Negative myoclonus as the leading symptom in acute cefepime neurotoxicity

Abstract

An 86-year-old woman was diagnosed with hospital-acquired pneumonia with Pseudomonas aeruginosa and treated with cefepime adjusted to her renal clearance. After 4 days, she developed acute-onset negative myoclonus without signs of altered mental status. After ruling out an acute intracranial haemorrhagic or ischaemic stroke as well as other metabolic and endocrine causes of negative myoclonus, the antibiotic was switched to piperacillin/tazobactam due to a suspicion of cefepime neurotoxicity. The patient improved within 24 hours and her symptoms fully resolved within 4 days. These observations suggest a link of the negative myoclonus to acute cefepime neurotoxicity, which may occur without or with minimal alteration of mental status, thus extending its spectrum of clinical presentation.

Background

Should new symptoms develop over the course of drug treatment, potential drug side effects should always be included in the differential diagnoses of possible causes. With increasing antibiotic resistance and rising numbers of very ill patients, newer broad-spectrum antibiotics, such as the fourth-generation cephalosporin cefepime, are increasingly being used. Cefepime neurotoxicity has been reported as a side effect in 1%–3% of treated patients.^1–3 Here, we report on a case of cefepime neurotoxicity that is distinct from the clinical picture described in the literature, thus extending the spectrum of potential side effects that should be anticipated when using this drug.

Case presentation

Our patient is an 86-year-old woman with multiple comorbid conditions, who was hospitalised for an influenza type A infection and who developed hospital-acquired pneumonia on hospital day 8. The patient was empirically treated with 2 g of cefepime twice daily, taking into account her chronic kidney failure with a baseline creatinine of 135 μmol/L and an estimated glomerular filtration rate (eGFR) according to Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) of 30–35 mL/min/1.73 m². The sputum culture grew Pseudomonas aeruginosa and Serratia marcescens. Both bacteria were sensitive to cefepime and piperacillin/tazobactam. Therefore, the therapy with cefepime was continued.

On day 10 of hospitalisation (day 3 of cefepime therapy), the patient developed acute-on-chronic prerenal kidney injury. The creatinine rose to 221 μmol/L with an eGFR (CKD-EPI) of 17 mL/min/1.73 m², thus the dose of cefepime was reduced to 2 g once daily. One day later (day 11 of hospitalisation and day 4 of cefepime therapy), the patient woke up with a shivering sensation, which evolved into an asynchronous loss of muscle tone (during a muscle contraction) of the whole body, predominantly in the upper limbs (video 1), throughout the day. We interpreted this clinical finding as a negative myoclonus. As a consequence of the negative myoclonus of the entire body, she could not perform smooth movements and her neurological examination revealed truncal and gait ataxia. Initially, there was also a direction-changing horizontal gaze nystagmus and horizontal smooth pursuit was slightly saccadic. While experiencing these symptoms, the patient was awake, alert and fully oriented to person, place, date/time and situation. Only a slight neuropsychological slowdown was appreciated in the examination.

Video 1.

Download video file ^{(17MB, mp4)}

DOI: 10.1136/bcr-2020-239744.video01

Other than cefepime, her prescribed medications included the following drugs, which she was taking prior to hospitalisation: rivaroxaban, pantoprazole, amlodipine, bisoprolol, amiodarone, insulin glargine, insulin aspart, linagliptin, vitamin D3, acetaminophen, quetiapine, formoterol/budesonide inhaler and salbutamol/ipratropium bromide nebulizer. Pregabalin, which she had been taking for years (daily dose of 100 mg in the evening), was stopped on day 10 of the hospitalisation (day 3 of cefepime therapy) and rivaroxaban was stopped 1 day later, to prevent toxic accumulation due to acute renal failure. Up to hospital day 7, she was treated with oseltamivir and prednisolone for the influenza type A infection with pulmonary wheezing. The intravenous diuretic furosemide was started on admission because of signs of congestive heart failure and later switched to oral torsemide. The diuretic therapy was stopped immediately on hospital day 10, because of the rising kidney injury.

