3 Methods

The study was approved by the Human Ethics Committee of the Medical Faculty Charité of the Humboldt University of Berlin in accordance with the declaration of Helsinki.

3.1 Experimental design

The influence of the menstrual cycle on the heart rate variability (HRV) was investigated in five different phases in trained and untrained women in this study. In addition, the influence of testosterone on the HRV was examined in five study days in trained and untrained men. The HRV was investigated at spontaneous breathing for 20 min at rest and during the orthostatic test, which consisted of three parts: 20 min supine, 10 min standing and 20 min. supine. The change of the body position was done actively by the subjects. The breathing frequency was controlled by respiration belts. The HRV evaluation was done in the time and the frequency domain including those parameters which were allowed for short time measurements in accordance to the Task Force [84].

The five menstrual cycle phases were individually determined based on the daily measurement of the basal body temperature (BBT) one month prior to the study. The BBT was orally determined by a digital thermometer in the morning before getting up. During this month, females also had to fill in a daily questionnaire to the mood state, the alcohol intake and the activity i.e. training pattern. Additionally, several blood borne parameters (INS, BG, WBC, RBC, Hc, Hb, Na+, K+, Mg²+, Ca²+ and Cl-) were analysed for each study day. This included the analysis of the endogenous hormones (P, E2, LH, [page 43↓]FSH) in women which served to control the menstrual cycle phases, along with the daily measured BBT. The study days of men were set at the same time interval as the menstrual cycle phases of women and thereby men served as control group. Additionally, the reliability of the method and procedure was veryfied by the five study days of men due to the repeated measurements. The blood borne parameters (INS, BG, WBC, RBC, Hc, Hb, Na+, K+, Mg²+, Ca²+ and Cl-) were analysed including the testosterone and the SHBG concentration in men for each study day. The subjects also had to fill in a daily questionnaire which served to control of the mood state, the alcohol intake and the activity i.e. training pattern during the study. Based on this, the preceding month and the test month of women were compared with each other and with the men’s month to ensure stable conditions.

60 volunteers, consisting of 30 men and women in the age of 25-38 years were included in this study. Subjects were without any kind of infection, arterial hypertension, and diabetes mellitus as well as without any psychological and/or psychosocial overstrain situations (e.g. overtraining in athletes) in the last 12 months. The male and female groups were separated into long term (≥2 years of training) endurance trained athletes and sedentary subjects. The athletes, consisting of marathon runners, cyclists and triathletes had to train ≥5h per week whereas the sedentary subjects were only included in moderate but not regular daily activity which was <2h per week. The maximal oxygen uptake in relation to the body weight was determined by a maximal effort test to establish the endurance capacity (EC) of athletes and sedentary subjects. The maximal effort test was done by bicycle for the sedentary subjects. Athletes who did not reach the maximal oxygen uptake on bicycle were additionally tested on a treadmill. The EC was set for male athletes at ≥55ml/min/kg, for female athletes at ≥50ml/min/kg, for sedentary men at <55ml/min/kg and for the untrained women at <45ml/min/kg.

Moreover, subjects were non-obese, non-smokers and did not take any kind of medications and/or drugs. The female subjects had regular menstrual cycles for at least 6 months prior to the study. They did not have any pre menstrual tension syndromes (PMS) before the menstrual bleeding and no woman was actually pregnant. Furthermore, none of the female subjects used oral contraceptives or any form of hormonal replacement therapy. Female athletes did not have menstrual dysfunctions or any other menstrual disturbances during high intensity training periods. Prior to the study, all subjects completed a health history questionnaire and a medical examination. [page 44↓]The subjects, who were finally included in the study, are described in details in the following chapter.

3.2 Subjects

Sixty subjects were tested in this study and forty nine male and female volunteers were included in the final study. Each subject was familiarized with the testing equipment and procedures used in the laboratory. They provided a written informed consent prior to participation. The use of tobacco products was forbidden and alcohol intake was restricted during the study. The athletes were allowed to train to their individual training pattern. The untrained subjects were moderate daily active, but never involved in regular exercise training. Every kind of physical activity was noted in the questionnaire during the study and thus all groups have a training anamnesis.

