Circadian, circannual and other rhythms in spontaneous nutrient and caloric intake of healthy four-year olds.

Intake of total calories, proteins, lipids and carbohydrates was measured individually in four separate groups each of 26 to 29 children studied during different seasons of the year and an additional group of eight children investigated during all seasons. The subjects came from broken homes, had been abandoned or were orphans; 60% were males, 4 +/- 1.5 years of age (mean and standard error) and were of similar body weight and height. They were residents in one of two separate institutions in Nancy, France, living on similar, rigorously enforced schedules unchanged throughout the year. Socio-ecologic synchronization involved "lights-on" at 0700 "lights-out" at 1830, meals at 0800, 100, 1400 and 1800 with food of similar origin and kind, and similar timing of mental and physical activities. Each child was asked at fixed meal times to select spontaneously the kind and amount of food desired, uninfluenced as far as possible by choices of others, and to consume the food in any order. During seven days of adherence to this request, the amounts of protein, lipid, carbohydrate and total calories consumed at each meal were carefully determined for each individual. Time series thus obtained were analyzed according to the mean cosinor method for detecting and characterizing bioperiodic phenomena. A statistically significant circadian rhythm was detected in each of the 4 variables for each day of the best fitting sine function used to approximate the rhythm occurred around noon in almost all the studied circumstances. In other words the spontaneously larger meals were usually taken at 0800 (breakfast) and 1800 (supper). The trough of both lipid and protein spontaneous intake was clustered around 1800 only on Sunday. Cosinor analysis of individual seven-day time series (of each variable of each season) indicated a statistically significant circaseptan (approximately equal to 7-days) rhythm, with a peak occuring on Saturday, Sunday or Monday, but never on Wednesday. Changes in the weekly mean adjusted levels obtained by this method (as well as changes in mean values resulting from other statistical methods) demonstrate: 1) a circannual variation in spontaneous intake of lipids, carbohydrates and calories (protein changes are not statistically significant): 2) a peak of lipid intake, occuring in spring time, and a peak of carbohydrate and calorie intake, occuring in summer time. The probability of both exogenous and endogenous components of these rhythms is suggested.