By Eric Kossoff, MD

This month’s segment of Ketogenic Diet News and Research highlights two recent articles which significantly add to the growing knowledge about the relatively new, modified Atkins diet. Although the use of this diet started at Johns Hopkins, it has increased in use around the world. Benefits of this diet include an absence of protein, fluid, or calorie restriction, as well as an outpatient, non-fasting initiation. The major drawback in 2007 is a lack of published evidence…but this appears to be changing rapidly.

In the January issue of Epilepsia, Drs. Kang and Kim and their group from Seoul, South Korea reported their experience using the modified Atkins diet in 14 children at Sanggye Paik Hospital. After six months on the diet, half remained on the diet and 36% were at least 50% improved. Although this improvement was only about half of what the previous Epilepsia paper from February 2006 reported in children, overall this diet both created a ketotic state and was very well-tolerated. Interestingly, this study found that levels of both blood and urine ketosis did fluctuate more in those with less seizure control.

In an upcoming issue of Epilepsy & Behavior, our group reported the second pediatric study of the modified Atkins diet from Johns Hopkins. In this study, we randomized 20 children to start the diet at either 10 grams per day of carbohydrates or 20 grams per day, with a crossover to the opposite carbohydrate limit after three months. We had noticed that in our first study one year earlier, many children were not only able to tolerate increased carbohydrates after several months, but despite occasional resultant drops in ketosis, seizures were not worsened by doing this. We predicted the starting carbohydrate limit would be of no importance, other than easier to follow at 20 grams per day.

Surprisingly, we were only partially right: the carbohydrate limit does seem to matter the first three months…but not after! Sixty percent of those who started at 10 grams per day were >50% improved at three months, compared to only 10% who started at 20 grams per day. The reasons for this aren’t clear – ketosis was about the same between groups. We were correct regarding the time from 3-6 months on the diet: 20 grams per day was certainly easier, and after three months a change either up or down in carbohydrates did not change seizure frequency.

Children in this study had a statistical rise in both total and LDL cholesterol, unlike the first study. By using detailed, 3-day food records, we also confirmed our suspicions that the modified Atkins diet is close to a 1:1 ratio (fat: carbohydrate and protein), and (at least in children) is not calorie restricted.

What do these two studies teach us about the modified Atkins diet? For one, with now 58 total children published in the literature, 34 (59%) have had a >50% reduction of seizures at 6 months. This is remarkably similar to the ketogenic diet. Secondly, although this diet is less strict than the ketogenic diet, there may be some value to starting at a more restrictive 10 grams of carbohydrates per day…but then planning to increase to 20 grams per day over time. Lastly, these studies continue to question ketogenic diet traditions that high ketosis, restricted calories, and fasting are necessary for seizure improvement. Stay tuned for more studies over the next few years!

Submitted: 4/02/07
Reviewed by Steven C. Schachter, MD