Epileptogenic Potential of Carbapenem Agents

Mechanism of Action, Seizure Rates, and Clinical Considerations

April D. Miller, Pharm.D.; Amanda M. Ball, Pharm.D.; P. Brandon Bookstaver, Pharm.D.; Emily K. Dornblaser, Pharm.D.; Charles L. Bennett, M.D., Ph.D.


Pharmacotherapy. 2011;31(4):408-423. 

In This Article

Carbapenem Agents


Imipenem-cilastatin was the first antimicrobial in the carbapenem class, receiving United States Food and Drug Administration (FDA) approval in 1985. It is available as a combination product in a 1:1 ratio of imipenem and the dehydropeptidase- 1 inhibitor, cilastatin. The presence of cilastatin decreases imipenem's degradation and inherent nephrotoxicity due to accumulation in renal tubules, making the combination both safer and more efficacious than imipenem alone.[6] Imipenem-cilastatin has been approved for a wide range of clinical infections with the noted exception of meningitis, due to both a lack of safety and the lack of efficacy data in this setting.[6] A more recent review describes the evidence of imipenem-cilastatin–induced seizures in the treatment of meningitis and urges caution in using the agent for CNS infections.[21]

The protein binding of imipenem and cilastatin is approximately 20% and 40%, respectively.[6] Imipenem-cilastatin has good penetration into most tissues, including CSF, with mean concentrations in the inflamed meninges of 2.6 μg/ml (range 0.5–5.5 μg/ml) 2 hours after dosing and 1 μg/ml (range 0.26–2 μg/ml) in the noninflamed meninges 4 hours after a 1-g dose.[6] Initial dosing recommendations for most infections are in the range of 2–3 g/day in 3–4 divided doses. The maximum daily dose is 50 mg/kg or 4 g; however, high-dose therapy is indicated only in severe life-threatening infections in patients with adequate renal function because of the dose-dependent seizure potential.[6] Because of imipenem-cilastatin's renal elimination, dosage reductions must be taken in patients with estimated creatinine clearances less than 70 ml/minute (using the Cockcroft-Gault equation) or in patients weighing less than 70 kg.

In the largest review of imipenem-cilastatin safety, published in 1986, the authors described the results of the first 3470 patients in phase III clinical trials.[22] Patients received daily doses of 1–4 g, and an overall 2% seizure rate among imipenem-cilastatin–treated patients was reported. Additional review of the phase III noncomparative trial populations (1754 patients) identified high-dose therapy (> 25 mg/kg), renal impairment, and preexisting CNS disorder including history of seizures or CNS lesions as the most important risk factors for seizures.[8] When patients with a creatinine clearance of less than 20 ml/minute were analyzed separately, the frequency of seizures was 16.1% in those receiving doses greater than 3 g/day compared with 11.8% in patients receiving 0.5–1.9 g/day. This dose-dependent relationship is further illustrated by a seizure rate of 0.3% in patients without renal or CNS disease who received appropriate imipenemcilastatin doses versus a seizure rate of 1.3% in patients receiving excessive doses.[8]

A review of the literature published since the 1986 safety analysis revealed 105 new prospective, randomized clinical trials of imipenemcilastatin therapy for a variety of infections.[3] Thirty-two of these studies reported at least one seizure episode related to imipenem-cilastatin therapy, with reported rates ranging from 0.4–33.3%.[23,24] Of note, the trial reporting the highest rate of 33.3% was a study of pediatric patients with meningitis.[24] Excluding that trial, the highest reported rate of seizures was 10.3%.[25] Twelve of these 32 studies reported at least one identifiable risk factor in patients experiencing seizures. As discussed below, seizures observed in these studies often occurred in patients with existing CNS disorders and decreased seizure thresholds.

