Sunday, November 10, 2013

Is Postprandial LPS From Fats a Cause for Concern?

Saturated fat is blamed for all kinds of things.  Generally the blame comes from: 
  1. Observational studies
  2. Studies that find SFA increases a risk factor (like LDL-C) followed by the researchers making inferences about disease (like CVD)
  3. In vitro studies
1) Observational studies have many confounding variables and often involve self-fulfilling prophecies.  For these reasons, there are instances where RCTs have found the opposite result that the observational studies predicted (such as HRT) 

2) Just because A causes B and B causes C, doesn’t mean A causes C.  A really ironic example of when A doesn't cause C is: 

  • Many studies show that saturated fats increases LPS in the bloodstream (see below)
  • LPS is a suspected cause of alcoholic and non-alcoholic fatty liver disease [1] [2] [3] [4]
  • However, SFA seems to be is protective in a number of animal models of alcoholic and non-alcoholic fatty liver disease [5] [6] [7] [8] [9] 

3) As for in the vitro studies, it’s not too difficult to find a paper where high levels of saturated free fatty acids cause disease (here’s one on NALFD).  So we have high dietary SFA being protective but high serum SFA being pathological in NAFLD.  The likely explanation is that high dietary SFA (within reason) doesn’t cause pathological elevations in serum SFA, but that the high serum SFA in NAFLD is caused by insulin resistance and elevated de novo lipogenesis. 

Macronutrients, LPS and Inflammation 

But still, the fact that SFA raises LPS in the bloodstream could be a cause for concern. 

You may have heard of this paper: ‘Metabolic Endotoxemia Initiates Obesity and Insulin Resistance’ where a high fat diet of 72% fat (corn oil and lard), 28% protein, and <1% carbohydrate caused what they described as metabolic endotoxemia (2-3x normal levels, but 10-50 times lower than sepsis), which then caused obesity and insulin resistance [10]. 

Long chain fatty acids postprandially increases LPS as it’s transported into the bloodstream via chylomicrons and long chain fats requires chylomicrons to be absorbed [11] (MCT and SCFA won’t have this effect because they don’t require chylomicrons) 

But the saturation also seems to matter.  In pigs 50ml of coconut oil doubled serum LPS concentration, while vegetable oil only had a non-significant increase and fish oil didn’t increase the concentration at all* [12].  What’s odd is that coconut oil is mainly made up of MCTs, which shouldn’t increase LPS concentrations since they don’t require chylomicrons for absorption 

There have been a few human trials on the effect of fat on postprandial LPS concentrations: 

Following a meal of a 900 calorie meal of a cup of tea and 3 slices of toast spread with a total of 50 g butter, plasma LPS levels increased from 8.2 pg/mL to 12.3 pg/mL (a 50% increase).  (The study was done in healthy men who occasionally smoked, aged 20-58) [13] 

Following 300 calories of either cream, glucose, orange juice or water, only cream increases plasma LPS (by about 45%).  (This study was done in healthy people with a normal BMI, aged 25–47) [14] 

This study also found that glucose and cream increased a few inflammatory markers, but orange juice didn’t 

TNFa and IL-1b mRNA in MNCs
NFkB Binding in MNCs
Plasma LPS

The team that did these human studies were looking at how diet can affect inflammation.  They also conducted other studies that found: 

  • Glucose increased ROS generation in MNCs and PMNL (two immune cells) and increased NFkB binding, but fructose and orange juice did not.  It appears two phytochemicals in orange juice (hesperetin and naringenin) are responsible (300 calories each) [15] 
  • 300 calories of glucose, but not ethanol (vodka), increased ROS generation from MNCs and PMNL cells and increased NFkB binding [16]
  • 300 calories of cream or casein increased ROS generation from MNCs and PMNL cells and increased NFkB binding [17] 

* Although the SFA rich diet resulted in higher levels of tight junction proteins, which may suggest better tight junction function than PUFA rich diets 

** LPS >> TLR4
ROS >> NFkB binding >> TNFa, IL-1b >> SOCS3 

*** These studies note that the increase of LPS is unlikely to be due to LPS found in food 

Other Things Matter 

20 healthy lean (BMI <25 kg/m2) subjects 20–50 years old ate a meal of 910 calories of either a high fat, high carb meal* (egg muffin and sausage muffin sandwiches and two hash browns, which contain 88 g carbohydrates, 51 g fat [33% saturated] and 34 g protein [carbohydrates 41%, protein 17%, and fat 42%]) or an American Heart Association (AHA)-recommended meal (oatmeal, milk, orange juice, raisins, peanut butter, and English muffin (carbohydrates 58%, protein 15%, and fat 27%).  Following the HFHC meal there was an increase in several markers of inflammation such as SOCS3, TLR4, TLR2 ROS generation, p47phox expression, NF-κB–binding activity, and plasma MMP-9, LPS (47% at 3 hours) and LBP concentrations.  These changes did not occur from the AHA meal (non-significant increases) [18] 

