Circadian alignment of food intake and glycaemic control by time-restricted eating: A systematic review and meta-analysis

Modern society is characterized in general by the ad libitum availability of food, continuous exposure to artificial light sources (especially devices with a screen), and continuous disruption of sleep and daily activities. All these factors disturb the day-night biological rhythm and are associated with the development of cardiometabolic diseases, especially type 2 diabetes [1, 2]. Intermittent fasting (IF) is a dietary regimen that consists of alternating periods of fasting and eating, either within a week – e.g. alternate-day fasting – or within a day – e.g. time-restricted eating (TRE) –. In recent years IF has become popular and many studies have been conducted to assess its effectiveness in improving the metabolic alterations induced by obesity and metabolic disorders.

TRE involves ad libitum eating in controlled time windows, emphasizing meal timing over calorie intake. A typical TRE protocol consists of 16:8 time windows [3], with the daytime feeding cycle comprising the shorter period (8 h), and a prolonged fasting period spanning the evening and continuing overnight (16 h). However, other meal timing windows have been tested in interventional studies [4, 5], and common fasting practices for religious purposes, such as the practice of Ramadan, involve an overnight eating window [6]. Some beneficial results of TRE have been described at the metabolic level in people with obesity or overweight, including an increase in lipid oxidation [7], a decrease in plasma glucose levels [8], and an improvement of insulin sensitivity [9]. Although the results obtained to date have been promising, some of the studies in question applied very short meal intervals (4 h, 6 h) [4, 5], which hinders their application to daily life. There are limited studies in subjects with type 2 diabetes [10, 11], and so the effects of TRE on metabolic health in that specific population remain unclear. In this sense, Andriessen et al. showed that three weeks of TRE improved mean 24 h glucose levels and time in range in adults with type 2 diabetes, but it did not improve insulin sensitivity [12].

TRE has beneficial effects at the metabolic level, both when there is weight loss [13] and when there is not [9], suggesting that the metabolic benefit may be due to other mechanisms underlying TRE and not solely to body weight changes. In this sense, some studies have pointed out that, compared to control lean subjects, people who have metabolic problems also develop alterations at the circadian cycle level in terms of the rhythmicity of metabolic processes, such as mitochondrial oxidative capacity [14], glucose homeostasis [15] and oxidation of substrates [14, 16]. It has, therefore, been proposed that disruption of the circadian rhythms can contribute to an impaired balance of substrate utilization and availability, which in turn is associated with metabolic diseases [17]. It has also been suggested that alterations in metabolic rhythmicity are the result of an altered cycle of feeding and fasting. Therefore, restricting food intake during the day to extend the fasting period may improve glycemic control.

Over the past few years, systematic review and meta-analyses have been conducted in order to shed light on the potential of TRE regimens to improve glucose metabolism related endpoints in different populations [6, 18, 19]. Although important contributions have been made in the field, it is necessary an update, since in the last 4 years a growing number of interventional studies in humans have been conducted. As a reference, only in 2022, 10 randomized clinical trials finalized and published results [11, 20–28]. Therefore, this systematic review was undertaken to evaluate whether TRE has beneficial effects on glycemic parameters. Further understanding of this subject would undoubtedly aid the development of nutrition strategies that prevent and even halt the development of metabolic diseases such as type 2 diabetes.

This systematic review was performed according to the guidelines of The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)[29]. The study has been registered in the International prospective register of systematic reviews (PROSPERO) with the registration number CRD42023405946.

Our systematic review and meta-analysis reveals that a TRE regimen of a window of between 6 to 10 h during 5 to 14 weeks is sufficient to improve glycemic parameters in subjects with overweight, obesity or type 2 diabetes. In particular, fasting insulin and HbA1c were significantly reduced in the TRE groups versus controls, regardless of calorie restriction or ad libitum regimens. Our analysis also suggests that the timing of food intake is a key factor in TRE dietary regimens, and that eTRE has greater metabolic benefits than lTRE, at least in terms of improving fasting glucose levels in subjects with overweight and obesity. In this sense, a significant reduction in fasting glucose levels was reported only by studies that followed eTRE, in which the regimen consisted of having breakfast before 9am or dinner before 4 pm, results that are in accordance with previous research, in which fasting glucose, insulin and HOMA-IR are reduced specifically by eTRE and not lTRE in healthy subjects [26, 42]. Similarly, in subjects with prediabetes, mean fasting glucose measured by continuous glucose monitoring was decreased by eTRE, and not with lTRE, when compared to baseline [43]. Despite not consisting of a prolonged fasting protocol (13:11 window), the study by Allison et al. demonstrates that eating early in the day versus eating late improves fasting glucose and insulin levels in lean and overweight subjects [44].

