The Physiological Effects of a Walking to Music Intervention in Adults with Intermediate Hyperglycemia

Purpose: Overweight individuals are at risk of intermediate hyperglycemia (IHG) [1]. Increasing physical activity (PA) in IHG is one method to reduce the risk of glucometabolic and cardiovascular (CV) complications [2]. This study investigates walking to music as a modality to increase moderate intensity PA and regulate glucometabolic disturbances. Patients and Methods: Participants were randomized to usual care (UC) or intervention group (IG) who completed a 6 month walking to music program. Physiological assess-ments for a range of variables (DEXA, flow mediated dilatation (FMD) and glucometabolic biomarkers) were completed at baseline, 4, 6 and 9 months (follow-up). Results: For IG group, walking compliance decreased with time however 71.4%, 79.5% and 73% of walking completed was moderate intensity at 4, 6 and 9 months. At 6 months IG FMD was significantly lower than UC and HbA 1C was 5% lower. Other physiological markers were not altered. Conclusion: Walking to music may be a novel method to encourage moderate intensity PA. However, the majority of results demonstrate that this intervention was not more effective than UC in managing glucometabolic and CV biomarkers in IHG. Future interventions should include additional support for the entire study duration however this has cost implications.


Introduction
Excess adiposity continues to present significant health and social problems [3]. Adults carrying excess adipose tissue are at higher risk of developing co-morbidities such as impaired glucose metabolism and cardiovascular disease (CVD) and have increased mortality rates [4]. As adiposity increases beyond a body mass index (BMI) of 25 kg/m 2 , the risk of the development of type 2 diabetes mellitus (T2DM) also increases [5]. Many adults with hyperglycemia remain in an asymptomatic state for a prolonged period before T2DM diagnosis [6]. This intermediate hyperglycemia (IHG) includes impaired glucose tolerance (IGT) and/or impaired fasting glucose [1]. In addition to the impaired handling of glucose, IHG can increase oxidative stress, inflammation and endothelium dysfunction [6] [7] [8]. Approximately 86 million (37% of the adult US population) meet the diagnostic criteria for IHG [9]. In the UK, around 7 million people are thought to have IHG [2] with estimates suggesting that 1 in 3 English adults have IHG [2] [10]. Like the majority of the population, less than half of this population report completing current PA guidelines [11]. It is well established that increasing PA in IHG is one method to reduce the risk of glucometabolic and CVD complications [2]. Moderate intensity exercise in particular is required to regulate hyperglycaemia by increasing GLUT-4 translocation to muscle tissue, enhancing uptake of glucose via PI-3K and MAP kinase pathways [12]. Similar effects are noted for the reduction of inflammation associated with diabetic progression [13]. Recent studies have calculated that £1 million per hour is spent from the National Health Service (NHS) budget on diabetes health care costs [14]. As such, preventing progression from IHG to T2DM using low cost, effective and sustainable methods is important for public health. Assisting individuals to achieve and maintain moderate intensity exercise remains problematic. Previous research has used heart rate monitors, self-report methods and accelerometers/ pedometers as tools, but many of these possess limitations from a user's perspective, but also a research reliability and validity perspective [15]. Using music with a pre-determined tempo to help stimulate and regulate cyclic PA intensity may ensure PA is regularly performed at the desired intensity [16].
The aims of this study were to examine the effects of a walking to music intervention to encourage a pace which achieves moderate intensity on glycemic control, oxidative stress, endothelial dysfunction and inflammatory markers in IHG.