The study population was first separated into a female and male group and secondly by the relative oxygen uptake (VO2rel) and the training anamnesis into a trained and untrained group.

Training anamnesis consisted of training age in years (TAge), training sessions per week (TSpweek) and of the training units in hours per week (TUpweek). The following 4 subgroups resulted: 13 trained (MT), 14 untrained men (MUT), 11 trained (WT) and 11 untrained women (WUT). Descriptive characteristics of the volunteers are presented in table 3-1 and data concerning training anamnesis in table 3-2.


[page 45↓]

Due to non-normal distribution, all data are presented in median, first and third quartile.

Table 3-1: Anthropometrical data of volunteers separated into 4 subgroups

Table 3-2: Anamnesis of training data including training age in years (TAge), training sessions per week (TSpweek) and training units per week in hours (TUpweek)


[page 46↓]

3.3  Study days of men and women

Female volunteers measured their basal body temperature (BBT) each morning and recorded the day of menses to verify a recent history of menstrual cycle regularity one month before entering the study and throughout the one month experimental period. The BBT was orally measured before getting up in the morning by the digital thermometer KD-132 (K-Jump Health Co., Lot-No: 07/02, Ltd. Taiwan Imp., SCALA Electro GmbH, Stahnsdorf, Germany). Thereby a temperature curve based on the individual determined BBT was evaluated for every woman. The menstruation (bleeding phase) and the ovulation with the characteristic enhanced BBT were chosen to determine these five phases individually; the menstruation, the middle of the follicular phase, the ovulation phase, the middle of the luteal phase and the pre menstruation phase. Females´ first experimental day was M and the last PreM. Women were involved in the study for two successive months starting with the preceding month i.e. BBT recording followed directly by the test month; ergo every woman had 5 study days.

The phases were verified as follows:

Menstruation phase (M): second or third day of bleeding

Middle of follicular phase (MidF): between M and O

Ovulation phase (O): ovulation ±1 day

Middle of luteal phase (MidL): between O and next M

Pre menstruation phase (PreM): 1 or 2 days prior to the next menstrual bleeding

Men could start the series of experimental investigations at any time. Their experimental period varied from 24 to 32 days similar to the women’s cycle length. In this period they had five experimental days every four to five days during one month. Although men do not have any cycle related changes during one month, they had to take part of five days including the daily BBT recording. Equal study conditions should thereby lead to comparable results between men and women.


[page 47↓]

3.4  Procedure

Volunteers had to fill in a POMS questionnaire every morning including the measurement of the resting heart rate and the basal body temperature before getting up. The alcohol consumption and sports activity of the previous day as well as the duration and the quality of sleep were required.

Subjects arrived at the laboratory at the individually equal time each morning, i.e. between 07.00 and 11.00 am. All of them were requested to refrain from eating and drinking for at least 8 hours before the experiment. They also had to avoid stress and exercise in the morning to be in a relaxed condition and quiet mood when arriving. Each of the five sessions were identical and were scheduled for the same time of day. The sessions were organized into groups of one to four persons.

The temperature in the laboratory was between 20 to 24°C and the humidity between 40 to 48%. The room was darkened and without acoustic disturbance. Volunteers had to feel comfortable during the experimental phase. They were instructed to be as relaxed as possible and to breathe spontaneously at their own rate.

After a resting period, the subjects rested 20 minutes in supine position, ten minutes in upright standing position and another 20 minutes in supine position. Volunteers stayed afterwards in supine position for the venous collecting of blood samples which was done by a doctor or a medical technical assistant (MTA).

3.5 Analysis

The analysis of the blood borne parameters was done by the “Laboratory 28” whereas the hormones and the lactate values were analysed in our laboratory.

