Cefepime overdosage with neurotoxicity recovered by high-volume haemofiltration

Abstract

Sir, Cefepime is a fourth-generation cephalosporin widely used for Gram-negative infections. Cefepime neurotoxicity with plasma overdosage may occur in the presence of impaired renal function. We report the case of a patient demonstrating seizure and encephalopathy with plasma cefepime overdosage in the presence of impaired renal function, in which drug clearance with high-volume haemofiltration enabled full neurological recovery. An adult male was admitted to an intensive care unit for surgical site infection following lumbar spinal stenosis surgery. Past medical history comprised hypertension, renal insufficiency, atrial fibrillation, respiratory insufficiency and ischaemic cerebrovascular accident. Usual medication included fluindione, digoxin, acetylsalicylic acid, losartan and amlodipin. Decompressive laminectomy, followed by instrumental fusion of all lumbar vertebrae, was performed. Two weeks later, there was discharge of purulent material from the scar. Surgical debridement was performed. High-level AmpC b-lactamase Morganella morganii was isolated in microbiological sampling. A combination of cefepime 6 g/day with amikacin 1.2 g/day oncedaily was started. Serum creatinine clearance was initially 50 mL/min, but then deteriorated. Four weeks after surgical debridement, the patient demonstrated intermittent obnubilation and myoclonic jerks of the right arm, and was transferred to intensive care. Chest, heart and abdominal examination were unremarkable. Laboratory findings were: blood urea, 30 mmol/L; serum creatinine, 460 mmol/L (measured creatinine clearance 4 mL/min); sodium, 132 mmol/L; calcium, 2.64 mmol/L; serum albumin, 25 g/L; urinary sodium, 53 mmol/L; and urinary potassium, 15 mmol/L. A brain CT scan showed sequelae of former middle cerebral artery ischaemic accident. The electroencephalogram showed status epilepticus requiring an intravenous clonazepam and valproic acid regimen. Mental status gradually returned to normal over the next hour. On the following day, the patient became lethargic and unresponsive. He had no fever and the remainder of his neurological examination was normal. CSF analysis was normal and culture was negative. The diagnosis of valproic acid encephalopathy was considered unlikely, but the epileptic drug was nonetheless switched to phenytoin. A second electroencephalogram was performed, showing slow symmetrical delta and theta activity, but no evidence of epileptic discharges. Cefepime-induced encephalopathy was suspected, and cefepime was stopped. Serum and CSF levels were measured by HPLC: on day 30, 3 h after the last injection, the serum cefepime level was 284 mg/L (usual levels are between 2.5 and 5.1 mg/L) and the CSF cefepime level was 18 mg/L. The antibiotic regimen was then switched to ertapenem 1 g/day once daily. In order to enhance cefepime clearance, renal replacement therapy with high-volume continuous veno-venous haemofiltration (‘CVVHF’) therapy was started using a synthetic polyethersulfone high-flux haemofiltration membrane (Aquamax HF 19, surface area 1.9 m; Edwards Lifesciences S.A.). After 24 h, the serum cefepime level had dropped to 5.6 mg/L (a 99% decrease in serum cefepime level). Haemofiltration was continued for 12 h, until the serum level reached 3.2 mg/L. By then, the patient had fully recovered anterior neurological status. b-Lactam-induced epilepsy is supposed to be a consequence of an inhibitory effect on g-amino-butyric acid (GABA) receptors. Penicillin and cephalosporin b-lactam ring structures share structural similarity with GABA neurotransmitters. Cefepime neurotoxicity may account for a higher risk of mortality compared with other b-lactams (risk ratio 1⁄4 1.26). Seizures, which are the most common clinical indicator of cefepime neurotoxicity, are reported at an incidence of 0.2% among patients treated with cefepime. Encephalopathy, or non-convulsive status epilepticus, has been less often described and is probably underestimated. Cefepime shows a linear pharmacokinetic profile, independent of treatment duration. It is primarily cleared by glomerular filtration as an unchanged drug. Neurotoxicity has most often been described in patients with impaired renal function. In such situations, the blood–brain barrier may demonstrate a higher permeability to drug crossing, due to blood urea increase, glycation, carbamylation or other chemical modification of plasma proteins. Our patient had a history of mild chronic renal insufficiency, and developed acute renal failure while receiving aminoglycosidic antibiotic, sartan therapy and suffering dehydration and sepsis. Cefepime demonstrates low drug-protein binding (,20%), which enables efficient removal of the unbound fraction by renal replacement therapies. Clinical data about cefepime removal by continuous veno-venous haemofiltration are scarce. – 8 Isla et al. showed that cefepime was significantly removed during low-volume continuous veno-venous haemofiltration and haemodiafiltration. In a patient undergoing haemofiltration, therapy accounted for 41.3% of cefepime total body clearance. Drug clearance increased with ultrafiltration flow rate with a 92% serum drug level decrease after 8 h of therapy. High-volume continuous veno-venous haemofiltration could therefore be considered superior to former low-volume therapies, with regard to drug removal. Indeed, in our patient, high-volume continuous veno-venous haemofiltration resulted in a 99% decrease in the serum cefepime level within 24 h. To our knowledge, this is the first report to demonstrate the clearance of toxic levels of cefepime by haemofiltration. Journal of Antimicrobial Chemotherapy doi:10.1093/jac/dkn256