Some of the authors reported concerns about adherence to eTRE regimens [45], but recent data by Steger et al. [46] endorses eTRE as a simple dietary intervention that allows subjects to maintain their normal eating habits by concentrating them in a smaller window of time, rather than changing the number, size, or content of meals and snacks.

The beneficial effects of TRE for individuals with prediabetes and diabetes is supported by studies evaluating continuous glucose monitoring, in which an increased time in the normal glucose range upon TRE compared with non-TRE group is witnessed, together with decreased mean 24 h glucose levels [12, 47]. Thus, lowering of HbA1c by TRE is expected, however, the results of our systematic review show that reduction in HbA1c induced by TRE is small, and clinically insignificant. It is important to note that many studies included non-diabetic subjects and that the average HbA1c was 5.5%. In addition, none of the studies lasted more than 14 weeks. These factors may have limited the magnitude of the decrease in HbA1c reported. In fact, the only included study that was performed in patients with type 2 diabetes did show a significant HbA1c decrease of -1.54% in 12 weeks [11].

Another TRE approach that is followed by a large world population and has been extensively studied is Ramadan fasting, in which subjects refrain from eating and drinking from sunrise to sunset for 30 days. Usually they eat two meals a day, one before the sunrise and the other shortly after sunset. Systematic reviews and meta-analyses have shown that Ramadan fasting promoted minimal improvements in fasting glucose in healthy subjects [6], although greater effects were found in subjects with type 2 diabetes [48, 49]. There is a significant but small body weight loss after Ramadan practice [50], though this lost weight is rapidly regained [51]. In addition, other lifestyle changes are modified during the Ramadan period, as for example smoking is forbidden during the daylight hours and sleep duration is reduced due to the overnight eating schedule, since subjects usually fast and work during the daytime [52]. All these factors make difficult to draw conclusions from studies involving Ramadan as a TRE regimen and thus, we have considered excluding them from our systematic review and meta-analysis.

Although some of the studies included in our meta-analysis reported high adherence to the TRE schedule, such as Cai et al., in which > 97% of subjects adhered to the TRE protocol, some of our impressions might have been biased by the inclusion of non-compliant subjects in some of the studies [37]. Brady et al. only included subjects in their final analysis when compliance with the TRE eating window had been greater than 80% [40]. In contrast, the study by Jamshed et al. [22] included a large proportion of non-adherent subjects in both the TRE and control arms, as evidenced by a secondary analysis published by the same group [53]. In this secondary analysis, as expected, the authors reported that the subjects who adhered to the TRE regimen displayed greater improvements in insulin resistance and glucose levels than the general group included in the primary study.

Most of the studies selected for our systematic review allowed ad libitum energy intake. However, some combined TRE with calorie restriction, compared to a control group that followed the same low calorie diet in a 12:12 fasting:feeding time window [22, 23, 25, 41]. Calorie restriction is understood as reducing daily caloric intake by 15–40% without reaching malnutrition, and has classically been associated with improvements in metabolism and increases in longevity [54, 55]. Interestingly, our meta-analysis reveals that, with the exception of one study [23], the effects of TRE exceeded those of calorie restriction alone in terms of reducing fasting glucose [29, 32] and fasting insulin levels [32], especially when a eTRE was adhered to [25].

Among the studies performed in patients with type 2 diabetes, that by Che et al. met our inclusion criteria [11] and was included in the meta-analysis. We found other interesting studies in individuals with type 2 diabetes, such as the one by Parr et al. which we excluded from our systematic review due to the lack of a control group [10]. However, when we compared the post vs. pre changes in fasting glucose and HbA1c in the TRE arm, a significant reduction was evident in both studies, though greater in that by Che et al. (glucose: 26.4 vs 5.4 mg/dl; HbA1c: -1.54% vs 0.2%). The different duration of the TRE interventions (12 weeks in Che et al. vs 4 weeks in Parr et al.) may account for this difference. However, it is important to notice the earlier food intake window followed by subjects in the Che vs Parr study (8am to 6 pm vs 10am to 7 pm), which may have influenced the more pronounced response in glucose and HbA1c.

