Relative Effects of Time-Restricted Eating, Energy-Restricted Eating, and Unrestricted Eating on Eating Patterns and Dietary Intake: Results from a Randomized Controlled Trial

Abstract

Background

Time-restricted eating (TRE) may be as effective as an energy-restricted (ER) diet for weight loss. But little is known about the effects of TRE on eating patterns and dietary intake.

Objective

The aims of this study were to examine the relative effects of a TRE, ER, and unrestricted eating (UE) diet on eating patterns and dietary intake.

Design

This study is a secondary analysis of data from a randomized controlled trial carried out between October 2020 and October 2023. Over this period 88 participants were randomized to a TRE, ER or UE diet group.

Participants/setting

Adults with obesity in the Minneapolis St Paul, Minnesota metropolitan area who completed study baseline and follow-up measures of dietary intake (n=73).

Intervention

The interventions were: 1) TRE with an 8-hour self-chosen window with ad-libitum diet; 2) ER diet with 15% reduction of energy intake; or 3) unrestricted eating (UE) in which self-monitoring of food intake was encouraged with no specific change to eating recommended. The intervention period was 12 weeks.

Main outcome measures

Outcomes included meals eaten and intake of vegetables, fruit, dairy, protein foods, grains, energy, added sugars, saturated fat, sodium, dietary fiber and potassium.

Statistical analyses

Multivariate linear regression analyses were carried out to compare change in food and nutrient intake between experimental groups. Logistic mixed effects models were constructed to examine change in meals eaten.

Results

The TRE group ate fewer daily meals at end-intervention (−1.1 meals/day; 95% CI: −1.6, −0.7) compared to baseline, whereas the ER and UE groups did not experience a change in eating occasions. Those in the TRE group were less likely to report eating breakfast during end-intervention compared to baseline (OR 0.13; 95% CI 0.05,0.33) whereas no statistically significant change in behavior was identified for the ER (OR 1.02; 95% CI 0.41,2.55) or UE (OR 0.68; 95% CI 0.28,1.68) groups. Between baseline and end-intervention those in the TRE group had a decrease in intake of energy (−469 kcal/day; 95% CI: −681,−257), saturated fat (−8.5 g/day; 95% CI:−12.9,−4.1), potassium (−496 mg/day; 95% CI: −729,−263), and total (−1.7 ounce equivalents/day; 95% CI: −2.9,−0.6) and refined grains (−1.6 ounce equivalents; 95% CI: −2.6,−0.6). These changes were more marked compared to changes in the UE group. There were no statistically significant differences found between those in the TRE and ER groups.

Conclusions

Findings suggest that TRE with an 8-hour window and ad libitum intake may have similar effects on food and nutrient intake as an energy-restricted diet.

INTRODUCTION

Time-restricted eating (TRE) is emerging as an effective approach to achieving weight loss¹ that has the potential to be as effective as an energy-restricted diet while requiring less counseling time due to its simplicity. TRE is a form of intermittent fasting where eating occurs within a daily window ranging from 4–12 hours, typically 8–10 hours.² In contrast to traditional caloric restriction recommendations for weight loss, TRE does not require counting calories or adhering to a particular eating pattern, with ad libitum intake allowed during the eating window³.

TRE as a standalone weight loss strategy may compromise diet quality, as individuals might interpret the ad libitum eating window as permission to consume foods indiscriminately. In addition, TRE might lead to the elimination of breakfast- a meal that tends to be nutrient-rich⁴. On the other hand, TRE could potentially improve diet quality by eliminating evening/late-night snacking which commonly involve salty snacks, desserts, candy, and sweetened beverages)⁵. Also, it could be speculated that people pursuing weight loss may maintain healthy eating habits while practicing TRE, rather than viewing it as permission to abandon good nutrition.

A limited number of experimental studies evaluating TRE as a weight loss strategy have examined the effect of this eating pattern on food and nutrient intake^6–13, and these studies have a number of limitations. Most had a small sample size (less than 30 participants in total)^7,10–13, which limits statistical power for detecting differences in food and nutrient intake between groups. One study had a non-randomized trial design¹⁰. Only one of the studies included a comparison group in which participants were instructed to adhere to an energy-restricted diet alone (not paired with TRE)¹³. Consequently, this study aimed to contribute to the limited literature on the effect of TRE on food and nutrient intake by addressing the following exploratory research questions using data from a randomized trial¹⁴ in which adults with obesity were randomized to a TRE, energy-restricted (ER) or unrestricted eating (UE) diet: 1) What is the effect of a TRE diet on the number of eating occasions and meals eaten compared to the effects of a ER or UE diet? and 2) What is the effect of TRE on key indicators of dietary intake in comparison to the effects of a ER or UE diet? It was hypothesized that TRE would reduce eating occasions and snacks more than ER and UE diets, while maintaining similar food and nutrient intake patterns to ER diet.

