734
ARTICLE Associations between heart rate variability, metabolic syndrome risk factors, and insulin resistance Appl. Physiol. Nutr. Metab. Downloaded from www.nrcresearchpress.com by University of Otago on 07/05/15 For personal use only.
Melanie I. Stuckey, Antti Kiviniemi, Dawn P. Gill, J. Kevin Shoemaker, and Robert J. Petrella
Abstract: The purpose of this study was to examine differences in heart rate variability (HRV) in metabolic syndrome (MetS) and to determine associations between HRV parameters, MetS risk factors, and insulin resistance (homeostasis model assessment for insulin resistance (HOMA-IR)). Participants (n = 220; aged 23–70 years) were assessed for MetS risk factors (waist circumference, blood pressure, fasting plasma glucose, triglycerides, and high-density lipoprotein cholesterol) and 5-min supine HRV (time and frequency domain and nonlinear). HRV was compared between those with 3 or more (MetS+) and those with 2 or fewer MetS risk factors (MetS–). Multiple linear regression models were built for each HRV parameter to investigate associations with MetS risk factors and HOMA-IR. Data with normal distribution are presented as means ± SD and those without as median [interquartile range]. In women, standard deviation of R–R intervals 38.0 [27.0] ms, 44.5 [29.3] ms; p = 0.020), low-frequency power (5.73 ± 1.06 ln ms2, 6.13 ± 1.05 ln ms2; p = 0.022), and the standard deviation of the length of the Poincaré plot (46.8 [31.6] ms, 58.4 [29.9] ms; p = 0.014) were lower and heart rate was higher (68 [13] beats/min, 64 [12] beats/min; p = 0. 018) in MetS+ compared with MetS–, with no differences in men. Waist circumference was most commonly associated with HRV, especially frequency domain parameters. HOMA-IR was associated with heart rate. In conclusion, MetS+ women had a less favourable HRV profile than MetS– women, but there were no differences in men. HOMA-IR was associated with heart rate, not HRV. Key words: heart rate variability, autonomic nervous system, metabolic syndrome, insulin resistance, cardiometabolic risk, diabetes prevention. Résumé : Cette étude se propose d’examiner les différences de variabilité du rythme cardiaque (« HRV ») dans le syndrome métabolique (SM) et d’établir des relations entre les paramètres de HRV, les facteurs de risque de SM et l’insulinorésistance (HOMA-IR). Les participants (n = 220; âgés de 23–70 ans) se soumettent a` l’évaluation des facteurs de risque de SM (tour de taille, pression sanguine, glucose, triglycérides et lipoprotéines de haute densité dans le plasma a` jeun) et HRV durant 5 min en position couchée (domaines temporel et fréquentiel ainsi que non linéaire). On compare HRV chez les sujets présentant 3 ou plus de facteurs de risque (SM+) par rapport aux sujets en présentant 2 ou moins (SM–). On élabore des modèles de régression linéaire multiple pour chaque paramètre de HRV afin d’examiner les associations entre les facteurs de risque de SM et d’insuline et de l’insulinorésistance selon le modèle homéostatique (« HOMA-IR »). Les données se conformant a` une distribution normale sont présentées par des moyennes ± des écarts-types et celles ne se conformant pas sont présentées par des médianes et [des écarts interquartiles]. Chez les femmes, l’écart-type des intervalles R–R (38,0 (27,0) ms, 44,5 (29,3) ms; p = 0,020), la puissance dans les basses fréquences (5,73 ± 1,06) ln ms2, 6,13 ± 1,05) ln ms2; p = 0,022) et l’écart-type de l’amplitude du graphique de Poincaré (46,8 (31,6) ms, 58,4 ± 29,9 ms; p = 0,014) sont plus bas et le rythme cardiaque plus élevé (68 ± 13 bpm, 64 ± 12 battements/min; p = 0,018) chez SM+ comparativement a` SM–, sans aucune différence chez les hommes. Le tour de taille est généralement associé a` HRV, plus particulièrement aux paramètres du domaine fréquentiel. HOMA-IR est associé au rythme cardiaque. En conclusion, les femmes SM+ présentent un profil HRV moins favorable que les femmes SM–, mais on n’observe aucune différence chez les hommes. HOMA-IR est associé au rythme cardiaque, mais pas a` HRV. [Traduit par la Rédaction] Mots-clés : variabilité du rythme cardiaque, système nerveux autonome, syndrome métabolique, insulinorésistance, risque cardiométabolique, prévention du diabète.