Twenty-four of the 105 studies compared imipenem-cilastatin with other antimicrobial therapy for treatment of neutropenic fever. Of these, 10 studies reported at least one seizure occurring in patients randomly assigned to receive imipenem-cilastatin.[23,25–33] A total of 2755 patients received imipenem-cilastatin therapy for neutropenic fever in the 24 studies reviewed, and 20 treatment courses resulted in seizures. The overall rate of seizures ranged from 0.5–3.4%. However, preliminary data from the first neutropenic fever trial initially used high-dose (1 g every 6 hrs) imipenem therapy and had an initial seizure rate of 10.3%.[25] The study protocol was amended to use a lower dose of 500 mg every 6 hours, and the rate of seizures was reduced to 0.9% in the amended protocol group. Only two studies reported the presence of risk factors in patients experiencing seizures, and among those patients, three had risk factors reported including renal failure, CNS hemorrhage, and progression of neoplasm into the meninges.[26,29]

Eighteen trials have been published on imipenemcilastatin administration for community-acquired, health care–associated, or ventilator-associated pneumonia. These studies primarily used a dosage of imipenem-cilastatin 1000 mg every 8 hours. Four trials reported seizures with rates ranging from 1.1–6%.[34–37] A total of 1661 patients received imipenem-cilastatin therapy, with 25 patients experiencing seizure activity. Of the 25 patients experiencing seizures, eight had risk factors: history of meningitis (one patient), history of seizures (one patient), and traumatic brain injury (six patients). The increased rate of seizure in these studies is interesting and is not completely explained by the use of a dosage of 1000 mg every 8 hours since the pharmacokinetics of 500 mg every 6 hours is considered equivalent to 1000 mg every 8 hours.[38] However, it is possible that increased peak concentrations produced by 1000-mg doses played a role by increasing peak CNS concentrations.

Imipenem-cilastatin administered intraperitoneally has been reported in five studies of patients with peritonitis who received continuous ambulatory peritoneal dialysis. Imipenemcilastatin therapy was administered to 373 patients. Two studies described a total of four patients who experienced seizures after alternating instillations of imipenem-cilastatin into the peritoneal dialysate solution.[39,40] Two of these patients had previous stroke as a reported risk factor. Also, two of the 27 studies in the setting of intraabdominal infections reported three patients who developed drug-associated seizures among 3306 patients who received imipenem-cilastatin therapy.[41,42] Only one patient was reported to have previous risk factors. Finally, among four studies that included 1124 patients who received imipenem-cilastatin for skin and soft tissue infection, only one reported a single patient as having a "significant" seizure; no risk factors were reported.[43]

In 26 studies evaluating imipenem-cilastatin as treatment for a variety of infections (community-acquired pneumonia, sepsis, skin and soft tissue infections, combined), 12 trials reported patients experiencing a seizure, with the rates ranging from 0.5–4%.[44–55] Almost 1000 patients received imipenem-cilastatin therapy, with 16 patients experiencing seizures. When compared with the neutropenic fever studies, these studies of multiple infection types were smaller in size and two of the comparator groups reported seizures as an adverse effect. Because of the heterogeneity of the infections, the dosages of imipenemcilastatin ranged from 500 mg every 12 hours to 1000 mg every 8 hours. Four trials reported presence of risk factors in patients who developed seizures.[44,48,51,54] Among the 16 patients who experienced seizures, six had risk factors including renal failure, CNS hemorrhage, and history of seizures or anoxic brain injury. Another patient was reported to have risk factors present that were not otherwise specified in the study.[51]

Case reports and single-center retrospective reviews of seizures related to imipenem-cilastatin therapy were published as early as 1984. Fourteen published reports representing 37 individual patients were identified for review (an appendix with full details is available on request from the corresponding author).[56–69] Nearly all patients had at least one documented risk factor, with renal insufficiency being the most common. In eight cases, dosages greater than 30 mg/kg/day were reported.[56,58,59,63] One case was reported with intraperitoneal administration, and one case described seizure activity with a single preoperative dose.[66,69] Eight cases reported seizures in patients receiving interacting drugs including theophylline and derivatives, and cyclosporine.[60,61] Although high-grade evidence is limited, the renal toxicity may have contributed to an increase in imipenem-cilastatin levels.[60]

