Key words diabetes - hypertension - obesity
Abbreviations
τ: time constant
ACE: angiotensin-converting enzyme
BMI: body mass index
CCFlag
: lag between maximum of autocorrelation function and cross-correlation function
CCFmax
: maximum in cross-correlation function
Q’: cardiac output
ECG: electrocardiography
HR: heart rate
HRmax
: maximal heart rate
mBP: mean arterial blood pressure
PRBSs: pseudo-random binary sequences
SV: stroke volume
T2DM: type 2 diabetes mellitus
T2D: type 2 diabetes mellitus patients without treatment with cardiovascular medication
T2Dc
: type 2 diabetes mellitus patients treated with centrally acting medication
T2Dp
: type 2 diabetes mellitus patients treated with peripherally acting medication
V’O2 max: maximal oxygen uptake
V’O2 musc: muscular oxygen uptake
V’O2 pulm: pulmonary oxygen uptake
WR: work rate
WRmax
: maximal work rate
Introduction
Type 2 diabetes mellitus (T2DM) is associated with increased cardiovascular morbidity
and mortality. This presents a major challenge to healthcare systems for several aspects
[1 ]
[2 ]
[3 ]
[4 ]. T2DM is often accompanied by comorbidities such as arterial hypertension [5 ]
[6 ], lipid metabolism disorders [7 ], overweight or obesity [8 ]. This cluster of diseases is termed the metabolic syndrome [4 ].
Centrally acting beta blockers, affecting directly the sympathetic nervous system
to decrease heart rate (HR) and consequently cardiac output (Q’), or peripherally
acting drugs as angiotensin-converting enzyme (ACE) inhibitors or calcium-channel
blockers, influencing the vascular tone, are commonly used to control arterial hypertension
in patients with T2DM.
There are indices, that both types of drugs can improve the responses of the cardiovascular,
respiratory and metabolic system to changing work rates (WR) at submaximal exercise
intensities [9 ]
[10 ]
[11 ]
[12 ], which can be described with oxygen uptake (V’O2 ) kinetics. V’O2 kinetics give information on the adjustment of the cardiovascular and respiratory
and metabolic system and therefore aspects of transport and metabolic processes to
changes in WR. Slower V’O2 kinetics are associated with lower exercise tolerance [13 ]. Using a circulatory model, considering venous volume and the V’O2 as well as perfusion of the non-working part of the body, muscular V’O2 (V’O2 musc) kinetics can be estimated from pulmonary V’O2 (V’O2 pulm) and HR [14 ]. This allows for a more detailed analysis of metabolic and circulatory processes.
To the best of our knowledge, no data are available regarding HR kinetics in patients
taking cardiovascular drugs. Although faster HR kinetics as indicators for kinetics
of Q’ have been considered as a potentially influencing factor for faster V’O2 pulm kinetics [10 ]
[11 ], they were not yet measured in this context. It was shown that beta blockers increase
RR interval variability and vagal tone in patients with former uncomplicated myocardial
infarction [15 ].
Apparently, no data have been published that show the influence of different cardiovascular
drugs on HR and V’O2 kinetics in patients with T2DM. Subjects with medication were either excluded from
the analysis [16 ]
[17 ]
[18 ]
[19 ], no detailed information was provided [20 ], or patients were included in the study (except patients taking beta blockers) but
not analyzed seperately [21 ]
[22 ].
The aim of the present pilot study is to investigate differences in HR and V’O2 kinetics between T2DM patients, considering their cardiovascular medication.
The following hypotheses were tested:
1) The kinetics responses of V’O2 musc and HR are faster in T2DM patients taking centrally acting medication compared
with T2DM patients not taking cardiovascular medication.
2) T2DM patients taking mainly peripherally acting medication show faster V’O2 musc and HR kinetics compared with T2DM patients not taking cardiovascular medication.
