Lets talk science again. So many bloggers and internet trollers out there getting their information from other bloggers. Leaves lots of people confused. They have even created a certain lingo. They love to say that epidemiology is correlation and not causation. This is true. There is more chocolate eaten in Sweden, and they have the most nobel laureates, so it seems that chocolate consumption and becoming a nobel laureate is correlated. Obviously this correlation is not causation. Problem with the bloggers is that they dont understand univariate vs multivariate analysis. Epidemiologists have very advanced statistical methods to remove any kind of confounding biases that may create a false correlation. They use techniques called multivariate regression analysis which basically tries to remove any factors that could create a false correlation. So, for instance, if you were trying decide if meat caused diabetes and you were looking at a large population of meat eaters and non meat eaters, then you might decide that you want to control for weight. If you have a meat eater who is really overweight then it is hard to determine whether his diabetes is from his weight or from his meat eating, so he is removed from the analysis. In many ways, these strict statistical methods over adjust. In the above example, it turns out that meat eaters in general are always heavier than plant eaters, so by removing the overweight ones you are diminishing the relationship between meat and diabetes, even though there really could be a correlation. The other important thing to understand with epidemiology is that it helps if you could look at multiple studies from multiple cultures. If just one study shows a correlation then you can wonder if the correlation does not equal causation. BUT, if you have multiple, prospective, long term, epidemiologic studies, done in multiple countries, all using advanced statistical methods, and they all say there is a correlation, then you better believe that where there is smoke there is fire. The Adventists Health study has followed 40,000+ people and looked at diet habits. The interesting thing is that Seventh Day Adventists believe that the body is the temple of the soul and should be treated as such, so they dont smoke, get exercise, and are healthier than your average Californian. In this study, the vegans had the lowest rate of diabetes, then the vegetarians, and the highest rates were with the meat eaters ncbi.nlm.nih.gov/pubmed/19351712. In a subset study they followed 8500 NON diabetics for 17 years! They found that weekly meat intake increased the risk of getting diabetes by 74%!!!!! ncbi.nlm.nih.gov/pubmed/18349528. Now if we travel from Cali to Boston, we find The Nurses Health Study done by Harvard. The followed 70,000 women for 14 yrs who initially did not have diabetes. They found that meat, and especially processed meat, greatly increased the risk of developing diabetes.ncbi.nlm.nih.gov/pubmed/15534160. When they combined the Nurses Health Study with the Health Profession study they found an even stronger risk of developing diabetes with just a little bit of meat. They suggest substituting grains or nuts to decrease risk. ncbi.nlm.nih.gov/pubmed/23779232, ncbi.nlm.nih.gov/pubmed/23779232. The Womens Health Initiative prospectively followed 38,000 women for 8.8 years and di an amazing multivariate analysis. Not only did they control for weight but they also controlled for calorie intake. Yet still, meat was strongly associated with an increase risk of developing diabetes. ncbi.nlm.nih.gov/pubmed/15333470 Now we can travel over to China for The Shanghai Womens study which followed 7000 women for 5 years. Yet again, meat was associated with significantly increased risk of diabetes, and early mortality. ncbi.nlm.nih.gov/pubmed/17435450. Interestingly increased fruit intake correlated with decreased diabetes. The largest ever epidemiologic study with the largest data base ever is the EPIC/Panacea study. They also looked at diabetes and meat. Guess what they found? ncbi.nlm.nih.gov/pubmed/22983636. ncbi.nlm.nih.gov/pubmed/19662376 Most interesting in these studies was the fact that glucose and fructose had NO relation to diabetes!!!!! ncbi.nlm.nih.gov/pubmed/23880355 The epidemiologic data is damning I am shocked that diabetics are ever told to eat more meat. In light of this evidence that seems to be against evidence based medicine. But these are just correlation studies. So what if we put it to the test? What happens if you design a 22 week study where you randomize a group of diabetics to either the ADA diet for diabetes, a diet that specifically limits carbs and fruit and is heavy with meat!!!! The other group was randomized to a vegan diet with lots of fruits and carbs. They ate pasta and cantaloupe etc. The results? Well if the epidemiology is correct then you should guess that the randomized control trial shows a superior improvement in diabetes control, and you would be right! ncbi.nlm.nih.gov/pubmed/19339401 So why does meat contribute to diabetes. I addressed this study in a prior post with cited studies. Basically, meat does stimulate insulin and is also associated with postprandial lipidemia which results in influx of fat into muscle cells that then interfere with insulin receptor production and therefore cause diabetes. The acid from the meat seems to be a big causative factor as does the inflammation. I have a lot of studies on this subject but they are very complex and I will save for another post. I do want to add that our carb fears are really ridiculous when it comes to diabetes. IF you cut meat and dairy down significantly and become more insulin sensitive than grains are an invaluable part of treating and preventing diabetes. ncbi.nlm.nih.gov/pubmed/12936955, In fact, eating a high grain diet is associated with a substantial decreased risk of developing diabetes in this meta analysis from the cochrane group. ncbi.nlm.nih.gov/pubmed/22649266
Posted on: Thu, 02 Oct 2014 01:49:26 +0000
Trending Topics
Recently Viewed Topics
© 2015