Introduction

Obesity remains a predominant public health concern. In recent decades, it`s prevalence has been rapidly growing worldwide, and nowadays has reached global epidemic proportions. In 2016 more than 1.9 billion adults were overweight or obese (WHO, 2018). The extent of risks associated with obesity has been highlighted in many countries. Increased risk of diabetes, high blood pressure, liver disease, gall bladder stones, coronary artery disease, cerebrovascular disease, psychological dysfunction, OSA, osteoarthritis, certain types of cancer, and infertility (Abdelaal et al., 2017).

A vast number of studies have been already undertaken to understand weight management in humans. One of the concepts, increasing meal frequency (MF) was proposed about 40 years ago, and since many large observational studies observed a favorable effect of increased MF on body weight and adiposity (Faby et al., 1964; Metzner et al., 1977; Drummond et al., 1998; Ma et al., 2003). Furthermore, numerous experimental trials have demonstrated evidence of increased MF benefits on body weight management, metabolism, and health (Palmer et al., 2009). It has been hypothesized, that the phenomenon of more frequent eating brings greater appetite control, improved glucose homeostasis, and increased food thermic effect (Schoenfeld et al., 2015). Alongside this, it has been shown that frequent macronutrient intake, particularly increased proteins, can beneficially contribute to lean mass development (Areta et al., 2013), enhancement of resting metabolic rate (Leidy and Campbell, 2011), and consequent fat mass reduction (Anciero et al.,2013).

However, there were several arguing reports from observational studies showing no association between food frequency occasions and body weight or body fatness (Kant et al., 1995; Yannakoulia et al., 2012). Some studies contrastingly stated that higher MF is associated with excess body weight and obesity (Howarth et al., 2007). Furthermore, several pilot studies did not observe a substantial difference in weight loss regarding various meal frequency (Berteus Forslund et al., 2008; Cameron et al., 2010). In contrast, some trials have suggested that a longer gap between meals and having fasting periods are beneficial (Stote et al., 2007; Kahleova et al., 2014; Gabel et al., 2018). The researches explained that in terms of weight loss the longer gaps between consuming foods spontaneously reduce daily caloric intake without calorie counting, and regarding health benefits, these gaps stimulate autophagy – the essential process to tackle cellular stress and save normal cell function (Paoli et al., 2019).

In addition, in recent years, concerning the mentioned appetite control in obesity management, there has been increasing interest in understanding hunger and satiety mechanisms (Suzuki et al., 2012). Numerous studies have been carried out to estimate the impact of MF on these complex body processes. While many trials have demonstrated that consumption of small, frequent meals positively affect appetite control (Louis-Sylvestre et al., 2003; Bachman and Raynor, 2012), others have stated the controversial concept, supporting the absence of significant evidence of mentioned association (Perrigue et al., 2016; Ingves et al., 2017). What is more, some studies have shown the opposite results, where increased meal frequency elevated feelings of hunger and desire to eat of examinees (Ohkawara et al., 2013).

Therefore, firstly, the present randomized control trial aimed to check the controversial hypothesis that increasing meal frequency results in higher efficacy of body weight reduction in comparison with a lower meal frequency under identic conditions. The difference between 3 and 4-5 eating occasions was examined in the present trial since the studies in previous literature focused on 1-3 meals a day versus 6 and more meals a day (Schoenfeld et al., 2015), thus little is known whether there is a difference in meal frequency effects if adding 1-2 meals only. Secondly, to discover whether a greater MF (also, the difference only in 1-2 meals) affects appetite and satiety sensation, particularly fasting and postprandial hunger and satiety levels at the baseline and in 4 weeks.

 2. Materials and methods

2.1 Study design

This parallel-group, randomized controlled pilot study was carried out in online mode from first to 28th of June 2020 in women population from different countries in two languages (English and Russian). Participants were assigned to higher (4-5 times a day) or rarer (3 times a day) meal frequency according to their will and dietary habits (most of the participants did not change their habitual frequency of eating).