Same HFHC meal, this time with either: water, 300 calories of glucose or 300 calories of orange juice.  The orange juice blunted much of the increase of many of the inflammatory markers (including a fairly large blunting of LPS) [19]. 

Fibre (30g) can also blunt the increase in ROS and inflammation to the same HFHC meal [20].  This is consistent with other research in rats where high fat diets lowered numbers of Bifidobacterium and increased LPS.  Then adding prebiotics to promote the growth of Bifidobacterium normalised LPS levels, perhaps due to the by-products of fermentation (SCFAs/butyrate) increasing barrier function and perhaps because Bifidobacterium are gram-positive are therefore don’t have LPS [21] 

Meals like the HFHC one mentioned above or pure fat also seem to postprandially cause endothelial dysfunction, but aerobic exercise 2h after [22] or supplementation with either arginine [23], vitamin C and E [24], fish oil [25], carnitine [26] and oats [27] prevented the postprandial endothelial dysfunction.  In case you were wondering whether this was just fat, high GI meals also seem to impair endothelial function postprandially [28] 

* Which to my way of thinking is moderate in protein, fat and carbs 


In summary: 

  • Protein, fat and glucose calories, but not fructose or alcohol, increases ROS generation and inflammation in MNCs and PMNL cells
  • Even though all calories increase inflammation, only long chain fatty acids increase LPS concentrations, although fats and glucose have similar ROS and inflammatory responses, including SOCS3
  • Orange juice and fibre quite strongly inhibit ROS generation, markers of inflammation and LPS
  • High fat and high GI meals impair endothelial function, but several things are protective 

So what to make of this?  I think there are a few possible explanations: 

  • Calories are inflammatory (fructose* and alcohol are not completely innocent), whereas many of the other substances in foods are anti-inflammatory (suggests whole foods > refined food)
  • But perhaps since protein, fat and glucose all increase ROS and inflammation, this could suggest that this is a physiological response to food for some reason (perhaps it’s a signal that the immune system has resources to function?).  And that the flavonoids in orange juice, fibre, etc are inhibiting this function.
  • In line with the previous point, the increase in LPS from fat is quite minor and transient compared to metabolic endotoxemia or sepsis.  Perhaps the increase in LPS functions to enhance immune activity [29]
  • We also probably shouldn’t infer something from the short term effects of meals.  Nor should we take a reductionist approach (nutritionism) because: (1) people eat meals/foods not nutrients so we lack the context and (2) nutrients have many effects so singling one out may miss the big picture (like the SFA rich diet resulting in higher LPS but also higher tight junction protein levels)
  • Finally, one thing I am convinced of, these studies can’t add much to the debate of whether you should get most of your calories from fats or carbohydrates, because while fats increased LPS, the reason to be concerned about LPS is inflammation, and glucose and fat had a similar inflammatory response.  Besides it seems like the presence/absence of beneficial compounds is more important (food quality > macros) 

* It’s interesting that some studies didn’t find purified fructose to have several of the negative effects that glucose did, considering that this blog post by Chris Masterjohn shows purified fructose increases trigs, decreases vitamin E, promotes inflammation and increases vulnerability to oxidative stress [30]


  1. I wonder if cream was the best example for that experiment! Not a very pure fat. Cocoa butter, ghee, or beef tallow might have been more neutral.

    This genius paper, I think, puts it in perspective: we are meant to get LPS after meals, because food can contain pathogens, and LPS improves immune "alertness" at a time of most danger.

    Re: the liver/SFA paradox; MCTs decrease LPS translocation from gut to liver, long-chain SFAs (cocoa butter) allow it but decrease liver immune cell sensitivity to LPS.

    1. True, but the total protein + lactose in the cream was quite low "the content of the dairy cream used includes 70% saturated fat, 28% unsaturated fat, <2% protein, and no carbohydrates"

      I read that paper too. It sure makes a lot of sense

      Is that is the context of alcoholic liver disease or in general?