Efficacy of abdominal vibration technique in Tuina in reducing depression, anxiety, and stress in patients with prediabetes

Abstract

BACKGROUND

Prediabetes is a high-risk precursor for diabetes development and is associated with increased psychological distress. Conventional pharmacological treatments for prediabetes have limitations, including adverse effects and poor patient compliance. Tuina’s abdominal vibration technique, a traditional non-pharmacological intervention, has shown promising results in glycemic regulation. However, its effects on psychological well-being remain largely unexplored.

AIM

To investigate efficacy of abdominal vibration at varying frequencies in Tuina for alleviating psychological and metabolic effects in prediabetes patients.

METHODS

A prospective cohort study (April 2025 to April 2026) at Dongfang Hospital of Beijing University of Chinese Medicine included 120 prediabetes patients. Participants were allocated to three groups based on vibration frequency: (1) Low-frequency (400 times/minute, n = 40); (2) Medium-frequency (500 times/minute, n = 40); and (3) High-frequency (600 times/minute, n = 40). All participants received 30-minute sessions three times weekly for 3 months, with follow-up at 6 months and 12 months. Primary outcomes included changes in depression (Beck Depression Inventory-II), anxiety (State-Trait Anxiety Inventory), and stress (Perceived Stress Scale) levels.

RESULTS

After a 3-month intervention period, with follow-ups at 6 months and 12 months, patients in all three groups showed significant improvements in depression, anxiety, and stress scores compared to baseline (P < 0.001). The high-frequency group demonstrated the most substantial psychological improvements (mean reduction in depression scores: 9.2 ± 2.3 points; anxiety: 8.7 ± 2.1 points; stress: 10.4 ± 2.5 points). These psychological improvements correlated significantly with reductions in insulin resistance (r = 0.68, P < 0.001). The high-frequency group also showed the most significant improvements in glycemic parameters, with mean reductions in fasting plasma glucose, 2-hour postprandial glucose, and glycosylated hemoglobin concentrations of 0.92 mmol/L, 1.87 mmol/L, and 0.51%, respectively. Correlation analysis revealed significant associations between improved psychological parameters and enhanced glycemic control.

CONCLUSION

Tuina’s abdominal vibration technique, especially at high frequency reduces depression, anxiety, and stress in prediabetes patients, correlating with enhanced glycemic control and insulin sensitivity, suggesting a bidirectional relationship between psychological and metabolic health.

Key Words: Prediabetes; Impaired glucose tolerance; Abdominal vibration technique; Tuina; Depression; Anxiety; Stress; Insulin resistance; Psychological well-being; Complementary therapy

Core Tip: This study explored the impact of different frequencies of abdominal vibration in Tuina on psychological well-being and insulin resistance (IR) in patients with prediabetes. The findings shows that all vibration frequencies significantly reduce depression, anxiety, and stress, with the high-frequency group showing the greatest improvements. These psychological benefits were strongly associated with better glycemic control and reduced IR, highlighting a potential bidirectional relationship between psychological health and metabolic outcomes in prediabetes management.

INTRODUCTION

Prediabetes, characterized by impaired glucose tolerance or impaired fasting glucose levels, represents a critical transitional state between normal glucose metabolism and type 2 diabetes mellitus (T2DM)[1]. Globally, the prevalence of prediabetes continues to rise dramatically alongside increasing rates of obesity and sedentary lifestyles[2]. Research indicates that approximately one-third of individuals with prediabetes will progress to T2DM within 5-10 years if left untreated[3]. Beyond its metabolic implications, prediabetes is increasingly recognized as a multifaceted condition associated with significant psychological burden, including elevated rates of depression, anxiety, and stress among affected individuals compared to normoglycemic individuals[4].

The diagnosis of prediabetes often triggers health anxiety, fear of disease progression, and feelings of loss of control, which can negatively impact self-care behaviors and adherence to lifestyle modifications[5].

Conventional management of prediabetes primarily focuses on lifestyle modifications, including dietary changes and increased physical activity, with pharmacological interventions such as metformin use reserved for high-risk individuals[6]. However, these approaches face significant challenges. Lifestyle modifications demand sustained behavioral changes which many patients struggle to maintain long-term, while pharmacological interventions carry risks of adverse effects[7]. The use of alpha-glucosidase inhibitors such as acarbose, while effective in reducing postprandial glucose excursions, frequently causes gastrointestinal side effects leading to poor adherence[8]. The limitations of current approaches underscore the need for complementary therapeutic strategies that address both the metabolic and psychological aspects of prediabetes.

Traditional Chinese medicine (TCM) offers several complementary approaches to prediabetes management. In TCM theory, prediabetes aligns with the concept of "spleen deficiency", and treatment principles aim to tonify the spleen and regulate glucose metabolism[9]. Among various TCM modalities, Tuina’s abdominal vibration technique represents a particularly promising non-pharmacological intervention that has gained increasing attention in recent years.

