Meropenem-induced liver injury and beta-lactam cross-reactivity

Abstract

A 63-year-old man admitted to hospital for the management of a frontal lobe abscess developed elevated liver enzymes within 48 hours of receiving meropenem. Liver enzymes reached a maximum at 5 days postadministration of meropenem, with alanine aminotransferase 1160 U/L, aspartate aminotransferase 787 U/L, alkaline phosphatase 297 U/L and gamma-glutamyltransferase 252 U/L. Meropenem was ceased and liver function normalised. Meropenem was administered for a second time later in the patient’s admission and again the patient developed rapidly increasing liver enzymes, with a mixed hepatocellular/cholestatic pattern. Other possible causes of liver injury were excluded following extensive investigations, and the patient’s liver enzymes continued to normalise following meropenem discontinuation. The patient was asymptomatic during the admission and was transferred to a rehabilitation facility. This case demonstrates that meropenem can cause severe liver injury and that early recognition of drug-induced liver injury is important.

Background

Idiosyncratic drug-induced liver injury (iDILI) accounts for 11% of cases of acute liver failure in the USA and is associated with a significant mortality and morbidity, and up to 10% of patients with the condition will die or require liver transplantation.¹ In paracetamol overdose, the liver injury that results is dose-related and has a predictable clinical course with onset at 24–36 hours, and peak at 72–96 hours, followed by recovery or death.² In contrast, iDILI follows a much more variable clinical course, with latency between drug exposure and onset ranging from 2 days to 90 days.¹ Diagnosis of iDILI can be challenging and requires a high index of suspicion and exclusion of alternative causes of liver injury.

Case presentation

A 63-year-old man presented to the hospital with fever and confusion on a background of endoscopic resection of sinonasal adenocarcinoma 6 months prior. CT and MR brain showed an abscess in the left frontal lobe and the patient was commenced on intravenous vancomycin and meropenem (2 g every 8 hours) as empiric antimicrobial therapy (day 0). A biopsy of the abscess was performed and Streptococcus constellatus was isolated in culture. On day 2, antibiotic therapy was changed to intravenous ceftriaxone 2 g two times per day and metronidazole oral 400 mg two times per day.

On day 3, he was noted to have elevated liver enzymes: alkaline phosphatase (ALP) 297 U/L (normal <110 U/L), gamma-glutamyltransferase (GGT) 252 U/L (normal <55 U/L), alanine aminotransferase (ALT) 168 U/L (normal <45 U/L) and aspartate aminotransferase (AST) 383 U/L (normal <35 U/L) (figure 1). Liver enzymes were within the normal range on day 1. Liver enzymes continued to rise and peaked on day 5, with ALP 658, GGT 575, ALT 1160 and AST 787. Bilirubin peaked at 39 μmol/L (24 μmol/L direct bilirubin) on day 4.

Figure 1.

Trend of liver enzymes before and after administration of meropenem (blue bar) and ceftriaxone (red bar). ALT (black line) and ALP (grey line) are shown. ALP, alkaline phosphatase; ALT, alanine aminotransferase.

The patient was asymptomatic with no abdominal pain, had no nausea or vomiting, and remained afebrile and haemodynamically stable. Abdominal examination was unremarkable and there was no jaundice or signs of chronic liver disease. He previously consumed 3–4 standard drinks of alcohol daily but had been abstinent for the previous month.

Drug-related acute liver injury was suspected and ceftriaxone and metronidazole were ceased on day 5. Vancomycin was restarted at this time. The patient underwent image-guided aspiration of the intracerebral abscess on day 9, and craniotomy and repair of the anterior cranial fossa defect on day 22. Multiple extradural tissue samples from the day 22 procedure had gram-negative bacilli on gram stain. Meropenem 2 g every 8 hours was restarted on day 23 for empirical treatment of gram-negative bacilli.

Marked elevation of liver enzymes was noted again on day 25, with ALP 498, GGT 864, ALT 2030 and AST 2660. A DILI was suspected and all antibiotics were ceased on day 25. Intravenous moxifloxacin 400 mg daily was commenced on day 26, and liver enzymes were significantly improved by day 27. The patient remained asymptomatic and afebrile during this period. Vancomycin was restarted on day 28 and liver enzymes continued to decrease.

There was no evidence of sepsis during the admission. There was no documented hypotension or hypoxia preceding either episode of liver injury.

