| Cardiovascular regulation in epilepsy with emphasis on the interictal state | ||
|---|---|---|
| Prev | Chapter 4. Subjects and methods | Next |
All patients were carefully interviewed and clinically examined and the epilepsy type was classified according to the recommendations of the International League Against Epilepsy (ILAE 1989). In an interview of each patient, clinical manifestations of ANS dysfunction such as arrythmias, dizziness due to ortostatic hypotension, sweating, abnormalities of urine and bowel function and sexual malfunction were evaluated in particular. Laboratory screening (liver and renal function, serum electrolytes, basic hematologic parameters) was normal in all the study patients. In general, the patients had neither excessive body weight nor elevated BP. EEG-recording was obtained from all the patients.
The exclusion criteria were met in male patients more often than in female patients which resulted in a smaller number of male patients in the study.
Cardiovascular autonomic reflex tests, based on HR and BP responses at rest and after various stimuli were performed in 122 study patients and 88 control subjects under standardized conditions, in a silent room with a temperature from 20 to 23°C between 9 a.m. and 12 a.m (Braune et al 1996, Suominen 1997). First, with the patient in a supine position on the tilt table, the maximum contraction power (handgrip) of the patient´s dominant hand was measured three times with a dynamometer for the isometric work test. At the end of a 30-minute resting period, the baseline BP was measured three times using an automatic arm sphygmomanometer. Thereafter, the following five tests were performed: normal breathing, paced breathing at six breaths per minute (deep breathing), the Valsalva maneuver, upright tilting (2 seconds, 90°) and isometric work. The interval between the various tests was standardized so that the next test was not started until the HR and BP had returned to the baseline level after the previous test. In the analyses of the deep breathing test, the thermistor signal ascertained that the paced deep but not maximum breathing was evenly performed. A small hole in the mouthpiece attached to a mercury manometer during the Valsalva maneuver guaranteed that thoracic pressure had to be used for blowing. The blowing pressure was monitored and only steadily maintained blowings were accepted. In the isometric work test the contraction power on the dynamometer was monitored graphically and thereby maintained at the optimal level. The ECG and breathing (nasal thermistor) signals were conveyed through an A/D converter with a sampling frequency of 320 Hz to a PC computer and were analyzed off-line using an automatic program package which allowed visual checking of the raw ECG and breathing signals. Test performances in deep breathing, the Valsalva maneuver and the isometric work tests were checked on-line and off-line, and only adequately accomplished test performances were accepted for analysis. (Suominen 1997, Myllylä et al. 2002)
In the normal breathing test the consecutive RR intervals for a period of one minute were measured from the ECG, and the standard deviation (SD) of the intervals was used as the test variable. Successive RR intervals were measured and the square root of the mean squares of the differences between successive intervals (rMSSD) was calculated, so reflecting the true beat-to-beat variation. (Myllylä et al. 2002)
In the deep breathing test the mean ratio of the longest (expiration) to the shortest (inspiration) RR interval of five consecutive breathing cycles was calculated. The test was performed twice and the higher RR interval ratio was used as the "maximum-minimum (max-min) ratio". (Myllylä et al. 2002)
In the Valsalva maneuver the ratio of the longest RR interval after blowing (at the pressure of 40 mmHg for 15 sec) to the shortest RR interval during blowing or immediately after it was calculated. The highest ratio of three maneuvers was used as the “Valsalva ratio”. (Myllylä et al. 2002)
In the tilting test the ratio of the longest RR interval around beat 30 (beats 20 to 40) to the shortest RR interval around beat 15 (beats 10 to 20) after quick passive upright tilting was used as the “30:15 ratio”. The systolic and diastolic BP responses were measured at rest, immediately after tilting, and at 2, 5, and 7 minutes after tilting. The difference between the BP at rest and the lowest BP after tilting was recorded. (Myllylä et al. 2002)
In the isometric work test the largest increase in systolic and diastolic BP during a 5- minute period of sustained handgrip with a dynamometer at 30% of the maximum voluntary power was recorded. BP was measured at 1, 2, 3, 4 and 5 minutes of work and the result was compared with the BP at rest. Male and female subjects were analysed separately. (Myllylä et al. 2002)
A two-channel 24-hour ambulatory ECG recording (Delmar Avionics electroscanner) was performed in patients and control subjects in studies III-IV (Task Force 1996). During the recording they continued to perform their normal daily activities. They were also asked to keep a diary of all the activities and possible seizures during the recording.
