Epidural analgesia is considered at present to be the most effective and innocuous technique for providing pain relief during labour and delivery. The quality of analgesia is far superior to that achieved by either parenteral or inhalation methods while the mother remains alert, thus avoiding the risks related to the airway and allowing her participation in the process of childbirth. The versatility of the continuous epidural anaesthesia using a catheter fulfils the variable analgesic requirements that occur in the dynamic process of labour, spontaneous delivery, instrumental delivery or Caesarean section (although for the latter spinal anaesthesia is being increasingly used). It also allows epidural analgesia to continue after operation. It is generally accepted that epidural analgesia, by decreasing or even abolishing physiological and biochemical changes caused by pain, benefits maternal and fetal status as well as labour progress in case of dynamic dystocias. Nevertheless, the possible increase in instrumental or surgical deliveries caused by motor block still causes controversy among anaesthesiologists and obstetricians and has provoked considerable pharmacological modifications in the technique over the last years. Epidural analgesia in obstetrics continues to be the origin of research, clinical interest and practical changes.
Development of epidural technique in obstetrics
The concept of selective or segmentary epidural analgesia, prevalent nowadays, originated in the works of Cleland, Bromage, and Doughty several decades ago [1-5]. It is based on administering low doses of local anaesthetics (LA) through an epidural catheter to achieve analgesia, mainly thoracolumbar, while preserving sensory painless perception and muscular and abdominal pelvic tone.
Despite the fact that the use of epidural analgesia for obstetrics has increased in the last decades as well as the general acceptance of its advantages, data reported by surveys in different European countries show that its use is far from being generalized. Rates of use range from the most optimistic, 37.2% in France , which is comparable to that of USA and Canada, 24% in the UK , to less than 10% in Germany . Although these surveys were published recently, the data were mostly collected in 1996 and increased use of epidural techniques may be expected at present. Differences in the use of epidural blocks for obstetrics are mainly due to structural and organizational factors. Obstetric services with more than 1500 deliveries per year offer a greater use of epidural blocks in the context of a provision of a 24-h anaesthetic service exclusively assigned to obstetric anaesthesia. However, maternity units with less than 500 deliveries per year can only offer a 24-h anaesthetic service shared with the other clinical specialities and so register a lower usage of epidural block. Differences also exist among teaching, non-teaching and private hospitals .
Indications and contraindications
Pain relief is a per se indication for epidural analgesia and this is recognized by professional obstetrical organizations . The risk and mortality associated with general anaesthesia in the parturient indicate the use of regional techniques whenever anaesthesia is required [10,11]. On the other hand, the recognized beneficial effects of epidural block on maternal, fetal and uterine dynamics as well as on uterine blood flow have made this technique particularly useful for mothers with associated diseases which might be decompensated by the sympathetic stimulation provoked by pain and Valsalva efforts. The maternal, obstetrical and fetal indications and contraindications for epidural block in obstetrics are summarized in Table 1. Many are common for other regional techniques such as single shot or continuous subarachnoid blockade, or even caudal, although epidural block remains the most advantageous technique in obstetrics [12,13]. The lack of adequate staff and facilities together with the fact that it is impossible for an expert anaesthesiologist to be present in the obstetric area almost immediately are formal contraindications for epidural analgesia.
Nowadays management techniques for epidural block used in obstetrics are based on the use of low doses of local anaesthetics (LA) combined with opioids; these have extended their indications to severely ill mothers [14,15]. Clinical experience supports the use of epidural block in mothers with neurological and cardiac diseases if systemic or local effects of the technique do not worsen the disease [14-17]. It can be used in Parkinson's disease and essential epilepsy as well as in spina bifida and established spinal cord injury . Multiple sclerosis remains a controversial indication . Most patients with a cardiopathy will benefit from the haemodynamic effect of epidural blockade during labour. Nevertheless, those who are highly dependent on cardiac load conditions, such as severe aortic stenosis or hypertrophic cardiomyopathy, present relative contraindications and need a meticulous management of the block and extensive cardiovascular monitoring [16,17]. Multidisciplinary careful evaluation and explanation to patients are essential in the management of epidural blockade in such difficult cases.
The substantial benefits to both mother and fetus of the provision of epidural analgesia for labour in pre-eclampsia - due to the suppression of sympathetic stimulation by pain and the increase in utero placental blood flow - are now recognized. Attention to the adequacy of an adequate intravascular volume and the absence of coagulopathies are essential in pre-eclamptic patients in order to minimize risk [18-20].
Advantages of peridural analgesia for labour
Pain and anxiety during labour are associated with adrenergic hyperactivity that increases heart rate, systemic vascular resistances and cardiac output. Hyperventilation and hypocapnia during uterine contractions cause a left shift of the oxyhaemoglobin dissociation curve so decreasing oxygen extraction by the fetus. Compensatory hypoventilation between contractions decreases maternal PaO2 below 9.1 kPa (70 mmHg) and can affect fetal well-being. Maternal and fetal tolerance to such effects can be low in the presence of disease. Uteroplacental blood flow is increased due to decreased catecholamine concentration and diminished systemic vascular resistance. Maternal and fetal oxygenation improves during epidural analgesia and prevents metabolic acidosis . A decrease in systemic vascular resistances mitigates the repercussion of the autotransfusion caused by uterine contraction and decreases arterial pressure in case of inadequate intravascular volume. Epidural blockage above T4 may be associated with bradycardia .
Doses of LA used in contemporary epidural techniques in obstetrics have not had deleterious effects on the fetus and newborn. The concentration of LA in maternal plasma is low and correspondingly so in the fetus. These concentrations do not adversely affect either myocardial function or heart rate of the fetus, nor the umbilical blood flow, Apgar test and neurobehavioral tests [13,21,22]. Neonatal outcome is better after epidural analgesia than parenteral analgesia with opioids . Although subtle fetal neurological alterations caused by opioids administered epidurally cannot be ruled out yet, the respiratory function of the neonate is unaffected by the doses used in contemporary epidural techniques. Umbilical artery blood flow is unaffected by the low concentrations of epinephrine reached. The most important effect of epidural block on the fetus is caused by hypotension due to the adrenergic block especially if there is associated aortocaval compression [12,21].