More recent studies comparing imipenemcilastatin with newer carbapenem agents have demonstrated a lower seizure rate than previously observed in early studies. This reduced rate is likely explained by the lower doses of imipenem-cilastatin (500 mg–1 g every 8 hrs) and increased awareness of the need for dosage adjustments in renal dysfunction. For instance, newer clinical trials comparing imipenemcilastatin with meropenem almost uniformly ensure dosage adjustments for both carbapenems based on renal function as a part of study protocols.[53,70,71]


Meropenem is the second carbapenem approved by the FDA for use against multidrugresistant infections. Currently, it is approved for complicated skin and skin structure infections, intraabdominal infections, and, unlike imipenemcilastatin, pediatric bacterial meningitis (patients aged ≥ 3 mo).[72] Meropenem is a hydrophilic molecule, is minimally protein bound (≤ 2%), and has one inactive metabolite. It is able to penetrate most human body tissues well, including the CSF. After administration of 20- and 40-mg/kg doses of meropenem, drug concentrations in the CSF of the inflamed meninges were a mean of 1.1 μg/ml (range 0.2–2.8 μg/ml) and 3.3 μg/ml (range 0.9–6.5 μg/ml), respectively; however, in noninflamed meninges, the drug concentration for both doses was lower at 0.2 μg/ml (range 0.1–0.3 μg/ml).[72]

Safety profile reviews of meropenem were reported in 1995, 1999, and 2007.[73–75] Each review assessed the reported frequency of seizures either in phases III–IV clinical trials or follow-up from patients who had received meropenem for approved indications. The latest safety review included 54 efficacy studies (6308 meropenem treatment exposures among 6154 evaluated patients), 51 of which were randomized, controlled, prospective, multicenter clinical trials; 47 studies were international. No seizures were reported among 1148 pediatric patients, including 383 pediatric patients who received meropenem for meningitis.[74]

In adults, very few seizures were reported during treatment (Table 1).[70,71,76–78] The clinical trials and safety analysis reported all seizures that occurred during treatment. However, seizures were also evaluated for causality to meropenem. All seizures deemed by study investigators as due to meropenem were then separately reported as drug related. In most cases of seizure, the dosages used were 500 mg every 8 hours in adults[70,71] and 20–40 mg/kg every 8 hours in pediatric patients.[76–78]

There is one detailed case report of meropenem-associated seizures developing in a 15-year-old girl who had a history of progressive myoclonic epilepsy.[79] Her initial electroencephalogram showed a high frequency of generalized spike-and-wave discharges. After receiving multiple antibiotics for pneumonia and persistent fever, her antibiotic regimen was changed to teicoplanin and meropenem 1 g every 8 hours on hospital day 3. The patient's initial seizures improved before meropenem initiation; however, on hospital day 6 (3 days after drug initiation), the patient had a seizure relapse and required an increase in her antiepileptic drug dosage. Meropenem was discontinued, and the patient's seizures subsequently improved by hospital day 12; she returned to her baseline mental status on hospital day 16. Of note, the patient was receiving concomitant valproic acid therapy. However, the authors reported that the patient maintained therapeutic levels throughout meropenem treatment.

Study criteria for most treatment trials involving meropenem, especially those assessing efficacy with meningitis, typically excluded patients with underlying CNS disorders. There is evidence that meropenem elimination from the CSF is slower than serum elimination in patients with hydrocephalus (mean ± SD half-life 7.4 ± 2.9 vs 1.7 ± 0.6 hrs).[20] This slower drug elimination may lead to drug accumulation over time and potential toxicity in this high-risk population. It is important to consider that data are limited on the safety of meropenem in patients with underlying CNS disorders, and these patients are at high risk for neurotoxic adverse events. However, a large variety of data indicate that meropenem carries a low seizure rate relative to imipenem-cilastatin, and a paucity of case reports of seizures exist despite its widespread use. This indicates that meropenem is safer than imipenem-cilastatin in patients with underlying CNS disorders.


