Efficacy of digital-based exercise intervention in adults with metabolic associated fatty liver disease: study protocol for a randomised controlled trial in Nanjing, China

Abstract

Introduction

Metabolic associated fatty liver disease (MAFLD) is internationally prevalent, resulting in considerable health and economic costs. Lifestyle adjustments, especially exercise, constitute fundamental treatments; nonetheless, sustained adherence poses significant challenges. This study is to assess the effectiveness of a 12-week digital exercise intervention in decreasing liver fat and enhancing metabolic outcomes in individuals with MAFLD.

Methods and analysis

This pragmatic, investigator-initiated study is a stratified, randomised controlled parallel-group clinical trial. Participants meeting the enrolment criteria will be stratified and randomised based on exercise habits and risk levels. They will be allocated into either the intervention group or the active-control group,with 50 participants per group. Participants in the intervention group will undergo a 12-week digital platform-based exercise intervention customised to their activity risk classification, encompassing exercise monitoring, dietary intervention and health education. Simultaneously, the active-control group will receive traditional exercise monitoring, dietary intervention and health education after the development of their health plan. The primary outcome measure will be liver fat fraction measured by magnetic resonance imaging-proton density fat fraction. Secondary outcome measures will encompass body composition, blood pressure, body weight, body mass index, blood lipids, liver function and additional factors. Evaluations will be performed at baseline and after the intervention (12 weeks).

Ethics and dissemination

Ethical approval of this trial was granted by the ethics committee of the First Affiliated Hospital of Nanjing Medical University (2024-SR-167). Informed written consent will be acquired from study participants prior to enrolment. The results will be published in peer-reviewed journals.

Trial registration number

ChiCTR2400083477.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The study uses a stratified randomised controlled trial design to reduce selection bias and guarantee equitable distribution of participants across varying exercise habits and risk levels.

• Magnetic resonance imaging-proton density fat fraction (MRI-PDFF) serves as a highly precise, non-invasive technique for measuring hepatic fat, providing dependable outcome evaluation.

• The digital intervention includes real-time monitoring and tailored feedback, improving the accuracy and customisation of exercise and dietary recommendations.

• The nature of a single-centre study limits the generalisation of the results.

Introduction

Metabolic associated fatty liver disease (MAFLD), previously referred to as non-alcoholic fatty liver disease (NAFLD), is characterised by the accumulation of triglycerides (TG) (steatosis) in the liver. This condition is currently the most widespread liver disease worldwide, affecting approximately one-quarter of the world’s adult population and presenting considerable health and economic issues to society.1,3 Forecasts indicate that by 2030, around 310 million individuals in China will be affected by MAFLD.⁴ The early phases of MAFLD may not significantly affect health; nevertheless, its advancement to non-alcoholic steatohepatitis (NASH) can result in grave consequences, including liver fibrosis, cirrhosis and liver tumours, all of which elevate mortality rates.⁵

Moreover, NAFLD/NASH significantly increases the incidence of other extrahepatic sequelae, such as type 2 diabetes mellitus, cardiovascular disease (CVD), chronic renal disease and various types of extrahepatic malignancies.⁶ Individuals with NAFLD/NASH exhibit a 64% heightened risk of acquiring CVD, with the occurrence of CVD being proportional to the severity of the hepatic disease.⁷ Additionally, these patients are susceptible to coronary atherosclerosis, myocardial alterations and arrhythmias, which all heighten the risk of heart failure.⁸ NAFLD/NASH markedly increases the likelihood of developing extrahepatic malignancies, such as colorectal tumours,⁹ gastric cancer,¹⁰ pancreatic cancer,¹¹ uterine cancer¹² and breast cancer.¹³ This array of potential health complications highlights the imperative for early detection and proper management of MAFLD to alleviate these significant health risks.

Early prevention, diagnosis, management and treatment of severe complications are economically efficient techniques for controlling MAFLD. Despite the identification of various pathophysiological pathways and genetic variations, pharmaceutical therapy options for fatty liver are still constrained.^{14 15} As a result, prevailing guidelines generally advocate for lifestyle modifications—specifically diet and exercise—as the initial approach in managing MAFLD to reduce fatty liver via weight reduction. ¹⁶A 3% reduction in weight improves steatosis; a 5% reduction diminishes inflammation and a 10% reduction mitigates fibrosis. A 7% weight reduction may be essential for the regression of NASH, especially in obese MAFLD patients.¹⁷ The Mediterranean diet, which decreases the average daily calorie intake by 500 to 1000 kcal, benefits not just NAFLD but also related comorbidities such as type 2 diabetes mellitus, dyslipidaemia and CVD.¹⁸ The principal factor in the enhancement of NAFLD is weight reduction, especially the decrease of visceral fat. Meanwhile, macro- or micronutrients^{19 20} and exercise type²¹ are secondary factors.

