Cefepime-induced generalized tonic–clonic seizures in traumatic brain injury

INTRODUCTION

Cefepime is widely utilized in the management of diverse infectious diseases, often in combination with other spectrum-expanding agents such as vancomycin. However, questions persist regarding the clinical circumstances in which cefepime is most appropriate. The Antibiotic Choice on Renal Outcomes trial found that piperacillin-tazobactam was not more likely to result in acute kidney injury, while cefepime was more likely to cause neurological dysfunction.^[1] A follow-up study analyzed the use of both antibiotics with vancomycin in patients over 90 days and found that piperacillin-tazobactam was associated with an increase in mortality in patients who did not require anaerobic coverage.^[2] These findings underscore the importance of judicious cefepime initiation, particularly in patients with comorbidities that may increase susceptibility to neurotoxicity.

CASE REPORT

A 76-year-old male with a past medical history of diverticulitis, Graves’ disease, pulmonary embolism on warfarin, hypertension, hypothyroidism, and prostate cancer (status post prostatectomy, radiation, and chemotherapy) presented to the emergency department (ED) with signs of weakness and altered mental status. While getting out of bed, he fell on the floor and lay on his back for 1 h before he was found by his wife. Emergency medical services were called, but the patient refused transport. Two hours later, the patient’s wife reported he was no longer answering her, and he was staring into space. She called 911 again, and he was subsequently taken to the ED. It was reported by his wife that the patient had a history of recurrent urinary tract infections (UTI) and had presented similarly during previous instances. Initial ED evaluation demonstrated normal renal function with serum creatinine measured at 0.95 mg/dL and an estimated glomerular filtration rate (GFR) of 83 mL/min/1.73 m². The patient was measured to be 6’1’ and weighed 277 lbs. In addition, he presented with a fever and laboratory evidence of a UTI; however, there were no reported or clinical manifestations of dysuria, which was corroborated by the patient’s wife. Upon arrival, the patient was conscious and could follow commands but exhibited altered mental status with expressive aphasia during triage.

It was determined by the ED team that the patient met systemic inflammatory response syndrome (SIRS) and septic criteria based on a lactic acid level of 2.5 mmol/L, heart rate of 106, and temperature of 101 ºF; thus, they initiated sepsis protocol. The patient was empirically started on piperacillin-tazobactam and vancomycin. In addition, non-contrasted computed tomography (CT) demonstrated an acute subdural hematoma (SDH) measuring 7 mm with a 3 mm left-to-right midline shift. Under the neurosurgery team’s recommendations, warfarin was reversed with prothrombin complex concentrate and phytonadione (vitamin K). In addition, he was initiated on a nicardipine drip for tight blood pressure control and levetiracetam 500 mg twice daily for seizure prophylaxis. One hour after admission, the patient’s lactate decreased to 1.7 mmol/L.

On the morning of hospital day 1, the antibiotics were changed from vancomycin and piperacillin–tazobactam to cefepime 2 g every 8 h to account for laboratory evidence of UTI and clinical concerns for aspiration pneumonia. This decision reflected a desire to maintain adequate gram-negative coverage while avoiding unnecessary anaerobic coverage, as the patient had no clinical indication for coverage of anaerobic pathogens. Review of the patient’s prior urinary cultures demonstrated pan-sensitive Escherichia coli, methicillin-sensitive Staphylococcus aureus, and Enterococcus faecalis, which further supported the selection of cefepime. The patient had no prior overnight events with stable imaging by CT; however, he remained clinically encephalopathic. The patient had macrocytic red blood cell indices and mild thrombocytopenia with a modestly elevated AST, suggesting a potential for EtOH use. On the morning of hospital day 2, the patient had a witnessed episode of right facial twitching. Neurology was consulted for suspicion of a focal seizure. By the early evening, the patient was observed having numerous grand mal seizures, with the longest lasting 3 min. The patient received 4 mg of lorazepam, and the levetiracetam dose was increased from 500 mg twice daily to 1,000 mg twice daily. Between hospital days 2 and 3, the patient was observed having multiple further focal seizures, which were supported by continuous electroencephalogram (EEG) monitoring. The levetiracetam dose was subsequently increased to 2,000 mg twice daily. On the afternoon of hospital day 2, in review of culture data from urine demonstrating pan-sensitive E. coli, cefepime was switched to ceftriaxone 2 g every 24 h. Antimicrobial therapy was continued during this period because a sepsis workup had already been initiated before our team’s evaluation. There was low suspicion for a central nervous system (CNS) source of infection, and antibiotic therapy was to be reassessed following review of culture data.

On the morning of hospital day 4, no further seizure activity was reported. At this time, the patient was able to follow simple commands and reacted to all stimuli, remaining substantially more awake and alert. However, he did experience dysarthria when speaking. Cerebrovascular accident was ruled out with serial CT scans as well as with brain magnetic resonance imaging, neck computed tomography angiography (CTA), and bilateral carotid artery ultrasound. By hospital day 6, the patient was clinically and neurologically stable enough to be transferred to the floor until discharge on hospital day 11. The discharge plan included outpatient imaging with ongoing neurosurgical guidance to determine the appropriate resumptive timing of anticoagulation therapy. Speech therapy was also added to the discharge plan in hopes of limiting any residual dysarthria. He was discharged on levetiracetam 500 mg 2 times/day.

DISCUSSION

In the present case of a 76-year-old man who developed recurrent seizures and altered mental status in the context of empiric antimicrobial therapy, several key features merit careful discussion. These include the significance, or lack thereof, of his minor acute SDH in causing neurologic events, the well-recognized phenomenon of cefepime-induced neurotoxicity (CIN), and broader issues of antibiotic stewardship in the context of infectious workups.

