Effect of slow deep breathing on cardiovascular autonomic neuropathy in type 2 diabetes mellitus patients

Abstract

Objective

This study aimed to investigate the effects of slow deep breathing (SDB) on heart rate variability (HRV) in patients with type 2 diabetes mellitus (T2DM) patients with cardiovascular autonomic neuropathy (CAN).

Methods

Sixty patients with CAN secondary to type T2DM were randomly assigned to SDB (n = 30) and control (n = 30) groups. All patients were treated with mecobalamin (0.5 mg; times daily). Patients in the SDB group underwent slow deep breathing training against a background of soothing music. Before treatment and 3 months after treatment, height, body weight, heart rate (HR), blood pressure (BP), HbA1c, LDL, TG, heart-rate-variability (HRV), and hart rate recovery (HRR) were measured 1 min after the 6-min walk test in the two groups.

Results

BMI, SBP, DBP, HbA1c, LDL, and TG levels were comparable between the two groups and remained unchanged after treatment in two groups (P > 0.05). In the SDB group, the standard deviation of N-N intervals (SDNN), the square root of the mean squared differences of successive RR (RMSSD), low-frequency power (LF), high-frequency power (HF), and LF/HF ratio changed significantly after treatment (P < 0.05), and these measurements remained statistically significant after adjusting for age (P < 0.05), but HR and the root mean square of the difference between adjacent R-R intervals (PNN50) remained unchanged (P > 0.05). In the control group, HR, SDNN, RMSSD, PNN50, LF, HF, and LF/HF ratio remained unchanged after treatment (P > 0.05). The SBP, SDNN, RMSSD, LF, HF, and LF/HF ratios were significantly different between the two groups after treatment.(P < 0.05), and these measurements remained statistically significant after adjusting for age (P < 0.05). In addition, the HRR1 levels were comparable between the two groups (P > 0.05).

Conclusion

With soothing music as the background, sustained slow deep breathing training can improve heart rate variability in patients with type 2 diabetes mellitus and cardiac autonomic neuropathy, independent of age, providing a new approach for the treatment of cardiac autonomic neuropathy in type 2 diabetes.

Clinical trial number

The study was registered in the Chinese Clinical Trial Registry (ChiCTR2200063260).

Introduction

Cardiovascular autonomic neuropathy (CAN) is a common chronic complication of diabetes mellitus (DM). The incidence of CAN is as high as 65% in patients with type 2 DM (T2DM) patients [1]. The clinical manifestations of CAN include sinus tachycardia, exercise intolerance, and postural hypotension others [2]. CAN can be found in the early stages of DM and is associated with an elevated risk of arrhythmia, asymptomatic myocardial ischemia, cardiac dysfunction, and sudden death [2, 3]. It has been confirmed that CAN is an important cause of increased mortality in DM patients [4]. The incidence of CAN increases with the prolongation of the disease course, increase in age, and elevation of urine protein levels. Thus, early diagnosis and treatment of CAN are crucial to prevent cardiovascular events (such as myocardial infarction and cardiogenic sudden death) and reduce mortality in patients with DM.

Heart rate variability (HRV) is a common noninvasive indicator for assessing autonomic nervous system function of the heart (sympathetic and vagus nerve activities) and has been used as an important parameter for the early prediction and diagnosis of CAN [5]. Studies have indicated that HRV is closely related to heart failure [6], death from all causes [7], and the prognosis of cancer patients [8]. Low HRV is also associated with an elevated risk of sudden coronary artery-related sudden death [9]. HRV has been widely used to investigate the effects of autonomous slow breathing on the regulation of cardiac function by the vagus nerve [10].

Regular physical activities (such as endurance and strength training) can increase the impact of the vagus nerve on the heart relative to the sympathetic nerve, thereby increasing HRV [11], which may exert protective effects on the heart and reduce cardiovascular risk [12]. Studies have found that slow deep breathing (SDB) can increase vagus nerve activity [13, 14], lower the heart rate (HR) [15], and alter HRV. SDB can also alter HRV and reduce mortality in patients with myocardial infarction and coronary heart disease [16, 17]. Therefore, SDB training is a non-invasive method for activating the parasympathetic nervous system.