Investigations and differential diagnosis

As the patient was being treated with rivaroxaban for atrial fibrillation, an unenhanced CT of the brain was obtained to rule out haemorrhagic stroke. In addition, an unenhanced MRI of the brain was ordered for the next day to address potential ischaemic stroke.

An emergency neurological consult was requested and the examination showed the abovementioned negative myoclonus without altered mental status (or with only slight neuropsychological slowdown). The neurological consultant proposed that the disturbances could be due to cerebellar dysfunction caused by a metabolic disorder, most likely uremic encephalopathy in this patient with worsening renal function. Therefore, we rehydrated the patient and monitored her closely. The hypernatraemia of 149 mmol/L (normal range=135–145 mmol/L) worsened up to 154 mmol/L 1 day after the negative myoclonus first appeared and then rapidly decreased to 148 mmol/L within 4 hours of rehydration on the same day. During the increase and decrease of the serum sodium level, the neurological status did not change at all.

The blood sugar at first presentation of the negative myoclonus was 5.5 mmol/L (normal range=3.3–5.6 mmol/L), which was the lowest value measured throughout the day. Thus, there was no relevant hypoglycaemia.

A blood gas analysis under room air showed global respiratory insufficiency with poor oxygenation and mild hypercapnia, but not yet severe hypoxia (pH 7.46 (normal range=7.35–7.45), pO₂ 62 mm Hg (normal range=83–108 mm Hg), pCO₂ 47 mm Hg (normal range=35–45 mm Hg)). The oxygen saturation on peripheral measurements was always stable. Furthermore, the chest X-ray from the same day showed resolving inflammatory infiltrates and demonstrated no signs of congestive heart failure. This was accompanied with a further decline of C reactive protein.

Because of the direction-changing horizontal gaze-evoked nystagmus with a coexisting ataxia, we also thought of Wernicke’s disease. The patient empirically received 900 mg/day of thiamine intravenously, without any improvement the next day. Retrospectively, the thiamine level (thiamine pyrophosphate) on that day was normal (110 nmol/L (normal range=67–200 nmol/L)).

Although being unlikely due to the patient’s history, we thought of hepatic encephalopathy or endocrinopathies as potential differential diagnoses. However, the ammonia and thyroid stimulating hormone levels in the serum were both normal. Serum liver function tests were all within normal range (alanine transaminase 9 U/L (normal range=10–35 U/L), aspartate transaminase 13 U/L (normal range=10–35 U/L), gamma-glutamyltransferase 21 U/L (normal range=5–39 U/L), alkaline phosphatase 93 U/L (normal range=35–105 U/L)).

The next day, an MRI of the brain showed no ischaemic lesions. An electroencephalogram (EEG) demonstrated mild-to-moderate background slowing, but no focal slowing or epileptic potentials were seen.

Because a potential metabolic disorder was suspected in the context of worsening kidney function with an elevated blood urea nitrogen (16.7 mmol/L (2.76–8.07 mmol/L)), a nephrology consult was ordered as well as a neurological follow-up consult. In an interdisciplinary evaluation and discussion, we postulated that the symptoms might be a side effect of the cefepime therapy, possibly due to cefepime accumulation in the context of the acute-on-chronic kidney disease. Cefepime was stopped on that day (day 12 of hospitalisation and day 5 of cefepime treatment) and 3 more days of piperacillin/tazobactam were administered for a total of 7 days of treatment with antibiotics for the pneumonia.

Other than cefepime, the only drug the patient was being treated with that includes neurotoxicity as a known side effect was pregabalin, which had been discontinued 1 day prior to the appearance of the negative myoclonus.

The following day, the negative myoclonus had clearly improved, and 1 day later, it was barely noticeable (video 1). At that time, we decided not to perform a lumbar puncture, which up until that point had been delayed due to the patient’s oral anticoagulation, as an infectious or inflammatory cause was no longer a likely differential diagnosis.

