Polymyoclonus, ventricular fibrillation and Takotsubo after accidental spinal injection of tranexamic acid

Abstract

Several factors have been identified as contributing to medication administration errors, including look-alike, sound-alike (LASA) errors. LASA errors are important causes of serious adverse events arising from spinal injection of tranexamic acid, which can be confused with ampoules of local anaesthesia.

We present a case of accidental injection of 250 mg of tranexamic acid rather than prilocaine during spinal anaesthesia. The patient developed lower extremities myoclonus, followed by generalised convulsions and ventricular fibrillation, that was reverted within 6 min. Severe cardiogenic shock requiring both inotropic and vasopressor therapy followed, along with a classic apical ballooning pattern on echocardiography and elevated myocardial injury markers, indicating Takotsubo cardiomyopathy. The patient’s condition progressively improved to full recovery, and she was discharged from hospital after 1 month with no neurological deficit or cardiac dysfunction.

To our knowledge, this is the 28th reported case of accidental spinal injection of tranexamic acid. We present a brief review of previously published cases.

Background

Adverse drug events are an important cause of injury and death in many health centres and hospitals. The most commonly reported errors include wrong dosage and administration of the wrong drug. In the perioperative setting, particularly if spinal anaesthesia is involved, drug administration errors can have devastating consequences for patients owing to the potential for elevated dosages being administered intrathecally or adjacent to critical structures when neuraxial anaesthesia and analgesia are performed.^1–4

Antifibrinolytic drugs are used worldwide to decrease the requirement for blood transfusions, reduce the risk of reoperation for bleeding and lower mortality associated with haemorrhage following major trauma. Tranexamic acid (TXA), a synthetic derivative of lysine, is a popular antifibrinolytic that exerts its haemostatic effects by binding to plasminogen. TXA has increasingly been used in elective surgical procedures owing to its safety and effectiveness.⁵ However, several cases of errors involving TXA have been reported, with mistakes made because of lack of attention, incorrect labelling or similar appearance to other medications leading to the wrong route of injection.^{4 6}

In this case report, we report an accidental injection of TXA rather than prilocaine for spinal anaesthesia as a result of confusion between two different ampoules. The error led to ventricular fibrillation, myoclonic seizures and intensive care unit (ICU) admission of a patient undergoing varicose vein surgery.

Case presentation

A woman in her 60s, weighing 70 kg and having American Society of Anaesthesiologists I physical status, was scheduled for excision of varicose veins. She had no relevant medical history of convulsions or cardiac pathology and no previous anaesthesia exposure. Her vital parameters and investigations were within normal limits. Spinal anaesthesia was performed under light sedation and analgesia with midazolam and fentanyl, with the patient in the left lateral position, while 2.5 mL of hyperbaric prilocaine was injected intrathecally. The procedure consisted of a single puncture with crystal clear cerebrospinal fluid (CSF). Sensorial blockade was present on neurological assessment, even though this examination may be confounded by the effect of intravenous drugs given prior to the procedure. Shortly after drug administration, the patient became extremely restless and developed subtle myoclonic movements in the lower extremities. Due to the patient’s pain complaints during surgery, intravenous propofol was given by subanaesthetic infusion to provide sedation, in addition to fentanyl that was reinforced for pain control, as well as midazolan to provide anterograde amnesia. Her pulse rate and blood pressure were kept within normal limits during surgery. Postoperatively, the patient had poorly controlled pain, and myoclonic movements of her lower limbs became more exuberant. This unusual response aroused suspicion of wrong drug administration, prompting immediate review. No broken ampoule of prilocaine was recovered; instead, a broken ampoule of TXA was found. Accidental subarachnoid injection of TXA was detected roughly 1 hour afterwards. Within a few minutes, polymyoclonus movements occurred, followed by generalised convulsions that were not effectively controlled by intravenous diazepam. About 5 min after, the seizures abated, but the patient had a depressed level of consciousness and was unable to protect the airway. Consequently, she received 500 mg of thiopental and 100 mg of rocuronium for tracheal intubation and was then mechanically ventilated. She rapidly deteriorated to cardiovascular collapse and cardiac arrest with ventricular fibrillation. External cardiac massage, intravenous epinephrine (1 mg) and intravenous fluids were instituted, and spontaneous circulation resumed after 6 min. A generalised rash was then noted, and dexamethasone was also administered. Intra-arterial access was secured, and norepinephrine infusion was started. The patient was immediately transferred to another hospital where she was admitted to the ICU. The initial ECG obtained at the emergency department showed multifocal premature ventricular contractions (10–12/min) with a heart rate of 89 beats per minute and minor QT prolongation (QTc 462 ms), and the patient was given magnesium sulfate. The maximum QTc of 479 ms was observed later in the ICU (figure 1). Intravenous amoxicillin and clavulanate were administered for aspiration pneumonia, as she developed respiratory compromise and presented bilateral radiological findings of aspiration. Her initial transthoracic echocardiogram showed severe left ventricular (LV) depression and akinesia of the middle and apical segments of LV, and troponin I maximum was 31.2 ng/mL. The echocardiographic appearance, the significantly decreased ejection fraction necessitating haemodynamic support and the elevated myocardial injury markers were compatible with Takotsubo cardiomyopathy (CMP). CT of the head did not show intracranial bleeding, structural anomalies or early ischaemic changes, and the patient was maintained on an infusion of propofol and midazolam.