Currently, studies on SDB mainly focus on patients with cardiovascular disease, neuropsychiatric disease, tumors, or healthy individuals, and little is known about the influence of SDB in T2DM patients. In the present study, T2DM patients with CAN who received glucose, lipid, blood pressure, and neurotrophic therapies were included. In addition, these patients were randomly divided into the SDB and control groups. In the SDB group, patients received SDB training with soothing music as the background for three months, and the parameters related to cardiac autonomic nervous function were assessed before and after SDB training, with the aim of exploring the impact of SDB on the cardiac autonomic nervous system in T2DM patients, which may provide a new direction for the treatment of CAN in T2DM patients.

Materials and methods

Study design

This was a single-center randomized controlled prospective clinical study. Patients were randomly divided into SDB and control groups at a ratio of 1:1 using a random number table.

Patients

Sample size: Patients in the SDB group underwent SDB training, while patients in the control group underwent simple chest expansion exercises. The outcome was assessed using the square root of the mean squared differences of successive RR (RMSSD) after training. The pilot study indicated the RMSSD was 21.54 ± 3.21 in the SDB group and 18.45 ± 4.17 in the control group. Assuming a bilateral a = 0.05 and a confidence level of 80%, the estimated sample size was 24 in each group. Considering the dropout rate of 20%, at least 30 patients were included in each group. Therefore, 60 patients were recruited in this study. T2DM patients with CAN were recruited between October 2022 and February 2024 at the Tongji Hospital, Tongji University.

Inclusion criteria

(1) The age ranged from 40 years to 70 years, and both males and females were included; (2) DM was diagnosed according to WHO Diagnostic Criteria for Diabetes Mellitus in 1999; patients were diagnosed with T2DM at least 1 year ago; (3) CAN was diagnosed according to the Expert Consensus on the Diagnosis and Treatment of Diabetic Neuropathy [18]; (4) The HbA1c was 6–8%, and the body mass index ranged from 18.5 kg/m² to 28 kg/m²; (5) Patients received baseline treatments (such as conventional glucose-lowering, blood pressure lowering, and lipid-lowering therapies) for more than 3 months; (6) The informed consent was obtained before study.

Exclusion criteria

1)There were obvious central nervous system diseases; there was peripheral nerve injury caused by medication, alcohol abuse, or severe liver and kidney diseases; 2) patients were administered medication that may affect the autonomic nervous system and heart rate in the past week; 3) patients had cognitive impairment and could not cooperate with training; 4) patients had severe organ dysfunction (such as heart, liver, and kidney), severe respiratory failure, or severe hyperlipidemia; 5) endocrine disorders (such as hyperthyroidism, hypothyroidism, hyperkalemia, or hypokalemia); 6) patients had organic heart diseases (such as coronary heart disease, hypertrophic cardiomyopathy, rheumatic heart disease, myocarditis, etc.); 7) hypertension was not controlled even after treatment with antihypertensive drugs for 4 weeks (systolic blood pressure ≥ 150 mmHg and diastolic blood pressure ≥ 95 mmHg); 8) there was any malignant tumor (cured or not); and 9) a history of drug abuse.

Criteria for exclusion, drop-out and termination of observation

(1) Patients could not participate in this study due to severe adverse effects, (2) patients developed severe complications during the study period, (3) patients had poor compliance to the treatment, and (4) patients actively withdrew from the trial.

All the subjects were treated for the routine risk factors of CAN (such as glucose-lowering, lipid-lowering, blood pressure lowering, etc.) and with mecobalamin (0.5 mg, thrice daily; Wei Cai [China] Pharmaceutical Co., Ltd.). The subjects underwent assessment and training at the Rehabilitation Sports Center of the Tongji Hospital. This study was approved by the Ethics Committee of Tongji Hospital, Tongji University (2021-LCYJ-009), and was registered in the Chinese Clinical Trial Registry (ChiCTR2200063260). Signed informed consent was obtained from each patient.

