The association among chronotype, timing of food intake and food preferences depends on body mass status

Abstract

Background/Objectives:

Previous studies have shown that individuals with circadian preferences for the evening (wake up later and reach maximum activity in the afternoon) have distorted dietary habits and misregulated body weight. Therefore, the present study was conducted to analyse the possible relationships between ‘morningness’ or ‘eveningness’ (chronotype), dietary habits and the level of obesity.

Subjects/Methods:

Among 400 participants, 171 subjects finished the follow-up period and were evaluated. Anthropometric, clinical and dietary parameters were analysed; the Horne–Östberg test was used to determine chronotype. A hypocaloric-behavioural intervention was performed in the overweight/obese subjects.

Results:

In normal-weight subjects, the morningness group ingested most of their energy and nutrients at breakfast and lunch, whereas the eveningness group showed a higher intake at dinner, corresponding with their chronotypes. A significant interaction was revealed between chronotype and body mass index regarding the energy and nutrients consumed at dinner (P<0.05 in all cases), as in the normal-weight subjects the evening food intake was higher in the eveningness group, but in the overweight subjects the situation was inverse. In addition, the food preferences were related to the chronotype, as the morningness subjects showed a higher intake of fruit (P<0.010).

Conclusions:

The timing of food intake corresponded to the chronotype in the normal-weight subjects; however, the overweight/obese subjects showed intake patterns removed from their physiological rhythms. These findings may indicate a need to design specific diets based not only on the total energy expenditure but also on the chronotype, as an indicator of the biological rhythms.

Introduction

It is widely known that most biological processes are rhythmically regulated in humans, including the sleep/wake and food intake cycles.¹ This regulation is performed through the interactions between endogenous and exogenous factors. In fact, although there is a biological system located in the suprachiasmatic nucleus for regulating human rhythms, the diet itself can modify homoeostasis and change the biological rhythms.^{2, 3}

Currently, there is a worldwide obesity epidemic, along with its associated diseases, such as diabetes mellitus, cardiovascular disease and metabolic syndrome (MetS). Several factors contribute to the increase in the prevalence of obesity, but obviously dietary habits and food composition directly influence the obesity development.⁴ Unfortunately, these factors cannot, in and of themselves, explain the increased prevalence of obesity over the last few decades, and other theories have been hypothesized. In this regard, several studies have shown that biological rhythm impairment, called chronodisruption, may imply the emergence of a number of metabolic diseases. This is the situation observed in shift workers and night-eating syndrome patients.⁵ Indeed, relationships between the disruption of the gene expression rhythms of the adipose tissue and the occurrence of obesity, MetS and other pathologies have been demonstrated.⁶

Recent studies have linked energy regulation to the circadian clock at the behavioural, physiological and molecular levels,⁶ emphasizing the fact that the rhythm or timing of food intake may have a significant role in weight regulation.⁷ In this way, the ‘eveningness’ chronotype seems to present a worse dietary pattern.⁸ To our knowledge, the relationship between nutrient or food intake and chronotype has not been studied in depth; on the other hand, the influence of chronotype on the effectiveness of weight regulation has not been analysed either. Therefore, in the present study, we sought to analyse the possible relationships between the biological chronotype and dietary habits, focussing on the timing of nutrient intake and food preferences, in order to evaluate whether these variables interact with the development of obesity. As a secondary objective, whether or not the chronotype affects a low-calorie intervention was also explored.

Subjects and methods

Design and subjects

A double design was used for the present study. On the one hand, it was a cross-sectional observational study, oriented to evaluate the relationships among dietary habits, chronotypes and body mass index (BMI). On the other hand, it was also an interventional longitudinal study using a hypocaloric-behavioural treatment for 3 months in overweight/obese subjects.

The target population was composed of the staff of the Catholic University of Murcia (UCAM), with an age range between 30 and 60 years. The selection procedure consisted of e-mail and advertisements following a probabilistic sampling procedure. Smokers, individuals with clinically significant illnesses (including type 2 diabetes, chronic heart failure, hepatitis and cancer) and those taking any medications known to affect body weight (thyroid hormones, corticosteroids and so on) were excluded from this study. In addition, those subjects previously diagnosed with eating disorders or with eating disorder symptomatology (eating attitudes test (EAT)-26 score>20) were also excluded. For the overweight/obese group, the subjects undergoing dietary treatment prior to participation were also excluded.

