Prednisolone: role in amoxicillin–clavulanate-induced cholestatic liver injury

Abstract

A 68-year-old patient presented with symptoms of a urinary tract infection. A deterioration in the patient’s liver function tests (LFTs) was noted 1 week following completion of a course of amoxicillin–clavunalate. This progressively worsened, reaching its peak by day 30. Our investigations excluded other possible causes for deranged LFTs and there was no improvement of same despite reduced dosing of potentially hepatotoxic medications.

A trial of 30 mg/day prednisolone was commenced, resulting in an immediate and progressive improvement in LFTs to baseline over a period of 22 days and an improvement in constitutional symptoms such as tiredness and poor appetite. Drug-induced liver injury (DILI) is one of the common causes of acute hepatitis and a leading cause of acute liver failure in the US and Europe. Patterns of DILI can be generally divided into: (1) hepatocellular injury, (2) cholestatic injury and (3) mixed injury.

Background

Hepatotoxicity is one of the most common side effects of antibiotics. Antimicrobials, including amoxicillin clavulanate, are implicated in approximately 45% of drug-induced liver injury (DILI) cases identified.¹ According to recent data from LiverTox, hepatotoxicity occurs in approximately 1 in 2500 patients treated with amoxicillin–clavulanate.¹ The typical onset of amoxicillin–clavulanate-induced liver injury is usually within a few days to as long as 8 weeks from the start of therapy and commonly occurs following completion of treatment.² Currently, there is no effective initial treatment for DILI apart from withdrawal of the causative medication and there are limited studies on the potential benefit of corticosteroids. Our case report showed a significant improvement in the patient’s liver function after commencement of prednisolone for the management of severe cholestatic liver injury induced by amoxicillin–clavulanate.

Case presentation

A 68-year-old Caucasian woman was admitted with symptoms of a urinary tract infection (UTI) and received a 5-day course of amoxicillin–clavunalate. Baseline liver function tests (LFTs) were elevated due to a history of chronic liver disease secondary to long-term alcohol use. Eleven days following completion of antibiotic therapy, a slight elevation of alkaline phosphatase (ALP) and gamma-glutamyl transferase (GGT) was noted; however, the patient was asymptomatic. Repeat LFTs 1 week later showed a further derangement of LFTs as follows: ALP 855 IU/L (50—135 IU/L); alanine aminotransferase (ALT) 351 IU/L (7–40 IU/L); GGT 757 IU/L (9–48 IU/L); total bilirubin 44.4 umol/L (4–19 umol/L) and serum albumin 19.8 g/L (35–55 g/L).

The patient reported fatigue and decreased appetite; however, denied pruritus, steatorrhea or dark urine. Physical examination revealed scleral icterus, jaundice, bilateral pitting lower limb oedema, abdominal distension, mild epigastric and right hypochondrium tenderness on deep palpation.

Her previous medical history was significant for stage IV nonsmall cell lung carcinoma for which she underwent multiple courses of chemotherapy in 2018. A recent Positron Emission Tomography (PET) scan performed in February 2019 had shown recurrent lung carcinoma resulting in an upper lobectomy. Other notable medical history included chronic hypoalbuminemia, atrial fibrillation, type 2 diabetes mellitus, recurrent UTI and chronic pancreatitis. A benign mucinous cyst adenoma was diagnosed in 2012 for which she underwent Whipple’s procedure with removal of the pancreatic head, gall bladder, duodenum and parts of the bile duct. She was an ex-smoker with a 20-year pack history and had a history of previous heavy alcohol use before 2018. No relevant family history was noted.

On day 28, LFTs worsened further: ALP 989 IU/L, GGT 1116 IU/L, ALT 136 IU/L, total bilirubin 52.6 umol/L and serum albumin of 19.8 g/L. The pattern of liver injury was identified by calculating the R factor (patient’s ALT/upper limit of normal ALT)/(patient’s ALP/upper limit of normal ALP). In this case, she had a value of 0.14, which indicated cholestatic liver injury. (R>5 hepatocellular injury, R<2 cholestatic injury and R 2–5 mixed liver injury).³ As other potential diagnoses were ruled out by investigation, the conclusion of DILI secondary to amoxicillin-clavulanate was reached.

Investigations

A series of diagnostic tests were carried out to identify the aetiology of deranged LFTs. Abdominal ultrasound revealed severe hepatic steatosis and moderate subphrenic ascites bilaterally without intrahepatic or extrahepatic biliary ductal dilation or splenomegaly. CT of thorax, abdomen and pelvis (CT-TAP) was performed to investigate possible liver metastasis on the background of her history of lung carcinoma. It showed new focal sigmoid circumferential thickening, descending and transverse colitis; however, there were no interval lung changes or liver lesions noted in comparison to a previous CT-TAP performed in April 2020.