Doripenem is a newer carbapenem, first approved by the FDA in 2007 and currently approved for treatment of complicated intra-abdominal infections and complicated urinary tract infections, including pyelonephritis in adults.[80] It is a hydrophilic molecule, minimally protein bound (8%) and is metabolized to one inactive metabolite (doripenem-M1).[80,81] Doripenem penetrates most human tissues well; however, to our knowledge, no data in humans evaluating doripenem penetration into CSF are available.[80,81] In patients with renal dysfunction, the normal half-life of 1 hour extends to 5 hours, which can be significant when considering drug accumulation and drug toxicities.[81]

Although no human data exist on the clinical effects of doripenem in the CSF, there are data about rats and canines. At all dosages administered either intravenously in rats or by intracerebroventricular injection in dogs, no significant changes in their electroencephalograms or in behavior for either group were noted.[82] Based on these data, doripenem is believed to have a low likelihood of inducing seizures.

The reported seizure rate of doripenem overall is very low, with only six seizures documented in manufacturer-sponsored clinical trials that included 1817 patients (0.3%).[83] None of these patients experienced status epilepticus. The overall rate of seizures with doripenem is described in Table 2. In a prospective, randomized, double-blind, controlled trial of intravenous doripenem 500 mg every 8 hours versus intravenous levofloxacin 250 mg every 24 hours for a complicated urinary tract infection, including pyelonephritis, no patients experienced doripenem-related seizures.[84] No seizures were reported in any patients in a prospective, randomized, double-blind, controlled trial of doripenem 500 mg every 8 hours (237 patients) versus meropenem 1000 mg every 8 hours (239 patients) for complicated intraabdominal infections.[85]

In an international, multicenter, prospective, randomized, open-label trial, doripenem was compared with piperacillin-tazobactam for nosocomial pneumonia among 225 and 223 patients, respectively.[86] Three patients in the doripenem group experienced seizures during treatment compared with six patients in the piperacillin-tazobactam group. Patients receiving doripenem who experienced seizures all had underlying CNS disorders. One patient had a subarachnoid hemorrhage and a history of hypertension, the second patient had known alcoholism and possible alcohol withdrawal with tremors, and the third patient had a history of arterial hypertension, stroke complicated by right and left hemiplegia, and seizures.[83]

A second international, multicenter, prospective, randomized, open-label trial compared doripenem 500 mg infused over 4 hours every 8 hours versus imipenem-cilastatin at varying doses for ventilator-associated pneumonia.[37] In 262 evaluated patients receiving doripenem, seizures occurred in 3 patients (1.1%) versus 10 (3.8%) of 263 patients in the imipenem-cilastatin arm. The doripenem-treated patients who experienced a seizure during treatment all had underlying CNS disorders, similar to previous studies. Two patients had a subarachnoid hemorrhage, and the third patient had a history of epilepsy.

In postmarketing reports through December 2008, the follow-up of pharmacovigilance studies had identified some additional instances of seizures related to doripenem.[87] In 508,176 patient exposures, 10 case reports of seizures were identified, with eight alone from Japan. Direct causality could not be established for these cases, and the full case details have not been reported. Three patients were female, six male, and one report did not include sex information. The mean age of the patients was 70 years (range 31–89 yrs). Two cases were likely not related to doripenem because the timing was deemed implausible (seizures 2–12 days after treatment), and two cases did not provide enough medical information for causality assessment. Of the remaining cases, five had noted seizure confounders including underlying disease states and electrolyte abnormalities, whereas in the final case the patient had a known CNS disorder and acute renal failure requiring dialysis. Even in postmarketing reports on doripenem, cases of de novo seizures without an underlying CNS disorder or risk factor are extremely uncommon.


Approved by the FDA for use in 2001, ertapenem is a 1-β-methylcarbapenem that possesses a narrowed spectrum of activity when compared with the other licensed carbapenems, with limited or no activity against several non–lactose-fermenting gram-negative bacilli, including Pseudomonas species. Ertapenem offers the advantage of a once-daily dosing schedule due to its prolonged half-life of approximately 4 hours, relative to 1 hour for the other class members. The approved dosage is 1000 mg once/day intravenously or intramuscularly in patients with normal renal function. Protein binding is reported to be 85–95%, which is inversely proportional to plasma concentrations.[18,88] The reversible nature of the albumin binding contributes to the prolonged half-life. To date, it has been primarily used as an agent in complicated skin and skin structure infections, complicated intraabdominal infections, and surgical prophylaxis, with limited investigation into its use in meningitis. One investigation demonstrated equal efficacy of ertapenem with standard-of-care ceftriaxone and vancomycin for penicillin-sensitive and penicillin-resistant pneumococcal strains in a rabbit model of meningitis.[89] After administration of doses of 60 mg/kg, CSF penetration was 7.1% and 2.4% in inflamed and noninflamed meninges, respectively.