An initial sample of 400 subjects was needed to reach a significance level of 95% and a statistical power of 90%, and a flow diagram of the participant selection and follow-up during the study is presented in Supplementary Information 1. The subjects were divided according to their BMIs into a normal-weight group (BMI=18.5–24.9) and an overweight/obese group (BMI>25). Similarly, according to the Horne–Östberg test median score, the population studied was classified into the ‘morningness’ group (score>51) and the ‘eveningness’ group (score⩽51).

The study was carried out from September 2014 to July 2015, after receiving written authorization from the Catholic University of Murcia (UCAM) Ethics Committee. The patients (subjects with BMI>25) and normal-weight volunteers were informed, both orally and in writing, about the design of this study. An ethical explanation of the research project was also given, informing the participants about the aim of the study and guaranteeing the confidentiality and anonymity of the data, as well as respecting the Helsinki Declaration Agreement.

Data acquisition

Anthropometric data

The assessment of obesity was carried out according to the criteria proposed by the Spanish Society for the Study of Obesity (SEEDO).⁹ Body weight was measured to the nearest 0.1 kg while the subjects were dressed in their underwear, and height was measured to the nearest centimetre. BMI was calculated using these data, and the total body fat (%) was measured using an electric bioimpedance scale (TANITA BF-350; TANITA Corporation of America, Arlington Heights, IL, USA).¹⁰ The body fat distribution was assessed by measuring the waist circumference. Each measurement was taken three times by a single operator.

Chronotype

The chronotype was assessed using the morningness/eveningness questionnaire.¹¹ This is the most referenced test for evaluating the biological rhythmicity in humans, and has been intensively validated in human studies as an adequate tool. The test is composed of 19 items related to sleeping and waking hours, schedule preferences and subjective alertness at different hours of the day; for example, at the most pleasant period of the day to work, do hard physical tasks or exercise (sports). The chronotype was calculated on the basis of the questionnaire, and the subjects were classified into two categories: low scores were associated with eveningness (16–51 points), and high scores were associated with morningness (52–86 points).¹¹

Food and nutrient intake

Each study subject recorded their dietary intake by means of a food frequency questionnaire, which had been previously validated for the Spanish population.¹² The nutrient intake was calculated using a computer program designed on the basis of Spanish food composition tables.¹³ To further analyse their dietary habits, especially with regard to the nutrient distribution throughout the day, a 24-h dietary recall was also completed by the subjects. The results of this test were analysed using the same software application.

Eating attitudes

In order to evaluate their eating habits and exclude patients with symptoms of eating disorders, the EAT-26 was used. The EAT-26 is a standardized measure of the symptoms and characteristics of eating disorders,¹⁴ in which the participants indicated the extent to which each item was true for them (‘I find myself worried about food’) using a Likert scale ranging from 0=’never’ to 3=’always’. The overall scores were calculated using a standardized system of summing the responses (α = 0.83).¹⁴

Metabolic syndrome parameters and clinical determinations

The definition used to determine the presence or absence of MetS was that proposed by the International Diabetes Federation.¹⁵ The plasma concentrations of glucose, triacylglycerides, total cholesterol, high-density lipoprotein cholesterol and low-density lipoprotein cholesterol were determined with commercial kits (Roche Diagnostics GmbH, Mannheim, Germany) following the manufacturer’s guidelines. Arterial blood pressure was also measured.

Hypocaloric treatment

The overweight/obese patients underwent a 3-month hypocaloric treatment, and the low-calorie diets were designed following the SEEDO and Spanish Federation of Nutrition, Food and Dietetics (FESNAD) guidelines.¹⁶ Simultaneously, a cognitive-behavioural treatment, designed to increase adherence and prevent the negative thoughts often associated with weight loss diets, was also administered.

The patients were instructed to replace their usual diet with a balanced diet using nutrient distributions based on the Mediterranean diet. This diet was estimated to provide patients with an adequate nutrient composition: 50–60% carbohydrates, 30–35% lipids and 15–20% proteins. The patients were monitored weekly to record their weight loss and body composition changes. Subsequently, but within the same weekly visit, the cognitive-behavioural therapies, including relaxation, social skills, stimulus control and self-esteem therapies, were performed for 12 sessions.

The diets were designed by a dietitian (JSGM) with the assistance of DietoWin 7.0 software (Bl-Biologica, Barcelona, Spain). This software includes the nutritional composition of more than 600 foods, according to the Spanish food composition tables.

