Suitability of fetal scalp electrodes for monitoring the fetal electrocardiogram during labour
This content has been downloaded from IOPscience. Please scroll down to see the full text. 1990 Clin. Phys. Physiol. Meas. 11 297 (http://iopscience.iop.org/0143-0815/11/4/002) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 220.127.116.11 This content was downloaded on 02/10/2015 at 11:57
Please note that terms and conditions apply.
Clin. Phys. Physiol. Meas., 1990, Vol. 11, No. 4, 297-306. Printed in the UK
Suitability of fetal scalp electrodes for monitoring the fetal electrocardiogram during labour J Westgatet, R D F Keith+, J S H Curnow§, E C Ifeachor+ and K R Greenet
t Department of Obstetrics and Gynaecology, Plymouth General Hospital, Freedom Fields, Plymouth PL4 7JJ, UK *Plymouth School of Electronic, Communication and Electrical Engineering, Polytechnic South West, PL4 8AA, UK
5 Department of Medical Physics, Plymouth General Hospital, Freedom Fields, Plymouth PL4 7JJ,UK Received 22 May 1990, in final form 20 August 1990 Abstract. As the limitations of heart-rate based intrapartum monitoring have become apparent, there is renewed interest in analysis of the fetal electrocardiographic waveform as obtained from a fetal scalp electrode. A high quality ECG signal is necessary for waveform analysis. This study examined the suitability of five commonly available scalp electrodes for collecting this signal by examining their physical and electrical characteristics, together with a randomised clinical trial in which the ECG trace quality was assessed in 50 patients. The frequency response of Copeland electrodes was such that they attenuate the ECG signal more than the baseline noise. Difficulties were experienced in obtaining optimum attachment and the long, semi-rigid design increased movement artefact resulting in significantly poorer quality ECG signals. Whilst the Hewlett-Packard double spiral electrode had a near ideal frequency response, certain design features made it difficult to apply and remain secure so the clinical signals were of intermediate quality. The Corometrics and Cetro single spirals had the most stable attachment to the scalp and a near ideal frequency response, so produced significantly better signal quality in the clinical trial. Currently, single spiral electrodes are the most suitable for electrocardiographic data collection.
1. Introduction Since the introduction of electronic fetal heart rate monitoring by Hon in the 1960s (Hon 1963) electrodes have been developed for direct application to the fetus to record the fetal electrocardiographic waveform for the calculation of heart rate only. The display of continuous fetal heart rate, together with information about uterine activity, comprises the cardiotocogram (CTG) which is now the standard method of monitoring the fetus in labour. However, the CTG has proved to be inadequate for the accurate diagnosis of intrapartum asphyxia having only a sensitivity of 20 38% and a specificity of 91 96% (Jongsma and Eskes 1989). The electrocardiographic waveform, however, contains more information, which is Probably more useful than heart rate alone (Greene 1987). At the present time two of the most useful features are changes in the relationship between time constants (Murray 1986) and changes in the S T waveform (Rosen 1986), but others may also be of interest in the future. A multi-centre E C Study to collect and analyse large numbers of .intrapartum fetal ECG records is about to commence (EC Concerted Action Project 1989) to study these and other ECG variables and relate them to fetal outcome.
0143-0815/90/040297 + 10 $02.50
0 1990 Institute of Physical Sciences in Medicine
J Westgate et a1
It has become apparent that past examination of the ECG waveform was hampered by lack of understanding of the effects of the filters used in the collection system. Before a large ECG database is collected and analysed the effects of the data collection system on the raw ECG signal need to be established and optimised. The first point in this is the tiSSue/electrode interface provided by the fetal scalp electrode (FSE). Two basic types of FSE are in use today the spiral FSES (single or double helix) and the Copeland PSES which are either reusable or disposable. They have been designed specifically to record signals suitable for fetal heart rate detection and their effect on the primary signal source of the ECG waveform has not previously been considered. The aim of this study was to compare and contrast five commonly available FSES by examining their physical and electrical characteristics and assessing the quality of the fetal ECG signal obtained from each in a randomised clinical trial.