Four days after cessation of cefepime (hospital day 16), the neurological symptoms had fully resolved (video 1). At this time, the creatinine was still rising to a maximum of 296 μmol/L (eGFR 12 mL/min/1.73 m²) and later returned to the patient’s baseline (135μmol/L) with an eGFR of 30–35 mL/min/1.73 m².

Treatment

The treatment of this patient involved considering a drug side effect as the cause of the patient’s negative myoclonus, stopping the suspected drug and monitoring the patient closely to move her up to an intermediate care unit in case of any further deterioration. Furthermore, we treated the slight to moderate hypernatraemia and the kidney injury with intravenous fluids.

Outcome and follow-up

The patient made a full recovery with regard to her neurological symptoms within 4 days (video 1). She had no further neurological symptoms during her remaining 18 days in hospital. Unfortunately, her pulmonary situation did not fully improve. She was prescribed a home oxygen supplementation therapy. In total, the patient was hospitalised for 34 days and she finally opted for palliative care in her nursing home. The patient passed away 8 months later. The cause of death was not investigated by autopsy. As far as we could investigate, her family doctor thought that the most likely cause of death was heart failure.

Discussion

The term myoclonus or positive myoclonus describes brief, involuntary muscle jerks or movements that might be regular or irregular due to uncontrolled muscle contraction.⁴ A tremor is defined as ‘involuntary, rhythmic, oscillatory movement of a body part’⁵ mostly of antagonising muscle groups.⁶ This definition for positive myoclonus opposed to tremor is easily understood and well accepted.

On the other hand, the term negative myoclonus describes a sudden, involuntary relaxation during a tonic muscle contraction or, in other words, an atonic phenomenon.⁴ This term is rarely used. Its effects are more often described by the terms asterixis or flapping tremor if seen in the hands and forearms (this is a misleading usage because myoclonus is not a tremor as stated above).^7–9 Especially in metabolic encephalopathy, internists often use the term flapping tremor to describe a negative myoclonus.¹⁰ Predominantly, the term negative myoclonus is used to refer to muscle relaxation in the limbs (upper and lower) although it can occur in any muscle of the body.⁴ When it affects the lower limbs or the torso, negative myoclonus is often described as ataxia than as negative myoclonus itself.^7–10 Thus, the overlapping use of these terms to describe distinct neurological symptoms has led to some confusion and imprecision.

Differential diagnosis for acute myoclonus

Myoclonus in adults is a symptom of an underlying disease in most cases (70%) and very rarely idiopathic.¹⁰ Its incidence is approximately 1.3/100 000 inhabitants.¹⁰ The differential diagnosis of acute (negative) myoclonus is wide (figure 1), including ischaemic or physical damage to the brain, metabolic encephalopathies, intoxication, drug side effects, infectious causes and psychiatric disorders.^{8 10}

Figure 1.

Differential diagnosis of myoclonus, modified after¹⁰ and reused with the permission of the publisher (Georg Thieme Verlag KG).

In our workup, we considered many secondary causes of myoclonus and were able to rule out (or empirically treat) various potential causes, including intracranial bleeding, stroke, hypoglycaemia, thyroid disease, thiamine deficiency, hypercapnia, hepatic encephalopathy and epilepsy. There was no evidence for trauma or dementia. Because of the patient’s age, an idiopathic myoclonic syndrome seemed very unlikely and was excluded early in the workup. Therefore, at the initial neurological consult, the worsening kidney function with the elevated uric acid in the blood was considered most likely to have triggered the acute onset of negative myoclonus. However, the fact that the symptoms improved later while the renal function was still worsening refutes this hypothesis. There were also no other signs of uremia, such as pruritus or altered mental status, and the patient still had sufficient renal function without the need for dialysis.

Hypernatraemia can cause irritability, restlessness, muscular twitching/tremor, spasticity and hyper-reflexia,¹¹ and, in more severe cases, lethargy, changes in consciousness/altered mental status and seizures.^{11 12} The serum sodium level at the first presentation of the negative myoclonus was only slightly elevated (149 mmol/L). This serum sodium level is unlikely to be the cause of the myoclonus. While the hypernatraemia worsened the next day to 154 mmol/L, the neurological symptoms were unchanged and did not improve due to rehydration and decreasing the sodium level to 148 mmol/L on the same day. This supports the hypothesis that the sodium level was not the cause of the negative myoclonus in our patient.

