Blood glucose prior to meal request in humans isolated from all temporal cues

The factors responsible for meal initiation in man are not known. Recently, a small, transient decline in blood glucose was found to precede meals in freely feeding rats (1). To determine whether a similar decline in blood glucose precedes meals in humans, we measured plasma glucose concentration before and after requests for lunch on 4 different days in each of 7 subjects living in a setting that eliminated temporal cues and most or all external prompts for eating. Under these conditions, subjects could request any meal at any time by naming it and specifying its content. Meals were promptly prepared and served. Blood was sampled at randomized intervals averaging 20 minutes. Blood samples from 120 minutes preceding the subject's request for lunch to 90 minutes after the request were analyzed for glucose by the hexokinase method. The mean plasma glucose concentration did not change significantly from about 1 hour before lunch request to 20-30 minutes after the request. A transient decline in glucose similar to that reported in the rat was observed prior to only 1 of the 28 lunches. Given the ease of obtaining food and the absence of the usual temporal and psychosocial cues for eating, the results are not consistent with the hypothesis that meal requests in humans are triggered by a transient decline of plasma glucose concentration. The hypothesis is not disproven, however, since the intersample intervals of about 20 minutes may have failed to detect decreases of blood glucose that were smaller or of shorter duration than those of the rat.