South Health Research Support Scheme Grant (#RSS_2022_012). KLH was supported by a Health Innovation, Investment and Research Office (HIIRO) Clinical Research Fellowship. All funding was external and the funding body had no role in the design of the study, collection, analysis and interpretation of data and in writing the manuscript. Authors’ contributionsEEP: Conceptualised, drafted and designed the work. PV: Designed the work, substantially revised the manuscript. LG: Consented the participants and administered the intervention. Collated data and performed the data analysis. Substantially revised the manuscript. KLH: Substantially revised the manuscript. KMI: Substantially revised the manuscript. All authors read and approved the final manuscript. EEP: Conceptualised, drafted and designed the work. PV: Designed the work, substantially revised the manuscript. LG: Consented the participants and administered the intervention. Collated data and performed the data analysis. Substantially revised the manuscript. KLH: Substantially revised the manuscript. KMI: Substantially revised the manuscript. All authors read and approved the final manuscript. The authors declare no conflicts of interest that pertain to this work. Please refer to the accompanying ICMJE disclosure forms for further details. Supplementary dataThe following are the supplementary data to this article: Download .pdf (.24 MB) Help with pdf files Multimedia component 1 The following are the supplementary data to this article: Download .pdf (.24 MB) Help with pdf files Multimedia component 1