Methodology
Following approval by Ulster University Research Ethics Committee and the Office for Research Ethics Northern Ireland (Approval number: 12/NI/0063), overweight and obese IHG adults who met the diagnostic criteria of a fasting plasma glucose between 5.6 -6.9 mmol/L −1 or a two hour glucose concentration between 7.8 -11.0 mmol/L −1 (n = 42), were recruited from consultant led diabetic clinics and from a university within Northern Ireland (Figure 1). Exclusion Open Journal of Endocrine and Metabolic Diseases criteria for the study included those who had progressed to T2DM/those who had normoglycaemia, those under the age of 18 or over the age of 65. Participants were also excluded if they were smokers, had a history of cardiovascular events, or were taking any medication which would interfere with the blood biomarkers assessed. Participants provided medical history details and informed consent before random assignment to an intervention group (IG) (n = 22) or usual care group (UC) (n = 20) (Figure 1). IG participants (n = 17) received an individualized 6 month walking program designed to increase moderate intensity walking to 30 minutes per day based on current PA guidelines [17]. Self-chosen music with a predetermined tempo was selected based upon height-related stride rate recommendations by Rowe et al. [18] to elicit moderate intensity exercise for each individual (50% -70% maximum HR, 90 -113 steps/min). Cadence (steps/min −1 ) was recorded using the inbuilt MP3 accelerometer and analyzed via MATLAB (MATLAB, The Mathworks Inc, Natick, MA, 2014). IG received Open Journal of Endocrine and Metabolic Diseases three physical activity consultations (PACS) during the 9 month intervention (0, 4 and 6 months) based on the work of Kirk and colleagues [19]. These were complimented by booster telephone sessions to revise goals, barriers to exercise and methods of achieving PA targets. UC participants (n = 20) received an information sheet detailing current PA guidelines and methods to achieve these on a weekly basis (this is the usual care those with IHG currently receive from the NHS). All measurements were completed at baseline, 4, 6 and 9 months ( Figure  1).

Physiological Testing
Height (cm) and body mass (kg) were recorded and BMI (kg/m 2 ) calculated [20]. Arterial blood pressure was measured in a supine position (Omron M5-1 BP monitor, Surrey, UK). DXA was used to measure whole and regional body composition and bone. Arterial stiffn-ness was assessed using a PulseTrace Pulse Contour Analysis (PCA) system (Micro Medical, UK). A 2D high resolution Sonoace Pico ultrasound system (GE Healthcare, Memphis TN) was used to assess endothelial function according to the methods of Harris et al. [21]. Physical fitness was measured via a sub-maximal graded exercise assessment using walking speeds determined by the ACSM metabolic equations [22]. Stages were 4 minutes in duration and consisted of three stages (12 minutes in total). Heart rate response and respiratory exchange values at each workload (km/hr −1 ) measured walking economy. PA was assessed via a self-administered International Physical Activity Questionnaire (IPAQ) short version [23]. IG intervention % walking compliance was assessed by comparing the app-recorded walks (for volume and duration) to walking prescribed in each program. Intensity compliance (%) was assessed by calculating the number of walks that, on average, were at/ above the predetermined target cadence. Both groups recorded PA in a logbook at each timepoint. At each timepoint, all participants completed a cadence walk around an elliptical track (≥eight minutes) to ensure that the desired stride rate and moderate intensity exercise was completed. Fasted blood samples were assessed for the following metabolites. Glucometabolic status: fasting plasma glucose, oral glucose tolerance test (OGTT [24]), glycosylated hemoglobin (HbA 1C ) and insulin; Lipoproteins: serum total cholesterol, high-density lipoprotein cholesterol (HDL-C), calculated low-density lipoprotein cholesterol (LDL-C) (Fried Wald et al. [25]) and serum triglyceride (TRIG). Inflammatory markers: tumor necrosis factor-α (TNF-α), interleukin 6 (Il-6) and high-sensitivity C reactive protein (Hs-CRP); Oxidative stress markers: Lipid hydroperoxides (LOOHs) (FOX assay), lipid soluble antioxidants (γ-tocopherol, α-tocopherol; Retinol; Lycopene; β-carotene, α-carotene) and ascorbyl free radical. Hematocrit and hemoglobin were also assessed via standard methods.

