Levobupivacaine-induced seizures and cardiovascular collapse treated with Intralipid®


Nigel Bedforth E-mail: nigel.bedforth@nuh.nhs.uk


Summary Lipid emulsion has been used in the successful treatment of local anaesthetic-induced cardiovascular collapse in animals and in two cases of cardiac arrest in humans. Previous reports of levobupivacaine toxicity in humans have been characterised by neurological signs and symptoms, without serious cardiovascular events. We present a case in which presumed intravenous injection of levobupivacaine led to neurological and cardiovascular consequences. This was treated successfully by resuscitation that included intravenous Intralipid® infusion.

Lipid emulsion has successfully treated local anaesthetic-induced cardiovascular collapse in animals [1, 2]. Intralipid® 20% (Fresenius Kabi, Uppsala, Sweden) has been used successfully in the resuscitation of two patients with prolonged cardiac arrest refractory to standard treatment [3, 4].

We report the administration of lipid emulsion to a 75-year-old lady who experienced seizures, cardiovascular collapse and electrocardiographic (ECG) abnormalities following a posterior (psoas compartment) lumbar plexus block with levobupivacaine. We believe this is the first report of the successful use of Intralipid in a pre-arrest situation, and the first report of cardiac arrhythmia and cardiovascular collapse following levobupivacaine administration in a peripheral nerve block.

Case report

A 75-year-old, ASA III, 85-kg woman presented for surgical repair of a fractured neck of femur following a fall at home; her fall was not precipitated by any complicating pathology. Her medical history included severe chronic obstructive pulmonary disease (FEV1 75% predicted) with breathlessness on minimal exertion, osteoporotic vertebral collapse and stable angina. A pre-operative ECG demonstrated normal conduction with frequent atrial ectopics and pathological q-waves in V1-V3.

Routine monitoring was commenced and 18-gauge peripheral venous access established. Baseline observations included an arterial pressure of 130/70 mmHg. Midazolam 1 mg and alfentanil 0.5 mg were administered and the patient was positioned in the sitting position with support provided by the operating department assistant. Following several unsuccessful attempts to place a subarachnoid block, a lumbar plexus block was commenced with the patient remaining in the sitting position. The lumbar plexus block was performed under aseptic conditions by an experienced consultant anaesthetist. A 22-gauge Whitacre-tip spinal needle was inserted 3 cm lateral to the spinous process of the fourth lumbar vertebra (L4); following contact with the transverse process of L4 at a depth of approximately 3 cm, the needle was inserted a further 2 cm into psoas muscle. Following negative aspiration, levobupivacaine 0.5% was injected slowly (with several further negative aspirations) to a total of 20 ml. No blood was aspirated via the needle during the performance of the block.

Within seconds the patient groaned and became unresponsive. She was positioned supine and immediately convulsed in a tonic-clonic fashion for several seconds. The patient's airway was supported and the lungs ventilated with 100% oxygen via a face mask. An immediate, presumptive diagnosis of local anaesthetic toxicity due to intravenous injection of local anaesthetic was made. Assistance was called for and Intralipid solution was requested from the theatre recovery area. Arterial pressure was measured at 60/40 mmHg and the continuous, lead II ECG trace demonstrated altered QRS morphology; QRS voltage was reduced and the complexes were broadened compared with baseline. The alteration in ECG was progressive with reducing voltage and broadening of the QRS complex. No radial pulse was palpable, although a carotid pulse was maintained. Heart rate was unchanged at 110 beats.min−1.

A second, short convulsion occurred 2 min after the first as assistance arrived. The patient appeared mottled and venous-congested and was cold to touch, despite the continued presence of a palpable carotid pulse. The QRS morphology continued to deteriorate, with complexes becoming difficult to distinguish from baseline ECG noise. An intravenous bolus of metaraminol 0.5 mg was administered and the decision taken to secure the airway. Propofol 80 mg and suxamethonium 100 mg were administered and the patient's trachea was intubated.

Immediately following tracheal intubation and within 4 min of the levobupivacaine injection, the Intralipid 20% solution arrived and 100 ml was infused over 5 min via the peripheral cannula. During the infusion, an arterial line was sited and arterial blood was sampled (see Table 1). Directly measured arterial pressure was 90/60 mmHg. During the Intralipid infusion there was a rapid normalisation of QRS morphology. Following the administration of propofol 80 mg, no further sedative/anaesthetic agents were administered over the subsequent 10 min; the patient did not convulse again or regain consciousness.

Table 1.  Arterial blood gas analyses following presumed intravenous injection of 20 ml levobupivacaine 0.5%.
 Time after block
6 min23 min93 min
Paco2; kPa11.28.75.1
Pao2; kPa70.451.410.6
HCO3; mmol.l−114.116.320.8
Base excess; mmol.l−1−13.3−10.2−4.4
Arterial lactate concentration; mmol.l−

The patient's arterial pressure, heart rate and ECG morphology remained stable and within normal limits for the 10 min following administration of the Intralipid and the decision was taken to proceed to surgery. Volatile anaesthesia was initiated and surgery continued without further incident. Repeated arterial gas analysis revealed a steady improvement (Table 1). Two hours later, surgery concluded and the patient regained consciousness. A repeat 12-lead ECG in the theatre recovery area revealed no changes from the pre-operative ECG. The patient remained stable, orientated and pain-free; she was discharged to the orthopaedic ward and recovered uneventfully.