Although we have long acknowledged that diet and exercise-based traditional lifestyle therapies are beneficial, implementing sustained lifestyle modifications presents significant hurdles for both clinicians and patients.²² Patients with NAFLD frequently experience a low rate of sustained weight loss, elevated dropout rates and inadequate adherence. Previous studies attribute failure to the following factors. Weight loss interventions for individuals with inadequate self-management frequently prove ineffective due to the tedium of prolonged weight reduction and the healthcare professionals’ inability to offer immediate out-of-hospital monitoring and psychological assistance.^{23 24} Furthermore, traditional treatment approaches depend on inperson consultations, leading to issues such as inconvenient appointment hours, substantial transportation expenses and complex offline medical processes, which significantly discourage treatment adherence.²⁵

Digital therapeutics (DTx) are health software programmes intended to treat or alleviate diseases, disorders, conditions or injuries by providing a medical intervention that has a demonstrable positive therapeutic effect on a patient’s health. Digital therapy may be used independently or in conjunction with other treatments, including pharmaceuticals and medical equipment. Digital therapy involves delivering patients text, images, animations, videos and other media to enhance disease treatment and management via software applications on smartphones, computers and similar devices, aiming to rectify detrimental behaviours or lifestyle habits for the prevention, management or treatment of diseases.^{26 27} Studies on the use of digital therapeutics for the management of NAFLD have been reported. Lim et al²⁸ discovered that a mobile app-facilitated lifestyle intervention resulted in substantial weight reduction and enhancements in anthropometric and clinical parameters among NAFLD patients, with the intervention group exhibiting a five-fold increased probability of attaining ≥5% weight loss. Research study by Kaewdech et al²⁹ indicated that a smartphone-assisted lifestyle intervention markedly enhanced liver stiffness in MASLD patients compared to usual therapy, with a more pronounced reduction noted in the intervention group. Sato et al³⁰ reported that a 48-week DTx intervention using the NASH app led to histological enhancements in patients with NASH, including a decrease in the NAFLD activity score and fibrosis stage, underscoring the potential of DTx as a therapeutic approach for NASH. These studies jointly highlight the efficacy of digital therapeutics in facilitating lifestyle modifications and improving clinical outcomes in individuals with fatty liver disorders.

The lack of digital exercise routines specifically designed for MAFLD sufferers is a significant gap. Moreover, considering that MAFLD patients frequently exhibit other comorbidities, including hypertension, diabetes, obesity and CVD, the risks associated with exercise are increased. Performing risk assessments prior to exercise can identify high-risk individuals, while the adoption of customised exercise plans can enhance safety during training sessions.

Primary objective

To evaluate the efficacy of a 12-week digital stratified exercise intervention compared to an active-control group in diminishing liver fat fraction, as quantified by MRI-proton density fat fraction (MRI-PDFF), in patients with MAFLD.

Secondary objectives

1. To evaluate the effects of the digital-based exercise intervention in comparison to the active-control group on body composition metrics, including body weight, body mass index, waist circumference and hip circumference.

2. To assess variations in metabolic parameters between the digital intervention group and the active control group, including blood lipids (TGs, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C)), blood glucose (fasting blood glucose (FBG), fasting insulin (FINS) and homeostatic model assessment for insulin resistance (HOMA-IR)) and liver function (aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), total bile acid (TBA), adenosine deaminase (ADA), ɑ-L-fucosidase (AFU), total protein (TP), albumin (ALB), globulin (GLO), white GLO ratio, total bilirubin (TBIL), indirect bilirubin (IBIL) and direct bilirubin (DBIL)).

3. To assess and contrast the impact of the intervention against the active-control group on blood pressure and body composition (fat mass, lean body mass and bone mass).

4. To assess and contrast patient adherence to the exercise intervention between the digital intervention group and the active-control group by evaluating participation rates, exercise frequency and duration.