METHODS

Overview

Data for this secondary analysis were obtained from a 12-week, prospective, parallel-group randomized controlled trial (RCT) carried out by the University of Minnesota (UMN) in the Minneapolis St Paul, metropolitan area (ClinicalTrials.gov: NCT04259632). The primary outcome for this trial was change in body weight, and sample size was determined based on this outcome. Participant enrollment in the study began in October 2020 with the final in-person visit in October of 2023. Participants were randomly assigned in a 1:1:1 ratio to one of the following intervention conditions: 1) TRE with an 8-hour self-chosen window with ad-libitum diet; 2) ER diet with 15% reduction of energy intake; or 3) unrestricted eating (UE) in which self-monitoring of food intake was encouraged with no specific change to eating recommended. The UE group served as the referent control group. Baseline and end-intervention measures included the collection of three interviewer administered 24-hour dietary recalls at each time point (baseline and end of intervention). The University of Minnesota Institutional Review Board approved this study (STUDY00008545), and all participants provided written informed consent. The study design is outlined in Figure 1.

Figure 1:

Flow diagram for enrollment of adults in the Minneapolis St. Paul, Minnesota Metropolitan Area into a randomized controlled weight loss trial carried out between October 2020 and October 2023

Participant Recruitment and Enrollment

Adults 18–65 years of age with a body mass index (BMI) of 30–55 kg/m² were recruited from the MHealth Fairview Health System in Minneapolis, MN. Study invitation letters were mailed out by MHealth Fairview Research Services to eligible patients. If interested, patients could reach out directly to the study team via phone call or email for further information. Additional inclusion criteria included self-reported waking times between 5am-9am; sleep duration of 6–9 hours per night; stable weight ± 5 pounds for > 3 months; eating window greater or equal to 12 hours/days; proficient in English; and demonstration of adequate logging of food intake using the smartphone app MyCircadianClock (mCC) app¹⁵ in the weeks before randomization. The mCC app enables real-time logging of food intake that captures the time of each eating occasion and the duration of eating within a day. Exclusion criteria included use of weight-affecting medications, shift work, significant medical conditions (e.g., cardiovascular disease, diabetes, uncontrolled pulmonary disease), magnetic resonance imaging (MRI) contraindications, current or expected pregnancy within six months, illiteracy, or history of an eating disorder. Initial screening for eligibility was carried out over the telephone with further screening carried out at a clinic visit. Participants were enrolled starting in October 2020 with the final in-person visit in October of 2023.

Randomization and Overview of the Intervention

A computer-generated code produced by the statistical team before study initiation was used to perform randomization in Research Electronic Data Capture (REDCap)^16,17. Permutated block randomization (block size: 3) to one of the three groups (TRE, ER, UE) was carried out with stratification by gender (male, female) and age (<45 years vs ≥ 45 years). Participants were informed of the experimental group to which they were randomized during an in-person study visit.

Registered dietitians (RD) provided telephone counseling to all participants. TRE and ER groups received weekly sessions for the first four weeks, then bi-weekly for eight weeks, totaling eight sessions over 12 weeks. UE participants had one initial session with instructions to maintain normal eating habits while using the mCC app¹⁵ for dietary tracking. Every RD interventionist provided counseling for all of the experimental groups.

Both TRE and ER groups received counseling from an RD interventionist that included goal setting and app-based diet tracking as strategies to support compliance. Participants in the UE group maintained their usual eating habits while tracking intake via the mCC app¹⁵.

Time Restricting Eating Intervention Protocol

The study interventionist RDs delivered the TRE support and education over a 12-week intervention period. An initial 30-minute counseling session was held in week 1; brief (up to 15 minutes) counseling sessions were then held in weeks 2–4; and brief (up to 15 minutes) counseling sessions were held every other week in weeks 6–12 (week 6, 8, 10, 12).

During the first counseling session, the Interventionist RD asked the participant about his/her weight loss goals and reasons for wanting to lose weight. The Interventionist RD explained that limiting eating and beverage intake to 8 hours may lead to weight loss, and that awareness of what one eats and drinks (self-monitoring) can also aid in weight loss and compliance with eating within 8 hours. The Interventionist RD assisted the participant in choosing an 8-hour eating window and explained the following eating and drinking-related TRE rules: Foods and beverages along with any dietary supplements they take should be consumed within their 8-hour eating window. Water and prescription or over-the-counter medications may be taken outside of the 8-hour eating window. Participants were encouraged to pick an 8-hour eating window that would facilitate their compliance with the window. Participants were encouraged to continuously self-monitor dietary compliance with their eating window by logging foods eaten using the mCC app¹⁵, with instructions provided for setting up their eating window in the app and viewing eating window reports available in the app. Also, the Interventionist RD recommended reading The Power of Habit book by Charles Duhigg.¹⁸ This book posits that habits are formed and hard to break due to a cue, routine, and reward loop, and strategies for altering this loop are provided in the book.