Introduction Heart rate variability (HRV) has been used extensively to noninvasively assess cardiac autonomic regulation. Alterations in HRV parameters, including reduced standard deviation of normal-tonormal R–R intervals (SDNN) and reductions in the HRV spectral
frequency bands, have been shown to predict cardiac and allcause mortality in patients with cardiovascular disease (Kleiger et al. 1987; La Rovere et al. 1998) and in the general population (Dekker et al. 2000; Makikallio et al. 2001). Additionally, low HRV may predict the onset of coronary heart disease in individuals with type 2 diabetes mellitus (Liao et al. 2002), but it remains
Received 27 November 2014. Accepted 26 February 2015. M.I. Stuckey. Lawson Health Research Institute, Aging Rehabilitation and Geriatric Care Research Centre, 801 Commissioners Road East, London, ON N6C 5J1, Canada; School of Kinesiology, University of Western Ontario, London, ON, Canada. A. Kiviniemi.* Verve Research, Department of Exercise and Medical Physiology, Kasarmintie 13, PO Box 404, FI-90101 Oulu, Finland. D.P. Gill. Lawson Health Research Institute, Aging Rehabilitation and Geriatric Care Research Centre, 801 Commissioners Road East, London, ON N6C 5J1, Canada; Faculty of Health Sciences, Western Centre for Public Health and Family Medicine, University of Western Ontario, London, ON N6A 3K7, Canada; School of Public Health, University of Washington, Seattle, WA 98195, USA. J.K. Shoemaker. School of Kinesiology, University of Western Ontario, London, ON N6A 3K7, Canada. R.J. Petrella. Lawson Health Research Institute, Aging Rehabilitation and Geriatric Care Research Centre, 801 Commissioners Road East, London, ON N6C 5J1, Canada; School of Kinesiology, University of Western Ontario, London, ON, Canada; Schulich School of Medicine and Dentistry, University of Western Ontario, London, ON N6A 5B9, Canada. Corresponding author: Robert J. Petrella (e-mail: [email protected]). *Present address: Medical Research Center Oulu, University of Oulu and Oulu University Hospital, PO Box 5000, FI-90014 Oulu, Finland. Appl. Physiol. Nutr. Metab. 40: 734–740 (2015) dx.doi.org/10.1139/apnm-2014-0528
Published at www.nrcresearchpress.com/apnm on 13 March 2015.
Appl. Physiol. Nutr. Metab. Downloaded from www.nrcresearchpress.com by University of Otago on 07/05/15 For personal use only.
Stuckey et al.
