Among the catalogue of procedures designed to improve pregnancy outcome cervical cerclage is one of the most controversial. Its purpose is clear, namely to prevent or reduce the incidence of miscarriage from a presumed diagnosis of ‘cervical incompetence’. On the face of it this seems entirely logical if the obstetrician believes that the cervix is likely to fail in its duty of remaining long and closed to retain the fetus in utero until the time is right for delivery. However, our ability to make a secure diagnosis of the condition is fraught with difficulty and our ‘belief’ may be built on false premises. The condition remains poorly understood and for upwards of 50 years investigators have striven to devise a reliable means of diagnosis. The clinician has often had to fall back on the clinical history of the pattern of previous pregnancy losses but even when there is seemingly compelling evidence of cervical incompetence uncertainty remains as to the best remedy amid a disappointing lack of persuasive evidence of benefit.
None of this appears to have discouraged clinicians from employing this intervention. Indeed it has often been grossly over used. In its simplest form it may seem a relatively harmless procedure but it is by no means without risk and needs to be employed with careful circumspection. One may sympathize with the clinician who, faced with a woman who has endured a succession of harrowing pregnancy losses, feels compelled to do something rather than nothing; but this is fuelled by an unfortunate paradox. The more sutures inserted in women who do not in fact have the condition, the better may the results appear to be because many of those women will have a good outcome despite the intervention. Furthermore they are likely to believe that the intervention saved the pregnancy. It is therefore imperative that the clinician should make an honest appraisal of each case before embarking on a procedure which, although seemingly innocuous, may in some cases expose the woman to serious risks.
The surgical options range from the comparatively straightforward insertion of a suture around the portio vaginalis of the cervix to insertion of one at the level of the fibro-muscular junction between the uterine corpus and cervix ( Fig 28-1 ). The latter is clearly a much more elaborate intervention, requiring a laparotomy. Not only does this cause pain and discomfort to the woman, but it requires a higher degree of experience and skill of the operator, and carries more serious implications for the subsequent management of the pregnancy. This will almost certainly mandate delivery by caesarean section.
A transvaginal approach is most commonly employed. Because of the manner in which effacement of the cervix occurs (see Chapter 1 ) it will be obvious that a suture placed too low on the portio vaginalis will be worse than useless. It is therefore important to place it as high as possible and this requires that effective traction be applied to the cervix. It used to be fashionable to employ sponge holding forceps for this in the belief that they would be atraumatic but experience has shown that they tend to slip off and to avulse the cervical epithelium in the process. A single tooth vulsellum is greatly to be preferred and while this may result in some bleeding this should not be significant. The advantage of an effective grip on the ectocervix greatly outweighs any concern about bleeding. The suture should be placed as high as feasible and some clinicians feel justified in opening the anterior vaginal fornix and pushing up the bladder in order to achieve as high a placement as is possible by a transvaginal approach.
Although some clinicians have considered inserting a suture between pregnancies this has little or nothing to commend it. In the first place, it may interfere with conception and secondly there is no merit in the presence of a suture in a woman destined to miscarry in the first trimester. Such miscarriages cannot be prevented by cervical cerclage. The appropriate timing for insertion of an elective suture is between 10 and 14 weeks’ gestation after an ultrasound scan has confirmed that the pregnancy is progressing normally.
Assuming the pregnancy has progressed satisfactorily it is usual to remove transvaginal sutures at around 37 weeks. In the event of spontaneous miscarriage or antepartum haemorrhage the suture may have to be removed earlier. In the case of transabdominal sutures it is usually possible to divide them with scissors introduced through the posterior fornix in the midline. This begs the question why this should not also be done at 37 weeks but it is considered better to deliver such pregnancies abdominally. The suture may then be left in place where it may serve again in a subsequent pregnancy.
Materials and Management
In past decades an exotic catalogue of different suture materials has been used including strips of the woman’s own fascia lata, tantalum wire and all manner of synthetic, non-absorbable materials. We feel there is no need to employ any other material than mersilene tape, which serves well regardless of the route of insertion. It has become conventional to cover the insertion period and subsequent few hours with an inhibitor of prostaglandin synthesis such as indomethacin and that particularly applies in the case of a decision to attempt a ‘rescue’ suture when the cervix has already dilated and allowed the amniotic sac to bulge into the vagina. When this occurs in the absence of uterine contractions an attempt of this sort may be worthwhile. It may seem that the cervix is already irrevocably dilated and delivery imminent, but this is by no means always so. With the sac filling the vagina it is easy to assume that the cervix is fully dilated but often there has been an ‘hour glass’ prolapsing of the sac through an only partially dilated cervix. In such circumstances strategies such as employing a steep head-down tilt, a balloon to persuade the sac back into the uterus and perhaps transabdominal amniocentesis to reduce the amniotic fluid volume may improve the prospects of success. Such cases are fraught with difficulty because an attempt at rescue may result in rupture of the membranes, leaving the clinician to feel that conservative management might have been preferable, but one has to consider all the risks and try to arrive at a balanced decision.