The interference of epidural analgesia on labour and delivery depends on the intensity of the motor block and on the obstetrical factors. It has caused much controversy among obstetricians and anaesthesiologists. Critical analysis of the literature [22-24] allows the following statements to be made about the effects of epidural analgesia on labour:
(a) The antispasmodic effect on the cervix increases the efficacy of labour;
(b) The first stage is prolonged for 1-2 h in spontaneous labour but the administration of oxytocin normalizes the duration of first stage;
(c) In the presence of dynamic dystocia, epidural analgesia co-ordinates uterine contractions and shortens the first stage.
(d) The second stage is prolonged in nulliparae and multiparae by 50% and this effect is more probable if the perineal analgesia and motor block have been produced in the first stage of the labour. This is due to retarded fetal descent and rotation, especially with posterior presentations, and a decrease in the effort of maternal pushing. The normal increase in release of oxytocin (Ferguson reflex) is abolished by epidural blockade of the sacral roots. A prolonged second stage under epidural block is not associated with fetal acidosis;
(e) The increased incidence of forceps instrumentation found in many studies can be reduced by delaying the injection of the doses for perineal analgesia until fetal head has descended and rotated, and by obstetricians allowing a longer second stage
(f) The increased incidence of Caesarean section cannot be attributed to epidural block.
Pain during labour
Labour is associated with severe pain in 60% of nulliparae and 50% of multiparae: great variability exists between individuals. During first stage of labour, from the beginning of regular contractions until complete dilation of the cervix, pain is caused by traction of the cervix and ischaemia of uterine muscle: it is visceral, diffuse and conducted to dermatomes T10 to L1. In the second stage, from complete dilation of the cervix until birth, the pain is somatic and mainly conducted by fibres Aδ to dermatomes S2-S3. The most painful period is the final phase of the first stage . Pain does not have any beneficial effect on labour, uterine contractions with or without pain are responsible for the progress of labour.
Technical peculiarities of lumbar epidural block in obstetrics
Epidural block is technically more difficult to perform in the pregnant patient due to lumbar lordosis, difficulty in obtaining good flexion of the legs, soft tissue oedema, changes in ligament texture as well as lack of co-operation in some patients. Engorgement of the epidural veins increases during uterine contraction and with pushing efforts. The risk of vascular puncture with the needle or the catheter in parturients is 10-12% compared to 1% in the non-pregnant population . The volume of the epidural space is reduced by 40% and together with lumbar lordosis favours cephalad spread of the injected solution and so makes caudal diffusion more difficult. Blood flow in the epidural veins, which lack valves, is greatly increased, and the accidental injection of LA produces rapidly high concentrations in the brain and heart [25,26].
Pressure in the lumbar epidural space is slightly positive at the beginning of labour. It reaches 0.95 kPa (10 cmH2O) during Valsalva manoeuvres, and increases even more in the dorsal or sitting position during delivery. Cerebrospinal fluid pressure increases during uterine contraction, frequently reaching more than 5.7 kPa (60 cmH2O) . Therefore, the risk of accidental dural puncture is high. The position of the patient when the block is established (lateral vs. sitting) does not affect the final level of the block. The lateral position is usually better tolerated by the mother especially if she is in pain; it is associated with less positive epidural pressure and less engorgement of the epidural veins. The sitting position facilitates the technique in cases of technical difficulty caused by obesity or spinal abnormality [25-27].
For all these reasons, epidural puncture in obstetrics must be achieved:
(a) At L2-L3 or L3-L4 level, where both the epidural space is widest (4-5 mm) and the interspace largest;
(b) By a medial approach since the epidural veins are mainly located laterally;
(c) The needle should only be advanced between contractions;
(d) The epidural space should best be detected by a loss of resistance technique to saline or air (not the hanging-drop technique);
(e) A catheter without a stylet should be used in an effort to reduce the risk of dural or vascular perforation;
(f) Injections should be made through the catheter after checking that it is not intravascular;
(g) Aliquots of small volumes of LA only may be injected, slowly, between contractions;
(h) Time is required to obtain a good block of the sacral roots because they are large and caudal diffusion is impaired [27-29].
In Table 2 different solutions used for epidural obstetric analgesia are presented. In clinical practice, LA combined with opioids is the most commonly used epidural solution. The advantages of adding epinephrine are controversial and other drugs such as clonidine are far from being the standard in obstetric anaesthesia practice.
The choice of LA is based on its analgesic efficacy and duration , the magnitude of its transfer across the placenta, its potential for fetal toxicity and the degree of motor block acceptable.
Fetal plasma concentrations of LA results from the total doses administered, the magnitude of vascular absorption (increased in pregnancy), protein binding in mother and fetus, fetal tissue binding, and maternal and fetal metabolism and excretion.
Non-ionized fractions of LA cross the placenta in both directions. The higher the protein binding in the mother the less the transfer across the placenta . Bupivacaine, with the highest protein binding of all LAs, presents the least transfer across the placenta. The fetal/maternal rate (quotient between fetal plasma level and maternal plasma level) is considered an index of the transfer across the placenta but fetal plasma concentrations are also influenced by drug fetal tissue binding. The fetal/maternal rate is 0.2-0.3 for bupivacaine and 0.7 for lidocaine. Since both LAs show a similar pattern of fetal tissue fixation (a little higher for bupivacaine), bupivacaine presents a better profile for obstetric use and remains the most popular LA for this indication . Fetal acidosis greatly increases the accumulation of LA in the fetus and causes redistribution of blood flow to brain and heart which will thus be exposed to more drug passage .
Bupivacaine. Its advantages in obstetrics are:
(a) High protein binding (95%);
(b) Long duration (more than 90 min with a concentration of 0.25% due to its elevated pKa);
(c) Absence of tachyphylaxis, and relative affinity for sensitive fibres that provides a good sensitive block with moderate muscular block with low concentrations (see Table 2 for doses and concentrations used in obstetrics).
Cardiotoxicity is the main disadvantage of bupivacaine; it is exacerbated during pregnancy attributed to decreased plasma levels of α1-acid glycoprotein that transports bupivacaine and to an increase of the free fraction . Cardiotoxicity precedes neurotoxicity in the parturient. Reported fatal toxic reactions have been produced by accidental intravascular injection. Total doses should be limited to 150 mg injected at any one time or 300 mg for 10 h. Concentrations higher than 0.5% should not be used in obstetrics due to potential cardiotoxicity [30,31]. Levobupivacaine, a promising analogue, shows same advantages as bupivacaine with less cardiotoxicity and neurotoxicity.
