Liver transplantation for acute liver failure due to efavirenz hepatotoxicity: the importance of routine monitoring

Case report

The risk of non-nucleoside reverse transcriptase inhibitor (NNRTI) hepatotoxicity is well recognised. It affects between 1 and 8% patients receiving efavirenz (EFV) but symptomatic disease is rare.^1–4 Prospective data suggest that in sub-Saharan African populations high plasma EFV levels and increased incidence of EFV hepatotoxicity have been associated with liver enzyme variant allele, CYP2B6*6, which has a major role in EFV metabolism.^5,6 The highest frequency of this allele has been identified in Zimbabwean populations. ⁷ EFV hepatotoxicity is most commonly attributed to hypersensitivity and onset is typically within the first eight weeks of therapy. ⁸ There is only one previous published case of EFV-induced adult acute liver failure (ALF). ⁹ In contrast, cases of ALF as a result of NVP therapy have changed post-exposure prophylaxis prescribing. ¹⁰ Cessation of antiretroviral therapy (ART) and safe resolution of liver inflammation depend on regular monitoring of liver function tests (LFTs) in patients initiating ART to detect subclinical hepatotoxicity. ¹¹ Adequate monitoring relies on patients engaging with their care provider. We present the first case report in the UK of ALF caused by EFV therapy culminating in liver transplantation where poor engagement contributed to late presentation of hepatotoxicity.

The patient is a 26-year-old Zimbabwean woman who was diagnosed with HIV infection in the UK on routine screening in 2003 during the 32nd week of pregnancy. She received ART and delivered a seronegative baby. Her ART was stopped after delivery in November 2003 as her CD4 count was greater than 350 cells per mm³. In March 2004, her CD4 count was 440 cells per mm³ and her HIV-1 viral load (VL) 15,000 copies per ml. At this time, her blood tests were negative for hepatitis B virus surface antigen (HBsAg) and hepatitis C virus antibody (HCVAb); and positive for HBV core antibody (HBcAb) and HBV surface antibody (HBsAb). In September 2011, her CD4 count was 289 cells per mm³ and her HIV-1 VL 1000 copies per ml. She commenced emtricitabine, tenofovir and efavirenz (Atripla) therapy at the end of December 2011. Prior to starting ART and 20 days after initiation her routine LFTs were normal. The patient did not attend her four-week genitourinary medicine follow-up appointment in February. She had changed her contact number and next attended in late March. At this time she was issued with a four-month prescription for Atripla and blood tests were taken. The biochemistry samples haemolysed, however, and were not processed. Her CD4 count was 439 cells per mm³. She missed another appointment and repeat blood tests in May. The patient was reached by phone on at least two occasions during May: she reported good adherence and felt well.

In July, nearly seven months after initiating ART, she attended her general practitioner (GP) giving a two-week history of lethargy, nausea and dark urine. She was empirically prescribed trimethoprim. She returned to her GP two days later after her aunt had noted that her eyes were becoming yellow. On examination, she was jaundiced with a tender liver edge. She was not encephalopathic. There was no examination evidence of a hypersensitivity reaction. She denied using any complementary medicines, alcohol or recreational drugs. She had been prescribed no other hepatotoxic medications. She was admitted to her local hospital with severe acute hepatitis. On admission, her aspartate transaminase (AST) was 1118 U/L, bilirubin 118 U/L and international normalized ratio (INR) 1.6. Her Atripla was stopped. An ultrasound scan of her liver was reported as normal. She was discharged after eight days with improving but abnormal LFTs. Outpatient LFTs nine days later showed worsening liver function: AST 581 U/L, bilirubin 393 U/L and INR 3.1. She was recalled to hospital.

The patient was transferred to the hepatology unit at Royal Free Hospital, London. Transjugular liver biopsy reported severe acute hepatitis with bridging and panacinar necrosis. Repeated blood tests investigating for autoimmune, infective and metabolic causes of her liver disease were unremarkable. Reactivation of hepatitis B virus infection was excluded by testing for hepatitis B virus DNA, which was undetectable. Hepatitis E virus serology was also negative. On 2 August her HIV-1 VL was undetectable and her CD4 counts was 819 cells per mm³ and not repeated until after she had restarted on ART. She received supportive therapy. Her LFTs continued to deteriorate and on 8 August she became frankly encephalopathic. She was admitted to the intensive care unit (ICU) and placed on the pathway for super-urgent liver transplantation. She became increasingly agitated and was intubated. She received an orthotopic liver transplantation (OLT) on 11 August. She began her recovery on ICU and re-started ART, tenofovir/emtricitabine and raltegravir, on 15 August. She was discharged from hospital on 24 August and continues to do well. Her CD4 count was 452 cells per mm³ and her VL was undetectable with normal liver function.

In this case, delayed onset of toxicity, isolated liver disease without rash or eosinophilia, non-specific histology findings and the absence of an alternative aetiology meant that EFV hepatotoxicity due to an idiosyncratic drug reaction or direct toxicity were considered the most likely explanations for the ALF. The contribution of liver enzyme variants to this particular presentation and her CYP2B6 genotype are unknown. However, given the high incidence of CYP2B6*6 in Zimbabwean populations it may prove instructive to explore this further. Irregular attendance and loss to follow up are independently associated with increased mortality in patients with HIV infection. ¹² In the UK, factors predicting loss to follow up include: female gender, non-white ethnicity, age 15 to 34 years, absence of AIDS-defining illnesses, and non-prescription of ART.^13,14 The patient was high risk for poor engagement with care. Efforts were made to address this by regular telephone contact. Limited evidence does suggest that telephone communication improves the likelihood of attending clinical appointments. ¹⁵ There were no language barriers to communication but the patient does have a young child and a restricted support network. She later reported that difficulties providing care for her child and the cost of transport were major barriers to her attending regular appointments. This case reminds us that ART can cause life-threatening injury if not monitored. Further, it shows that engagement with care is fundamental to patient safety. We are also reminded that a clinical emphasis on ART adherence should not overlook routine but essential biochemical monitoring of the medication.