The technique involves rhythmic vibration manipulations of the abdominal region and has demonstrated potential efficacy in improving glucose metabolism[10]. The technique is simple to perform, cost-effective, and well-accepted by patients. Preliminary studies suggest that the abdominal vibration technique may improve insulin sensitivity, reduce fasting and postprandial glucose levels, and enhance overall glycemic control in individuals with prediabetes[11]. The proposed mechanisms behind these effects include enhanced circulation to pancreatic tissues, improved digestive function, and reduced systemic inflammation[12].

Recent research has begun to elucidate the molecular pathways through which abdominal vibration technique might exert its metabolic effects. Studies demonstrated that abdominal vibration may activate the silent information regulator 1/peroxisome proliferator-activated receptor gamma coactivator-1 alpha signaling pathway in skeletal muscle, a key regulatory pathway for energy metabolism and mitochondrial biogenesis[13]. This activation leads to improved insulin sensitivity and glucose utilization. Additional research has shown that the abdominal vibration technique may downregulate pro-inflammatory cytokines such as tumor necrosis factor-alpha (TNF-α) and interleukin 6 (IL-6), thereby reducing low-grade systemic inflammation characteristic of prediabetes[14].

While existing research provides valuable insights into the metabolic effects of this abdominal vibration technique, its impact on the psychological well-being of prediabetes individuals remains largely unexplored. Given the established bidirectional relationship between psychological distress and metabolic dysfunction, it is reasonable to hypothesize that abdominal vibration technique might confer psychological benefits alongside its metabolic effects. To address these knowledge gaps, the present study aimed to investigate the efficacy of different frequencies of the abdominal vibration technique in Tuina in reducing depression, anxiety, and stress among patients with prediabetes while simultaneously evaluating its effects on insulin resistance (IR) and oxidative stress markers. By exploring both psychological and metabolic outcomes, this study seeks to provide a comprehensive assessment of abdominal vibration technique as a complementary therapeutic approach for prediabetes management.

MATERIALS AND METHODS

Study design and setting

A prospective cohort study was conducted from April 2025 to April 2026 at the Prediabetes Management Center of Dongfang Hospital of Beijing University of Chinese Medicine. The study protocol was reviewed and approved by the Institutional Ethics Committee and registered in the Chinese Clinical Trial Registry. The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. All participants provided written informed consent before enrollment.

Participants

Sample size calculation: Sample size calculation was performed using G*Power 3.1 software. Based on previous studies on psychological interventions in prediabetes, we anticipated a medium effect size (f = 0.25) for the primary outcome measures. With an alpha error of 0.05, a statistical power of 0.80, and accounting for three intervention groups, with two primary measurement time points (baseline and 12 months), the required sample size was calculated as 108 participants. To accommodate an anticipated dropout rate of 10%, we aimed to recruit 120 participants (40 per group).

Inclusion and exclusion criteria

Eligible participants were adults aged 18-65 years who were diagnosed with prediabetes according to the 2020 Chinese Diabetes Society guidelines, defined as impaired glucose tolerance (2-hour plasma glucose levels of 7.8-11.0 mmol/L following a 75 g oral glucose tolerance test) and/or impaired fasting glucose [fasting plasma glucose (FPG) levels of 6.1-7.0 mmol/L].

The exclusion criteria included: (1) Diagnosis of T2DM; (2) Acute illness or severe metabolic disorders within the previous 3 months; (3) Severe cardiovascular, hepatic, renal, or other systemic diseases; (4) Active or severe psychiatric disorders requiring current medication (mild depression or anxiety without medication was not exclusionary); (5) Pregnancy or lactation; (6) Participation in other clinical trials within the previous 3 months; (7) Ongoing use of medications known to affect glucose metabolism or psychological status; and (8) Contraindications to the abdominal vibration technique, including recent abdominal surgery, abdominal hernia, intestinal obstruction, or abdominal malignancy.

Recruitment and group allocation

Participants were recruited through outpatient clinic referrals, community health screenings, and advertisement in local media. Following initial screening, eligible participants were allocated into three groups based on the frequency of abdominal vibration: (1) Low-frequency (400 times/minute, n = 40); (2) Medium-frequency (500 times/minute, n = 40); and (3) High-frequency (600 times/minute, n = 40). Group allocation was conducted using a computer-generated randomization sequence with stratification by age, sex, and baseline glucose levels to ensure comparability between groups.

Intervention protocol

Abdominal vibration technique in Tuina: Abdominal vibration was administered by licensed Tuina practitioners with at least 5 years of experience and specialized training in the standardized protocol developed for this study. The technique was performed with the patient in a supine position with knees slightly flexed.

The core technique involved using the forearm to actively drive the wrist joint flexion and extension to complete the palm vibration. The palm center (Laogong acupoint) was aligned with the patient's navel (Shenque acupoint), with fingers naturally relaxed, performing rhythmic vibrations at specific frequencies with relatively consistent amplitude according to group allocation: (1) 400 times/minute for the low-frequency group; (2) 500 times/minute for the medium-frequency group; and (3) 600 times/minute for the high-frequency group. The frequency was controlled and monitored using a metronome application. The frequency selection was based on traditional Tuina practice guidelines and preliminary mechanistic studies. The 400-600 times/minute range represents the physiologically effective spectrum for abdominal vibration, with 400 times/minute considered the minimum threshold for therapeutic effect, 500 times/minute representing standard clinical practice, and 600 times/minute approaching the upper limit of comfortable manual vibration that can be sustained by practitioners. Previous studies suggest that 600 times/minute may represent a biological threshold where mechanoreceptor activation and tissue perfusion reach optimal levels without causing excessive stimulation or practitioner fatigue.

