Comparative study of the early predictive values of the heart rate variability parameters, neurophysiological and metabolic markers for the early detection of type two diabetic visceral autonomic neuropathy

Abstract

Background:

Early detection of diabetic neuropathy is crucial for both health and economic stability. Clinically, timely diagnosis and intervention can prevent the progression to severe complications, significantly preserving a patient’s quality of life and mobility. Economically, this proactive approach reduces the long-term burden on healthcare systems.

Objective:

To assess and compare the early predictive value of different heart rate variability (HRV) parameters for the early detection of type 2 diabetes mellitus (T2DM)-induced neuropathy. And to evaluate their correlation with the novel serum parameters, Brain-Derived Neurotrophic Factor (BDNF), and Nesfatin-1.

Results:

HRV parameters showed early changes in type 2 diabetics, even before the development of overt neuropathy. We investigated early changes in heart rate variability parameters in a cohort of type 2 diabetic patients compared to age-matched healthy controls. We found that heart rate variability changes occur even before the development of neuropathic symptoms, suggesting that it can be used as an early detection measure in diabetic neuropathy. Also, we measured serum brain-derived neurotrophic factor and Nesfatine-1 in all study subjects and found brain-derived neurotrophic factor to show a stronger correlation with the HRV parameters than Nesfatine-1. This correlation was more significant in the diabetic patients than the healthy controls.

Conclusion:

We concluded that the HRV metrics can serve as sensitive early biomarkers for detecting type 2 diabetes mellitus-induced neuropathy, especially cardiac autonomic neuropathy, even in T2DM patients without clinical signs of neuropathy. Additionally, BDNF is superior to Nesfatin-1 for early detection of visceral autonomic neuropathy.

Introduction

Increasing evidence supports the need for the early detection of complications associated with type 2 diabetes mellitus (T2DM), which is a complex, multisystem metabolic and vascular disorder with a high global prevalence. Implementing early detection strategies for these complications is important for secondary prevention, and can have a profound impact on health outcomes and socioeconomic factors.[1]

Diabetic neuropathy is a gradually progressive pathophysiological condition that can begin as early as the prediabetic stage or during the impaired glucose tolerance phase, even before the onset of clinically diagnosed diabetes.[2]

Early detection of diabetic complications, particularly diabetic neuropathy, confers significant health and economic benefits. Identifying neuropathy at an early stage can prevent severe health outcomes, such as ulcers, infections, amputations, and loss of mobility, thus improving patient quality of life and reducing the healthcare burden.[3]

Heart rate variability (HRV) and respiratory variability (RV) are important physiological parameters that reflect autonomic nervous system function, particularly the balance between sympathetic and parasympathetic activities. They can be altered in different etiologies of dysautonomia.[4]

HRV refers to the physiological fluctuations in the intervals between consecutive heartbeats. Reduced HRV is a well-recognized marker of autonomic dysfunction, and has been associated with increased cardiovascular risk, neuropathy, and other diabetic complications, which often occur even before clinical symptoms become evident.[5]

HRV holds a promising potential as an early diagnostic tools to detect subclinical autonomic dysfunction in diabetic patients. By assessing these variabilities, clinicians can potentially identify individuals at greater risk of developing severe complications, such as cardiovascular diseases, diabetic neuropathy, and sudden cardiac death, enabling early intervention and improved management strategies to prevent disease progression and enhance patient outcomes.[6]

The primary objective of this study was to elucidate the patterns associated with early alterations in HRV among adult patients diagnosed with T2DM and to systematically evaluate and contrast the characteristic changes observed between traditional autonomic HRV markers. In addition, this study evaluated the novel biomarkers, brain-derived neurotrophic factor (BDNF) and nesfatin-1, to examine their potential relevance and implications in the context of diabetic autonomic neuropathy.

Methods

Methodology and research design

This study used a cross-sectional, observational, non-interventional study design.

Variables

Serum neurophysiological markers (BDNF and nesfatin 1) and HRV metrics.

Data collection procedures

Subject selection and recruitment

A total of 60 participants, including 30 age-matched healthy controls (healthy group) and 30 patients diagnosed with T2DM (diabetic group), were included in this study. The healthy subjects were recruited through an announcement and email invitations. The diabetic subjects were selected from the regular attendants of the outpatient clinic in the King Khalid University Hospital (KKUH) and the King Saud University Medical City (KSUMC), Riyadh, Kingdom of Saudi Arabia. All the measurements and procedures of this research were conducted in accordance with the ethical guidelines of the King Saud University IRB: E-23-8253. Ethic cimmittee the approval was obtained, with an approval date of 30 March 2022.