In subsequent counseling sessions, the Interventionist RD reviewed the participant’s mCC tracking record and revisited goals from the previous session within the context of the mCC data. Positive feedback was given for progress/accomplishments. The Interventionist RD worked with participants in developing strategies to overcome adherence challenges, such as adjusting their eating windows. Beyond the 8-hour eating window restriction, participants received no dietary or weight loss advice and were instructed to maintain their usual food choices.

Weekly feedback on tracking using the mCC app¹⁵ was provided by the Interventionist RD. Feedback, which was sent via text message, encouraged compliance with self-monitoring and eating within the target time window.

Energy Restriction Intervention Protocol

The study interventionist RD delivered the ER intervention over a 12-week intervention period. An initial 90-minute counseling session was held in week 1; 30-minute counseling sessions were scheduled for weeks 2–4 and then every other week in weeks 6–12 (weeks 6, 8, 10, 12).

Before the first counseling session the interventionist RD calculated a daily calorie consumption target 15% below calculated daily energy needs. This calorie target was calculated by the interventionist RD by first calculating total daily calorie requirement using the Calorie Calculator at Calorie.net.¹⁹ The participant’s gender, age, height, weight, and activity level were entered into the calculator using the ‘Miffler St Jeor’ option. To determine the participant’s activity level the Interventionist RD calculated the participant’s average daily Metabolic Equivalents of Tasks (METs) using the daily METs values in the ‘MET Rate’ portion of the participant’s Actigraph Clinical Report (participants wore an Actigraph accelerator²⁰ for one week during the baseline measurement period). The total daily calorie requirement estimate generated by the calculator was then multiplied by 0.85 to estimate the total daily calorie intake goal for a 15% energy-restricted diet. Based on the total daily calorie intake goal and usual eating pattern information identified from the mCC app¹⁵ food logs, the Interventionist RD drafted a healthy food plan using the food exchange system in the ‘Choose Your Foods: Food Lists for Weight Management’ booklet published by the Academy of Nutrition and Dietetics and the American Diabetes Association.²¹ This booklet along with the food plan was sent to the participant along with related education materials.

At the first session the interventionist RD asked participants about their weight loss goals and interest in losing weight. The Interventionist RD also discussed the principles of an energy-restricted diet and the use of the food exchange system and food plan²¹ developed for them for controlling energy intake. The interventionist RD encouraged dietary self-monitoring via the mCC app¹⁵ to support weight loss. If the participant expressed a strong desire to use another app that provides calorie intake data, this was allowed. The Intervention RD also recommended reading The Power of Habit.¹⁸

In subsequent counseling sessions, the Interventionist RD reviewed mCC tracking data and prior goals and helped participants develop strategies to overcome adherence challenges.

The Interventionist RD texted participants weekly to provide feedback and encouragement on mCC app¹⁵ tracking and ER food plan compliance.

Unrestricted Eating Control group

The interventionist RD held a single 30-minute initial session with participants in the UE group. The initial session focused on weight loss goals and introduced dietary tracking through the mCC app¹⁵. It emphasized that increased dietary self-monitoring improves awareness of dietary intake and may aid weight loss. Otherwise, no additional dietary advice was provided during the intervention. Post-intervention, participants were offered eight free nutrition counseling sessions based on their tracked eating patterns. Additionally, they would be provided with a copy of The Power of Habit book.¹⁸ After the initial session there was no further contact between the Interventionist RD and participants in the UE group outside of the weekly text to provide feedback and encourage compliance with the mCC app¹⁵ tracking.

Fidelity of the Intervention

As measures of intervention fidelity counseling session attendance and duration were recorded as was a summary of information provided and discussed at each session. As reported in the primary results paper, participants in the TRE and CR groups completed a similar number of counseling sessions with a study interventionist RD. Due to the simplicity of the intervention the TRE had less allocated dietitian time (mean of 20.9 minutes per visit; 95% CI: 17.8–23.9) compared to the ER group (mean of 36.2 minutes per visit; 95% CI 33.1–39.3)¹⁴.

Outcome Measures

Three interviewer administered 24-hour dietary recalls were collected at baseline and end-intervention to document dietary intake. Each set of three recalls was collected unannounced over the telephone using the Nutrition Data System for Research (NDSR) dietary analysis software program.²² The versions of NDSR used over the study period were 2020, 2021 and 2022. Staff trained and certified in the collection of dietary recalls using NDSR conducted the interviews and were blinded to intervention assignment. Recalls were collected all days of the week to capture intake on one weekend and two weekdays during each measurement period. The time and name of each eating occasion (e.g. breakfast, lunch, snack, etc.) were assessed as part of the 24-hour dietary recall interview.

As documented elsewhere¹⁴, a variety of additional outcome measures were collected including measured height and weight and self-reported age, gender (male, female), ethnicity (Hispanic or Latino, Not Hispanic or Latino), and race (American Indian/Alaskan Native, Asian, Hawaiian or other Pacific Islander, Black or African American, White).