3.5.1 Blood borne parameters and electrolytes

The analysis of the haemoglobin (Hb), hematocrit (Hc), leukocyte (WBC), erythrocyte (RBC), blood glucose (BG) and electrolytes (Na+, Ca²+, K+, Mg²+ and Cl-) was carried out by the “Labor 28” (Labor 28, Berlin, Germany). Hb, Hc, WBC, RBC and BG were analysed from EDTA fresh blood. BG was stored in a sodium-fluoride tube whereas Na+, Ca²+,, K+, Mg²+ and Cl- were determined in the blood serum. The blood sample rested for 30 min before it was centrifuged with 2000 revolutions/min for 15 min by the centrifuge EBA 8 (Centrifuge EBA 8, A. Hettich, Tuttlingen, Germany). Then, the [page 48↓]blood serum was separated by a pipette from the blood clot, filled in primary tubes and stored at 4-8°C till analysis at the same day. The analyses of the electrolytes and the blood borne parameters were different and thus several methods are described in the following part.

3.5.1.1 Analysis of the electrolytes

Na+, K+ and Cl- were determined in mmol/l by the indirect potentiometer (Hitachi 747-400, Tokyo, Japan) with the following test principle: Electrodes of selective ions provoke an electrometrical power which is related to the activity of the ions in hydrous solution. The electrometrical power can be converted into mmol/l by a formula. The specification of the analysis of Na+ was 80 mmol/l, K+ 1.5 mmol/l and of Cl- 60 mmol/l.

Ca²+ was analysed by the colour test with end-point and blank value determination (Hitachi 747-400, Tokyo, Japan). Ca²+ forms with Kresolphthalein-complexion a violet complex in alkalise solution. The intensity of the violet colour is direct proportional to the concentration of Ca²+ which results in the calculation of Ca²+ in mmol/l with a precision of 2.12 mmol/l intra- and 2.1 mmol/l interassays.

Mg²+ was analysed by the atom absorption spectrometer with flame technique which is based on the proportion of the element’s concentration and its absorbed radiation after atomising (Beer’s Law). The measurement of the element’s concentration at specific wavelength happens after atomising.

BG [mmol/l] was determined by the glucose dehydrogenise (B-D-Glucose) which catalyses the oxidation of blood glucose (BG) to the following equation:

The amount of NADH which results from this equation is proportional to the glucose concentration.

Finally, the analysis of the WBC [G/l], RBC [G/l], Hc [%] and Hb [mmol/l] concentrations has been done by the analysis system of Sysmex 9500 which consists of different measurement methods.

WBS are analysed by the galvanic current and resistance principle, the RBC by the galvanic current with hydrodynamic focussing, the Hb with sheaths flow and photometric measurement whereas the Hc value is calculated by the cumulative summation of the RBC impulses. These analysis methods are all based on the principle that the resistance between the electrodes changes by suspension of the blood cells which produces electrical impulses. These impulses which are proportional to the cell [page 49↓]size are counted and imply a distribution which reflects the cell size. Additional high frequency resistance to the galvanic current results in impulses which reflect the specific gravity. Based on these data the cells are counted with the help of a scatter plot. The sheaths flow enhances the suspension and thus, the counting gets more specific for the Hb concentration. WBC has an intra- and an interassay of 1.15% and 3.45 %, RBC of 0.89% and 0.55 %, Hc 0.96% and 0.79% and finally Hb 0.32% and 0%.

3.5.2 Hormonal analysis

Hormone status of progesterone (P), estradiol-17β (E2), luteinizing hormone (LH), follicle stimulating hormone (FSH), total testosterone (tT), insulin (INS) and sex-hormone binding globulin (SHBG) was determined from blood serum which was frozen till analysis. The shelf-life of the hormones was 2 months frozen at -20°C. The menstrual cycle was monitored by five analyses of the P, E2, LH and FSH levels whereas tT and SHBG were analysed for men and finally the INS level for both genders.

3.5.2.1 Analysis of the hormones

The hormonal analysis was done in our laboratory by the Immunoassay Analyser Immulite 2000 of the Diagnostic Products Corporation DPC (DPC Cirrus Inc., Los Angeles, USA; German agency DPC Biermann GmbH, Bad Nauheim, Germany). Immulite 2000 is a continuous Random Access System to manage chemiluminescent immunoassays automatically. Immulite uses test tubes, reagents and substrate to analyse the hormonal levels. A synthetic globe with parameter specific antigens is inside the test tubes. The immunoreactions, the incubation, the washing steps and the development of the signal take place in the test tubes. The reagents which are identified by barcodes are marked by an alkalised phosphatase. At least the substrate used is a chemiluminescent enzyme substrate.