Heart rate recovery assessment

6-Minute Walk Test (6 MWT) was conducted according to the 2002 guidelines of the American Pulmonary Association [19]. The test was conducted in a quiet and ventilated corridor, 30 m in length, with a hard ground. The corridor was marked orange every 3 m and orange traffic cones were placed at the turning point. The patients underwent finger oxygen saturation detection, 12-channel portable electrocardiogram monitoring, and blood pressure monitoring before exercise. Emergency management, such as nitroglycerin, oxygen, blood pressure monitoring, and defibrillators, was immediately available. The subjects walked as quickly as possible in a straight line, turned around, and continued walking until both ends were reached. The investigator reminded patients at the end of each minute. At the end of 6th minute, the investigator reminded them: “Time is up.” The participants immediately stopped walking and sat at rest. The electrocardiogram was monitored while walking and the heart rate per minute was recorded. After the test, the HR) and finger oxygen saturation were detected immediately, clinical symptoms were observed, and the heart rate reserve in the first minute (HRR1) was recorded.

HRV assessment

At rest, the electrocardiogram was recorded with a 24-hour dynamic electrocardiogram recorder (CT-08, Hangzhou Baihui Medical Equipment Co., Ltd., China). The R-wave was employed to analyze the time- and frequency-domain variables of HRV. HRV analysis software (V1.2.0, Hangzhou Baihui Medical Equipment Co., Ltd., China) was used for automatic analysis of HRV. HRV indicators can effectively evaluate the autonomic nervous system function of the heart. Time domain variables, such as the root mean square of the difference between adjacent R-R intervals (RMSSD) and the number of adjacent R-R intervals with a difference greater than 50 ms (pNN50), can reflect parasympathetic regulation. The standard deviation of the N-N interval (SDNN) reflects the overall variability of HR and cannot distinguish whether changes are related to the elimination of vagal nerve tension or an increase in sympathetic nerve tension. As a frequency-domain indicator, low-frequency power (LF) represents the combined effect of the sympathetic and parasympathetic nerves on the heart, high-frequency power (HF) is an indicator of vagus nerve activity, and the LF/HF ratio represents the balance between the sympathetic and vagus nerves. All HRV variables provide information on the relative effects of sympathetic and parasympathetic nervous tensions on the heart [20].

A 24-h electrocardiogram (ECG) was recorded, and HRV was analyzed. Participants were advised to avoid caffeinated substances (including tea, coffee, and cola drinks) within 12 h prior to the assessment, as well as to avoid alcoholic beverages and smoking within at least 24 h before assessment. The patients were instructed to avoid any additional physical activity (activities beyond normal lifestyle). They discontinued taking β-blockers within 48 h prior to the HRV assessment. The HRR and HRV tests were independently completed by two professionals, who explained the procedures and precautions of each test to every patient before the test.

Training protocols

Before and after training, height, body weight, HR, blood pressure, blood glucose, blood lipids, HRV, and heart rate recovery in the first minute (HRR1) after the 6-minute walk test. After the baseline assessment, all subjects received treatment with mecobalamin (0.5 mg) three times daily for 3 months and were randomly divided into the SDB and control groups. In the SDB group, patients underwent slow and rhythmic deep breathing training for 20 min once daily for a total of 3 months. Rhythmic SDB training was performed using soothing music as the background, and deep exhalation and inhalation were performed under voice guidance. Soothing music can keep subjects in a relaxed state and voice guidance can establish auditory respiratory biofeedback. The respiratory rate (6 breaths/min) and inhalation exhalation ratio (1:2) were maintained consistently. In the control group, patients underwent simple chest expansion training for 20 min once daily for a total of 3 months. Before the initial training, the patients attended two courses to familiarize themselves with SDB. Compliance with the interventions was also evaluated. Follow-up was conducted by telephone once weekly to remind participants to adhere to the training, and follow-up was performed in the clinics once monthly to confirm that the respiratory training procedures were correct in these patients.

