British Journal of Obstetrics and Gynaecology April 1992, Vol.99. pp 302-306

FETAL AND NEONATAL MEDICINE

A randomized controlled trail to compare three types of fetal scalp electrode

LINDA NEEDS Midwifery Sister Maternity Unit, Royal Berkshire Hospital, Reading ADRIAN GRANT Epidemiologist National Perinatal Epidemiology Unit Oxford JENNIFER SLEEP District Research Coordinator West Berkshire Health Authority Reading SARAH AYERS Computer Programmer National Perinatal Epidemiology Unit Oxford GAYE HENSON Consultant Obstetrician The Whinington Hospital, London

Abstracts Objectives To compare three scalp electrodes in respect of need for reapplication, trace quality and scalp trauma. Design Randomized controlled trial. Setting Maternity Unit, Royal Berkshire Hospital, Reading. Participants 780 women in labour at term, with cephalic fetal presentation, deemed to require electronic fetal heart rate monitoring using a scalp electrode. Interventions Three types of fetal scalp electrode: Rocket-Rolon (double-helix spiral); Hewlett-Packard (double-helix spiral); and Surgicraft-Copeland (clip). Main outcome measures Need for more than one electrode; reason for replacement; trace quality; scalp trauma; direct financial costs. Results Reapplication was significantly least likely in the Surgicraft-Copeland group (odds ado vs Hewlett-Packard = 0-61, 9'~% CI 0414)-90; odds ratio vs Rocket-Rolon = 043, 95% CI = 0-2943-62). This reflected a much lower ate of replacement because of detachment. The Hewlett-Packard was, however, least likely to be replaced because of poor trace quality. Interpretable traces were equally common, but trace quality tended to be superior in the Hewlett-Packard group. There was no serious scalp trauma. Difficulty removing the electrode was least often reported in the Hewlett-Packard group, and a mark on the scalp persisted least often in the Rocket-Rolon group. Conclusions The Surgicraft-Copeland performed the best overall, and particularly in respect of attachment to the scalp. It appears to be the electrode of choice of the three tested

Fetal monitoring with a scalp electrode is commonly used during labour to provide a continuous recording of the heart rate. There are several different types of electrode, but no consensus about which performs best. The principal differences between the electrodes are in the mechanism by which they are attached to the fetal scalp. Spiral electrodes, which may be a single helix (for example. the Corometrics electrode), or double helix (for example, the Rolon electrode manufactured by Rocket, and the Hewlett-Packard electrode), are applied with an applicator and attached by a screwing action. The Surgicraft-Copeland electrode has a skin clip on a small disc (Ghosh & Tipton 1976). The clip is spring-loaded in a closed position but once the disc is apposed to the fetal scalp the clip can be opened by twisting the body of the electrode and then released to secure attachment

The study reported here was prompted by an observation made by midwives at the Royal Berkshire Maternity Unit that the Rocket-Rolon fetal scalp electrode then in routine use frequently needed replacing several times during the course of a single labour. Informal introduction of alternative electrodes suggested that these might perform better. To test whether this was true, a large randomized controlled trial was mounted to

compare the three electrodes (Rocket-Rolon, Hewlett-Packard and Surgicraft-Copeland) that are most commonly’ used within the UK.

Subjects and methods

The trial was conducted at the Maternity Unit of the Royal Berkshire Hospital during a 7-month period in 1988. The study design was approved by the research and ethics committee for the West Berkshire Health District.

The principal aims of the study were to assess which of the three electrodes (i) required least frequent reapplication throughout the progress of a woman's labour; (ii) best maintained a good quality trace of the fetal heart rate pattern; and (iii) caused the least trauma to the baby's scalp.

Steps were taken to minimize the possibility that previous experience with one particular electrode might influence the outcome of the study. First, the midwives were instructed in the correct method of application of each type of electrode; and second, during a 3-month training period before the trial each device was used systematically for one month according to the manufacturer's instructions.