The previous year, the patient had a similar episode of elevated liver enzymes that developed 48–72 hours after receiving cephazolin, a first-generation cephalosporin.

Investigations

Numerous investigations were performed to exclude common causes of elevated liver enzymes. Ultrasound abdomen performed on days 6 and 26 showed hepatic steatosis and a small haemangioma but no other abnormalities. Hepatitis B core antibody was positive but the surface antibody was negative and repeated hepatitis B quantitative PCR was negative. Hepatitis A, C and E serologies were negative. Hepatitis C PCR was negative. Paracetamol level on day 25 was 17 mg/L (the patient was on regular paracetamol for analgesia), this level was within the expected range. Antinuclear antibody, antismooth muscle antibody, anti-liver–kidney microsomal type 1 antibody, antineutrophil cytoplasmic antibodies and antitissue transglutaminase antibodies were negative. Alpha-1-antitrypsin levels, serum ceruloplasmin and serum ferritin were normal.

A liver biopsy was not performed in this case, as it was felt that it would not alter the patient’s management and that a diagnosis of DILI was clearly established on the basis of the temporal relationship between drug exposure and liver injury, and normalisation of liver enzymes on cessation of the implicated drug.

Outcome and follow-up

The patient remained well and was transferred to another facility for ongoing rehabilitation. Liver enzymes progressively improved since the cessation of beta-lactam antibiotics, and the patient was advised to avoid beta-lactam antibiotics in future.

Discussion

Meropenem is a carbapenem antibiotic with a broad spectrum of activity and is stable to most beta-lactamases, other than the carbapenemases.³ Meropenem is the preferred carbapenem for use in central nervous system infections because it is a lower risk of inducing seizures compared with imipenem, but retains excellent penetration into the meninges.⁴

DILI is a frequent differential diagnosis in patients with an acute liver injury. Antibiotics have been found to be the most common drugs leading to hepatotoxicity.⁵ In a prospective study of 1257 patients with suspected DILI, 71% of the 41 cases of DILI with latency ≤7 days were caused by antimicrobial agents, with 1 case attributed to meropenem.⁶ The case we present featured a latency of approximately 48 hours between the first administration of meropenem and biochemical evidence of liver injury.

Meropenem is a safe and well-tolerated drug that has been on the market for >20 years.⁷ Liver injury has very rarely been reported with meropenem. A recent review on the categorisation of drugs with potential hepatotoxicity classified meropenem as category D (low likelihood of hepatotoxicity), with only two published cases of liver injury associated with meropenem and no documented drug re-challenge.⁸

Liver injury associated with meropenem has been reported in a 3-year-old patient, with elevated AST and ALT noted 2 weeks after commencement of meropenem and normalisation of liver enzymes 2 weeks after cessation of meropenem.⁹ Cholestasis was reported in a 70-year-old patient treated with meropenem, with ALP and GGT elevated at day 7 of treatment, with normal AST and ALT.¹⁰ One case of vanishing bile duct syndrome associated with meropenem use has been reported, with a mixed hepatocellular and cholestatic liver injury.¹¹

Interestingly, our patient developed liver injury following multiple different beta-lactam antibiotics. Immunological cross-reactivity has been documented between penicillins and cephalosporins, and cross-reactivity is likely the result of similar antigenic R1 side chains in the drug molecule,¹² however, an R1 side chain is not shared between these three drugs. The mechanism by which our patient developed liver injury to both meropenem, ceftriaxone and cefazolin is unclear. All three drugs share the beta-lactam ring, and therefore, the beta-lactam ring or a component of its degradation may be the antigenic moiety in our patient. In a study of 98 patients with documented immediate IgE-mediated responses to cephalosporins and a positive skin test to cephalosporins, only 1 patient (1%) had a positive skin test to meropenem.¹³ That patient also had positive reactions to all other tested beta-lactams, and it was postulated that IgE antibodies were directed against the beta-lactam ring.

The mechanism of iDILI involves activation of both the innate and adaptive immune system, leading to hepatocyte death via apoptosis and necrosis.¹⁴ Notably, our patient exhibited greater peak levels of liver enzymes after the second exposure to meropenem, consistent with an adaptive immune response.

Learning points.

• Meropenem can cause severe liver injury.

• Patients who develop liver injury following one beta-lactam antibiotic may develop liver injury following a different beta-lactam drug.

• Early identification recognition of drug-induced liver injury is important, and prompt cessation of the offending drug is recommended.