Data Analysis
Levene's test for equality of variances (p > 0.05) were performed to investigate homogeneity of variance. Independent-sample t-tests (P) (p < 0.05) were used to compare themeans of two groups' (IG vs. UC). A repeated measures GLM was utilized to assess the difference between variables with one between-subjects variable (two groups IG and UC), one within subject variable with two or more levels (e.g.4 time points: baseline, 4, 6 and 9 months) and one continuous dependent variable (parameter measurement) (p < 0.05). All data is presented as mean value ± standard deviation (SD) unless otherwise stated. Descriptive statistics (means or frequencies) were calculated where appropriate. All statistical analysis was completed using Microsoft Excel and SPSS version 22.0, for windows (SPSS, Chicago, IL).

Blood Biomarkers
No significant interaction (time × group) and no main effect for group was . * A repeated measures GLM (p < 0.05) examined differences between groups across assessment points. † Denotes a significant difference between groups at 6 months (p < 0.05). ^D enotes a within group significant difference between given value and 6 month value (p < 0.05) Continuous IPAQ scores are presented as median and interquartile range. A repeated measures GLM found no statistically significant differences between groups across assessment points (p > 0.05).

Insulin
No significant interaction (time × group) effect with no main (time) effect or no significant main (group) effect were found for insulin (p > 0.05; Table 3). No significant interaction (time × group) effect with no main (time) effect or no main (group) effect (p > 0.05) were found for hematocrit (%) and hemoglobin (g/dL −1 ). (Table 4  CHOL = Serum total cholesterol, HDL-C = High-density lipoprotein cholesterol, LDL-C = Low-density lipoprotein cholesterol, TRIG = Serum triglyceride, TNF-α = Tumor necrosis factor-α, IL-6 = Interleukin, Hs-CRP = High-sensitivity C-reactive protein, LOOHs = Lipid hydroperoxides, Ascorbyl = Ascorbyl free radical. A repeated measures GLM (p < 0.05) examined differences between groups across assessment points. † Denotes a significant difference between groups (p < 0.05). ^D enotes a within group significant difference between given value and 6 month value (p < 0.05). ¥ Denotes a within group significant difference between given value and 9 month value (p < 0.05).

Oxidative Stress Markers
No significant interaction (time × group) effect was found for LOOHs (p > 0.05).  Table 4 for full antioxidant results. No significant interaction (time × group) effect, with no significant main (time) effect and no main (group) effect was observed for ascorbyl free radical (p > 0.05).

Physical Fitness Assessment
A significant interaction (time × group) effect was found for heart rate response at given speeds throughout the study [F(3, 32) = 3.073, p = 0.044, multivariate partial eta squared = 0.224]. Post hoc analysis found no significant difference between IG and UC at any assessment points (p > 0.05). Significant differences were found within UC for HR response at 4.2 km/hr −1 between 4 and 9 months (p = 0.34), for HR response to 5.2 km/hr −1 between 4 and 9 months (p = 0.016) and for HR response to 6.2 km/hr −1 between 6 and 9 months (p = 0.008) ( Table  5). No significant interaction (group × time) effect and no significant main (group) effect were observed for RPE. RPE displayed a significant main (time) effect. Post hoc comparisons found a statistical significance between baseline and 4 month (p = 0.024), baseline and 6 month (p < 0.001), baseline and 9 month (p < 0.001) and 4 and 6 month RPE values (p = 0.013). Although no significant interaction (time × group) effect was observed between groups (p > METs = metabolic equivalents, RPE = rate of perceived exertion. Repeated measures GLM examined differences between groups across assessment points (p < 0.05). ¥ Denotes a within group significant difference between given value and 9 month value (p < 0.05).
0.05), at 5.2 km/hr −1 and 6.2 km/hr −1 RPE displayed decreasing trends for both IG and UC across the study, IG RPE at 4.2 km/hr −1 decreased from baseline to 9 months, while UC RPE at 4.2 km/hr −1 decreased from baseline to 6 months before displaying an increasing trend between 6 and 9 months (p > 0.05).