When examining the relevance of acute SDH, imaging and clinical context must be considered. The non-contrasted head CT revealed an acute SDH measuring approximately 7 mm in thickness with a 3 mm left-to-right midline shift. While any acute intracranial hemorrhage has the potential to lower the seizure threshold, these radiographic findings are typically considered insufficient to produce severe encephalopathy or recurrent seizures.^[3] In a systematic review produced by Won et al., midline shift, hematoma size, and laterality were not found to be independent predictors of post-traumatic seizures. Conversely, a lower Glasgow coma scale (GCS) score and surgical intervention were associated with greater seizure risk.^[3] These findings suggest that the patient’s hematoma, though potentially contributory, was unlikely to fully explain the observed profound and sustained neurological symptomatology. In addition, fever can be seen commonly in up to 70% of patients with acute brain injury.^[4] This, too, should have been considered before the initiation of antibiotics.

The diagnostic framework for sepsis and associated need for antibiotic therapy has evolved to improve early recognition while avoiding overtreatment. Under sepsis-3 criteria, sepsis is defined by suspected infection with an acute increase in Sequential Organ Failure Assessment (SOFA) score of two or more points, a model shown to better predict mortality than earlier SIRS-based definitions, particularly in ED settings.^[5,6] However, neurologic abnormalities and fever, both present in this patient, may independently elevate SOFA components and mimic infectious deterioration, complicating early diagnostic decision-making.

Cefepime neurotoxicity represents a well-documented but often underrecognized complication of antibiotic therapy, especially in older adults or those with underlying neurological pathology, which may compromise the integrity of the blood–brain barrier (BBB).^[7] CIN occurs most commonly in patients with renal impairment due to renal clearance of the drug; however, cases have also been reported in patients with normal renal function where higher cefepime exposure is associated with an increased risk of neurotoxicity.^[8] Reported incidence varies, with a prospective cohort study demonstrating CIN in 1% of medically ill patients, increasing to 4.5% in those with a GFR 60–15 mL/min and 16.6% in patients with a GFR <15 mL/min.^[9] Clinically, CIN most often presents with altered mental status, myoclonus, aphasia, or seizures, and may manifest as non-convulsive status epilepticus, mimicking metabolic or structural encephalopathies.^[7] In a systematic review of 135 cases, 73% demonstrated EEG abnormalities, with symptom resolution occurring within a median of 2 days following drug discontinuation.^[7] Notably, nearly one quarter of cases occurred despite appropriate renal dosing, highlighting the role of patient-specific and pharmacodynamic factors such as age, BBB integrity, and CNS vulnerability.^[7] Early recognition and prompt cessation of cefepime remain the cornerstone of management, and long-term anticonvulsant therapy is generally not necessary.^[7] In the setting of mild or moderate traumatic brain injury, long-term anticonvulsant therapy has not been discussed in recent literature.

The neurotoxicity caused by cefepime is closely linked to its ability to penetrate the CNS, which is significantly enhanced when the BBB is disrupted.^[10] Animal studies elucidate the relationship between acute SDH and degeneration of the BBB. Rat studies conducted by Jussen et al., demonstrate that injection of whole blood into the subdural space significantly increased immunoreactivity, leading to measurable neuroinflammation and subsequent BBB leakage.^[10] Similarly, Payne emphasizes that CNS pathology markedly increases CIN risk due to altered BBB permeability and elevated cerebral drug concentrations.^[11] This mechanism plausibly explains the clinical course in the present patient, whose acute SDH and encephalopathy would have facilitated heightened CNS exposure even at standard dosing consistent with the patient’s adequate renal function. As such, initiating cefepime in a patient with active neurological pathology carried a potentially predictable and preventable risk.

The clinical timeline further supports this interpretation. Seizure activity developed within 24 h of cefepime initiation and resolved promptly after the medication was discontinued and replaced with ceftriaxone. This pattern is consistent with the classic presentation of CIN, including rapid onset after exposure, EEG evidence of cortical irritability, and prompt recovery after drug withdrawal.^[11] In the setting of mild, radiographically stable SDH, CIN provides the most plausible unifying explanation for the observed neurologic decline and subsequent recovery. Levetiracetam was continued per neurology recommendations following resolution of seizure activity, with the decision regarding long-term antiepileptic therapy deferred to specialist management. More broadly, this case underscores the necessity of appropriate antibiotic use. Empiric broad-spectrum therapy is frequently initiated in older adults with altered mental status, but this practice must be weighed against the risks of iatrogenic harm, particularly when sepsis is suspected but not yet confirmed. The Infectious Diseases Society of America recommends against treating asymptomatic bacteriuria or non-specific findings in the absence of systemic signs of infection.^[12] In neurologically vulnerable patients, the threshold for initiating an antibiotic with known neurotoxic potential should be even higher. Likewise, the Society of Healthcare Epidemiology of America does not include delirium as a reason to treat suspected UTI in non-catheterized patients.^[13] Taken together, these factors suggest that a more careful approach to cefepime may be appropriate in patients with subtle neurologic vulnerabilities, which could influence tolerance of neurotoxic effects.

CONCLUSION

This patient’s modest SDH was radiographically stable and insufficient to explain his neurologic decline. The timing, progression, and reversibility of symptoms instead point towards CIN potentiated by enhanced vulnerability of the BBB in the setting of intracranial pathology. Clinicians should remain vigilant for medication-induced causes of encephalopathy, particularly when agents capable of penetrating the CNS are administered to patients with preexisting brain injury.