unclear whether HRV could provide an early marker of cardiovascular disease and type 2 diabetes mellitus risk. The metabolic syndrome (MetS) is an important clinical clustering of cardiometabolic risk factors, which increases the risk of developing cardiovascular disease and type 2 diabetes mellitus (Alberti et al. 2009); hence, alterations in HRV in this population may provide insight into potential associations between HRV and the progression of disease along the cardiometabolic risk continuum. A systematic review reported that the SDNN and both high- (HF) and low-frequency (LF) power spectral bands were generally reduced in those with compared with those without MetS (Stuckey et al. 2014). Similar differences were shown in women, but not in men, though findings were inconsistent between studies (Stuckey et al. 2014). The systematic review also reported inconsistent findings among studies that used correlation or regression to examine the relationships between individual MetS risk factors and HRV parameters. Waist circumference and blood pressure were most often associated with SDNN, the root mean square of successive differences (RMSSD), and all spectral bands, though there were sex differences (Stuckey et al. 2014). Fasting plasma glucose and triglycerides were also associated with SDNN and spectral bands in a number of studies, but high-density lipoprotein (HDL) cholesterol was only associated with frequency domain measures of HRV in 1 study (Stuckey et al. 2014). Relationships with Poincaré plot parameters have only been examined in a young population aged 18–21 years (Soares-Miranda et al. 2012). Nonlinear analysis of HRV has been used to examine the qualitative properties of heart rate (HR). Detrended fluctuation analysis short-term scaling exponent (␣1) quantifies the fractal-like characteristics of HR. Approximate entropy quantifies the complexity or regularity of time series data by calculating the likelihood that runs of patterns that are close will remain close on the next incremental comparison. These HRV indices may provide additional insight into HR dynamics have not been examined in MetS. Insulin resistance has been implicated as an important mechanism linking the clustering of MetS risk factors and autonomic dysfunction, ultimately leading to cardiovascular diseases and type 2 diabetes mellitus (Esler et al. 2001; Lambert et al. 2010). Chang et al. (2010) noted that HRV was reduced in individuals with 1 risk factor, but insulin resistance was not apparent until 2 risk factors were present. It was suggested that autonomic impairment may precede insulin resistance and, therefore, may be mechanistically linked to the development of insulin resistance and MetS. However, the relationships between HRV, insulin resistance, and MetS have not been studied. Therefore, the purpose of this study was 3-fold: first, to determine whether individuals with MetS would have abnormal linear and nonlinear HRV compared with those without MetS; second, to determine which MetS risk factors were most strongly associated with HRV parameters (both linear and nonlinear) in a population of North American adults; and third, to examine whether insulin resistance strengthened the model to explain HRV. It was hypothesized that all HRV would be reduced in MetS, MetS risk factors except for HDL cholesterol would be associated with HRV parameters, and insulin resistance would be associated with HRV parameters.
Materials and methods This study was part of a multicentre trial conducted between August 2009 and December 2011 at the Gateway Rural Health Research Institute (Seaforth, Ont., Canada) and the Laboratory for Brain and Heart Health at the University of Western Ontario (London, Ont., Canada). Participants were eligible if they were aged 18–70 years and presented with at least 1 MetS risk factor according to published guidelines (National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATPIII) 2002). Exclu-
735
sion criteria were systolic blood pressure (BP) >180 mm Hg and/or diastolic BP >110 mm Hg; type 1 diabetes; history of myocardial infarction, angioplasty, coronary artery bypass, or cerebrovascular ischemia/stroke; symptomatic congestive heart failure; atrial flutter; unstable angina; unstable pulmonary disease; use of medications known to affect HR (such as beta blockers), or use of other medication that may interfere with study objectives; second or third degree heart block; history of alcoholism, drug abuse, or other emotional cognitive or psychiatric problems; pacemaker; unstable metabolic disease; and orthopedic or rheumatologic problems that could impair the ability to exercise. In total, 224 participants provided informed consent and volunteered for this study, which was approved by the University of Western