Ropivacaine. Is chemically similar to bupivacaine but with clinical differences especially regarding toxicity. It produces local vasoconstriction that compensates for its lower protein binding (65%) so the duration of the block is only slightly shorter than that of bupivacaine and the motor block is significantly less. The cardiotoxicity of ropivacaine is intermediate between lidocaine and bupivacaine [33,34]. In clinical practice differences in hypotension and motor block between epidural bupivacaine and ropivacaine are minor although a lower potency has been found with ropivacaine [33,34]. The vasoconstrictor effect of ropivacaine makes the addition of epinephrine unnecessary (see Table 2 for dosage).
Lidocaine. Compared to bupivacaine and ropivacaine, lidocaine has a faster onset (5 min for analgesia and 15 min for motor block), a shorter duration (60-80 min) of effect, more motor block, greater transfer across the placenta and higher fetal accumulation in the presence of fetal acidosis . A maximal total dosage of 400 mg is recommended . Tachyphylaxis appears after four or five doses. Despite the fact that lidocaine is not the first choice LA, it is still useful when there is only a short time available to establish the block (e.g. short labour in multiparae, urgent Caesarean section), a need for a deep sensitive motor block (instrumental delivery), and to improve a patched block due to its better diffusion in the epidural space. The addition of epinephrine increases the duration of block by 25% (see Table 2 for dosage).
2-Chloroprocaine. This drug is widely used in the USA but little in Europe. It is an ester, metabolized by plasma cholinesterase with a low maternal and fetal toxicity. It produces a very good block of short onset (8 min) and short duration (35-50 min) that is increased 40% by the addition of epinephrine. The motor block is intense. A concentration of 2% is used for analgesia in labour and 3% for Caesarean section and forceps delivery. 2-Chloroprocaine antagonizes the analgesic effects of both opioids and bupivacaine .
Intraspinal opioids administered alone alleviate pain during the first stage of labour but, except for meperidine, do not show sufficient efficacy by the epidural route . All the opioids have been shown to increase the analgesic quality of the epidural block but the highly lipophilic fentanyl and sufentanil are the most commonly used opioids in obstetrics. During labour, the main mechanism of action of epidural opioids is a spinal one although systemic absorption from the epidural space produces plasma detectable concentrations of drug . A large experience is now available with the combination of these opioids with both bupivacaine and lidocaine, but less is known about their combination with ropivacaine. The results of the addition of opioids to epidural LA in the parturient are: a shorter onset and better quality of analgesia, less incidence of incomplete block, and better block of the sacral roots during delivery - so allowing a decrease of LA dose (therefore less motor block) and favouring pushing efforts by the mother [13,24,26,27,31,35,38].
Local anaesthetics given in sufficient dose produce very satisfactory clinical effects that are not improved by adding opioids. Therefore, the advantages of combining opioids with LA for obstetric epidural blockade are clinically only significant when the LA doses are lower than required when used alone, thus allowing less potential for toxicity and motor blockade. With the recommended doses (Table 2) maternal respiratory depression is very improbable although it has been described with 100 μg of fentanyl and 50 μg of sufentanil in Caesarean section . Accidental subarachnoid injection of a solution containing opioids can produce respiratory depression. Nausea, and especially pruritus, are very frequent complications after epidural opioids during delivery. On the other hand, shivering is decreased when opioids are combined with LA [26,28].
Epinephrine. Epinephrine 1:200 000 (5 μg mL−1) or 1 : 400 000 shortens the latency period, improves the quality and duration of the block of lidocaine and low concentrations of bupivacaine, but not of ropivacaine. It will also increase motor blockade [26,36]. Uterine vasoconstriction is a risk of doses over 100 μg, therefore epinephrine is not used for continuous infusion techniques although it is useful for anaesthetic epidural block, for Caesarean section with bupivacaine and lidocaine [36,39].
Alkalinization of the LA solution. This is especially useful to decrease the latency of bupivacaine, ropivacaine and lidocaine by about 33%. One mL of sodium bicarbonate 8.4% per each 10 mL of the solution is recommended .
Clonidine. Clonidine added to LA alone or with fentanyl and epinephrine improves quality and duration of the epidural block in obstetrics and allows the decrease LA doses. Effective doses are not determined and seem to range within 120-150 μg. Its use is not recommended in obstetrics due to frequent dose-dependent secondary effects that include maternal hypotension, bradycardia and sedation, labour prolongation and hypotony of the baby [40,41].
Performance of epidural block in obstetrics
Appropriate moment to perform epidural block and preparation of the patient
At present, it is generally accepted that epidural analgesia with low concentrations of LA and minor motor block does not affect the progress of labour. The accepted criterion for performing the epidural block is pain associated in established labour with a 3-cm dilatation of the cervix [25,26,28]. High-risk patients or those with complicated labours will benefit from epidural block provided early: it will avoid stress in the mother and allow obstetrical intervention at any time. The agreement of the obstetrician is always required before performing an epidural block.
Although many parturients are first seen by the anaesthesiologist in the obstetric area, ideally, preanaesthetic assessment must be made and informed consent obtained in a preanaesthetic consultation during the last month of pregnancy . Decompensation, or appearance of some maternal diseases, frequently occurs during the last weeks or days of pregnancy or during labour; most coagulopathies are acquired during labour. Therefore, clinical assessment must be made immediately before the epidural block is established. It is important to know if systemic opioids have been given to the mother during the 6 previous hours in order to decrease or even avoid epidural administration of opioids. The intravascular volume status should be adequate. Crystalloids such as 0.9% saline or Ringer's solution, 15 mL kg−1 for analgesic block and 25 mL kg−1 for anaesthetic block, should be administered before the test dose. Glucose-containing solutions should be avoided. Fetal cardiac rate and uterine dynamics should be monitored.
A test dose has a double aim: to detect any intravascular position of the catheter (or the needle), and to recognize accidental subarachnoid injection. Due to the severity of these complications, the test dose must be very sensible and specific [43,44]. Epinephrine is the most commonly used drug to detect intravascular injection in epidural analgesia. Pregnant patients without uterine contractions receiving epidural blocks for elective Caesarean section respond with an increase of 30% in heart rate lasting more than 30 s to the accidental injection of 15 μg of epinephrine (3 mL 1:200 000) as do other surgical patients. However, this response is not reliable for women in active labour who present both variable heart rates and baseline tachycardia caused by pain and stress. Other symptoms of intravascular epinephrine are jitters, pale complexion, headache, palpitations, chest pain and restlessness. Therefore, 15 μg of epinephrine as a test dose lacks sufficient specificity and sensibility during labour and higher doses would affect uterine blood flow. Isoprenaline 5 μg or Doppler detection of 1-2 mL of air injected into the epidural space has been suggested as a 'test dose' but are not used in practice [26,45]. Moreover, testing intravascular injections using closed tip multiport catheters is more difficult.