In this study, commercial kits of DPC Biermann were used with the following quality numbers: P 267, 277; E2 241, 249; LH 254, 259; FSH 267, 272; fT 157, 158, 163; INS 126, 126A, 130; SHBG 237, 238, 241. Additionally, twice distilled water mixed with a washing module (quality number 063, 064, 069) in the ratio of 1:9 i.e. 100 ml washing module added to 900 ml distilled water was used beside the chemiluminescent enzyme substrate with the quality numbers 179, 181, 190. The serum of the patients was filled in probe tubes which were identified by numbers.


[page 50↓]

Prior to the analysis, the calibration, the daily routine and the analysis of the controls had to be done. The calibration included a high and low calibration level, which was determined four times to calculate the slope (max. 10% variation) and the intercept (max. 2% variation) in accordance to the master standard curve of DPC Biermann. The control serums CON4, CON5 and CON6 (quality number 019) and the special SHBG control (quality number 017) were used to measure the controls which were done twice (high and low level). The analysis was only started with approved calibrations and controls, which had to be renewed after the analysis of 500 test tubes.

The run of the automatically Immulite analysis is now described step by step:

  1. The patient’s serum and the reagents are incubated in Immulite test tubes.
  2. Then, the incubation time, which varies from 30-60 min starts at 37°C.
  3. After the incubation, water is added to the test tubes to wash the globe. Then, the test tubes are rotated around the vertical axis to separate serum, water and reagent liquid from the globe. Finally the globe is free of unbound markers and liquids.
  4. Then the bound marker is quantized by a luminescent Dioxetane-substrate.
  5. Finally, the photo multiplier measures the emitted light and the software calculates the results of each probe.

After the analysis, the syringes had to be cleaned, waste liquids had to be taken away and data printed or stored on the disc. Immulite remained in idle to keep the accumulated data and the analysis instrument ready.

3.5.2.2 Intra- and interassay variance

Intra- and interassay variance was calculated in accordance to the General Medical Council of Germany (December 2003) and to IMMULITE reference values: LH (4.9% / 8.9%), FSH (5.7% / 7.4%), E2 (8.4% / 8.8%), P (7.6% / 8.6%), SHBG (6.2 %/ 9.8%), tT (7.2% / 7.2%), INS (4.8% / 6.0%).

3.6 Recording of electrocardiograms

Different data were recorded during the study including the ECG, the respiration frequency and the spiroergometric tests on bicycle and treadmill. The procedure of each recording is described in the following chapters.


[page 51↓]

3.6.1  Electrocardiogram

Subjects arrived in the laboratory. Three ARBO (H124) electrodes (Kendall, Medizinische Erzeugnisse GmbH, Neustadt/Donau, Germany) were attached according to the second derivation of Nebh A (anterior). The first electrode was mounted to the fourth intercostal space and the second to the apex of the heart on the left side of the midclavicular line. The third electrode was attached at the right collarbone as earthing. Electrode diameter was 55 mm and contact surface 20 mm. Men with significant body hair were shaved before mounting the electrodes. Electrodes were connected with an electrocardiogram amplifier (ECG-Amplifier) from Biovision (Biovision, Bleichstrasse 6a, 61273 Wehrheim, Germany).

An elongation cable of 2 metres conducted the electrocardiogram (ECG) signal with an entry impedance of 10 gigaΩ to the input box from Biovision (Biovision, Wehrheim, Germany). The input box had 16 channels, of which a maximum of eight were used. A wide-bend cable connected the input box to the A/D converter DAQ card (DAQ card 700, National Instruments, München, Germany). The sample rate was set to 1000 Hz and the resolution to 12-bit. TheECG recordings were made using a commercial computer running the Daisylab software (Daisylab version 6.0, National Instruments, München, Germany).