2. Methods The FSES investigated were the single spiral of Corometrics (Corometrics Medical Systems Inc., USA) and Cetro (Cetro AB, Sweden), the double spiral of Hewlett-Packard (HewlettPackard Medical, UK) and both the Copeland Reusable and Disposable (Surgicraft Ltd, UK). The data collection system used was the ST Analyser (STAN), Cinventa AB, Sweden. The STAN is a microprocessor-based system for on-line analysis of the S T segment of the fetal ECG waveform (Rosen and Lindecrantz 1989). It filters the raw ECG signal for waveform analysis, between 0.05 Hz (1st order high-pass filter) and 100 Hz (4th order Butterworth low-pass filter), and for heart rate analysis, between 4 and 26 Hz. The signal is digitised to 8 bits at a sampling frequency of 500 Hz. The functioning of this processor was independently tested by us (Plymouth Fetal Research Group Report 1989) prior to this study. It has been used merely as a preprocessor to obtain raw ECG data output we have not been concerned with any quantitative measures of the S T waveform, nor the concepts surrounding this measurement. The gain setting of the analyser was normally just over 2000, but was halved or increased by 2 or 4 as required. The FECG signal was obtained by measuring differentially between the electrode on the scalp and a standard skin electrode placed on the maternal thigh with the vaginal electrode as earth. This lead system has a sensitivity vector in the longitudinal plane of the fetus compared with the sagittal plane of the standard FSE connection and should reduce ECG vector change resulting from rotation of the fetus (Lindecrantz et aZ 1988).
2.1. Physical characteristics
The construction of each FSE was examined under a hand lens and the clinical method of application to the fetus critically reviewed. 2.2. EZectrical characteristics
The frequency responses of the electrodes were obtained in an in vitro experiment to assess their ability to reproduce the ECG waveform faithfully (figure 1). Optimal attachment to the piece of liver was ensured for each electrode. The Solartron 1170 Frequency Response Analyser was used to obtain the frequency response for each electrode between 0.01 and 200 Hz.
Suitability of fetal scalp electrodes
Solarton 1170 analyser
Area immersed in 1 crn of saline
Return signal Figure 1. Experimental set-up for in vitro measurement of the frequency response of scalp electrodes.
2.3. ClinicaI trial
The five types of FSE were randomly allocated to a group of 50 women in labour. Women were entered into the trial after the decision to use a FSE was made by labour ward staff and informed consent obtained. The electrode to be used was determined by opening a sealed opaque envelope. Only new electrodes of each type were used. All spiral electrodes and most Copelands were applied by one investigator (JW). T h e others were applied by experienced midwives. On 20 patients the raw ECG data from the STAN was stored onto a PC-based optical disc for later analysis. Signal quality was assessed subjectively at the bedside a t intervals throughout data collection. If the CTG or ECG traces showed interference such as excessive signal loss, large baseline swings or high frequency noise content the electrode was re-applied. At a later stage the complete STAN CTG and ECG waveform trace was divided into 30 min epochs and analysed. The first and second stages of labour were considered separately. An example of a trace is shown in figure 2. The CTG trace was graded 1 (optimal), 2 (fair, adequate for monitoring purposes), or 3 (poor, inadequate for clinical use). These gradings were based on the amount of signal loss and interference seen on the trace. The STAN computes the ratio of the T wave height to the QRS complex height (TIQRS ratio) from an average of 30 accepted ECG complexes. This ratio is plotted on the trace and is updated every time the next 30 accepted complexes have formed a new average. Every W O to three minutes the analogue signal of the most recent average of 30 complexes is Printed onto the trace and this gives a measure of signal quality (figure 2). Each analogue output was graded on the basis of its shape: 1 (optimal, well formed ECG complexes with a stable baseline and no high frequency noise); 2 (fair, some high frequency noise present but waveform components still well defined); or 3 (poor, baseline shift, excessive high frequency noise or very low amplitude signals, unreliable for monitoring). When the signal quality is very poor it will take some time before 30 complexes are accepted to form an average. During this delay the TlQRS ratio will not be plotted. Both the number of
J Westgate et al
Figure 2. A STAN fetal monitor trace showing a Grade 1 fetal heart rate trace (with some decelerations after contractions), analogue Grade 1 ECG plots (on side) and T/QRS measurements as both digital printout and continuous plots.