The detection of accidental subarachnoid injection requires a test dose that produces a sensitive motor block of limited extension. Low concentrations of LA lack sensitivity, isobaric solutions can spread too high with the risk of hypotension, and motor block with bupivacaine can take as long as 10 min to appear. Nevertheless, 3 mL of 1.5-2% hyperbaric of lidocaine or 2-chloroprocaine injected into the subarachnoid space will produce a deep motor block of lower limbs within 3 min. Hyperbaric bupivacaine 0.5% is the other choice.
Despite its limitations, a test dose containing epinephrine 1:200 000 is widely used. It should be injected immediately after an uterine contraction with close monitoring of the patient. The risk of undetected intravascular injection of the test dose and the potential for subarachnoid catheter migration at any time throughout labour obliges the anaesthesiologist to: consider every epidural injection as a test dose; aspirate gently before, during and after the injection; giving the total doses in aliquots of 5 mL between uterine contractions with an interval between injections of 2 min; monitor closely after each dose; and consider anaesthetic epidural block with high doses of LA for a Caesarean section in active labour as a high risk situation for this complication .
Induction of epidural block
If the test dose has been injected with the patient in the lateral position, a block of the dependent side will appear earlier and be more dense after the induction dose. Injecting the induction dose through the catheter with the patient on the opposite side can compensate for this. Satisfactory epidural analgesic blockade for the first stage of labour requires a minimum T10 level sensitive block that is accomplished with a volume of 10-12 mL (including test dose) of either bupivacaine 0.25%, ropivacaine 2%, lidocaine 1% or 2-chloroprocaine 2% (Table 2). The induction dose must be fractionated into aliquots of 5 mL. Re-injection of 20-25% of the initial dose of LA 10-15 min after completion of the initial dose increases the density and quality of the block without elevating the level of sensory blockade . It is essential that the quality and level of the block obtained with the initial dose is satisfactory. The patient should be able to move her legs. The sacral roots will not be blocked but successive intermittent injections or continuous infusion of LA will block them given sufficient time.
If epidural blockade is performed in the second stage of labour, higher doses of LA will be needed to block the sacral roots and provide perineal analgesia. Carbonated solutions, addition of epinephrine, and the use of lidocaine or 2-chloroprocaine will help to obtain a more potent and earlier block, but significant motor blockade will be apparent [25,29]. Caesarean section requires dense anaesthetic block at a minimum level of T4 even when transverse suprapubic incisions are made in which traction on the uterus is intense (see Table 2). Lower levels are associated with pain, nausea, retching, vomiting and dissatisfaction for mother and obstetrician. Opioids will improve the quality of the block but do not allow for reducing the dose of LA since motor block is required. When an urgent or emergent Caesarean section is necessary without an adequate epidural block already established, the indication for epidural anaesthesia depends on the time available to obtain the surgical block. If it is less than 15 min, subarachnoid block or combined spinal epidural techniques are more effective and safer. If the patient already has satisfactory epidural analgesia the doses of LA required to obtain surgical block are similar to those required for a Caesarean section without an epidural block for labour [25,29,46].
Maintenance of epidural blockade
Epidural blockade should be continuously maintained until completion of delivery and repair of episiotomy. There is no rationale to allow the block to dissipate at the end of labour in favour of the patient pushing to deliver, although it has had defenders. Three techniques can be used to maintain epidural block: intermittent injections or 'top-up doses', continuous infusion and patient-controlled epidural analgesia (PCEA) with or without a background continuous infusion. The aim of all these methods is not to allow the block to regress more than two dermatomes, and to prevent pain. The need for continuous monitoring of the patient and adaptation of the block to changing conditions of labour is common with all techniques.
Intermittent injections. These still remain the most favourite practice. 'Top-up' doses must be 75-100% of the initial dose according to the initial level of block obtained. The main disadvantage of the technique compared with continuous infusion is that pain can return between doses but if re-injections are given 'by the clock' according to the duration of the administered solution (usually every 1.5-2 h) that disadvantage is overcome (Table 2). Close monitoring of cardiovascular responses and possible neurological signs of toxicity should follow every dose since accidental subarachnoid injection with this technique has more severe consequences than with continuous infusions [13,47]. Frequently, the need for re-injection is manifested before pain reappears by some discomfort during contractions.
Continuous infusion. Is started only after a satisfactory analgesic block has been obtained with the objective to maintain it, not to elevate or intensify it. Diluted LA solutions are used (Table 2). There are no advantages of using lidocaine or 2-chloroprocaine. Bupivacaine and ropivacaine in the recommended doses give a good sensory block with undetectable motor block at the beginning; however, progressive motor block develops within a period of time. Perineal analgesia is frequently insufficient in the second stage of delivery, especially if the infusion time has been less than 2 h and instrumentation has to be used; a bolus of more concentrated LA is necessary in such cases [48,49]. When unexpected changes in the block appear, malfunction of the infusion pump should immediately be suspected. Gutsche  has given the following recommendations to solve possible problems with continuous infusions:
(a) If the block begins to dissipate check that the catheter is not intravascular, then re-inject the initial dose and increase infusion rate
(b) If the sensory block begins to rise, check that the catheter has not migrated into the subarachnoid space, then decrease infusion rate by 25-50% but do not stop the infusion
(c) A breakthrough of pain despite an adequate sensory level of the block is typical of posterior and transverse presentation. Re-inject the initial dose and increase the infusion rate. For Caesarean section the infusion must be discontinued and an appropriate dose of LA given, although the infusion may be recommenced to provide analgesia in the postoperative period.
Compared to intermittent injections, the advantages of continuous infusion are better haemodynamic stability, stable analgesia, less motor block and lower potential for toxicity. Subarachnoid migration or intravascular injection will not cause serious problems if they are recognized before a bolus of LA is injected. Disadvantages are frequently insufficient perineal analgesia (reaching 50% of cases), a larger total dose of LA and potential for malfunction of the infusion pump .
