Gretchen's postprandial diet experiment

Gretchen sent me the results of a little experiment she ran on herself. She measured blood glucose and triglycerides after 1) a low-fat diet and 2) a low-carb diet.

Gretchen describes her experience:

Several years ago I received a windfall of triglyceride strips that would expire in a week or so. I hated to waste them, so I decided to use them to test my triglyceride and BG responses to two different diets: low carb and low fat.

The first day I followed a low-fat diet. For breakfast I ate a lot of carbohydrate, including 1 oz of spaghetti cooked al dente and ¾ cup of white rice. For the rest of the day I ate less carbohydrate but continued to eat low fat.

The second day I followed a low-carb diet. For breakfast I ate a lot of fat, including a sausage, mushrooms fried in butter, 2 slices of bacon, and ¼ cup of the creamy topping of whole-milk yogurt. For the rest of the day I ate less fat, especially less saturated fat, but continued to eat low carb.

Both days I measured both BG and triglyceride levels every hour until I went to bed. On the low-carb day I had 3 meals. On the low-fat day, I was constantly hungry, had 4 meals, and kept snacking.

You can see the results in Figure 1. On the low-fat diet, after a “healthy” low-fat breakfast of low-glycemic pasta with low-fat sauce, my BG levels shot up to over 200 mg/dL and took more than 6 hours to come down. My triglycerides, however, remained low, and at first I thought perhaps the low-fat diet might be better overall. However, after about 6 hours, the triglyceride levels started to increase steadily, and by the next morning, they were higher than they had been the day before.
On the low-carb diet, my BG levels stayed low all day. However, after meals, the triglyceride levels skyrocketed. After meals they came down, and by the next morning they were lower than they had been the day before.

As I interpret these results, the high triglyceride levels after eating the high-fat meals represent chylomicrons, the lipoproteins that transport fat from your meals to the cells of your body. The high triglyceride levels the morning after eating the low-fat meals represent very low density lipoprotein, which takes the cholesterol your liver synthesizes when your intake of dietary cholesterol is low and distributes it to cells that need it, or again, to the fat for storage.

There are several interesting factors to consider here. First, when you have a lipid test done at the lab, it’s usually done fasting, which means first thing in the morning after not eating for 8 to 12 hours. It tells you nothing about what your triglyceride levels were all day.

Second, the low-carb diet resulted in lower fasting triglyceride levels, but much higher postprandial triglyceride levels. Which are more dangerous? I’m afraid I don’t know. You should also note that the high-fat, low-carb breakfast was extremely high in fat, including saturated fat. I don’t normally eat that much fat but wanted to test extremes.

Third, although the low-fat diet didn’t produce the very high postprandial triglyceride levels that the high-fat diet did, it produced extremely high BG levels that persisted for 6 hours. Some people think that it’s oxidized and glycated lipids that are the dangerous ones, so high BG levels and normal triglyceride levels might be more dangerous than very high triglyceride levels and normal BG levels. Note that high BG levels also contribute to oxidation rates.

Fourth, this shows the results of an experiment with a sample size of one. My physiology might not be typical. If you want to know how your own body’s lipids respond to different types of diets, you should get a lipid meter and test yourself. Unfortunately, your insurance is unlikely to want to pay for this, so it will be an expensive experiment.

The main point of this is that the results of different diets are complex. We have to eat. And what we eat can affect many different systems in our bodies. Finding the ideal diet that matches our own physiology and results in the best lipid levels as well as BG levels is a real challenge.

This was a lot of effort for one person. Thanks to Gretchen for sharing her interesting experience.

Gretchen makes a crucial point: Some of the effects of diet changes evolve over time, much as triglyceride levels changed substantially for her on the day following her experiment. Wouldn't it be interesting to see how postprandial patterns develop over time if levels were observed sequentially, day after day?

The stark contrast in blood sugars is impressive--Low-carb clearly has the advantage here. Are there manipulations in diet composition in low-carb meals that we can make to blunt the early (3-6 hour) postprandial lipoprotein (triglyceride) peak? That's a topic we will consider in future.