It is important to stress the danger if labour should become established with a suture still in place and the woman must be advised that, if that should happen, its prompt removal is essential. The risks include further damage to the cervix if the suture tears out or rupture of the uterus. The latter is a particular hazard if there is a previous caesarean scar. There have also been cases reported of annular detachment of the cervix and of delivery through the posterior vaginal fornix bypassing the cervix. A special warning should be highlighted of a circumstance where a second higher suture has been inserted because the original one has been deemed inadequate but left in place. We know of an instance in which a suture was later removed without recognizing that the other remained in place; a disastrous outcome was only narrowly avoided. This also highlights the need for accurate and prominent recording in the case records of the presence of sutures and their location, together with clear instructions about their subsequent management.
In certain obstetric complications, rapid uterine relaxation is required for their successful resolution. Until relatively recently acute uterine tocolysis required the institution of general anaesthesia with halogenated agents and involved rapid sequence induction and tracheal intubation with its associated maternal risks. The recent application of β-adrenergic drugs, oxytocin antagonists and nitroglycerine to achieve rapid uterine relaxation has obviated the need for general anaesthesia in these circumstances.
‘I found it impossible to get my hand into the uterus to deliver the placenta. Bearing in mind the remarkable power which nitrite of amyl possesses in relaxing tension in the blood-vessels, I determined to test its action on the uterine spasms. The patient had three drops of the nitrite of amyl given her on a handkerchief to inhale. During the inhalation, the ring of muscular fibres around the os interna, which had been so rigid as to be absolutely undilatable, steadily yielded, until I could pass the whole hand into the uterus …’
Hour-glass contraction of the uterus treated with nitrate of amyl. BMJ 1882; 1:377
Nitroglycerine is an ester of nitric acid and exerts its relaxant effect on smooth muscle by the formation of nitric oxide. The drug is rapidly metabolized by the liver so that the half-life is only 2 minutes. It has a low molecular weight of 227 and therefore crosses the placenta, but no adverse fetal or neonatal effects have been described in either animal studies or humans. In addition to uterine tocolysis the smooth muscle relaxation effect causes peripheral vasodilatation and reduced venous tone. Provided there is no hypovolaemia this causes only a mild and clinically insignificant hypotension. However, if there is associated hypovolaemia, rapid infusion of intravenous crystalloid is necessary to avoid acute and severe maternal hypotension. The peripheral vasodilation responds to epinephrine (adrenaline) and the uterine relaxation responds to oxytocin.
The biggest experience with this group of drugs has been with ritodrine, terbutaline and hexoprenaline. These selective β-2 receptor agonists have the main effect of relaxing arteriolar and uterine smooth muscle. These drugs are given subcutaneously or by slow intravenous injection diluted in saline; they do not inhibit uterine activity when given as nebulizers. They may cause mild hypotension and tachycardia and there have been isolated reports of atrial fibrillation. These agents cross the placenta and may cause a mild transient fetal tachycardia. They abolish uterine activity for 15–30 minutes, and their effect can be reversed by β-blockers.
Atosiban is a synthetic analogue of oxytocin that acts as an oxytocin antagonist by binding to myometrial cell oxytocin receptors. The half-life is about 12 minutes and it crosses the placenta, but umbilical vein levels are only 10% of those in the maternal uterine vein. The advantage of atosiban is that it has fewer cardiovascular side effects compared with the β-adrenergic agents. Atosiban has been used mainly in attempts to suppress preterm labour but it may also have an application for acute tocolysis in intrapartum fetal distress.
Rapid uterine relaxation may be necessary in the following clinical situations.
Excessive Uterine Action
Excessive uterine action in labour is defined as more than five uterine contractions in a 10-minute period averaged over 30 minutes (tachysystole) or contractions exceeding 2 minutes in duration (hypertonus). This uterine hyperstimulation is most likely to occur in response to oxytocic drugs, either oxytocin or, more likely, the longer acting prostaglandins used for cervical ripening and induction of labour. Although less common, tachysystole and uterine hypertonus can occur in spontaneous labour and more so with pathological conditions such as placental abruption.