Ontario Research Ethics Board (no. 15828). Participants reported to the laboratory following an overnight fast where they were assessed for MetS risk factors. Waist circumference was measured at the midpoint between the iliac crest and last rib (Alberti et al. 2005). Supine BP was calculated as the average of the last 2 of 3 measurements taken at 1-min intervals. Measurements at Gateway Rural Health Research Institute were automated (BpTRU, VSM MedTech Ltd., Coquitlam, B.C., Canada), and manual measurements were completed at the Laboratory for Brain and Heart Health. Blood was drawn and sent to a central laboratory for fasting plasma glucose, triglycerides, HDL cholesterol, and insulin analysis. Those presenting with 3 or more risk factors were categorized as with MetS (MetS+), and those with 2 or fewer risk factors were considered without MetS (MetS–) (National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATPIII) 2002). The homeostasis model assessment for insulin resistance (HOMA-IR) was calculated with standard methods (Matthews et al. 1985). Following a light, standardized snack, participants were instrumented for a lead II electrocardiogram. A respiratory belt (Pneumotrace II, ADInstruments, Colorado Springs, Colo., USA) was secured around the thorax for collection of respiratory rate. R–R intervals were collected during 10 min of supine rest. External stimuli, such as light and noise, were controlled to ensure signal stability. Participants were instructed to remain awake and still. All measures were sampled at 1000 Hz, input into a data acquisition board (PowerLab ML795, ADInstruments) for analog-to-digital signal conversion with LabChart7Pro software (ADInstruments), and stored for offline analysis. LabChart files were converted to text files for analysis with HRV software (Hearts version 7, Heart Signal Co., Oulu, Finland). A predictive stepping test was used to estimate maximal oxygen uptake (V˙O2max) (Stuckey et al. 2012). Briefly, participants stepped up and down a set of two 20-cm stairs, 20 times, at a pace considered normal. Age, sex, body weight, radial pulse measured immediately upon completing the step test, and time to complete test were entered into the predictive equation. HRV analysis Editing of the HR time series was performed by a single investigator. All electrocardiograms were manually scanned for ectopic or non-sinus beats, which were deleted from the time series. 90% of data was needed for inclusion. Time domain HRV analyses included HR, SDNN, and RMSSD. The HRV spectrum was computed with the nonparametric fast Fourier transform method. LF (0.04– 0.15 Hz), HF (0.15–0.4 Hz), and LF/HF were examined as well as both LF and HF expressed in normalized units (LFnu = LF/(LF + HF); HFnu = HF/(LF + HF)). A Poincaré plot was formed by plotting each R–R interval against the following one to create a scatter plot. The standard deviation of the width (SD1) and length (SD2) were calculated. The detrended fluctuation analysis method was used to examine fractal characteristics of HR fluctuations. The root-mean square fluctuations of integrated and detrended data were measured in observation windows and then plotted against the size of the window on a log-log scale. The ␣1 was calculated from the slope of the line (from 4–11 beats). Approximate entropy quantiPublished by NRC Research Press
736
Appl. Physiol. Nutr. Metab. Vol. 40, 2015
Table 1. Participant characteristics.
Appl. Physiol. Nutr. Metab. Downloaded from www.nrcresearchpress.com by University of Otago on 07/05/15 For personal use only.
n Male Female Age (y) Male Female WC (cm) Male Female SBP (mmHg) Male Female DBP (mmHg) Male Female FPG (mmol/L) Male Female TG (mmol/L) Male Female HDL (mmol/L) Male Female V˙O2max (mL/(kg·min)−1 Male Female HOMA-IR Male Female Breathing frequency Male Female
MetS−
MetS+
p
95 21 74 58 [12] 57 [12] 58 [11] 103.3±14.6 105.6±13.0 102.6±15.0 131 [25] 135 [22] 130 [23] 80±11 84±11 79 [11] 5.1 [0.8] 5.4 [0.5] 5.0 [0.6] 1.01 [0.56] 1.13 [0.63] 1.00 [0.31] 1.57 [0.50] 1.21 [0.30] 1.65 [0.42] 30.93±6.31 35.78±4.17 29.60±6.17 1.56 [1.19] 1.77 [0.58] 1.33 [1.23] 13.4±3.3 13.3±3.1 13.5±3.3
129 41 88 59 [10] 59 [9] 59 [11] 107.3±11.1 109.0±10.6 106.5±11.3 135 [17] 134 [15] 138 [20] 85±11 86±10 84 [13] 5.3 [1.2] 5.4 [1.2] 5.3 [1.2] 1.75 [0.98] 1.77 [1.08] 1.75 [0.93] 1.12 [0.36] 0.96 [0.21] 1.19 [0.31] 30.35±5.87 34.88±5.23 28.26±4.91 2.80 [2.47] 2.95 [2.59] 2.62 [2.41] 14.1±3.4 13.9±3.4 14.3±3.4
0.78 0.45 0.89 0.04* 0.31 0.09 0.02* 0.91 0.01* 0.008* 0.60 0.009* 0.02* 0.73 0.02*
ARTICLE Associations between heart rate variability, metabolic syndrome risk factors, and insulin resistance Appl. Physiol. Nutr. Metab. Downloaded from www.nrcresearchpress.com by University of Otago on 07/05/15 For personal use only.
