This study was a parallel-group, randomized controlled trial conducted over a period of 12 weeks. Participants were randomly assigned to the time-restricted eating intervention group and the control group. All outcomes were assessed at baseline and at the end of the 12-week intervention. This study was approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Approval No. 2022–052) and registered with the Chinese Clinical Trial Registry (Registration No. ChiCTR2200060631). This study was conducted in accordance with the principles of the Declaration of Helsinki and adhered to the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline for RCTs. All participants provided written informed consent.

Participants were recruited from the sleep laboratory of the otolaryngology out-patient department at a tertiary hospital in Chongqing, China. The inclusion criteria were: (1) participants with mild or higher severity of OSAHS (AHI ≥ 5 events/h); (2) BMI of 28–45 kg/m²; (3) age 18–60 years; (4) ability to use a smartphone or mobile device; (5) willingness to participate in this study and provide written informed consent. Exclusion criteria were as follows: (1) current or planned treatment with either surgery or CPAP for OSAHS; (2) participants with severe cardiopulmonary disease, hepatic or renal dysfunction, malignant tumors, other serious organ diseases, or those currently taking weight-loss medications; (3) individuals with psychiatric disorders or cognitive impairment; (4) those who were pregnant or planning pregnancy. The withdrawal criteria included: (1) voluntary withdrawal from the study; (2) lost to follow-up for various reasons.

Participant recruitment and randomization took place from April to December 2022, with allocation to the TRE group or the control group at a 1:1 ratio. A statistician not involved in the study used a computer to generate a random number sequence, which was then printed on allocation cards and placed in sealed, opaque envelopes. Another researcher subsequently opened the envelopes according to participant enrollment order, and participants were assigned to either the TRE group or the control group based on the instructions inside the respective envelope.

The TRE intervention was systematically developed using the Behavior Change Wheel (BCW) framework, a comprehensive model for designing behavior change interventions (44). The process involved three core stages: (1) Understanding the Behavior: The target behavior was defined as “adhering to an 8-hour eating window daily.” Behavioral analysis using the COM-B model (Capability, Opportunity, Motivation-Behavior) (44) and the Theoretical Domains Framework (TDF) (45) identified key determinants to address. (2) Identifying Intervention Options: Based on pre-trial structured interviews with obese OSAHS patients, key behavioral determinants were identified, guiding the selection of intervention functions: education, incentivization, environmental restructuring, and training. (3) Identifying Content and Implementation Options: Specific Behavior Change Techniques (BCTs) were selected corresponding to the chosen intervention functions (46). These BCTs included, but were not limited to, information about health benefits, instruction on how to perform the behavior, self-monitoring of behavior, feedback on behavior, social support (practical), and review of behavior goals.

Participants in the TRE group were instructed to adhere to an 8-h eating window (10:00 a.m. to 6:00 p.m.). During the eating window, participants could consume food in any quantity and of any composition. Outside the window, calorie intake was prohibited, with the stipulation that only water and energy-free beverages like unsweetened coffee and tea were permitted. Participants were instructed to maintain the same eating window throughout the 12-week intervention period. Adherence was monitored using standardized paper forms where participants recorded their first and last eating times each day. Researchers reviewed these records every 4 weeks. Adherence was defined as confining intake to the 8-h window for at least 6 days per week. To reduce participant burden and enhance compliance, detailed food logging was not required. Additionally, participants received guidance on physical activity recommendations outlined in obesity prevention and management guidelines during the 12-week study period.

Besides standard care of OSAHS management, the specific intervention measures for the TRE group were developed using the BCW guidebook. All intervention components were delivered through face-to-face consultation, telephone call, and online platform. The primary components included face-to-face in-person educational sessions delivered by trained intervention researchers to participants. These sessions covered topics such as the health risks associated with OSAHS and obesity, establishing accurate disease awareness, correcting patient misconceptions, and teaching participants how to calculate their BMI and understand the normal range for self-monitoring purposes. Additionally, participants were informed about the fundamental principles and health benefits of time-restricted eating and advised to set periodic weight loss goals. Researchers established a peer support group based on online platform to foster communication and mutual encouragement between participants. Every 4 weeks, researchers conducted online support meetings to document adverse events, review participants’ weight loss progress milestones, and provide solutions to emerging issues. Furthermore, participants collaborated with researchers to develop weight management plans and goals for the next phase.

During the 12-week study period, participants in the control group were offered standard care protocols from the sleep laboratory. This included basic OSAHS health education and general nutritional counseling aligned with the Chinese Dietary Guidelines. The nutritional counseling advised a balanced diet without specific caloric limits or personalized meal plans, emphasizing increased intake of vegetables, fruits, whole grains, and lean protein, while reducing consumption of cooking oil, added sugars, and ultra-processed foods. Education also highlighted regular physical activity, proper sleeping positions, smoking cessation, and alcohol moderation. No specific caloric targets or daily food intake monitoring was provided. After the study, participants in the control group were offered an introduction to the intervention program and obtain electronic materials.

