Myoclonic status epilepticus induced by cefepime overdose requiring haemodialysis

Abstract

We report a case of cefepime neurotoxicity characterised by myoclonic status epilepticus with coma, in a context of acute renal failure and requiring one discontinuous conventional haemodialysis. Cefepime is a fourth-generation broad‐spectrum cephalosporin mainly used to treat hospital-acquired Gram-negative infections. Acute neurotoxicity is an increasingly reported adverse effect which occurs predominantly in patients with renal impairment. Renal replacement therapy has been proposed to treat this condition.

Background

Cefepime is a fourth-generation parenteral cephalosporin with bactericidal activity against Gram-positive and Gram-negative organisms. Neurological adverse side effects, particularly seizure, were first described in 1999. Cefepime has a low binding protein ratio and 80% unmodified renal excretion. These are two conditions supporting the use of renal replacement therapy (RRT) to treat severe cefepime toxicity.

Case presentation

An elderly woman with bronchiectasis was administered cefepime, 2 g two times per day, for 10 days to treat an Acinetobacter baumannii superinfection. Her medical history included a pacemaker for cardiac arrhythmia, arterial hypertension and stage 3 chronic kidney impairment (Cockcroft-Gault glomerular filtration rate (GFR) estimated at 46 mL/min). She was also treated with furosemide for acute cardiac failure during her hospital stay. The infection improved with the antimicrobial treatment.

After the tenth day, the patient presented with alertness disorders. She had no fever and the cefepime treatment was complete. On day 12, she developed a myoclonic status epilepticus and received fosphenytoin. She next went into coma and required intubation and mechanical ventilation. She was then admitted to the intensive care unit (ICU).

Her initial blood tests revealed a serum creatinine level of 48.1 mg/L (acute renal failure AKIN 3), a urea level of 1.94 g/L, hyponatraemia with a sodium level of 130 mmol/L, leucocytosis of 23×10⁹/L and metabolic acidosis (pH 7.2, HCO₃₋ 10.8 mmol/L and Paco₂ 26 mm Hg). Her liver tests were within the normal range and her serum albumin level was 15 g/L.

Investigations

The patient’s brain CT scan and cerebrospinal fluid (CSF) analysis were normal. Her metabolic disorders were treated appropriately. Her first electroencephalogram (EEG), recorded during midazolam sedation, showed a pattern dominated by permanent occurrence of periodic generalised activities at 1 Hz, which was then replaced by a slow delta and theta plot with diffuse arrhythmic paroxysmal anomalies compatible with a status epilepticus requiring intravenous administration of thiopental in addition to levetiracetam (figure 1).

Figure 1.

Diffuse arrhythmic paroxysmal abnormalities related to status epilepticus.

Treatment

The patient underwent a 6-hour intermittent haemodialysis (HD) (sustained low-efficiency daily dialysis) 2 days after her admission to the ICU because of persistent renal failure and suspicion of cefepime toxicity. Other causes of status epilepticus were excluded.

Thiopental was stopped on the same day due to evidence of barbiturate impregnation on a second EEG (figure 2).

Figure 2.

Non-reactive isoelectric tracing under thiopental therapy.

Her serum cefepime level, determined on ICU admission, was 86 mg/L, that is, three times the optimum residual therapeutic value (which is around 25–30 mg/L for sensitive Pseudomonas aeruginosa and Acinetobacter) and much higher than the toxicity threshold (35 mg/L).

The first HD lowered her serum cefepime level to 1.7 mg/L. No secondary reascension was observed.

Outcome and follow-up

The patient underwent continuous renal venous-venous HD (CVVHD) with ultrafiltration treatment for the next 3 days for persistent renal impairment. (Her cefepime serum level was measured at 0.8 mg/L before the CVVHD.)

At the end of the RRT, her EEG improved, with the appearance of a slow continuous activity (5 Hz) interrupted by periodic flashes of activity (1 Hz).

Her ICU stay was complicated by an arterial catheter-related Pseudomonas aeruginosa bacteraemia, which was treated with continuous infusion of ceftazidime. Her renal function stabilised, but she had persistent renal impairment (Cockcroft-Gault GFR at 25 mL/min).