Intervention schedule and compliance monitoring

All participants received abdominal vibration technique three times per week for 3 months as the treatment period, with follow-up assessments at 6 months and 12 months. Each session lasted 30 minutes and was conducted at approximately the same time of day (between 9:00 AM and 11:00 AM) to minimize circadian variations in metabolic parameters. To ensure consistent technique delivery across practitioners, regular calibration sessions were conducted monthly.

Participants were instructed to maintain their usual activities of daily living and not to initiate any new exercise programs or significant dietary changes during the study period. Compliance with the intervention was monitored through attendance records, and participants missing more than three consecutive sessions or more than 10% of total sessions were excluded from the per-protocol analysis.

Outcome measures

Primary outcomes (psychological parameters): Psychological status was assessed using validated Chinese versions of the following instruments: (1) Depression: The Beck Depression Inventory-II (BDI-II), a 21-item self-report questionnaire measuring the severity of depressive symptoms. Scores range from 0 to 63, with higher scores indicating more severe depression; (2) Anxiety: The State-Trait Anxiety Inventory (STAI), consisting of two 20-item scales measuring state and trait anxiety. Scores range from 20 to 80 for each scale, with higher scores indicating greater anxiety; and (3) Stress: The Perceived Stress Scale (PSS), a 10-item scale measuring the degree to which situations in one's life are appraised as stressful. Scores range from 0 to 40, with higher scores indicating higher perceived stress.

These psychological assessments were conducted at baseline, 3 months, 6 months, and 12 months by trained psychologists who were blinded to participants' group allocation.

Secondary outcomes (metabolic parameters): The following metabolic parameters were measured: (1) Glycemic parameters, including FPG levels, 2-hour postprandial glucose levels following a standardized meal, and glycosylated hemoglobin values were evaluated; and (2) Insulin parameters were also evaluated, comprising fasting insulin (FINS) levels, 2-hour postprandial insulin levels, and IR index calculated using the homeostasis model assessment of insulin resistance (HOMA-IR) formula: (FPG × FINS)/22.5.

Various oxidative stress markers were measured, specifically serum cysteine-C (Cys-C) levels, beta-2-microglobulin (β2-MG) levels, malondialdehyde (MDA) levels, and total antioxidant capacity (T-AOC).

Inflammatory markers were assessed, including high-sensitivity C-reactive protein (hs-CRP), TNF-α, and IL-6 Levels.

Additionally, anthropometric measurements were recorded, consisting of body weight, body mass index (BMI), waist circumference, hip circumference, and waist-to-hip ratio. Blood samples were collected after an overnight fast (at least 10 hours) at baseline, 3 months, 6 months, and 12 months. Additional blood samples were collected 2 hours after a standardized meal containing approximately 500 kcal. All laboratory analyses were performed at the central laboratory of our institution by technicians blinded to participants' group allocation.

Safety monitoring

Safety was monitored through: (1) Regular assessment of vital signs (blood pressure, heart rate, respiratory rate, and body temperature); (2) Electrocardiogram (ECG) at baseline and 12 months; (3) Comprehensive blood chemistry, including liver and kidney function tests, at baseline, 6 months, and 12 months; and (4) Any adverse events occurring throughout the study period were documented.

Statistical analysis

All statistical analyses were performed using Statistical Package for the Social Sciences version 26.0 (IBM Corp., Armonk, NY, United States). A P value of < 0.05 was considered statistically significant. Data were presented as means ± SD for continuous variables and as frequencies and percentages for categorical variables.

The Shapiro-Wilk test was used to assess the normality of continuous variables. For normally distributed data, one-way analysis of variance (ANOVA) was used to compare baseline characteristics and changes in outcome measures across groups, followed by Bonferroni post-hoc tests for pairwise comparisons. For non-normally distributed data, the Kruskal-Wallis’ test was used for between-group comparisons, followed by Dunn's post-hoc tests.

To evaluate changes in outcome measures over time within and between groups, repeated measures ANOVA was performed, with time (baseline, 3 months, 6 months, and 12 months) as the within-subject factor and group (low-frequency, medium-frequency, and high-frequency) as the between-subject factor. The Greenhouse-Geisser correction was applied when the assumption of sphericity was violated.

Pearson's or Spearman's correlation coefficients were calculated to examine associations between changes in psychological and metabolic parameters. Multiple linear regression analyses were performed to identify predictors of changes in psychological and metabolic outcomes, with adjustment for potential confounding factors, including age, sex, BMI, duration of prediabetes, and baseline values of respective outcome measures.

Both intention-to-treat (ITT) and per-protocol analyses were conducted. For the ITT analysis, missing data were handled using the last observation carried forward method. Sensitivity analyses were performed to assess the robustness of the results under different missing data scenarios.