Data Analysis

For this paper the analytic sample was limited to randomized study participants who had at least 2 dietary recalls at both baseline and end-intervention and had had mean energy intake between 500 kilocalories (kcals) and 7,000 kcals per day (Figure 1).

Means (standard deviation) and frequencies (percentages) were calculated to describe the baseline characteristics of participants overall and by experimental group. One-way analysis of variance analysis (ANOVA), Chi-square, or Fisher’s exact tests (as appropriate) were carried out to determine whether differences in characteristics across experimental groups were statistically significant.

Logistic mixed-effects models were constructed to examine changes in meal type (breakfast, lunch, dinner, and snack) eaten between baseline and end-intervention across experimental groups. The model included fixed effects for time point (pre or post-intervention), the experimental group, and the interaction between the two terms. The models included a random intercept to account for correlations among repeated measures within participants (up to six recalls). The interaction term for these models can be interpreted as a ratio of odds ratios, indicating whether the odds of a dietary recall containing the meal at end-intervention compared to baseline were different between experimental groups. P-values from type 3 tests for the interaction of time point (pre versus end-intervention) and treatment are provided.

Multivariate linear regression analyses were carried out to compare change in food and nutrient intake from baseline within and between experimental groups. Covariates included in the models were age, gender, and baseline values for the variable being modeled. Pairwise comparisons were conducted using Tukey’s method to account for multiple comparisons. Food groups examined included key components of the eating pattern recommended in the 2020–2025 Dietary Guidelines for Americans (DGAs)²³: total grains, whole grains, refined grains, fruit, vegetables, dairy, and protein foods. Nutrients and foods for which limited consumption is advised (added sugars, saturated fatty acids (SFA), sodium, alcoholic beverages, and sugar-sweetened beverages) were examined. Also, nutrients of public health concern due to inadequate intake among US adults (dietary fiber and potassium) were examined.²³

All analyses were carried out using SAS (Version 9.4, SAS Institute Inc., Cary, NC)²⁴. The significance threshold was 2-sided P < .05.

RESULTS

Participants (N=73, with 28 in the TRE group, 19 in the ER group, and 26 in the UE group) predominately identified as white (83.6%) and had a mean (standard deviation) age of 43.7 years (10.6). Approximately one-half (53.4%) identified as female and mean body mass index (BMI) (kg/m2) among all participants was 36.3 (5.1) at baseline (Table 1). Baseline demographics and dietary intake measures were similar between experimental groups with one exception. Whole grain intake at baseline was statistically significantly lower in the TRE compared to the UE group.

Table 1.

Baseline characteristics and dietary intake of participants in a weight loss trial overall and by experimental group (n=73 adults with obesity from the Minneapolis St Paul Minnesota metropolitan area)

	Overall (n=73)	Time-restricted eating (n=28)	Energy-restricted diet (n=19)	Unrestricted eating (n=26)	P valuea
Age (years), mean (SD)	43.7 (10.6)	43.7 (11.4)	42.4 (9.7)	44.5 (10.7)	0.81
Gender- n (%)					0.80
Female	39 (53.4)	16 (57.1)	9 (47.4)	14 (53.8)
Male	34 (46.6)	12 (42.9)	10 (52.6)	12 (46.2)
Ethnicity- n (%)					0.72
Hispanic or Latino	2 (2.7)	1 (3.6)	1 (5.3)	0 (0)
Not Hispanic or Latino	71 (97.3)	27 (96.4)	18 (94.7)	26 (100)
Race- n (%)					0.37
American Indian/Alaska Native	1 (1.4)	0 (0)	1 (5.3)	0 (0)
Asian	5 (6.8)	2 (7.1)	0 (0)	3 (11.5)
Black or African American	3 (4.1)	2 (7.1)	0 (0)	1 (3.8)
Mixed	2 (2.7)	2 (7.1)	0 (0)	0 (0)
Native Hawaiian or Other Pacific Islander	1 (1.4)	1 (3.6)	0 (0)	0 (0)
White	61 (83.6)	21 (75.0)	18 (94.7)	22 (84.6)
Weight (kg), mean (SD)	109.3 (19.3)	109.0 (20.7)	110.6 (19.9)	108.8 (18.0)	0.95
BMI (kg/m2), mean (SD)	36.3 (5.1)	36.2 (5.5)	36.5 (5.7)	36.3 (4.3)	0.98
Total number of eating occasions per day, mean (SD)	5.5 (1.2)	5.5 (1.4)	5.3 (1.1)	5.7 (1.0)	0.49
Energy (kcal/day), mean (SD)	2288 (663)	2149 (523)	2252 (690)	2464 (757)	0.21
Added Sugars (g/day), mean (SD)	58 (50)	47 (30)	49 (31)	75 (72)	0.09
Saturated Fatty Acids (SFA) (g/day), mean (SD)	36 (14)	33 (13)	35 (11)	39 (17)	0.32
Sodium (mg/day), mean (SD)	3802(1173)	3689 (992)	3909 (1455)	3847 (1162)	0.80
Dietary Fiber (g/day), mean (SD)	20.0 (5.6)	18.7 (4.9)	19.9 (5.8)	21.5 (6.1)	0.18
Potassium (mg/day), mean (SD)	2631 (710)	2416 (546)	2725 (856)	2793 (718)	0.12
Total Grains (ounce equivalents/day), mean (SD)	7.7 (3.2)	7.2 (2.9)	7.7 (3.9)	8.2 (3.2)	0.55
Whole Grains (ounce equivalents/day), mean (SD)	1.4 (1.3)	1.0b (0.9)	1.3 (1.0)	1.9b (1.6)	0.03
Refined Grains (ounce equivalents/day), mean (SD)	6.3 (2.8)	6.2 (2.8)	6.3 (3.3)	6.3 (2.4)	0.99
Fruit (cup equivalents/day), mean (SD)	0.6 (0.5)	0.4 (0.4)	0.7 (0.7)	0.6 (0.5)	0.29
Vegetables (cup equivalents/day), mean (SD)	1.7 (0.9)	1.6 (0.8)	1.9 (0.9)	1.7 (0.9)	0.41
Dairy (cup equivalents/day), mean (SD)	1.8 (1.4)	1.7 (1.4)	1.9 (1.1)	2.0 (1.7)	0.78
Protein foods (ounce equivalents/day), mean (SD)	7.0 (3.4)	7.6 (3.3)	6.2 (3.5)	6.8 (3.5)	0.43
Alcoholic beverages (servings/day), mean (SD)	0.7 (1.0)	0.7 (1.0)	0.7 (0.9)	0.6 (1.2)	0.90
Sugar sweetened beverages (servings/day), mean (SD)	0.7 (1.5)	0.4 (0.6)	0.3 (0.6)	1.2 (2.4)	0.10

