By Dr. Geoffrey Modest
This is the second part of the series on NAFLD, a review of therapies, with more detail on a couple of topics (e.g. the roles of fructose and the microbiome)
(see Hannah WN. Dig Dis Sci 2016; 61:1365)
It is pretty clear that weight loss in those overweight/obese is the best documented therapy for NAFLD. Those who can lose 3-5% of their body weight tend to improve hepatic steatosis; and those who can lose >5% in one large study found 58% had resolution of NASH and 82% had a 2-point reduction in their NAS (NAFLD Activity Score on biopsy). Those who lost >10% of their body weight (n=29) had 100% had resolution of NAS, 90% had resolution of NASH, and 45% had regression of fibrosis. Other studies suggest that the type of diet is not so important: it’s the weight loss. But, we know that losing weight is really hard, and maintenance of weight loss is even harder (e.g., see http://blogs.bmj.com/ebm/2016/08/17/primary-care-corner-with-geoffrey-modest-md-weight-loss-and-resting-metabolic-rate/ ). So, it is useful to look at individual components (micronutrients, macronutrients, etc.) to see if there can be some utility of specific dietary changes, even in patients who cannot lose sufficient weight.
- Small studies suggest that, comparing Mediterranean diet vs low fat diet and achieving comparable weight loss and with a cross-over design: there was decreased hepatic steatosis and improved insulin sensitivity with the Mediterranean diet.
- Observational and some intervention studies find exercise without weight loss in patients with NAFLD improves markers of insulin resistance and liver enzymes, some find decrease in hepatic steatosis, some suggest a dose-response curve with the more exercise the better. But other studies did not find much effect.
Diet and exercise
- One study of 293 patients with histologic NASH on a 52-week diet and exercise program leading to a mean weight loss of 4.6 kg found: 25% had resolution of NASH, 47% a reduction in NAS, and 19% a resolution of fibrosis, all with a dose-response relationship.
Monounsaturated fats (an important component of the Mediterranean diet)
- Very small studies suggest decrease of steatosis with high monounsaturated fat diet, but hard to disassociate this from other dietary changes (see Mediterranean diet, above)
Polyunsaturated fatty acids (PUFA’s, esp n-3)
- One study found that supplementation with 4 g/d of PUFAs led to a significant improvement in hepatic fat and ALT levels. Another, at lower dosages (up to 2.7 g/d) found no benefit, including in biopsies
Fructose: see next article, but probably really important causally in NAFLD
- A few studies have found some benefit: the PIVENS study (see N Engl J Med 2010;; 362: 1675) randomized 247 nondiabetic patients with NASH to vitamin E 800 IU vs pioglitazone 30mg vs placebo, found that vitamin E led to improvement in NAS (43% vs 19% on placebo) though nonsignificant improvement with pioglitazone vs placebo (34% vs 19%, which actually had p=0.04 but not considered significant in this study). Both led to improvement in LFTs, as well as decreases in hepatic steatosis and lobular inflammation, but there was no improvement in fibrosis with either therapies. Another trial in kids comparing vitamin E 800 IU vs metformin 1000mg vs placebo found that vitamin E led to significant improvement in hepatocellular ballooning and NAS, and those with NASH had significantly more resolution of the NASH with vitamin E than metformin. There are some concerns that higher doses of vitamin E is associated with increased all-cause mortality (see Miller ER. Ann Intern Med 2005; 142:37), leading some specialists to suggest the unproven benefit of using only 400 IU/day, and pioglitazone was associated with significant weight gain in the PIVENS and other pioglitazone studies.
- The wonder drug…. though in this case the data are not so compelling: mostly retrospective analyses, with a couple of studies: a population-based study finding that those drinking >= 3 cups/d had lower odds of significant fibrosis; another found less NASH on biopsy, and those who developed NASH had lower fibrosis scores
- There are interesting studies finding disruptions in the gut microbiome are associated with the pathogenesis of NAFLD and the progression to NASH (see http://blogs.bmj.com/ebm/2016/03/18/primary-care-corner-with-geoffrey-modest-md-microbiome-changes-and-severity-of-nafld/). And see the review below. A few remarkably small trials have found some benefit from probiotics: one had 20 patients with NASH followed 6 months finding improvements in intrahepatic triglycerides, and reduction in ALT; another of 44 obese kids found significant improvement in hepatic steatosis by giving a product containing one strain of streptococcus thermophilus, three strains of bifidobacterium and four strains of lactobacillus in defined ratios.