RESULTS

Participant flow and baseline characteristics

From a total of 168 individuals screened for eligibility, 120 participants meeting the inclusion criteria were enrolled in the study and allocated to three intervention groups (40 participants per group). During the 12-month follow-up period, 11 (9.2%) participants discontinued the intervention (4 in the low-frequency group, 3 in the medium-frequency group, and 4 in the high-frequency group). Reasons for discontinuation included relocation (n = 4), personal reasons unrelated to the intervention (n = 5), and development of type 2 diabetes requiring pharmacological treatment (n = 2). The final analysis included 109 participants who completed the 12-month intervention.

The baseline demographic, clinical, and psychological characteristics of the participants are presented in Table 1. There were no significant differences between groups in terms of age, sex distribution, BMI, duration of prediabetes, glycemic parameters, insulin parameters, oxidative stress markers, inflammatory markers, or psychological parameters at baseline, indicating successful randomization and comparable groups at study initiation.

Table 1 Baseline characteristics of study participants.

Characteristic	Low-frequency group (n = 40)	Medium-frequency group (n = 40)	High-frequency group (n = 40)	P value
Demographic and clinical characteristics
Age (years)	52.4 ± 8.7	53.1 ± 9.2	51.8 ± 8.5	0.814
Sex (male/female)	18/22	19/21	17/23	0.925
Body mass index (kg/m²)	26.8 ± 3.6	27.1 ± 3.9	26.5 ± 3.4	0.763
Waist circumference (cm)	92.4 ± 10.2	93.1 ± 11.5	91.8 ± 9.8	0.842
Duration of prediabetes (months)	18.7 ± 10.3	19.5 ± 11.2	17.9 ± 9.5	0.789
Glycemic parameters
Fasting plasma glucose (mmol/L)	6.38 ± 0.42	6.41 ± 0.39	6.35 ± 0.41	0.810
2-hour postprandial glucose (mmol/L)	9.23 ± 1.15	9.17 ± 1.09	9.31 ± 1.18	0.847
Glycosylated hemoglobin (%)	6.14 ± 0.32	6.17 ± 0.29	6.11 ± 0.31	0.688
Insulin parameters
Fasting insulin (μIU/mL)	14.76 ± 5.21	14.92 ± 5.38	14.58 ± 5.12	0.957
2-hour postprandial insulin (μIU/mL)	78.45 ± 25.73	79.18 ± 26.41	77.69 ± 24.93	0.968
Homeostasis model assessment of insulin resistance	4.17 ± 1.48	4.23 ± 1.52	4.09 ± 1.43	0.910
Oxidative stress markers
Cysteine-C (mg/L)	0.92 ± 0.18	0.93 ± 0.19	0.91 ± 0.17	0.873
Beta-2-microglobulin (mg/L)	1.98 ± 0.41	2.01 ± 0.44	1.95 ± 0.39	0.804
Malondialdehyde (nmol/mL)	5.27 ± 1.23	5.34 ± 1.29	5.21 ± 1.18	0.886
Total antioxidant capacity (U/mL)	9.84 ± 2.63	9.78 ± 2.59	9.91 ± 2.68	0.975
Inflammatory markers
High-sensitivity C-reactive protein (mg/L)	3.14 ± 1.52	3.19 ± 1.57	3.08 ± 1.48	0.947
Tumor necrosis factor-alpha (pg/mL)	12.67 ± 3.98	12.78 ± 4.05	12.51 ± 3.89	0.953
Interleukin-6 (pg/mL)	3.28 ± 1.21	3.34 ± 1.25	3.22 ± 1.17	0.901
Psychological parameters
Beck Depression Inventory-II score	15.82 ± 5.73	16.04 ± 5.91	15.69 ± 5.62	0.959
STAI-State score	41.35 ± 8.67	41.73 ± 8.94	40.98 ± 8.42	0.928
STAI-Trait score	43.19 ± 9.25	43.58 ± 9.44	42.87 ± 9.13	0.952
Perceived Stress Scale score	19.85 ± 5.37	20.12 ± 5.49	19.67 ± 5.28	0.924

Values are presented as mean ± SD or frequencies. STAI: State-Trait Anxiety Inventory.

Changes in psychological parameters

All three groups showed significant improvements in psychological parameters over the 12-month follow-up period, with varying degrees of improvement according to vibration frequency (Table 2). The high-frequency group demonstrated the most substantial reductions in depression, anxiety, and stress scores, followed by the medium-frequency group and then the low-frequency group.

Table 2 Changes in psychological parameters from baseline to 12 months.