All outcomes were measured at both baseline and Week 12. Primary outcomes were BMI and sleep quality. Secondary outcomes included changes in excessive daytime sleepiness, quality of life, and the change in STOP-Bang questionnaire score. Furthermore, sociodemographic information and disease indicators were collected at baseline, including age, gender, educational level, occupational status, marital status, family monthly household income, and AHI. As a process measure, adherence to the TRE protocol was calculated as the percentage of days where self-reported caloric intake was confined to the 8-h window.

The calculation of BMI used a standard formula (weight in kilograms divided by height in meters squared). Participants were asked to record their weight in the morning upon waking, barefoot and wearing light clothing.

Sleep quality change was measured with the Pittsburgh Sleep Quality Index (PSQI) (47). This scale comprises seven components, including subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, sleep medicine use, and daytime dysfunction. Each component is scored on a scale of 0 to 3, and the sum of the scores for each dimension constitutes the PSQI total score. The total score ranges from 0 to 21 points, with higher scores indicating poorer sleep quality.

The STOP-Bang questionnaire (SBQ) was utilized to assess the severity of OSAHS (48). This instrument has been validated in Chinese populations and demonstrates good diagnostic performance for OSAHS (49). It consists of 8 items, yielding a total score of 0–8. A score of ≥3 indicates a high risk of moderate to severe OSAHS.

The Epworth Sleeping Scale (ESS) is an 8-item instrument designed to evaluate excessive daytime sleepiness in patients (50). It has also been validated and is frequently employed in patients with OSAHS (51). Excessive daytime sleepiness was selected as a secondary outcome because it serves as a key clinical marker of OSAHS severity and directly impacts quality of life (52). Measuring changes in ESS scores allowed to explore whether TRE could alleviate this specific symptom beyond weight loss. The total score ranges from 0 to 24, with higher scores indicating more severe daytime sleepiness.

The Calgary Sleep Apnea Quality of Life Index (SAQLI) was primarily used to measure the quality of life in adult patients with OSAHS (53). This instrument consists of 35 items, grouped into 5 subscales: daily functioning, social interactions, emotional functioning, symptoms, and treatment-related symptoms. Since none of the participants in this study had undergone any treatment, only the first four subscales were involved. The total score was the average of four subscales, ranging from 1 to 7 points, with higher scores indicating better quality of life.

For the sample size calculation, we employed a two-sample mean comparison superiority trial approach for estimation. Based on data from a prior report suggesting the efficacy of similar interventions for weight loss (54), we anticipated that the TRE group would achieve a 5% greater reduction in BMI compared to the control group after 12 weeks. With a type-I error rate of 0.05, an estimated sample size of 37 participants per group would provide at least 80% power at week 12 to detect a statistically significant difference in BMI between the TRE group and the control group. Regarding the 10% dropout rate, the final total sample size was 82 participants (41 per group). However, due to recruitment constraints imposed by the COVID-19 pandemic, the final sample size was limited to 68 participants. A post-hoc power analysis using G*Power software, based on the final sample size (n = 60), an observed effect size (Cohen’s d) of approximately 0.35 for the primary outcome (BMI), and α = 0.05, indicated a statistical power of approximately 58%, which is below the conventional 80% threshold. To address methodological limitation, the trial design, statistical analysis plan, and this specific issue were reviewed by an independent statistician. It was acknowledged that the reduced power limits the definitive conclusions; however, the data are valid for estimating effect sizes and generating hypotheses. This limitation is explicitly considered in the interpretation of the findings.

Statistical analysis was performed using SPSS version 26.0 software (IBM, Chicago, IL, United States). Unless otherwise specified, continuous variables were presented as mean ± standard deviation (SD), while categorical variables were presented as frequencies (percentages). Differences between the two baseline groups were analyzed using t-tests for continuous variables or Chi-squared tests for categorical variables. For within-group comparisons, paired sample t-tests or Wilcoxon signed-rank tests were used to analyze the differences before and after the intervention, respectively. The primary analysis for between-group comparisons over time was conducted using linear mixed models, following the intention-to-treat principle. The models contained fixed effects for group, time, and the group-by-time interaction, with participant included as a random intercept. All models were adjusted for the respective baseline outcome value. To control for potential confounding, the following baseline covariates were also included in all primary models: age, gender, education level, employments status, marital status, family monthly income, and AHI. Additionally, in exploratory analyses, categorical and continuous outcomes were analyzed with the chi-square test and one-way ANOVA, respectively. Associations were expressed as odds ratio (OR) with 95% confidence intervals from logistic regression. A two-sided p < 0.05 was considered statistically significant.

