Cefepime-associated neurotoxicity

Introduction
Cefepime is a fourth-generation cephalosporin antibiotic, initially approved for use in 1996 in the United States in IV formulation. It has wide efficacy against both gram-positive and gram-negative bacteria, as well as antipseudominal activity.

In the initial trials of cefepime including 2,032 patients, 2.4% of patients had headache and 0.2% of patients had seizures, but these were mostly felt to not be due to the study drug, and no other neurotoxicological side effects were noted.

In 1998, French physicians published a case of seizure associated with cefepime administration in a patient with baseline renal insufficiency (creatinine clearance 20 mL/min). The first English-language report of cefepime-induced neurotoxicity was published one year later in 1999, describing a 40-year-old patient with end-stage renal disease who developed altered mental status, myoclonus, and a generalized tonic-clonic seizure 5 days after starting cefepime therapy for pneumonia in the setting of a serum cefepime level of 105 mg/L. Symptoms entirely resolved after discontinuing the medication and having two sessions of dialysis. These reports were followed by two reports of status epilepticus in patients treated with cefepime, one with acute renal failure, and the other with mild chronic failure and low body weight, both given cefepime at doses higher than recommended, and both with resolution of symptoms after cefepime was held.

The first case series was published in 2000, describing 19 patients, all of whom were given cefepime in the setting of renal insufficiency, and all of whom had encephalopathy as well as 2-3 Hz continuous sharp-slow-wave activity with triphasic morphology on EEG, with resolution of EEG findings in 1-2 days and resolution of encephalopathy 2-8 days after discontinuation of cefepime. Also in 2000, Saurina et al. published two cases of patients who received seemingly appropriate dosages of cefepime for their renal insufficiency and developed encephalopathy and periodic discharges on EEG anyway. In one of these cases, treatment with benzodiazepines led to improvement in both the EEG and clinical state, suggestive of possible nonconvulsive seizures and similar immediate improvement with benzodiazepines occurred in subsequent case reports.

In 2005-2006 case reports described elderly patients with cefepime-associated encephalopathy in the absence of renal dysfunction. . In 2011, a large series of patients on 1120 patients treated with cefepime noted that continuous epileptiform discharges were noted in 14 patients but that serum creatinine concentrations were normal in the majority of the patients. In 2012 the Food and Drug Administration (FDA) released a safety announcement revised the label to include the concern for nonconvulsive status epilepticus in patients with renal impairment, recommending dosing adjustment in all patients in whom creatinine clearance is less than or equal to 60 mL/min.

Pharmacokinetics
The pharmacokinetics of IV cefepime are notable for the following:
 * Elimination half-life is approximately 2 hours
 * Peak serum concentrations were 82 mg/L for a 1g dose, and 164 mg/L for a 2g dose.
 * Total body clearance was 122-155 mL/min independent of dose
 * 80% renal clearance
 * Average volume of distribution at steady state is 18L in the average-sized adult
 * Special populations
 * Elderly: slightly longer half-life, less clearance, and large volume of distribution, but no change in dose is needed because degree of changes is small
 * Renal failure: quite significant effects on mean serum concentration and half-life, requiring dose adjustment. Half-life prolonged to up to 13 hours in those with ESRD. Even despite dose adjustment, cefepime may still accumulate.
 * Dialysis: Clears approximately 160 mL/min, with half-life of 2.3 hours during dialysis, removing 70% of dose in the first 3 hours of dialysis
 * CSF penetration: CSF-serum concentration ratio was initially thought to be 10%. However, this ratio appears to be increased in patients with toxicity and recent animal studies suggest a true CSF-serum concentration ratio of approximately 20%. Moreover this ratio may be substantially higher in patients with renal failure.

Epidemiology
In a retrospective cohort of 498 patients taking cefepime in Brazil, 5 of 498 (1%) were diagnosed with cefepime-induced encephalopathy, and a study of 1120 patients showed an incidence of 14 in 1120 (1.25%).