More of Gretchen's thoughts can be found at:

Comments (11) -

  • ET

    11/27/2009 5:10:43 PM |

    Interesting results.  I follow a low-carbohydrate (<60 g/day) eating plan and I recently had blood drawn and sent to a lab for glucose and cholesterol testing.  It was supposed to be fasting, but the sample was drawn 4.5 hours after I had gotten up for the day.  By then, I'd already eaten breakfast several hours earlier and my coffee with coconut creamer and  half-and-half which represents around 60 grams of fat and 9 grams of carbs.  Both my  glucose and triglycerides were 91.  My total cholsterol was unchanged.  My fasting triglycerides are usually around 45.  I did exercise prior to having the sample taken which could influence my triglycerides.

    On an earlier occassion, I also had a non-fasting cholesterol test performed by a lab and the sample was taken mid-afternoon.  I'd consumed around 150g of fat total that day, starting nine hours earlier and my triglycerides were 79.

  • DrStrange

    11/27/2009 6:07:28 PM |

    Having done similar myself though only testing blood sugar, I can say for certain that if you ate a truly low fat diet (<10% calories from fat) for 2 weeks prior to the test and then ate a truly low fat meal, your blood glucose curve would have been similar to that for the low carb meal.  The spike comes from insulin resistance, largely caused by dietary fat.  It takes about 2 weeks on a low fat diet for that component of IR to be reversed.  I have repeated this same experiment several times with identical results.

    There are two ways to keep a fairly flat sugar curve.  One is a very low carb diet, the other a very low fat diet.  For the low fat diet to work however it must be constant without cheating.  In my experiments I found that only one meal of "normal" fat content, increased IR and caused sugar spikes for many days after.  Over about 10-14 days, my post postprandial sugar curve returned to normal.

  • Nigel Kinbrum BSc(Hons)Eng

    11/27/2009 8:09:00 PM |

    Any chance of persuading someone (Oxford Group, say) to do a randomised crossover intervention trial with a suitable washout period and using different Carb/Fat percentages e.g.
    15P, 55C, 30F (Standard American/English Diet) alternating with:-
    15P, 5C, 80F
    15P, 15C, 70F
    15P, 25C, 60F
    15P, 35C, 50F
    15P, 45C, 40F
    15P, 65C, 20F
    15P, 75C, 10F but without adding extra sugar? The trial previously mentioned has received criticism on a board I post on for adding extra sugar to exaggerate results.

  • Gretchen

    11/28/2009 2:12:20 PM |

    DrStrange, perhaps it was not clear from the quoted material, but I am diabetic. My BG would go up on a low-fat diet. That's what I was on for about 6 months after diagnosis, and I certainly did not have normal blood glucose levels. My A1c was much higher than it was on low-carb, high-fat.

    However, it would be very interesting if nondiabetic people repeated my experiment. I think exaggerated TG responses may be caused by whatever it is that causes the diabetes.

  • Anonymous

    11/28/2009 3:01:50 PM |

    I think DrStrange illustrates that there must be momentum to physiological responses to what is a typical diet for any one person. A low carber's insulin levels may not be ready to handle an untypical (for them)carbohydrate load while a person with a low fat diet maintains higher insulin at all times.

  • Gretchen

    11/28/2009 3:32:59 PM |

    My experiment was flawed in several ways. I was not planning on publicizing the results, so I didn't weigh food or make sure the protein amounts were the same.

    And I didn't do the usual 3-day high-carb eating to make sure I was producing carb-producing enzymes. Not eating carbs can produce diabetes blood glucose (BG) levels in a nondiabetic: called "starvation diabetes" because starvation is the ultimate LC diet.

    However, I already knew I was diabetic. And people with diabetes have different lipid responses, so someone without diabetes would probably not see such a dramatic difference.

    A good experiment for a nondiabetic who has been on a LC diet would be the following:

    1. Measure BG and TGs for a day on your usual LC diet. Weigh the food so you can get exact nutrient levels. This would tell you what happens when you're adapted to the LC diet. Probably measuring every 2 hours would be enough.