All women, booked to be delivered in the hospital, were given a letter in the last trimester of pregnancy seeking their collaboration in the study. It mentioned that the standard policy of monitoring for low-risk labours was intermittent auscultation and emphasized that an electrode would only be applied if there were a medical indication for electronic fetal monitoring, for example, if labour was artificially induced or epidural analgesia was in progress. The women were also assured that none of the electrodes was known to be less reliable than the others and that a decision not to participate would not jeopardize their care in any way.

Recruitment took place on the labour ward. Women were eligible for entry to the trial if (i) the fetus was a singleton and presenting cephalically, (ii) the gestation was more than 37 completed weeks, and (iii) a decision had been made to apply a scalp electrode. At the time of the study, the routine policy in the hospital was to apply a scalp electrode once a decision had been taken to monitor labour electronically; an external transducer was only used if it was not possible to monitor labour satisfactorily with a scalp electrode. The monitors used were Hewlett-Packard 80/30As and 80/40As. Of the 3316 women who were delivered during the study period, 802 fulfilled these entry criteria and 780 (97%) of them were subsequently recruited to the study. The commonest reasons for non-recruitment were that the midwife forgot the trial, and that 'labour advanced too rapidly. Entry to the trial) which was signalled by opening the next in a series of consecutively numbered sealed opaque envelopes, occurred at the time a decision was made to apply a fetal scalp electrode.

Once the envelope seal was broken, the woman had irrevocably entered the trial. To minimize selection bias a woman remained in the study group allocated for the purpose of analysis, regardless of subsequent management. If a woman was recruited into the study but the envelope remained unopened, the midwife was asked to note the reason why.

Each envelope contained a random allocation to one of the following electrodes throughout the course of a labour Rocket-Rolon, Hewlett-Packard or Surgicraft-Copeland. Random allocation was computer-generated in balanced blocks that randomly varied in size between 3 and 15.

Data describing the women at trial entry and the subsequent performance of the scalp electrodes were recorded during labour. The midwives were asked to record the date, time and reason for the initial, and each subsequent, electrode application. They also noted the position and dilatation of the cervix, and whether or not an application was accomplished at the first attempt. The experience of operators was assessed by their status and by the number of times they had previously applied a particular electrode. The total number of vaginal examinations throughout labour, and whether or not a fetal blood sample had been obtained, were also recorded, as these procedures may affect the performance of an electrode or may themselves be affected by the performance of an electrode. Following delivery, the midwives noted any problems with removal of the device, and the degree of any trauma to the baby's scalp.

Scalp trauma was reassessed by community midwives (who were unaware of the group allocation) on the 1001 day postpartum, when the mothers were asked to record their experiences and opinions in relation to the use of the electrodes.

At the end of the study the quality of each cardiotocographic recording was assessed by an obstetrician (O.H.) who was 'blind' to the trial allocation. The recording for each stage of labour was graded for quality and total duration of trace. Each trace was divided into sections of 'good' quality, 'adequate' quality, 'poor' quality and 'external' (recorded via an abdominal transducer) recording. A good quality trace had minimal signal loss or artefactual recording, and could be interpreted easily. An adequate quality trace had some signal loss or arte-fact but could still be interpreted. A poor quality trace had so much signal loss or artefact that it could not be interpreted. The length of time in the trial with 'no tracing' was calculated by subtraction. The type and age of the monitor was also noted.

A sample size of 750 women was pre-specified. This was based on an expected reapplication rate for the Rocket-Rolon electrode of 50%. A trial of this size would have an 80% (β error 0-20) likelihood "C identifying a reduction to 37.5% associated with one or other alternative electrode (a = 0-05).

