Antibiotic-associated encephalopathy (AAE) may cause delirium more often than is recognized, according to research published March 8 in Neurology. AAE should be considered as a diagnosis for all patients who develop delirium after the initiation of antibiotics, the authors suggest. Improved recognition of AAE could reduce the period of delirium and improve patient outcomes through the early discontinuation of antibiotics.
Delirium occurs in as much as half of hospitalized patients and in as much as 80% of patients in intensive care. “People who have delirium are more likely to have other complications, go into a nursing home instead of going home after being in the hospital, and are more likely to die than people who do not develop delirium,” said Shamik Bhattacharyya, MD, Instructor in Neurology at Brigham and Women’s Hospital in Boston. “Any efforts we can make to help identify the cause of delirium have the potential to be greatly beneficial.”
Investigators Reviewed Case Reports
Although encephalopathy is reported as an adverse effect of antibiotics with a frequency of less than 1%, a 2013 retrospective study of 100 critically ill patients reported a 15% rate of encephalopathy associated with cefepime, a cephalosporin. The result suggests that AAE is underdiagnosed.
Dr. Bhattacharyya and colleagues conducted a comprehensive review of reported cases of AAE to define the specific clinical features, EEG changes, and neuroimaging findings associated with encephalopathy from particular antibiotic classes and individual antibiotics. The investigators excluded case series that did not describe individual patient data. For each described case, the researchers collected data on demographics, antibiotic used, comorbidities, clinical symptoms accompanying encephalopathy, time to toxicity onset and improvement, and laboratory investigations. They calculated a Naranjo score for each case to assess the likelihood that the given antibiotic caused encephalopathy.
Psychosis Was Common
The investigators identified 292 articles describing 391 individual cases from 1946 through 2013. Toxicity reports mentioned 54 antibiotics from 12 classes of antibiotic. Approximately 54% of the 391 cases were male, and the median age was 54.
About 47% of cases had psychosis, which was defined as the presence of delusions or hallucinations. Psychosis was most common in encephalopathy associated with sulfonamides (68%), quinolones (67%), macrolides (63%), and penicillin procaine (68%).
Overall, 14% of cases had seizures. This symptom was most common in AAE associated with penicillin (38%) and cephalosporins (35%). Seizures associated with cephalosporin-associated encephalopathy were nonconvulsive in 54% of patients, and nearly all other reported seizures were clinically apparent.
Approximately 15% of cases had myoclonus. Myoclonus was most common in cases of encephalopathy associated with penicillin (71%) and cephalosporins (41%), but infrequent (≤10%) with other antibiotic classes.
MRI of the brain was abnormal in AAE associated with metronidazole, but normal in all others, with the exception of one case of cefditoren pivoxil toxicity in the setting of acquired carnitine deficiency. Similarly, CT of the brain was normal in all cases except for one case of cerebellar hypodensity with metronidazole toxicity and one report of left thalamic hypodensity with imipenem toxicity associated with generalized seizures and epileptiform discharges on EEG.
EEG, however, was abnormal in 70% of cases of AAE in which EEG was performed, including 95% of cases of cephalosporin-associated encephalopathy. EEG was abnormal in 83% of cases of AAE associated with penicillin, 83% of cases associated with ciprofloxacin, and 69% of cases associated with isoniazid, but the test was performed much less frequently in patients with encephalopathy associated with these antibiotics, thus limiting interpretation of this finding.
AAE was apparent within a median of five days after antibiotic initiation for all individual antibiotics except isoniazid and metronidazole. For these antibiotics, the median length of time from initiation to emergence of encephalopathy was approximately three weeks. Time to resolution of encephalopathy after antibiotic discontinuation was within a median of five days for most antibiotic classes. Metronidazole, for which median time to resolution was 13 days, was an exception.
The median Naranjo scale score for all antibiotic classes was 4, indicating a possible association between the antibiotic and AAE. All reported cases had, by definition, an infection that could not be excluded as a potential cause of encephalopathy. When non-CNS infections were not considered a potential cause of encephalopathy, the median Naranjo scale score was 5, indicating a probable association.
Three Clinical Phenotypes of AAE
Based on these data, the investigators identified three clinical phenotypes of AAE. They characterized Type 1 AAE by onset within days of antibiotic initiation, common occurrence of myoclonus or seizures, abnormal EEG, normal MRI, and resolution within days. Type 1 AAE is associated with penicillin (as an individual antibiotic) and cephalosporins.