The Bristol Third-Stage Trial
by John Stevenson
© 2005 Midwifery Today, Inc. All rights reserved.
[Editor's note: This article first appeared in Midwifery Today Issue 73, Spring 2005.]
If you set out to compare a policy of intensive, precipitous intervention with a policy of sitting back and watching the patient bleed, obviously the former will be seen to be safer. This was the approach and conclusion of the Bristol Third-Stage Trial. But the trial, based on false premises, is completely misleading, and numerous criteria are mistaken, misunderstood or misinterpreted.
The British Medical Journal of 19 November 1988 published a six-page report of an important trial conducted in Bristol Maternity Hospital concerning active versus physiological management of third stage of labour. The stated objective was to compare the effects on foetal and maternal morbidity of routine active management with expectant or physiological management, particularly regarding postpartum haemorrhage. The trial was discontinued early because of excessive incidence of serious haemorrhages in the physiologically managed group. Analysis of the results showed that the policy of active management not only reduces hemorrhages, but has the additional benefits of shortening third stage and reducing neonatal packed cell volume. This trial clearly appears to be disastrous for the proponents of restrained management.
This research report is important because of its far-reaching implications. From then on, obstetricians all over the world would tell their patients they will not stand for any nonsense about third-stage management, because a carefully planned, expertly structured and thoroughly monitored trial has proved conclusively that routine active management of third stage is safe and efficient, whereas the unorthodox requests of the minority of nuisance patients for physiological management have been proved to be dangerous; so much so that the trial had to be stopped before it got halfway through its planned course. Any patient who wants to continue arguing will be referred to this report, where she will get bogged down amongst overpowering rationalizing and scientific jargon.
The trial was planned and set up by the Academic Department of Obstetrics and Gynaecology (whose members are experts at obstetrical intervention), with the help of a research midwife and a social statistician. It was a progressive randomized trial, in which women who were expected to deliver vaginally were allocated randomly to either active or physiological management. As each woman approached delivery, a sealed envelope was opened, revealing the group to which she was allocated. Active management consisted of injection of oxytocin as the anterior shoulder was delivered, immediate clamping and cutting of the cord and delivery of the placenta by cord traction. The physiological group was managed by NOT giving oxytocin prophylactically, leaving the cord intact until the placenta was delivered, refraining from cord traction or any manual interference with the fundus, encouraging early breastfeeding and encouraging a posture favourable for placental delivery.
The purpose of the trial, according to the authors, was to test contentions published in medical journals that questioned the scientific bases of certain obstetrical routines.
The obvious alternative to active management is passive management, to see whether a woman's body can cope without assistance; that should settle the matter conclusively. Here is a very common blunder in logic: to go from the sublime to the ridiculous, to justify one inordinate extreme by ridiculing the opposite inordinate extreme.
To devise the protocol for physiological management, because there is no standard definition of it, the research midwife took advice from "certain midwives known to practice it" and then trained the labour-ward midwives before the trial commenced. I question whether a research midwife from the hospital establishment could faithfully obtain a clear assessment of methods used by midwives outside the system and then responsibly reduce those methods to protocol rules and teach them to sceptical labour-ward staff in the hospital setting. Those "certain midwives known to practice it" were not called in to check the finalized protocol for "physiological management."
Please note also another vital missing factor—the patients' wishes. The trial was planned and controlled by experts without any input from any patient. The correct and valid basis for this trial should have been patient participation, one group wanting a share in the decision-making and the other content to leave all decisions to the staff.
The mother has a very basic prerogative. She has grown this child within her body, and when she has brought him forth she will continue to nourish him at her breast. Of all the notable events in the life of her child, the most momentous for her is his birth. She needs and deserves attendants who are sympathetically involved, emotionally supportive and respectfully helpful—and this especially concerns her medical assistants.
Homebirth is so safe because we use common sense rather than rule of thumb, and we tackle haemorrhages in good time but by gentle methods. The organizers of this trial could have and should have seen that it was wrongly conceived and would need to be stopped because of disastrous results.
A pilot study was conducted for six weeks in November and December 1985 to ensure that the management was feasible and to train the staff. The trial proper commenced in January 1986. By May 1986 excessive haemorrhages in the physiological group caused the management committee to modify the protocol by increasing the criteria for exclusion from the trial, clarifying the conditions for transferring patients from the physiological to the active group and increasing encouragement of attempts at early breast feeding. However, the data monitoring committee became increasingly concerned about the alarming haemorrhages in the physiological group, until in January 1987 they recommended that the trial be stopped.
Virtually all the women allocated to active management received it, but less than half the women allocated to physiological management were successful, the rest needing intervention. Patients had been reclassified as low-risk if labour was spontaneous, duration less than twelve hours and delivery spontaneous, but higher-risk if induction, augmentation or epidural was "necessary" and if labour was longer and "needed" assistance. Further analysis showed that, even then, active management was "preferable," regardless of first- and second-stage criteria. In fact, if anything, and contrary to prior expectation, the advantage of active management was consistently greater in the low-risk group.
Of those allocated to the physiologically managed group, many failed when the cord was clamped and cut immediately, either because it was tight around the baby's neck or because the amniotic fluid was meconium-stained. I believe the cord should not be cut for either of these indications. In the rare event of baby needing resuscitation, this can be done effectively and efficiently without cutting the cord. In any case, the cord should never be cut until it is empty and has ceased pulsating.