Table 1. Description of groups at trial entry	Rocket-Rolon (n=260)	Hewlett-Packard (n=261)	Surgicraft Copeland ·(n=259)
Maternal age (years) mean (SD]	26-3 (5-2)	27-0 (5-2)	26-3 (5-2)
Primiparous n (%)	174 (67)	180 (69)	178 (69)
Gestational age (weeks) mean (SD]	39.5 (1-7)	39-5 (1-9)	39-6 (1-5)
Birth weight (g) mean (SD]	3327 (532)	3349 (522)	3339 (522)
Length of labour before trial entry (mm) median	L84-5	178-0	199-0
(inter-quartile range)	(364~383-0)	(11-0 - 409-0)	(28-0 – 369-0)
Effective epidural n (%)	92 (35)	94 (36)	103 (40)
Reason for scalp electrode attachment n (%)
Auscultated FHR abnormality	109 (42)	109 (42)	120 (47)
Meconium	77 (30)	90 (34)	74 (29)
Intravenous oxytocin	61 (23)	46 (18)	50 (19)
Suspected IUOR	12 (5)	16 (6)	14 (5)
Other	1 (0)	0 (0)	1 (0)

ERR = fetal heart rate; IUGR intrauterine growth retardation.

Table 2. Details of first application	Rocket-Rolon (n=260) n(%)	Hewlett-Packard (n=261) n(%)	Surgicraft Copeland ·(n=259) n(%)
Electrode actually used
Rocket-Rolon	250 (96)	2 (1)	1 (0)
Hewlett-Packard	3 (1)	257 (98)	7 (3)
Surgicraft-Copeland	4 (2)	2 (1)	251 (97)
Not known	3 (1)	0 (0)	0 (0)
Operator
Midwifery sister	90 (35)	88 (34)	99 (38)
Staff midwife	104 (40)	96 (37)	92 (36)
Doctor	64 (25)	68 (26)	61 (24)
other	2 (1)	9 (3)	7 (3)
Operator experience (>20 applications)	150 (58)	135 (52)	139 (54)
Cervix posterior	86 (33)	88 (34)	89 (34)
Cervical dilatation <3cm	43 (17)	40 (15)	42 (16)
Head above ischial spines	201 (77)	192 (74)	196 (76)
More than one attempt	41 (16)	43 (16)	42 (16)
Application
Difficult	32 (12)	28 (11)	22 (8)
Impossible	15 (5)	11 (4)	7 (3)

Table 3. Description of the trial policies and performance of electrodes	Rocket-Rolon (n=260) n(%)	Hewlett-Packard (n=261) n(%)	Surgicraft Copeland ·(n=259) n(%)
Types of electrode actually used
Allocated electrode used throughout	232 (89)	249 (95)	247 (95)
Allocated plus other electrode	18 (7)	9 (3)	5 (2)
Non-allocated electrode used throughout	19 (4)	3 (1)	7 (3)
No. of applications per labour
1	161 (62 )	183 (70)	206 (80)
2	70 (27)	62 (24)	41 (16)
3	12 ( 9)	16 (6)	10 (4)
4	5 (2)	0 (0)	2 (1)
5	1 (0)	0 (0)	0 (0)
Total applications
Reason for first replacement
Poor or absent trace	26(10)	12 (5)	24 (9)
Detachment	64(25)	62(24)	25(10)
Other	.9 (3)	4 (2)	4 (2)
Performance of electrode
Not replaced
Detached <5 mm before birth, or after delivery	116(45)?	130(50)	142(55)
Detached >5 min before delivery	40(15)	46(18)	52(20)
Time of detachment not known	5 (2)	7 (3)	12 (5)
Replaced	99 (38) ‡	78 (30)‡	53 (20)‡

*Rocket-Rolon vs Hewlett-Packard, x² 7.00, P<0-01. Rocket-Rolon vs Surgicraft-Copeland, x² 6-87, P<0-0

†Rocket-Rolon vs Surgicraft-Copeland, , x² 5-41, P<0-05.

