Non-alcoholic fatty liver disease (NAFLD) is a global epidemic with a prevalence of 25%. Currently therapies for NAFLD patients without diabetes mellitus (DM) are limited, and are associated with various adverse side effects. Sodium-glucose cotransporter type-2 (SGLT2) inhibitors can reduce hepatic fat content in patients with DM. However, the role of SGLT2 inhibitors in NAFLD patients without DM has not been investigated. Magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) and liver stiffness measurement (LSM) are non-invasive methods to diagnose hepatic steatosis and fibrosis/cirrhosis, respectively.

The investigators propose a double-blind, randomized, placebo-controlled trial to compare the effects of empagliflozin (a type of SLGT2 inhibitors) versus placebo (in a 1:1 ratio) in reducing hepatic fat content as measured by MRI-PDFF in NAFLD patients without DM. A total of 98 adult patients will be randomly sampled from the liver clinic in our local hospital. Empagliflozin 10mg daily will be given to the treatment arm. The placebo pill will be manufactured to be identical in appearance to the study drug. Eligible subjects will be followed up until week 52, and will undergo clinical, anthropometric and laboratory assessments (including liver function test and fasting blood) at baseline, week 6, 12, 26, 40 and 52. They will undergo LSM at baseline, week 26 and 52, and MRI-PDFF at baseline and week 52. The primary outcome will be a difference in change of liver fat content (measured by MRI-PDFF) at week 52 from baseline between the two groups.

The study results will determine whether SGLT2 inhibitors can reduce hepatic steatosis in NAFLD patients without DM.

Non-alcoholic fatty liver disease (NAFLD) is a global epidemic with a prevalence of 25%. Currently therapies for NAFLD patients without diabetes mellitus (DM) are limited, and are associated with various adverse side effects. Sodium-glucose cotransporter type-2 (SGLT2) inhibitors are antidiabetic drugs that reduce hepatic fat content in patients with DM, which is independent of glycemic control. However, the role of SGLT2 inhibitors in NAFLD patients without DM has not been investigated. Magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) is an emerging non-invasive imaging technique, and is more sensitive than liver biopsy/histology in quantifying liver fat change. Liver stiffness measurement (LSM) by transient elastography is a non-invasive method to diagnose fibrosis/cirrhosis with high accuracy.

The novelty of utilizing the concept of "drug repositioning" by changing the role of SGLT2 inhibitors in treating DM to treating NAFLD in patients without DM deserves exploration. The investigators propose a double-blind, randomized, placebo-controlled trial to compare the effects of empagliflozin (a type of SLGT2 inhibitors) versus placebo (in a 1:1 ratio) in reducing hepatic fat content as measured by MRI-PDFF in NAFLD patients without DM. A total of 98 adult patients will be randomly sampled from the liver clinical in our local hospital. Empagliflozin 10mg daily will be given to the treatment arm. The placebo pill will be manufactured to be identical in appearance to the study drug. Eligible subjects will be followed up until week 52, and will undergo clinical, anthropometric and laboratory assessments (including liver function test and fasting blood) at baseline, week 6, 12, 26, 40 and 52. They will undergo LSM at baseline, week 26 and 52, and MRI-PDFF at baseline and week 52. The primary outcome will be a difference in change of liver fat content (measured by MRI-PDFF) at week 52 from baseline between the two groups. The secondary outcomes will be remission of steatosis (MRI-PDFF <5%) at week 52, reduction of liver fibrosis (LSM) at week 26 and 52, improvement of laboratory results (including liver transaminases and ductal enzymes, fasting glucose, HbA1c, lipid profile), improvement of anthropometric measurements, and combined cardiovascular and cerebrovascular events.

The study results will determine whether SGLT2 inhibitors can reduce hepatic steatosis and regress fibrosis in NAFLD patients without DM.

Inclusion Criteria:

• Potential study subjects will first be screened by transient elastography for the presence of hepatic steatosis (defined as a measurement of controlled attenuation parameter [CAP] >= 248 db/M).
• They will be recruited into study if steatosis is >= 5% as confirmed by MRI-PDFF

Exclusion Criteria:

• DM (defined as hemoglobin A1c [HbA1c] >= 6.5% or fasting glucose >= 7.0 mmol/L)
• alcohol intake > 20g within past 2 years
• concurrent chronic liver diseases (including chronic viral hepatitis infection, autoimmune hepatitis, Wilson's disease, hemochromatosis, congestive hepatopathy, primary biliary cholangitis, primary sclerosing cholangitis, biliary tract obstruction)
• drug-induced liver disease
• usage of drugs that can lead to hepatic steatosis (e.g. steroids, amiodarone, valproate, methotrexate, tamoxifen)
• decompensated cirrhosis (including ascites, hepatic hydrothorax, variceal bleeding, hepatic encephalopathy, hepatorenal syndrome, hepatopulmonary syndrome)
• history of malignancy including HCC
• recreational substance abuse
• pregnancy
• contraindications to empagliflozin use (estimated glomerular filtration rate [eGFR] <45mL/min/1.73m2 as measured by the MDRD equation, history of recurrent genitourinary tract infections, gangrene, or allergy)
• contraindications to MRI (e.g., claustrophobia, certain cardiac pacemakers, implanted medical devices with ferromagnetic properties).

IPD will be made available in the form of excel files upon request by other researchers

The placebo pills will be manufactured in identical appearance to the study drug (empagliflozin)