October 4, 2009

What Is Really Known about Fatty Liver?

The latest fad diets promise to reduce the fat in your liver. This is a noble aim--the amount of fat in your liver correlates closely to cardiovascular risk. Liver fat also appears to increase insulin resistance, though it is interesting to note that very recent research has found that pancreatic fat has a strong relationship to the development of Type 2 diabetes and may in fact be more important than liver fat.

The problem with the premise behind the fad diets, though, is that there is no significant research showing any treatment, be it a drug, diet or exercise, to be uniformly effective in reducing pre-existing liver fat.

The studies usually cited to advance one or the other dietary approach as a cure for fatty liver are either correlational studies, rodent studies, or studies that examine only surrogate markers--lab tests--rather than studies that biopsy the liver to see how much fat is in it.

The correlational studies are those where the researchers link a population's reported food intake with their risk of developing fatty liver. This kind of research may point to factors that cause fatty liver--though it also may not, given the poor track record of these kinds of studies, since they are based on questionnaires that do not accurately reflect what people really eat.

The correlation studies suggest that eating a high carbohydrate diet raises the risk of developing fatty liver. Fructose in particular seems to pile on the liver fat but any diet that results in high triglyceride levels looks like one that can damage the liver.

But the problem in applying this data to reversing liver disease is this: Once you develop a condition that correlates to a food intake pattern, it is not at all clear that eliminating the causative foods will eliminate the condition.

In the case of fat-related problems, you have only to remember that once new fat cells form, they never go away. You can reduce the amount of fat in each cell, but the cells themselves remain. (This is one reason why Avandia and Actos are such worrisome drugs, because they work by creating new fat cells in the arms and legs, cells that you are stuck with for life after taking these drugs.)

So once you have created a new population of fat cells in your liver, while it is possible dietary interventions may reduce the fat they contain, you are not going to get rid of the cells.

Beyond that, there is the problem that if the fat becomes inflamed--which it often is in NAFLD--you are likely to develop fibrosis--scarring--in the liver. This scarring is a large part of what destroys liver function. "Reversing" scarring is a whole different proposition than preventing that scarring from occurring.

With this in mind, I have spent quite a lot of time looking for studies that demonstrate that specific treatments cause significant real improvement in existing fatty liver disease. If we define "significant" not in the statistical manner but in terms of whether it makes any difference in your health. They are rare.

To understand this you need to understand that statistical significance refers to any difference between two numbers that is large enough not to be attributable to chance fluctuations. So one job pays $50,000 and another $50,500 the difference in salary here may well be statistically "significant." But you wouldn't leave your current job just because you could earn that extra $500. For the difference to be significant in terms of value, you'd probably want the second job to pay at least $55,000. Maybe more.

So when you see that a drug made a "significant" difference in fatty liver, but the research (often paid for by a drug manufacturer) avoids quantifying the actual difference in the abstract--the only part of the study most people ever see--you can safely bet that the actual amount of change wasn't anything you'd pay the inflated price of the drug to achieve.

Almost all studies claiming that this or that treatment reduces liver fat report surrogate markers--lab test results, rather than biopsy findings. This is not surprising. It's easy to take blood, not so easy to remove slices of a human liver. So almost all of the studies about fatty liver treatment evaluate the effect of the treatment on liver enzymes, specifically ALT.

Because liver biopsy is a massively expensive undertaking, the few human studies that do it are often done in people with other conditions that require abdominal surgery--conditions that might skew the results. Or they involve a very small sample--small enough that the results are not compelling.

The majority of studies that present liver biopsy results are rodent studies, and as we know, rodents are not furry little humans. Invariably they are done with rodents fed a very high fat diet--one that is very different from the diet rodents have evolved to eat and which cause changes in the rodent body often quite different from the effect they have on human bodies.

So the bottom line is this: there is a very small amount of evidence showing that low carb diets and exercise might prevent fatty liver, though even here it is hard to know how seriously to take them because they are so often based on surrogate markers.

There is virtually no evidence that conclusively proves that any treatment reverses fatty liver in everyone or even the majority of those who pursue it.

There is an excellent balanced review of Fatty Liver diagnosis and treatment on Medscape which I urge you to read if you are interested in this issue. It includes an extensive list of research references. I will cite some points made in it, but the whole thing is worth your attention.

Medscape: How Should We Manage Patients With Non-alcoholic Fatty Liver Disease in 2007? Henry L-Y Chan; H. Janaka de Silva; Nancy W-Y Leung; Seng-Gee Lim; Geoffrey C. Farrell; the Asia-Pacific Working Party on NAFLD (2007)

Do You Have Fatty Liver?

An important point this review brings out is that though they are usually used to diagnose fatty liver, it turns out that ALT levels (ALT is a liver enzyme released when the liver is damaged) do not correlate well with how much fat or scarring you have in your liver.