‡Rocket-Rolon vs Hewlett-Packard, , x² 3-90, P<0.05. Rocket-Rolon vs Surgicraft-Copeland , x² 19-44, P<0.01

‡Hewlett-Packard vs Surgicraft-Copeland, x' 6-12, P<0-02

Table 4. Labour, delivery and neonatal outcome	Rocket-Rolon (n=260) n(%)	Hewlett-Packard (n=261) n(%)	Surgicraft Copeland ·(n=259) n(%)
Vaginal examinations-mean (SEI	3-84 (0-09]	3.79 (0-09]	3-68 (04)9]
Fetal blood sample taken	13 (5)	16 (6)	7 (3)
Mode of delivery
Normal vaginal	163(63)	l73(~)	185(71)
Instrumental vaginal	69(26)	65(25)	50(19)
Caesarean	28(11)	25(10)	24 (9) -t
Apgar score
0 – 3	3 (1)	1 (0)	1 (0)
4 – 6	3 (1)	2 (1)	1 (0)
7 - 10	254 (98)	258 (99)	257 (99)
Electrode difficult to remove at delivery	12 (5)	3 (1)	10 (4)
Scalp trauma at delivery more than 'broken skin'	1 (0)	1 (0)	4 (2)
Scalp bleeding following delivery	16 (6)	13 (5)	17 (7)
Scalp trauma visible at 10 days postpartum	38/217 (18)*	59/224 (26)*	56/206 (27)*

*Rocket-Rolon vs Hewlett-Packard x² = 5-0l P = 0-025. Rocket-Rolon vs Surgicraft-Copeland. x² =5-21.P<0-025.

Results

Table I describes the 780 women recruited to the trial and demonstrates the comparability of the three groups generated by randomization. Overall, the main reasons for the initial scalp electrode attachment were a fetal heart rate abnormality suspected from external auscultation (43%), meconium staining of the amniotic fluid (32%), and intravenous oxytocin for induction or augmentation of labour (21%). Fourteen electronic fetal heart rate monitors were employed during the study and these were used equally frequently in the three trial groups.

Similar numbers of women in the three groups received the allocated electrode at the first application Table 2). The type and previous experience of the operator, and the degree of difficulty of application, as judged by the position and dilation of the cervix and by the station of the head, were also similar (Table 2). More than one attempt was needed for application for 16% of women but this did not differ between the groups. Application of the Rocket-Rolon electrode, however was more often described as 'difficult' or proved impossible (Table 2) (odds ratio vs Surgicraft-Copeland = 1.75, 95% CI 1.07-2-82; P<0-05); and women allocated to the Rocket- Rolon were significantly more likely to be monitored with one of the other electrodes at some time during labour.

Women in the Surgicraft-Copeland group were significantly least likely to need a reapplication (odds ratio vs Hewlett- Packard = 0-61,95% CI 0.4143-90; odds ratio vs Rocket-Rolon = 0.43,95% CI = 0-2~4)-62) (Table 3). The difference in reapplication rates persisted: a third application was necessary to 12 allocated to the Surgicraft-Copeland, to 16 allocated to the Hewlett-Packard, and to 29 women allocated the Rolon electrode

The Surgicraft-Copeland was much less likely to become detached (odds ratio vs Hewlett-Packard = 0-36,95% CI 0.2~ 0-58; odds ratio vs Rocket-Rolon =0-35,95% CI =0.2243- whereas the Hewlett-Packard was least likely to be replaced signify because of poor or absent fetal heart rate trace (odds ratio vs Surgicraft-Copeland =049,95% CI 0-2543-95; odds ratio vs Rocket-Rolon = 0-45, 95% CI 0-2Y4>87), This pattern of reasons for reapplication also persisted

The rate of success of the electrode in terms of persistent attachment to the fetal scalp up to at least 5 mm before delivery was greatest in the Surgicraft-Copeland group (Table 3), and this was significantly higher than in the Rocket-Rolon group -(odds ratio = 0-67.95% Cl 0-4743-94). The differences between the electrodes in the need for reapplication were reflected in the frequency of vaginal examination in the trial groups (Table 4); when compared with the Surgicraft-C6peland group, there were an extra 11 examinations for every l00~women monitored in the Hewlett-Packard group, and an extra 16 examinations per 100 women in the Rocket-Rolon group. The differences were not- however, statistically significant

There were no clear differences between the groups in the use of fetal blood sampling, mode of delivery and low Apgar score, although fetal blood sampling, instrumental delivery and low Apgar score were least common in the Surgicraft Copeland group (Table 4).