Figure 1.

ECG on intensive care unit admission showing sinus rhythm (86 beats per minute) and QT prolongation (QTc 479 ms).

Outcome and follow-up

The patient’s haemodynamic status improved, and she was weaned off vasopressors. The propofol dosage was reduced during 12 hours of a clinical and electroencephalographic seizure-free period, and then both propofol and midazolam were gradually tapered over 24 hours. The patient was put on a maintenance dose of levetiracetam. Her sensorium started to improve, and she was extubated after 3 days. On neurological examination, she had no sensorium alterations or cranial pair deficits; however, lower extremity motor function was impaired (motor score of 2/5 bilaterally), and deep tendon reflexes were diminished. MRI of the brain and spinal cord was performed, excluding intracanal collections and medullar signal alterations indicative of myelopathy. She was regularly assessed by Neurology and Physical Medicine and Rehabilitation and followed an individualised rehabilitation plan during hospitalisation. She was re-examined by cardiology, and a transthoracic echocardiogram was repeated and found to be normal. After 4 weeks of hospital admission, the patient was discharged with no neurological deficit, as she fully recovered strength in her lower extremities.

Discussion

The intrathecal space has become an important anatomic site for medical intervention. Accidental intrathecal injections are generally rare, with clinical effects depending on drug characteristics and the dose administered intrathecally. Consequences range from no adverse effect to profound neurological consequences and/or death.

Little is known about the effect of direct intrathecal administration of TXA in humans. Wong et al presented the first report of an inadvertent intrathecal injection of TXA. In that case, 75 mg of TXA was injected intrathecally in a young man scheduled for appendectomy. He developed clonic convulsions that progressed to a generalised seizure; he recovered without sequelae.⁷ Since that report, several similar cases have been described. The problem was highlighted by Patel et al,⁴ who reviewed accidental spinal administration of TXA from 1988 to 2018 and found 21 case reports. Outcomes ranged from short-duration seizures treated with a single antiepileptic and followed by full recovery to life-threatening neurological and/or cardiac complications, requiring resuscitation and intensive care, and death. The typical signs and symptoms that were reported included no sensory or motor block; severe pain in the back, buttocks and legs; myoclonus starting in the legs and progressing to generalised convulsions; severe tachycardia and hypertension; and ventricular arrhythmia. Death was reported in 10 patients, and in the remaining 11 patients, 10 required ICU admission for management of refractory convulsions and/or tachyarrhythmias.⁴ We have reviewed accidental intrathecal administration of TXA since 2018 and found seven additional case reports,^{6 8–12} including one paediatric fatal outcome.¹¹ Death occurred in three cases, and ICU admission was necessary in the remaining four patients. Overall, spinal administration of TXA had a mortality rate of about 50%, and there was a high incidence of permanent neurological injury in survivors.