Discussion
To the author's knowledge, this study is the first to demonstrate that walking to music with an individualized predetermined beat is a novel method to ensure PA guidelines for health can be achieved in a free-living environment for up to 9 months. The results have shown that although walking compliance (volume and duration) decreased over the study period from 70(+39)% to 37.5(43.5)%; the majority of participants were able to attain their individualized moderate inten-sity target cadence between baseline to four months, four to six months and six to nine months (71.4%, 79.5% and 73.3% respectively). In other words, over 70% of the physical activity which they did complete was at least moderate intensity or higher. Achieving moderate intensity PA in those with IHG is of particular importance as it is well recognized to reduce the risk of glucometabolic and CVD complications [2]. Previous research by Tudor-Locke and Rowe [16] has indicated that regulating intensity of PA in a free-living environment remains problematic. As such these results in terms of attaining moderate PA in an "at-risk" population who should be targeted for PA behaviour change interventions could be described as promising.

Vascular Function and CVD Risk
Despite these encouraging results in terms of attaining moderate intensity PA, one of themain findings of this study was that the current intervention was not effective in eliciting changes to the majority of glucometabolic and CVD biomarkers assessed. Some significant improvements were found in vascular function; however these should be interpreted with care. SI, which reflects the measurement of stiffness in the large arteries, displayed a significant main effect for time with values decreasing during the study (baseline to 6 months). RI values which reflect the vascular tone of the small arteries did not change throughout the study, however values positively decreased by 7.37% for the IG during the intervention in contrast to UC RI values which increased by 13.6% over the same period. FMD analysis demonstrated a significant interaction effect between groups at six months (IG, 6.51 ± 4.38 vs UC, 11.21 ± 5.72). Interestingly, significant increases were observed within UC FMD (%) between 4 and 6 months (7.40 ± 2.85 and 11.21 ± 5.72 respectively). Whilst FMD in IG displayed a lowering trend across all timepoints. Improvements in arterial stiffness and FMD are crucial in this population which are predisposed to increased CVD risk [26].
Research has highlighted the inverse relationship which exists between FMD and future cardiac event risk [27]. Indeed, a recent meta-analysis showed that for every 1% or 1 standard deviation decrease in FMD there is an associated increased risk of 8% or 22% respectively of suffering a future CVD event [28]. Previous research has shown the reliability of FMD is a particularly intricate technique [29] and as such it is felt an important methodological concern which should be published in relation to the current population. FMD technique relies on skilled acquisition and interpretation of results [30] and as such requires replication in a larger cohort to confirm these findings. Several measurement issues are possible when assessing FMD in obese IHG populations, which require consideration. These include sizing issues with cuff and plinth, inability to lie supine for required sampling and larger quantities of adipose tissue introducing difficulty in isolating the brachial artery for analysis. Previous research has shown that improvements in SI and FMD may be the result of reduced circulating lipids and increased NO availability [26]. This study observed favorable improvements in FMD in the absence of improvements in lipoproteins, adipose tissue mass and Open Journal of Endocrine and Metabolic Diseases Hs-CRP (mg/L −1 ). Reductions in Hs-CRP are normally associated with increased nitric oxide (NO) expression which leads to enhanced NO bioavailability and improved FMD [31]. Although Hs-CRP did not change significantly, the observed decreasing trends may have been sufficient to have impacted on FMD. However, it is important to consider the methodological issues encountered and note that statistical power of this study was based on two-hour glucose rather than FMD.  [36]. Therefore, improving glycemic control via a reduction in HbA 1C is advantageous in IHG and has been achieved to some degree in this study.