With a history of an obstructive lung disease, most likely bronchial asthma, former infection with tuberculosis in 1956, pulmonary hypertension shown on echocardiography and the absence of an acute drop of the peripheral oxygenation, we postulated the oxygenation disorder to be chronic and stable. The very slight hypercapnia and respiratory alkalosis were not significantly altered from baseline and therefore highly unlikely to be the cause of the negative myoclonus.

Looking at drug side effects, there were no new drugs taken other than cefepime. Of her regular drugs, only the pregabalin is known to frequently cause drowsiness (28%), sleepiness (16%), as well as more rarely ataxia, tremor, occasionally myoclonus and, of course, other (neurological) side effects.¹³ It is cleared via the kidney and thus the appearance of side effects due to accumulation may arise in acute kidney failure.¹³ However, this theory is refuted by the fact that pregabalin had been taken by the patient for years, that the dose remained unchanged and that it was discontinued 1 day prior to the manifestation of the negative myoclonus. Thus, while negative myoclonus has been described with pregabalin toxicity,¹⁴ this seems unlikely in our patient. Furthermore, myoclonus is not a feature of pregabalin withdrawal.¹⁵

Cefepime neurotoxicity

In a subsequent multidisciplinary evaluation, we considered cefepime as a possible cause of the patient’s acute neurological problems. Typically, symptoms begin after 4 days of treatment with cefepime and are more common in patients with chronic renal insufficiency, acute kidney injury and critically ill patients.^{16 17} Further risk factors are older age, altered blood–brain barrier and drug overdose (especially in renal insufficiency).¹⁷ Resolution usually starts 2–7 days after the cessation of cefepime.^{2 17 18} In two recent studies, the symptoms improved in 81%–89% of patients after stopping or adjusting cefepime therapy^{17 18} and in 50% of patients who had fully recovered.¹⁷

The Food and Drug Administration and the Swiss Agency for Therapeutic Products report encephalopathy, seizures, myoclonus and ataxia as potential signs and symptoms of cefepime neurotoxicity, but give no precise information on how frequent these symptoms are.^{16 19} In a recent Swiss study designed to develop thresholds for cefepime intoxication, the most common neurological symptoms were confusion, agitation and hallucination (62%), followed by reduced consciousness and coma (43%). Myoclonus was observed in 6 of 72 patients (8%), ataxia and flapping tremor were noted in 3% and dystonia was found in 1%.¹⁸ Existing publications emphasise a characteristic progression of symptoms with initial changes in mental status followed later on by seizures and/or myoclonus.¹⁷ This combination of altered mental status and asterixis or myoclonus (positive and negative) is already described for third-generation cephalosporins²⁰ and seems to continue in fourth-generation cephalosporins. A recently published adult case describes only positive myoclonic movements of the upper limbs and trunk due to cefepime neurotoxicity without changes in consciousness or mental status.² To our knowledge, no reports of cefepime intoxication with negative myoclonus as the leading symptom have been published. Also, we did not identify any publications on cefepime neurotoxicity with ataxia or negative myoclonus/asterixis without altered mental status or with just slight neuropsychological slowdown. This might explain why cefepime toxicity was not in our initial differential diagnosis, resulting in a diagnostic delay of about 24 hours.

Because of the remarkably fast and complete recovery after terminating the cefepime therapy, we believe that the neurological symptoms described were due to cefepime intoxication. Third-generation cephalosporins, such as cetazidime, are known to provoke several neurological side effects. Asterixis (negative myoclonus) with confusion or altered mental status is one of them.²¹ Thus, it is very likely that the next-generation cephalosporin might cause the same neurological symptoms.