Early dose-finding and safety evaluation studies of 250 healthy volunteers examined doses ranging from less than 1 g up to 3 g.[90] No seizures were reported. To date, 16 studies of adults receiving ertapenem therapy have been published. Fourteen were randomized controlled trials with active comparator arms including cefepime, piperacillintazobactam, and ceftriaxone with or without metronidazole.[91–104] Two studies included open-label experience in treating early ventilator-associated pneumonia.[105,106] Overall, 2927 patients received a single 1-g dose or greater of ertapenem in the 16 studies. The breakdown of the studies was as follows: intraabdominal infection with the comparators piperacillintazobactam or ceftriaxone plus metronidazole (seven studies); community acquired pneumonia with the comparator ceftriaxone (three studies); hospital-acquired or ventilator-associated pneumonia with the comparator cefepime with or without metronidazole (two studies); urinary tract infection with the comparator ceftriaxone (two studies); and skin and soft tissue infection with the comparator piperacillin-tazobactam (two studies). Safety as an end point was evaluated in each report, typically extending to 14 days after therapy discontinuation.

Of the nearly 3000 total patients evaluated, three seizures were reported as likely drug related, resulting in a rate of around 0.1%.[92,97,98] Two patients were older than 75 years, and the third had a history of seizure disorder and frontal lobe resection.[97,98] Also, the third patient with significant neurologic history also received an increased dose of 2 g/day the day before seizure activity.[98] Seizures occurred on day 10 of therapy in two patients; the timing for the third case was unknown. The manufacturer reports a rate of 0.2% of drug-related seizures, and an overall rate of 0.5% in patients receiving ertapenem in clinical trials.[18,90] A safety and tolerability analysis published by the manufacturer in 2004 described seven additional study patients who experienced seizures after ertapenem therapy. These occurred at or more than 2 days after therapy and were not deemed drug related.

Several case reports and case series have been published that associate ertapenem use with seizures. A total of nine reports in the literature are detailed in Table 3.[107–111] Available data for patients with reported seizures deemed drug related during ertapenem clinical studies are also included in Table 3.[92,97,98] The first seizure episode occurred a mean ± SD of 6.7 ± 2.5 days after start of ertapenem therapy. Of the 11 patients who had age and sex data reported, 9 (82%) were male and the average age was approximately 64 years. Nine patients had a previous history of CNS disorders, including four who had a documented history of seizure disorder. Details on renal function are not completely known, but at least five patients had some degree of renal dysfunction (creatinine clearance < 60 ml/min) including one who required continuous ambulatory peritoneal dialysis. One patient with progressive terminal disease and a standing "do not resuscitate" order died after one of the seizure episodes.

Ertapenem has also been evaluated for surgical prophylaxis in patients undergoing colorectal procedures. To our knowledge, there are no reports of postsurgical seizures in patients who received ertapenem; although, it is conceivable that a high-risk patient with a history of seizure disorder may be at risk for additive proconvulsant effects.[112]

These data suggest that elderly patients may be at higher risk for ertapenem-induced seizures. A manufacturer review of the influence of age (≥ 65 vs < 65 yrs) on occurrence of adverse drug events did not identify age-related differences in seizure rates.[90,113] A pharmacokinetic analysis of ertapenem in elderly patients demonstrated only slight, nonsignificant increases in area under the concentration-time curve (AUC) and decreases in clearance compared with those parameters in healthy adult volunteers, with no need for dosage adjustment.[114,115] A more plausible explanation may lie in complexity of the patients, with most having underlying CNS disease and/or seizure disorders. Also, the need for ertapenem in these patients was predicated by presumed or confirmed multidrug-resistant pathogens, likely due to previous antibiotic courses and recurrent hospitalizations.