The ECG data from the recordings were sampled digitally and transferred from the Oxford Medilog scanner to a microcomputer for analysis of HRV. All RR interval time series were first edited automatically, after which careful manual editing was performed by visual inspection of the RR intervals. Each RR interval time series was passed through a filter that eliminates premature beats and artifacts and deletes the filling gaps (Huikuri et al 1993, Huikuri et al 1994, Huikuri et al 1996). In the final analysis of linear components of HRV, 24-hour measurements were divided into segments of 3600 RR intervals, and in the analysis of non-linear components of HRV, 24-hour measurements were divided into segments of 8000 and only segments with >85% sinus beats were included. (Myllylä et al. 2002)
The mean length of all RR intervals and standard deviation (SDNN) of all RR intervals were computed as time domain measures of HRV. The power spectra of HRV (Figure 1) were quantified by measuring the area in 3 frequency bands: 0.005 to 0.04 Hz, VLF, 0.04 to 0.15 Hz, LF and 0.15 to 0.4, HF. (Myllylä et al. 2002)
For quantitative two-dimensional vector analysis, the standard deviation of instantaneous beat-to-beat RR interval variability (SD1) and continuous long-term RR interval variability (SD2) were analysed, and visually presented as Poincaré plot scattergrams (Figure 2), in which each RR interval is plotted as a function of the previous one (Tulppo et al. 1996, Korpelainen et al. 1999). In the computerised analysis, the Poincaré plot was first turned clockwise, and then standard deviation of the plot data was then computed around the horizontal axis, passing through the data centre (SD1). The SD of the continous long-term RR intervals was quantified by turning the plot 45º counterclockwise (SD2) and by computing the data points around the horizontal axis, passing through the centre of the data. (Myllylä et al. 2002)
A value of ApEn is a measure that quantifies the regularity of time series data. It measures the logarithmic likelihood that runs of patterns (beat to beat difference of RR interval length) are close in the next incremental comparison. A time series containing many repetitive patterns has a relatively small ApEn, whereas more random data produce higher values. Two input variables, m and r, must be fixed to compute ApEn, and m = 2 and r = 20% of the SD of the data sets were chosen as suitable values on the basis of previous findings of good statistical validity. (Pincus 1991, Pincus & Viscarello 1992, Pincus & Goldberger 1994, Myllylä et al. 2002)
To quantify fractal correlation properties of HR, the detrended fluctuation analysis technique, which is a modified root-mean-square analysis of random walk, was used. The HR correlation properties were defined separetely for short-term (≤ 11 beats, α1) and for long-term (> 11 beats, α2) correlations of RR interval data (short- and long-term scaling exponents). (Peng et al 1995, Iyengar et al 1996, Mäkikallio et al 1997, Myllylä et al. 2002)
The power-law relationship of RR interval variability, a spectral measure reflecting the distribution of the spectral characteristics of the RR interval oscillations, was calculated from the frequency range of 10-4 to 10-2. The point power spectrum was logarithmically smoothed in the frequency domain, and the power was intergrated into bins spaced 0.0167 log(Hz) apart. A robust line fitting algorithm of log(power) on log(frequency) was then applied to the power spectrum between 10-4 and 10-2, and the slope of this line was calculated. This frequency band was chosen on the basis of previous observations regarding the linear relationship between log(power) and log(frequency) in this frequency band. (Saul et al 1987, Bigger et al 1996, Myllylä et al. 2002)
In Studies II - IV, MRI was performed in all except three patients with well-controlled TLE who had claustrophobia. CT was performed in those three patients. In Study I CT was performed in all patients to exclude symptomatic epilepsy. MRI was performed on a 1.0 Tesla unit (Magnetom SP 42, Siemens, Erlangen Gemany). The protocol consisted first of sagittal, T1-weighted scans with a 5.0-mm section thickness. TR = 570 msec, TE = 15 msec, 2 acquisitions. Second, a fast spin-echo with TR of 6000 msec, effective TE of 90 ms, 8 echoes and FOV of 230 mm was used to image axial slices. Third, a coronal, three-dimensional (3-D) gradient-echo fast low-angle shot (FLASH) acquisition was performed. Parameters of the sequence were 30/5/1 (TR/TE/excitations), field of view was 25 cm and matrix size was 256 x 256. Section thickness was 3 mm and the slices were contiguous. Fourth, coronal multiple spin-echo (SE) imaging with 8 echoes was performed with TR of 2000 msec and TE from 20 msec to 125 msec with 15 ms interval. the field of view was 230 mm and section thickness was 5.0 mm with a 1.0 gap. Contrast was not routinely used.