Biochemical examinations

Basic clinical data and lifestyle information of the study population were collected, including age, sex, height, and body weight. Blood biochemical parameters, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), serum creatinine (Scr), triglyceride (TG) and low-density lipoprotein cholesterol (LDL-C) were detected on an automatic biochemical analyzer (AU 5800, Beckman Coulter, USA). glycosylated hemoglobin (HbA1c) was detected by high-performance liquid chromatography (HLC-723G8, TOSOH CORPORATION, Japan). Statistical analysis.

Data analyses were conducted using SPSS version 23.0. All variables were subjected to the Shapiro–Wilk test for normality of distribution. For variables with a normal distribution, an independent sample t-test was used to compare the demographic characteristics and parameters between the two groups before and after treatment. A paired-sample t-test was used to compare parameters before and after treatment in the same group. For variables without a normal distribution, nonparametric tests (such as the Wilcoxon test) were used to compare the parameters before and after treatment between the two groups. Non parametric tests (such as the Mann–Whitney U test) were used to compare parameters before and after treatment in the same group. All blood routine parameters, were measured using an automated hematology analyzer (Cobas 8000; Roche, Switzerland). The measured parameters included white blood cell (WBC) count, red blood cell (RBC) count and platelet (PLT) count.

Statistical analysis

Data analyses were conducted using SPSS version 23.0. All variables were subjected to the Shapiro–Wilk test for normality of distribution. For variables with a normal distribution, an independent sample t-test was used to compare the demographic characteristics and parameters between the two groups before and after treatment. A paired-sample t-test was used to compare parameters before and after treatment in the same group. For variables without a normal distribution, nonparametric tests (such as the Wilcoxon test) were used to compare the parameters before and after treatment between the two groups. Non parametric tests (such as the Mann–Whitney U test) were used to compare parameters before and after treatment in the same group. Analysis of covariance (ANCOVA) was used to compare the changes in each measurement after adjusting for age. Statistical significance was set at P < 0.05.

Results

Sixty patients were included in this study. In the control group, two patients withdrew from the study because of perceived poor efficacy. Ultimately, 30 and 28 patients were included in the SDB and control groups, respectively. The baseline demographic characteristics, blood pressure, blood glucose, blood lipid, Liver and kidney functions levels were compared between the two groups. The results showed no significant differences in BMI, SBP, DBP, HbA1c, LDL, and TG between the two groups at baseline (P > 0.05). After respiratory training for 3 months, there were no marked differences in BMI, DBP, HbA1c, LDL, or TG between the two groups (P > 0.05), but significant differences were noted in SBP (P < 0.05). BMI, SBP, DBP, HbA1c, ALT、AST、Scr, LDL, and TG levels were compared before and after treatment in each group, and the results showed that these parameters remained unchanged in the two groups after three months of training (P > 0.05) (Tables 1 and 2).

Table 1.

Baseline clinical characteristics of patients in tow groups

Group	N	Sex (male/female)	Age (years)	Diabetes Duration (years)	eGFR (mL/min/1.73m2)	ALT (U/L)	AST (U/L)	RBC х1012/L	WBC х109/L	PTL х109/L
SDB	30	9/21	58.87 ± 6.93	13.36 ± 4.85	69.37 ± 17.57	43.21 ± 6.15	27.91 ± 8.45	4.81 ± 0.56	6.51 ± 1.38	214 ± 38
Control	28	10/18	60.82 ± 8.01	14.02 ± 3.71	73.17 ± 15.39	45.08 ± 7.89	29.34 ± 5.79	4.59 ± 0.74	6.71 ± 1.52	209 ± 40
P		/	0.368	0.780	0.235	0.214	0.273	0.187	0.611	0.582

Table 2.