Unless the uterus is relaxed there may be difficulty with delivery of extended arms and/or the after-coming head of the preterm breech. This can occur with both vaginal and caesarean delivery but is more likely with the latter. The fetal breech, trunk and legs may be easily delivered through the uterine incision because of their smaller diameter, but the unrelaxed uterine muscle may clamp down around the fetal head, leading to delay, asphyxia and potential trauma. This is particularly likely in the preterm breech, but can also occur with the term fetus.
External Cephalic Version
Short-term uterine relaxation may help facilitate external cephalic version of breech presentation at term. In this context the beta-adrenergic drugs are more effective than nitroglycerine, because of their longer duration of action.
Intrapartum Version of Fetal Malpresentations
Uterine relaxation is usually necessary to safely and effectively carry out both external cephalic and internal podalic version at both vaginal and caesarean delivery of the second twin (see Chapter 17 ). With regional anaesthesia uterine relaxation is often inadequate and additional tocolysis is necessary.
If uterine contractions interfere with cord decompression after cord prolapse, acute tocolysis may be needed. This is more likely if there is an anticipated delay in the diagnosis-to-delivery interval (see Chapter 18 ).
In rare cases of shoulder dystocia that cannot be resolved with the traditional manoeuvres, cephalic replacement may be considered (see Chapter 12 ). In this situation, replacement of the head is facilitated by uterine relaxation, which may also help restore the utero-placental circulation and fetal oxygenation.
Acute tocolysis may aid relaxation of a contraction ring and allow spontaneous delivery of the separated but retained placenta. On other occasions, tocolysis may be necessary to facilitate the manual removal of a non-separated placenta, in which a contraction ring prevents manual access.
Acute Uterine Inversion
Manual replacement of acute uterine inversion may be possible using acute tocolysis, obviating the need for general anaesthesia.
In many of the above situations, regional anaesthesia in the form of epidural or spinal analgesia may be in effect. However, these regional techniques, while providing excellent analgesia, do not provide uterine relaxation, so that additional tocolysis may be necessary.
Administration of Tocolytics
This can be given by the sublingual or intravenous routes. Sublingual administration is via an aerosol spray in doses of 400µg. However, the mucosal absorption is not as predictable or precise as the intravenous route and the latter is usually chosen for acute tocolysis.
Nitroglycerine comes in an ampoule containing 5 mg in 1 ml solution. If this is added to a 100 ml bag of normal saline it produces a solution of 50 µg per ml. Draw up 20 ml into a syringe – this allows the precise titration of 50 µg per ml administered. Nitroglycerine is rapidly degraded and its effect is usually obvious within 90 seconds and lasts for a further 1–2 minutes. At the time of its administration intravenous crystalloid should be running rapidly, particularly if there is any question of hypovolaemia. The dose is titrated depending on the initial response and the clinical situation. For cases of fetal entrapment a rapid response is necessary and one usually starts with a dose of 200 µg, repeating this at about 2-minute intervals until appropriate uterine relaxation is achieved. In cases with retained placenta or acute uterine inversion, hypovolaemia should be corrected and smaller initial doses (100 µg) may be given.
In all cases in which oxytocin or prostaglandins have been previously administered, higher doses of nitroglycerine may be required.
It is most commonly given as 250 µg subcutaneously, but can be given intravenously in 5 ml saline, slowly over 5 minutes. The antagonist to terbutaline is propranolol 1–2 mg IV.
This is given as 6 mg mixed in 10 ml normal saline and administered intravenously over 3 minutes.
Give 5 µg in 10 ml normal saline intravenously over 5 minutes.
Mix 6.75 mg of atosiban in 5 ml normal saline and give intravenously over 1 minute.
For rapid short-lived tocolysis necessary to deal with malpresentations and third stage complications, nitroglycerine is the drug of choice. For more sustained uterine relaxation needed with uterine hypertonus or external cephalic version atosiban or terbutaline are more appropriate.
Version refers to the procedure that changes the presenting part of the fetus. This may be carried out by external manipulation of the fetus, internal manipulation or combined ( bipolar ) internal and external manoeuvres. Bipolar version is discussed in the chapter on antepartum haemorrhage.
External Cephalic Version
External cephalic version (ECV) is used to alter the presentation of the fetus in the later weeks of pregnancy. Breech presentation, transverse or oblique lie may be converted to cephalic presentation. This may also be achieved in early labour in some women, particularly the multiparous. Following delivery of the first twin, ECV may be undertaken for breech or oblique lie of the second twin.
‘In all cases, where, after eight months of gestation the head occupies either the iliac fossa or the superior uterine segment, we should perform cephalic version by external manoeuvres … The first stage of the operation consists in rendering the foetus moveable … To facilitate this displacement of the breech, we may at the same time exert slight pressure in an opposite direction over the cephalic extremity. The two fetal extremities being moveable and accessible, and the hands being applied over them, we should make slow and continued pressure in such a manner as to cause the breech to ascend and the head to descend by the shortest route.
