Rapid bupropion-induced hepatotoxicity: a case report and review of the literature

Medications are a common cause of liver injury because the liver is the main site of drug clearance, biotransformation, and elimination. Bupropion is a commonly used medication to aid with smoking cessation and was also widely used in the past as an antidepressant. The most common adverse effects are dry mouth, headache, insomnia, and agitation. Drug-induced hepatotoxicity, however, is a very rare adverse event, with only seven cases reported in the world literature. Having reviewed these cases, we note that the presentation, mortality, and treatment varied in each of these cases, creating both diagnostic and management dilemmas. We report a patient who experienced significant derangement of her liver tests with a clinical vignette highly suggestive of bupropion-induced hepatotoxicity. We believe further additions to the already existing small literature base on this topic will help practicing physicians when dealing with the possibility of bupropion-induced hepatotoxicity.

A 50-year-old Caucasian woman with a 6-month history of Crohn’s disease and receiving methotrexate for this disease presented with deranged liver function tests to our gastroenterology clinic. She had recently been discharged from the hospital 1 week earlier, following a flare of her Crohn’s disease. On discharge, she was sent home with nasogastric feeding to help with her malnutrition, and as part of recognizing any refeeding syndrome, she was having regular blood tests in the community. Her routine blood tests on 19 December 2016 showed marked derangement in her transaminases, with aspartate transaminase (AST) of 787 U/L and alanine transaminase (ALT) of 1032 U/L. Her bilirubin and alkaline phosphatase (ALP) were normal, and her γ-glutamyl transferase (GGT) was only mildly raised at 51 U/L. Results of her liver tests done 3 days prior, on 16 December, following discharge from the hospital were completely normal. Of note, she was started on 150 mg of bupropion on 13 December, which was increased to 150 mg twice daily 3 days later, on 16 December, to assist with smoking cessation. Results of her remaining blood tests, apart from long-standing stable normocytic anemia, were unremarkable.

The patient had a history of osteoporosis, palpitations, and depression. Her regular medications for these diagnoses included methotrexate 15 mg weekly for Crohn’s colitis, folic acid 5 mg weekly, cholecalciferol, sertraline, melatonin, propranolol, conjugated estrogen (Premarin; Wyeth Pharmaceuticals, Philadelphia, PA, USA) as hormone replacement therapy for menopause, pantoprazole, oxycodone, Coloxyl (Aspen Australia, St. Leonard’s, Australia), and a 7-week weaning course of prednisone (from 35 mg daily) for her recent Crohn’s flare with co-trimoxazole cover to continue until weaned off prednisone.

The patient’s blood tests on 20 December, done following the initial derangement on 19 December, were even more markedly deranged, with AST of 4006 U/L and ALT of 5007 U/L. Her GGT was also mildly more raised at 73 U/L. However, her bilirubin and ALP remained normal. She was reviewed on 22 December in the gastroenterology clinic. She had weaned to 30 mg of prednisone at that point. There had been no new medications other than the bupropion added since her discharge from the hospital. On examination, she did not have any jaundice, bruising, or any signs of chronic liver disease. Because she was feeling subjectively well with a normal bilirubin level, she was managed in the community setting. Her bupropion had already been stopped on 20 December by her general practitioner. All her usual medications, including prednisone and co-trimoxazole, were continued. At her 22 December clinical review, apart from an increase in GGT to 138 U/L, the remaining markers showed significant improvement with the AST decreased to 263 U/L and ALT reduced to 2643 U/L. Hence, a liver biopsy was not pursued, and instead she was followed with further surveillance liver tests, which continued to show improvement with subsequent near-normalization almost 4 weeks later, on 12 January, with ALT of 39 U/L and GGT of 94 U/L. Her AST had already normalized by 27 December. The trends of her liver tests are shown in Fig. 1.