The reason why 20% of the physiological group received prophylactic oxytocin is not stated. Most of the "failures" in this group obviously occurred because those midwives were so experienced and confident with "active" management. The authors predictably conclude that the policy of active management is justified, and not just in terms of the primary criterion of blood loss exceeding 500 ml. The increased risks of "physiological" management applied also to maternal blood loss greater than one litre, the "need" for blood transfusions, longer third stage and raised neonatal packed cell volume.
I will first consider the complaint that physiological management causes third stage to last longer—more than half an hour in 25% of cases. Homebirth practitioners perceive baby and the birth process to be far more valuable than their valuable time, and they do other things while watching out for haemorrhage, like looking to mother's comfort, writing up their notes or just admiring baby. Why is third stage longer than two minutes deemed a disadvantage? The policy of the College of Obstetricians is delivery of babies safely and efficiently. We argue that their concept of efficiency is incompatible with safety, and I suggest that the superior safety of homebirth is precisely due to our policy of patience and careful restraint, which the researchers should investigate and evaluate.
Secondly, I will discuss the problem of excessive haemorrhages. This report states the blood loss in volume categories, as though it were always a bad thing, assuming that the human body is so badly designed that it needs obstetricians to apply "active" management to save women's lives. While I flatly object to their "active" management, and I also strongly disagree with the so-called "physiological" management used in the Bristol trial, neither do I agree with their assessment of blood loss, as though it were always bad. A woman's body produces over a litre of extra blood during pregnancy. A loss of two or three deciliters with no disability is quite normal. Some women lose a litre without disability. There is not a single mention in this report of how a patient was affected by blood loss, nor of the criteria on which a transfusion was deemed necessary.
When there is minimal bleeding, or none at all, there is very good reason not to disturb the placenta. Besides allowing the parents to contemplate, caress and bond with baby undistracted, it is important to remember that the part of the uterine wall underlying the placenta is consolidating and becoming less likely to bleed, so that after half an hour, haemorrhage becomes so unlikely that one could regard it as effectively prevented. As a very basic safety measure, this principle could be formulated: In the absence of significant bleeding or expulsive contractions or cervical discomfort, the placenta should be left undisturbed for a good half hour. After that, it can be delivered when the mother wishes.
When a haemorrhage threatens, the correct policy is, first, deliver the placenta; second, help the uterus contract by massaging it directly. (This is preferable to nipple stimulation, though that might help. But don't expect baby to do it in the first half hour—rather, ask the husband.) Third and lastly, give oxytocic injections—rarely needed. At the first convenient opportunity after the birth, the underpad should be changed, so that amniotic fluid and meconium are removed, and any blood loss will be obvious. If bleeding occurs, it should be scooped up and placed in a measuring-jug, and when the cumulative loss approaches half a litre, it is time to warn the mother that the placenta needs delivering. The reason for instigating the action before the loss reaches 500 ml is that by the time the placenta is delivered, particularly if it is stubborn, and when the blood in the membranes is included, the total loss is considerably more, possibly even double that.
Let us thirdly look at the problem of increased neonatal packed cell volume, which implies that baby's circulation is overloaded with too many red cells. The babies of the physiological group averaged 85 g heavier than those of the active group. Three times as many had a packed cell volume of over 0.65, and they were much more likely to have blood specimens taken to check for polycythaemia or jaundice.
When the cord is not clamped immediately, allowing the placenta to drain, baby receives about half to one cupful of its own blood as drainage from the placenta. The experts feel they are doing baby a favour by depriving it of this extra blood, thus reducing jaundice and possibly increasing other benefits, especially saving time. There is no way I will accept without evidence that baby is better off without that extra blood. I believe mild or moderate physiological jaundice is absolutely no threat to a healthy breastfed baby. Ultraviolet lights are unnecessary. I am sure baby makes good use of that extra blood; it is a valuable body substance. There is also a glaring inconsistency in the way obstetrical experts view baby's surplus blood from the placenta (which they blithely discard), compared with mother's surplus generated during pregnancy but redundant once baby is born (which they strain to conserve). I say, take it easy; the body's natural processes handle all that.
The statistician, nowadays retitled "epidemiologist," without noticing the abysmal flaws in the structure of this trial, has dressed it in jargon, giving it an air of respectability and credibility.
The Bristol Third-Stage Trial, conceived and executed by "experts" with no patient input, is very bad news for proponents of gentle, optimal birth. It rationalizes and justifies the alleged safety and efficiency of the horrendous "active management" of third stage. No woman who has experienced or witnessed or read about our careful, gentle management of third stage would agree with the standard "active" management. Good mothering is enhanced and facilitated by inner peace, confidence and bonding to baby, and these are not favoured by thoughtless, hasty obstetrical routines.
Editor's Note: For more about natural hemorrhage prevention, see Marion's Message, "Unchanging Protocols," page 9.
John Stevenson lives in Maryborough, Victoria, Australia, where he practiced medicine for 20 years before giving up general practice to focus on homebirth. He is currently working on a book about childbirth management.
If you enjoyed this article, you'll enjoy Midwifery Today magazine! Subscribe now!