|
|
Flaws |
Research flaws, disputes with published findings
The following are examples of results reported which we believe to be misleading. We base our opinions on careful reading of the published article, a review of its citations (the research upon which the published article is based), our analysis of test results, other relevant research, comments by other scientists, personal communications with researchers and authors, presentations at scientific meetings, and our research and clinical experience.
Due to the complex issues involved in biomedical research and lack of access to unpublished data, we present only our personal OPINIONS based on our interpretations of the data. Scientists who disagree are welcomed to provide their data for review and discussion, or to submit a dissenting view.
We frequently ask the scientists involved to allow us to review their methods and data on individual subjects (studies usually publish only statistical averages), to verify whether or not their conclusions are valid. Unless stated otherwise, authors have refused to provide the raw data used to publish their research.
Measuring Essential Fatty
acids
Essential fats cannot be measured accurately using dietary intake questionnaires. The PUFA/TFA content of the foods we eat depends on how the food was processed (before buying it), storage conditions, and how it is cooked. It also depends on where the food was grown, local temperatures, and season (which affects PUFA incorporation in foods). Fluctuations in PUFA content of animal feed + variations in animal feed caused by changes in market prices can dramatically alter the composition of poultry, pork, eggs, fish, and many other animals.
These changes in food composition are not known and therefore cannot be incorporated in the formulas used to calculate PUFA intake. Food-frequency questionnaires cannot measure the great variability in similar foods (brands, composition). Also, no one can accurately recall exactly what he eats. Thus, I submit that studies that rely in on food questionnaires to calculate intake of PUFAs and PUFA status are misleading.
The only way to accurately measure someone's PUFA status is to measure levels of fats in their blood; that is, to perform an ACCURATE fatty acid profile. It is difficult to do accurate fatty acid analysis. There are many potential sources of errors in studies and tests using fatty acid profiles.
Common errors in lipid
lowering studies (studies to lower cholesterol and triglycerides, and to
increase HDL)
The usual design consists of an "experimental group" that takes substance X (such as a drug) whose effects are to be studied, and a "control group" that does not eat X but eats a "placebo" (something that is similar in looks, taste, and appearance) and is otherwise as similar as possible to the experimental group.
Ideally, one would choose identical people for each group. Unfortunately, this is not possible with humans (although it is possible with some animals). The next best choice would be identical twins, but there are not enough of them. Next choice would be a group of randomly chosen individuals, but that is not practical (not everyone wants to disrupt his life for the sake of Science). Therefore, scientists use volunteers. Unfortunately, if these people have abnormal lipids, they have already been treated or have been following widely different diets. Also, the volunteers might have other things in common, such as a lot of free time, proximity to a hospital, interest in the problem being studied, etc, which makes them a somewhat non-random (atypical) group.
Every subject is supposed to eat the same diet every day for a period of many weeks during the study. Because it is practically impossible to monitor eating habits for very long periods of time (people start "cheating" at parties, etc), most studies last 6 to 12 weeks.
It is known that there are huge differences in the way people respond to different substances (drugs or foods) to lower lipids. There are several ways to compensate for those individual differences. A common approach is to use a very large number of people, let us say 2,000. The idea is that two groups of 1,000 randomly chosen people will be fairly similar; that is, individual differences will cancel each other out. The flaw is that the people are not randomly chosen, but are those who self-selected to participate in the study.
To compensate for these issues, researchers use several strategies:
All subjects go through a baseline period of
about 4 weeks during which they eat the same diet. Usually this is similar to
the American Heart Association diet, with ~30% of total calories as fat (10%
SFA, 10% MUFA, and 10% PUFAs). Also, scientists often give each subject a
multivitamin and mineral with 100% of the RDA, to maintain equal body levels of
these substances in all subjects.
1/2 subjects start on X and the other 1/2
start on the placebo (subjects do not know which one they are getting). After
about 6-8 weeks, they rotate. (Sometimes they go through a "wash"
period of 4 weeks of eating the standard diet again between changing
treatments.) In this manner, all subjects experience the same treatment, and so
it is possible to compare the performance of one subject with himself on both
treatment programs.
The researchers then average and report the
results of each group. These averages are practically never identical (that is,
they are almost always different). Sometimes the difference is too small to be
clinically relevant. However, other times the statistical methods may find that
the difference is not due to chance. With a large sample size, even a tiny
difference is statistically different (i.e., not due to chance), because the
statistical methods can tell that the two numbers are not equal.
"Statistically significant" differences between the averages of the
groups often means that two numbers are different and that the probability that
the difference is due to chance (random variation) is less than 5%.
In a study to lower cholesterol, a 10% change can easily be achieved by many factors or by random fluctuation. A "substantial" change is >20%. However, even a 10% change may be important if it is consistent across all subjects.
In practice, a few subjects influence the group's average because they changed a lot during the study. We do not know why those subjects changed so much. Perhaps they started to eat better. Perhaps they exercised more. There are too many uncontrolled variables to know.