Materials and Methods
Subjects
18 male subjects participated in the study. All subjects declared that they were not
diagnosed with diabetic nephropathy, retinopathy, neuropathy, and/or other cardiovascular
complications other than arterial hypertension. None of the subjects performed regular
physical activity, and no contraindications for participation in exercise testing
were evident. The subjects were selected for 3 subgroups, according to their medication:
6 patients were diagnosed with T2DM and did not take any cardiovascular medication,
6 subjects were diagnosed with T2DM and took mainly centrally acting antihypertensive
medication (T2Dc , 2 of these subjects had dyslipidemia); and 6 patients were T2DM patients treated
with mainly peripherally acting drugs (T2Dp , one subject had dyslipidemia). One subject taking the calcium-channel-blocker ‘verapamil’
was included in the T2Dc group, because this drug is known to act at heart level. Anthropometric data and
differentiation of the subgroups are specified in [Table 1 ].
Table 1 Anthropometric data of all subjects divided into subgroups.
Group (N=18)
Age [years]
BMI [kg·m−2 ]
Group of cardiovascular medication
Cardiovascular agent
Duration of type 2 diabetes mellitus since diagnosis [years]
T2D (n=6)
Mean
60
33.0
−
−
3.5
SD
8
5.9
3.1
T2D
p (n=6)
Mean
56
32.8
ACE inhibitors, angiotensin 1 blockers, calcium-channel blockers
Ramipril, enalapril, irbesartan, amlodipine
8.0
SD
10
4.0
7.2
T2D
c (n=6)
Mean
61
32.6
ß-blockers or any combination of ß-blockers and other antihypertensive drugs
Bisoprolol, verapamil, combination of bisoprolol or verapamil with other antihypertensive
drugs
4.3
SD
9
8.0
3.9
BMI: body mass index; SD: standard deviation; T2D: subjects with type 2 diabetes mellitus
not taking additional medication; T2Dp : subjects with type 2 diabetes mellitus taking peripherally acting medication; T2Dc : subjects with type 2 diabetes mellitus taking centrally acting medication
Considering the subjects’ anti-diabetic treatment, 4 of the T2D subjects took metformin
and 2 did not take any medication. In the T2Dc group 3 subjects took metformin, one took sitagliptin and 2 did not take anti-diabetic
drugs. 5 of the T2Dp subjects took metformin and one did not take any anti-diabetic medication. Subjects
visited the laboratory twice: The first time, anthropometric measurements, resting
electrocardiography (ECG), and a V’O2 max test were performed. Given that no contraindications in the ECG and during V’O2 max test were identified, the subjects returned to the laboratory a second time for
a cardiorespiratory kinetics test.
A positive vote of the ethics committee of the German Sport University Cologne, in
accordance with the Declaration of Helsinki (1964 including the amendments until 2013),
was available before the beginning of the tests. All subjects gave their written informed
consent prior to the testing procedures.
V’O2 max testing
The subjects were tested using the protocol recommended by the World Health Organization
in a seated position on a cycle ergometer (Ergoline ER 900, Ergoline GmbH, Bitz, Germany).
WR was increased by 25 W every 2 min until subjective exhaustion or the occurrence
of one of the common reasons for test termination (e. g., ST segment depression or
couplets of premature heart contractions).
HR was measured continuously via 12-lead ECG (GE Medical Systems, Information Technologies,
Munich, Germany) and recorded by an AMEDTEC ECGpro® V.3.66 (MedizintechnikAue GmbH, Aue, Germany). Pulmonary data were assessed breath
by breath via a ZAN 600 (ZAN Messgeräte GmbH, Oberthulba, Germany) including the algorithms
of Beaver et al. [23 ]. All instruments were calibrated according to the manufacturer’s suggestions before
all tests. For maximal oxygen uptake (V’O2 max) the highest 30 s averaged value of the highest achieved WR was determined as
the maximum value. The Achievement of true V’O2 max was assumed, if a plateau in V’O2 (increase in V’O2 ≤2.1 ml kg−1 ·min−1 ) despite an increase in WR (as the primary criterion) appeared. When no plateau occurred,
V’O2 max was assumed when HRmax was higher than 200 beats min−1 minus the years of age [24 ] and the maximal respiratory exchange ratio was not lower than 1.06 [25 ]
[26 ]. All subjects included in this study achieved V’O2 max according to the predefined criteria, which was then normalized to body mass.