There was no serious scalp trauma 'Difficult removal' was least often reported for the Hewlett-Packard electrode (odds ratio vs Surgicraft-Copeland =0-33, 95% CI 0.1143.99; odds 55) ratio vs Rocket-Rolon =0-29,95%CI0-lO4)-81). Marks at the site of electrode insertion 10 days after delivery were significantly least commonly reported- in the Rocket- Rolon group (Table 4) (odds ratio vs Surgicraft-Copeland = 0-st95% 'CI 0-36--0-90; odds ratio vs Hewlett-Packard = 0-60, 95% a. 0-3843-93).

The type of fetal heart rate monitoring and the quality of the traces recorded using scalp electrodes, a-re summarized in Table 5. There was no difference between the electrodes in the c length of interpretable trace ('good' and 'adequate' combined). However, interpretable traces were more often of 'good' quality in the Hewlett-Packard group. Similarly, there was no difference between the electrodes in the length of interpretable trace ('poor' quality or no trace), but an external transducer was used significantly least often in the Surgicraft-Copeland group.

Table 5. Type and quality of fetal heart rate tracing	Length of trace (mm)
	Rocket-Rolon	Hewlett-Packard	Surgicraft Copeland
Good or adequate quality tracing from scalp electrode	163 (80,261)	164 (78.291)	157 (89, 280)
Poor quality tracing or no trace	23 (1, 44)	19 (8, 39)	22 (10, 43)
Via external transducer*	0 (0, 14)	0 (0, 13)	0 (0, 0)

*Rocket~Rolon vs Surgicraft-Copeland. Mann-Whitney-U =25054 P<0-05.

'Hewlett-Packard vs Surgicraft-Copeland. Mann-Whitney-U =24032, P<0-05.

Results are median (inter-quartile range) values.

Questionnaires completed at 10 days after delivery were returned by 647 (83%) women. Similar numbers in the three trial groups remembered the electrode attachment (90%); 44 (7%) described it as 'moderately uncomfortable' and 42(6%) as 'severely uncomfortable'. Fewer women who had been allocated to the Rocket-Rolon group found monitoring reassuring (61% vs 66% in the Hewlett-Packard group, and 67% in the Surgicraft-Copeland) and fewer would 'definitely have-a fetal scalp electrode again' (43% vs 49% and 51%, respectively); but these differences were not statistically significant

Discussion

This study was initiated following our observation that fetal scalp electrodes frequently needed to be replaced This had raised concerns about the quality of continuous fetal heart rate monitoring in the -hospital, about use of midwives' time to replace electrodes, and about possible adverse effects on the baby.

We believe that we have addressed these concerns reliably. We ensured before the study that those responsible for intrapartum monitoring were trained to use each of the electrodes interchangeably. Random allocation generated groups that were comparable in important respects (Table 1); compliance with the allocated electrode was satisfactory; those applying the electrodes were of similar status; and the application presented similar difficulties as judged by the position and dilatation of the cervix and the station of the fetal head (Table 2). Furthermore, the trial was sufficiently large to ensure identification of clinically important differences between the electrodes, and outcome was assessed 'blind' to the allocation, where possible.

We identified -clear differences between the electrodes in respect of the need for replacement this applied to 38% of those allocated to the Rocket-Rolon compared with 30% allocated to the Hewlett-Packard and 20% allocated to the Surgicraft-Copeland electrode. This was reflected in differences in the total number of replacements (Table 3), and in the number of electrodes that remained attached up to (or nearly up to) delivery Table 3). The reason why the Hewlett-Packard per-formed better than the Rocket-Rolon was that it was less likely to require replacement because of a poor quality or absent trace. The Hewlett-Packard was also better than the Surgicraft-Copeland in this respect but this advantage was more than offset by the fewer replacements of the Surgicraft-Copeland because the electrode had become detached Table 3). These differences between the electrodes persisted at later attachments.