Methods and analysis

Trial design

This pragmatic, investigator-initiated study was a stratified, randomised controlled trial. The trial was conducted at the First Affiliated Hospital of Nanjing Medical University. The trial was registered on 26 April 2024 in the Chinese Clinical Trial Registry (ChiCTR2400083477). The study will commenced on 1 April 2024, and conclude on 1 May 2027. Study flow and investigations are stated in figure 1.

Figure 1. Trial flow. MAFLD, metabolic associated fatty liver disease.

Trial setting

The trial will be conducted in the rehabilitation department and the fatty liver specialist clinic of the First Affiliated Hospital of Nanjing Medical University.

Recruitment

In this study, participants will be recruited through two pathways. First, recruitment will occur in the hospital by putting up posters. Second, recruitment will be conducted online through WeChat Moments, WeChat groups, Weibo and other network platforms to publish and disseminate recruitment information. Recruitment is scheduled to start on 1 August 2025 and finish by 1 May 2027.

Eligibility criteria

Inclusion criteria:

1. Patients diagnosed with MAFLD in the outpatient department.

2. 24 kg/m²≤body mass index (BMI) ≤30 kg/m².

3. Age: 18–60 years old.

4. There is no liver fibrosis, and the controlled attenuation parameter value of fatty liver detected by Fibrosan assay is ≥238 dB/m.

5. Weekly drinking volume (alcohol content) in the past half year was <140 g for men and <70 g for women.

6. Ability to act independently; able to actively cooperate with doctors’ follow-up and conduct regular review according to medical advice.

7. Voluntary signed the informed consent, agreed and actively cooperated with this study.

Exclusion criteria:

1. Patients with viral hepatitis, drug hepatitis, alcoholic fatty liver, organ failure, severe liver and renal insufficiency.

2. Patients with severe hypertension, heart disease, diabetes mellitus, mental and neurological diseases, motor dysfunction and other exercise contraindications.

3. Patients who have undergone bariatric surgery in the past 1 year or have taken weight-loss drugs in the last 3 months and lost more than 10% of their body weight.

4. Patients or family members are unable to perform digital intervention programmes using smartphones. Patients or family members lack the fundamental digital literacy required to engage with digital intervention programmes via smartphones, including the ability to instal apps, upload data and navigate the platform.

5. Subjects are also participating in other clinical research trials.

6. Claustrophobia, patients too large to enter the MRI scanner and patients with MRI contraindications (such as certain metal implants).

Randomisation methods

Randomisation Process

Randomisation is stratified by exercise risk and the randomisation sequence will be computer-generated using random blocks to balance enrolments over the intervention and active-control groups. A statistician who is not enrolled in the study takes charge of the randomisation.

Blinding

In this study, we have separate intervention and assessment teams and all assessment staff are blinded to participant group allocation. The physicians delivering the study treatment will not be blinded and will therefore not be involved in data assessments related to this study.

Interventions

Digital intervention group

Following baseline assessments during outpatient visits, participants will obtain personalised exercise prescriptions and engage in digital-based exercise guidance via a platform specifically developed for MAFLD patients. At the onset of the intervention, participants will be registered on an online management platform through a smartphone application. They will furnish pertinent information (eg, personal information, exercise habits, eating habits, etc) and upload examination results to avert the loss of medical data. They will be thereafter allocated to a team comprising an internist, a sports medicine specialist, a physical therapist responsible for exercise training and a nutritionist to develop a personalised diet programme. During the initial week, participants will be instructed on platform utilisation and become acquainted with accessing their customised training programme details. The 3-month exercise training phase commences thereafter. Participants will follow their individualised training regimens, will measure their heart rates in real time with a fitness tracker (the Xiaomi Mi Band 8) and input data . They will obtain dietary recommendations from the nutritionist via the platform and will adhere to nutritional instruction. Participants will be required to take and upload pictures of their meals to enable calorie intake assessment. Simultaneously, they will receive health education messages to enhance their understanding of MAFLD and exercise, partake in post-training quizzes and take weekly weight measurements, uploading the data to the weight tracking module. Healthcare professionals (eg, internists, sports medicine specialists, physical therapists, nutritionists) will have access real-time data submitted by participants. This data includes exercise metrics (eg, duration, intensity, heart rate), dietary intake and weight measurements. Healthcare professionals will review these data weekly to monitor participants’ progress and adherence to the intervention. Healthcare professionals will offer comments to participants through the platform at least once a week. This feedback will be tailored and derived from the analysed data. If a participant’s exercise intensity continually falls below the specified level, the physical therapist may offer targeted recommendations to safely enhance intensity. Likewise, if dietary intake reveals a trend of excessive calorie consumption, the nutritionist can provide customised guidance to assist the participant in making better selections. This constant contact will guarantee that participants have continual support and direction, hence improving their engagement and compliance with the intervention.