A Treatise on Abdominal Palpation, as Applied to Obstetrics, and Version by External Manipulations. English Edition. New York: J. H. Vail & Co, 1885, p75–78
Potential associations with malpresentations such as placenta praevia should be considered and excluded.
For antepartum cases it is usual to delay version until 36−37 weeks’ gestation. From 37 weeks only 5–10% will revert to breech presentation after successful version. In addition, rare complications of version may necessitate immediate delivery, which is more acceptable if the fetus is mature. However, a randomized-controlled trial comparing early ECV at 34−36 weeks with ECV at 37−38 weeks showed more cephalic presentations at delivery in the early ECV group but no difference in caesarean delivery or preterm birth.
A detailed ultrasound examination is undertaken to confirm gestational age, exclude fetal anomalies, identify the type of breech presentation, locate the placenta and assess the amount of amniotic fluid.
If the woman has a single previous transverse lower segment caesarean scar, and is planning labour and vaginal delivery, it is permissible to attempt gentle ECV in selected cases.
Some obstetricians feel that these patients should be fasting because of the rare cases that require immediate delivery due to complications of the version. It is unpleasant for pregnant women to fast and, in addition, the fetus is often very quiet with maternal fasting. As it is helpful to have fetal movement to assist the version (see below) a reasonable compromise is to offer the woman a light snack 1–2 hours before the planned version.
Ideally, a 20 minute pre-version cardiotocograph should show a reactive fetal heart rate.
Perform an ultrasound scan to confirm the malpresentation, localize the placenta and determine the amount of amniotic fluid.
The woman should be reasonably comfortable with slight elevation of the head of the bed and a minor left lateral tilt. After the ultrasound, remove the gel from the woman’s abdomen and substitute talcum powder. This allows one’s hands to move easily and smoothly over the abdomen and to avoid the use of excess force. It also facilitates the changing positions of the hands during the version. Others may use extra gel for the same purpose.
It is usual to turn the fetus by promoting flexion, so that the fetus is moved in the direction it is ‘looking’. This principle should be explained to the woman. Tell her she will feel pressure from your hands and that this will be sustained but she should not feel pain.
The essential first step, upon which success depends, is the displacement and elevation of the breech from the pelvis. This can require sustained, albeit gentle pressure from both hands ( Fig 28-2 ). Once this has been achieved the right hand continues to elevate the breech while the left hand is moved behind the fetal head ( Fig 28-3 ). With the hands working in unison, intermittent pressure is applied to both fetal poles. It is at this point that fetal movement may help propel the fetus in the right direction. Throughout the procedure, gentleness is the guiding rule. Discretion is the better part of valour and the procedure is either going to be achieved gently and without pain, or it should be abandoned. If these principles are followed it is extremely rare for there to be any maternal or fetal complications.
Once the fetus has been turned ultrasound is again used to confirm the new presentation and assess the fetal heart rate. It is not uncommon for there to be transient bradycardia due to the manipulations. Some will use the ultrasound or doptone over the fetal thorax to monitor the fetal heart rate during version. After version the fetal heart rate should be monitored for about 30 minutes and, provided this is normal and the woman is clinically well and without abdominal pain or vaginal bleeding, she can be discharged.
If the woman is Rh negative, blood should be drawn for Kleihauer testing and an appropriate dose of Rh immune globulin given.
Although there are some studies that support the use of uterine tocolysis to aid ECV the results are mixed and most will not use routine tocolysis. However, if attempted version without tocolysis fails, and the failure is thought to be due to good uterine tone, it is rational to consider tocolysis. The best tocolytics for this purpose are the β-adrenergic drugs and this is discussed in the section on acute tocolysis above.
On rare occasions, if the need to achieve ECV is felt to be compelling, epidural or spinal anaesthesia may increase the chance of success. Obviously the risks of the regional anaesthetic have to be balanced against the ultimate goal. In addition, with the removal of maternal pain sensation the obstetrician has to be very careful to avoid using excessive force in these circumstances.
A reasonable compromise in cases of failed ECV for breech presentation is to assess the cervix on the morning of the booked elective caesarean section. If the cervix is favourable, epidural analgesia is established and ECV performed. If the fetus is successfully converted to a cephalic presentation, oxytocin and amniotomy induction of labour can be undertaken and caesarean delivery potentially avoided.
There are a number of factors which increase or diminish the chance of successful version:
Parity is the most important and success rates are higher in multiparous women, probably due to diminished abdominal and uterine muscle tone.