The differences were also reflected in the type and quality of fetal heart rate monitoring. The Hewlett-Packard tended to produce more good quality traces compared with the Surgicraft-Copeland; nevertheless, the total time with an interpret-able trace ('good' or 'adequate quality) was similar in the groups (Table 5). Reason to monitoring with an abdominal transducer was least common in the Surgicraft-Copeland group.

There have been reports of serious neonatal complications of fetal scalp electrodes (Overturf & Balfour 1975; Plavidal & Werch 1976; Okado & Chow 1977; D'Souza et at. 1982: Cordero er at. 1983). In this trial, however, there was no serious scalp trauma, and bleeding from the scalp after delivery was equally common in the three groups (Table 3). The only advantage for the Rocket-Rolon identified in the trial was that it was least likely to leave a mark on the scalp that was still visible 10 days after delivery.

The purchase price (1990) of each electrode is as follows:

Rocket-Rolon £3.65; Hewlett-Packard £4.42; and Surgicraft-Copeland £4.00. Taking into account the difference in the need for reapplications, the avenge costs of electrodes per woman9 as allocated, were £5j5, £6.01 and £5.03, respectively. This calculation does not take account of other differences in direct costs, such as those incurred in vaginal examinations (about 20 fewer per 100 women monitored using the Surgicraft-Copeland), and possibly in caesarean or instrumental deliveries.

The trial has demonstrated clinically important differences between the performances of the three electrodes. The findings are consistent with the results of previously reported small trials (Calvert & Newcombe 1980; Nickelsen et al. 1989). The Surgicraft-Copeland was significantly least likely to become detached. This led to fewer vaginal exanimations to apply a new electrode. In this respect, it was clearly preferable to the Hewlett-Packard and Rocket-Rolon both to clinical care givers and to the women.

On the basis of the results of this trial, we feel that the Surgicraft-Copeland is the electrode of choice.

Acknowledgments

We thank the many midwives and obstetricians in the West Berkshire Health District who worked so hard to make this study a success; the mothers who responded so enthusiastically to requests for information about their views; Pamela Dean for clerical assistance, Carole Harris for data entry, and Pauline Carter for typing the manuscript Funding for this study was obtained through the locally organized research scheme (Oxford Region). The National Perinatal Epidemiology Unit is funded by a grant from the Department of Health.

References

Calvert I. P. & Newcombe R. Q. (1980), Which fetal scalp electrode?

· Lancet i, 371.

Cordero L, Anderson C. W. & Zuspan. R (1983) Scalp abscess: a benign and infrequent complication of fetal monitoring. Am J

Obstet Gynecol 146, 126-130.

D'Souza S. W. Black R & MacFarlane T. (1982) Fetal scalp damage and neonatal jaundice: a risk of routine fetal scalp electrode monitoring. J Obstet Gynaecol 1, 161-164.

Ghosh A- K. & Tipton It H. (1976) Fetal scalp electrode. Lancet i, 1075.

Nickelsen C., Weber T.. Parnell C., Nim 1., Kemp A. M. & Junge I (1989) Cardiotocographic monitoring of deliveries. A prospective investigation of two types of electrodes Ugeskr Laeger 151, 44(M42.

Okada D. M. & chow A- W. (19fl) Neonatal scalp abscess following intrapartum fetal monitoring: prospective comparisons of two spiral electrodes. Am J Obstet Gynecol 127, 875-878.

Overturf G. D. & Balfour 0. (1975) Osteomyelitis and sepsis: severe complications of fetal monitoring. Paediatrics 55,244247.

Plavidal 1. & Werch A. (1976) Fetal scalp abscess secondary to intrauterine monitoring. Am J Object Gynecol 125, 65-70.

Correspondence: Dr A. Grant, Perinatal Trials Service, National Perinatal Epidemiology Unit, Radcliffe Infirmary. Oxford 0X2 6HE.

Received 26 April 1991

Accepted 25 October /991