Active-control group

Following the baseline assessments during the enrolment visit, individuals will be provided with conventional exercise prescriptions, nutritional advice and health education. During the initial week, they will be required to keep a record of their diet and exercise . Subsequently, they will commence a 3-month fitness programme founded on a customised regimen provided by the physical therapist. They will also receive monthly handouts with health education materials to improve their understanding of MAFLD. Furthermore, they will be instructed to adhere to the nutritional guidelines and sustain a healthy diet to manage energy consumption. Throughout the 3-month intervention, participants will be required to document their exercise and dietary intake every day, with these records being collected during their monthly appointments.

Exercise

Patients may possess several simultaneous underlying disorders. Data regarding individual exercise preferences, current habits, existing comorbidities, disease-related symptoms and signs will be collected through the app for pre-exercise screening. Pre-exercise assessments of CVD risk and cardiopulmonary function will be performed to identify potential abnormalities, hence mitigating the risk of exercise-related incidents. Exercise programmes will be tailored according to screening and assessment outcomes, adhering to guidelines from the American College of Sports Medicine (ACSM) for exercise testing and prescription (11th edition).³¹ Each exercise session will comprise a 10-min warm-up, formal training, a 5-min cool-down and 5 min of stretching, totalling 40 min. The regimen will mainly comprises aerobic exercises (eg, jogging and swimming), supplemented with resistance exercises (eg, push-ups, planks) 2–3 times per week. The app will feature exercise videos that can be customised for varying intensities and presented by a coach. Patients can modify intensity based on heart rate and tolerance or advance incrementally. The application will synchronise with wearable devices such as fitness bands to track heart rate and calories expenditure in real-time. If the heart rate exceeds the target range, patients will be prompted to reduce intensity for safety. Exercise during the training phase will be modified according to established exercise routines (minimum of 3 days per week, 30 min per day of moderate-intensity physical activity for at least 3 months) and the presence or history of cardiovascular or renal disease symptoms.

Dietary intervention

Nutritional prescriptions will be formulated following a baseline assessment of the patient’s current dietary intake and weight by the nutritionist. These prescriptions aim to moderately reduce total energy intake and guide the proportions of macronutrient intake. Patients will generally receive nutritional advice throughout their appointments. Patients will first be asked about their typical eating patterns, including food types, preferences, history of food allergies and nutritional needs associated with concurrent illnesses. Thereafter, according to each patient’s established weight loss objective and physical activity level, the monthly weight loss targets and daily caloric needs will be determined. Patients will receive dietary advice tailored to their daily energy requirements, accompanied by a food exchange table to aid in the selection of energy-equivalent foods. Patients in the digital intervention group will receive their prescriptions online and will be instructed to log into the MAFLD app daily to upload photos of their food intake, enabling the app to automatically calculate their actual daily energy intake.

The virtual dietician will offer immediate oversight and direction. During the approach, energy adjustments will be implemented as patients lose weight and their activity level changes until they attain their optimal weight and waist circumference. Participants in the active-control group will receive conventional dietary counselling during their monthly consultations with the same dietitian.

Health education

Health education is essential for addressing the misunderstandings and misconceptions regarding the causes and effects of MAFLD. By providing pertinent information, it improves patients’ comprehension of MAFLD’s pathophysiology, risks,and treatment alternatives, promoting increased collaboration during the intervention process. The health education materials are developed in accordance with the guidelines for the prevention and treatment of metabolic dysfunction-associated (non-alcoholic) fatty liver disease (V.2024),³² integrating pertinent domestic and international research and clinical expertise. This approach guarantees that the information presented is both substantiated by evidence and clinically pertinent. The MAFLD management application delivers health education via push notifications, providing patients with essential information. Furthermore, it periodically disseminates popular science articles or videos to enhance patient understanding. Conversely, the active-control group will receive conventional health education at each visit, accompanied by handouts that include identical information to that offered in the digital intervention group.

Outcome measures

Primary outcome

The primary outcome is liver fat fraction measured by MRI PDFF.