Gestational age: the closer to term, the lower the success rate. This is particularly evident when version is attempted at 40 weeks’ gestation or later, as the relative fetus to amniotic fluid volume ratio works against successful version.
Anterior placenta may reduce the chances of success but this is not a major factor.
Obesity reduces the chances of success.
Frank breech presentation with the legs splinting the fetal body is less amenable to version compared with complete or footling breech.
Feto-maternal bleed which may initiate or worsen isoimmunization.
Umbilical cord entanglement, which may cause abnormal fetal heart rate patterns with variable deceleration, prolonged deceleration or fetal bradycardia which, if sustained, may lead to fetal asphyxia. If this occurs one may have to turn the fetus back to the previous malpresentation in an attempt to alleviate the cord entanglement. On rare occasions this is the complication that leads to immediate caesarean delivery.
With the appropriate safeguards as outlined, there should be no increased perinatal morbidity or mortality and a less than 1% need for emergency caesarean section.
Internal Podalic Version
Internal podalic version involves the obstetrician placing one hand inside the uterus to turn the fetus from transverse or oblique lie to breech presentation. Probably the only justifiable indication for internal podalic version and breech extraction in modern obstetrics, when caesarean section is an available alternative, is in delivery of the second twin. The risk of uterine rupture and fetal trauma with internal version associated with singleton pregnancy and neglected transverse lie or shoulder presentation in advanced labour is no longer acceptable.
A sine qua non for the performance of internal version is adequate analgesia and uterine relaxation. Epidural or spinal anaesthesia will provide good analgesia but no uterine relaxation. Thus, this procedure should either be carried out under general anaesthesia with halogenated agents for uterine relaxation or under regional anaesthesia with uterine tocolysis, usually in the form of intravenous nitroglycerine (see section on ‘ Acute tocolysis ’ above). The importance of good uterine relaxation for the safe performance of internal podalic version and breech extraction cannot be overemphasized. The ease of this procedure and the lack of trauma to the fetus and the uterus depend entirely upon adequate uterine relaxation.
‘And then let him put his hands gently into the mouth of the womb, having first made it gentle and slippery with much oil; and when his hand is in let him find out the form and situation of the child … and so turn him that his feet may come forwards … and when he hath them both out, let him join them both together, and so by little and little let him draw the whole body from the womb.’
The Works of Ambroise Paré (1549). Translated by T. H. Johnson. London: Clark, 1678
Recognition of Fetal Landmarks
Practice is required to become familiar with palpation of fetal anatomical landmarks to delineate the correct position and lie of the fetus. This familiarity can be gained by practising palpation of the newly delivered infant with one’s eyes closed. The main landmarks and potential pitfalls are as follows:
Foot and hand: The toes of the foot are roughly equal in length and the separation between the big toe and the others is less distinct. The fingers of the hand are not equal and the separation between the thumb and the remaining fingers is obvious. The heel of the foot is more prominent than the heel of the hand. It is possible to tell whether the hand is right or left by ‘shaking hands with the fetus’ ( Fig 28-4 ). To a degree the same can be said for ‘shaking hands with the foot’.
The shoulder can be recognized by the confluence of the humerus, scapula and clavicle. If in doubt the adjacent ribs can often be palpated.
Knee and elbow: The flexed elbow has a prominent olecranon process, whereas the flexed knee has a dimple between the patella and tibia ( Fig 28-5 ).
Mouth and anus: Usually these can be easily distinguished but in an oedematous face presentation the mouth may be mistaken for the anus. The anus is smaller with greater tone and the adjacent ridges of the spines of the sacrum can usually be felt. In the case of the mouth the softer lips and adjacent gums and tongue should help make the distinction.
With a long glove and the hand and forearm well lubricated the operator introduces the hand gently into the vagina and uterus. The external hand rests on the abdomen. Provided adequate uterine relaxation has been achieved the membranes should not be tense and one should be able to establish the lie and position of the fetus, with particular reference to the anatomical landmarks noted above.
The anterior fetal foot should be identified, grasped and steady but gentle downward traction applied. The external hand helps by manipulating the fetal trunk and head to the vertical position ( Fig 28-6 ). If it is possible to grasp both feet this is most desirable ( Fig 28-7 ).
At some point during this downward traction the membranes will rupture. However, by this time the fetus has been converted to a longitudinal lie and the breech will already be entering the pelvis. If it has only been possible to grasp the posterior leg, then during the downward traction there should be 180° rotation, such that the posterior leg becomes anterior. This prevents the anterior buttock from becoming impacted above the symphysis ( Fig 28-8 ). The remainder of the delivery is accomplished by breech extraction as outlined in Chapter 16 .