Cardiorespiratory kinetics test
Subjects were tested on a semi-recumbent cycle ergometer (Cardiac Stress Table, Lode
B.V., Netherlands; backrest at 45°, legs at 42°, relative to ground level). Pseudo-random
binary sequences (PRBSs) were used as the WR protocol. The protocol consisted of 180 s
of rest; 200 s of 30 W, as low steady state (Low); followed by two 300 s periods of
PRBS (PRBS1 and PRBS2), with changing WRs between 30 and 80 W; and ended with 200 s
of 80 W, as high steady state (High) ([Fig. 1a ]). The cadence was maintained at 60 rpm. HR was assessed beat to beat via electrocardiography;
stroke volume (SV) was measured beat to beat via impedance cardiography (Task Force® Monitor, CNSystems Medizintechnik AG, Graz, Austria). Pulmonary gas exchange data
were determined breath by breath (ZAN 680, ZAN Meßgeräte GmbH, Oberthulba, Germany),
incorporating the algorithms of Beaver et al. [23 ]. From SV and HR, Q’ was calculated. The instruments were calibrated before each
measurement, according to the manufacturer’s guidelines. For reduction of noise, data
were filtered with a low-pass filter (0.1 Hz). Data were synchronized via trigger
signals and interpolated to 1 s intervals for homogeneous sampling [27 ].
Fig. 1 Demonstration of data acquisition and analysis. a : Data acquisition during the work rate protocol; b : Data after time series analysis. The arrows indicate the respective maximum of the
cross-correlation course (CCFmax ). Lag: lag of cross-correlation function; ACF: autocorrelation function; CCF: cross-correlation
function; HR: heart rate; V’O2 musc: muscular oxygen uptake; V’O2 pulm: pulmonary oxygen uptake; Rest: resting period; Low: 30 W constant phase; PRBS:
pseudo-random binary sequence; High: 80 W constant phase; Recovery: recovery phase.
For analysis of cardiorespiratory kinetics, time series analysis was applied and V’O2 musc kinetics were estimated from HR and V’O2 pulm [14 ]. Briefly, the PRBS WR protocol was auto-correlated, which resulted in a triangular
shape and each parameter was cross-correlated with the WR protocol ([Fig. 1b ]). The autocorrelation can be approximated as a WR impulse. The cross-correlation
function was interpreted as the response of the respective parameter to this impulse.
The kinetics of the parameters were summarized by the maximum in cross-correlation
function (CCFmax, compare [Fig. 1b ]) and the related lag (CCFlag ). Higher CCFmax indicate faster response times of the particular parameter. From CCFmax , the time constant τ can be estimated. Further, V’O2 musc and the corresponding kinetics were calculated using the backward calculation
method. This method is based on a circulatory model with 2 compartments (working and
remainder part). V’O2 musc was estimated considering a certain venous blood volume between muscle and mouth,
as well as V’O2, and perfusion of the remainder of the body (see [14 ] for further details on the method). This method makes it possible to distinguish
between V’O2 musc and V’O2 pulm, which leads to a more detailed analysis of the cardiorespiratory and metabolic
regulation considering transport processes. For kinetics comparisons, V’O2 musc and HR kinetics have been considered.
Statistical analysis
Between-group comparisons for the factors ‘CCFmax (V’O2 musc)’ and ‘CCFmax (HR)’ (‘Parameter x Group’) were performed via 2-factorial ANOVA. The following post
hoc comparisons were implemented via LSD test. For the means of HR, mean arterial
blood pressure (mBP), V’O2 pulm, V’O2 musc, SV and Q’ during the different steps (Low, PRBS1, PRBS2, High) of the PRBS protocol,
2-factorial ANOVA (‘Step x Group’) were applied. Since each group included only 6
subjects, Kruskal-Wallis tests were used to compare V’O2 max, HRmax , maximal WR (WRmax ), body mass index (BMI), age, resting mBP, fasting blood glucose and glycosylated
hemoglobin (HbA1c ) between the groups. When applicable, post hoc tests were adjusted via Bonferroni
correction.
Results
Anthropometric data as well as cardiorespiratory and metabolic capacities are shown
in [Table 2 ].