The ECG recordings were detected and corrected by hand for ectopic, electrical disturbances and missed beats by the automatic detection algorithm implemented in RASCHlab (RASCHlab-mfc version 0.1, using LibRASCH version 0.3.1 of Ralph Schneider, Technical University of Munich, Germany). After shortening the ECG values to ten minutes segments, RR interval distances not classified as sinus rhythms and outside the range of 27 to 200 beats per minute were eliminated; no interpolation was done. Sequences with more than 5 eliminated beats per minute were not taken for the evaluation.

The time domain parameters meanNN, SDNN, RMSDD and pNN50 were calculated. Afterwards the heart rate time series was band-pass filtered by a Hamming window as an apodization function in order to reduce aliasing. Next a Fast Fourier Transformation (Source: Proc. Algorithms in Matlab, Disk., 1996) was performed on the filtered signal; each heart rate time series was expressed by a sequence of amplitudes describing the time-dependent magnitude of the oscillations.

The main frequency bands LF, HF and VLF were defined at the following bands according to the Task Force [84]: LF (0.04-0.15 Hz), HF (0.15-0.4 Hz) and VLF [page 52↓](0.0033-0.04 Hz). Due to short time records, ULF was not evaluated. TP consisted of LF, HF and VLF power.

3.6.2 Breathing frequency

The spontaneous breathing frequency was recorded by the respiratory belt Effort Sensor of Viasys (Viasys Health Care GmbH, order number 706428, Höchberg, Germany). Piezoelectric parts in the respiration belt register the tension in the belt produced by inspiration and expiration, i.e. the rise and the sink of the abdominal wall. Different tension results in different electrical signals which were conducted by an elongation cable of 2 metres to the input box, the A/D converter and finally the computer.

The belt was fixed around the abdomen between belly and costal arch. Volunteers were not disturbed by the belt and felt comfortable. In supine position the respiration is more abdominal than thoracal and thus the use of only one belt was sufficient to record the abdominal breathing. The inspiration and the expiration as an expression of the breathing rhythm i.e. the respiratory frequency were measured at rest and during the orthostatic test. The evaluation of the breathing excursions per unit of time was registered at the screen.

3.6.3 Spiroergometric tests

All volunteers underwent a bicycle maximal-effort test to measure the maximal (VO2max) and the relative (VO2rel) oxygen uptake as a measure of cardio respiratory endurance capacity. The VO2max is the maximum uptake of oxygen in millilitres per minute (ml/min). The VO2rel is the maximum uptake of oxygen relative to body weight in millilitres per minute per kilo body weight. Spirometric data were recorded breath by breath. Men could carry out the bicycle test at any time around the experimental period. Women were tested in the middle of the luteal phase (MidL) because of missing premenstrual tension syndrome before menstruation. The sedentary subjects were not familiar with a maximal effort test and therefore the bicycle instead of the treadmill was chosen for the test. The athletes which consisted of marathon runners, cyclists and triathletes were tested on bicycle and on treadmill. The VO2rel was determined by the results of the bicycle and the treadmill test to have comparable results for the three kinds of athletes.

One day prior to the test hard training had to be restricted; on the day of the assessment exertion, exercise and caffeine had to be avoided. The procedure of the bicycle and [page 53↓]treadmill test as well as the ventilation recording was similar. The protocols of both tests are described separately in the following chapters whereas the ventilation recording was described only once. The maximal ergometric tests ended with subject’s exhaustion.

3.6.3.1 Protocol of bicycle ergometric test

The bicycle exercise test was done by an Ergometrics 800 S bike (Ergo Line, Medical Measurement Systems, Binz, Germany). Ergometrics 800 S was a computer-controlled high performance bicycle. Prior to the bicycle test, a 12-lead standard ECG recording (ECG-recorder Quinton, model Q 7/0 sx, serial 00334-165-0107, Quinton instrument co., Seattle, Wa., USA) was obtained at rest in supine position.