Lumbar plexus blockade is widely used for intra-operative and postoperative analgesia for hip and knee surgery. Intravascular injection has been reported previously [5, 6], and dural/extradural spread are also recognised complications [7, 8].

In this case, there was a rapid loss of consciousness followed by two seizures immediately after injection of levobupivacaine. This strongly suggests the intravascular administration of levobupivacaine. Negative aspiration of blood prior to injection has been noted in previous reports of local anaesthetic toxicity following posterior lumbar plexus block and does not exclude intravascular needle placement [5, 6]. The differential diagnosis includes intrathecal and extradural spread of local anaesthetic leading to cardiovascular instability, although these are less credible as causes of the convulsions. Cardiac arrest has been reported following intrathecal spread of local anaesthetic during a posterior lumbar plexus block [9]; however, in this reported case, there were no seizures and sensory block was noted to ascend very rapidly to T2, with associated bilateral papillary dilation immediately prior to the arrest. In our case, no mydriasis or bilateral motor or sensory block was noted. Further evidence against the intrathecal/extradural injection of local anaesthetic is provided by the rapid recovery of cardiovascular stability following Intralipid administration.

Levobupivacaine contains the single S(–) enantiomer of bupivacaine and was introduced to improve on the cardiac safety profile of racemic bupivacaine. To our knowledge, there have been no published case reports of fatal or serious cardiac toxicity with levobupivacaine. Generalised tonic-clonic seizures have been reported in a 27-year-old patient following an interscalene brachial plexus block after presumed intravenous injection of levobupivacaine [10]. However, there were no ECG abnormalities or cardiovascular instability. There have also been case reports of central nervous system toxicity and grand mal seizures following accidental intravascular injection of levobupivacaine after posterior lumbar plexus block [6]. In neither of these cases were any cardiovascular or ECG changes suggestive of cardiovascular toxicity. In our case, the ECG demonstrated dynamic widening of the QRS complexes with progressively decreasing voltage, suggesting a serious, evolving and widespread cardiac conduction defect.

We did not measure plasma levobupivacaine concentrations, so we are unable to comment on the effect of Intralipid on serum concentrations of local anaesthetic. Previous authors have highlighted the variability in plasma concentration required to cause toxicity following the administration of other local anaesthetics [6, 11].

This is the first report of the use of lipid emulsion in a peri-arrest situation. The use of lipid emulsion has previously been recommended after standard resuscitative measures have proven ineffective as its safety is unknown when administered in the high doses used in lipid rescue [12, 13]. However, pretreatment with a lipid infusion has been shown to increase the bupivacaine dose required to induce asystole in rats [2]. There have also been no documented adverse effects from the use of Intralipid to resuscitate the two patients following local anaesthetic-induced cardiac arrest [3, 4].

This case occurred shortly after the introduction of Intralipid to our department. We have placed boxes containing the solution with administration equipment and a simple protocol in all our theatre areas. We are satisfied that the Intralipid was obtained quickly from our theatre recovery area, allowing early institution of treatment.

A similar case was reported of cardiac arrest following injection of ropivacaine for posterior lumbar plexus blockade in a 66-year-old man scheduled for elective hip arthroplasty [5]. Similar to our case, shortly after completion of the injection the patient became unresponsive and had tonic-clonic generalised seizures. He then developed asystole from which he was successfully resuscitated. Cardiac activity was restored after 5 min with the ECG displaying bradycardia with widened QRS complexes. There was a metabolic acidosis (arterial pH 7.10) with a plasma lactate concentration of 11.7 mmol.l−1. Blood samples showed high concentrations of ropivacaine. Drawing parallels to our case, it seems possible that the patient may have progressed to cardiac arrest if left untreated; she demonstrated a progressive deterioration of her ECG combined with signs of severely compromised cardiac output and seizures. All physiological derangement appeared to resolve following Intralipid treatment.

It is impossible to predict the course of events had we not administered Intralipid and carried out standard resuscitation. However, evidence from animal studies suggests that levobupivacaine-induced cardiac arrest may be less susceptible to such treatment than that induced by ropivacaine [14]. Evidence from animal studies suggests a beneficial effect of Intralipid administration in local anaesthetic toxicity, and there are sound theories for its putative mode of action [13].

Our case represents the third report of the use of Intralipid in the clinical setting of local anaesthetic toxicity and the first of its use in a near cardiac-arrest situation following levobupivacaine administration. Because of the difficulty of examining the issue of local anaesthetic toxicity and its treatment with Intralipid, we should continue to collate its reported uses, both successful and unsuccessful, so that we may reach a consensus on the efficacy and appropriateness of this novel treatment.