2.2 Participants

Eighty-four females from 18 to 60 years old without serious health problems or special health conditions were recruited via an advertisement in social networks, with BMI more than 25 for Study 1, and with BMI between 18 and 25 for Study 2. The participants were screened online for any health problems by the standard checklist. At the screening, the individuals` height and weight were self-reported and body mass index (BMI) was calculated. Participants with BMI<18 and were excluded, as well as pregnant and breast-feeding women and volunteers with chronic diseases, which could affect the outcomes. All the recruited participants signed the Consent Form approved by Roehampton University, London. Of 84 participants, taken to the experiment, 23 withdrew after the start, and 61 completed it, whose results are presented herein.

2.3 Experimental procedure

The whole experiment was undertaken online over 28 days in two parallel Studies. Study 1 was checking the association of Meal Frequency (MF) and weight loss hypothesis, whereas Study 2 was discovering the effects of different MF on hunger and satiety ratings.  For both Studies on Day0 (baseline, before the start) and Day29 (control day, after the completion) the participants filled up the detailed questionnaires.

The questionnaires contained: Personal details (name, age, place of living, physical activity level); Self-reported anthropometric measurements (height, weight, waist, and hip circumferences); Visual Analogue Scales (VAS, Flint et al., 2000) for Hunger and Satiety self-assessments at 3 points a day (10 minutes before breakfast, lunch and dinner – for hunger, and 10 minutes after breakfast, lunch, and dinner – for satiety). The 0-meaning for hunger self-assessment was “not hungry at all”, and meaning at value 10 was administered as “starving”. For estimating satiety sensation, value 0 meant “not full at all, would eat the same amount”, and “ate too much, feel nauseous” – for grade 10; VAS for reporting the diet compliance on Day29. VAS was performed by scale from 0 to 10 for choosing relevant diet adherence, where value 0 was “I have not followed the given diet at all” and value 10 was “I have been following the diet very strictly”.

To avoid the limitation due to the impact of different levels of physical activity on weight loss (Swift et al., 2014), all the participants were recommended to apply 150 minutes a week of any exercise. Furthermore, for achieving better experiment’s diet adherence, before the start and every week of the experiment, the researcher discussed difficulties and questions regarding the trial on Webinars and was in touch with the participants via email and social channels every day of the study.

 Study 1

Thirty-five overweight or obese (BMI >25 kg/m2) females were observed in Study 1. All the subjects for 4 weeks were following the identic moderately restricted (1500-1700 kcal/day) higher in protein hypocaloric diet (Supplementary Data) split by 4-5 or 3 meals. The women were divided into two groups accordingly to the willing type of meal frequency: Frequently Eating Overweight/obese individuals (FEO) and Seldom Eating Overweight/obese individuals (SEO), n=18, and n=17 respectively. Anthropometric measurements, such as height, weight, waist, and hip circumferences, were self-reported at baseline (Day0) and in 4 weeks (Day29) in the given questionnaires. Alongside this, the volunteers were examined regarding hunger and satiety rates via VAS.

Study 2

Twenty-six healthy-weight (BMI 18-25 kg/m2) females took part in the second study of the trial. All the participants were given the identic eucaloric balanced diet for weight maintenance, and were organized into two groups relevantly to the type of meal frequency: Frequently Eating Healthy individuals (FEH, n=10) and Seldom Eating Healthy individuals (SEH, n=16). The imbalance in the number of participants in each group occurred due to the volunteers withdrawing. The diets of each group can be seen in Supplementary Data. For analyzing hunger and satiety levels around three main meals the rating system (VAS) was performed in the questionnaires, which the participants filled up at baseline and control day after 4 weeks.

2.4 Diet

All types of diets were developed by using DietPlan software (Version 7). In both studies, the first meal of a day (breakfast) was introduced within 1 hour after waking up and the last meal (dinner) was applied no closer than 3 hours before sleep. This is due to the meal timing concept, which has been vastly studied to play a substantial role in obesity and metabolic health (Xiao et al., 2019; Kahleova et al., 2017). Furthermore, we applied a greater amount of protein in the diet (25-30% of daily energy intake) for Study 1, as earlier scientists observed significantly higher total and abdominal fat loss in those who consumed an increased in protein diet (Anciero et al., 2013). Also, the weight loss diet had a moderate (300-500), but not a large calorie deficit, to minimize lean mass losses (Bopp et al., 2008). 