^aP values from one-way ANOVA for continuous variables, Chi-square test for gender, and Fisher’s exact test for race.

^bindicates groups that were significantly different with p value < 0.05

Number of Eating Occasions and Meals Eaten

The change in number of eating occasions was found to differ between experimental groups (p=0.02) (Table 2). On average, those in the TRE had fewer eating occasions (−1.1 eating occasions/day; 95% CI: −1.6, −0.7) at end-intervention compared to baseline, and this decline was statistically significantly greater than the change observed in the ER group. No other between group differences in change were statistically significantly different.

Table 2.

Mean change from baseline in food and nutrient intake and number of eating occasions by experimental group and between group differences in change^a (n=73 adults with obesity from the Minneapolis St Paul Minnesota metropolitan area)

	TREb change from baseline (N=28) mean (95% CI)	ERc change from baseline (N=19) mean (95% CI)	UEd change from baseline (N=26) mean (95% CI)	Difference in change between TREb vs. ERc mean (95% CI)	Difference in change between TREb vs. UEd mean (95% CI)	Difference in change between ERc vs. UEd mean (95% CI)	P value e
Total number of eating occasions per day	−1.1 (−1.6,−0.7)	−0.2 (−0.8,0.3)	−0.4 (−0.9,0)	−0.9f (−1.8,−0.1)	−0.7 (−1.5,0.1)	0.2 (−0.7,1.1)	0.02
Energy (kcal/day)	−469 (−681,−257)	−314 (−569,−58)	28 (−195,251)	−156 (−553,242)	−497f (−870,−124)	−341 (−751,68)	0.01
Added Sugars (g/day)	−16.0 (−28.0,−4.0)	−10.9 (−25.4,3.6)	−5.9 (−18.5,6.8)	−5.1 (−27.5,17.3)	−10.1 (−31.3,11.0)	−5.0 (−28.4,18.4)	0.52
Saturated Fatty Acids (SFA) (g/day)	−8.5 (−12.9,−4.1)	−5.8 (−11.0,−0.5)	−0.5 (−5.1,4.1)	−2.7 (−11.0,5.5)	−8.0f (−15.7,−0.3)	−5.3 (−13.7,3.2)	0.05
Sodium (mg/day)	−680 (−1086,−274)	−539 (−1031,−48)	4 (−417,424)	−141 (−907,626)	−684 (−1385,17)	−543 (−1321,234)	0.06
Dietary Fiber (g/day)	−2.8 (−5.1,−0.4)	−1.7 (−4.4,1.1)	−1.1 (−3.5,1.4)	−1.1 (−5.5,3.3)	−1.7 (−5.8,2.4)	−0.6 (−5.0,3.8)	0.60
Potassium (mg/day)	−496 (−729,−263)	−213 (−492,66)	−19 (−260,221)	−283 (−722,157)	−476f (−883,−69)	−194 (−634,246)	0.02
Total Grains (ounce equivalents/day)	−1.7 (−2.9,−0.6)	−0.5 (−1.9,0.9)	0.5 (−0.7,1.6)	−1.2 (−3.4,0.9)	−2.2f (−4.2,−0.2)	−0.9 (−3.1,1.2)	0.04
Whole Grains (ounce equivalents/day)	0.1 (−0.5,0.6)	0.2 (−0.4,0.8)	0.1 (−0.4,0.6)	−0.2 (−1.1,0.8)	0 (−0.9,0.9)	0.2 (−0.8,1.1)	0.89
Refined Grains (ounce equivalents/day)	−1.6 (−2.6,−0.6)	−0.7 (−1.9,0.5)	0.2 (−0.8,1.3)	−0.9 (−2.8,1.0)	−1.8f (−3.6,−0.1)	−0.9 (−2.9,1.0)	0.05
Fruit (cup equivalents/day)	−0.1 (−0.4,0.1)	0 (−0.2,0.3)	0.1 (−0.2,0.3)	−0.2 (−0.6,0.2)	−0.2 (−0.6,0.2)	0 (−0.4,0.4)	0.33
Vegetables (cup equivalents/day)	−0.2 (−0.4,0.1)	−0.2 (−0.5,0.1)	−0.3 (−0.6,0)	0.1 (−0.4,0.5)	0.2 (−0.3,0.6)	0.1 (−0.4,0.6)	0.70
Dairy (cup equivalents/day)	−0.5 (−0.8,−0.1)	0.1 (−0.3,0.6)	−0.4 (−0.8,0)	−0.6 (−1.2,0.1)	0 (−0.6,0.5)	0.5 (−0.1,1.2)	0.07