Patient controlled epidural analgesia (PCEA). Less clinical experience is available with this technique compared to intermittent injections and continuous infusion. After a satisfactory analgesic block has been obtained with the initial dose described above, it is maintained either by autoadministered demand injections alone or by a background infusion plus demand injections signalled by the patient as soon as there is a recurrence of minimal or undesirable discomfort. The latter is preferred since it avoids periods of pain and allows the mother to rest more. The main advantage of PCEA is that mothers assume an active role in the control of their analgesic treatment and they can decide the level of sensation they want to feel, and that in turn gives a greater degree of satisfaction; a decrease in the total dose of LA has not been uniformly demonstrated [50,51]. Disadvantages are the need for technically safe PCEA pumps, the possibility of pump malfunction that can be very dangerous if the lock-out mechanism fails, insufficient time for the mother to learn pump management and need for well trained personnel in the technique. In deciding the background infusion rate and the demand doses, it is important to know the mother's desires for the amount of sensation she desires to experience [25,50,51]. In Table 2 the two different regimens used in the author's department are described. Perineal analgesia is usually sufficient; for Caesarean section, appropriate doses to obtain an anaesthetic block must be administered as described above. Whatever the maintenance technique used, an additional bolus of LA solution without opioids should be injected if the patient complains of pain.
Combined spinal epidural block (CSE) and the 'walking epidural'
This is the newest technique introduced in obstetric analgesia. It was described by Brownridge in 1981 for Caesarean section using two different interspinous spaces for the spinal and epidural punctures, respectively. The method was improved by Carrie in 1984 who described the 'needle through needle' technique that led to the 'spinal epidural sequential blockade' for Caesarean section and labour. The background for the present use of CSEs is the low incidence of postdural headache with the spinal needles used nowadays, the development of sophisticated equipment for the CSE technique and the subarachnoid injection of opioids as analgesia for the first stage of labour [52-54].
With CSE immediate analgesia or anaesthesia is obtained by subarachnoid administration of LA, opioids or their combinations, and continued by the epidural route. If CSE is performed in the first stage of labour, 10-25 μg of fentanyl or 5-10 μg of sufentanil provides an average of 2 h of sufficient sensitive block without motor blockade . When the pain reappears, epidural analgesia is provided by one of the techniques described above (Table 2). If CSE is performed for Caesarean section, LA in sufficient doses is injected into the subarachnoid space to provide an anaesthetic block, and just in case of insufficient analgesia or for postoperative analgesia, epidural injections will follow. In this last instance, accidental subarachnoid position of the catheter will be difficult to check since motor block will have been already established.
The spectacular technique of a 'walking epidural' described by Barbara Morgan from Queen Charlotte's Hospital, London, is based on a CSE technique with subarachnoid opioids and low doses of bupivacaine for the first stage of labour that allows the patient to walk around . Although women accept very well, and the use of this technique and the demand for it are increasing, some facts limit its extensive use. Fetal and uterine activity should be monitored during a 'walking epidural', that means earlier monitoring and the need for telemetry monitoring. Even 2.5 mg of subarachnoid bupivacaine can give some degree of motor blockade or loss of proprioceptive sensitivity in feet or legs making ambulating unsafe in some patients. Subarachnoid administration of opioids has frequent side-effects as pruritus and nausea, and others less frequent such as vomiting, hypotension, urinary retention, fetal bradycardia, uterine hyperstimulation and maternal respiratory depression. Postdural headache is always possible even using very thin spinal needles. An alternative to CSE for obtaining a 'walking epidural' without the disadvantages of CSE is the epidural administration of low concentrations of bupivacaine or ropivacaine.
Protocols for 'walking epidurals' should be strictly followed. The mother, the labour and the fetus must all be monitored. Ambulation (supported by a companion and an intravenous (i.v.) infusion stand on wheels) is only allowed for periods of 15 min, after 30 min of CSE so long as a motor block has been excluded (the patient should be able to lift her legs while in the supine position, to stand up and to perform a deep knee bend) and arterial pressure is stable when in the upright position . One satisfactory and less complicated option is to allow the mother to sit in a chair.
Complications of epidural analgesia in obstetrics
All possible complications of epidural analgesia  can appear in obstetric patients but only those with a high incidence in this field will be described here. Acceptance of even minor complications by patients, their relatives and the obstetrical team is low compared to the use of epidural blocks in surgical or chronic pain patients. Minor complications such as back pain or postdural headache accounted for 38% of the claims in obstetrical anaesthesia in the USA .
The complications that are more frequent and their causes are summarized in Table 3 together with the recommended procedures to treat them. Puncture of a blood vessel, 'bloody tap' and unintentional dural puncture are more frequent in the obstetric patient due to the changes described in the anatomy of the epidural space and the presence of technical difficulties. The incidence of dural puncture is considered as a quality index of an anaesthesia service in obstetrics and must remain below 2%; it increases with operator inexperience and during night work . Dural taps are usually made by the needle because catheters without stylets are judged safe. The usual practice following the detection of a dural tap is to attempt the block through the interspinous space immediately above; catheter holes will remain far from the dural tap but small incremental doses of LA must be administered slowly to avoid any increase in epidural pressure that may force part of the dose to enter the subarachnoid space. Sometimes, a test dose will produce a significant amount of pain relief, and in the case of Caesarean section the total amount of LA should be reduced by 25%. Even if cerebrospinal fluid does not exit through the needle, dural tear and subsequent subarachnoid catheterization at any time of the block is possible. Therefore the test dose and close monitoring after each epidural injection are crucial. If 10-12 mL of the usual epidural solutions are injected through such a subarachnoid catheter, a high or total subarachnoid block (depending on dose) will appear in 1-5 min, with severe hypotension. Treatment should be immediately started consisting in turning the mother into the left lateral position, administration of 100% oxygen, crystalloids (glucose-free i.v. solutions), i.v. ephedrine; tracheal intubation and assisted ventilation may be necessary. Repercussions on the fetus will depend on the response to maternal resuscitation but prompt delivery or Caesarean section should be carried out.
The interspace where the dural puncture occurred should be documented in the anaesthetic record to allow an epidural blood patch to be appropriately placed, if it becomes necessary . In the parturient, a dural puncture with an 18-gauge or larger needle without prophylaxis results in a 70% incidence of headache, which is often severe and incapacitating; patients must be informed of this risk. This high frequency has been attributed to a greater loss of cerebrospinal fluid pressure caused by pushing efforts, collapse of epidural veins following delivery and rapid loss of fluids after delivery from blood loss, lactation and diuresis . In cases of accidental dural puncture the patient should be informed of the possibility of a postdural headache and its lack of longterm consequences. Vigorous oral or i.v. hydration, bed rest and analgesia alleviate the symptoms in most patients. The epidural infusion of crystalloids or dextrans  delay rather than avoid a postdural headache. Caffeine reduces symptoms in almost 90% of cases either by the i.v. or oral routes . An autologous blood patch with 15-20 mL of blood is only recommended after 3 days of severe headache (Table 3).
