Has carbohydrate-restriction been forgotten as a treatment for diabetes mellitus? A perspective on the ACCORD study design

Eric C Westman1  and Mary C Vernon2 

1Department of Medicine, Duke University Medical Center, Durham, NC, USA; 2Private Practice, Lawrence and Lenexa, KS, USA.

Originaly published in Nutrition & Metabolism 2008, 5:10. April 2008.


Prior to the discovery of medical treatment for diabetes, carbohydrate-restriction was the predominant treatment recommendation to treat diabetes mellitus. In this commentary we argue that carbohydrate-restriction should be reincorporated into contemporary treatment studies for diabetes mellitus.


In the early 20th century, before any medications were available for the treatment of diabetes mellitus, experts recommended dietary carbohydrate-restriction [1,2]. The dietary recommendation for diabetes in a prominent internal medicine textbook from 1923 was 75% fat , 17% protein , 6% alcohol and only 2% carbohydrate [3]. The recommended total daily energy intake was 1,795 Calories per day. After the discovery of insulin and oral hypoglycemic medications, experts gradually changed the dietary recommendations to include more carbohydrate intake because most experts reasoned that the medications could be used to keep the glucose in control.

The NIH NHLBI Action to Control Cardiovascular Risk in Diabetes (ACCORD) group recently announced termination of the intensive insulin therapy arm of their study after an interim analysis showed that mortality was significantly higher in this group than in the other two less intensive glucose control groups [4,5]. Because lead investigators from the ACCORD trial and other experts have stated how unexpected this finding was, and have suggested that the concept of normal glucose control among patients with type 2 diabetes may not be desirable, we feel compelled to provide an alternative view.


From our perspective of familiarity with dietary carbohydrate-restriction and diabetes, these results are not surprising–in fact, they are predicted. We believe that it is unlikely that the increased mortality was due to the tight glucose control but rather due to the particular method for trying to achieve it. When high carbohydrate diets are consumed and intensive medication therapy is used to "cover the carbohydrate," it is very difficult to achieve normal glycemic control without hypoglycemic reactions. In our clinical practices, we frequently see individuals who are instructed to eat high carbohydrate diets and use intensive injectable hypoglycemic therapy, and they are susceptible to hypoglycemic reactions. Severe hypoglycemic reactions are associated with an increased morbidity and mortality [6].

There are other ways to improve glycemic control without the risk of hypoglycemic reactions; one of these is carbohydrate-restriction. Carbohydrate-restriction makes pathophysiological sense because type 2 diabetes is, in essence, a case of carbohydrate intolerance. We have observed that the same patients who have hypoglycemic reactions with high carbohydrate diets and aggressive medication therapy no longer have hypoglycemic reactions with carbohydrate-restriction. Moreover, the continued concerns about carbohydrate-restricted diets have never materialized and recent scientific studies show general health benefits including reduced cardiometabolic risk factors [7-10].