Epidemiology is strongly tied to renal function. In the Brazilian series of 498 patients, there were no cases of encephalopathy in patients with a normal GFR. In patients with a GFR of 15-60 mL/min, the cumulative incidence reached 4.5%, and with GFR <15 mL/min, cumulative incidence was 16.6%. In a retrospective cohort of 100 patients with renal insufficiency in the ICU, cefepime neurotoxicity occurred in 15 patients (15%) despite the dose having been appropriately adjusted in four of these patients. However, renal function appears to be unimpaired in approximately 10-20% of cases.

In a French study, most of the reported neurological serious adverse events of cephalosporins from 1997 to 2017 were reported for cefepime (33%) and ceftriaxone (30%).

Pathophysiology
Part of the cephalosporin molecule is similar to bicuculline, a known GABA antagonist. Therefore, cephalosporins may partially block GABA and cause symptoms when at high concentrations in the brain and CSF. However, animal studies suggest that cefepime is only a weak GABA-A blocker and was not able to induce seizures in and of itself in rat brain slices. It was, however, able to cause seizures when other abnormalities were present, however, such as altered ionic composition of the perfusing media, which may be the case in patients with renal insufficiency. Cephalosporin has also been hypothesized to possibly decrease GABA release from nerve terminals or increase excitatory amino acid release. Renal dysfunction not only increases serum levels but the accumulation of toxic organic acids is hypothesized to also compete with cefepime for active transportation back into the blood via the blood-brain barrier, further increasing CSF concentrations.

Most recently, Fernandez-Fernandez and Ameneiros-Lago suggested that carnitine deficiency, as cefepime partially inhibits the OCTN2 carnitine transport, leading to an increased loss of carnitine in the urine and possible carnitine deficiency.

Risk Factors

 * Renal insufficiency (acute and chronic).    In systematic reviews, this was present in 80-87% of cases.
 * Dosing above recommended dosage range. In systematic reviews this was present in approximately 50% of cases.
 * Hematological malignancy
 * ICU stay
 * Advanced age (possibly), although it has also been reported in children .  However, systematic reviews average age was 67-69 years and this was not significant in a multivariate analysis.
 * Low body weight (possibly)
 * Pre-existing nervous system disease (possibly seen in one study, but a larger series showed only a borderline association ).

Signs/Symptoms
Time course: Usually 1-7 days after onset but has been reported further out at 12-16 days. In a systematic review, median was 4 days.
 * Altered mental status
 * Diminished level of consciousness. In systematic reviews, this occurred in 47-80% of cases.
 * Agitation. In systematic reviews this occurred in 11-47% of cases
 * Aphasia. In systematic reviews this occurred in 9-15% of cases.
 * Hallucinations
 * Seizure activity:
 * Generalized nonconvulsive status epilepticus.  In systematic reviews this occurred in 25-31% of cases.
 * Generalized seizures, and convulsive status epilepticus  . In systematic reviews seizures occurred in 11-13% of cases.  Cefepime may carry a 5-fold increased risk of seizure compared with meropenem.
 * Focal seizures and status epilepticus -- these may be either from cefepime itself or from lowered seizure threshold in the setting of a prior lesion. Animal models show that cefepime can lower seizure threshold.
 * Abnormal movements
 * Myoclonus: in systematic reviews this occurred in ~40% of cases.
 * Limb myoclonus
 * Jaw myoclonus
 * Tremor
 * Chorea-athetosis
 * Opsoclonus: reported in one case

Diagnosis
Diagnosis is based primarily on clinical scenario, and a high level of clinical suspicion is important as in 2003 a review suggested a median 5 day delay between symptom onset and diagnosis.