Inadequate block occurs in 10% of epidural blocks in labour; opioids decrease the incidence (Table 3). Inadequate efficacy of the first dose must induce the repetition of the block [60,61]. Excessive motor block is frequent and can be considered a complication in obstetrics . Subdural block is rare but needs to be suspected to avoid too high of a block [62,63].
Hypotension defined as systolic blood pressure below 100 mmHg or a decrease of 30% of baseline values is the complication with greatest potential for deleterious effect on the fetus. Its incidence and severity depends on block extension, the circulating blood volume and any associated aortocaval compression. With adequate hydration arterial pressure decreases only moderately if the level of block is T10 due to sympathetic stimulation during labour; a block to the level of T4, required for Caesarean section, is frequently associated with hypotension especially in elective procedures without active labour. Prophylaxis is based on prehydration with crystalloids (10-15 mL kg−1 for labour and 20-25 mL kg−1 for Caesarean section) [25,31]. Treatment should start immediately and consists of 500-1000 mL of 0.9% saline, oxygen and ephedrine as a 5-mg i.v. bolus. Ephedrine produces direct and indirect stimulation of α- and β-adrenergic receptors and it has been shown not to affect uterine and placental blood flow; when β-adrenergic effects are not desirable, phenylephrine 100 μg is the best alternative. The effects of direct α-adrenergic drugs are decreased in pregnant woman but they produce marked uterine vessel vasoconstriction. Ephedrine, like all adrenergic stimulants, increases fetal heart rate decreasing its spontaneous variability. Prophylactic i.v. or intramuscular (i.m.) administration of ephedrine has been recommended in epidural blockade for Caesarean section; the most predictable effects of ephedrine are obtained with an i.v. continuous infusion [25,31].
The frequency of LA toxicity is 0.03-0.5% . It is due to high plasma concentrations caused by accidental intravascular injection, or overdose or accumulation of repeated injections. Neurological symptoms usually precede cardiovascular symptoms, and are progressive and related to plasma levels of LA: lightheadedness, confusion, tinnitus, circumoral numbness, metallic taste, shivering, visual disturbances, agitation and loss of consciousness. Seizures can appear at any moment causing great compromise of both maternal and fetal oxygenation.
Hypertension and tachycardia can accompany seizures but rapidly evolve to severe hypotension, ventricular dysrhythmias and cardiac arrest. Bupivacaine is the LA with greatest cardiotoxicity during pregnancy. The severity of the reaction will dictate treatment intensity that will include 100% oxygen, i.v. thiopental, tracheal intubation using neuromuscular blockade, left shift of the uterus, i.v. fluids and vasopressors . Cardiopulmonary resuscitation is more difficult in pregnant women especially in cases of bupivacaine toxicity. Toxic levels of LA in the baby will produce similar symptoms to those described in the mother.
Recommended doses of liposoluble opioids fentanyl and sufentanil (Table 2) are safe provided systemic opioids have not been administered in the previous few hours. Nevertheless, subarachnoid administration of sufentanil for CSE has produced instances of severe respiratory depression . Nausea and vomiting are more frequent in epidural blockade above T10, especially if there is concomitant hypotension, and during Caesarean section. Association of opioids with LA increases the incidence of these complications.
Claude de Bernard-Horner syndrome, a rare and benign complication of lumbar epidural analgesia has been mainly reported in obstetrics with an incidence of 1.33-4%; it is believed to occur more frequently than reported but is often undetected [66-68]. Shivering of short duration and variable intensity appears in 30-50% of epidural blocks in obstetrics, usually after the first epidural dose. Different therapy of low efficacy has been suggested [25,26](Table 3). Peripheral neuropathies, postpartum back pain and postpartum urinary retention are all very frequent after delivery and although they are attributed to the epidural block their aetiology is related to obstetrical and postural factors [26,28,68].
Some controversies in obstetric analgesia
Informed consent for epidural block. Written informed consent obtained in the delivery room, especially when the patient is in pain is inappropriate so it can only fulfil formal requirements if given antenatally. That carries organizational problems since not all pregnant women receive a preanaesthesia consultation [26,42] and it is not always predictable who will receive an epidural block.
Who can give top-up doses? One-third of inadvertent subarachnoid injections occurs not with the first epidural dose, but with subsequent doses due to catheter migration. On the other hand, the quality of epidural analgesia requires adaptation to individual needs. In some countries, e.g. Switzerland or the UK, 'top-up' injections are usually administered by midwives but in others such as Germany or Spain 'top-ups' are only administered by anaesthesiologists. Remarkable differences in training midwives about anaesthesia are responsible for these different practices in Europe . If midwives are allowed to monitor the block and administer 'top-up' doses, they must be trained in the specific risks and complications of epidural analgesia by anaesthesiologists; the LA doses in each injection must be sufficient to maintain adequate analgesia but insufficient to produce a total spinal block in case of inadvertent subarachnoid injection. Bupivacaine 0.125% with fentanyl 20 μg in 10 mL 0.9% saline will produce a high subarachnoid block but not a total block if given into the subarachnoid space . Continuous epidural infusions need a close control and well trained midwives can be very helpful in providing good quality of analgesia. This a very controversial field and difficult to standardize in all countries, but one thinks that women would benefit more from well-trained staff who closely supervise the effects of the epidural analgesia and can adjust it to her needs rather than from an anaesthesiologist who is not present in the obstetric unit .
Platelet count. The minimum platelet count for performing a safe epidural block is a difficult, controversial and confusing field. In many texts, the limit is empirically established in 100 000 mm−3[26,61] although lower platelet counts have been advocated as safe in parturients. This field will be impossible to investigate prospectively and it seems more judicious to consider the cause of a thrombocytopenia more than a specific figure, and especially important is to ascertain if a thrombopathy is also associated, as happens in pre-eclampsia, Von Willebrand disease or diabetes . Most coagulopathies in pregnant women appear at the end of the pregnancy or during labour. Pre-eclamptic patients showing acceptable platelet counts before an epidural block can develop a severe coagulopathy or even a HELLP (haemolysis, elevated liver enzymes and low platelet count) syndrome afterwards with a risk of epidural bleeding . With patients at risk of coagulopathy, the decision to establish an epidural block should be based on aetiological and clinical evaluation together with coagulation tests including bleeding time, performed as close as possible to the moment of epidural block performance [20,70].