Anthropometry
Previous IHG and T2DM interventions have focused on decreasing body mass in overweight individuals [34]. However, it is now clear that improved glucose metabolism and CVD risk reduction can be achieved in the absence of weight loss [37]. This study did not find a significant difference in body fat percentage. IG android fat (%) decreased from 53.8% ± 9.2% (baseline) to 52.9% ± 8.7% at the end of the intervention (6 months), decreases which were maintained at follow up and within the UC group. Centrally located android (%) and gynoid (%) has been associated with increased risk of CVD [38] therefore these results sug-gest that IG android and gynoid fat (%) displayed favorable improvements.

Inflammatory and Oxidative Stress Markers
IL-6 and TNF-α are cytokines associated with skeletal muscle inflammation and feature in obesity related insulin sensitivity [39]. These increases cause the chronically stimulation of the NF-κB pathways which increases insulin resistance and β-cell dysfunction [40]. In turn, increased oxidative stress can be caused by ROS directly damaging cellular macromolecules [40]. In the current study a significant main (time) effect was observed for IL-6 (pg/ml −1 ) from 6 to 9 months. While IL-6 (pg/ml −1 ) displayed a significant increase over time between 6 months and 9 months, it is worth noting follow-up concentrations remained lower for IG participants suggesting a reduced inflammatory response. Harris et al. [41] found that IL-6 (pg/ml −1 ) production during acute exercise is dependent on the intensity of activity and has the ability to directly inhibit TNF-α. In contrast to moderate (50% VO 2max ) and vigorous itensity (75% VO 2max ) activity (p < 0.05), one hour post low intensity activity (25% VO 2max ) elicited no significant effects on IL-6 (pg/ml −1 ) production [21]. In the current study no significance was observed for TNF-α (pg/ml −1 ) however, IG TNF-α (pg/ml −1 ) displayed lowering trends from baseline to 4 months and recorded its lowest value at 9 months. In contrast UC TNF-α (pg/ml −1 ) values displayed non-significant increasing trends throughout the intervention and from baseline to follow-up (0 -6 months and 0 -9 months). This may be due to the observed lack of reduction in fat mass parameters as has been found by Katsuki et al. [42]. Ascorbyl free radical and LOOHs are indicators of oxidative stress [43]. In the current study no statistical significance was found between groups for either of these markers. IG LOOHs increased from baseline to 9 months at each time point. Interestingly, a significant interaction effect was observed for numerous lipid soluble antioxidants (γ-tocopherol and α-tocopherol, retinol and lycopene. IG γ-tocopherol and α-tocopherol displayed similar patterns, increasing continually from baseline to follow-up. In contrast UC values decreased throughout the intervention (0 -6 months). Similarly, IG retinol increased from baseline to 9 months (43.61%) across all assessment points while UC retinol decreased throughout the study. IG lycopene displayed favorable increases from baseline to 6 months and in contrast UC lycopene decreased throughout the study. It may be suggested that the increased antioxidant defense observed above may have been sufficient to control any oxidative stress caused by increased moderate intensity activity. Dietary manipulation did not form part of the study but it is possible that the increase in PA led to changes in antioxidant intake as observed in previous research [44].

Conclusion
To the authors' knowledge, this is the first study to demonstrate that using music with a predetermined tempo can assist overweight adults with IHG to regulate their walking cadence to moderate intensity in a free-living environment. Furthermore, we have demonstrated that it is feasible to complete a 6 month walking to music intervention in an "at risk" population assessing a comprehensive range of physiological measures and biomarkers at 4 time points (including 9 month follow up). This study showed the novel walking intervention and individualized PACs were not more effective than standard NHS care in managing glucose metabolism overweight adults with IHG. This study did achieve significant improvements in FMD (mainly in the UC group); however, as discussed previously, these findings should be accepted with caution. A number of significant metabolic health improvements were observed over time but not between groups. It is suggested that future research examines the effect of walking to music with these parameters as primary outcome measures. It is strongly recommended that behavior change support embedded within PA interventions should be maintained for as long as feasibly and financially possible to determine whether effects on similar parameters may be improved.

Funding
This work was supported by a Department of Education (Northern Ireland) grant as part of the PhD project of Maria Faulkner.