NAFLD: utility of a lower cutpoint of ALT levels

 There are a few new studies on non-alcoholic fatty liver disease (NAFLD) that challenge/illuminate some clinical issues in our conception of NAFLD and the approach to the disease.

 

The current blog is from a French study that found that “normal” ALT levels with a cutpoint of >40 IU/L were insufficient in diagnosing NAFLD (see NAFLD NASH in DM with nl ALT DiabCare2023 in dropbox, or doi.org/10.2337/dc22-2048). The second (to follow soon) reviews the relatively common issue of NAFLD in “lean” individuals, who may well have a worse prognosis.

 

Details:

-- 713 patients with diabetes and NAFLD documented by ultrasound who had been referred to a hepatologist for further workup by FIB-4 and VCTE (vibration-controlled transient elastography)

-- all participants were asked to get a liver biopsy if their ALT levels were persistently >20 IU/L in females or >30 IU/L in males, and with no other evident cause of liver disease

-- 63% male, BMI 32, waist circumference median of 109cm (81% had elevated waist circumference of >102cm for men or >88cm for women), HTN 68%, DM duration 9yrs, retinop 19%/neurop 21%/nephrop 43%

-- ischemic heart dz in 15%, at least one macrovasc complication in 20%

-- meds: oral diabetic meds 97%, insulin without GLP-1 in 9%, GLP-1 without insulin in 23%, both GLP-1 and insulin in 34%, lipid lowering med 65%, cardiovasc drug 73%

-- AST 35, ALT 49, ALT >40 in 64%

    -- the only demographic differences between those with ALT <40 vs >40 was that the former were more likely to be female; and they had lower GGT, bilirubin, serum albumin and ferritin levels

-- total cholesterol 3.97 (153.5 mg/dL)/LDL 2.07 (80 mg/Dl)/HDL 1.03 (40 mg/dL), TG 1.73 (153 mg/dL), A1c 7.5%, eGFR 95, FIB-4 1.20, NAFLD fibrosis score -0.7 (<0.675=normal)

-- Exclusion criteria: pregnant women, other causes of chronic liver disease than NASH, alcohol consumption of >20g/d for women or >30g/d for men

-- main outcomes: predictive values of a variety of noninvasive parameters (as below), as compared to liver biopsy (the definitive evaluation of NASH or fibrosis)

Results:

-- 330 patients received an evaluable liver biopsy, finding:

        -- NASH: 192 patients, 58% (53-64%)

        -- advanced fibrosis (defined as fibrosis stages F3-F4): 124 patients, 38% (32-43%)   

            -- stage F3 fibrosis: 91 patients (28%)

            -- cirrhosis: 33 patients (10%)

 

-- Associations, based on multivariate analysis:

    -- advanced fibrosis:

        -- if did not have VCTE done: waist circumference of >102 cm in men and >88cm in women; age, BMI, HDL <1.03 (40 mg/dL) in men or <1.29 (50 mg/dL) in women, increased GGT and FIB-4

            -- the area under the ROC curve (AUROC): 0.77 (0.72-0.83) by combining waist circumference, HDL, increased GGT, and FIB-4

        -- if did have VCTE done: HDL <1.03 (40 mg/dL) in men or <1.29 (50 mg/dL) in women, FIB-4, and liver stiffness measurement by VCTE

              -- the area under the ROC curve (AUROC): 0.84 (0.80-0.89) if combine these 3  factors   

               -- there was a strong statistically significant benefit for adding the VCTE, p=0.0007 

        -- of note: the ALT and AST measurements by themselves were not statistically significant (though both are represented in the FIB-4)

 

    -- NASH:

        -- if not have VCTE: waist circumference of >102 cm in men and >88cm in women; hypertension, triglycerides >1.7 (153 mg/dL); AST, and low serum albumin and creatinine levels

            -- AUROC: 0.81 (0.76-0.86), by combining these 6 factors

        -- if have VCTE: increased CAP (controlled attenuation parameter by VCTE), triglycerides >1.7 (153 mg/dL), AST, low serum albumin and creatinine levels

              -- AUROC: 0.82 (0.77-0.87) combining these 5 factors   

                -- there no statistically significant benefit for adding the VCTE

        -- of note: the ALT and AST measurements by themselves were not significant (though both are represented in the FIB-4)