1 Cleland JG. Paravertebral anesthesia in obstetrics, experimental and clinical basis. Surg Gynecol Obstet
2 Hingson RA, Edwards WB. Continuous caudal analgesia; an analysis of the first 10 000 confinements thus managed with the report of the authors' first 1000 cases. JAMA
3 Doughty A. Selective epidural analgesia and the forceps rate. Br J Anaesth
4 Bonica JJ. Principles and Practice of Obstetric Analgesia and Anesthesia.
Philadelphia: Davies, 1969.
5 Cousins MJ, Veering BT. Epidural neural blockade. In: Cousins MJ, Bridenbaugh PO, eds. Neural Blockade in Clinical Anaesthesia and Management of Pain.
Philadelphia: Lippincott-Raven, 1998: 243-322.
6 Palot M, Chale JJ, Colladon B et al.
Anesthésies et analgesies pratiqueés dans las maternites françaises. Ann Fr Anesth Reanim
7 Burnstein R, Buckland R, Pickett JA. A survey of epidural analgesia for labor in the United Kingdom. Anaesthesia
8 Stamer UM, Messerschmidt A, Wulf H, Hoeft A. Practice of epidural analgesia for labour pain: a German survey. Eur J Anaesthesiol
9 ACOG Committee on Obstetrics. Maternal fetal medicine, ACOG Committee Opinion. No 118. Int J Gynecol Obstet
10 Crowhurst JA, Plaat F. Why mothers die - Report on Confidential Enquiries into Maternal Deaths in the United Kingdom 1994-96. Anaesthesia
11 Glassenberg R. General anesthesia and maternal mortality. Semin Perinatol
12 Cheek TG, Gutsche BB, Gaiser RR. The pain of childbirth, its effect on the mother and fetus. In: Chesnut DH, ed. Obstetric Anesthesia: Principles and Practice.
St. Louis: Mosby, 1994: 314-329.
13 Naulty JS. Epidural analgesia for labor. In: Seebacher J, ed. Anesthesie-Réanimation En Obstetrique.
Paris: Arnette, 1993: 115-147.
14 Eldridge AJ, Kipling M, Smith JW. Anaesthetic management of a woman who became paraplegic at 22 weeks' gestation after a spontaneous spinal cord haemorrhage secondary to a presumed arteriovenous malformation. Br J Anaesth
15 Aveline C. Peut-on réaliser des anesthésies périmedúllaires chez des patients souffrant d'une sclérosis on plaques? Ann Fr Anesth Réanim
16 Paiz B, Cyna A, Belperio P, Simmons S. Epidural analgesia for labour and delivery in a parturient with congenital hypertrophic obstructive cardiomyopathy. Anaesth Intensive Care
17 Brauneis S, Apponi F, Commisso C, Pinto G, Fedele F. Epidural anesthesia for cesarean section in patients with hypertrophic cardiomyopathy: a report of three cases. Anesthesiology
18 Kanayma N, Bewlayet HM, Khatum S et al.
A new treatment of severe pre-eclampsia by long-term epidural anaesthesia. J Hum Hypertens
19 Hogg B, Hauth JC, Caritis SN et al.
Safety of labor epidural anesthesia for women with severe hypertensive disease. Am J Obstet Gynecol
20 Blasi A, Gomar C, Fernández C, Nalda MA. Indicación de la anestesia espinal para la cesárea en pacientes con coagulopatía por síndrome de HELLP. Rev Esp Anestesiol Reanim
21 Thorburn J. Assessment of the effects of epidural anesthesia. In: Van Zundert A, Ostheimer GW, eds. Pain Relief and Anesthesia in Obstetrics.
New York: Churchill Livingstone, 1996: 288-295.
22 Halpern S, Leighton B, Ohlsson A, Barrett J, Rice A. Effect of epidural vs parenteral opioid analgesia on the progress of labor. JAMA
23 Chesnut D. Epidural analgesia: impact on progress and outcome of labor. ASA Refresher Course Lectures.
Park Ridge, IL: American Society of Anesthesiologists, 1999: 221.
24 Olofsson Ch, Ekblom A, Ekman-Ordeberg G, Irestedt L. Obstetric outcome following epidural analgesia with bupivacaine 0.125% with sufentanil: a prospective randomized study in 1000 parturients. Acta Anesthesiol Scand
25 Gutsche BR. Lumbar epidural analgesia in obstetrics: taps and patches. In: Reynolds F, ed. Epidural and Spinal Blockade in Obstetrics.
London: Baillière Tindall, 1990: 95-106.
26 Dailland P, Cahusis P, Landru J, Belkacem H. L'anesthésie péridurale pour le travail. Cah Anesth
27 Ostheimer GW, Leavitt KA. Epidural anesthesia: technique. In: Van Zundert A, Ostheimer GW, eds. Pain Relief and Anesthesia in Obstetrics.
New York: Churchill Livingstone, 1996: 288-295.
28 Gutsche B. How to make an epidural safe and effective. ASA Refresher Course Lectures.
Park Ridge, IL: American Society of Anesthesiologists, 1999: 142.
29 Hess LW, Morrison JC. Antepartum management and analgesia for normal labor and vaginal delivery. In: Bonica JJ, McDonald JS, eds. Principles and Practice of Obstetric Analgesia and Anesthesia,
2nd edn. Baltimore: Williams & Wilkins, 1995: 751-782.
30 Wulf H, Munstedt P, Maier C. Plasma protein binding of bupivacaine in pregnant women at term. Acta Anesthesiol Scand
31 Shnider SM, Levinson G, Ralston D. Regional anesthesia for labor and deliver. In: Shnider SM, Levinson G, eds. Anesthesia for Obstetrics,
3rd edn. Baltimore: Williams & Wilkins, 1995: 135-153.
