A 31-yr-old female presented at 38 weeks gestation for induction of labour because of intrauterine growth retardation. Dinoprostone pessaries were administered and the contractions became regular and painful later that day. An epidural block for pain relief was requested when the cervix was 4 cm dilated.
Intravenous (i.v.) access was established and a blood pressure of 120/70 mmHg recorded. A multi-orifice epidural catheter (Portex, Hythe, Kent, UK) was sited easily, using loss of resistance to saline. After negative catheter aspiration, a test dose of bupivacaine 0.25% 3 mL was administered. Approximately 5 min later the pain was unchanged, being rated as 8/10. The patient had no sensory block to cold stimulation and had no subjective leg weakness. The blood pressure remained unaltered. Another bupivacaine 0.25% 7 mL was administered to effect analgesia before starting an epidural infusion. Shortly after this second dose, the patient's blood pressure dropped to 90/60 mmHg, despite the left lateral position and continued crystalloid administration. This was associated with a fetal bradycardia of 90 beats min−1, lasting 3 min. Further i.v. fluids were administered together with ephedrine 6 mg. Oxygen, 6 L min−1 through a facemask was commenced. The blood pressure and fetal heart rate quickly returned to normal. At this point, 15 min after administration of the epidural local anaesthetic, contractions were still painful, though lessening in intensity. The pain was rated as 5-6/10. There was reduced sensation to ice-cold stimulation from T8 to L2. There was no motor block. Blood pressure was 110/70 mmHg.
Despite maternal normotension, the continued administration of oxygen (6 L min−1) and the maintenance of the left lateral position, a further 5-min period of fetal cardiac deceleration of 80 beats min−1 occurred. On vaginal re-examination, the cervix was 7 cm dilated. Fetal blood sampling was temporarily unavailable, and in the absence of such monitoring and with a second prolonged (and incompletely resolved) bradycardia, the obstetric decision was to proceed to Caesarean section.
Epidural anaesthesia was to be established with a mixture of lidocaine 2% 20 mL plus epinephrine 1 mL (1/10 000) and sodium bicarbonate 8.4% 1 mL. A total of 5 mL of this mixture was administered with the patient in the left lateral position. After a few minutes, another 5 mL was given and the patient was turned into the right lateral position. Shortly after this second dose, the patient complained of left-sided deafness. She remained otherwise well and did not complain of tinnitus, peri-oral tingling, or abnormal motor or sensory functions. Within minutes she became tachypnocic and could only whisper that she could not breathe. She became cyanosed, but could communicate at all times, and she specifically denied chest pain. She was normotensive, but tachycardic. Although her legs felt weak, the patient had normal upper limb function and power. A differential diagnosis at this time included developing a total spinal block, subdural catheter placement, amniotic fluid embolism or allergic response to the local anaesthetic mixture administered.
The patient was transferred to the operating room where standard monitoring of continuous electrocardiography, non-invasive blood pressure measurement and oxygen saturation were established. Continuous fetal heart rate monitoring was continued. A fetal heart rate of 90-100 beats min−1 was recorded. Oxygen saturation was 40% despite breathing 100% oxygen. Auscultation of the chest revealed little air entry throughout all lung fields. There were no added sounds. The primary concern was to improve maternal oxygenation. General anaesthesia was induced with thiopental and succinylcholine was used to facilitate tracheal intubation with a size 7.5 tube. Manual ventilation of the lungs was easy, with good air entry throughout both lungs. Airway pressure was <20 cm H2O. Oxygenation improved quickly and an SPO2 of 98% while breathing 50% oxygen was achieved within 5 min. Anaesthesia was maintained with 1% isoflurane in a mixture of oxygen and nitrous oxide.
During this time, the patient was re-examined and her cervix found to be fully dilated. A live female infant in good condition was delivered by Ventouse extraction. After the delivery, spontaneous respiration returned. The patient was turned into the left lateral position and allowed to regain consciousness. She extubated herself approximately 20 min after induction of anaesthesia and had a normal respiratory pattern and rate of 12 breaths min−1. A bilateral sensory block to level T4, with full lower limb motor block but normal upper limb power, was found on examination. There was no evidence of Horner's syndrome, and the patient's hearing was subjectively normal. Forty-five minutes had elapsed since the main epidural dose had been administered.
An electrocardiogram and chest radiograph were normal. Full blood count, clotting profile, urea and electrolyte concentrations, and liver function tests were within normal limits. No fetal cells were seen in the maternal serum. The epidural catheter was left in situ and 3 mL radio-opaque dye (Omnipaque 240®; Nycomed SA, Paris, France) was later injected under fluoroscopic control.