Parameter	Group	Baseline	3 months	6 months	12 months	P value (within group)	P value (between groups)
Beck Depression Inventory-II	Low-frequency	15.82 ± 5.73	13.54 ± 5.12	10.98 ± 4.58	9.27 ± 4.11	< 0.001	< 0.001
	Medium-frequency	16.04 ± 5.91	12.87 ± 4.98	9.73 ± 4.21	7.82 ± 3.76	< 0.001
	High-frequency	15.69 ± 5.62	11.25 ± 4.65	8.17 ± 3.82	6.46 ± 3.42	< 0.001
STAI-State	Low-frequency	41.35 ± 8.67	37.62 ± 7.95	34.18 ± 7.32	31.94 ± 6.87	< 0.001	< 0.001
	Medium-frequency	41.73 ± 8.94	36.41 ± 7.73	32.29 ± 6.95	29.53 ± 6.41	< 0.001
	High-frequency	40.98 ± 8.42	34.75 ± 7.38	30.12 ± 6.51	27.24 ± 6.03	< 0.001
STAI-Trait	Low-frequency	43.19 ± 9.25	40.57 ± 8.81	37.84 ± 8.23	35.92 ± 7.94	< 0.001	< 0.001
	Medium-frequency	43.58 ± 9.44	39.82 ± 8.67	36.31 ± 7.89	33.75 ± 7.41	< 0.001
	High-frequency	42.87 ± 9.13	38.14 ± 8.32	34.52 ± 7.65	31.38 ± 7.12	< 0.001
Perceived Stress Scale	Low-frequency	19.85 ± 5.37	17.43 ± 4.98	14.89 ± 4.51	13.25 ± 4.18	< 0.001	< 0.001
	Medium-frequency	20.12 ± 5.49	16.78 ± 4.85	13.61 ± 4.32	11.84 ± 3.96	< 0.001
	High-frequency	19.67 ± 5.28	15.41 ± 4.63	11.92 ± 4.05	9.28 ± 3.72	< 0.001

Values are presented as mean ± SD. P value (within group): Repeated measures analysis of variance (ANOVA) comparing changes over time within each group. P value (between groups): One-way ANOVA comparing changes from baseline to 12 months between groups. STAI: State-Trait Anxiety Inventory.

Repeated measures ANOVA revealed significant time × group interactions for all psychological parameters (P < 0.001), indicating that the rate and magnitude of improvement differed according to vibration frequency. Post-hoc analyses showed that the high-frequency group had significantly greater reductions in depression (BDI-II), anxiety (STAI-State and STAI-Trait), and stress (PSS) scores compared to both the medium-frequency and low-frequency groups at all time points after baseline (all P < 0.05). The medium-frequency group also showed significantly greater reductions compared to the low-frequency group (all P < 0.001).

The mean reduction in BDI-II scores from baseline to 12 months was 6.55 ± 1.87 points in the low-frequency group, 8.22 ± 2.14 points in the medium-frequency group, and 9.23 ± 2.31 points in the high-frequency group. Similarly, the mean reduction in PSS scores was 6.60 ± 1.92 points in the low-frequency group, 8.28 ± 2.18 points in the medium-frequency group, and 10.39 ± 2.52 points in the high-frequency group. Table 2 demonstrates the progressive improvements in all psychological parameters across the 12-month study period. The most notable finding was the dose-dependent effect of vibration frequency, with the high-frequency group achieving approximately 60% reduction in depression scores, 35% reduction in anxiety scores, and 53% reduction in perceived stress scores from baseline to 12 months. These improvements were not only statistically significant but also clinically meaningful, with most participants in the high-frequency group moving from moderate to mild or minimal symptom categories.

Changes in metabolic parameters

Parallel to improvements in psychological parameters, all three groups demonstrated significant improvements in glycemic and insulin parameters over the 12-month follow-up period (Table 3). Consistent with psychological outcomes, the high-frequency group showed the most substantial improvements in metabolic parameters.

Table 3 Changes in glycemic and insulin parameters from baseline to 12 months.

Parameter	Group	Baseline	3 months	6 months	12 months	P value (within group)	P value (between groups)
Fasting plasma glucose (mmol/L)	Low-frequency	6.38 ± 0.42	6.19 ± 0.39	5.97 ± 0.36	5.82 ± 0.35	< 0.001	< 0.001
	Medium-frequency	6.41 ± 0.39	6.12 ± 0.37	5.84 ± 0.34	5.65 ± 0.32	< 0.001
	High-frequency	6.35 ± 0.41	5.98 ± 0.35	5.67 ± 0.31	5.43 ± 0.29	< 0.001
2-hour postprandial glucose (mmol/L)	Low-frequency	9.23 ± 1.15	8.87 ± 1.08	8.42 ± 0.98	8.12 ± 0.93	< 0.001	< 0.001
	Medium-frequency	9.17 ± 1.09	8.65 ± 1.02	8.09 ± 0.92	7.69 ± 0.86	< 0.001
	High-frequency	9.31 ± 1.18	8.43 ± 0.97	7.76 ± 0.85	7.44 ± 0.81	< 0.001
Glycosylated hemoglobin (%)	Low-frequency	6.14 ± 0.32	6.03 ± 0.30	5.89 ± 0.28	5.81 ± 0.27	< 0.001	< 0.001
	Medium-frequency	6.17 ± 0.29	5.98 ± 0.28	5.79 ± 0.26	5.69 ± 0.25	< 0.001
	High-frequency	6.11 ± 0.31	5.87 ± 0.27	5.67 ± 0.24	5.60 ± 0.23	< 0.001
Fasting insulin (μIU/mL)	Low-frequency	14.76 ± 5.21	13.92 ± 4.95	12.87 ± 4.63	12.15 ± 4.38	< 0.001	< 0.001
	Medium-frequency	14.92 ± 5.38	13.61 ± 4.87	12.32 ± 4.45	11.47 ± 4.12	< 0.001
	High-frequency	14.58 ± 5.12	12.93 ± 4.62	11.51 ± 4.18	10.68 ± 3.87	< 0.001
2-hour postprandial insulin (μIU/mL)	Low-frequency	78.45 ± 25.73	73.21 ± 24.15	67.14 ± 22.26	63.54 ± 21.08	< 0.001	< 0.001
	Medium-frequency	79.18 ± 26.41	71.47 ± 23.84	64.25 ± 21.58	59.46 ± 19.87	< 0.001
	High-frequency	77.69 ± 24.93	68.35 ± 22.31	60.42 ± 19.75	55.13 ± 18.02	< 0.001
Homeostasis model assessment of insulin resistance	Low-frequency	4.17 ± 1.48	3.82 ± 1.36	3.41 ± 1.22	3.14 ± 1.12	< 0.001	< 0.001
	Medium-frequency	4.23 ± 1.52	3.68 ± 1.33	3.19 ± 1.15	2.87 ± 1.03	< 0.001
	High-frequency	4.09 ± 1.43	3.42 ± 1.24	2.89 ± 1.05	2.57 ± 0.93	< 0.001