A total of 92 individuals were screened, and 68 were randomized to the TRE group (n = 34) or the control group (n = 34). Sixty participants completed the study (Figure 1). Baseline demographic and clinical characteristics were similar between the two groups, as shown in Table 1. Furthermore, scores for outcome measures at baseline were comparable, which are summarized in Table 2.

Linear mixed model analysis found no significant effect of TRE on BMI change over time [F (1,128) = 1.56; p = 0.214]. The between-group comparison showed no statistically significant difference in BMI change at 12 weeks (adjusted mean difference [MD]: −0.18 kg/m², 95% CI: −0.96 to 0.59; p = 0.641; Table 3). However, a distinct pattern emerged from the within-group analyses. Intra-group analysis revealed a significant reduction in BMI from baseline within the TRE group (−0.86 ± 0.55 kg/m²; p = 0.033; Figure 2A), with a mean (SD) adherence of 5.2 (0.7) days per week (74.8% of days over 12 weeks) to the eating window. In contrast, no significant change was observed within the control group (0.02 ± 0.05 kg/m²; p = 0.125, Figure 2B).

Given this heterogeneity in responses, exploratory analyses were conducted. When participants were categorized by a clinically meaningful BMI change threshold (±0.5 kg/m²), the distribution of weight loss, stable, and weight gain categories differed significantly between groups (p = 0.001). Weight gain was more frequent in the TRE group (29.0%) than in the control group (10.3%) (odds ratio: 2.81, 95% CI: 0.84 to 9.36; p = 0.071). Furthermore, within the TRE group, both higher baseline BMI and greater adherence were associated with weight loss (p = 0.010 and p < 0.001, respectively).

Regarding sleep quality, following the 12-week intervention, intra-group analysis showed that the TRE group exhibited significant improvements in the sleep efficiency and daytime dysfunction components of the PSQI compared to baseline (p = 0.030 and p = 0.048, respectively; Figures 3A,B). In contrast, no significant changes in PSQI component or total scores were observed within the control group from baseline (Table 3). At 12 weeks, the linear mixed model analysis indicated that the intervention effect (group-by-time interaction) was not statistically significant for the PSQI total score (adjusted MD: 0.66, 95% CI: −0.55 to 1.87; p = 0.283; Figure 3C). However, exploratory analysis of individual PSQI components suggested specific benefits of TRE. At week 12, compared with the control group, the TRE group demonstrated significantly greater improvement in sleep latency (adjusted MD: 0.44, 95% CI: 0.12 to 0.76; p = 0.007), sleep duration (adjusted MD: 0.54, 95% CI: 0.19 to 0.89; p = 0.003), and sleep efficiency (adjusted MD: -0.55, 95% CI: −0.88 to −0.22; p = 0.001). No significant between-group differences were observed for the remaining components (Table 3).

Linear mixed model analysis showed no significant intervention effects on either the SBQ score or excessive daytime sleepiness (assessed by the ESS). For the SBQ score, the group-by-time interaction was not significant [F (1, 128) = 1.514, p = 0.221], with an adjusted mean difference of −0.27 at 12 weeks (95% CI: −0.77 to 0.24; p = 0.293). Similarly, for the ESS, the group-by-time interaction was not significant [F (1, 128) = 0.748; p = 0.389], and the adjusted mean difference was 1.49 (95% CI: −0.07 to 3.05; p = 0.060) (Table 3). Changes from baseline within each group were also not statistically significant (Table 3).

Quality of life was evaluated using the SAQLI questionnaire. For the total SAQLI score, the interaction between group and time was not statistically significant [F (1,128) = 3.082; p = 0.082]. However, at week 12, the TRE group had a significantly higher total score than the control group (adjusted MD: 0.36, 95% CI: 0.04 to 0.68; p = 0.027; Figure 4A). Exploratory analysis of the SAQLI components showed distinct patterns. The TRE group showed greater improvement than the control group at 12 weeks in emotional functioning (adjusted MD: 0.58, 95% CI: 0.16 to 1.00; p = 0.008; Figure 4B). In contrast, no significant between-group differences were observed for the daily functioning (adjusted MD: 0.47, 95% CI: −0.01 to 0.96; p = 0.055), social interactions (adjusted MD: 0.41, 95% CI: −0.07 to 0.89; p = 0.091), and symptoms (adjusted MD: 0.01, 95% CI: −0.11 to 0.14; p = 0.831) (Table 3). In within-group analyses, the TRE group showed significant improvements from baseline in the total SAQLI score and the social interactions component (p = 0.040 and p = 0.015, respectively; Table 3). In contrast, the control group did not show significant changes from baseline in any SAQLI component or the total score (see Table 3).