Secondary outcomes

1. Liver function: AST, ALT, ALP, GGT, TBA, ADA, AFU, TP, ALB, GLO, white GLO ratio (A/G), TBIL, IBIL and DBIL.

2. Blood lipid: TG, TC, LDL-C and HDL-C.

3. Blood glucose: FBG, FINS and HOMA-IR.

4. Liver stiffness measurement (LSM) value measured by FibroScan.

5. Anthropometric parameters: body weight, body mass index, waist circumference, hip circumference.

6. Body composition: fat mass, lean body mass and bone mass.

7. Blood pressure.

8. Compliance.

Figure 1 presents an overview of procedures and measures used according to the Standard Protocol Items: Recommendations for Interventional Trials statement.

Methods of measurements

Hepatic fat content assessment

MRI-PDFF measurement and grouping: the patients will undergo MRI 3.0 upper abdominal non-contrast scanning and MRI-PDFF evaluation. Operation method: using the GE Architect 3.0T nuclear magnetic resonance machine, the examination was completed by one technologist with standardised training in liver MRI, using Dual-Echo Hybrid Iterative Decomposition of water and fat with Echo Asymmetry and Least-squares estimation Quantification fat measurement sequence, relevant parameters: field of view (FOV)=40, phase FOV=0.8, slice=6 mm, repetition time (TR)=5.7 ms, repetition time (TE)=Minimum, and three different loci in the largest layer of liver imaging (left hepatic lobe, right anterior lobe, right posterior lobe, avoiding large blood vessels, bile ducts and obvious artefacts). The MRI images will undergo quality control by an additional highly skilled radiologist, wherein the quality control indicators comprise the complete labelling of patient information, the qualification of the image acquisition and the precision of the measured values. The average of the 3-site MRI-PDFF will be taken.

LSM value measured by FibroScan

The Haskell FIBROTOUCH-FT5000 will be used to carry it out. The patient will be positioned supine with the right hand behind the head, exposing the auxiliary space in the right lobe region of the liver. The detection area is usually taken as the area enclosed by the horizontal line of the xiphoid process, the midline of the right axillary and the lower border of the rib cage. The examiner will apply the probe perpendicular to the skin surface in the selected intercostal area and presses the probe button to initiate image acquisition and gather measurement. The final measurement will be the median of ten successful measurements. The degree of liver fibrosis will be determined based on the LSM value.

Blood samples

Venous blood samples will be collected at 07:00–08:00 under standardised fasting conditions. FBG, FINS, glycosylated haemoglobin, TC, TGs, HDL, LDL, ALT, AST, GGT, ALP, TBAs, ADA, fructosidase, TP, ALB, GLO, TBIL, DBIL and IBIL will be measured from serum samples by conventional methods.

Anthropometric parameters

Height will be measured using a wall-mounted stadiometer, and body weight will be measured using a calibrated scale. BMI will be calculated as weight in kilograms divided by the square of height in metres (kg/m²). Waist circumference will be measured at the midpoint between the lowest rib margin and iliac crest using a non-elastic tape, recorded after normal expiration. Hip circumference will be assessed at the maximal protrusion of the buttocks. Both measurements will be taken in a standing position by trained staff, repeated twice (average recorded; third measurement if discrepancy >1 cm), with subjects in light clothing. Blood pressure will be recorded after 5 min of seated rest.

Body composition

Body composition will be assessed using a bioelectrical impedance analysis device. All measurements will be performed by trained staff from the hospital's nutrition department. The measurement protocol is as follows: (1) the examination room temperature will be adjusted to ensure participant comfort; (2) participants will be instructed to empty their bladder and to fast for at least 2 hours prior to measurement; (3) the instrument will be preheated according to manufacturer guidelines; (4) participants will remove shoes and socks, etc, and wipe the palm surface of the subject with alcohol and the contact surfaces of hands and feet will be cleaned with alcohol before electrode contact; (5) after entering participant information into the device, the measurement will be initiated. The result will be printed upon completion, and the device will be returned to standby mode.

Compliance

The intervention adherence will be evaluated using various methods. Adherence in the digital intervention group will be assessed by monitoring the frequency and duration of exercise sessions recorded via the smartphone application and fitness tracker. The platform will record the frequency of dietary data submissions and engagement with health education resources. Furthermore, the number of feedback exchanges between participants and healthcare professionals on the platform will be recorded. In the active-control group, adherence will be assessed through the completeness and consistency of physical activity and dietary logs, in addition to participation in monthly appointments. The adherence rate in both groups will be calculated as the percentage of prescribed intervention activities that were successfully performed.