- The data on interventions to improve NAFLD is quite impressive for weight loss. Unfortunately, the studies on most of the more specific dietary and other interventions is quite limited by very small short-term studies and of mixed quality. That being said, I think there are enough data to support the following interventions:
- which also helps with initiating and maintaining weight loss through diet, and has a multitude of physical and psychological benefits
- Mediterranean diet (low fat, high fruit/vege) seems to help with NAFLD, as well as improving insulin sensitivity/diabetes/metabolic syndrome, and general cardiovascular risk reduction (and the old secondary prevention Lyon Heart Study found decreased recurrent cardiovascular events, on the order of taking statins).
- Monounsaturated fats. Really minimal data on benefit for NAFLD per se, but pretty clear that they improve lipid profiles and improve glycemic control. Also data that they have anti-inflammatory properties and improve endothelial function. And this is a relatively easy intervention for many people: change their oils to olive/canola/etc., avoid saturated fats…
- N-3 polyunsaturated fats (e.g. fish oils). Also not much data on NAFLD, but PUFAs are depleted in the liver in those with NAFLD, and there are limited data that they improve steatosis. Also there are reasonably good data that eating more fish is good for many clinical outcomes (heart, etc.). In terms of supplements, it is not really clear how much is needed to take to get a positive effect (or the specific composition of those supplements), and this also depends on background consumption of fish in the diet. And the caveat: high fat fish, rich in PUVAs, also tend to concentrate more environmental toxins in the oils (many toxins are fat-soluble), leading to potential concerns about mercury poisoning, etc.
- There are other medical therapies in the wing, though it is clear that diet is the most effective so far.
Role of fructose: (see or Softic S. Dig Dis Sci 2016; 61:1282)
- Fructose may well be the worst dietary actor:
- Fructose is the most commonly consumed sugar, largely through high fructose corn syrup
- The major hepatic abnormality leading to increased hepatic lipids in NAFLD is enhanced de novo lipogenesis (DNL)
- Dietary fructose increases DNL more strongly than a high-fat diet
- Fructose metabolism: absorbed via portal vein (higher concentrations in liver than other tissues); increases protein levels of all DNL enzymes in its conversion to triglycerides; fructose does not require insulin for its metabolism, furthering lipogenesis in setting of insulin resistance, fructose leads to ATP depletion and suppression of mitochondrial fatty acid oxidation (and more reactive oxygen species); and it promotes uric acid production (which itself may promote hepatic steatosis by the generation of mitochondrial oxidative stress). And fructose potentiates its own metabolism (i.e., no negative feedback loop).
- Short-term carbohydrate overfeeding in overweight humans for 3 weeks led to 27% increase in liver fat, while total body weight increased only 2%; a 6-month hypocaloric diet in the same people led to 25% of loss of liver fat and 4% decrease in body weight: all reinforcing the profound and disproportionate effect of carbohydrates on liver fat. High fat also has a role, and some studies suggest that this is predominantly by saturated trans fats. But the data for fructose may be more robust, given the above. Studies have found that fructose-sweetened drinks (but not glucose) is associated with increased visceral adiposity, insulin resistance and hepatic de-novo lipogenesis (as noted above); also more hepatic inflammation and higher likelihood of progression of liver disease to fibrosis. Conversely, looking at adults with biopsy-proven NAFLD, their fructose consumption is 2-3 times higher. In kids, sugar and fructose consumption is higher than adults (and even though sucrose is 50% fructose, this is a lesser source than high-fructose corn syrup)
- Animal models show that a high fructose diet leads to NAFLD (as well as changes in microbiome and increased endoplasmic reticulum stress and apoptotic activity)
- 1 billion people worldwide have NAFLD.
- The consumption of fructose has dramatically increased in the US: before 1900, it was 15g/d (4% of calories, mostly from fruits and vegetables), by 1994 it was 55g/d (10% of calories). And in adolescents it is 15% of calories, but 10% of teens consume >25% of their calories from refined sugar (and high fructose corn syrup is the major source)
- As mentioned in prior blogs (see, for example, http://blogs.bmj.com/ebm/2015/12/11/primary-care-corner-with-geoffrey-modest-md-fructose-restriction-and-cardiometabolic-and-weight-improvement/ , though a search for “fructose” at that site finds many additional blogs), I am very concerned about the amount of fructose (largely high fructose corn syrup) in the diet. I have had several patients with hyperuricemia and gout who were able to stop consumption of soft drinks, and had significantly lowered their uric acid levels. This physiologic relationship between fructose and NAFLD is another strong reason to encourage decreasing fructose consumption (especially since NAFLD is still pretty common in patients without glucose intolerance and who are not overweight…). And the low-hanging fructose fruit for most of my patients is soft drink consumption, which seems to be the easiest profound dietary change for many.