The images were transferred to a HP 9000/730 workstation. Volumetric measurements were performed using a custom made software. Hippocampal volume was measured on sections of the coronal 3-D FLASH aqcuisition that was not perpendicular to the axis of the hippocampal formation. The images were magnified and the area of the hippocampal formation was zoomed. The structures to be measured were outlined manually with a mouse-driven cursor. The measurements included entire rostrocaudal extent of the hippocampus. The landmarks used for the definition of the hippocampus have been described in detail (Lehericy et al 1994, Hasboun et al 1996).
Measurements of the hippocampal formations at the level of the body of the structure were usually easy. These measurements included Ammon´s horn, the subiculum, the dentate gyrus, and the white-matter tracts of the alveus and the fimbria. The border between the subiculum and parahippocampal gyrus was arbitrarily defined as the most medial extent of the junction of subiculum and the parahippocampal gyrus. Measurements at the level of the head and tail were more difficult. The delineation was guided by the band of high signal intensity generated by the alveus. If the alveus was not clearly visible the most accurate anterior limit was confirmed with the 3-D sagittal plane. Caudally, the posterior boundary of the hippocampal formation was chosen as the last section containing Ammon´s horn, which corresponded to the section where the crus of the fornix was visible.
After the segmentation process, the hippocampus was marked with colour and the entire hippocampus was portrayed with the 3-D cursor in the sagittal plane. Reformatting in this fashion allowed confirmation of the anterior coordinates of the hippocampus and correlation with the original coronal images to ensure accuracy of measurements. The hippocampal area in one section was then calculated by pixel counting. The volumes were calculated by adding areas and multiplying by section thickness. Volume measurements were performed by one operator unaware of the clinical diagnosis.
Table 8. Number of patients epilepsy participating in various examinations in different substudies.
| Substudies | Cardiovascular reflex tests* | Analysis of HRV from ambulatory 24 hour ECG recordings# | CT | MRI |
|---|---|---|---|---|
| Study I | ||||
| Untreated patients | 37 | - | 37 | |
| Patients with idiopathic generalized epilepsies taking AED(s) | 11 | - | 11 | |
| Patients with partial epilepsies taking AED(s) | 36 | - | 36 | |
| Study II-III | ||||
| Patients with refractory TLE | 19 | 19 | 19 | |
| Patients with well controlled TLE | 25 | 25 | 3 | 22 |
| Study IV | ||||
| Patients with hippocampal sclerosis | 8 | 8 | 8 | |
| Patients without hippocampal sclerosis | 31 | 31 | 31 | |
| *, SD of RR intervals, max-min ratio, Valsalva ratio, 30:15 ratio, isometric test; #, spectral analysis, dynamic measures. | ||||
The demographic data of the patients and control subjects were analysed with Student´s unpaired or paired t-test in the case of normal distribution when comparing quantitative data between different groups. The Mann-Whitney two-sample test and the Kruskal-Wallis test (comparison of three groups) were used in analysing unevenly distributed data or ordinal variables.
In the analyses of the cardiovascular reflex tests, analysis of covariance was used to determine the significance levels in comparison between the patients and the control subjects. Cardiovascular autonomic reflexes have been shown to be dependent upon both age and baseline HR (Bannister & Mathias 1988, Kajser et al 1985, McLeod & Tuck 1987, O´Brien et al 1986, Robinson et al 1983, Vargas & Lye 1980, Vita et al 1986, Yokohama et al 1991). Therefore, the values of the HR and BP responses (after logarithmic transformation) were corrected for age and baseline HR separately for the patients and the control subjects using multiple regression. The significance levels for the comparison between the patients and the controls were obtained using analysis of covariance (ANCOVA). The values of the HR responses in all groups were separately adjusted both for the mean age and for the mean baseline RR interval (880 ms) of the patients by multiple regression analysis, and the differences between the results of the regression analysis and those between the regression values were compared by ANCOVA. Statistical evaluation of the various HR variability analyses was performed with using the Kruskall-Wallis test and the Mann-Whitney two-sample test to compare the values of the control subjects and those of the patients.
Correlations between ApEn, the slope of the powerlaw relationship of the HR variability and the traditional spectral measures of HR variation in patients with TLE (Study III) and correlations between the measures of HR responses and measures of HR variability (Study IV) were assessed with Pearson´s bivariate correlation test. Values p < 0.05 were considered significant. All analyses were made on observed cases and calculated using the SPSS Windows.
The mean (± SD) values of the volumetric measures of the hippocampi and amygdalas were calculated using with Student´s t-test.