The patient developed respiratory insufficiency as a complication after a dose of local anaesthetic had been administered through what was assumed to be a catheter sited in the epidural space. One possibility was that the catheter entered the subarachnoid space, particularly given the rapidity of onset of symptoms. However, the lack of immediate motor blockade, relative sparing of the sympathetic system, negative aspiration to cerebrospinal fluid and limitation of the height of the block following a large dose of local anaesthetic mitigate against the diagnosis of massive spinal block. There is some evidence to suggest that at least part of the catheter was in the epidural space. Forty-five minutes after the administration of a large volume of local anaesthetic, a motor block of the lower limbs with bilateral sensory block to T4, tested with cold sensation, was present. The sensory block coupled with the duration of the motor block of approximately 2 h suggested an epidural blockade.
Although there is a possibility of a high sensory block after giving the epidural local anaesthetic, it is unusual to experience transient severe respiratory distress resulting in desaturation together with unilateral deafness. Consequently, the hypothesis of normal epidural placement requires an alternative explanation for the respiratory compromise experienced by the patient. 'Classical' causes of shortness of breath such as pulmonary or amniotic fluid embolism do not correlate with the rapid resolution of symptoms and the lack of cardiovascular sequelae. Bronchospasm is unlikely in view of the low airway inflation pressures with ventilation and the absence of clinical signs. A pneumothorax was excluded later with a normal chest radiograph.
Subdural injection has been widely reported though definitions of likely effects have varied. Collier  suggested diagnostic criteria for subdural injection based on a survey of all published data and a radiographic investigation. Slow onset over 20-35 min is characteristic when bupivacaine is used, with a high sensory block that could extend intracranially. Respiratory embarrassment can occur, most often with discoordination and weakness. Relative sparing of the sympathetic and motor systems may occur . Respiratory compromise and difficulty in phonation was a most striking feature of our case, with the sparing of upper limb motor power. Hypotension was not a sustained problem, suggesting the sparing of the sympathetic system. These features imply that at least part of the top-up dose given in our case for the Caesarean section entered the subdural space.
The patient demonstrated a rapid onset of respiratory distress shortly after injection of lidocaine 2% 10 mL. The duration of this compromise is uncertain because the patient was anaesthetized and her lungs ventilated. This rapid onset and transient nature is considerably shorter than many proven subdural blocks, but Manchanda and colleagues  described a subdural block occurring with chloroprocaine 2% 6 mL after only 7 min. This particular block lasted 45 min. There have been no reported subdural blocks with lidocaine, but it seems reasonable to suppose that the onset - with such a large volume of local anaesthetic - would be rapid.
Lubenow and colleagues  suggested an incidence of subdural block of 0.8% using retrospective clinical criteria. In a population of chronic pain patients, Mehta and Salmon  reported that in 7% of cases, radio-opaque dye entered both the epidural and subdural spaces, causing a 'composite block'. To ascertain the position of the epidural, radiographic contrast medium was injected. Signs of subdural block are a typical 'railroading' of dye; subarachnoid contrast appears as beading and epidural injection can be recognized by the extension of contrast through the intervertebral foramina . This revealed evidence of subdural communication together with epidural block (Figs 1 and 2). There was no evidence of subarachnoid contrast.
A multi-orifice epidural catheter was used. If holes of such a catheter lie in different spaces, it is possible that the distribution of injected substances depends on the injection pressure . Consequently, our initial test dose, administered slowly, entered the epidural space. However, on administering the 'top-up' before the presumed urgent Caesarean section, we may have used greater force and thus preferential injection through the distal (subdural) orifice occurred. The most likely diagnosis for the respiratory distress is that part of the top-up dose for the epidural block entered the subdural space causing a high block and intercostal muscle weakness - a 'composite' block . Radiographic investigations remain important in the diagnosis of regional blocks that are unusual in nature.
M. K. Sanders
M. Y. K. Wee
Department of Anaesthetics, Poole General Hospital; Poole, UK
1. Beck H, Brassow F, Doehn FL, Dziadzka A, Schulte AM, Esch J. Epidural catheters of the multi-orifice type: dangers and complications. Acta Anaesthesiol Scand
2. Collier CB. Accidental subdural block: four more cases and a radiographic review. Anaesth Intensive Care
3. Russell IF. Total spinal anaesthesia: the effect of spinal infusions. In: Reynolds F, (ed.) Epidural and Spinal Blockade in Obstetrics.
London, UK: Baillière Tindall, 1990: 100-111.
4. Manchanda VN, Murad SH, Shilyansky G, Mehringer M. Unusual clinical course of accidental subdural local anaesthetic injection Anesth Analg
5. Lubenow T, Keh-Wog E, Kristof K, Ivankovich O, Ivankovich AD. Inadvertent subdural injection: a complication of an epidural block. Anesth Analg
6. Mehta M, Salmon N. Extradural block. Confirmation of the injection site by X-ray monitoring. Anaesthesia
7. Power I, Thorburn J. Differential flow from multihole epidural catheters Anaesthesia