Participants

The participants were selected based on specific inclusion and exclusion criteria to ensure the validity and reliability of the HRV measurements.

Inclusion criteria

Age between 40 and 65 years.

No comorbidity, specifically excluding the diagnosis of T2DM in the diabetic group.

Exclusion criteria

Age less than 40 or more than 65 years.

Presence of comorbidities (other than the diagnosis of T2DM in subjects of the diabetic group).

Devices and software

Measurement

HRV data was collected using the Polar Verity Sense armband.

Data analysis

HRV data were analyzed using the Kubios HRV software.

Sample handling, storage, and measurements were performed in the Physiology lab, College of Medicine, King Saud University.

Enzyme-linked immunosorbent assay (ELISA)

The levels of human nesfatin-1 and BDNF were determined using a solid-phase sandwich enzyme-linked immunosorbent assay (ELISA) kit (Catalog# EH333RB and EH42RB, respectively) based on the manufacturer’s instructions (Invitrogen/Thermo Fisher Scientific, Waltham, Massachusetts, USA).

Procedures and data collection

Timeframe

All subjects were sent to the College of Medicine at King Saud University (KSU) in the morning between 7:30 AM and 10:00 AM. Before performing any physical tests, they were seated in a comfortable chair in a quiet room with dimmed lights, and instructed to remain still and quiet for HRV measurement.

Following the completion of HRV measurement, the subjects proceeded to KKUH to have blood samples collected by a certified phlebotomist. The blood samples were collected into yellow gel tubes and then transferred for centrifugation. The processed samples were stored at −80°C in the Physiology Department at KSU for further analysis. The data collection process was completed during a single visit for each subject.

Heart rate variability data acquisition

The Polar Verity Sense armband was placed on the right arm, just below the elbow, ensuring proper contact with the skin and consistent environmental conditions. The HRV test was performed for 20 min, while the subject was in a seated position without moving. The first 5 min of the 20-min recording were discarded to avoid collecting noisy data because of any change in body position (from standing to sitting), and the data of the following 15 min were selected for analysis based on the standards of HRV analysis.

Precautions:

Before the procedure: Subjects were instructed to avoid smoking and caffeine intake for 24 h before the examination.

During the procedure: Paced breathing rate, same body position, avoiding standing, avoiding stressors, and minimal swallowing.

Statistical methods

The data were analyzed using IBM SPSS Statistics for Windows, Version 30.0 (IBM Corp., Armonk, NY, USA) and Microsoft Excel for Microsoft 365 (Microsoft Corp., Redmond, WA, USA). Based on the distribution of the variables, which included normally and non-normally distributed data, either the student’s t-test or the Mann-Whitney test was applied as appropriate. Descriptive statistics were presented as means and standard deviations (SDs) for normally distributed continuous variables, whereas medians and interquartile ranges (IQRs) were used for non-normally distributed continuous variables. Normality tests were conducted for all variables, and comparisons between subgroups were performed using parametric or non-parametric methods as applicable. A P value of < 0.05 was considered statistically significant.

Results

Enzyme-linked immunosorbent assay (ELISA) results

The results of the Shapiro–Wilk test indicated a normal distribution of BDNF in both groups. A two-sample t-test revealed that healthy controls showed significantly higher mean BDNF concentrations (mean = 31.92 ng/mL) compared to diabetic patients (mean = 22.34 ng/mL, P = 0.0008). The effect size indicated a large practical difference. For serum nesfatin-1 levels, both groups exhibited highly skewed distributions (Shapiro–Wilk P < 0.001); therefore, medians were more representative compared to means. Healthy subjects showed a wider spread and slightly higher central tendency (median = 2.4) compared to diabetic participants (IQR = 25.48, median = 2.96); however, the non-parametric Mann–Whitney test (P = 0.6522) indicated no statistically significant difference in nesfatin-1 levels between the two groups, with a small effect size, indicating a negligible practical difference [Figure 1].

Figure 1.

Comparison of BDNF and nesfatin-1 levels between the study groups. BDNF = brain-derived neurotrophic factor

HRV Parameters

Mean heart rate (HR)

Average heart rate (HR) was approximately 3.8 bpm higher in the diabetic group; however, the intergroup difference is not statistically significant. Both groups showed normal distribution.