32 Johnson RF, Herman NL, Johnson V, Arney TL, Paschall RL, Downing JW. Effects of fetal pH on local anesthetic transfer across the human placenta. Anesthesiology
33 Polley LS, Columb MO, Naughton NN, Wagner DS, van de Ven CJ. Relative analgesic potencies of ropivacaine and bupivacaine for epidural analgesia in labor: implications for therapeutic indexes. Anesthesiology
34 McCrae AF, Westerling P, McClure JH. Pharmacokinetic and clinical study of ropivacaine and bupivacaine in women receiving extradural analgesia in labour. Br J Anaesth
35 Haberer JP, Diemunsch P. Anesthésie obstétricale. In: Bach JF, Imbert JC, Jasmin C, eds. Encyclopédie Médico-chirurgicale. Anesthésie-Reanimation.
Paris: Editions Techniques, 1992.
36 Maestre ML, Miranda A, Capogna G, Celleno D. Medicación espinal analgo-anestésica. In: Miranda A, ed. Tratado de Anestesiología Y - Reanimación En Obstetricia Principios Fundamentales Y Bases de Aplicación Practica.
Barcelona: Masson, 1997: 185-228.
37 D'Angelo R, Gerancher JC, Eisebach JC, Raphael BL. Epidural fentanyl produces labor analgesia by spinal mechanism. Anesthesiology
38 Porter JS, Bonello E, Reynolds F. The effect of epidural opioids on maternal oxygenation during labour and delivery. Anaesthesia
39 Alahuhta S, Räsänen J, Jouppila R, Jouppila P, Hollmén AI. Effects of extradural bupivacaine with adrenaline for caesarean section on uteroplacental and fetal circulation. Br J Anaesth
40 Claes B, Soetens M, Van Zundert A, Datta S. Clonidine added to bupivacaine-epinephrine-sufentanil improves epidural analgesia during childbirth. Reg Anesth Pain Med
41 Cigarini I, Kaba A, Bonnet F et al.
Epidural clonidine combined with bupivacaine for analgesia in labor. Regional Anesthesia
42 Affleck PJ, Waisel DB, Cusick JM, Van Decar T. Informed consent in labor. Anesth Analg
43 Norris C, Ferrenbach D, Dalman H et al.
Does epinephrine improve the diagnostic accuracy of aspiration during labor epidural analgesia? Anesth Analg
44 Mulroy M, Norris M, Liu S. Safety steps for epidural injection of local anesthetics: review of the literature and recommendations. Anesth Analg
45 Leighton BL, De Simone CA, Norris MC, Chayen B. Isoproterenol is an effective marker of intravenous injection in laboring women. Anesthesiology
46 Levison G. Avoiding obstetric anesthesia lawsuits. Relevant Guidelines and Standards, ASA Refresher Course Lectures
Park Ridge, IL: American Society of Anesthesiologists, 1999: 532.
47 Seebacher J, Darbois Y. Comment choisir entre les différentes techniques et méthodes d'analgésie peridurale. In: Seebacher J, ed. Anesthesie-Réanimation En Obstetrique.
Paris: Arnette, 1993: 99-113.
48 Elliot R. Continuous infusion epidural analgesia for obstetrics: bupivacaine versus bupivacaine-fentanyl mixture. Can J Anaesth
49 Chesnut DH, Laszewski LJ, Pollack K, Bates J, Manago NK, Choi W. Continuous epidural infusion of 0.0625% bupivacaine-0.0002% fentanyl during the second stage of labor. Anesthesiology
50 Gambling DR, Huber CJ, Berkowitz J et al.
Patient-controlled epidural analgesia in labour: varying bolus dose and lockout interval. Can J Anaesth
51 Paech MJ. Patient-controlled epidural analgesia during labor: choice of solution. Int J Obstet Anesth
52 Birnbach D. Combined spinal epidural (CSE) and other new techniques for labor analgesia. ASA Refresher Course Lectures.
Park Ridge, IL: American Society of Anesthesiologists, 1999: 223.
53 Tsen LC, Thue B, Datta S, Segal S. Is combined spinal-epidural analgesia associated with more rapid cervical dilation in nulliparous patients when compared with conventional epidural analgesia? Anesthesiology
54 Collis RE, Davies DW, Aveling W. Randomised comparison of combined spinal-epidural and standard epidural analgesia in labour. Lancet
55 Morgan B. Mobile regional analgesia in labour. Br J Obstet Gynaecol
56 Aya AG, Mangin R, Robert C, Ferrer JM, Eledjam JJ. Increased risk of unintentional dural puncture in night-time obstetric epidural anesthesia. Can J Anaesth
57 Salvador L, Carrero E, Castillo J, Villalonga A, Nalda MA. Prevention of postdural puncture headache with epidural-administered Dextran 40. Reg Anesth
58 Camann W, Murray R, Mushlin P, Lambert D. Effects of oral caffeine on postdural puncture headache. Anesth Analg
59 Crawford JS. Experience with epidural blood patch. Anaesthesia
60 Le Coq G, Ducot B, Benhamou D. Risk factors of inadequate pain relief during epidural analgesia for labour and delivery. Can J Anaesth
61 Beilin Y, Bemstein H, Zucker-Pinchoff Zahn J, Zenzen WJ. Treatment of incomplete analgesia after placement of epidural anesthetic for the woman in labor. Anesth Analg
62 Reynolds F, Speedy H. The subdural space. The third place to go astray. Anaesthesia
63 Carrero E, Agustí M, Fabregas N, Valldeoriola F, Fernandez C. Unilateral trigeminal and facial palsies associated with epidural analgesia in labor. Can J Anaesth
64 Laishley RS. Local anaesthetic toxicity. In: Reynolds F, ed. Epidural and Spinal Blockade in Obstetrics.
London: Baillière Tindall, 1990: 81-94.
65 Katsiris S, Williams S, Leighton BL, Halpern S. Respiratory arrest following intrathecal injection of sufentanil and bupivacaine in a parturient. Can J Anaesth
66 Segura P, Speeg-schatz C, Wagner JM, Lern O. Le syndrome de Claude de Bernard-Horner et son contrarie, le syndrome de Pourfour du Petit en anesthèsie-réanimation. Ann Fr Anesth Réanim
67 Zarcos A, Díaz F, Cuéllar E, Pérez F, Villegas E, Rodríguez S. Síndrome de Horner en el trabajo de parto. Rev Esp Anestesiol Reanim
68 Russell R, Reynolds F. Epidural analgesia and chronic backache. Anesth Analg
69 Cox M, Lawton G, Gowrie-Mohan S, Morgan B. Let us reexamine the safety of midwife epidural top-ups. Anaesthesia
70 Yuen TS, Kua JS, Tan IK. Spinal haematoma following epidural anaesthesia in a patient with eclampsia. Anaesthesia