The review explains,
Serum ALT level cannot accurately reflect the severity of NAFLD [non-alcoholic fatty liver disease]; a significant proportion of patients with normal ALT have bridging fibrosis and even cirrhosis on histology. An AST to ALT ratio more than 1.0,[19] increased serum hyaluronic acid[20] and high homeostasis model assessment–insulin resistance (HOMA-IR) score, which is calculated by serum glucose × serum insulin (both in mmol/L) divided by 22.5,[21] are the common biomarkers associated with liver fibrosis.
However the review goes on to state that the predictive value of all biomarkers, including some fancy ones tested only in studies, is only 75-80%, and the only truly reliable diagnostic method is liver biopsy.

If your doctor has told you that you have fatty liver disease, ask on what basis he or she makes this diagnosis. Chances are it is your ALT level or the ratio of ALT to AST.

If You Did, Why Your "Cure" Probably Wasn't A Cure

The problem with ALT is that you can make it go down with many different treatments, without having any actual impact on the composition of your liver.

The review reports later,
Although elevated ALT is a well-recognized feature of NAFLD, ALT elevation can be present or absent in both simple steatosis and NASH [nonalcoholic Steatohepatitis--liver inflammation]. In most studies on lifestyle modification and pharmacological treatment in NAFLD, improvement in ALT is usually accompanied by improvement in histologic steatosis grading. The relationships between changes in ALT and changes in necroinflammation or fibrosis are less consistent. In a study in Hong Kong comparing patients with simple steatosis and NASH, patients with more advanced disease tended to have lower ALT levels.[7] Lowering of ALT should not therefore be considered a valid marker for improvement of liver disease in the follow up of patients with NAFLD.[Emphasis mine]
Once you realize that lowering ALT levels may not mean you have improved the state of the liver, several "cures" lose their efficacy.

For example, Metformin. Metformin can dramatically lower the ALT liver enzyme, but a recent study that did the liver biopsies showed no actual improvement in the subjects' livers.

Metformin in patients with non-alcoholic fatty liver disease: A randomized, controlled trial.
HAUKELAND John Willy, et al. Scandinavian journal of gastroenterology 2009, vol. 44, no7, pp. 853-860

There are a few small studies that claim that Avandia or Actos made "significant" improvements in NAFLD on biopsy, but because the abstracts do not present the numbers, the "significance" might well be a tiny amount of change that doesn't mean much clinically. There are conterbalancing studies that show no effect from Avandia or Actos, too. My guess is that examination of who paid for these studies--drug company or others, may explain the difference in their findings.

The Medscape review suggests, rightfully, that the weight gain side effect of Actos and Avandia may counterbalance any small change they make in liver status.

Weight Loss Surgery is being promoted very heavily as a cure for NAFLD (and just about everything else.) The studies demonstrating improvement involve very few subjects and because of the huge profits to surgeons and hospitals involved in this surgery one should approach these studies with the same skepticism with which you approach drug company sponsored studies. The other side effects of WLS can be life-ruining and/or even fatal, so it is not a "cure" to be adopted without much prior study and investigation.

Fatty Liver and Insulin Resistance

Most researchers seem to believe that fatty liver is linked to insulin resistance though whether it is caused by the IR or causes it is not at all clear to me. Nor is it clear that the idea common in research that lowering insulin resistance will improve fatty live is realistic.

Why? Because truly reducing insulin resistance, rather than blood sugar, appears to be almost impossible. This truth is obscured by faulty methods of measuring IR that do not account for the impact of lowering carbohydrate intake on blood sugar levels.

Most people who have Type 2 diabetes remain seriously insulin resistant no matter what drug they take or what diet they eat. Diet, especially the low carb diet, can normalize their blood sugar, eliminating most diabetic complications, but the amount of insulin the person with Type 2 Diabetes needs to inject to lower their blood sugar 10 mg/dl is always much higher than that a normal person would need. Anywhere from five times higher to twenty times higher.

Losing weight doesn't normalize insulin sensitivity in most people with Type 2 Diabetes either. We know thin, active young relatives of people with Type 2 already show higher than normal insulin resistance. So it is possible the IR associated with fatty liver precedes the depositing of liver fat rather than, as is commonly assumed, the fat is causing the IR. In any case, because we have no reliable method of reducing insulin resistance that doesn't have side effects worse than the conditions (viz Avandia and Actos) fad diets that claim to reduce insulin resistance, which actually only reduce blood sugar, are not going to reverse your fatty liver.

What we can conclude, tentatively is that diets that lower blood sugar and triglycerides are better than those that raise them. There is some possibility that they may prevent further deterioration of fatty liver disease though until we understand the real causes of insulin resistance, this is only a guess.

Such diets, and metformin, often will normalize liver enzymes, the significance of which is unknown. What we do not know is whether any diet will lower the amount of fat stored in the liver.

But there is another huge issue to consider when reviewing a fad diet: Diets that cause very fast weight loss have been linked to raising the risk of developing fatty liver. So the last thing you want to do if you are fighting fatty liver is adopt a diet that promises to take off a lot of weight in a very short time. If you are losing more than a pound a week (two if you are seriously obese) you may be stressing your liver.

If your fad diet is high in carbs, you are most certainly stressing your liver and making your fatty liver worse.