Our patient fulfils many risk factors for cefepime intoxication, including renal dysfunction, older age and possible altered blood–brain barrier due to the infection. Though we found no formal evidence that cefepime was overdosed, the calculated eGFR might have overestimated renal clearance and thus the prescribed dosage may have inadvertently led to accumulation. Notably, a recent review found that 26% cases of cefepime neurotoxicity occurred in situations without documented overdosing of cefepime. As the cefepime clearance of our patient decreased to <30 mL/min/1.73 m², her risk for neurotoxicity was clearly elevated (54%).¹⁸

A cefepime trough level was not obtained due to the lack of in-house laboratory testing. An elevated trough level could have supported our thesis of cefepime intoxication. What remains unclear is, whether the extent of neurotoxicity correlates with the cefepime trough level. Cefepime neurotoxicity is thought to be related to a concentration-dependent competitive γ-aminobutyric acid (GABA) antagonism. Probably, the higher the concentration in the cerebrospinal fluid is, the more likely is the toxicity. Renal insufficiency, for example, with a decreased cefepime clearance leads to higher cefepime concentrations. The lower the protein binding of cefepime and the higher the accumulation of organic acids, the higher is the penetration of cefepime through the blood–brain barrier and thus its concentration and toxicity.¹⁷ In summary, the sicker a patient is, the more prone to cefepime intoxication he or she is. Unfortunately, cefepime is an antibiotic that is used in very sick patients and thus probably in patients with more chronic illnesses.

If there are rapid changes in the cefepime concentration in the cerebrospinal fluid, the appearance of neurological symptoms might be higher, due to a sudden and rapid inhibition of GABA receptors, thus making the patient’s brain more rapidly susceptible to uncontrolled neurological excitation.

Our patient had renal insufficiency and older age as risk factors for cefepime neurotoxicity. But she received an adequate dose and was not critically ill, thus her blood–brain barrier is expected to be normal, as well as her plasma protein levels and organic acid levels. Thus, we expect a slow change in the cefepime cerebrospinal fluid levels and, therefore, rather mild neurological changes. This hypothesis is supported by the non-specific EEG findings in our patient. A recent case report with positive myoclonus described a non-critically ill patient with slightly decreased renal function, who also had no changes in his mental status, but experienced solely positive myoclonus.²

Patient’s perspective.

When I initially experienced dyspnoea and was admitted by my family doctor to the hospital, I was happy to go there, because I did not like my retirement home. The hospital staff was nice to me and I got better quickly. The doctors organised rehabilitation for me and I was looking forward to this.

Then everything turned out differently. My kidney function began to worsen and I got another pneumonia. This was quite a setback for me. The doctors assured me that I would recover from this rapidly, but for me, this was not so easy. I did not have much strength left and the little I had, I needed to cough, breathe and mobilise myself in and out of bed.

One day, I felt a shaking sensation. It was really strange. It was inside me and everything was sort of trembling. At that time, I could not describe well what was going on. The doctors did not take it for granted at the beginning, but a few hours later, I was not able to control my body’s movements well. My arms repetitiously lost their strength. I could hardly eat or drink because I was spilling the food and beverages all over. Even sitting was difficult. I was very worried by this and as the doctors were not sure what was going on, I got even more frightened. I thought I had had another stroke and, at some point, I thought I was going to die. When the doctors thought they knew what was going on, I could not believe them. It was hard for me to imagine that a drug side effect could have such an overwhelming and tremendous effect on someone’s body. Fortunately, the recovery was fast and I gained back control of my arms and legs.

Despite the improvement, I had lost some faith in the doctors and even more in myself and my health. In the further course of hospitalisation, I realised that I will not fully recover and I expressed the wish of a palliative care plan for my future health. Thankfully, this wish was fully respected by my family and the treating physicians.

Learning points.

• The differential diagnosis of the clinical sign of negative myoclonus should always include drug side effects.

• Neurotoxicity is a frequent side effect of cefepime treatment and physicians should have a high index of suspicion when using it.

• The rate of neurotoxicity and mortality is clearly higher in patients with an eGFR lower than 35 mL/min/1.73 m².

• Even correctly dosed cefepime might cause neurotoxicity. Trough levels might help to confirm cefepime neurotoxicity.