Root mean square of successive differences (RMSSD)

Both Shapiro–Wilk tests rejected normality (P < 0.001) with marked right-sided skewing; therefore, non-parametric inference was performed, which revealed a significant difference (healthy > diabetic; P value: 0.00069), and a moderate-to-large effect size (Cohen’s d: 0.7).

Standard deviation of NN intervals (SDNN)

The Shapiro–Wilk normality (P < 0.05) for both groups indicated a nonnormal distribution. Therefore, the Mann–Whitney U test was used. SDNN values in the healthy group were significantly higher compared to those in the diabetic group, with a mediumsized difference (P = 0.008), and a moderate effect size.

NN50

Healthy group mean = 64 ± 40 and median = 51 (very wide spread). Diabetic group mean = 39 ± 32 and median = 33.

Shapiro–Wilk P < 0.05 for both groups indicated a non-normal distribution. Therefore, the nonparametric Mann–Whitney U test was applied, yielding a higher NN50 in the healthy group (P = 0.0035), with a mediumtolarge effect size.

pNN50

Non-normal distribution; therefore, the Mann-Whitney U test was applied. A significant difference was observed between the two groups, with markedly lower values in the diabetic group (P value = 0.00030), with a large effect size.

Para-sympathetic nervous system index

For the Shapiro–Wilk normality test. Both P values were less than 0.05; thus, neither group was normally distributed.

The Mann–Whitney U test was used, which revealed a significant difference between the two groups (P value = 0.0223), with a moderate effect size.

Sympathetic nervous system (SNS) index

The measurements (mean = 1.37 and median = 1.58) for the diabetic group exceeded those for the healthy group (mean = 0.78 and median = 0.56).

Both distributions were normally distributed, and an independent t-test suggested that the mean differences were mildly significant (P = 0.065).

Overall, measurements from the diabetic group clustered around higher values with moderate variability, whereas those from the healthy group were lower and more heterogeneous; however, the intergroup difference was not statistically significant.

Stress index

Normality (Shapiro–Wilk) P values were 0.17 and 0.58 for the healthy) and diabetic groups, respectively, indicating a normal distribution, with a statistically significant intergroup difference and a higher stress index in the diabetic group.

Discussion

Diabetic neuropathy is a common complication of diabetes, which affects both somatic and autonomic nerves.[7]

Early detection of diabetic neuropathy is important as it significantly reduces the risk of severe complications, including foot ulcers, infections, and amputation. Early identification enables timely intervention and enables patients to adopt improved glycemic control. HRV analysis is a promising novel tool for early detection.[8]

One of the simplest and most reliable methods for assessing cardiac autonomic function is to measure HRV, or the variation in time between consecutive heartbeats, with greater variability indicating higher parasympathetic activity. A high HRV suggests that an individual can adapt effectively to subtle environmental changes, whereas low HRV is considered a marker of increased cardiovascular risk.[9]

Our findings emphasized the correlation between HRV parameters and markers of neuropathy. We found that HRV analysis might serve as a potential predictor of the onset of neuropathy in patients with type 2 diabetes. Incorporating this information might provide significant diagnostic and prognostic benefits, as well as enhance future guidelines, diagnoses, and clinical monitoring strategies [Figures 2-6 and Tables 1-2].

Figure 2.

Bar chart comparisons of the HRV results between the study groups. HRV = Heart rate variability

Figure 6.

Correlation scatterplots from the diabetic group

Table 1.

Main descriptive statistics for the healthy group

		Mean		Standard Deviation		n
BDNF		31.92		10.991		30
Nesfatin-1		20.30		27.329		30
Mean HR		78.23		12.574		30
RMSSD		56.30		32.798		30
SDNN		55.09		23.995		30
NN50		64.13		40.277		30
pNN50		29.21		17.471		30
PNS index		−0.17		1.482		30
SNS index		0.78		1.380		30
Stress index		9.62		3.930		30

RMSSD, root mean square of successive differences; SDNN, standard deviation of NN intervals; SNS, sympathetic nervous system; HR, heart rate

Table 2.

Main descriptive statistics for the diabetic group

		Mean		Std. Deviation		n
BDNF		22.3423		10.00868		30
Nesfatin-1		14.7007		22.70380		30
Mean HR		82.0000		10.54710		30
RMSSD		36.6333		21.97097		30
SDNN		40.0567		17.27673		30
NN50		38.6000		31.94672		30
pNN50		15.1700		13.28472		30
PNS index		-0.8667		0.98730		30
SNS index		1.3683		0.99274		30
Stress index		12.5287		4.04156		30

RMSSD, root mean square of successive differences; SDNN, standard deviation of NN intervals; SNS, sympathetic nervous system; HR, heart rate

Figure 3.