Her neurologic status improved progressively. The patient was extubated after 10 days of mechanical ventilation. Furosemide was reintroduced because of dilated cardiopathy.

The patient died 3 weeks after she left the ICU because of a new acute diastolic cardiac dysfunction.

Discussion

The neurotoxicity of cefepime was first reported in 1999.¹ The latency in the development of this neurotoxicity varies from 1 to 10 days. The neurological symptoms typically resolve within 2–7 days after drug discontinuation, but an early diagnosis is a major key to a favourable neurological outcome.^2–4 Cephalosporin-induced neurotoxicity is viewed as being related to concentration-dependent competitive ϒ-aminobutyric acid antagonism.¹ The clinical signs of neurotoxicity encompass altered mental status, confusion, myoclonus, aphasia and seizures.¹ The patients in the cohort examined by Payne et al all had EEG abnormalities, including non-convulsive status epilepticus (25%) and myoclonic status epilepticus (7%). Nevertheless, the extent of status epilepticus is probably underestimated. The available data suggest that the primary risk factor for cefepime neurotoxicity is renal impairment, particularly when the treatment dose is not appropriately reduced.

Cefepime is a 480 Da protein. Its volume of distribution is approximately 0.25 L/kg. It has an unmodified renal elimination and is 16%–20% protein bound. However, an important point to emphasise is that renal impairment, through hypoalbuminaemia, increases the available unbound biologically active fraction of cefepime that may reach the central nervous system.¹ Renal dysfunction can also extend the cefepime half-life from 2 to 13 hours, thereby increasing the circulating drug concentrations.⁵ No toxic concentration threshold has yet been clearly established, but Payne et al suggested that neurotoxicity is associated with serum cefepime levels exceeding 35 mg/L.^{1 6} Furthermore, according to Chaïbi et al, cefepime neurotoxicity may account for a higher risk of mortality when compared with other β-lactams (risk ratio: 1.26).^{7 8} In rare cases, neurotoxicity occurs without renal dysfunction or with a theoretically appropriate renal dose adjustment.^{6 9 10} In 2012, the Adverse Event Reporting System database of the Food and Drug Administration (FDA) identified 59 cases of non-convulsive status epilepticus during cefepime administration from 1996 to 2012. Most cases were patients over 65 years of age and 69% were women. All cases had renal dysfunction. For 95% of the patients, the prescribed dosage of cefepime had not been appropriately adjusted for renal function. The FDA therefore issued a safety warning regarding the necessity of making adjustments for cefepime dosage based on renal function.¹¹

The use of RRTs should be discussed for cefepime-induced status epilepticus, in addition to other symptomatic treatments. Because of its low drug-protein binding (<19%), the unbound fraction of cefepime may be efficiently removed by RRTs.

The clinical outcomes of cefepime removal by continuous venovenous haemofiltration (CVVHF) and CVVHD are increasingly reported.^12–17 For example, Isla et al showed a significant removal of cefepime by continuous RTT with a serum elimination half-life of 4.6±0.9 hours. The report also underlined the absence of significant differences between the polyacrylonitrile and polysulfone membranes used in the CVVHF or CVVHD procedures in terms of drug elimination or diffusion through the membrane.¹⁶

Based on the fact that up to 70% of a given dose might be removed during a 3-hour HD session,¹⁸ CVVHD or CVVHF might not perform better than short conventional discontinuous HD. However, this supposition only takes into consideration the plasma clearance of cefepime, without considering the clearance in a multi-compartmental model with possible transfer between compartments during and after the RRT session. This type of transfer could have been particularly valid in our case, since the cefepime dosage was not determined in the CSF either before or after the CVVHD session. However, the hydrophilic properties of cefepime and the sustained low serum cefepime levels after CVVHD do not support this hypothesis. Overall, the available published data and the present case report support the early initiation of HD in patients with cefepime-induced neurotoxicity, since HD may rapidly lower cephalosporin blood and CSF levels, thereby shortening the cerebral exposure to the drug and improving neurological outcomes.

Learning points.

• Diagnosis of β-lactam-induced encephalopathy must be made early, when neurological disorder occurs, especially in cases of renal impairment.

• Careful cefepime dosage adjustment is essential in cases of renal failure.

• Haemodialysis should be considered early, before confirmation of high serum levels, even in frail elderly patients.