Effects of SDB on blood pressure, blood glucose, and blood lipids

Group	N	BMI (kg/m2)		P	SBP (mmHg)		P	DBP (mmHg)		P	HbA1c (%)		P	LDL (mmol/L)		P	TG (mmol/L)		P
		Before	After		Before	After		Before	After		Before	After		Before	After		Before	After
SDB	30	24.58 ± 2.57	23.75 ± 2.05	0.178	131.86 ± 9.43	127.96 ± 7.25	0.064	89.21 ± 6.80	86.42 ± 4.96	0.050	7.51 ± 0.54	7.35 ± 0.49	0.209	2.25 ± 0.84	2.13 ± 0.71	0.567	1.69 ± 0.50	1.47 ± 0.44	0.081
Control	28	24.80 ± 3.73	24.70 ± 2.42	0.898	131.52 ± 9.63	132.05 ± 6.24	0.755	86.67 ± 6.25	87.93 ± 5.25	0.470	7.47 ± 0.67	7.57 ± 0.48	0.556	2.32 ± 0.66	2.22 ± 0.60	0.586	1.79 ± 0.54	1.72 ± 0.49	0.654
P'		0.786	0.108		0.893	0.025		0.145	0.263		0.805	0.101		0.746	0.594		0.486	0.052

P' < 0.05: Control group vs. SDB group; P< 0.05: Before vs. After treatment

In the SDB group, SDNN and RMSSD changed significantly after training for three months (P < 0.05), There were still statistically significant differences after adjusting for the effect of age, whereas HR and PNN50 remained unchanged (P > 0.05). In the control group, the HR, SDNN, RMSSD, and PNN50 remained unchanged after treatment (P > 0.05). At baseline, there were no significant differences in HR, SDNN, RMSSD, or PNN50 between the two groups (P > 0.05). After three months of training, there were no significant differences in the HR and PNN50 between the two groups (P > 0.05), but significant differences were observed in the SDNN and RMSSD between the two groups (P < 0.05). There were still statistically significant differences after adjusting for the effect of age (Table 3).

Table 3.

Time domain variation of HRV in each group before and after treatment

Group	N	HR(bpm)		P	SDNN(ms)		P	RMSSD(ms)		P	PNN50(%)		P
		Before	After		Before	After		Before	After		Before	After
SDB	30	78.66 ± 11.44	72.52 ± 10.20#	0.062	(24.50, 45.85)	(34.35, 53.87)# *	0.021	(17.48, 22.35)	(19.57, 25.01)#*	0.023	(1.26, 3.60)	(1.96, 4.18)	0.192
Control	28	77.84 ± 12.72	76.09 ± 13.50	0.670	(25.84, 42.27)	(23.45, 43.37)	0.773	(16.22, 20.00)	(16.72, 21.13)	0.305	(1.66, 3.38)	(1.25, 3.64)	0.909
P'		0.797	0.258		0.586	0.004		0.135	0.013		0.913	0.050

P' < 0.05: Control group vs. SDB group; P< 0.05: Before vs. After treatment

#: Before vs. After treatment, after adjusting for age, P < 0.05; *: Control group vs. SDB group, after adjusting for age, P < 0.05

In the SDB group, LF, HF, and the LF/HF ratio changed significantly after training for 3 months (P < 0.05), There were still statistically significant differences after adjusting for the effect of age, but LF, HF, and the LF/HF ratio remained unchanged in the control group after training (P > 0.05). At baseline, there were no significant differences in the LF, HF, or LF/HF ratio between the two groups (P > 0.05). After training for 3 months, there were significant differences in LF, HF, and the LF/HF ratio between the two groups (P < 0.05). There were still statistically significant differences after adjusting for the effect of age (P < 0.05). (Table 4)

Table 4.