At times the membranes rupture during the manipulations and the fetal hand and arm will prolapse. It is difficult and often traumatic to try and replace the hand and arm. Thus, one should just proceed and grasp the foot, or feet, and pull them down through the introitus. As the fetal body turns the arm will automatically follow and can be delivered with the shoulders.
Following delivery of the infant the uterus, cervix and upper vagina should be carefully explored for any lacerations.
Uterine and vaginal tamponade
When oxytocic agents fail to control postpartum haemorrhage, examination under anaesthesia is warranted. These cases may end in laparotomy and, if conservative surgical methods fail, hysterectomy may be necessary. A survey in the UK found that hysterectomy was the most common surgical procedure in women who did not respond to combinations of uterotonics, and that methods to reduce the need for hysterectomy were urgently needed.
Uterine tamponade is a less invasive procedure which is simple, does not require major surgery, can be done within minutes, and will often immediately reduce or stop the bleeding. It can be tried as soon as uterotonics are found to be ineffective. If it stops the bleeding this will be apparent within minutes and the need for definitive surgery either averted or confirmed promptly. Thus, it may avoid the need for laparotomy and hysterectomy as well as reducing the need for blood transfusion with its inherent risks. It is ideal for postpartum haemorrhage due to non-traumatic causes and for those without any retained tissue in utero. It is important that conservative surgical procedures such as uterine tamponade or compression sutures are performed before coagulopathy sets in. It may be useful to consider clot stabilization with antifibrinolytic agents such as tranexamic acid 1−2 mg IV.
A principle of first aid to stop bleeding is to apply pressure to the bleeding site sufficient to compress the blood vessels. This can be end-on or side-on compression and must be greater than the pressure of blood flow in that vessel. Once the applied pressure stops the bleeding, the blood can clot and form a permanent seal. Blood flows into the uterus with a mean arterial pressure of about 90 mmHg, although the spiral arteriolar arrangement in the uterus probably lowers the arterial pressure as the blood flows through the uterine muscle. After placental separation the venous sinuses and spiral arterioles are exposed, which results in bleeding from the placental bed if the uterus does not contract and retract efficiently enough to compress these vessels.
If, despite the appropriate use of oxytocic drugs, uterine atony continues, bimanual compression is undertaken after excluding any obvious lower genital tract trauma. If this does not stop the bleeding the uterus should be explored under anaesthesia to exclude retained products within the uterus and any uterine or lower genital tract trauma. If the bleeding is due to uterine atony a ‘tamponade test’ is useful to decide whether uterine tamponade itself would be therapeutic or whether laparotomy is needed to arrest the bleeding.
In the past, uterine tamponade could only be achieved by packing the uterus with cotton gauze. While this can be life-saving, there are a number of disadvantages; general, spinal or epidural anaesthesia is needed, the packing is done blindly, and it is hard to be sure that the entire uterine cavity is tightly packed. Incomplete and ineffectual packing can occur due to the fear of perforation ( Fig 28-9 ). Packing the uterus in this manner requires several metres of 10 cm gauze tightly packed into the uterine cavity manually and with ring forceps. The vagina is also firmly packed and a Foley catheter placed in the bladder. The gauze packing is removed in 12–24 hours.
Whether packing has been effective is not known for several minutes as the blood has to first soak through the pack before revealing itself at the cervix. To overcome some of these difficulties a sterile plastic bag may be introduced into the uterus first and the bag then packed with gauze. This facilitates more complete packing of the uterine cavity and makes for easier removal of the gauze. However, gauze packing can be cumbersome to place and remove, takes time, and is not always effective. The use of balloon tamponade may overcome some of these drawbacks.
There are several reports in the literature describing success using hydrostatic balloon tamponade either alone or in combination with additional surgical methods. Different types of balloons have been used, including the Sengstaken–Blakemore tube, Rüsch urological balloon, and the Bakri balloon ( Fig 28-10 ) filled with sterile water or saline at room or higher temperature. When uterotonics and uterine massage do not stop the bleeding, causes of local trauma or retained tissue in the uterus should be excluded. Balloon insertion may be achieved by performing a vaginal examination, identifying the cervix and manually passing the balloon into the uterine cavity without need for complex anaesthesia. Alternatively, under direct vision a vaginal speculum is passed and the anterior lip of the cervix is secured with sponge forceps. The balloon is held with another sponge forceps and inserted into the uterine cavity. The balloon is filled with about 200–500 ml of warm saline or sterile water. Care must be taken not to overfill the balloon, which may cause it to pass out through the cervix and to be expelled. Once the balloon is seen to bulge at the cervix it is adequately filled.