Bupropion is an antidepressant that is also used as a nonnicotine method to aid in smoking cessation, and it acts by inhibiting the reuptake of dopamine and noradrenaline (norepinephrine) [1]. Preclinical animal toxicology studies have demonstrated reversible hepatotoxicity, hepatocellular hypertrophy, and focal hepatic hyperplasia related to hepatic enzyme induction [2]. Bupropion-induced hepatotoxicity is quoted to affect between 0.1% and 1% of treated patients with a hepatocellular and/or a cholestatic pattern of damage [3]. The mechanism of damage is considered to be predominantly immune-mediated with the presence of a hypersensitivity syndrome (fever, rash, eosinophilia, autoantibodies) and a short latency period (1–6 weeks) [3]. Looking through the literature, we found seven reported cases of hepatotoxicity considered to be induced by bupropion and one further case of hepatotoxicity due to conjoint therapy with bupropion and doxycycline. The features of these cases are summarized in Table 1.

After the Alonso Rodríguez et al. [4] case, our patient’s case involved the next shortest time frame (6 days) from initiation of bupropion therapy to the onset of liver enzyme derangement. Given the very small number of cases, owing to the idiosyncratic nature of the drug reaction, there appears to be a large variation in the time frames for the onset of liver derangement after the start of treatment. It is possible that there was an add-on effect from the 5 months of methotrexate use because chronic low-dose methotrexate use has been associated with hepatotoxicity on a histological level, even despite unremarkable liver function tests [5]. However, this is usually expected only with cumulative doses of 1.5 g and above [5], which our patient had not reached (approximately 300 mg in total) by the time of onset of her liver derangement. She was also receiving long-term sertraline and propranolol, which can cause mild liver test abnormalities, including elevations in serum aminotransferase levels in approximately 1–2% of cases [6]; however, the fact that she improved despite continuation of these and her completely normal liver function tests prior to initiation of bupropion make these being partly responsible less likely. The other medication of possibility was the co-trimoxazole started during her hospital admission approximately 2 weeks prior to the liver derangement. Again, the co-trimoxazole cover was continued throughout the subsequent improvement of her liver function tests, which makes it an unlikely cause. The damage with co-trimoxazole is typically cholestatic or mixed with jaundice rather than being a solely hepatocellular picture of liver derangement, as was the case in our patient [7].

Our patient had a good recovery following the initial insult without requiring inpatient admission and with no decompensation of her liver. Using “Hy’s rule” [7], the lack of bilirubinemia did suggest a good prognosis in this case. Our patient subsequently had rapid resolution of her liver function tests, which showed significant improvement within 3–5 days of cessation of bupropion therapy. It is difficult to tell whether the ongoing moderate-dose prednisone therapy for her recent Crohn’s flare had helped in her recovery; however, she had been receiving 30 mg daily of prednisone in the week prior to the derangement of her liver tests and 35 mg daily 2 weeks prior. An antinuclear antibody (ANA) measurement would have been useful to know if there were autoimmune features of her liver derangement and hence the utility of steroid therapy, but this was not carried out; however, there is a possibility of a false-negative result in this case, given the concomitant prednisone therapy for 2 weeks prior to the insult. A liver biopsy was also not carried out in our patient, owing to her rapid improvement after cessation of bupropion. Previous cases in the literature published in 2000 [8], 2001 [9], and 2007 [10] demonstrated liver biopsies in bupropion-induced hepatotoxicity to involve eosinophilic inflammatory infiltration of the portal tracts along with neutrophils and lymphocytes. However, in the case reported by Hu et al. [8], there was complete resolution without the use of steroids, but the ANA was negative. In contrast, Alvaro et al. [9] and Humayun et al. [10] did use steroid treatment, and both cases also had a positive ANA at onset. Interestingly, in the case of Humayun et al. [10], there was relapse following withdrawal of steroids after 6 weeks of treatment, with subsequent death occurring despite reinitiation of steroids.

In summary, given the rapid onset following initiation of bupropion and fast resolution following its cessation, we strongly believe our patient’s case was one of bupropion-induced hepatotoxicity. Bupropion is an effective and safe treatment option for smoking cessation. However, as with many other medications in clinical practice, it does pose a risk of very rare idiosyncratic reactions of hepatotoxicity. The presentation and the clinical outcome following such insult to the liver is varied with bupropion, and management, apart from immediate cessation of the drug, remains unclear, given the very small number of cases in the literature. Steroid treatment is not clearly proven, although there may be a role for steroids in cases of positive of ANA and/or SMA. Owing to the potential for rare but fatal outcomes of bupropion hepatotoxicity, practicing physicians should be aware of this adverse reaction when assessing patients receiving bupropion therapy.