  

Commentary:

-- NAFLD is becoming the most common cause of liver disease worldwide, affecting 25% of the global population, and >70% of those with type 2 diabetes (it is also found in 22% of those with type 1 diabetes, but might be attributable to obesity):

    -- NASH (non-alcoholic steatohepatitis, defined as at least 5% hepatic steatosis, and inflammation with hepatocyte injury, with or without fibrosis), is found in at least 20-30% of those with NAFLD, and is the condition associated with progressive liver disease/cirrhosis/hepatocelllular cancer (HCC)/increased risk of cardiovascular disease/increased cardiovascular and all-cause mortality. Other associations found with NASH include: chronic kidney disease and extra-hepatic cancers

    -- NASH is found in 1.5-6.5% of the general population and 37% of those with diabetes, and is expected to increase by 63% between 2015 to 2030 (which translates to 4.9-21 million Americans and >100 million globally)

    -- NASH is the number 1 indication for liver transplantation in women, those >54yo, and Medicare recipients

    -- NAFLD and NASH are progressive diseases overall, with progression of 1 stage of fibrosis every 7 and 14 years, respectively (though this is not linear: fibrosis can progress or regress over 5 years, in one study)

    -- for a strong argument to assess and treat NASH aggressively (including many of the points above), see “Preparing for the NASH epidemic: a call to action”, a multi-specialty document, referenced in the ADA 2023 practice guidelines), which also documents significant gaps in knowledge about NAFLD/NASH in both primary care providers and endocrinologists

-- Risk factors for NAFLD/NASH: especially obesity and diabetes (though the presence of NAFLD is associated with increased risk of developing diabetes as well as cirrhosis, HCC, liver-related mortality and cardiovascular disease)

    -- the risk of bad outcomes with NAFLD is more related to the stage of fibrosis, especially if fibrosis stage is >2. the risk of all-cause mortality and liver-related mortality are, respectively:

        -- stage 1: risk ratios 1.58 and 1.41

        -- stage 2: risk ratios 2.52 and 9.57

        -- stage 3: risk ratios 3.48 for all-cause mortality (no report on liver-related mortality)

        -- stage 4: risk ratios 6.40 and 42.30

-- Proposed screening and diagnostic approach:  targeted case finding in high-risk (BMI>35, diabetes >10yrs, metabolic syndrome): ALT, AST (though if elevated, exclude other causes, eg HBV and HCV serology, auto-antibodies of ANA, AMA, ASMA antibodies, serum ferritin, a1-antitrypsin), and order liver ultrasound (though not sensitive for mild steatosis). The general recommendations also include calculated FIB-4 and NAFLD fibrosis scores; and a second-line test of elastography in those with high calculated scores (elastography, by the way, can be inaccurate since it is operator-dependent and fibrosis can be spotty)

    -- though the definitive test for defining NASH is the degree of steatosis, inflammation and hepatocyte injury by liver biopsy, this test is invasive and has potentially serious adverse outcomes: hence the approach of relying on non-invasive strategies

-- the main treatments: lifestyle changes, specifically weight loss and not drinking alcohol. and several meds do decrease steatosis and inflammation, with some having effect on fibrosis: esp pioglitazone if patient has diabetes, vitamin E 800IU/d in those without diabetes; GLP-1 agonists semaglutide (the best studied), liraglutide, dulaglutide, and exenatide are all associated with decreasing hepatic steatosis/inflammation. Though anything that decreases weight helps, including bariatric surgery (the most effective intervention) and SGLT-2’s (though only small studies done)

  

-- This study was novel in that they performed liver biopsies in those patients with diabetes who had ALT levels lower than 40 IU/L

    -- they found on a multivariate analysis model including the combination of hypertension, waist circumference  >102cm in men and  >88cm in women, triglycerides >1.7 (153 mg/dL), AST level, and low serum albumin and serum creatinine levels (but without VCTE) was highly predictive of NASH (area under the ROC curve 0.82, versus 0.68 by just assessing even the lower ALT levels). The FIB-4 score for the group had a mean of 1.20 points (a score of >1.45 is considered low risk for advanced fibrosis, >3.25 is considered high risk). In multivariate analysis, these models with combinations of these risk factors had the highest predicted value of any of the single  markers tested, with or without VCTE being done (and VTCE did not really add anything for their NASH group though did help for those with advanced fibrosis)