Melanie I. Stuckey, Antti Kiviniemi, Dawn P. Gill, J. Kevin Shoemaker, and Robert J. Petrella
Abstract: The purpose of this study was to examine differences in heart rate variability (HRV) in metabolic syndrome (MetS) and to determine associations between HRV parameters, MetS risk factors, and insulin resistance (homeostasis model assessment for insulin resistance (HOMA-IR)). Participants (n = 220; aged 23–70 years) were assessed for MetS risk factors (waist circumference, blood pressure, fasting plasma glucose, triglycerides, and high-density lipoprotein cholesterol) and 5-min supine HRV (time and frequency domain and nonlinear). HRV was compared between those with 3 or more (MetS+) and those with 2 or fewer MetS risk factors (MetS–). Multiple linear regression models were built for each HRV parameter to investigate associations with MetS risk factors and HOMA-IR. Data with normal distribution are presented as means ± SD and those without as median [interquartile range]. In women, standard deviation of R–R intervals 38.0 [27.0] ms, 44.5 [29.3] ms; p = 0.020), low-frequency power (5.73 ± 1.06 ln ms2, 6.13 ± 1.05 ln ms2; p = 0.022), and the standard deviation of the length of the Poincaré plot (46.8 [31.6] ms, 58.4 [29.9] ms; p = 0.014) were lower and heart rate was higher (68 [13] beats/min, 64 [12] beats/min; p = 0. 018) in MetS+ compared with MetS–, with no differences in men. Waist circumference was most commonly associated with HRV, especially frequency domain parameters. HOMA-IR was associated with heart rate. In conclusion, MetS+ women had a less favourable HRV profile than MetS– women, but there were no differences in men. HOMA-IR was associated with heart rate, not HRV. Key words: heart rate variability, autonomic nervous system, metabolic syndrome, insulin resistance, cardiometabolic risk, diabetes prevention. Résumé : Cette étude se propose d’examiner les différences de variabilité du rythme cardiaque (« HRV ») dans le syndrome métabolique (SM) et d’établir des relations entre les paramètres de HRV, les facteurs de risque de SM et l’insulinorésistance (HOMA-IR). Les participants (n = 220; âgés de 23–70 ans) se soumettent a` l’évaluation des facteurs de risque de SM (tour de taille, pression sanguine, glucose, triglycérides et lipoprotéines de haute densité dans le plasma a` jeun) et HRV durant 5 min en position couchée (domaines temporel et fréquentiel ainsi que non linéaire). On compare HRV chez les sujets présentant 3 ou plus de facteurs de risque (SM+) par rapport aux sujets en présentant 2 ou moins (SM–). On élabore des modèles de régression linéaire multiple pour chaque paramètre de HRV afin d’examiner les associations entre les facteurs de risque de SM et d’insuline et de l’insulinorésistance selon le modèle homéostatique (« HOMA-IR »). Les données se conformant a` une distribution normale sont présentées par des moyennes ± des écarts-types et celles ne se conformant pas sont présentées par des médianes et [des écarts interquartiles]. Chez les femmes, l’écart-type des intervalles R–R (38,0 (27,0) ms, 44,5 (29,3) ms; p = 0,020), la puissance dans les basses fréquences (5,73 ± 1,06) ln ms2, 6,13 ± 1,05) ln ms2; p = 0,022) et l’écart-type de l’amplitude du graphique de Poincaré (46,8 (31,6) ms, 58,4 ± 29,9 ms; p = 0,014) sont plus bas et le rythme cardiaque plus élevé (68 ± 13 bpm, 64 ± 12 battements/min; p = 0,018) chez SM+ comparativement a` SM–, sans aucune différence chez les hommes. Le tour de taille est généralement associé a` HRV, plus particulièrement aux paramètres du domaine fréquentiel. HOMA-IR est associé au rythme cardiaque. En conclusion, les femmes SM+ présentent un profil HRV moins favorable que les femmes SM–, mais on n’observe aucune différence chez les hommes. HOMA-IR est associé au rythme cardiaque, mais pas a` HRV. [Traduit par la Rédaction] Mots-clés : variabilité du rythme cardiaque, système nerveux autonome, syndrome métabolique, insulinorésistance, risque cardiométabolique, prévention du diabète.