Frequency domain variation of HRV in each group before and after treatment

Group	N	LF (ms2)		P	HF (ms2)		P	LF/HF		P
		Before	After		Before	After		Before	After
SDB	30	(238.02,307.73)	(229.32,267.34)# *	0.022	(100.62,136.83)	(112.78,146.65)#*	0.012	(1.98, 3.04)	(1.63, 2.08)# *	0.002
Control	28	(246.05,315.21)	(253.56, 319.25)	0.649	(99.42,135.79)	(102.46,135.83)	0.600	(1.86, 2.91)	(1.90, 2.79)	0.432
P'		0.901	0.002		0.663	0.042		0.720	0.010

P' < 0.05: Control group vs. SDB group; P< 0.05:Bbefore vs. After treatment

#: Before vs. After treatment, after adjusting for age, P < 0.05; *: Control group vs. SDB group, after adjusting for age, P < 0.05

Using heart rate recovery at 1 min after the 6-minute walk test (HRR1) as the evaluation index, the changes before and after deep breathing training were analyzed. The results showed that there was no significant difference in HRR1 measurements between the experimental group and the control group after slow deep breathing training (P > 0.05). (Table 5)

Table 5.

Difference in HRR1 before and after treatment in two groups

Group	N	HRR1		P
		Before	After
SDB	30	14 ± 5.07	14.17 ± 5.06	0.134
Control	28	13.36 ± 5.19	13.71 ± 4.99	0.152
P'		0.287	0.203

P' < 0.05: Control group vs. SDB group; P< 0.05: Before vs. After treatment

Discussion

This study found that slow breathing exercise training with soothing music can improve heart rate variability (HRV) in patients with type 2 diabetes mellitus (T2DM) and cardiac autonomic neuropathy (CAN), independent of age.

CAN is one of the common chronic complications of T2DM [21], prone to causing myocardial infarction, sudden cardiac death, and painless myocardial ischemia, serving as a significant factor in mortality among T2DM patients [1]. Early cardiovascular autonomic neuropathy in T2DM primarily affects the parasympathetic nerve, leading to baroreceptor and HRV abnormalities, while the late stage involves both sympathetic and parasympathetic nerves, causing sympathetic-parasympathetic imbalance [2]. Currently, there is no effective treatment for cardiovascular autonomic neuropathy in T2DM. Thus, non-invasive vagus nerve stimulation to improve vagal function and maintain sympathetic-parasympathetic balance may be a promising direction for preventing and treating cardiovascular autonomic neuropathy in T2DM.

In this study, the experimental group performed slow deep exhalation and inhalation exercises with soothing music as the background under voice guidance. Soothing music can make subjects stay in a relaxed state, while voice guidance can establish auditory-respiratory biofeedback to ensure the consistency of respiratory rate (6 breaths per minute) and inhalation-exhalation ratio (1:2). This biofeedback allows subjects to stably control the respiratory training process and maintain a good psychological state, so as to effectively exert the effect of slow deep breathing training on HRV. Slow deep breathing training produces strong stimulation to vagal afferent neurons, which is integrated in the brainstem through baroreflex, respiratory sinus arrhythmia, pulmonary afferent neurons and emotional regulation network, and then reflected in the cardiac vagal efferent nerve [22]. In this study, after slow deep breathing training, the vagus nerve-mediated HRV indices (RMSSD, HF) and the overall HRV index (SDNN) significantly increased in T2DM patients with CAN, suggesting that slow breathing training improved cardiac vagal function. The LF (reflecting the combined effect of sympathetic and parasympathetic nerves) and LF/HF ratio (evaluating sympathetic-parasympathetic balance) decreased. When LF decreases while HF increases or remains stable, the reduced LF/HF ratio indicates enhanced vagal activity and autonomic balance shifting toward the vagus nerve, consistent with our results. PNN50 showed no change, possibly due to the small sample size, which only reflected changes in primary indices (RMSSD, SDNN) rather than PNN50.

A randomized controlled trial in coronary heart disease patients found increased SDNN after slow breathing training [23]. Another trial in 46 post-myocardial infarction patients showed short-term elevation of HRV indices (SDNN, HF) with slow breathing training [17]. In type 1 diabetes, reduced baroreflex sensitivity (BRS)—an indicator of early cardiovascular autonomic dysfunction—can be restored by deep breathing regardless of disease duration [24]. In T2DM patients with or without renal impairment, BRS significantly improved during slow breathing and hyperoxic therapy, suggesting partial reversal of autonomic dysfunction by slow breathing [25], consistent with our findings.