Another problem is that subjects are told to eat a percent of their calories as PUFAs. Subjects who eat more food and exercise more do not gain weight, but eat a higher amount, in grams, of PUFAs. Very few studies control for this situation. According to Dr. Siguel’s research, subjects who ate more PUFAs would have lower TC/HDLC.
Another problem is that subjects start with huge variabilities in their body status for essential fats. Some are significantly deficient in EFs, while others have adequate levels. Subjects who were very deficient and started to eat better would improve much more than subjects that had adequate levels from the beginning would. Most studies do not measure the fatty acid status of subjects, and therefore we cannot correct for this difference.
Sometimes, there are commercial interests to prove a particular result. The maker of substance X would prefer to find that X is useful. Researchers do not have money to evaluate all relevant factors, and the maker of X may prefer that those certain factors not be known. The result is a study that intentionally omits key factors or whose design is biased to find desirable results.
For example, the maker of a fat substitute would not like to know that eating its product lowers the levels of essential fats in the body. Thus, the maker will not pay for fatty acid analyses (which are expensive) and these "problems" will never be found. Alternatively, the researchers may use a (cheaper) test that is not sensitive enough to detect deficiencies or abnormalities of EFs. (See Siguel EN. Diagnosing essential fatty acid deficiency. Circulation, Jun 30 1998; 97(25): 2580-2583.)
These factors combine in many ways to render the results of many studies suspicious. To evaluate test results, look first for at the underlying biochemistry. Is there a very good biochemical hypothesis that explains how substance X acts? Are the subjects eating enough PUFAs? Do all subjects have similar levels of PUFAs, vitamins, and minerals in their bodies?
Siguel, E. A new relationship between
Polyunsaturated Fatty Acids and Total/HDL Cholesterol. Lipids, 1996; 31,
S51-S56. A new diagnosis and treatment approach to abnormal cholesterol ratios.
The following describes studies which I believe
present misleading results.
Vitamin E and Alzheimer's
disease
In 1997, publications reported that taking Vitamin E in large doses (about 2,000 IU per day) helped prevent Alzheimer's disease.
Our review suggests that the alleged improvements were very small and inconsistent. They could be due to random variability among subjects, or to random variations in measurement (accurate measures of changes in mental status are difficult to obtain; those frequently used are highly variable and can be unreliable). The alleged improvements might not represent substantive improvements over a long period of time. They could also be due to changes in diet not measured by the researchers.
Using 2,000 IU/day of vitamin E does not make sense in light of other studies, which show that 200 IU/day is better for the immune system. Taking 2,000 IU/day of vit. E could interfere with oxidation processes needed to kill infectious agents, toxins and cancer. These matters need to be investigated.
I do not recommend taking large doses of vitamin E (over 400 I.U.) for Alzheimer's disease. Instead, I suggest that patients be tested to identify and correct for abnormalities of EFs. Optimizing one's levels of EFs (particularly the w3 fats), plus taking about 200 IU/day of vitamin E and other antioxidants, seems a better strategy.
Eating more monounsaturated fatty
acids (MUFAs) and olive oil
Researchers recommend that people eat more monounsaturated fatty acids (MUFAs) and olive oil. We disagree. Olive oils, particularly virgin olive oils, contain many compounds (linoleic acid, antioxidants, other substances) besides MUFAs. Some of these substances may help prevent CAD, but according to Dr. Siguel's research, the MUFAs themselves contribute to, rather than prevent, CAD. We need to distinguish between the benefits of olives (which contain antioxidants and other nutrients) and the benefits of MUFAs (the key fatty acids in olive oil).
Some people believe that MUFAs help to prevent or treat cardiovascular disease. Some base their beliefs on alleged studies of people in Mediterranean countries who presumably eat MUFAs and have low incidence of heart disease.
We reviewed many of the studies linking MUFAs to heart disease. We spoke with many of the researchers who wrote the original studies. We learned that many of the papers supporting the "Mediterranean diet" relied on studies conducted in the island of Crete, and were based on what those people ate since 1960. The diet in 1960 was very different from the diet today. It relied on homegrown vegetables, fruits, and fish, which contain EFAs and many other nutrients.
The low rate of CAD in Mediterranean countries is due to many factors, including a warm temperature (climate), good exercise habits, and high intake of vegetables and essential fats when compared with populations with high rates of CAD. We found that people from Mediterranean countries with low incidence of heart disease have higher than average levels of w3 and w6 essential fats in their bodies.
An often-cited NEJM study on MUFAs relied on food frequency questionnaires. Dietary questionnaires cannot accurately measure the intake of trans fatty acids, w3s, or w6s. In the opinion of Dr. Siguel, any study that relies primarily on dietary intake questionnaires to measure w3 and w6 essential fats is flawed (see above). Trans, w3, and w6 fatty acids cannot be measured accurately by questionnaires; instead, these fatty acids must be measured in blood and tissues using a fatty acid profile.