Table 2 Means and standard deviations of anthropometric and glycemic data, and parameters
of cardiorespiratory capacities.
Group (N=18)
V’O2 max [ml·min−1 ·kg−1 ]
WRmax [Watt]
HRmax [min−1 ]
Fasting blood glucose [mg·dl−1 ]
HbA1c [% (mmol·mol−1 )]
Resting mBP [mmHg]
T2D (n=6)
Mean
21.3
146
146
160
7.1 (54)
108
SD
5.2
25
12
81
1.9 (21)
7
TD2
p (n=6)
Mean
23.0
158
153
150
7.2 (56)
105
SD
6.7
34
25
29
1 (11)
7
TD2
c (n=6)
Mean
19.2
133
131
140
6.4 (47)
97
SD
5.1
13
17
24
1.4 (15)
9
V’O2 max: maximum oxygen uptake; WRmax : maximum work rate; HRmax : maximum heart rate; HbA1c: glycosylated hemoglobin; mBP: mean arterial blood pressure;
T2D: subjects with type 2 diabetes mellitus not taking additional medication; T2Dc : subjects with type 2 diabetes mellitus taking centrally acting medication; T2Dp : subjects with type 2 diabetes mellitus taking peripherally acting medication
No significant differences between groups were observed regarding V’O2 max, WRmax , HRmax , resting mBP, fasting blood glucose or HbA1c .
For the absolute values of HR during the WR protocol, a significant between-group
effect was found (p=0.01). T2Dc was significantly different to both T2D (p=0.014) and T2Dp (p=0.005), which did not differ. These group differences for HR were found for all
analyzed phases. For all other parameters, no statistical group differences were found,
as listed in [Table 3 ].
Table 3 Analyses of cardiorespiratory parameters of all WR steps.
Group (N=18)
WR step
HR [min−1 ]
V’O2 pulm [L·min−1 ]
V’O2 musc [L·min−1 ]
mBP [mmHg]
SV [mL]
Q’ [L·min−1 ]
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
T2D (n=6)
Low
96
8
1.0
0.1
0.7
0.1
115
11
85.0
24.8
8.1
1.8
PRBS1
103
8
1.3
0.1
0.9
0.1
117
10
86.8
21.7
8.7
2.0
PRBS2
105
9
1.4
0.1
1.0
0.1
116
6
86.2
19.4
8.9
2.0
High
116
10
1.6
0.2
1.2
0.2
122
11
84.4
12.6
9.8
2.4
T2D
p (n=6)
Low
98
10
1.1
0.1
0.7
0.1
112
8
85.7
28.0
8.4
2.0
PRBS1
106
11
1.3
0.1
1.0
0.1
114
10
86.7
26.7
9.1
2.1
PRBS2
108
11
1.3
0.1
1.0
0.1
112
10
89.7
32.6
9.2
2.1
High
119
15
1.6
0.1
1.3
0.1
117
12
90.8
35.4
10.2
2.2
T2D
c (n=6)
Low
80*
12
1.1
0.2
0.7
0.1
106
9
101.6
16.5
8.1
1.9
PRBS1
86*
12
1.3
0.2
0.9
0.1
111
7
99.7
12.2
8.6
1.9
PRBS2
88*
11
1.3
0.2
1.0
0.1
111
8
98.1
14.3
8.8
2.0
High
96*
12
1.6
0.2
1.2
0.1
114
8
98.0
22.3
9.7
2.1
HR: heart rate; V’O2 pulm: pulmonary oxygen uptake; V’O2 musc: muscular oxygen uptake; mBP: mean arterial blood pressure; SV: stroke volume;
Q’: cardiac output; SD: standard deviation; T2D: subjects with type 2 diabetes mellitus
not taking additional medication; T2Dc : subjects with type 2 diabetes mellitus taking centrally acting medication; T2Dp : subjects with type 2 diabetes mellitus taking peripherally acting medication; Low:
30 W low steady state; PRBS1: 53.3 W, mean value of first pseudo-random binary sequence;
PRBS2: 53.3 W, mean value of second pseudo-random binary sequence; High: 80 W high
steady state; WR: work rate; *indicates significantly different from T2D and T2Dp (p<0.05)
Static linearity, as a prerequisite for the application of time series analysis, was
analyzed and proved for all groups. The respective regression functions for HR and
V’O2 pulm (during Low, PRBS1, PRBS2 and High; n=4) for each group are shown in [Table 4 ].