analogue printouts in 30 min (maximum 14) was counted and the duration of the T/QRS ratio plot in 30 min (maximum 30 min) was measured to give a further measure of signal quality. However, the fact that the STAN prints a grade 3 complex and makes a TlQRS measurement from it means that the assessment of the analogue printout is the best arbiter of signal quality. Neither the actual value of the T/QRS ratio nor its clinical interpretation were considered. We were only concerned with the quality of the signal from which the STAN was making the calculations. Statistical analysis based on ten observations per electrode type was carried out by oneway analysis of variance using the Minitab Statisical Analysis Package (Version 7) on a Prime mainframe computer. Sections of raw ECG data from each electrode were randomly selected for each of the ECG complex gradings and visually assessed to identify the reasons for differences in signal quality. 3. Results
3.1. PhysicaZ characteristics
All FSES were bipolar in design with an active element attached directly to the fetus and the other element making contact with the maternal vaginal tissues via cervical secretions and amniotic fluid. Both elements were stainless steel. In the spiral FSES the stainless steel elements were soldered to insulated wires within a 5 mm by 10 mm cylinder of plastic encapsulant. In the Copeland FSES the elements made a non-soldered contact connection with their lead wires within the base of the FSE (figure 3). The spiral electrodes are attached to the fetus with the aid of two PVC plastic tubes a driving tube and a guide tube. These slip respectively on to and over the electrode facilitating safe introduction into the vagina and rotation of the spiral into the fetal scalp. These tubes are withdrawn once the electrode is applied leaving only the electrode wires emerging from the vagina. All of these FSES are designed to be used once only. The Copeland FSES are either reusable or disposable. They have a spring action twisting the base of the electrode anticlockwise against resistance at the head rotates the semicircular needle into a recess in the electrode head. Once the head of the electrode is
Suitability of fetal scalp electrodes
Figure 3. Spiral and Coper,. w s . ( a )Single (left) and double (right)spiral PSES. The PVC applicator tubes have been removed. (b) Copeland Reusable (lower) and Disposable (upper) PSES. (c) Copeland PSE head showing the semicircular active element held in a half open position.