The protocol started with a warm-up period consisting of three minutes of cycling with the minimum of 60 revolutions per minute (60 rpm) at a workload of 0 watt. The increase in workload was 50 W every three minutes without any breaks. The rotation frequency of 60 rpm had to be maintained throughout the test.

3.6.3.2 Treadmill testing

Subjects were monitored using the 12-lead ECG recording during the test. Volunteers wore a security belt of Saturn on treadmill (Saturn, HP Cosmos, Sports Equipment GmbH, Nussdorf/Traunstein, Germany) during the test. The protocol started with a warm-up period consisting of three minutes walking at a speed of 1.5 metres per second (m/s). The increase of the speed was 0.5 m/s every three minutes. Interruptions of 30 seconds after each level served for the measurement of blood pressure. The treadmill test was also a maximal-effort test.

The exercise test ended when subjects were exhausted. The treadmill test could be stopped at any time with a delay within five seconds.

3.6.3.3 Ventilation recording

The ventilation, oxygen uptake, carbon dioxide excretion and derived parameters changing with increasing workload were measured by Oxycon Champion Record manufactured by Viasys (Viasys Health Care GmbH, Höchberg, Germany). After a running up time (adjustment of the measuring instruments) of 30 min, volume and gas had to be confirmed by calibration. The calibration of volume was done by a 2 litre calibration pump. Gas calibration was done by twice solidified calibration gas. 15 percent by volume of oxygen, 5 percent by volume of carbon dioxide and the reminder of nitrogen were the components of the gas mixture (Calibration gas of Messer [page 54↓]Griesheim GmbH, bottle number A022941, Krefeld, Germany). The result of both calibrations was a correction factor calculated by computer to correct the measurement. Dry atmospheric air was the reference for the oxygen analyser. The ergometric data were analysed by the software Lab Manager, version 4.5.2.

Volunteers had to wear a synthetic basic face mask of Viasys (Viasys Health Care GmbH, Höchberg, Germany) connected to a Triple V Volume Sensor of Viasys. The analysis of oxygen was based on the differential paramagnetic principle and the analysis of carbon dioxide was based on infra-red absorption principle. The inspiration and expiration air flow was analysed during the bicycle and treadmill test. Subjects were monitored using the 12-lead ECG during the test. The Borg Scale was used to measure the subjects’ exhaustion at each level. Blood pressure was measured at each level and five minutes after the test.

3.7 Statistical evaluation

Statistical evaluation was performed using the STATISTICA 1999 Edition Kernel-Version 5.5 A (Stat Soft Inc., 2300 East 14th Street, Tulsa, OK 74104, USA).

3.7.1 Profile of mood states

The Brown-Forsythe-Test was used to compare the homogeneity of variances of all female profiles between the preceding and the test month. The Brown-Forsythe-Test was also used to compare the results between the groups i.e. men versus women and trained versus untrained in the test month. Significance levels were statistically determined at p-values of <0.05 (significant), <0.01 (very significant) and <0.001 (highly significant) in accordance to Hartung et al. [31].

Kruskal-Wallis ANOVA with repeated measurements was used to compare the difference of the POMS profiles between men and women and also between trained and untrained men and women in the 5 different phases of menstrual cycle.

3.7.2 Heart rate variability

Due to non-normal distribution and heterogeneity of the variance of variables, non parametric tests were selected for the statistical analysis and data are reported as medians and quartiles 1st and 3rd). Inter- and intragroup comparisons were done by the [page 55↓]Mann-Whitney and the Wilcoxon test. The significance level was adjusted to Bonferroni because of multiple testing and repeating measurement designs.

Kruskal-Wallis ANOVA with measurements repeated five times was used for inter- and intragroup comparison in the menstrual cycle trend. 2x2 MANOVA with measurements repeated five times was used to compare men versus women and trained versus untrained subjects. 2x2x3 MANOVA with measurements repeated five times was used to compare men versus women and trained versus untrained during the orthostatic test. Correlation was tested by Spearman’s rank correlation.

3.7.3 Reliability

The reliability of the study was controlled by the intragroup correlations coefficient (ICC), the coefficient of variation and the Bland-Altman plots [5].


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 3.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
06.01.2005