Table 4 Static linearity for HR (heart rate) and V’O2 pulm (pulmonary oxygen uptake) during the PRBS WR protocol.
Group (N=18)
HR
V’O2 pulm
Slope
Intercept
R2
Slope
Intercept
R2
T2D (n=6)
0.40
83.46
0.98
0.017
0.7
0.99
TD2
p (n=6)
0.43
84.76
0.99
0.011
0.74
0.99
TD2
c (n=6)
0.32
70.43
0.98
0.011
0.71
0.99
HR: heart rate; V’O2 pulm: pulmonary oxygen uptake; T2D: subjects with type 2 diabetes mellitus not taking
additional medication; T2Dp : subjects with type 2 diabetes mellitus taking peripherally acting medication; T2Dc : subjects with type 2 diabetes mellitus taking centrally acting medication
ANOVA Parameter x Group with repeated measures on CCFmax (HR) and CCFmax (V’O2 musc), presented in [Fig. 2 ], showed a significant main effect on the Parameter x Group (p=0.004, partial; ŋ2 =0.515) and Parameter (p=0.005, partial; ŋ2 =0.424). Post hoc tests revealed significant differences between CCFmax (V’O2 musc) and CCFmax (HR) for T2D (p=0.001) and T2Dp (p=0.009) but not for T2Dc . Between groups, no significant differences were identified following ANOVA.
Fig. 2 Means and standard errors for CCFmax of HR and V’O2 in the T2D, T2Dc and T2Dp group. CCFmax (V’O2 musc) was significantly different from CCFmax (HR) in T2D (p=0.001) and T2Dp (p=0.009) but T2Dc was not. CCFmax : maximum of cross-correlation function; CCFlag : lag of cross-correlation function; HR: heart rate; V’O2 musc: muscular oxygen uptake. *Significantly different.
For comparisons with data from other publications, CCFmax values of V’O2 musc, V’O2 pulm and HR were converted into time constants (τ) (see [14 ]). These time constants should be regarded as rough estimates, since they were obtained
from CCFmax values [Table 5 ] .
Table 5 Time constants of HR, V’O2 musc and V’O2 pulm kinetics and model parameters.
Group (N=18)
τHR
τV’O2 musc
τV’O2 pulm
Vv
Q'rem
V’O2 rem
[s]
[s]
[s]
[ml·min−1 ]
[ml·min−1 ]
[L·min−1 ]
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
T2D (n=6)
69.5
28.2
49.7
23.8
66.4
20.3
2816.7
892.0
3111.1
1294.0
0.38
0.06
T2D
p (n=6)
53.8
12.7
41.8
10.5
50.9
15.4
2308.3
1081.9
3219.8
1676.5
0.34
0.04
T2D
c (n=6)
52.1
17.6
57.1
18.3
58.0
16.9
3100.0
834.9
2386.0
536.8
0.35
0.07
HR: heart rate, V’O2 musc: muscular oxygen uptake; V’O2 pulm: pulmonary oxygen uptake; Vv : venous volume; Q’rem : perfusion of the remainder of the body; V’O2 rem: oxygen uptake of the remainder of the body; τ: time constant; SD: standard deviation;
T2D: subjects with type 2 diabetes mellitus not taking additional medication; T2Dp : subjects with type 2 diabetes mellitus taking peripherally acting medication; T2Dc : subjects with type 2 diabetes mellitus taking centrally acting medication
Discussion
The aim of this study was to investigate differences in HR and V’O2 musc kinetics between groups of patients with T2DM, considering their cardiovascular
medication. For comparisons between the groups, a standardized WR protocol was used
and V’O2 musc kinetics were estimated from HR and V’O2 pulm applying a circulatory model.