Viral hepatitis screen including hepatitis B, hepatitis C, hepatitis E, cytomegalovirus and Epstein-Barr virus was negative. An autoimmune liver screen including antinuclear antibody, antimitochondrial antibody, antismooth muscle antibody, antiliver–kidney microsomal antibodies and antineutrophil cytoplasm antibody) was negative. Alpha-1 antitrypsin and alpha-fetoprotein were normal. Faecal calprotectin was mildly elevated at 90 µg/g (normal range 0–40 µg/g); however, rectosigmoidoscopy and biopsy showed no inflammation or evidence of metaplasia or dysplasia. Stool alpha-1-antitrypsin was normal. Immunoglobulin levels were all normal with the exception of IgA, which was 6.21 g/L (normal range 0.7–4 g/L). Cerulosplasmin was mildly reduced at 0.13 g/L (normal range 0.16–0.45 g/L).

Differential diagnosis

Several diagnoses were considered based on her presentation and medical history. These included alcoholic hepatitis, liver metastases, viral hepatitis, autoimmune hepatitis, primary sclerosing cholangitis, primary biliary cholangitis, alpha-1 antitrypsin deficiency, choledocholithiasis and DILI secondary to amoxicillin–clavulanate. CT-TAP results and negative viral hepatitis screen excluded liver metastasis and viral hepatitis, respectively. Primary sclerosing cholangitis was considered due to evidence of descending and transverse colitis on CT, however, liver autoantibodies were normal, hence reducing the likelihood. Absence of ductal dilation on abdominal ultrasound excluded choledocholithiasis. We also considered cholangiocarcinoma as a differential diagnosis as several case reports had highlighted a correlation between lung adenocarcinoma and primary cholangiocarcinoma.⁴ She was not keen for further investigations with liver biopsy, MRI or magnetic resonance cholangiopancreatography (MRCP), therefore, these avenues were not explored. Additionally, as her LFTs improved following commencement of a trial of corticosteroids, the possibility of cholangiocarcinoma was deemed unlikely.

After excluding other potential causes of liver injury, the diagnosis of DILI was made. This was based on the fact that amoxicillin–clavulanate is well known for its hepatotoxicity and typically causes a cholestatic pattern of liver injury.² Second, the timeline of onset of amoxicillin–clavulanate liver injury is typically within a few days to 8 weeks (average of 3 weeks) following initiation of treatment with the antibiotic and frequently occurs following completion of the antibiotic course.² This is consistent with the onset of LFT deterioration seen in our patient’s case.

Treatment

The patient was asymptomatic in the first 2 weeks following completion of amoxicillin–clavulanate. Her LFTs continued to worsen despite reduction of potentially hepatotoxic medications on her prescription. Lactulose was commenced to prevent hepatic encephalopathy. By day 28, there was a moderate improvement in ALT, however, ALP and GGT levels remained significantly high at 989 IU/L and 1116 IU/L, respectively. A trial of corticosteroids was introduced with an initial dose of 30 mg of prednisolone per day from day 30. By day 32, the first improvement LFTs was noted as follows: ALP 775 IU/L, GGT 1003 IU/L, ALT 74 IU/L and total bilirubin 28. The dose of prednisolone was slowly tapered to 20 mg/day on day 45 and 15 mg/day on day 52. LFTs continued to improve and by day 52, LFTs were as follows: ALP 172, ALT 25, GGT 287, total bilirubin 11.5.

Outcome and follow-up

The patient’s LFTs continued to show significant improvement following the introduction of prednisolone. On day 33, prednisolone was held, which resulted in a rebound worsening of LFTs. Prednisolone was restarted and the baseline LFTs were attained by day 52. Her appetite, fatigue and jaundice also improved significantly. Figure 1 demonstrates LFT trend throughout admission. The patient was maintained on 10 mg of prednisolone per day and her LFTs remained at baseline. Her blood sugar level was also tightly monitored by our diabetes nurse specialist due to her history of type 2 diabetes mellitus. She was discharged on 10 mg of prednisolone/day for 2 weeks and planned for an outpatient review within 2 weeks. Unfortunately, shortly after discharge, the patient developed a UTI and passed away from complications of urinary sepsis.

Figure 1.

Patient’s liver function test (LFTs) results. ALP, alkaline phosphate; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase. Hold/reduce hepatotoxic medication on day 19. Prednisolone was started on day 30. Prednisolone was on hold on day 33. Prednisolone was restarted on day 36.