Introduction Heart rate variability (HRV) has been used extensively to noninvasively assess cardiac autonomic regulation. Alterations in HRV parameters, including reduced standard deviation of normal-tonormal R–R intervals (SDNN) and reductions in the HRV spectral
frequency bands, have been shown to predict cardiac and allcause mortality in patients with cardiovascular disease (Kleiger et al. 1987; La Rovere et al. 1998) and in the general population (Dekker et al. 2000; Makikallio et al. 2001). Additionally, low HRV may predict the onset of coronary heart disease in individuals with type 2 diabetes mellitus (Liao et al. 2002), but it remains
Received 27 November 2014. Accepted 26 February 2015. M.I. Stuckey. Lawson Health Research Institute, Aging Rehabilitation and Geriatric Care Research Centre, 801 Commissioners Road East, London, ON N6C 5J1, Canada; School of Kinesiology, University of Western Ontario, London, ON, Canada. A. Kiviniemi.* Verve Research, Department of Exercise and Medical Physiology, Kasarmintie 13, PO Box 404, FI-90101 Oulu, Finland. D.P. Gill. Lawson Health Research Institute, Aging Rehabilitation and Geriatric Care Research Centre, 801 Commissioners Road East, London, ON N6C 5J1, Canada; Faculty of Health Sciences, Western Centre for Public Health and Family Medicine, University of Western Ontario, London, ON N6A 3K7, Canada; School of Public Health, University of Washington, Seattle, WA 98195, USA. J.K. Shoemaker. School of Kinesiology, University of Western Ontario, London, ON N6A 3K7, Canada. R.J. Petrella. Lawson Health Research Institute, Aging Rehabilitation and Geriatric Care Research Centre, 801 Commissioners Road East, London, ON N6C 5J1, Canada; School of Kinesiology, University of Western Ontario, London, ON, Canada; Schulich School of Medicine and Dentistry, University of Western Ontario, London, ON N6A 5B9, Canada. Corresponding author: Robert J. Petrella (e-mail: [email protected]). *Present address: Medical Research Center Oulu, University of Oulu and Oulu University Hospital, PO Box 5000, FI-90014 Oulu, Finland. Appl. Physiol. Nutr. Metab. 40: 734–740 (2015) dx.doi.org/10.1139/apnm-2014-0528
Published at www.nrcresearchpress.com/apnm on 13 March 2015.
Appl. Physiol. Nutr. Metab. Downloaded from www.nrcresearchpress.com by University of Otago on 07/05/15 For personal use only.
Stuckey et al.