Statistical analysis

The data represent the mean±s.d. The nutrient intake data were estimated as the mean value of the 24-h dietary recall and the food frequency questionnaire, whereas the food preferences were analysed considering only the food frequency questionnaire-derived data. The main differences with regard to the chronotype or BMI level were evaluated using Student’s t-test, and the χ²-test was used to evaluate the association between the chronotype and MetS. Moreover, to analyse the possible interactions between the chronotype and degree of obesity, a linear general model was performed. However, a repeated measures analysis of variance test was conducted to analyse the influence of the chronotype on weight loss (baseline weight vs weight after 3 months). Finally, the Mann–Whitney U-test was conducted to evaluate the possible food intake differences depending on the chronotype. The data were analysed using SPSS 22.0 (SPSS Inc., Chicago, IL, USA) for Windows, and a P-value <0.05 was set as statistically significant.

Results

General and clinical characteristics according to chronotype

Table 1 shows the clinical and demographic characteristics of the population studied. All of the clinical parameters were similar between the two chronotypes, with only diastolic blood pressure and plasma high-density lipoprotein cholesterol levels being significantly higher in the morningness group, although in both parameters the mean values were within the normal range.¹⁵ On the other hand, the plasma triglycerides were significantly higher in the eveningness subjects, and in this case the mean value was slightly higher than the MetS cut-off point for this parameter. Moreover, the eating attitudes were practically the same between the chronotypes, as indicated by the EAT-26 test.

Table 1 General characteristics of the population studied attending to their chronotype

Nutrient intake according to chronotype and degree of obesity

The energy and macronutrient intake data for the chronotype and body weight groups are shown in Table 2. It is important to highlight the fact that none of the chronotype or BMI groups showed an energy intake pattern adjusted to the Spanish recommendations (Figure 1)¹⁷ for consuming an excess of energy at lunch and dinner. In addition, in both chronotypes, overweight/obese patients showed lower energy intakes at breakfast compared with normal-weight subjects.

Table 2 Total energy and macronutrient intake and the percentage consumed at different meals attending to chronotype and BMI level

Figure 1

Energy distribution at different meals in the (a) morning and (b) evening groups. Dotted line (····) represents normal-weight subjects and dashed (- - -) line reflects patients with overweight/obesity. The Spanish recommendations are also represented as continuous lines (^___). Significant differences between the normal-weight and overweight/obesity groups were analysed with Student’st-test. *P<0.05; **P<0.010; ***P<0.001.

When considering only the morningness group, the normal-weight subjects consumed a higher quantity of energy at breakfast and at their mid-morning, lunch and mid-afternoon meals, indicating an agreement between the chronotype preferences and energy consumed during the day. In contrast, the overweight/obese subjects consumed a statistically significantly higher quantity of energy at dinner (Figure 1), which may indicate a disruption between the chronotype and energy distribution throughout the day. However, the eveningness subjects showed a different pattern, depending on their degree of obesity. In this regard, the normal-weight subjects ate a higher quantity of calories at dinner, whereas the overweight/obese subjects consumed a higher quantity of energy at lunch. Again, the overweight subjects had energy distributions removed from their chronotypes.

The macronutrient intake patterns differed in similar ways with regard to the chronotype and obesity level (Figure 2). Essentially, the normal-weight subjects consumed a higher quantity of macronutrients at breakfast and during their mid-afternoon snacks, independently of their chronotype (Figure 2). Greater differences were observed in relation to the percentage of carbohydrates consumed at lunch, which was significantly higher in the overweight/obese patients in both morningness and eveningness groups.

Figure 2

The percentage of carbohydrates (a), lipids (b) and proteins (c) consumed at different meals in the morning (left-panel) and evening (right-panel) groups. White () bars represent normal-weight subjects and black solid (▪) bars reflect patients with overweight/obesity. Significant differences between the normal-weight and overweight/obesity groups were analysed using Student’s t-test. *P<0.05; **P<0.010; ***P<0.001.

The intakes of the other nutrients, such as alcohol or the different fatty acid families, were also studied (Supplementary Information 3). In the overweight subjects, alcohol consumption was significantly higher in the morningness group (P=0.006). Moreover, the percentage of alcohol consumed at lunch was significantly higher in the morningness subjects (both normal-weight and overweight groups, P<0.001 in both cases), whereas the percentage of alcohol consumed at dinner was higher in the eveningness subjects, especially in the overweight subjects (P<0.001). On the other hand, the fatty acid families had intake distributions quite similar to the total fat content. Specifically, morningness subjects showed higher saturated fatty acid, monounsaturated fatty acid and polyunsaturated fatty acid intakes at breakfast than did the eveningness subjects (P<0.015 in all cases).

