Based on the clinical experience of others, and published clinical trials, we use carbohydrate-restriction in clinical practice for the treatment of diabetes mellitus [11-15]. At the end of our clinic day, we go home thinking, "The clinical improvements are so large and obvious, why don't other doctors understand?" Carbohydrate-restriction is easily grasped by patients: because carbohydrates in the diet raise the blood glucose, and as diabetes is defined by high blood glucose, it makes sense to lower the carbohydrate in the diet. By reducing the carbohydrate in the diet, we have been able to taper patients off as much as 150 units of insulin per day in 8 days, with marked improvement in glycemic control-even normalization of glycemic parameters. Due to the potent effect of carbohydrate restriction in decreasing blood glucose levels, we must reduce the insulin by 50% on the first day of dietary carbohydrate-restriction to avoid hypoglycemia. As the weeks pass, most patients achieve normoglycemia without medication, obese patients lose weight, and patients save money because they are not paying for medications. It is not so far-fetched to predict that these savings will also be passed along to the health care system and self-insured companies because there will be less expenditure on medications and the long-term diabetic complications.

Conclusion

The inattention to potent dietary therapy in all recent major diabetes studies, including the recent ACCORD trial, should not lead us to forget about carbohydrate-restriction as a means to achieve weight loss and glycemic control without hypoglycemia. We urgently need controlled studies comparing the newer "higher-carbohydrate diet with or without medication" approach to the earlier "carbohydrate-restricted diet without medication" approach for type 2 diabetes mellitus. One of the important advantages of carbohydrate-restriction is that there is no risk of hypoglycemia if medications are not used. We believe that carbohydrate-restriction has come of age for the treatment of obesity and diabetes mellitus and should be urgently translated from clinical practice to intensive testing in studies relating to mechanism, health services research, and public health.