Values are presented as mean ± SD. P value (within group): Repeated measures analysis of variance (ANOVA) comparing changes over time within each group. P value (between groups): One-way ANOVA comparing changes from baseline to 12 months between groups.

The mean reduction in FPG levels from baseline to 12 months was 0.56 ± 0.13 mmol/L in the low-frequency group, 0.76 ± 0.15 mmol/L in the medium-frequency group, and 0.92 ± 0.17 mmol/L in the high-frequency group. Similarly, the mean reduction in HOMA-IR was 1.03 ± 0.38 in the low-frequency group, 1.36 ± 0.47 in the medium-frequency group, and 1.52 ± 0.53 in the high-frequency group.

Repeated measures ANOVA revealed significant time × group interactions for all glycemic and insulin parameters (P < 0.001), indicating that the rate and magnitude of improvement differed according to vibration frequency. Post-hoc analyses showed that the high-frequency group had significantly greater improvements compared to both the medium-frequency and low-frequency groups at all time points after baseline (all P < 0.05). The medium-frequency group also showed significantly greater improvements compared to the low-frequency group (all P < 0.05).

Changes in oxidative stress and inflammatory markers

Significant improvements in oxidative stress and inflammatory markers were observed in all three groups, with the high-frequency group demonstrating the most substantial improvements (Table 4).

Table 4 Changes in oxidative stress and inflammatory markers from baseline to 12 months.

Parameter	Group	Baseline	6 months	12 months	P value (within group)	P value (between groups)
Cysteine-C (mg/L)	Low-frequency	0.92 ± 0.18	0.86 ± 0.16	0.81 ± 0.15	< 0.001	< 0.001
	Medium-frequency	0.93 ± 0.19	0.83 ± 0.15	0.76 ± 0.14	< 0.001
	High-frequency	0.91 ± 0.17	0.78 ± 0.14	0.70 ± 0.12	< 0.001
Beta-2-microglobulin (mg/L)	Low-frequency	1.98 ± 0.41	1.85 ± 0.38	1.76 ± 0.35	< 0.001	< 0.001
	Medium-frequency	2.01 ± 0.44	1.82 ± 0.36	1.69 ± 0.33	< 0.001
	High-frequency	1.95 ± 0.39	1.73 ± 0.34	1.58 ± 0.30	< 0.001
Malondialdehyde (nmol/mL)	Low-frequency	5.27 ± 1.23	4.82 ± 1.15	4.53 ± 1.07	< 0.001	< 0.001
	Medium-frequency	5.34 ± 1.29	4.67 ± 1.12	4.28 ± 1.03	< 0.001
	High-frequency	5.21 ± 1.18	4.38 ± 1.05	3.92 ± 0.94	< 0.001
Total antioxidant capacity (U/mL)	Low-frequency	9.84 ± 2.63	10.95 ± 2.84	11.68 ± 3.02	< 0.001	< 0.001
	Medium-frequency	9.78 ± 2.59	11.36 ± 2.92	12.47 ± 3.19	< 0.001
	High-frequency	9.91 ± 2.68	12.05 ± 3.13	13.85 ± 3.48	< 0.001
High-sensitivity C-reactive protein (mg/L)	Low-frequency	3.14 ± 1.52	2.78 ± 1.36	2.53 ± 1.25	< 0.001	< 0.001
	Medium-frequency	3.19 ± 1.57	2.65 ± 1.31	2.31 ± 1.18	< 0.001
	High-frequency	3.08 ± 1.48	2.42 ± 1.23	2.05 ± 1.09	< 0.001
Tumor necrosis factor-alpha (pg/mL)	Low-frequency	12.67 ± 3.98	11.34 ± 3.62	10.52 ± 3.41	< 0.001	< 0.001
	Medium-frequency	12.78 ± 4.05	10.89 ± 3.53	9.84 ± 3.26	< 0.001
	High-frequency	12.51 ± 3.89	10.12 ± 3.31	8.95 ± 3.05	< 0.001
Interleukin-6 (pg/mL)	Low-frequency	3.28 ± 1.21	2.87 ± 1.08	2.61 ± 0.98	< 0.001	< 0.001
	Medium-frequency	3.34 ± 1.25	2.75 ± 1.04	2.42 ± 0.92	< 0.001
	High-frequency	3.22 ± 1.17	2.54 ± 0.96	2.18 ± 0.83	< 0.001

Values are presented as mean ± SD. P value (within group): Repeated measures analysis of variance (ANOVA) comparing changes over time within each group. P value (between groups): One-way ANOVA comparing changes from baseline to 12 months between groups.