1) Descriptively, HR kinetics, represented by CCFmax , were faster in T2Dc patients, compared with T2D patients not taking cardiovascular medication. V’O2 musc kinetics seemed slower in T2Dc subjects, but the results were not significant.
2) V’O2 musc kinetics of the T2Dp patients were slightly faster compared with the T2D patients without cardiovascular
medication, but this was not significant. HR kinetics of the T2Dp were slightly faster compared with the T2D groups, but this difference was also not
significant.
Although no statistical differences regarding direct group comparisons were identified,
a significant main effect for Parameter x Group including V’O2 musc and HR kinetics was found. V’O2 musc kinetics were significantly faster than HR kinetics within the T2Dp and T2D groups, but not within the T2Dc group. For the T2Dc group, V’O2 musc kinetics seemed slower than HR kinetics. For comparison, HR kinetics have been
shown to be faster than V’O2 musc kinetics in healthy young subjects [14 ]
[28 ]. In sedentary aged subjects, V’O2 musc kinetics were faster (but not significantly) than HR kinetics [29 ]. This is in line with the results of the T2Dp and T2D group, in the present study. Taniguchi et al. [10 ] showed a positive effect of beta blockers (after one year of treatment) and Dayi
et al. [11 ] for ACE inhibitors (administered shortly before the exercise test) on V’O2 pulm kinetics in hypertensive subjects and in patients with dilated cardiomyopathy.
They explained this effect by improved cardiac function (improved left ventricular
ejection fraction). Descriptively, the results of this study show this positive effect
of the cardiovascular medication on HR kinetics compared with the group not taking
cardiovascular medication ([Fig. 2 ]). Nevertheless, the effect was insignificant and did not result in faster V’O2 musc kinetics compared with the T2D group.
Overall, the disease status of the T2Dc group might have been worse, since some of them were treated with more than one cardiovascular
medication ([Table 1 ]). Between the 3 groups, no obvious differences in V’O2 max were evident. Hence, regulatory processes to achieve the same V’O2 max seem to be different and influenced by cardiovascular medication and/or disease
status.
The very slow HR kinetics in the T2D patients (no cardiovascular medication) in the
present study were also observed in other studies, comparing T2DM patients with healthy
controls [19 ]
[21 ]
[22 ]. It has been shown, that T2DM influences cardiac mechanoenergetic efficiency and
cardiac hypertrophy [30 ]
[31 ]
[32 ]
[33 ]. The respective medication in the T2Dc and T2Dp group might have improved cardiac function, as has been supposed in previous studies
[9 ]
[10 ]
[11 ]. However, as can be observed in [Fig. 2 ] this did not lead to faster V’O2 musc kinetics compared to the T2D group.
The patients analyzed in this study were selected for the subgroups, considering the
group of cardiovascular medication they were taking ([Table 1 ]). The possiblity of the underlying disease to influence the obtained results cannot
be excluded. Anyway, the applied method showed that differences can be detected between
the analyzed patient groups even in a small sample size. Regulatory processes between
T2Dc and the other 2 groups were different.
To be comparable with other studies, time constants were calculated as rough estimates
from CCFmax . The time constants calculated in this study were within the given ranges from the
literature, where values for τV’O2 musc (in the literature represented by the phase 2 τ of V’O2 pulm) as a response to WRs vary from 41 s to 58 s and values for HR vary from 51 s
to 81 s for T2DM [16 ]
[17 ]
[19 ]
[20 ]
[21 ]
[22 ]
[34 ]. Since the applied test delivers plausible results within a single test session,
without the need to adjust WR ranges or the need to fit data to an explicit model,
the applied test might be relevant for clinical routine.
Conclusion
Even though this study can only be treated as a pilot study, the different effects
of the peripherally and centrally acting medication and/or disease conditions on HR
and V’O2 musc kinetics without any obvious differences in V’O2 max are worth being considered. In the T2D and T2Dp group, but not the T2Dc group, V’O2 musc kinetics were significantly faster than HR kinetics. This shows that regulatory
processes for the achievement of a similar V’O2 max are different between the groups. Future, larger studies analyzing T2DM patients
should consider the influence of cardiovascular medication on HR and V’O2 musc kinetics, rather than excluding those patients from analysis.