pressed firmly against the fetus the base is twisted clockwise causing the needle to emerge from the recess and rotate into the fetal scalp. The length of electrode which protrudes from the vagina increases as descent of the fetal head occurs through the maternal pelvis. The Reusable Copelands are cleaned, tested for electrical continuity and re-sterilised before reuse. The long, semi-rigid design of the Copelands made them susceptible to movement artefact as the entire length of the electrode was suspended from its point of insertion into the fetal scalp. It was found that the degree of fixation to the fetal scalp was an important
J Westgate et a2
feature in reducing this movement artefact. But the needle of the Copeland FSE is round bodied and approaches the tough fetal scalp at an oblique angle and it often failed to penetrate the skin completely or merely pinched it without penetration. Both non. penetration and incomplete penetration resulted in a low amplitude ECG signal as only a small area of electrode was in contact with fetal subcutaneous.tissues. Thus both noise and low amplitude signal contribute to a poor signal to noise ratio leading to difficulties in fetal ECG processing. The Hewlett-Packard double spirals became detached more frequently than the other spirals and there are a number of reasons to explain this. It was found that the larger diameter and more rigid applicator tubes made ideal positioning and application to the fetal head difficult, especially early in labour with a posterior cervix. Despite providing a large surface area for signal collection, the double spiral only allowed a 180' clockwise rotation for attachment so that it was more likely to unwind than the 360' rotation of the single spiral. Interestingly, the wires from both the Corometrics and Hewlett-Packard electrodes were wound in the opposite direction to the needles which encouraged unwinding with maternal movement and traction on the wires and this was especially noticeable with the Hewlett-Packard. This effect was confirmed by screwing the electrodes into a piece of thick card and applying traction to the wire. 3.2.EZectrical characteristics The magnitude frequency response for each electrode is shown in figure 4. The phase response was considered negligible as only a maximum phase change of 14' was measured (Copeland Reusable). The electrodes fell into two groups. The magnitude response for the spiral FSES approached the ideal electrode response, being nearly flat within the frequency band of interest (0.05 100 Hz). The Copeland FSES attenuated the whole signal more and had a response that attenuated higher frequency components 4 to 8 dB more than the very low frequency components, especially the baseline shifts. This signal attenuation was confirmed in the clinical trial as the STAN had to be used with higher gain settings to obtain a satisfactory signal with the Copelands compared to the spirals.
F i r e 4. Electrode frequency responses from the in witro experiment illustrated in figure 1.
Suitability of fetal scalp electrodes
To investigate the effects of the frequency responses of the electrodes on the ECG waveform, the responses of the Corometrics and Copeland Reusable FSES were modelled using digital filters. These filters were used on ECG data collected from four neonates all within one hour of birth using precordial silver/silver chloride electrodes, The data collection system used had 12-bit resolution, a bandwidth of 0.05 200 Hz and a sampling frequency of 1000 Hz. Typical unfiltered and filtered sections from each electrode simulation are shown in figure 5. As expected, visual examination shows that both electrodes attenuated the signal, but the peak to peak amplitude of the Copeland signal was only 54% of that of the Corometrics. While there did not appear to be any major changes in the waveform shape, there were small modifications in the P and T wave regions with the Copeland FSE, though this has not been examined in detail.
” n c 0
0 Time (s) 1.7 Figure 5. The filtering effects of the Corometrics and Copeland Reusable Electrodes. (a)A section o f neonate ECG signal before filtering, ( b ) after filtering by Corometrics, and (c) after filtering by Copeland Reusable electrodes.
3.3. Clinical trial results
The first stage results are summarised in table 1. The Copeland Disposable electrode had a Significantly lower proportion of Grade 1 CTG epochs compared to the spiral FSES (pcO.05) and a significantly higher proportion of Grade 2 epochs compared to the Corometrics single spiral and Hewlett-Packard double spiral ( P < 0.05).
J Westgate et al
Table 1. First stage results by electrode type. 10 traces from each electrode type, each assessed over 30 min epochs and averaged. Average results for each electrode type shown.
Corometrics Cetro Hewlett-Packard Copeland Reuseable Copeland Disposable (a)
Total no 30 min epochs studied 81 69 65 73 64
CTG quality 70of epochs Grade Grade Grade
1 84.0 85.5 83.1 57.5 39.1(')
6.1 9.9 14.5 0.0 10.8 6.1 19.2 23.3 39.1(b) 21.8
Total no complexes per 30min 12.5 13.2 13.2 12.2 10.7
complexes TlQRS ratio % of complexes duration of Grade Grade Grade plot per 1 2 3 30 min ~ S D
87.8 80.3 76.5 61.6(c) 39.1(d)
8.9 3.3 23.8f6.0 17.3 2.4 25.5f3.8 11.6 11.9 24.9f2.9 18.9 19.5(') 21.5f6.5 22.2 38.1@) 15.0f6.5(')
P < 0.05, compared to Corometrics, Cetro and Hewlett-Packard spirals