We attended many symposiums and conferences on the role of different types of fat (including MUFAs) on heart disease, cholesterol, HDL and triglycerides. We also conducted our own studies using subjects from the Framingham Heart Study, using blood tests to measure fatty acid status (not food frequency questionnaires!).
We found that scientific research has established that MUFAs are not essential fats. Humans can make MUFAs, either from SFAs, or from excess calories from carbohydrates and proteins, as needed. However, it is impossible for humans to transform MUFAs into essential fats (that's why essential fats are essential). The body will substitute MUFAs for PUFAs if it is desperate (leading to suboptimal function), but it would be far better just to provide your body with PUFAs.
Our research has shown that MUFAs are more similar to saturated fats than to essential fats. Eating more MUFAs will not prevent heart disease if one fails to eat enough essential fats.
Some rural populations have low intake of EFs and total calories, and low rates of CAD. The low rates of CAD in these populations are probably due to low body fat associated with low caloric intake and high exercise. Even though the body has small absolute levels of PUFAs, these people have a high ratio of PUFA/total fat in their bodies.
Olive oil is useful in cooking. When cooking, you want an oil low in essential fats, because you need something stable that can withstand heat. [Oils high in EFs will deteriorate and form isomers and undesirable trans fatty acids when used for cooking.] However, we suggest that you avoid cooking with oils, because of the extra calories. Try grilling, broiling, steaming, or microwaving instead. If you like, add a tiny amount of EF-rich oil to the cooked product for flavor (adding the oil afterwards avoids deterioration of the oil). You can also substitute oils rich in EFs anytime you use oil "raw" olive oil (for example, in a salad dressings) to supplement your diet with EFs.
Siguel EN, Lerman, RH. Fatty Acid Patterns in
Patients with Angiographically Documented Coronary Artery Disease. Metabolism
1994; 43:982-993. Shows link between EFA abnormalities and heart disease and
abnormal cholesterol ratios.
Are low fat foods much
healthier than foods with fat?
Our research indicates that many foods with essential fats (such as chicken, eggs, turkey, fish, tofu, and lean meat) are healthier than low fat foods without essential fats (such as low-fat cakes, pastries, cookies, sweets, yogurts, ice creams/sherbets/frozen yogurts, breads, pastas, and processed cereals).
In 1997, the NEJM published reports that agree with studies published by Dr. Siguel several years before showing that low fat diets can increase the risk of CAD by lowering HDL and increasing TGs. However, replacing foods high in total or saturated fat and cholesterol with low-fat foods high in processed carbohydrates can help when people lose weight and release EFs from adipose tissue.
Siguel E, Lerman RH. The role of Essential
Fatty Acids: Dangers in the USDA dietary recommendations ("pyramid")
and in low fat diets. Am. J. Clin. Nutrition, 1994; 60:973-9.
Siguel E, Lerman, RH. The role of Essential
Fatty Acids: Reply to the USDA. Am. J. Clin. Nutrition, 1995; 63:973-9.
Selenium (Se) may help delay some
cancers but promotes others
One key study found that selenium might help prevent prostate cancer in some patients, particularly those who come from areas where the food is low in selenium. This is a problem in a few places around the world. However, food in US supermarkets comes from many places; products grown in selenium-deficient soil are probably rare and are difficult to identify.
On the other hand, selenium may increase the risk of breast cancer and other disorders.
Misconceptions about Herpes
Some people believe that anti-Herpes drugs kill the virus, and that one cannot transmit the virus if one has no lesions.
The truth is, anti-Herpes drugs make it difficult for the virus to reproduce rapidly, but you need a healthy immune system to kill the virus. We know that the virus can be shed and transmitted even without any evidence of lesions (asymptomatic shedding). It is particularly dangerous to self-transmit (to spread the virus from one part of your body to another), because this expands the area of "action" of the virus. Even if you have Herpes, you should try to avoid contact with another persons with Herpes, so that you do not get a different form of the virus that would be resistant to antiviral drugs.
Breast cancer and w3/w6s
Ref: Bagga, D, Capone, S, Wang HJ, et al. Dietary Modulation of w3/w6 Polyunsaturated Fatty Acid Ratios in Patients with Breast Cancer. Journal of the National Cancer Institute, 89(15), August 6, 1997: 1123-1131.
The study shows that the ratio of w3/w6 can be changed by eating more w3 in proportion to w6. Meaning: Increasing the ratio of w3/w6 in the body may reduce the growth of cancer cells, because w3s may interfere with cancer cell reproduction. Dietary changes can influence w3/w6 ratio.
Siguel EN. Dietary modulation of
omega-3/omega-6 polyunsaturated fatty acid ratios in patients with breast
cancer. J Natl Cancer Inst. 1998 Apr 15; 90(8): 629-631.
|
All information on this website is copyrighted; see use and permission to reproduce. The information in this website is not medical advice, merely a general scientific discussion. See warnings & disclaimers. |
|
|
© 1998 Edward Siguel. All rights reserved |
modified 9/9/98 |