The presence of hypoalbuminemia may also be an independent risk factor for seizures associated with ertapenem compared with the other carbapenems, increasing the amount of free drug and subsequently maximum concentration and AUC.[89,115] The clinical effect of this difference on seizure propensity is unknown. This increase in free drug concentration could potentiate the movement of ertapenem across the blood-brain barrier. Conversely, the increase in free drug increases clearance and volume of distribution, making it difficult for patients to achieve time above minimum inhibitory concentration targets.[115] If doses were increased to achieve optimum pharmacodynamic markers in these patients with increased drug clearance, such as those in the intensive care unit, caution should be taken in high-risk patients, given the possibility of potentiating seizure activity.

Agents Available Outside the United States and Investigational Agents

Panipenem-betamipron is a parenteral carbapenem compound indicated for the treatment of lower respiratory tract infections, urinary tract infections, surgical site infections, and obstetric and gynecologic infections in Japan, China, and Korea, but not in the United States. Similar to imipenem, panipenem is hydrolyzed in the kidney by dehydropeptidase-1 and is coformulated with the renal tubular transport inhibitor betamipron to reduce nephrotoxic potential. Animal models investigating the potential neurotoxicity of this agent suggest that the proconvulsant potential related to GABA interactive binding is less than that of imipenemcilastatin, but slightly greater than that of meropenem.[17] Cumulative clinical trial data demonstrate a low propensity for seizures, with reports of less than 0.1%. However, at least one report of a fatal seizure was attributed to panipenem-betamipron administration.[116]

Currently available only in Japan and China, biapenem is approved for the treatment of serious bacterial infections.[117] It possesses a similar spectrum of activity to meropenem and imipenemcilastatin. Several murine and rat models have compared the proconvulsant activity of biapenem with imipenem-cilastatin, panipenem-beta-mipron, meropenem, and cefazolin. Biapenem demonstrated a lower proconvulsant potential than all comparators, including cefazolin, and seems to lack substantial neurotoxic potential. Investigators attribute this to its weaker affinity for GABA and structural differences including a methyl group at the 1-β position and a triazolam radical side chain at position 2. In clinical trials, drug-attributable seizures were not reported, and further postmarketing cases have not been reported in the literature.

Several additional carbapenem compounds, both oral and parenteral agents, are being evaluated in phases I and II trials. The first oral carbapenem, tebipenem, is in development primarily in Japan.[118] Although human data are quite limited, in mice and rat models investigating proconvulsant properties of carbapenems, animals exposed to oral tebipenem 1000 mg/kg (or 300 mg/kg intravenously) demonstrated no convulsive activity. Intraventricular administration of tebipenem 100 μg induced clonic seizures in 7 of 10 animals. Compared with imipenemcilastatin in the same model, the proconvulsant potential is very low, and it is believed that this will translate well into human use.[118]

Tomopenem is a novel 1-β-methylcarbapenem with added in vitro activity against methicillin-resistant Staphylococcus aureus (MRSA). Similar to tebipenem, other than single-dose pharmacokinetic studies, human data are not available. In rat and mice models, tomopenem administered in doses up to 1000 mg/kg intravenously did not result in any proconvulsant activity, unlike comparators including imipenem-cilastatin. The potential for seizure activity in humans is hypothesized to be very low in clinical trials.[119] Phase II trials are under way in the United States for a second anti-MRSA carbapenem, razupenem (PZ-601). No data are available regarding its proconvulsant potential.

Several drugs in a related class of antimicrobials, the penems, are in development, including faropenem and sulopenem. Faropenem received a nonapproval letter from the FDA in late 2006 due to inadequate trial comparators and questionable data sets. Available data suggest a low potential for serious adverse events and no reports of drug-induced seizures.[120] This is consistent with in vitro data suggesting a lower proconvulsant potential of the penems compared with the carbapenem class.[121] Limited published data do not allow for further discussion with regard to seizure activity of these agents.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.