The purposes of the present experiments were to examine (A) the time course of endocrine changes with exposure to a ketogenic diet and (B) the effects of returning to a high carbohydrate diet after a period of consuming a ketogenic diet. We demonstrate that immediately following the discontinuation of the ketogenic diet, rats developed a brief period of hypophagia that was accompanied by a transient increase in plasma ghrelin levels, and a subsequent hyperphagia that developed 6 weeks after the dietary change and persisted through the end of the study. The increased chow intake was accompanied by significantly greater increases in body weight after the dietary switch, as compared to rats consuming chow for the entirety of the experiment. Whereas KD rats had significantly increased fat pad weights and plasma leptin levels as compared to CH rats, resuming the chow diet prevented a further increase in adiposity and leptin over time. Rats that consumed chow for the entirety of the study increased fat pad weight and leptin to resemble those of KD:CH rats by the end of the study. In addition, plasma insulin levels in KD:CH rats were not different from CH rats one week after returning to the chow diet, although it was significantly increased after 8 weeks of consuming chow after the ketogenic diet.

Previous studies in which rats with a history of consuming a ketogenic diet are switched to a high-carbohydrate, low-fat chow diet have observed an initial, short-term decrease in caloric intake following the diet switch [15, 16, 21, 22]. This effect has been found to last a short duration, with caloric intake returning to pre-switch levels after one week or less [15, 16]. In the present experiment, KD:CH rats initially decreased caloric intake during the first week after changing diets. This difference was transient and dissipated by the second week after the dietary change. It is likely that the increased ghrelin measured in this group one week after the dietary switch is a reflection of the transiently decreased caloric intake. Following the early hypophagia, caloric intake increased over time in KD:CH-fed rats after 6 weeks of consuming chow, as compared to caloric intake by CH rats.

Levels of leptin in plasma circulate in proportion to levels of body adiposity [23]. In the present experiment, KD rats exhibited increased plasma leptin levels, as compared to CH. These data are in accordance with previous observations that chronic consumption of a ketogenic diet induces an increase in this anorexigenic signal [17, 24]. As was the case with adiposity levels following the switch to CH, plasma leptin levels of KD:CH rats remained similar to levels prior to the diet switch. These results are in agreement with prior studies in which the short-term effects of a low- to high-carbohydrate diet switch were examined. Del Prete and colleagues have previously observed that plasma leptin levels remained steady two days after a similar diet switch [16]. In contrast, CH controls exhibited a significant increase in plasma leptin levels over time, and ultimately increased to the level of KD:CH leptin by the end of the study.

Ingestion of a ketogenic diet, at least in the rat, produces a unique endocrine state. While there is generally a positive correlation between levels of leptin and insulin [25], consuming a ketogenic diet alters this relationship such that leptin is exceedingly high while insulin levels are depressed. In this experiment, as in previous studies [17, 26], KD rats had lower plasma insulin levels than did CH controls despite increased levels of body adiposity. Although we report that insulin was significantly increased in KD rats after one week on the diet, it is likely that this was due to low levels of insulin in the CH rats at this time point, rather than a diet-induced rise in insulin from consuming a ketogenic diet. Interestingly, a significant decrease in plasma insulin was not measured until rats had been consuming KD for 8 weeks. Levels of circulating insulin are known to fall during other states of ketosis, such as starvation [27]. In addition, injections of an insulin anti-serum have been shown to increase concentrations of ketone bodies in blood [28]. Here, although levels of blood β-hydroxybutyrate were significantly elevated within one week of consuming the ketogenic diet, insulin was not decreased until week 8. The relationship between ketosis and insulin requires further investigation. It may be the case that over time, the process of ketosis has an inhibitory effect on insulin production and result in a dissociation between adiposity and plasma insulin levels.

In these experiments, returning to a carbohydrate diet after 8 weeks of consuming a ketogenic diet initially resulted in normalization of KD:CH insulin levels, compared to CH controls. After 8 weeks of consuming chow, the KD:CH rats had significantly elevated plasma insulin levels. It is likely that the increase in insulin is associated with increased caloric intake by KD:CH rats at this time point, as this increase corresponds with the increase in caloric intake exhibited by these rats in the last 3 weeks of the study.

Collectively, the present data demonstrate that following the switch from a ketogenic diet to one that is high in carbohydrates, rats gained more weight than did rats that had never consumed the ketogenic diet. Levels of epididymal fat were maintained at pre-switch levels throughout the chow-feeding period, although rats that underwent the dietary switch increased caloric intake over time and became hyperinsulinemic as compared to CH-fed controls. Whether these effects are observed in humans that choose to return to a high-carbohydrate diet after periods of consuming diets low in carbohydrates is currently unknown, however the current data suggest that continued maintenance on a low carbohydrate diet may play a role in prevention of increased body weight and caloric intake.