Abstract
Aims: International guidelines recommend non-invasive screening for non-alcoholic fatty liver disease (NAFLD) in people with type 2 diabetes mellitus. Several readily available biomarker scores have been developed to estimate the risk of liver fibrosis. These include the Fibrosis-4 score (FIB4), NAFLD fibrosis score (NFS), and AST to platelet ratio index (APRI). In a cohort of individuals with type 2 diabetes, we aimed to describe the distribution of these scores and the association between risk categories and all-cause mortality.
Materials and Methods: This was a retrospective cohort study of 12,589 patients with follow-up from January 2012 until November 2021. The cut-points used to identify low risk were: FIB4 =65 years; NFS =65 years; APRI < 1 (independent of age). High risk cut points were FIB4 >2.67; NFS >0.676; APRI >=1 (all independent of age). Multivariable Cox regression analysis was performed to assess the association between liver fibrosis scores and all-cause mortality.
Results: Mean±SD age was 65.2±12.1 years. 54.5% were men and median (IQR) diabetes duration was 5.8 (2.8-9.3) years. Prevalence of high risk categories was 6.1% for FIB4, 23.5% for NFS and 1.6% for APRI. During median follow-up of 9.8 years, 3925 patients (31.1%) died resulting in a crude mortality rate of 40.4 per 1000 patient-years. Overall adjusted all-cause mortality hazard ratios (95% CIs) in the high compared with low fibrosis risk groups were 3.69 (1.95-2.75) for FIB4, 2.32 (2.88-4.70) for NFS, and 3.92 (2.88-5.34) for APRI. Stratified adjusted all-cause mortality hazard ratios for individuals under 65 years and people over 65 years of age at cohort entry were 3.89 (2.99-5.05) and 1.44 (1.28-1.61) for FIB4, 2.50 (1.89-3.18) and 1.35 (1.24-1.48) for NFS and 3.74 (2.73-5.14) and 1.64 (1.24-2.17) for APRI.
Conclusions: All three fibrosis risk scores were positively associated with all-cause mortality in people with type 2 diabetes, with higher relative risks in younger than older people. Effective interventions are required to minimise excess mortality in people at high risk of liver fibrosis.
Materials and Methods: This was a retrospective cohort study of 12,589 patients with follow-up from January 2012 until November 2021. The cut-points used to identify low risk were: FIB4 =65 years; NFS =65 years; APRI < 1 (independent of age). High risk cut points were FIB4 >2.67; NFS >0.676; APRI >=1 (all independent of age). Multivariable Cox regression analysis was performed to assess the association between liver fibrosis scores and all-cause mortality.
Results: Mean±SD age was 65.2±12.1 years. 54.5% were men and median (IQR) diabetes duration was 5.8 (2.8-9.3) years. Prevalence of high risk categories was 6.1% for FIB4, 23.5% for NFS and 1.6% for APRI. During median follow-up of 9.8 years, 3925 patients (31.1%) died resulting in a crude mortality rate of 40.4 per 1000 patient-years. Overall adjusted all-cause mortality hazard ratios (95% CIs) in the high compared with low fibrosis risk groups were 3.69 (1.95-2.75) for FIB4, 2.32 (2.88-4.70) for NFS, and 3.92 (2.88-5.34) for APRI. Stratified adjusted all-cause mortality hazard ratios for individuals under 65 years and people over 65 years of age at cohort entry were 3.89 (2.99-5.05) and 1.44 (1.28-1.61) for FIB4, 2.50 (1.89-3.18) and 1.35 (1.24-1.48) for NFS and 3.74 (2.73-5.14) and 1.64 (1.24-2.17) for APRI.
Conclusions: All three fibrosis risk scores were positively associated with all-cause mortality in people with type 2 diabetes, with higher relative risks in younger than older people. Effective interventions are required to minimise excess mortality in people at high risk of liver fibrosis.
| Original language | English |
|---|---|
| Journal | Diabetes, Obesity and Metabolism |
| Early online date | 13 Jun 2023 |
| DOIs | |
| Publication status | E-pub ahead of print - 13 Jun 2023 |