Atkins diet: Atkins 1972. The five adult studies: Brehm et al. 2003; Foster et al. 2003; Samaha et al. 2003; Yancy et al. 2004; Gardner et al. 2007 [the Stanford study]. The adolescent study: Sondike et al. 2003.
In two of these studies—Foster et al. 2003, and Gardner et al. 2007—the investigators also included a comparison of weight maintenance at the end of a year. In the former, those randomized to the Atkins diet maintained a greater weight loss than those assigned to the low-calorie, low-fat diet (4.4 © 6.7 vs. 2.5 © 6.3 percent of initial body weight), but the difference was not “statistically significant.” In the latter, those randomized to the Atkins diet maintained, on average, a 10.4-pound weight loss, compared with 5.7 pounds for a “lifestyle” program that included both a low-fat, low-calorie diet and exercise. The difference between these numbers and the larger numbers that Kemp reported could be due to the fact that these modern studies were more rigorous in following up on patients and measuring weight. It could be because the modern studies provided no counseling after the first two months of the trial, whereas Kemp categorized patients as having “defaulted” if they did not continue to appear at his monthly counseling sessions. Perhaps for this reason, the carbohydrate restriction after the first couple of months in the modern trials was at best modest. In Gardner et al. 2007—see p. 973 (table 2)—the subjects randomized to the Atkins diet consumed, on average, 30 percent of their calories from carbohydrates at six months, and 35 percent at twelve. This problem was particularly significant in the one trial I omitted: Dansinger et al. 2005:46 (table 2). Here the subjects randomized to the Atkins diet consumed an average of 137 grams of carbohydrates per day (32 percent of calories) after only two months of the trial, and 190 grams (over 40 percent) at six months and a year—the equivalent of four to five large baked potatoes every day—effectively identical to subjects randomized to low-fat, low-calorie diets. Thus, the similarity in weight loss on the different diets in this particular trial may simply reflect the similarity in carbohydrate consumption.

“first published synthesis…”: Bravata et al. 2003.

“Calories are all alike…”: Quoted in Berland 1983:7. “the effect of specific…”: Rubner 1982:36.

Bistrian and Blackburn instructed: Palgi et al. 1985. “thousands of patients…”: Interview, Bruce Bistrian.

Paradox relating to hunger: Leith 1961 (“nagging discomfort”); Peña et al. 1979; Hanssen 1936; Krehl et al. 1967 (“more than amply satisfied”). “Isn’t the proof…”: Interview, Bruce Bistrian.

Sims’s overfeeding experiments: Goldman et al. 1976:167.

Bloom’s articles on starvation therapy: Azar and Bloom 1963 (“At a cellular level…”); Bloom and Azar 1963.

“little hunger”: Bloom 1958. “In total starvation…”: Keys, Brozek, et al. 1950:829. “The most astonishing aspect…”: Drenick et al. 1964. “The gratifying weight loss…”: Anon. 1964c.

“mighty stimulant…”: Pennington 1954. “Many individuals spontaneously…”: Dwyer 1985:185.

“The fact remains…”: Anon. 1973.

“If you put a restaurant-size…”: Mayer 1975a:30–31. Fallback for the NIH recommendation: Ernst and Levy 1984:733–34; interview, William Harlan.