Did You Really Reverse Your Fatty Liver? Normal ALT is Not What The Lab Range Tells You

Keeping in mind that ALT levels don't tell you if you have reversed your fatty liver, I discovered something rather disturbing. It turns out that the lab ranges currently used for "normal" on ALT tests are (like glucose reference ranges) much too high, probably because the populations used to establish them included large numbers of people with undiagnosed fatty liver disease. It also turns out that males and females have very different truly normal ranges. The true female normal range is much, much lover.

A study that established true normal for ALT values is found here:

Re-evaluation of Serum Alanine Aminotransferase Upper Normal Limit and Its Modulating Factors in a Large-Scale Population Study Moshe Leshno et al. Liver International Full Text on Medscape.

When I looked at my own liver function tests over the years, keeping in mind the data presented here, and also using a biomarker the review suggested, the ratio of ALT to AST, I discovered that in liver function tests performed at several different times back when I had been eating a ketogenic low carb diet of no more than 65 grams of carbohydrate a day for more than 2.5 years, my AST was well over the normal range for females and my AST/ALT ratio was high enough to suggest NAFLD.

These biomarkers normalized completely when I started Metformin--at which point my high end of normal triglycerides also plummeted, but as the biopsy results suggest, this doesn't mean that my liver is in any better shape.

I also learned that another sign of fatty liver disease may be very high iron levels in the absence of the gene for hemochromatosis. Another member of my family who has a different kind of diabetes--one that has the pattern of very high fasting blood sugar but normal post-prandial blood sugar, was diagnosed with this very high iron level. He is at a normal weight but carries a gene that is associated with lipid abnormalities. This makes me wonder about whether the minority pattern of diabetes characterized by high fasting glucose but normal post-meal numbers points to fatty liver disease even in normal weight people.

So What to Do?

There really is no easy answer here. Slow steady weight loss might help, except that the weight of the evidence relating to weight loss is that it is almost impossible to maintain a weight loss of more than 20% no matter what diet you eat. I had made that kind of weight loss several years before my ALT/AST values tested ugly.

Exercise may help if it truly reduces insulin resistance rather than just blood sugar. Whether this happens has a lot to do with your own unique physiology and what causes your individual insulin resistance.

Avoiding the high carb foods and especially fructose that appear to worsen liver fat is a good idea.

There is also some question about the impact of the statin drugs given that they often worsen liver enzymes. Very little of the research addressing this examined the livers of those taking statins.

A small study that did came up with conflicting data:

Statins in non-alcoholic fatty liver disease and chronically elevated liver enzymes: A histopathological follow-up study Mattias Ekstedt et al. Journal of Hepatology Volume 47, Issue 1, Pages 135-141 (July 2007).

Here are the findings of this study:
At baseline, patients that later were prescribed statins had significantly higher BMI and more pronounced hepatic steatosis. At follow-up patients on medication with statins continued to have significantly higher BMI. Diabetes was significantly more common among patients on medication with statins and they had significantly more pronounced insulin resistance. However, they exhibited a significant reduction of liver steatosis at follow-up as opposed to patients not taking statins. Despite exhibiting a high risk profile for progression of liver fibrosis, only four patients on statin treatment progressed in fibrosis stage.

Final verdict

Sometimes it helps to know what it is you don't know. What we don't know is how to reverse fatty liver in a way that is safe and improves health. So be very wary of any mass market diet book that claims that its diet will reverse your fatty liver. There is not enough good quality data to make any prescription here.



JD said...


"People on low-carbohydrate diets are more dependent on the oxidation of fat in the liver for energy than those on a low-calorie diet, researchers at UT Southwestern Medical Center have found in a small clinical study."

JD said...


"A new study shows that an imaging technology developed by Mayo Clinic researchers can identify liver fibrosis with high accuracy and help eliminate the need for liver biopsies. Liver fibrosis is a common condition that can lead to incurable cirrhosis if not treated in time."

"Patients with nonalcoholic fatty liver disease and no significant inflammation or fibrosis were identified with 97 percent accuracy."

Appears that the tools to detect NAFLD without liver biopsies are starting to become available.

Jenny said...

The Mayo technique may work but it is one of those very expensive new techniques that isn't available to most people or if it is, the copay can be $1000 or more.

The low carb diet study cited was a two week study. All of us who have low carbed for any significant amount of time know that the physiology of those first two weeks is very different from what we experience after several months when the body has adapted to being in a ketogenic state.

Kathy W. said...

Good points re the "evidence" and deficiencies in the studies!

As to fat cells, however, there are none in the liver--fatty acids are taken up by the liver for processing, and so liver cells naturally contain varying amounts of fat. Fat from liver cells is released (or not) into the bloodstream, where it can end up in adipose tissue or muscle elsewhere in the body. Likewise, any fat in the pancreas is in the form of droplets in the beta cells themselves or elsewhere in the pancreas.

Even the true fat cells aren't forever. New studies have indicated that fat cell apoptosis goes on all the time. But the fat cells are, apparently, longer-lived than many other cells.

It seems like there's a strong genetic component, as you've noted, to the liver (and pancreas) storing too much free fat from the bloodstream. But since different foods and exercise can turn genes on and off, there's probably a combination that's optimal for people with this particular tendency...