A study with an intent-to-treat prospective design was published in 1998 by a team from the Johns Hopkins Hospital[20] and followed-up by a report published in 2001.[21] As with most studies of the ketogenic diet, no control group (patients who did not receive the treatment) was used. The study enrolled 150 children. After three months, 83% of them were still on the diet, 26% had experienced a good reduction in seizures, 31% had had an excellent reduction, and 3% were seizure-free.[Note 7] At 12 months, 55% were still on the diet, 23% had a good response, 20% had an excellent response, and 7% were seizure-free. Those who had discontinued the diet by this stage did so because it was ineffective, too restrictive, or due to illness, and most of those who remained were benefiting from it. The percentage of those still on the diet at two, three, and four years was 39%, 20%, and 12%, respectively. During this period, the most common reason for discontinuing the diet was because the children had become seizure-free or significantly better. At four years, 16% of the original 150 children had a good reduction in seizure frequency, 14% had an excellent reduction, and 13% were seizure-free, though these figures include many who were no longer on the diet. Those remaining on the diet after this duration were typically not seizure-free, but had had an excellent response.[21][22]

"You can find a lot of "fat bomb" recipes on the Internet," Wittrock says. "These are very good at satisfying your sweet tooth, and are a great way to increase fat consumption without going over on protein. Also, I'm a huge fan of salted pumpkin seeds and salted sunflower seed kernels. Believe it or not, pork rinds are also a very good keto snack."
In 1921, Rollin Turner Woodyatt reviewed the research on diet and diabetes. He reported that three water-soluble compounds, β-hydroxybutyrate, acetoacetate, and acetone (known collectively as ketone bodies), were produced by the liver in otherwise healthy people when they were starved or if they consumed a very low-carbohydrate, high-fat diet.[10] Dr. Russell Morse Wilder, at the Mayo Clinic, built on this research and coined the term "ketogenic diet" to describe a diet that produced a high level of ketone bodies in the blood (ketonemia) through an excess of fat and lack of carbohydrate. Wilder hoped to obtain the benefits of fasting in a dietary therapy that could be maintained indefinitely. His trial on a few epilepsy patients in 1921 was the first use of the ketogenic diet as a treatment for epilepsy.[10]
Early studies reported high success rates; in one study in 1925, 60% of patients became seizure-free, and another 35% of patients had a 50% reduction in seizure frequency. These studies generally examined a cohort of patients recently treated by the physician (a retrospective study) and selected patients who had successfully maintained the dietary restrictions. However, these studies are difficult to compare to modern trials. One reason is that these older trials suffered from selection bias, as they excluded patients who were unable to start or maintain the diet and thereby selected from patients who would generate better results. In an attempt to control for this bias, modern study design prefers a prospective cohort (the patients in the study are chosen before therapy begins) in which the results are presented for all patients regardless of whether they started or completed the treatment (known as intent-to-treat analysis).[19]
The first modern study of fasting as a treatment for epilepsy was in France in 1911.[12] Twenty epilepsy patients of all ages were "detoxified" by consuming a low-calorie vegetarian diet, combined with periods of fasting and purging. Two benefited enormously, but most failed to maintain compliance with the imposed restrictions. The diet improved the patients' mental capabilities, in contrast to their medication, potassium bromide, which dulled the mind.[13]