Protein foods (ounce equivalents/day)	−0.8 (−1.8,0.2)	−1.0 (−2.2,0.2)	0.4 (−0.6,1.4)	0.2 (−1.7,2.1)	−1.2 (−2.9,0.5)	−1.4 (−3.3,0.5)	0.13
Alcoholic beverages (servings/day)	−0.1 (−0.5,0.3)	0.1 (−0.3,0.6)	0.1 (−0.2,0.5)	−0.2 (−0.9,0.5)	−0.2 (−0.9,0.4)	0 (−0.8,0.7)	0.64
Sugar sweetened beverages (servings/day)	0.1 (−0.4,0.6)	−0.3 (−0.9,0.4)	0 (−0.6,0.5)	0.3 (−0.6,1.3)	0.1 (−0.8,1.0)	−0.2 (−1.2,0.8)	0.74

^aAdjusted for age, gender, and the baseline value of the variable being modeled

^bTime-restricted eating

^cEnergy-restricted diet

^dUnrestricted eating

^ep-value from one way analysis of variance of differences between groups

^findicates between group difference was significantly different with p value < 0.05

In comparing meals eaten at end-intervention compared to baseline by experimental group (Table 3), breakfast was the only meal for which statistically significant between-group differences were observed (p=0.005). Those in the TRE group had lower odds of reporting eating breakfast during a dietary recall conducted during end-intervention compared to baseline (OR 0.13; 95% CI 0.05,0.33) whereas the odds of eating breakfast did not change for those in the ER ER (OR 1.02; 95% CI 0.41,2.55) and UE (OR 0.68; 95% CI 0.28,1.68) groups. These differences in odds were statistically significantly different between the TRE and the ER groups and the TRE and UE groups.

Table 3.

Odds of end-intervention dietary recalls including a meal compared to baseline dietary recalls among adults with obesity from the Minneapolis St Paul Minnesota metropolitan area by experimental group and in comparison to experimental group (n = 429 recalls - 215 at baseline and 214 at end-intervention)

Meal Type	TREa change from baseline to end-intervention Odds Ratio (95% CI)	ERb change from baseline to end-intervention Odds Ratio (95% CI)	UEc change from baseline to end-intervention Odds Ratio (95% CI)	Ratio of TREa to ERb Odds Ratio (95% CI)	Ratio of TREa to UEc Odds Ratio (95% CI)	Ratio of ERb to UEc Odds Ratio (95% CI)	P valued
Breakfast	0.13 (0.05,0.33)	1.02 (0.41,2.55)	0.68 (0.28,1.68)	0.13e (0.04,0.48)	0.20e (0.05,0.71)	1.50 (0.42,5.43)	0.005
Lunch	0.78 (0.33,1.84)	1.60 (0.62,4.11)	0.77 (0.29,2.02)	0.49 (0.14,1.75)	1.02 (0.28,3.71)	2.08 (0.54,8.06)	0.46
Dinner	0.53 (0.11,2.44)	0.80 (0.21,3.10)	0.84 (0.19,3.83)	0.66 (0.09,5.10)	0.62 (0.07,5.37)	0.95 (0.12,7.20)	0.89
Snack	0.40 (0.18,0.89)	0.68 (0.29,1.60)	0.84 (0.36,1.95)	0.60 (0.19,1.90)	0.48 (0.15,1.53)	0.81 (0.24,2.69)	0.43
Beverage only	0.90 (0.45,1.80)	0.81 (0.36,1.82)	1.34 (0.65,2.75)	1.11 (0.38,3.20)	0.68 (0.25,1.84)	0.61 (0.21,1.80)	0.62

^aTime-restricted eating

^bEnergy-restricted diet

^cUnrestricted eating

^dP values from logistic mixed-effects models examining changes in meal type between baseline and end-intervention across experimental groups.