unclear whether HRV could provide an early marker of cardiovascular disease and type 2 diabetes mellitus risk. The metabolic syndrome (MetS) is an important clinical clustering of cardiometabolic risk factors, which increases the risk of developing cardiovascular disease and type 2 diabetes mellitus (Alberti et al. 2009); hence, alterations in HRV in this population may provide insight into potential associations between HRV and the progression of disease along the cardiometabolic risk continuum. A systematic review reported that the SDNN and both high- (HF) and low-frequency (LF) power spectral bands were generally reduced in those with compared with those without MetS (Stuckey et al. 2014). Similar differences were shown in women, but not in men, though findings were inconsistent between studies (Stuckey et al. 2014). The systematic review also reported inconsistent findings among studies that used correlation or regression to examine the relationships between individual MetS risk factors and HRV parameters. Waist circumference and blood pressure were most often associated with SDNN, the root mean square of successive differences (RMSSD), and all spectral bands, though there were sex differences (Stuckey et al. 2014). Fasting plasma glucose and triglycerides were also associated with SDNN and spectral bands in a number of studies, but high-density lipoprotein (HDL) cholesterol was only associated with frequency domain measures of HRV in 1 study (Stuckey et al. 2014). Relationships with Poincaré plot parameters have only been examined in a young population aged 18–21 years (Soares-Miranda et al. 2012). Nonlinear analysis of HRV has been used to examine the qualitative properties of heart rate (HR). Detrended fluctuation analysis short-term scaling exponent (␣1) quantifies the fractal-like characteristics of HR. Approximate entropy quantifies the complexity or regularity of time series data by calculating the likelihood that runs of patterns that are close will remain close on the next incremental comparison. These HRV indices may provide additional insight into HR dynamics have not been examined in MetS. Insulin resistance has been implicated as an important mechanism linking the clustering of MetS risk factors and autonomic dysfunction, ultimately leading to cardiovascular diseases and type 2 diabetes mellitus (Esler et al. 2001; Lambert et al. 2010). Chang et al. (2010) noted that HRV was reduced in individuals with 1 risk factor, but insulin resistance was not apparent until 2 risk factors were present. It was suggested that autonomic impairment may precede insulin resistance and, therefore, may be mechanistically linked to the development of insulin resistance and MetS. However, the relationships between HRV, insulin resistance, and MetS have not been studied. Therefore, the purpose of this study was 3-fold: first, to determine whether individuals with MetS would have abnormal linear and nonlinear HRV compared with those without MetS; second, to determine which MetS risk factors were most strongly associated with HRV parameters (both linear and nonlinear) in a population of North American adults; and third, to examine whether insulin resistance strengthened the model to explain HRV. It was hypothesized that all HRV would be reduced in MetS, MetS risk factors except for HDL cholesterol would be associated with HRV parameters, and insulin resistance would be associated with HRV parameters.
Materials and methods This study was part of a multicentre trial conducted between August 2009 and December 2011 at the Gateway Rural Health Research Institute (Seaforth, Ont., Canada) and the Laboratory for Brain and Heart Health at the University of Western Ontario (London, Ont., Canada). Participants were eligible if they were aged 18–70 years and presented with at least 1 MetS risk factor according to published guidelines (National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATPIII) 2002). Exclu-
735
sion criteria were systolic blood pressure (BP) >180 mm Hg and/or diastolic BP >110 mm Hg; type 1 diabetes; history of myocardial infarction, angioplasty, coronary artery bypass, or cerebrovascular ischemia/stroke; symptomatic congestive heart failure; atrial flutter; unstable angina; unstable pulmonary disease; use of medications known to affect HR (such as beta blockers), or use of other medication that may interfere with study objectives; second or third degree heart block; history of alcoholism, drug abuse, or other emotional cognitive or psychiatric problems; pacemaker; unstable metabolic disease; and orthopedic or rheumatologic problems that could impair the ability to exercise. In total, 224 