Comparisons of the HRV results between the study groups. HRV = Heart rate variability

Figure 4.

Boxplot comparisons of the HRV results between the study groups. HRV = Heart rate variability

Figure 5.

Group one correlation scatterplots

Our results indicated that the heartratevariability parameter (RMSSD) was markedly lower in the diabetic group compared to healthy participants. Power calculations revealed a statistically significant intergroup difference.

Our results also indicated that SDNN showed markedly lower values in the diabetic group. The mean SDNN for the healthy subjects was significantly higher than that for the diabetic group. This finding suggested that individuals with T2DM exhibit reduced overall heartratevariability, indicating decreased autonomic flexibility relative to the healthy controls. These results support the reduced vagal tone in diabetes, which warrants larger, covariate-controlled studies for validation.

The NN50 analysis showed markedly greater beat-to-beat variability in the healthy group compared to the diabetic group. Our results also showed that pNN50, which is a marker of short-term, parasympathetic heartratevariability, was considerably lower in individuals with diabetes compared to the healthy controls with a large effect size indicating that diabetics experience a marked reduction in parasympathetic modulation of HR compared to healthy individuals.

Collectively, these results support that diabetics exhibit significantly reduced HR variability, consistent with autonomic dysregulation, and the magnitude of the difference is statistically and clinically relevant.

These results, in general, are consistent with the recent literature. As the HRV parameters especially RMSSD indicates better vagal functioning.[10] Moreover, lower SDNN suggests impaired sympathetic cardiac control.[11] Our results were consistent with other studies on HRV in T2DM patients.[12,13]

In a study involving continuous electrocardiogram and glucose monitoring of 43 patients with T2DM, HRV measures were significantly reduced during episodes of hyperglycemia compared to periods of better glycemic control.[14]

Our results were consistent with a previous study, that reported that among 160 T2DM patients stratified by disease duration, those with early (<5 years) T2DM exhibited markedly reduced HRV parameters compared to healthy controls, with more pronounced reductions in those with a longer-standing disease.[12] and[15] demonstrated 97% sensitivity and approximately 90% accuracy in the time-frequency HRV metrics for detecting autonomic neuropathy, providing a simpler alternative to conventional reflex testing in patients with T2DM with nephropathy.

Autonomic indices (PNS, SNS, and stress indexes) suggest that individuals with diabetes in this cohort have materially higher physiological or perceived stress, together with impaired parasympathetic function. This finding warrants further investigation into potential mediators, such as glycemic control, medication regimen, and psychosocial factors, and may motivate stress-reduction interventions to improve diabetic health outcomes.

Our results showed only a modest, non-significant increase in resting HR in the diabetic group, with no statistical significance or apparent clinically important differences. This finding suggested that mean HR, although frequently used as a general indicator of physical fitness in gyms and non-clinical environments, lacks utility as an early predictor of nervous system integrity. Despite its widespread use, the mean HR appears insufficiently sensitive to subtle physiological changes associated with early nervous system dysfunction. In contrast, emerging HRV parameters demonstrate greater prognostic potential in this context.

Our results indicated that healthy adults showed markedly higher statistically significant serum BDNF concentrations than diabetic adults. Several recent studies have examined the differences in the BDNF levels in T2DM patients relative to the healthy population. The results were conflicting; however, it is believed that their levels were decreased in T2DM individuals compared to healthy age-matched controls.[16]

Our results were consistent with those of several studies showing a decrease in BDNF levels in T2DM patients,[17] compared to control subjects. With reciprocal correlation with the fasting glucose levels, they concluded that BDNF levels correlate with obesity and T2DM, and are independent indicators of diabetic complications.

Uzel et al.,[18] found that the serum and aqueous humor levels of BDNF decrease in patients with diabetes mellitus even before the emergence of clinical signs of diabetic retinopathy. In contrast, Boyuk et al.,[19] found that serum BDNF levels were higher in patients with T2DM, and suggested that BDNF might contribute to glucose and lipid metabolism dysregulation and subsequent inflammation. Similarly, Suwa et al.,[20] who evaluated serum BDNF levels in newly diagnosed T2DM female patients, found similar results and linked the increased levels of BDNF to metabolic and glycemic pathophysiological dysregulation.