In 27 patients, an ampoule error was the cause of the accidental use of TXA, and it mainly resulted from the similar appearance of the TXA and local anaesthetic ampoules. In the remaining patient, TXA was mistakenly administered via a spinal catheter that was in place for pain management instead of the intravenous line.¹³ This latter case was the only episode that occurred in the ward; all other accidental spinal injections happened in the operating theatre environment.

Overall, TXA has been increasingly available in operating rooms and labour wards and has been used in multiple clinical procedures, including cardiovascular, obstetric and orthopaedic surgeries due to its effects in reducing blood loss and its safety and effectiveness.⁵ This increased usage was paralleled by the increasing number of drug error reports in recent years. In fact, 13 cases, almost half of those reported, occurred after 2015.

Spinal injection of TXA seems to result in fatal outcomes more often in pregnant women than in other patients. We found nine reports of TXA administered in the context of caesarean section/labour.^{4 8 10} Seven deaths occurred after TXA spinal administration, and two women survived after ICU stabilisation.^{10 14} A possible reason for a higher rate of deaths in pregnant women may be that the volume of CSF decreases in pregnancy and the dural sac surface area is reduced, which may facilitate the spread of local anaesthetics.¹⁵ In addition, CSF density is lower at term gestation, which is an important determinant of the extent of subarachnoid block and could explain the rapid, high sensory block observed after administration of local anaesthetics for labour/caesarean section analgesia; it could possibly also account for differences in TXA distribution in CSF.¹⁶ Such changes may result in higher concentrations of TXA in CSF in pregnant women; furthermore, reduced subarachnoid space volume may potentiate toxicity. Most maternal deaths occurred within hours of receiving spinal TXA. There were no reported adverse effects in any of the newborns, even though TXA can cross the placenta.

Although little is known about the pharmacokinetics of TXA, it seems reasonable that a high drug concentration in CSF would determine the extent of spinal and cerebral neurotoxicity. All 13 patients who died had received a TXA dose of 200 mg or more. Also, those who received a high dose experienced ventricular arrythmias more often, as was the case in our patient.

The mechanism by which TXA induces convulsions or ventricular fibrillation is not completely understood. However, TXA-associated seizures have been previously described in patients undergoing cardiac and non-cardiac surgery.^{17 18} Polymyoclonus may arise not only from the cerebral cortex but also directly from spinal cord, and it has been reported after trauma or subarachnoid injections of irritating agents during myelography.^{17 19}

Lecker et al described TXA binding to glycine receptors, resulting in a decrease in the inhibitory anion current and a subsequent increase in excitability leading to seizures.¹⁷ Massive sympathetic discharge induced by TXA may then suppress inhibitory gamma-amino butyric acid-A receptors in the cerebral cortex or lower cerebral blood flow with consequent cerebral ischaemia.²⁰

Stress or Takotsubo CMP is a common entity in critical illness, resulting from severe catecholamine surges that affect the myocardium. The high beta-adrenoreceptor density in apical regions of the heart may explain the classic apical ballooning pattern consisting of akinesia of the apex and hypercontractility of the base. Identifiable stressors are present in 70% of cases, and these can be either physical or emotional. While this entity has never been reported in spinal injection of TXA, a case of Takotsubo CMP associated with intravenous exposure to TXA has described. The patient presented clinically with hypertension and tachycardia shortly after the first dose of TXA, similarly to most patients after spinal injection. She then developed refractory cardiogenic shock requiring both inotropic and vasopressor therapy.²¹ Although the time course and known pathophysiology and reports support the relationship of Takotsubo CMP to TXA administration, Takotsubo CMP may result from other stressors or the combination of several stressors, including patient anxiety before surgery.