participants provided informed consent and volunteered for this study, which was approved by the University of Western Ontario Research Ethics Board (no. 15828). Participants reported to the laboratory following an overnight fast where they were assessed for MetS risk factors. Waist circumference was measured at the midpoint between the iliac crest and last rib (Alberti et al. 2005). Supine BP was calculated as the average of the last 2 of 3 measurements taken at 1-min intervals. Measurements at Gateway Rural Health Research Institute were automated (BpTRU, VSM MedTech Ltd., Coquitlam, B.C., Canada), and manual measurements were completed at the Laboratory for Brain and Heart Health. Blood was drawn and sent to a central laboratory for fasting plasma glucose, triglycerides, HDL cholesterol, and insulin analysis. Those presenting with 3 or more risk factors were categorized as with MetS (MetS+), and those with 2 or fewer risk factors were considered without MetS (MetS–) (National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATPIII) 2002). The homeostasis model assessment for insulin resistance (HOMA-IR) was calculated with standard methods (Matthews et al. 1985). Following a light, standardized snack, participants were instrumented for a lead II electrocardiogram. A respiratory belt (Pneumotrace II, ADInstruments, Colorado Springs, Colo., USA) was secured around the thorax for collection of respiratory rate. R–R intervals were collected during 10 min of supine rest. External stimuli, such as light and noise, were controlled to ensure signal stability. Participants were instructed to remain awake and still. All measures were sampled at 1000 Hz, input into a data acquisition board (PowerLab ML795, ADInstruments) for analog-to-digital signal conversion with LabChart7Pro software (ADInstruments), and stored for offline analysis. LabChart files were converted to text files for analysis with HRV software (Hearts version 7, Heart Signal Co., Oulu, Finland). A predictive stepping test was used to estimate maximal oxygen uptake (V˙O2max) (Stuckey et al. 2012). Briefly, participants stepped up and down a set of two 20-cm stairs, 20 times, at a pace considered normal. Age, sex, body weight, radial pulse measured immediately upon completing the step test, and time to complete test were entered into the predictive equation. HRV analysis Editing of the HR time series was performed by a single investigator. All electrocardiograms were manually scanned for ectopic or non-sinus beats, which were deleted from the time series. 90% of data was needed for inclusion. Time domain HRV analyses included HR, SDNN, and RMSSD. The HRV spectrum was computed with the nonparametric fast Fourier transform method. LF (0.04– 0.15 Hz), HF (0.15–0.4 Hz), and LF/HF were examined as well as both LF and HF expressed in normalized units (LFnu = LF/(LF + HF); HFnu = HF/(LF + HF)). A Poincaré plot was formed by plotting each R–R interval against the following one to create a scatter plot. The standard deviation of the width (SD1) and length (SD2) were calculated. The detrended fluctuation analysis method was used to examine fractal characteristics of HR fluctuations. The root-mean square fluctuations of integrated and detrended data were measured in observation windows and then plotted against the size of the window on a log-log scale. The ␣1 was calculated from the slope of the line (from 4–11 beats). Approximate entropy quantiPublished by NRC Research Press
736
Appl. Physiol. Nutr. Metab. Vol. 40, 2015
Table 1. Participant characteristics.
Appl. Physiol. Nutr. Metab. Downloaded from www.nrcresearchpress.com by University of Otago on 07/05/15 For personal use only.
n Male Female Age (y) Male Female WC (cm) Male Female SBP (mmHg) Male Female DBP (mmHg) Male Female FPG (mmol/L) Male Female TG (mmol/L) Male Female HDL (mmol/L) Male Female V˙O2max (mL/(kg·min)−1 Male Female HOMA-IR Male Female Breathing frequency Male Female
MetS−
MetS+
p
95 21 74 58 [12] 57 [12] 58 [11] 103.3±14.6 105.6±13.0 102.6±15.0 131 [25] 135 [22] 130 [23] 80±11 84±11 79 [11] 5.1 [0.8] 5.4 [0.5] 5.0 [0.6] 1.01 [0.56] 1.13 [0.63] 1.00 [0.31] 1.57 [0.50] 1.21 [0.30] 1.65 [0.42] 30.93±6.31 35.78±4.17 29.60±6.17 1.56 [1.19] 1.77 [0.58] 1.33 [1.23] 13.4±3.3 13.3±3.1 13.5±3.3
129 41 88 59 [10] 59 [9] 59 [11] 107.3±11.1 109.0±10.6 106.5±11.3 135 [17] 134 [15] 138 [20] 85±11 86±10 84 [13] 5.3 [1.2] 5.4 [1.2] 5.3 [1.2] 1.75 [0.98] 1.77 [1.08] 1.75 [0.93] 1.12 [0.36] 0.96 [0.21] 1.19 [0.31] 30.35±5.87 34.88±5.23 28.26±4.91 2.80 [2.47] 2.95 [2.59] 2.62 [2.41] 14.1±3.4 13.9±3.4 14.3±3.4
0.78 0.45 0.89 0.04* 0.31 0.09 0.02* 0.91 0.01* 0.008* 0.60 0.009* 0.02* 0.73 0.02*