The details of enrolments, interventions and assessments are presented in table 1.

Table 1. Scheduled events and timeline.

	Study period
	Enrolment	Allocation	Post-allocation
Time point	Baseline	0	Week 0	…	Week 12
Enrolment:
Eligibility screen	X
Informed consent	X
Allocation		X
Interventions:
Digital intervention group
Active-control Group
Assessments:
MRI-PDFF	X				X
AST	X				X
ALT	X				X
ALP	X				X
GGT	X				X
TBA	X				X
ADA	X				X
AFU	X				X
TP	X				X
ALB	X				X
GLO	X				X
A/G	X				X
TBIL	X				X
IBIL	X				X
DBIL	X				X
TG	X				X
TC	X				X
LDL-C	X				X
HDL-C	X				X
FBG	X				X
FINS	X				X
HOMA-IR	X				X
LSM value	X				X
Anthropometric parameters	X				X
Body composition	X				X
Blood pressure	X				X
Compliance					X

ADA, adenosine deaminase; AFU, ɑ-L-fucosidase; A/G, white globulin ratio; ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; DBIL, direct bilirubin; FBG, fasting blood glucose; FINS, fasting insulin ; GGT, glutamyl transpeptidase; GLO, globulin; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, homeostatic model assessment for insulin resistance index; IBIL, indirect bilirubin; LDL-C, low-density lipoprotein cholesterol; LSM, Liver stiffness measurement; MRI-PDFF, MRI-proton density fat fraction; TBA, total bile acid; TBIL, total bilirubin; TC, total cholesterol; TG, triglyceride; TP, total protein.

Sample size

Sample size was calculated with Stata V.14 (Stata Corp, College Station, Texas, USA), drawing on clinical experience; liver fat fraction measured by MRI-PDFF from the two groups served as the basis for this calculation. To attain 80% power and account for a 20% dropout rate, a two-tailed significance test with a 0.05 alpha level was used. A minimum of 50 participants per group, totalling 100 participants, is required for the study.

Exit criteria and termination criteria

Subjects, according to the patient management protection rules, are entitled to withdraw from the study at any time and for any reason. If a patient requests to withdraw from the trial, the reason(s) for the withdrawal should be meticulously documented as an assessment index at the moment of termination. If a participant fails to return for scheduled follow-up, the investigators will attempt to contact them via telephone or letter to ascertain the reason, as this information is crucial for interpreting the study results. Clinical trial participants who wish to assess the potential effects of termination cases on research conclusions should meticulously record the reasons for trial termination and their connection to the trial.

After the trial concludes, all original documentation for every participant that was withdrawn or excluded should be preserved and archived for audit purposes. Treatment will also be discontinued, and the participant will be withdrawn from the study, if any of the following occurred: (1) any expected adverse events such as fractures, heart disease, stroke, etc; (2) hepatic decompensation during enrolment; (3) compliance <80% or > 120%, non-compliance with treatment; (4) pregnancy during this study; (5) intolerable adverse events (the patient or investigator may decide to withdraw from the study; (6) diseases or factors unrelated to treatment; (7) loss to follow-up; (8) breaking the blind.

Data monitoring

To safeguard the trial’s validity and integrity, a data safety monitoring committee (DSMC) will be established. Meeting on a regular basis throughout the trial, the DSMC’s responsibilities include monitoring and reviewing patient safety; requesting, if necessary, the conducting of interim data analysis and reviewing patient recruitment, accrual and withdrawal. The DSMC is tasked with more than just making suggestions regarding whether to keep the trial going or change it. If the DSMC determines that the rehabilitation programme or the evaluations are to blame for any serious adverse occurrences, the trial may be terminated.