The observed inconsistencies across studies examining BDNF levels in T2DM may be attributed, in part, to heterogeneity in study design, population demographics, degree of glycemic control, and the methodologies used for BDNF quantification. In addition, the presence of confounding variables, such as the use of antidiabetic or neuroactive medications, coexisting medical conditions, and lifestyle factors, including diet and physical activity, may have influenced the reported outcomes. To determine the complex relationship between BDNF and T2DM, further studies are warranted, particularly through well-controlled, large-scale longitudinal studies that systematically account for these potential confounders.

Nesfatin-1 is a newly identified neuropeptide that exerts diverse physiological effects by interacting with specific G-protein-coupled receptors. It plays a role in regulating feeding behavior, glucose metabolism, and the activity of the hypothalamic-pituitary-adrenal axis.

Our results showed no statistically significant differences in nesfatin-1 levels between the two study groups, although they were slightly higher in the diabetic group. Recent studies have shown conflicting results, with some reporting higher nesfatin-1 levels in healthy individuals compared to those with diabetes.[21,22,23]

Other studies have reported opposite results.[24]

A meta-analysis revealed no significant differences in nesfatin-1 levels between the healthy population and T2DM patients,[25] which was consistent with our results.

The discrepancies in nesfatin-1 levels among diabetic patients compared to healthy subjects may be affected by factors such as disease stage, treatment status, and individual metabolic differences. Elevated nesfatin-1 levels are more prominent in newly diagnosed, untreated T2DM patients, and may act as a compensatory mechanism. As the disease progresses and with the introduction of anti-diabetic treatments, nesfatin-1 levels may decrease.

The inconsistency in the literature highlights a key challenge in the current understanding of the role of nesfatin-1 in glucose metabolism and energy homeostasis. One possible explanation for the lack of a significant difference in the present study could be the high inter-individual and inter-study variability in nesfatin-1 levels reported in the literature.

Nesfatin-1 does not have a consistent or direct association with T2DM. Its physiological role is more nuanced and influenced by a complex network of neuroendocrine factors. Therefore, while our results do not support a precise diagnostic or differentiating role for nesfatin-1 in T2DM, they highlight the need for studies using standardized protocols and larger, more diverse cohorts to clarify its significance in metabolic disorders.

The abovementioned findings suggested that BDNF is more useful for the early detection of visceral or autonomic neuropathy compared to nesfatin-1. This inference is further supported by the correlations observed in our study.

Our results suggested that HRV parameters may be dysfunctional in T2DM, indicating their potential role as early biomarkers for the detection of diabetic neuropathy. The observed positive correlations with neurotrophic factors, such as BDNF, further support the combined use of HRV indices and neurotrophic markers as an approach for the early diagnosis and prognostic assessment of neuropathy in the context of T2DM.

Conclusion

HRV metrics may serve as sensitive indicators for the detection of T2DM-induced neuropathy and early biomarkers for diabetes-associated neuropathies, particularly cardiac autonomic neuropathy, even in T2DM subjects with no clinical evidence of neuropathy. Among the altered plasma molecules, BDNF exhibits superior predictive value compared to nesfatin-1 for the early detection of visceral autonomic neuropathy. It is also correlated with cardiac autonomic dysfunction reflected by alterations in HRV parameters, in particular parasympathetic-related neuropathy.

Study Limitations and Future Aspects

The mail limitation of our study is the small sample size, so further larger sample-sized studies, with longer prospective course and deeper multivariate and correlation analysis, are needed to elucidate the specific patterns of diagnostic validity of the HRV and neurotrophic parameters and to validate their use as diagnostic tools.

Abbreviations

• AGEs: Advanced Glycation End-Products

• BDNF: Brain Derived Neurotropic factor

• BPM: Beat Per Minute

• ELISA: Enzyme linked Immunosorbent Assay

• HR: Heart Rate

• HRV: Heart Rate Variability

• NN50: Number of successive Normal - Normal intervals differing by more than 50 ms

• PN: Peripheral Neuropathy

• pNN50: Percentage of successive Normal - Normal intervals differing by more than 50 ms

• PNS: Para-Sympathetic Nervous System

• RMSSD: Root Mean Square of Successive Differences

• SD: Standard Deviation

• SDNN: Standard deviation of normal-to-normal intervals

• SNS: Sympathetic Nervous System

• T2DM: Type Two Diabetes Mellitus

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.