If the Sengstaken–Blakemore tube, Rüsch urological balloon or Bakri balloon are not available a simplified version can be fashioned as follows. Tie the cuff of a surgical glove or condom to a straight plastic urinary catheter with any suture material. Place the glove or condom and attached catheter into the uterus and fill via the catheter with a large syringe or intravenous fluid bag using the same principles outlined above. For uterine bleeding following second trimester termination a simple Foley catheter balloon may provide adequate volume for intrauterine tamponade.
In the case reports available in the literature the volume used has been arbitrary (200–500 ml) and the practice has been to fill the balloon until part of it is visible via the cervix. At this stage, if there is no bleeding through the cervix and none through the drainage channel of the catheter the test of tamponade is pronounced successful and no further fluid is added. When the intra-balloon pressure exceeds the arterial pressure of the patient no additional fluid need be added and the bleeding should stop. If bleeding continues the test is a failure and further surgical treatment is needed, in the form of laparotomy, provided one is sure the bleeding is not from the lower genital tract.
Whether tamponade is going to be successful or not is known within minutes. Once found to be successful the uterine fundus is palpated abdominally and a mark is made with a pen to provide a reference line from which any enlargement or distension of the uterus would be judged during the period of observation.
Care after Uterine Tamponade
The woman should be kept fasting and under close surveillance after insertion of the balloon. Her pulse, blood pressure, uterine fundal height and signs of any vaginal bleeding or bleeding via the lumen of the catheter should be monitored. If there is doubt whether blood is oozing above the balloon, ultrasound may help to confirm or deny this. Her temperature should be recorded every 2 hours and urinary output measured hourly via an indwelling Foley catheter. The woman should receive broad-spectrum antibiotics from the time of insertion for up to 3 days. A low dose infusion of oxytocin, 40 units in 1 L of normal saline, is continued to keep the uterus contracted over the balloon. After 6 hours, if the uterine fundus remains at the same level and there is no active bleeding around the balloon via the cervix or via the central lumen of the catheter, it is safe to remove the catheter provided the woman is stable and adequate blood replacement has been given if required.
First the balloon is deflated but is not removed for 30 minutes. Provided there is no bleeding the oxytocin infusion is stopped for another 30 minutes and if there is still no bleeding the catheter is removed. These precautions are taken in case the woman starts bleeding when the balloon is deflated or the oxytocin is stopped, in which case the balloon can be re-inflated. In our experience there has not been an instance when the balloon needed to be refilled. Six hours is usually sufficient for the placental bed to clot and stop bleeding. The reported success rate varies from 80 to 95%.
Balloon tamponade and its associated ‘tamponade test’ can be used in cases of primary and secondary postpartum haemorrhage, and for cases of bleeding after second trimester miscarriage. Both gauze packing and balloon tamponade may also be considered after caesarean section in cases without discrete bleeding points or retained tissue: for example, to provide tamponade of the poorly contracting lower uterine segment in placenta praevia.
About 80–90% of cases of primary postpartum haemorrhage are due to uterine atony. In those cases unresponsive to the usual oxytocic drugs, laparotomy and surgical measures, such as uterine compression sutures, major vessel ligation and hysterectomy may be required. Another alternative is interventional radiological embolization of the internal iliac arteries. These procedures are covered elsewhere in this chapter but all require sophisticated facilities. If these facilities are not available, balloon tamponade and uterine packing can be used as an alternative or as a stop-gap measure until the woman can be transferred to an appropriate hospital or the equipment and personnel brought to her.
There are occasions when vaginal tamponade is necessary; for example with vaginal lacerations that continue to ooze despite suturing, or paravaginal haematomas with no discrete bleeding points that can be oversewn. The options are:
Vaginal packing with gauze. This may be facilitated (both insertion and removal) by packing into a plastic bag or lubricating the gauze pack itself. The vagina that has recently accommodated the passage of a term infant can be very capacious and require a lot of gauze.
The Bakri, or similar balloon, can be used.
For the very capacious vagina a blood pressure cuff can be placed in the vagina inside a surgical glove or sterile plastic bag and inflated.
Uterine compression sutures
Uterine compression sutures of an improvised type have been used for decades: for example, figure-of-eight sutures in the lower uterine segment in cases of placenta praevia. In recent years more specific techniques for the application of compression sutures have been developed. In most cases these haemostatic sutures are used at the time of caesarean section, although they are occasionally used when all other methods of haemostasis have failed following vaginal delivery, and laparotomy is undertaken with a view to definitive arrest of haemorrhage by hysterectomy. In such cases major vessel ligation and/or uterine compression sutures may be used as a last-ditch attempt before resorting to hysterectomy. As with uterine tamponade, major vessel ligation and other rarely performed procedures, it is wise for each labour unit to have the equipment and instruments readily available in an identifiable pack so that it can be made rapidly available when needed. Diagrams of the various techniques of compression sutures can be added to the obstetric haemorrhage pack.