Statistical analysis

Statistical analyses will be conducted using SAS V.9.4 software. Continuous variables will be presented as mean±SD for normally distributed data or median (interquartile range, IQR) for non-normally distributed data. Categorical variables will be expressed as numbers (percentages). Baseline characteristics will be compared between groups using χ² tests or Fisher’s exact tests for categorical variables and independent t-tests or Mann-Whitney U tests for continuous variables, depending on data distribution. For outcomes measured at two time points (baseline and post intervention at 12 weeks), paired t-tests or Wilcoxon signed-rank tests will assess within-group changes, while between-group differences will be evaluated using linear mixed-effects models adjusted for stratification factors. For the anthropometric parameters measured repeatedly (body weight, BMI, waist circumference, hip circumference), mixed-effects models with random intercepts for participants and fixed effects for time, group and time-group interaction will be employed to account for repeated measures. Both intention-to-treat and per-protocol analyses will be performed. Missing data will be handled multiple imputation by chained equations, creating 20 imputed datasets. Sensitivity analyses, including complete-case and pattern-mixture models, will be conducted to assess robustness of the results under missing-not-at-random assumptions. Effect sizes will be reported with 95% CIs, and two-tailed p values<0.05 will indicate statistical significance. Subgroup analyses will explore intervention effects across pre-specified risk strata by including interaction terms in adjusted models.

Ethics and dissemination

This study has been approved by the institutional review board of the First Affiliated Hospital of Nanjing Medical University (reference number: 2024 SR-167). The study will adhere to all relevant guidelines for good scientific and clinical practice. None of the measurements is known to have any significant health risk. All individuals will voluntarily participate in this study and obtain informed consent prior to enrolment in the study, and an example consent form is available as online supplemental material 1. Participants will be informed that they are free to exit the study or choose to withdraw consent at any time during the participation process without any repercussion. All data will be handled and archived confidentially. All samples will be coded using an ID number without any personal identifiers. All image scans and background information will be electronically transferred and stored in a secure, password-protected database. If analyses are performed outside the data collection site, sample transportation will comply with relevant safety standards. The duration of the sample storage is generally 10 years. If the required storage period is exceeded, renewed permission will be sought from the participants and the ethical committee. If a participant is deceased, permission will be sought from their next of kin. This study will be conducted in accordance with the ethical principles of the Declaration of Helsinki. The final study results will be disseminated through publication in peer-reviewed journals. The public will be informed about the results of the study through media reports, briefings, etc.

Discussion

With the rising prevalence of obesity and diabetes, the incidence of MAFLD continues to escalate, which profoundly affects human health. Although exercise-focused therapies are considered advantageous for MAFLD, inadequate long-term adherence frequently diminishes their practical efficacy. Our study aims to evaluate the efficacy of digital exercise intervention delivered via a smartphone app for MAFLD, compared with traditional exercise programmes. Given the heterogeneity of the MAFLD population, we implemented a risk-stratified exercise regimen as advised by the ACSM’s guidelines for exercise testing and prescription (11th edition). This programme is designed based on patients’ cardiovascular risk profiles to ensure exercise intensity and safety. Consequently, we devised a randomised, stratified clinical trial for MAFLD patients, considering their exercise habits and risks, to explore treatment effects, compliance and training programme complications, thus providing a reliable clinical foundation for MAFLD treatment.

Recent studies on exercise interventions for fatty liver disease have primarily concentrated on singular exercise modalities (eg, specific regimens of aerobic exercise, resistance training or combined aerobic-resistance exercise) or non-stratified personalised exercise prescriptions.^{33 34} For instance, research conducted by Charatcharoenwitthaya et al³⁵ and Guo et al³⁶ used uniform exercise prescriptions for all participants, while Mascaró et al³⁷ employed individualised prescriptions. Patients with MAFLD encounter considerable exercise-related hazards, especially cardiovascular risks, due to their significantly heightened prevalence and incidence of CVD events. Classifying exercise interventions according to pre-evaluated cardiovascular risk profiles facilitates customised prescriptions for various risk groups, effectively reducing exercise-related cardiovascular problems and maintaining safety. In contrast, single-mode or non-stratified personalised prescriptions may overlook safety considerations for high-risk populations, increasing the risk of exercise-related adverse events. As of now, no research has conducted risk-stratified exercise therapies for MAFLD. This study addresses this gap by categorising participants according to cardiovascular risk evaluations. For low-risk individuals, universally applicable exercise prescriptions eliminate the need for pre-exercise health screenings, streamlining care, reducing unnecessary medical resource utilisation and enhancing patient convenience. Conversely, for high-risk individuals, health screenings identify exercise-related risks, enabling individualised prescriptions that balance efficacy with safety, thereby minimising injuries and hazards. The widespread adoption of this stratified approach could establish an effective self-management model for the broader MAFLD population.