Through our general linear model, several interactions between the degree of obesity and chronotype were also observed (Table 2), especially those related to the percentages of energy, lipids and proteins consumed at dinner. As a whole, these interactions revealed that in the normal-weight group, the eveningness subjects tended to eat a higher quantity of energy, lipids and protein at dinner, as expected. In contrast, in the overweight/obese group, the situation was just the opposite as the highest intake at dinner was shown in the morningness subjects. This suggests that a misalignment between chronotype and dietary intake pattern could promote an excess of body weight. These interactions were also observed in the variables of alcohol consumption and saturated fatty acid, monounsaturated fatty acid and polyunsaturated fatty acid intakes.

Food preferences according to chronotype, and interactions with BMI

In addition to nutrients, a possible relationship between chronotype and food preferences was also analysed in this research. Figure 3 shows the relative contributions of the different food groups in the morningness and eveningness subjects, whereas the detailed results of all foods studied are shown in Supplementary Information 2. The main differences were observed with regard to fruit intake, in the sense that the morningness subjects showed a significantly higher intake of fruit, specifically of citrus and citrus juice, during breakfast. Likewise, eveningness subjects were characterized by a higher intake of drinks (beer, soda and light soda) and added fats (Figure 3). As a whole, our data indicated that, in spite of a similar daily energy intake, the food intake pattern in the eveningness subjects was characterized by eating less healthy foods.

Figure 3

Relative contribution of the different food groups to the usual diet of the subject studied. White () bars represent morning subjects and black solid (▪) bars show evening subjects. Significant differences between the normal-weight and overweight/obesity groups were analysed through Student’s t-test. *P<0.05; **P<0.010.

Effectiveness of the weight loss intervention according to chronotype

As the chronotype seems to be associated with dietary pattern, whether or not the chronotype affects a low-calorie diet was also explored in the present study. Table 3 shows that body weight and BMI losses were higher in the morningness group; however, no statistically significant differences were observed, which is likely due to the heterogeneity of the data obtained in both groups. Nevertheless, the effectiveness of the dietary intervention appeared to be greater in the morningness subjects.

Table 3 Weight loss intervention effectiveness attending to the chronotype in those subjects with overweight/obesity

Discussion

The present study was performed with the aim of analysing the possible relationship between dietary intake and chronotype—that is, on the preference towards morningness or eveningness and its interaction with obesity. At first, it would be logical to assume that the eveningness subjects consumed a greater amount of energy and nutrients later in the day, whereas the morningness subjects ate larger early meals. Following our observations, this situation was especially evident in the normal-weight subjects, whereas the relationship between chronotype and dietary pattern was not manifested in the overweight/obese subjects.

To date, most studies have suggested a trend towards worse dietary behaviours in eveningness subjects, at least in young adolescents.¹⁸ However, in order to accurately analyse dietary habits, it is necessary to consider the degree of obesity of the patients. In fact, in the present study, the normal-weight eveningness subjects presented the most adequate dietary pattern, even though it was still far from the Spanish recommendations.

The other data revealed by our observations, which were in line with that of previous studies, were related to the total energy intake. As previously described by Garaulet et al.¹⁹ there were no differences regarding the daily energy intake between early and late eaters or, as in the present study, between the morningness and eveningness subjects. It is worth noting that the morningness group presented higher energy and macronutrient intakes at breakfast compared with the eveningness group. This situation could be a key point because, as described by Leidy et al.²⁰ subjects eating a higher quantity of energy and, particularly, protein at breakfast have better appetitive, hormonal and neural control over their food intake regulation.

Alcohol consumption also seems to be related to chronotype. In fact, the overweight morningness subjects showed a higher alcohol intake than the eveningness overweight subjects, especially at lunch. This issue is quite important, as recently Rose et al.²¹ suggested that, although alcohol may not increase food consumption per se, it may disrupt appetite signals, resulting in alterations in the food intake regulation.