    2. Eat a low-fat diet without preparation and follow BG and TGs for a day.

    3. Continue to eat the low-fat diet for a minimum of 3 days, maybe a week or 2 as suggested by DrStrange. Then repeat the experiment in No. 2, eating exactly the same thing.

    4. If you wish, repeat the experiment in No. 1, to see if your lipids went high when you weren't adapted to the LC diet.

    I do know someone else who was concerned that she might be diabetic, and while on a LC diet, her BG levels in a home "glucose tolerance test" approached diabetic levels. She then went on some kind of vegan diet, and on this diet, her BG levels stayed low on the GTT.

    So yes, not eating carbs and eating more fat (two different inputs) can produce a result that looks like diabetes. Which of these two factors is more important is not clear because when you reduce one nutrient you have to increase another to keep the calorie content constant.

  • DrStrange

    11/28/2009 3:45:39 PM |

    Gretchen, were you consistently at about 10% total calories from fat on your "low fat diet"?  Fat intake for most needs to always be not more than about 10% of total calories  for insulin resistance to be dramatically reduced. If someone is insulin dependent, the insulin they take will be maximally effective in that case but the fat intake must be at that low level for every meal ongoing.  One meal with excess fat and IR can shoot up and take many days to come back down.  This is what I found in my body.  Also, this is presupposing that all carbs are coming from very few fruits, vegetables, whole/intact grains (not flour, not sugar, etc)

    I really do not know which is healthier in the long run for someone who is insulin dependent, low fat/high carb or high fat/low carb.  There are studies showing increases, long term, of a number of health problems w/ high fat intake.  On the other hand, w/ the low fat diet you would need to use more insulin which is inflammatory.

  • DrStrange

    11/28/2009 3:59:33 PM |

    Anonymous, I do not have high insulin levels, in fact the opposite!  My A1c was creeping up a couple years ago (peaked at 6.4) so i tried low carb for 9 months and felt worse and worse the entire time.  More research and switched to low fat/high carb.  I have been on a very low fat, vegan (McDougall) diet for about 20 months now.  My A1c last tested 5.1 fasting insulin is <2.00 uIU/ml.  Recent glucose tolerance test with insulin values showed this after fasting ingestion of 75 gm glucose (I am only 5'3" 110 pounds:

    BG (mg/ml) fasting=82, 1/2 hour=114 1 hour=103, 2 hour=86, 3 hour=100

    insulin (uIU/ml) fasting<2, 1/2 hour 8.5, 1 hour=11.0, 1 1/2 hour=13.4, 2 hour=18.1

  • Gretchen

    11/28/2009 9:56:57 PM |


    Maybe we should discuss this privately, as the best way to treat diabetes is really not the topic of this blog. There are so many variables. You can get my e-mail on my blogsite, which is at the end of the stuff that Dr. Davis cited. Use Wildly Fluctuating.

  • buy jeans

    11/2/2010 8:21:09 PM |

    Gretchen makes a crucial point: Some of the effects of diet changes evolve over time, much as triglyceride levels changed substantially for her on the day following her experiment. Wouldn't it be interesting to see how postprandial patterns develop over time if levels were observed sequentially, day after day?

  • Dharini

    11/13/2011 11:56:52 PM |

    De novo lipogenesis is almost non-existant in humans. Trigs being high the day following a high carb meal probably reflect the regular trigs levels for a particular person. This would be true for any macronutrient composition that is not low in carbs.

    Naturally after a low carb meal there is an increased turnover of lipids because the brain does not have any glucose or glycogen sources for fuel. Thus, triglyceride levels fall ad fasting triglyceride levels are low.

    More importantly, fasting triglycerides DO NOT independently predict risk of heart disease once you adjust for post-prandial triglycerides. (Bansal et al_JAMA 2007_Fasting Compared With Nonfasting Triglycerides and Risk of Cardiovascular Events in Women)

    The effects of a low carb diet on postprandial triglycerides has been measured over 4 weeks in a study by Natalie et al_Diabetes Care 2009. They also found a worsening of triglycerides post prandially.