For FEO and SEO groups the energy-restricted diet was designed, identic in kcal and macronutrients, but split into 4-5 or 3 meals respectively. The diet contained between 1500-1700 kcal, with 30% of the protein in total energy intake, 30% of fats, and 40% of carbohydrates. For FEH and SEH groups the identic eucaloric diet, split into 4-5 or 3 meals accordingly, was developed. The energy maintaining diet contained 1800-2000 kcal with a balanced formula of macronutrient distribution: 20% proteins, 30% fats, 50% carbohydrates were applied (see Supplementary Data).

2.5 Statistical analysis

All data from the Initial (on Day0) and Final (on Day29) questionnaires were placed into the Excel file, where several formulas were applied to count needed results, such as BMI=weight/height2, (on Day0 and Day29); Wight loss in kg = Weight on Day0 – Weight onDay29; Weight loss in % of body mass = Weight loss in kg x 100 / Weight on Day0; Excess body weight = Height 2 / 10000 x (-25-(-BMI)), (on Day0 and Day29); Weight loss in % of Excess body weight = Weight loss in kg x 100 / Excess body weight; Waist/Hip ratio = Waist circumference / Hip circumference (on Day0 and Day29); Mean hunger (throughout a day) = (Hunger rate before breakfast + Hunger rate before lunch + Hunger rate before dinner) / 3; Mean satiety (throughout a day) = (Satiety after breakfast + Satiety after lunch + Satiety after dinner) / 3; Where Weight loss in kg in overweight/obese participants was more than their Excess weight in kg, the outcome “Weight loss in % of Excess weight” was counted as 100%.

IBM SPSS Statistics software for Windows, Version 26 was used for analyzing the study`s outcomes.

For the first study in overweight and obese participants, to understand whether the reduction of body weight was significant or not after the 28-day diet intervention, the paired t-test was applied for values of Excess body weight on Day0 and Day29. Furthermore, the independent student t-test was used for comparison of the weight changes (in the percentage of both body mass and excess body weight) in the groups with different meal frequencies (FEO and SEO). In the second study, to compare hunger and satiety levels at baseline and control day around the main three meals in two healthy-weight groups with different MF (FEH and SEH), a paired t-test was applied to the values of Mean Hunger on Day0 and Day29. The same was administered in overweight/obese individuals (FEO and SEO groups) for additional outcomes. In addition, by ANOVA-test was identified the difference in mean hunger and satiety rates on Day29 between all four groups.

 3. Results

 3.1 Participant characteristics      

Sixty-one women out of 84 recruited have completed the experiment. Most of the volunteers were from the Republic of Belarus (68,9%) and the UK (18%), the remaining distributed between Russia (9,8%), Hungary (1,6%), and Singapore (1,6%). Thirty-five subjects with BMI more than 25 were participating in Study 1, and 26 healthy-weight women were observed in Study 2.

Study 1. Effects of MF on weight loss.

Average participant characteristics and anthropometrics at baseline and control day overall and in the groups are shown in Table 1a. The mean age of the overweight/obese participants was 37.7±8.3 years old, and the range of BMI laid between 25.1 and 42.58.

Table 1a.  Anthropometrics and characteristics of overweight/obese participants at baseline and control day, also in the groups of 2 types of meal frequency (FEO- Frequently Eating Overweight/obese; SEO- Seldom Eating Overweight/obese)

Study 2. Effect of MF on hunger and satiety. 

The 26 healthy-weight participants were aged between 27 and 44, and their mean BMI was 22.6 (SD=1.5). The anthropometric characteristics and Hunger/Satiety rates at Day0 (baseline) and Day29 (control day) are performed in Table 1b 

Table 1b.  Anthropometrics and characteristics of participants with healthy weight at baseline and control day, also in the groups of 2 types of meal frequency (FEH- Frequently Eating Healthy-weight; SEO- Seldom Eating Healthy-weight)

3.2 Weight loss in overweight/obese group after 28-day diet intervention (Study 1)

Following the 4-week intervention, the participants` body Excess Weight significantly decreased versus baseline for both FEO and SEO groups (FEO: t=3.56, DF=17, p=0.002, SEO: t=5.99, DF=16, p=0.001) (Figure 1). Similarly, the significant reduction happened in total body weight, waist, and hip circumferences (p=0.001-0.005) in both meal frequency groups of overweight/obese participants (Table 1a). However, due to the high dropout, the frequent eaters initially had more excess weight than seldom eaters did. This can appear to be a limitation of group equality.