Microbiome. (see Abdou RM. Dig Dis Sci 2016; 61: 1268)
- 20-30% of adults with NAFLD develop NASH, and a subgroup of these develop severe morbidities of cirrhosis, hepatocellular carcinoma and liver failure.
- The human intestinal microbiome has 150-fold more genes compared to the human host: the intestinal microbiome (IM) is sometimes referred to as the “missing organ”
- The liver may be a more important target of action for the IM since 70% of its blood supply is from the portal vein, directly draining the intestine
- Data summary of microbiome studies for obesity:
- Germ-free (GF) mice who were resistant to developing obesity on a high fat/high sugar diet, then had their cecal IM colonized by conventionally raised non-GF mice, had a 60% increase in body fat and increased insulin resistance (IR) within 2 weeks despite decreasedfood intake.
- Astrain of mice with the ability to increase the amount of energy extracted from food were able to transfer this ability to other mice by transferring their IM.
- GF mice colonized with IM from a lean human twin had less adiposity and weight gain vs mice colonized from the obese twin of the dyad.
- In humans, there are differences in the IM of people with obesity or NASH, with the ratio of Firmicutes and Bacteroidetes species (the vast majority of known species) disturbed in these patients. Bacteroidetes have higher levels of complex carbohydrate-digesting enzymes with the potential to harvest more energy. [SCFAs (short-chained fatty acids), in particular acetate, butyrate, and propionate) are the end-product of this polysaccharide digestion, and higher SCFAs are found in obese and overweight adults on a Western diet]. High fructose diets affect the microbiome, as mentioned above.
- Data summary for microbiome studies and NASH
- One interesting finding is that adolescents with NASH have more Escherichia species in the gut, this species is associated with more mixed-acid fermentation and the endogenous production of alcohol. And those non-alcohol drinking adolescents who have higher blood alcohol levels have more NASH. Other studies have found that obese children with ultrasound-diagnosed NAFLD have more ethanol metabolites in their stool. Also, alcohol dehydrogenase is one of the most dramatically upregulated hepatic genes in adolescents with NASH. Higher serum alcohol levels are also found in a small study adults with histologically-proven NAFLD vs controls, with NAFLD subjects actually drinking somewhat less alcohol than the controls (1.9 gm/d in the NAFLD group vs 2.4 gm/d in controls). Another study found that blood alcohol levels are positively associated with IR (insulin resistance). There are differences in LFTs between alcoholic liver disease and NAFLD (the former with AST/ALT ratio typically >2, then latter <1), which makes this a bit hard to interpret. But AST reflects more mitochondrial injury to the hepatocytes (ALT is from the hepatocyte cytoplasm, AST is 20% cytoplasmic and 80% mitochondrial). Perhaps there are nutritional differences between those who have very high alcohol intake and alcoholic liver disease vs NAFLD?? Perhaps related to the array of micro/macronutrients often found to be depleted in alcoholic patients?? Or perhaps there direct mitochondial toxic effects of higher-dose alcohol causing mitochondrial damage. I.e., it conceivable that the increased endogenous alcohol production related to microbiome changes and associated with NAFLD is causative or contributory to NAFLD in some patients, despite the different LFT presentations.
- The microbiome changes also lead to increase in permeability of the intestinal barrier, making it even easier to get endogenous alcohol-related hepatic damage.
- Also, fibrosis (likely the bad actor, as above) is often independent of steatosis, and fibrosis itself is associated with changes in the microbiome. Some studies have found some specific microbiome changes in those with higher levels of fibrosis (more Bacteroides and Ruminococcus, less Prevotella)
- So, bottom line, it does make sense to protect our “missing organ”, the microbiome. And the best protection is probably eating healthfully, exercise, avoiding antibiotics whenever possible (including those put into meats in agribusiness), avoiding added chemicals to foods whenever possible (e.g., artificial sweeteners), etc. (see http://blogs.bmj.com/ebm/category/microbiome/ for an array of blogs on the potential health effects of changes in the microbiome. but it often comes down to, as Michael Pollan says, sticking to foods your grandmother would eat…..)