Digital interventions possess the capacity to enhance adherence via multiple ways. Initially, behavioural modification strategies such as goal establishment, progress monitoring and incentives for sustained participation can be integrated into the application, encouraging users to adhere to their exercise and eating plans. Second, individualised feedback derived from real-time data obtained from wearable devices and dietary inputs facilitates prompt modifications and the reinforcement of beneficial behaviours. Remote oversight by healthcare specialists guarantees ongoing assistance and responsibility, facilitating the timely resolution of unforeseen concerns. In contrast to conventional approaches that depend on regular inperson consultations and self-reported information, digital interventions provide more prompt and interactive involvement, may improving long-term adherence. Particular features of the MAFLD app designed to enhance adherence include its intuitive design that simplifies the monitoring of exercise and nutritional practices. The application offers instantaneous feedback and reminders to help users stay committed to their designated regimens. The platform’s capability to facilitate interaction with healthcare experts provides a sense of ongoing support and oversight, frequently absent in conventional exercise regimens. The application integrates educational resources and quizzes, improving participants’ comprehension of MAFLD and the significance of lifestyle changes, therefore potentially increasing their motivation to adhere to the intervention.

The randomised controlled trial approach used in this study provides robust evidence for causal inference regarding intervention efficacy. However, we recognise its intrinsic limits in addressing the multifaceted complexity of digital health interventions. In contrast to pharmaceutical trials, digital solutions necessitate simultaneous assessment of clinical results and the dynamics of human-technology interaction, encompassing usability, engagement and adherence across time. Obstacles to usability, sustained engagement trends and contextual elements affecting compliance. Future research may employ hybrid effectiveness-implementation designs, combining mixed-methods approaches (eg, nested qualitative interviews, usage analytics) with adaptive trial protocols that allow for real-time intervention adjustment informed by user input. Embedding a contemporaneous embedded design would provide quantitative evaluation of hepatic fat reduction while qualitatively examining how patients’ perceived usability, cultural acceptance of app-based monitoring and trust in data security affect long-term adherence. Moreover, using implementation science frameworks like Reach, Effectiveness, Adoption, Implementation and Maintenance framework could methodically assess scalability factors, encompassing healthcare provider adoption rates and institutional infrastructure prerequisites for practical application.

Older adults will be excluded from this study due to concerns that mobile phone applications might be perceived as uncomfortable or cumbersome, potentially compromising intervention compliance. This exclusion may introduce selection bias and limit the generalisability of our findings, despite the study's aims for inclusivity. This constraint highlights the necessity for forthcoming research to create more accessible digital tools specifically designed for older adults, guaranteeing their participation in digital health treatments.

The research contrasts digital interventions with conventional MAFLD management techniques, developing offline health plans and using internet-based remote monitoring to guarantee the effectiveness and safety of the intervention. Nonetheless, internet-based management platforms may jeopardise patient personal information and privacy through data collection, transfer, storage and administration. Therefore, it is essential to safely handle data and protect patient confidentiality. In our study, all data communicated via the app will be protected using advanced encryption standards to avert interception and unwanted access. The identity of participants will be anonymised in the database to safeguard their privacy. Only anonymised data will be used for research and reporting. Data storage will comply with rigorous security protocols, including restricted access and periodic security audits. The database will be situated on secure servers equipped with firewalls and intrusion detection systems. We will perform regular security audits to guarantee adherence to data protection requirements and to identify and mitigate any potential vulnerabilities.

The primary outcome measure of this study is MRI-PDFF. Among non-invasive imaging modalities, MRI-PDFF currently demonstrates the highest diagnostic accuracy for hepatic steatosis in MAFLD patients and is acknowledged as the most accurate approach for quantifying liver fat.³⁸ MRI-PDFF is a non-invasive and quantitative method that delivers precise and swift evaluations of hepatic fat content, demonstrating high sensitivity, specificity, reproducibility and repeatability, so rendering it appropriate for the dynamic monitoring of hepatic steatosis. In contrast to histological analysis, MRI-PDFF facilitates an extensive assessment of fat distribution throughout the entire liver parenchyma.³⁹

The necessity for specialised equipment and skilled specialists restricts its extensive clinical use. Furthermore, the exorbitant expense of examinations places financial strain on both consumers and healthcare systems.^{40 41} As a result, MRI-PDFF is predominantly employed as a research instrument rather than in standard clinical applications. Ongoing technology developments and anticipated cost reductions suggest that MRI-PDFF may be viable for future clinical implementation.