The improvements in oxidative stress markers were characterized by reductions in Cys-C, β2-MG, and MDA levels, along with increases in T-AOC. Similarly, inflammatory marker (hs-CRP, TNF-α, and IL-6) levels showed significant reductions in all groups. Consistent with the pattern observed for psychological and glycemic parameters, the high-frequency group demonstrated the most substantial improvements in oxidative stress and inflammatory marker values, followed by the medium-frequency group and then the low-frequency group.

Correlations between psychological and metabolic parameters

Pearson correlation analyses revealed significant associations between changes in psychological parameters and changes in metabolic parameters (Table 5). Reductions in depression, anxiety, and stress scores were significantly correlated with improvements in glycemic control, insulin sensitivity, oxidative stress marker values, and inflammatory marker values.

Table 5 Correlations between changes in psychological and metabolic parameters from baseline to 12 months.

Parameter	Change in Beck Depression Inventory-II	Change in STAI-State	Change in STAI-Trait	Change in Perceived Stress Scale
Change in fasting plasma glucose	0.52	0.49	0.47	0.55
Change in 2-hour postprandial glucose	0.57	0.54	0.52	0.58
Change in glycosylated hemoglobin	0.49	0.46	0.44	0.51
Change in fasting insulin	0.54	0.52	0.49	0.56
Change in 2-hour postprandial insulin	0.58	0.56	0.53	0.59
Change in homeostasis model assessment of insulin resistance	0.63	0.61	0.57	0.65
Change in cysteine-C	0.48	0.45	0.43	0.50
Change in beta-2-microglobulin	0.47	0.45	0.42	0.49
Change in malondialdehyde	0.51	0.48	0.46	0.53
Change in total antioxidant capacity	-0.53	-0.50	-0.47	-0.54
Change in high-sensitivity C-reactive protein	0.62	0.59	0.56	0.64
Change in tumor necrosis factor-alpha	0.60	0.58	0.55	0.62
Change in interleukin-6	0.59	0.57	0.54	0.61

Values represent Pearson correlation coefficients (r). STAI: State-Trait Anxiety Inventory.

The strongest correlations were observed between changes in HOMA-IR and changes in psychological parameters, particularly PSS scores (r = 0.65, P < 0.001) and BDI-II scores (r = 0.63, P < 0.001). Similarly, changes in inflammatory markers, particularly hs-CRP levels, were strongly correlated with changes in psychological parameters. These findings suggest a bidirectional relationship between psychological distress and metabolic dysfunction, with improvements in one domain potentially facilitating improvements in the other.

Safety and adverse events

No serious adverse events related to the intervention were reported during the study period. Minor adverse events included temporary mild discomfort during initial vibration sessions (n = 7), which resolved spontaneously within the first few weeks of the intervention. There were no significant changes in routine blood tests, liver function tests, kidney function tests, or ECG parameters in any of the groups, indicating the safety of abdominal vibration technique across different frequencies.

DISCUSSION

This prospective cohort study is the first to comprehensively evaluate the effects of different frequencies of the abdominal vibration technique in Tuina on the psychological well-being and metabolic parameters in patients with prediabetes. Our findings demonstrate that Tuina’s abdominal vibration technique is an effective complementary therapeutic approach for reducing depression, anxiety, and stress in this population, with concurrent improvements in glycemic control, insulin sensitivity, oxidative stress, and inflammatory markers. Importantly, the study reveals a clear frequency-dependent effect, with high-frequency abdominal vibration (600 times/minute) yielding superior outcomes compared to medium-frequency (500 times/minute) and low-frequency (400 times/minute) interventions. The substantial reductions in depression, anxiety, and stress scores observed in our study align with previous studies. However, the present study extends these findings to individuals with prediabetes, a population for whom psychological distress is both prevalent and prognostically significant. The magnitude of improvement in psychological parameters (approximately 40%-50% reduction from baseline scores in the high-frequency group), suggests that abdominal vibration may be particularly effective for addressing psychological distress in prediabetes.

Several mechanisms may explain the psychological benefits observed in our study. First, vibration therapy has been associated with reduced cortisol levels and increased production of mood-enhancing neurotransmitters such as serotonin and dopamine[15]. Second, the physical touch involved in Tuina therapy may activate oxytocin release, which has anxiolytic effects[16]. Third, the relaxation response induced by abdominal vibration may counteract the stress response, leading to reduced sympathetic nervous system activity[17]. Finally, the improvements in glycemic control and reduction in IR observed in our study may have contributed to enhanced psychological well-being, given the established bidirectional relationship between metabolic dysfunction and psychological distress[18].