Strong suture material is required for compression sutures, such as No. 1 polyglactin 910 (Vicryl), polyglycolic acid (Dexon) or poliglecaprone (Monocryl). If available, No. 2 chromic catgut is also suitable. For most compression sutures a curved needle of at least 70–80 mm and sometimes larger is required; straight needles should be 8–10 cm long. In many of the standard packaged suture materials the needles are not of adequate dimension. It may therefore be advisable to have larger-eyed needles available in the haemorrhage equipment pack.
With all of the techniques for uterine compression sutures it is important to assess the efficacy of the technique. To this end the patient should be placed in the Lloyd Davies (frog-legged) position so that an assistant can remove any clots from the vagina. With both hands providing compression of the uterus it can be seen whether or not this will stop the bleeding. If it does, the compression suture is applied and, upon its completion, a careful appraisal will confirm whether the bleeding has been controlled – the ‘test of tamponade’.
The first standardized technique of uterine compression sutures was described and named after himself by B-Lynch in 1997. This type of suture is performed following low transverse caesarean section and is usually done for uterine atony unresponsive to oxytocic agents. With the uterus out of the abdominal incision and using a large (≥70 mm) round bodied, preferably blunt needle the first suture is placed from outside in to the uterine cavity approximately 3 cm below the lateral margin of the lower transverse caesarean incision and guided through the uterine cavity and out 3 cm above the caesarean incision. The suture material is then looped over the fundus of the uterus down to the posterior wall of the uterus opposite the caesarean incision. The suture is carried through the posterior wall into the uterine cavity and out on the other side roughly opposite the lateral margins of the caesarean incision. This suture is then looped over the posterior wall of the uterus down the anterior wall and placed through the uterus 3 cm above the other lateral margin of the caesarean incision and out 3 cm below ( Fig 28-11a ). Each of the suture insertion points is placed about 4 cm from the lateral border of the uterus. The two ends of the suture are then progressively tightened, with an assistant applying continuous anteroposterior compression to the uterus with both hands. The loops of the sutures over the fundus of the uterus are placed approximately 4 cm from each lateral border of the uterus. It is very important that there be progressive compression and tightening of the suture which may take 1–2 minutes to be completely effective. Once it is achieved the two ends of the suture are tied across the midline below the transverse caesarean incision ( Fig 28-11b ). At this point the assistant carefully checks the blood loss from the vagina to ensure that the suture has arrested the bleeding. If so, the low transverse incision is closed in the routine fashion, followed by closure of the abdomen. The placement of the suture is further illustrated in Figure 28-11c .
Modified B-Lynch Suture
A simplified adaptation of the B-Lynch suture has been proposed by Bhal et al. This retains the same principles but uses two separate sutures, one for each side. Figure 28-12a shows the entry and exit points of both sutures. As in the B-Lynch technique, progressive manual compression and tying down of the suture is vital. As shown in Figure 28-12b each of the sutures is tied across in the midline. The advantages of this technique are that it is a little easier to remember and by using one suture on each side the standard length (70 cm) of a polyglactin 910 suture is adequate on each side. With the B-Lynch technique one may have to tie two suture lengths together to achieve the entire suture placement.
Vertical Uterine Suture
This is an even simpler modification proposed by Hayman et al. Using a straight needle and ensuring that the bladder is well reflected the individual vertical sutures are passed from anterior to posterior approximately 3 cm below the site of a transverse caesarean incision. Depending on the width of the uterus, anything from 2 to 5 of these sutures may be placed. The usual manual compression is carried out and the sutures are tied at the fundus ( Fig 28-13 ). The advantages of this technique are its simplicity, the fact that it can be done with or without a caesarean incision and the variable number of sutures which can be placed depending upon the width of the uterine fundus.
Square Compression Sutures
In this technique the anterior and posterior walls of the uterus are compressed together using a straight needle passing from front to back through the entire uterine wall and moving laterally approximately 3 cm coming back through the uterine wall from back to front. From this exit point the needle is placed 3 cm below from front to back and then laterally 3 cm to a final pass of the needle from back to front to complete the square ( Fig 28-14 ). The anterior and posterior walls are then compressed together and the suture snugly tied down. This technique can be repeated at a number of sites, if necessary, to cover the entire uterine cavity.