^eindicates between group difference was significantly different with p value < 0.05

Dietary Intake (Table 2)

The changes in energy intake (kcals/day) between baseline and end-intervention were statistically significantly different between experimental groups (p=0.01). The TRE group and ER groups both reduced energy intake from baseline (−469 kcal/day; 95% CI: −681,−257 and −314 kcal/day; 95% CI: −569,−58, respectively) whereas the change observed for the UE group (28 kcal/day; 95% CI: −195,251) was not statistically significant.

Regarding nutrients and foods for which limited consumption is encouraged (added sugars, saturated fatty acids (SFA), sodium, sugar-sweetened beverages, alcoholic beverages, and refined grains), the only between group differences in change that were statistically significant were for SFA (p=0.05) and refined grains (p=0.05). The TRE group and ER groups both reduced SFA intake from baseline (−8.5 g/day; 95% CI:−12.9,−4.1 and −5.8; 95% CI: −11.0,−0.5 g/day) whereas the change observed for the UE group (−0.5 g/day; 95% CI:−5.1,4.1) was not statistically significantly different. In comparing change in SFA between groups, the difference in change between the TRE and UE groups was statistically significant (−8.0; 95% CI −15.7, −0.3). With respect to refined grains, the TRE group reduced intake from baseline (−1.6 ounce equivalents/day; 95% CI −2.6, −0.6) and this change was statistically significantly different from the change observed in the UE group (−1.8 ounce equivalents/day; 95% CI −3.6,−0.1).

For nutrients of public health concern due to inadequate intake (dietary fiber and potassium), potassium was the only nutrient with a difference in change between groups that was statistically significant (p=0.02). The TRE group reduced potassium intake from baseline (−496 mg/day; 95% CI: −729,−263 mg/day) whereas the change observed for the ER (−213 mg/day; 95% CI: −492,66) or UE (−19.3; 95% CI: −260,221) groups were not statistically significant. The difference in change between the TRE and UE groups was statistically significant (−476 mg/day; 95% CI −883, −69).

In examining changes in food groups for which there are recommended intake amounts in the 2020–2025 DGAs²³, no notable or statistically significant between group changes were observed for whole grains, fruit, vegetables, and protein foods. However, changes in intake of total grains (p=0.04) and refined grains (p=0.05) differed statistically significantly between groups. On average the TRE group reduced intake of total grains (−1.7 ounce equivalents/day; 95% CI: −2.9,−0.6) between baseline and end-intervention whereas the changes observed for ER (−0.5 ounce equivalents/day; 95% CI: −1.9,0.9) and UE (0.5 ounce equivalents/day; 95% CI: −0.7,1.6) groups were not statistically significantly different. Similarly, the TRE group reduced intake of refined grains between baseline and end-intervention (−1.6 ounce equivalents; 95% CI: −2.6,−0.6) whereas the changes observed for ER (−0.7, 95% CI: −1.9,0.5) and UE (0.2, 95% CI: −0.8,1.3) groups were not statistically significantly different.

DISCUSSION

Findings suggest a TRE diet allowing ad libitum intake with an 8-hour self-selected eating window may lead to fewer meals eaten per day in comparison to an ER diet. In addition, a TRE diet may reduce the odds of eating breakfast, with a greater reduction in odds than with a ER diet or UR diet. Also, a TRE diet may lead to a more pronounced decrease in intake of energy, saturated fat, potassium, and total and refined grain intake in comparison to an unrestricted diet. Changes in intake of these nutrients and food groups observed in the TRE and ER groups were not statistically significantly different, which suggests that TRE and ER diets may have similar effects on dietary intake. These findings concur somewhat with weight loss and energy intake change results included in the paper that reported primary results from this trial¹⁴. The mean weight loss difference between the TRE and UE groups was −1.4 kg (95% CI: −4.5 to 1.7; p=0.53) and the weight loss difference between the ER and UE groups was −2.5 kg (95% CI: −5.8 to 0.8; p=0.18). The weight loss difference between the TRE and ER groups was 1.1 (−2.0 to 4.2; p=0.69). Energy intake changes and differences were also reported in the main results paper and were also mostly consistent with findings reported in this paper. Differences likely reflect differences in the analytic sample (present analysis excluded those with fewer than two dietary recalls at baseline or follow up whereas the main results paper did not have this exclusion criteria) and differences in covariates included in the model (present analysis included age and gender as covariates whereas these variables were not included in the model reported in the main results paper).