Discussion

Amoxicillin–clavulanate has been implicated in many cases of DILI across Europe and the USA.⁵ The mechanism of hepatotoxicity is not well known; however, it is likely to be an immunoallergic reaction. Genetic studies have shown that the human leucocyte antigen class II DRB1*1501 allele is associated with a higher risk of amoxicillin–clavulanate-induced liver injury.⁶ Additional risk factors identified include Caucasian race, female sex, age >65 and repeated treatment with the antibiotic and regular alcohol use.^{7 8}

According to the clinical practice guideline from EASL (European Association for the Study of the Liver), the most crucial initial step in the management of in DILI is the withdrawal of the offending medication.⁸ In most cases, there is a total or near total resolution of DILI within days to weeks without the need for any specific measure. In our case, however, we saw a continued worsening of the patient’s LFTs. Studies on the benefit of corticosteroid therapy in the treatment of DILI injury are scarce. Herrero-Herrero et al reported the case of a patient who developed cholestatic liver following a 7-day course of amoxicillin–clavulanate for the treatment of pneumonia. He commenced prednisolone 1 mg/kg/day, resulting in both a significant improvement of his clinical condition and LFTs within 5 days. Baseline LFTs were attained after a 10-week course. Although our patient commenced a lower dose of corticosteroids, the clinical outcome is similar in both cases.⁹

Risk of long-term corticosteroid use and overall outcomes of DILI had been compared in table 1. Clinicians should weigh the risks and benefits prior to commencement of corticosteroid in patients with DILI. In our case, the patient’s LFTs responded well to treatment with prednisolone with progressive improvement and achievement of baseline levels by day 52. As previously mentioned, discontinuation of prednisolone on day 33 resulted in a rebound worsening of LFTs, which improved following recommencement. This clearly demonstrates in our case, the benefits of introducing prednisolone in order to prevent further deterioration of liver function in the patient. Not long after discharge, our patient suffered another UTI and was hospitalised due to urinary sepsis, subsequently passing away.

Table 1.

Risk of long-term corticosteroid use (>1 month) vs overall outcomes of DILI

Risk of long-term corticosteroid use11		Overall outcomes of DILI12
Hypertension	>30%	Complete recovery within 6 months	58.7%
Bone fracture	21%–30%	Mortality	9.3%
Cataract	1%–3%	Liver transplantation	1.3%
Nausea/vomiting/gastrointestinal conditions	1%–5%	Chronic liver disease (greater than 6 months)	30.7%
Weight gain/hyperglycaemic/type 2 diabetes mellitus	4-folds the risk vs control
Infection	37.99 per 1000 person-years13

DILI, drug-induced liver injury.

Although we cannot exclude the potential influence of corticosteroid use in the development of sepsis in our patient, it serves as an important adverse effect to consider the use of steroids in DILI. However, with a significant mortality of 9.3% from DILI, the benefit of starting steroid outweighs its complications.

To date, studies on the benefits of corticosteroid in DILI are scare, with the majority of data describing positive benefits in the form of case reports.^{9 10} Further studies are necessary in order to clarify the benefit of corticosteroids in the treatment of DILI, to identify patients who might benefit and to determine a standardised administration protocol.

Patient’s perspective.

I started experiencing increased tiredness and reduced mobility few weeks after my admission. I was not sure what was going on in my body. I was worried due to my previous history of lung cancer and the chance of spreading. I have told my scan shown no cancer in my liver, but my liver function was not good. I was told that it was likely to be caused by the antibiotic that I was given on admission. At one stage, I felt so tired that I started to bound to bed and my mobility had progressively worsened. After starting steroid, I was told my blood result and liver were improving. Now, (2 August 2020), I felt a lot better and my appetite started to return. I am planning to start my physiotherapy next week to help myself walk again.

Learning points.

• The potential benefit of corticosteroids in the management of cholestatic drug-induced liver injury (DILI) secondary to amoxicillin–clavulanate and its potential use in the treatment of other DILI.

• Considering the possibility of DILI when commencing amoxicillin–clavulanate, especially in patients >65 years of age, history of chronic alcohol use and repeated treatment with antibiotics.

• Monitoring patients’ liver function tests during and after the course of amoxicillin–clavulanate as the majority of patients tend to be asymptomatic initially.

• Preventative measures had been taken for the above patient by adding amoxicillin–clavulanate to the ‘allergic/adverse reaction section’ of the in-patient electronic record, and additionally, her general practitioner was informed on discharge. The patient was also advised to inform other healthcare professionals such as dentists or any specialists regarding sensitivity to amoxillicin–clavulanate to prevent re-exposure.

• Due to significant risk of liver injury (1 in 2500),¹ the decision of starting amoxicillin–clavulanate should be reconsidered when treating urinary tract infections (UTIs). National Institute for Health and Care Excellence (NICE) guideline recommends nitrofurantoin/trimethoprim as first-line treatment for lower UTI in men and nonpregnant women aged 16 years and over.¹⁴