Diagnostic criteria (non-formalized)

 * 1) Neurological symptoms developing several days after initiation of cefepime treatment
 * 2) EEG findings consistent with GPD with triphasic morphology
 * 3) Symptoms and abnormal EEG both resolve within several days of cefepime discontinuation
 * 4) No more likely alternative etiology of toxic-metabolic encephalopathy
 * 5) Elevated serum cefepime levels (if available)

Cefepime levels

 * Serum: One study of 30 patients suggested that the risk of neurological toxicity increased beginning at a serum cefepime trough of 16 mg/L, with 50% of patients having toxicity at a level of 22 mg/L or higher and 100% of patients developed toxicity with serum trough levels of 32 mg/L or higher. However, following cefepime levels is not clearly useful for predicting neurotoxicity as they are not sufficiently sensitive or specific and suggested threshold levels have been variable. However in a larger study of 319 patients in whom cefepime levels were drawn, no patients had neurotoxicity at trough levels <7.7 mg/L, and all patients with trough levels over 38.1 mg/L had neurotoxicity. They proposed that clinicians avoid cefepime troughs over 7.5 mg/L, but acknowledged that some Pseudomonal infections and other drug-resistant infections may require higher doses if they have MICs of 4-8 mg/L.
 * CSF: In several cases in which CSF cefepime levels were measured, they were 6-18 mg/L.

EEG

 * Generalized slowing
 * Generalized periodic discharges, often at 2-3 Hz, often with triphasic morphology, often with frontotemporal predominance. This lies on the ictal-interictal continuum and there is debate about whether these should be treated.  In a series of 19 patients, the EEG pattern resolved with benzodiazepines but did not reverse the confusional state. However, others have reported immediate improvement with benzodiazepines.
 * One series of patients with mixed cephalosporin-toxicity (mostly cefepime) suggested that the triphasic waves had a negative component of the sharp wave (upgoing deflection) exceeding the amplitude of the preceding slow wave, which is less common in other metabolic encephalopathies. This has been described in other studies as well, but no direct comparison with patients with other encephalopathies has yet been done.
 * Compared with patients on meropenem, those on cefepime were 5-fold more likely to have periodic epileptiform discharges on EEG.
 * Paroxysmal myoclonic or convulsive bursts.
 * Stimulus-induced rhythmic sharp waves (SIRPIDs)
 * Lateralized period discharges and focal seizures

NCHCT
Normal.

CSF
Normal.

Prevention
Dose adjustment is necessary in renal dysfunction as per the following table:Other considerations:
 * Critically ill patients: dose adjustment according to serum cefepime trough levels may be useful. It has been proposed that clinicians avoid cefepime troughs over 7.5 mg/L, but some Pseudomonal infections and other drug-resistant infections may require higher doses if they have MICs of 4-8 mg/L.
 * CRRT: no dose adjustment is necessary

Treatment

 * Cessation of treatment with cefepime. Prolonged exposure to cefepime after toxicity occurs may correlate with mortality.
 * Hemodialysis if needed. This has proven effective in a meta-analysis and numerous case reports, as cefepime is rapidly dialyzable
 * Treatment with antiepileptic medications, as needed. Based on pathophysiology it is reasonable to favor benzodiazepines and/or phenobarbital given hypothesized GABA-antagonist properties of cefepime. However, as noted above, there is debate about whether many of the EEG patterns, which lie on the ictal-interictal continuum should be treated.  Tcahpyjnikov and Luedke argue against treatment.  However, many (including the author of this website) would argue that, since true seizures have been associated with cefepime neurotoxicity as well, treatment of these patterns on the ictal-interictal continuum may often be warranted if they appear more consistent with nonconvulsive status epilepticus. Moreover, there are clearly some patients reported in whom both EEG and mental status have improved with benzodiazepines.

Prognosis
In one series of 19 patients, after discontinuation of cefepime epileptiform discharges on EEG resolved after 1-2 days, and confusion state resolved after 2-8 days. A systematic review also confirmed clinical improvement at a median of 2 days (IQR 1-3) after cessation of cefepime. In a retrospective series of 8 patients all of them died, but all of them had sepsis and other severe comorbidities and 6 of 8 were over age 70, so death may not have been related to cefepime at all. A much larger study comparing patients with cefepime with those on meropenem showed equivalent mortality rates, and no increase in mortality with cefepime use has been shown in a meta-analysis.