 

    -- this all suggests:

        -- ALT is an unreliable marker for NASH and advanced fibrosis: with the standard cutpoint of >40 IU/L, the majority of patients would not be evaluated further, even though advanced fibrosis and NASH were quite common

            -- though the lower cutpoints of ALT were better at identifying NASH and advanced fibrosis, there are concerns that this would lead to a very large number of liver biopsies, an impractical and dangerous procedure for huge numbers of people

        -- FIB-4 scores are helpful but not completely reliable

        -- the individual risk factors, as noted above, were less predictive of either advanced fibrosis or NASH vs the combinations noted of 3-6 variables. From a primary care perspective, it really does not matter so much whether the patient has advanced fibrosis or NASH, since these conditions both dramatically increase liver-related problems and mortality.  Finding either of these leads to more intensive followup (eg ultrasound screening for HCC) and more aggressive approaches to weight loss, stopping alcohol, and meds as above

-- one other issue is that many people with NAFLD do drink some alcohol, and even alcohol consumption within the recommended limits is associated with accelerated adverse liver disease outcomes (see NAFLD any ETOH is bad BMJ2022 in dropbox, or doi.org/10.1136/ bmjopen-2021-049767). as in many studies, NAFLD is considered a diagnosis even with moderate alcohol consumption....

-- so, based on the study, if VCTE is in fact available, the study would support its use, perhaps even more so in combination with HDL and FIB-4 scores, particularly in those with higher fibrosis risk on screening.

 

Limitations:

-- this study included patients referred to hepatologists for NAFLD work-up. This may well be a select group of people, different from the many we see in primary care

-- this study did not exclude patients with alcohol consumption of <20 g/d for women or <30g/d for men, and a few studies have found that even small amounts of alcohol are deleterious. (and this cutpoint of alcohol consumption is not rigorously supported, especially with the above-mentioned study and others on cardiovascular effects of alcohol suggesting that there is no safe level of alcohol consumption: eg see http://gmodestmedblogs.blogspot.com/2023/04/alcohol-consumption-small-amounts-not.html ). And there was no information about the level of alcohol consumption in this study, or if there were different outcomes based on levels of that consumption

-- no one in the study was on an SGLT-2 inhibitor, currently a strong choice for people with diabetes, which might limit the generalizability of their results

-- it is always a bit shocking to me that only 65% of the patients with diabetes were on lipid-lowering agents: the American Diabetes Association recommendations that all adults with diabetes from age 40-75 be on at least a moderate-intensity statin as well as aggressive promotion of a healthy lifestyle, likely related to the higher likelihood of the much more atherogenic small, dense LDL particles (eg see http://gmodestmedblogs.blogspot.com/2023/10/update-ascvd-risk-factor-critique.html )

-- many of the risk factors noted above were classified at binary variables, and there might be a large gradation of risk depending on the actual numbers. Are there important differences if the BMI or waist circumference was assessed as a gradation vs above or below a single value? Or the lipid values? etc

-- this study was based on patients with diabetes and results may not be generalizable to those with NAFLD who do not have diabetes

 

so, a few points:

-- NAFLD is really common and increasing in prevalence over time, with an associated increase to the more pathologic evolution to advanced fibrosis and non-alcoholic steatohepatitis (NASH)

-- given this prevalence and the many liver-related as well as cardiovascular and mortality increases in those with NASH, we really do need a systematic approach to screening, diagnosis and treatment, though this is hampered by the inaccuracies of the non-invasive measurements and the problems with performing the definitive liver biopsy diagnosis (adverse effects, limited ability to do this many biopsies, cost...)

-- this article adds some important conceptual and practical approaches to developing a more accurate non-invasive strategy of combining an array of measurements (eg FIB-4 along with some other values) to help predict the presence of fibrosis.  these newer algorithms would need to be assessed in larger studies, with documentation by liver biopsy

-- the next blog on "lean" NAFLD will review the data on this entity and further suggestions on screening and diagnosis of NAFLD

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

if you would like access to the dropbox for articles, go to https://www.dropbox.com/scl/fo/vj803z91w1trd471h9fj8/h?rlkey=klpxdjpdhcdt3sahnpirzz730&dl=0

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list