These findings contribute to a limited literature^6–13, and it is difficult to compare findings from this study with those of other experimental trials because of heterogeneity in study designs. Two of the eight prior studies evaluated the effect of designated early TRE diet on food and nutrient intake^8,11. Early TRE is a form of TRE in which the time window is assigned and designed to encompass breakfast while excluding eating in the late afternoon. In four of the studies an eating window that began later in the day was assigned^6,10,12,13. In contrast, participants in the TRE group in the present study were allowed to self-select an 8- hour eating window. Other notable differences included the use of a shorter^7,11 or longer¹⁰ eating window in comparison to the 8-hour eating window in the present study.

In line with the heterogeneity in the design of studies, findings across studies have been variable with some finding statistically significant differences in change of food and nutrient intake between the TRE and control groups^6,11–13 and others finding no differences that were statistically significant^7–10. Consistent with our finding of a greater decrease in energy intake among those in the TRE compared to the usual energy intake control group, Cienfuegos et al.⁶ found a greater decline in energy intake among those assigned to a 4 or 6-hour TRE window in comparison to a control group that had no eating restrictions. Likewise, Gabel et al.¹² found a greater reduction in energy intake among those assigned to an 8-hour TRE window compared to a control group that was instructed to maintain normal energy intake.

In this study decreased intake of some foods and nutrients where limited consumption is recommended was found among those in the TRE group, and these changes were greater than differences observed in the usual energy intake control group. This is consistent with findings reported for a study in which those assigned to a 6-hour TRE diet had a greater improvement in diet quality in comparison to a control group that was instructed to eat within one hour of waking and for up to 16 hours subsequently¹¹.

Only one prior study compared a TRE diet with an energy-restricted diet alone¹³. Mengi et al. carried out a trial in which 23 healthy overweight females were randomized to an 8-hour TRE window (10 a.m.– 6:00 p.m.) or an energy-restricted diet for eight weeks. In this study, energy intake decreased in both groups, with a greater decline observed in those in the energy-restricted diet group. Greater declines in total fat, saturated fat, vitamin E, and sodium were likewise found for those in the energy-restricted diet group. In contrast, vitamin C and potassium intake declined to a greater extent among those in the TRE compared to the energy-restricted diet group. In the present study none of the changes in food and nutrient intake between those in the TRE and energy-restricted diet groups were statistically significantly different. The differences in findings between the present study and the Mengi study could be due to the numerous methodological differences between studies such as differences in demographics of the study population (e.g. males and females in the present study and females only in the Mengi study) and differences in the duration of the intervention (12 week in the present study and 8 weeks in the Meni study).

The present study has numerous strengths, which include the use of a rigorous measure of dietary intake (collection of multiple interviewer-administered 24-hour dietary recalls at each time point); inclusion of both energy-restricted and unrestricted diet groups for comparison to TRE; and use of an effectiveness trial design²⁵ so that effects under normal clinical conditions are more closely approximated in comparison to an efficacy trial. Intervention compliance data reported in the main results paper indicated that adherence to the 8-hour eating window among those in the TRE group averaged 63.8% of days during the intervention and 73.4% of days at the end of the intervention. The mean eating window among those in the TRE group during the final two weeks of the intervention period was 9.1 hours per day which was an average reduction of 5.1 hours per day (95% CO: 4.5–5.7) relative to baseline¹⁴.

Limitations of the present study include use of a self-report dietary assessment tool that may be susceptible to intervention-related bias in reporting dietary intake at end-intervention.²⁶ Also, although the present study had a larger sample size than most prior studies (n=73 in present study compared to sample sizes ranging from 18–90 in prior studies^6–13), statistical power for detecting within-group changes and between-group differences in change was likely low as the study was powered for detecting change in body weight (primary study outcome) and variance in food and nutrient intake is notoriously high.²⁷ Another limitation is the way in which gender was assessed in the study. A response option was not provided for participants who do not identify as male or female, and as a result these individuals would have been misclassified. An additional limitation is that the study was conducted in one geographical area and the study population was limited in racial identity and ethnic diversity.

Conclusion

Findings from the present study suggest that self-selected TRE with an 8-hour window and ad libitum intake may have similar effects on food and nutrient intake as an energy-restricted diet. Future trials evaluating TRE should include larger sample sizes along with rigorous measures of food and nutrient intake so that effects on dietary intake may be better understood. Also, trials should be carried out in more geographically and racially and ethnically diverse samples.

Research Snapshot.

Research Questions

What is the effect of time-restricted eating (TRE) on dietary intake in comparison to energy-restricted (ER) and unrestricted eating (UE) diets?

Key findings

After the 12 week intervention period, the TRE diet was found to have led to a more pronounced decrease in intake of energy, saturated fat, potassium, and total and refined grain intake in comparison to the UE diet. The changes between baseline and follow up in the TRE and ER groups were found to be not statistically significant.