The frequency-dependent effect observed in our study, with high-frequency vibration yielding superior outcomes, is a novel finding with important clinical implications. The differential effects of vibration frequency may be explained by varying impacts on tissue perfusion, metabolic activity, and neural signaling. Higher-frequency manipulations may enhance blood flow to abdominal organs, including the pancreas, thereby improving insulin secretion and action. Additionally, higher-frequency stimulation may activate mechanoreceptors and proprioceptors more effectively, potentially enhancing afferent signaling to the central nervous system regions involved in mood regulation.

The significant improvements in glycemic control and insulin sensitivity observed in our study align with the finding of previous research on the metabolic effects of the abdominal vibration technique[10-12]. However, our findings provide more detailed insights into the specific mechanisms involved. The reductions in oxidative stress marker (Cys-C, β2-MG, MDA) levels and increases in anti-oxidant capacity (T-AOC) suggest that abdominal vibration may enhance cellular defense against oxidative damage, a key pathophysiological feature of prediabetes. Similarly, the reductions in inflammatory marker (hs-CRP, TNF-α, IL-6) levels indicate anti-inflammatory effects, which may contribute to improved insulin signaling and glucose metabolism.

The strong correlations observed between changes in psychological parameters and changes in metabolic parameters highlight the interconnectedness of psychological and metabolic health in prediabetes. These findings support the concept of a "psychology-metabolism axis", wherein improvements in psychological well-being may facilitate metabolic improvements and vice versa. The particularly strong correlations between changes in HOMA-IR and changes in psychological parameters suggest that IR may be a key mediator in this bidirectional relationship[19]. Recent research has demonstrated multiple mechanisms linking inflammatory processes to IR, providing a potential pathway through which psychological stress may affect metabolic function in prediabetes patients.

From a practical perspective, our findings have several important clinical implications. First, they provide evidence for including Tuina’s abdominal vibration technique as a complementary therapeutic approach in comprehensive prediabetes management programs. Second, they suggest that high-frequency abdominal vibration should be preferred for optimal outcomes. Third, they highlight the importance of addressing both psychological and metabolic aspects of prediabetes in an integrated manner[20]. Understanding the role of oxidative stress regulation pathways, such as the nuclear factor-E2-related factor 2 system, may provide additional therapeutic targets for comprehensive management approaches that address both the psychological and metabolic dimensions of prediabetic conditions.

Our study demonstrates that regular high-frequency abdominal vibration therapy can effectively improve patients' clinical indicators and psychological scores. However, it is important to note that the hospital-based treatment process is cumbersome and inconvenient for many patients. Additionally, the vibration frequency of 600 times/minute is technically challenging for practitioners to consistently maintain and difficult to implement widely in clinical practice. These limitations provide a clear foundation and need for the development of portable or home-use abdominal vibration treatment devices that could deliver consistent, high-frequency vibrations while improving accessibility and compliance for patients with prediabetes.

The present study presents several important innovations in prediabetes management research. First, this is the first study to systematically evaluate different frequencies of abdominal vibration, revealing a clear frequency-dependent therapeutic effect. Second, we demonstrate for the first time that non-pharmacological manual therapy can simultaneously improve both psychological and metabolic parameters in prediabetes, with effect sizes comparable to or exceeding those of conventional interventions. Third, the strong correlations between psychological and metabolic improvements provide novel evidence for the bidirectional psychology-metabolism axis in prediabetes, suggesting that targeting one domain can facilitate improvements in the other. Finally, the sustained benefits observed at 12-month follow-up indicate that relatively brief interventions can produce long-lasting effects, challenging the current paradigm of continuous long-term treatment.

However, some limitations should be acknowledged. First, as a non-randomized cohort study, causal inferences must be made with caution. Second, the absence of a non-intervention control group limits our ability to definitively attribute the observed improvements to the vibration intervention rather than to other factors such as regression to the mean or placebo effects. Third, while we assessed a wide range of metabolic parameters, we did not directly measure pancreatic β-cell function or tissue-specific insulin sensitivity, which would have provided more detailed insights into the metabolic effects of abdominal vibration technique. Fourth, the study population was from a single center, potentially limiting the generalizability of our findings to other populations and settings.

CONCLUSION

The abdominal vibration technique in Tuina is an effective complementary therapeutic approach for reducing depression, anxiety, and stress in patients with prediabetes, with concurrent improvements in glycemic control, insulin sensitivity, oxidative stress, and inflammatory markers. The beneficial effects are frequency-dependent, with high-frequency vibration (600 times/minute) yielding superior outcomes compared to medium-frequency and low-frequency interventions. The strong correlations between psychological and metabolic improvements highlight the interconnectedness of psychological and metabolic health in prediabetes and underscore the value of integrated approaches to prediabetes management. Given its efficacy, simplicity, safety, and patient acceptability, abdominal vibration technique, particularly at higher frequencies, merits consideration as a component of comprehensive prediabetes management programs.