Other data of note included the scarce consumption outside of the three main meals in the morningness group, whereas the eveningness subjects displayed a more adequate intake at these meals. In fact, a higher energy intake of snacks has been associated with a lower daily energy intake, better glycaemic control and more effective appetite regulation, which reinforces the relevance of dividing the daily energy intake as much as possible.^{17, 22}

Neither the morningness nor the eveningness subjects followed adequate dietary patterns. However, the eveningness subjects showed impaired metabolic control when compared with the morningness group, especially with regard to plasma triglycerides, fasting glucose and high-density lipoprotein cholesterol values, a situation that has been described by Reutrakul et al.²³ This issue might be associated with an impairment in the circadian metabolic control, as the timing of food intake could be more relevant than the total energy intake in the regulation of the circadian processes of fat physiology, as reported by Garaulet et al.¹⁹ Therefore, although the chronotype groups had similar daily energy intakes, the increased energy consumption at breakfast and reduced energy consumption at dinner in the morningness group exerted a protective role against metabolic impairment, confirming the role of breakfast in preventing metabolic alterations.^{20, 24}

As a whole, both chronotype and obesity appeared to influence the dietary patterns. In fact, a significant interaction between these variables was observed, especially in the percentage of energy and macronutrients consumed at dinner, a situation that was conserved with regard to the different fatty acid families. Curiously, in the normal-weight group, the eveningness subjects consumed more energy and macronutrients compared with the morningness group in this meal, as expected. Conversely, when the data on overweight/obese patients were analysed, the situation was just the opposite.

It may be hypothesized that this increased intake at night in the overweight/obese morningness individuals could be due to an alteration in the systems regulating appetite control. Indeed, it is well known that the leptin synthesis rhythms are impaired in patients with obesity.²⁵ In a previous study, we were able to show an impairment of the GLP-1 daily rhythms in parallel with the progression of body weight.²⁶ Therefore, it is tempting to speculate that, although the subject is gaining weight, an imbalance between the physiological systems regulating the biological rhythms and appetite control occurs. This, in turn, is manifested by metabolic derangement.

The different dietary patterns between the chronotype groups were observed not only in the nutrient intake but also in the food preferences. Based on our data, the morningness subjects ate a higher quantity of fruit, especially citrus, at breakfast. In contrast, the eveningness subjects showed a greater preference for meat products, alcohol consumption and sugary drinks. In the revised bibliography, there were few data regarding food preferences according to chronotype. For example, Arora and Taheri²⁷ recently showed that eveningness subjects exhibited an inadequate daily fruit and vegetable intake when compared with morningness subjects, confirming our observations. Similarly, Kanerva et al.⁸ also observed a higher intake of soda in eveningness subjects. After considering our data and that of previous works, it seems clear that eveningness subjects have an impaired food intake pattern when compared with morningness subjects, despite their similar energy consumption.

On the other hand, the timing of one’s food intake has been associated with the effectiveness of a weight loss programme.¹⁹ To confirm this hypothesis, the association between body weight loss and chronotype was also analysed. Our data confirmed higher body weight and BMI losses in the morningness subjects, which corresponds with the results of previous studies.¹⁹ Nevertheless, no statistically significant differences were observed in this study.

With the lack of specific dietary recommendations according to chronotype, the low-calorie diets in the present study were designed following the SEEDO recommendations,¹⁶ regardless of chronotype. Consequently, the nutrient and food contents of the different diets were similar in all of the subjects studied, which could explain the lack of statistical differences regarding the weight loss effectiveness. Otherwise, as suggested by Ross et al.²⁸ the chronotype might be associated with weight loss maintenance rather than with weight loss itself. A long-term evaluation of the subjects studied may shed some light on this issue.

There are several limitations in the present work that deserve consideration. First, the number of subjects was limited when compared with other large population studies,⁸ although it was similar to previous cross-sectional works. In addition, it would have been interesting to examine whether the weight loss was maintained over the long term based on the chronotype. Moreover, the data regarding each subject’s intake and weight loss showed large dispersion within the subjects, and for this reason any extrapolation of our observations should be carried out with some degree of caution.

In summary, the chronotype and degree of obesity were not only related to the total daily intake but also to the timing of the energy and nutrient intakes, in the sense that normal-weight morningness subjects consumed more energy and macronutrients at breakfast and lunch, whereas normal-weight eveningness subjects consumed higher percentages at dinner, in line with their chronotypes. In contrast, the overweight/obese subjects showed a worse dietary pattern, misaligned with their physiological rhythms, which in our opinion reinforces the need to design specific diets based on the chronotype as an indicator of biological rhythms.

An important conclusion that emerges from these results is that the design of low-calorie diets should take into account not only the total energy requirements but also the timing of the different meals, as a hypocaloric diet may be inadequate in the event that a patient eats at inappropriate times. Further studies are necessary to elucidate whether a dietary pattern adjusted to chronotype is better for losing weight than the usual hypocaloric diets.