Management of the adverse effects of accidental TXA administration remains fundamentally supportive, as there is no known TXA antagonist. There have been attempts to use multiple anticonvulsants and antiarrhythmics. Sedatives such as propofol or thiopental have commonly been used, and electroencephalographic monitoring may help to titrate infusion.⁴

However, seizures could often not be terminated even with multiple drug associations, and severe ventricular arrhythmias were often refractory. In some patients, including the case we report, seizures and arrhythmias persisted/recurred for days despite aggressive management attempts, suggesting that their resolution is dependent on the decline of the TXA concentration in CSF following spinal injection, which is probably slow.⁴

The toxic effects of spinal TXA may be reduced by diminishing its concentration in CSF, which was the rationale behind the CSF lavage that was performed in three patients. The first was performed in afemale 30 min after injection of 300 mg TXA, using a single 20 mL spinal shot of normal saline.³ The second patient was a 62-year-old man, who underwent CSF lavage less than an hour after 250 mg of TXA was administered in the subarachnoid space, using a total of 25 mL of normal saline administered via 5 mL injections.²² The most recent report of CSF lavage involved a 14-year-old boy who received 300 mg of TXA. Lavage was done with aliquots of normal saline adding up to a total volume of 120 mL.⁹ All patients were subsequently discharged with no neurological deficits after 1–2 weeks of hospital admission. In view of the terrible prognosis associated with this drug error, some authors suggest early use of lavage.⁴

To our knowledge, this is the first report of drug error resulting from confusion between prilocaine and TXA ampoules, as all other reported cases involved bupivacaine.

Look-alike, sound-alike (LASA) medicines have a high risk of confusion due to similar-looking packaging or a similar-sounding name.^{23 24} Similar-looking inner or outer packaging can lead to error and serious adverse events and has been the most reported cause of errors involving TXA.⁴ TXA and prilocaine ampoules have similar packaging and labels, and their trade names sound similar (figure 2).

Figure 2.

Ampoules of hyperbaric prilocaine (left) and tranexamic acid (right).

The following measures can help to minimise errors involving LASA medication: warning labels, especially in high-risk LASA medications with a narrow therapeutic margin, such as cardiovascular or anaesthetic drugs;^{24 25} use of a barcode-driven medication process;²⁶ a standardised organisation of drugs in the operating theatre and physical segregation of TXA from commonly administered spinal medications; use of different shapes, sizes, and colours of ampoules and avoidance of similar packaging and presentation where possible; continuous review of medication errors locally to identify predisposing factors; and development of systematic interventions for prevention.²⁵ Finally, we believe medication errors must be reviewed and acknowledged, with reports shared with the healthcare and scientific communities to raise awareness and to promote timely institution of preventative actions globally.

Learning points.

• The consequences of direct intrathecal administration of tranexamic acid (TXA) range from short-duration seizures and full recovery to profound neurological and cardiac complications and death. Massive sympathetic discharge may be the underlying mechanism behind TXA neuronal and cardiac toxicity.

• As of the date of writing, death was reported in 13 patients, giving a mortality rate of just under 50%. Most patients required intensive care unit admission for management of refractory convulsions and tachyarrhythmias.

• There is a high incidence of permanent neurological injury in survivors, and fatal outcomes are more common in pregnant women (seven deaths in nine cases) likely due to higher concentrations of TXA in cerebrospinal fluid (CSF).

• Most errors resulted from ampoule confusion. This is the first report of drug error resulting from confusion between prilocaine and TXA ampoules, which are both similar looking (packaging and label) and similar sounding.

• Management remains fundamentally supportive. As high concentration in CSF may determine the extent of neurotoxicity, early CSF lavage has been described in three patients with favourable outcomes and is suggested by some authors.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).