Our 3-month study detected improved HRV in T2DM patients with cardiovascular autonomic neuropathy after training. Slow breathing repeatedly stimulates vagal afferent nerves, inducing functional changes that chronically enhance vagal efferent function. These changes may stem from optimized bodily functions, including subcortical and cortical modifications (e.g., improved baroreflex and respiratory sinus arrhythmia) and neural network adjustments in emotion regulation [26]. Such chronic changes can reduce blood pressure [27], improve cardiopulmonary function in asthma, relieve stress, and lower catecholamine and cortisol levels [28].

Aging reduces cardiac parasympathetic activity, a key factor affecting HRV [29]. Adjusting HRV results for age, we found that significant differences remained in all HRV indices after slow breathing training, indicating age did not interfere with the training’s effects.

Heart rate recovery (HRR) after exercise termination is a critical indicator of autonomic function (e.g., vagal and sympathetic nerves). Studies show HRR indirectly reflects sympathetic-vagal tone and balance [30]. HRR is regulated by sympathetic-parasympathetic interactions, with the vagus nerve slowing heart rate early after exercise cessation [31]. Our study found no significant difference in HRR1 post-training, possibly due to interference from environmental factors, joint mobility, physical strength, and cardiopulmonary function, which introduced subjectivity and masked true training effects. In contrast, HRV measurements were more objective.

Systolic blood pressure (SBP) decreased in the training group vs. the control group, suggesting slow breathing may affect blood pressure regulation. A randomized trial in 43 prehypertensive patients showed increased HRV and reduced SBP/DBP after 3 months of slow breathing training [32]. Another trial in 65 hypertensive patients found reduced SBP with slow breathing [33], consistent with our results.

Andreas Vosseler et al. reported that slow breathing regulates cardiac autonomic activity without affecting blood glucose or insulin secretion [34]. Our study also found no changes in HbA1c, LDL, or TG after training, indicating slow breathing has limited impact on peripheral glucose and lipid metabolism.

Limitations

This study has limitations: multiple factors influenced outcome measures; no specific regulations on hypoglycemic, antihypertensive, or lipid-lowering medications; failed to record the specific medication conditions of each subject in detail; small sample size without stratification by hypertension status; lack of urinary protein assessment for diabetic nephropathy; and a female-dominated sample prone to bias. These issues warrant improvement in future research.

Conclusion

Based on comprehensive treatments (hypoglycemic, lipid-lowering, antihypertensive, and neurotrophic therapies), rhythmic slow breathing training with soothing music can improve HRV in patients with type 2 diabetes mellitus and cardiac autonomic neuropathy, independent of age. This approach holds promise as a non-invasive novel treatment for cardiac autonomic neuropathy in type 2 diabetes.

Abbreviations

SDB

Slow deep breathing

HRV

Heart rate variability

T2DM

Type 2 diabetes mellitus

CAN

Cardiovascular autonomic neuropathy

HR

Heart rate

BP

Blood pressure

HRR1

Heart rate recovery in the first minute

RMSSD

The square root of the mean squared differences of successive RR

SDNN

The standard deviation of N-N intervals

PNN50

He root mean square of the difference between adjacent R-R intervals

LF

Low-frequency power

HF

High -frequency power

BRS

Baroreflex sensitivity

Author contributions

M.X.: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Validation, Writing – original draft, Writing– review & editing. Y.S.: Conceptualization, Methodology, Project administration, Validation, Visualization, Writing – review & editing. W.T.: Conceptualization, Methodology, Project administration, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing. PF: Data curation, Project administration, Resources, Visualization, Writing– review & editing. G.L.: Data curation, Investigation, Methodology, Project administration, Resources, Visualization, Writing – review & editing. G.L., Q.F., L.P.: Data curation, Investigation, Methodology, Project administration, Resources, Visualization, Writing – review & editing.

Funding

This study was supported by the 2021 Shanghai Municipal Health Commission Project (202140292).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Tongji Hospital, Tongji University (2021-LCYJ-009). The study was conducted according to the Declaration of Helsinki, and written informed consent was obtained from each patient.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.