Ideally, nutrition should be synchronized with clock-regulated metabolic functions. The human body's metabolism is optimized for energy intake early in the day, as glucose tolerance and insulin sensitivity are higher upon waking than in the evening [56], partially because of the suppressive effect of melatonin on insulin secretion [57]. Molecular clocks in metabolic tissues such as adipose tissue or skeletal muscle are regulated by food intake. A clear example is the liver, where the peripheral clock system synchronizes gluconeogenesis and glucose release with the habitual fasting period [58]. Moreover, it should not be forgotten that our human ancestors were not sedentary and did not eat snacks or several meals during the day. In this respect, humans have evolutionary conserved cellular responses to adapt to prolonged fasting periods and physical activity [59]. For all these reasons, the TRE paradigm should be physiologically advantageous by acting on multiple organ systems.

Recent evidence suggests that most of the metabolic benefits of TRE are independent of weight loss [60]. Sutton et al. clearly demonstrated that TRE exerts beneficial effects on insulin sensitivity and improves beta-cell function without altering body weight [9]. In addition, a study by Andriessen et al. [12] reported modest effects of TRE on weight loss, but significant improvements in 24 h mean and fasting glucose, as well as glucose time in range after only 3 weeks of intervention. Given the crossover design of both studies, and despite the small sample sizes, their results are of great value. Mechanisms underlying glucose metabolism improvements by TRE besides body weight loss are not likely related to peripheral and hepatic insulin sensitivity or lipid content, muscle mitochondrial function, or changes in energy metabolism [12]. Instead, TRE has been described to affect the rhythmicity of serum and muscle amino acid and lipid metabolites and to regulate the rhythmicity of genes controlling amino acid transport [61]. Nevertheless, further research is warranted to decipher the molecular mechanisms triggered by TRE and that contribute to the metabolic health improvements.

Our findings highlight that, in the context of TRE, eating early in the day confers greater benefits on some glycemic parameters than delaying eating times. This is in line with a recent review by Lotti et al. [62] in which subjects with evening choronotypes were associated with a worse cardiometabolic risk profile and higher risk of diabetes. Considering that TRE does not necessarily imply a reduction of calorie intake, one can assume that confining all caloric intake to within a defined time window of a few hours has multiple health benefits, particularly if that window is early in the day. Further research is needed in subjects with prediabetes and type 2 diabetes to demonstrate whether greater benefits can be achieved.

Below is the link to the electronic supplementary material.

We thank Brian Normanly (University of Valencia/CIBERehd) for his editorial assistance.

SRL and VMV contributed to the concept and design; SRL, CLM, LPG, CB and CM contributed to the acquisition and analysis of data; SRL and VMV drafted the manuscript; all authors: made important contributions to editing and critically revising the manuscript. SRL and VMV guarantee this work and have primary responsibility for final content. All authors have read and approved the manuscript.

Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This research was supported by the European Regional Development Fund (ERDF ‘‘A way to build Europe’’); FISABIO (Foundation for the Promotion of Health and Biomedical Research in the Valencian Region, grants UGP15-220 and UGP-21–236); Generalitat Valenciana, Conselleria de Innovación, Universidades, Ciencia y Sociedad Digital (grants PROMETEO CIPROM/2022/32 and CIGE/2022/122), Spanish Ministry of Science and Innovation (grants CIBERehd CB06/04/0071, PI22/00424 and PI21/01160 from Carlos III Health Institute) and “Ayuda SED a Proyectos de Investigación en Diabetes dirigidos por Jóvenes Investigadores 2023", of the “Sociedad Española de Diabetes”. S.R-L is recipient of a Maria Zambrano fellowship [ZA21-049] from the requalification of the Spanish university system (Ministry of Universities, Government of Spain, European Union, NextGeneration EU). CB is recipient of a Miguel Servet contract [CP19/00077] from Carlos III Health Institute.

All data analysed during this study are included in this published article and its supplementary information files.

Susana Rovira-Llopis, Email: susana.rovira@uv.es.

Victor M. Victor, Email: victor.victor@uv.es