Discordance between biochemical and histological severity in persistent immune checkpoint inhibitor induced liver injury

Summary

We report a case of a man in his early 70s referred for Grade 2 immune checkpoint inhibitor induced liver injury (ChILI) post nivolumab and ipilimumab for metastatic melanoma. Despite treatment with immunosuppression, the liver enzymes improved but failed to normalise completely. Liver biopsy performed 6 months after diagnosis identified a severe hepatitis with interface and perivenular necroinflammatory activity. Mycophenolate was up-titrated to 1 g twice daily and prednisone 40 mg daily was commenced. Once liver enzymes normalised immunosuppression was then gradually withdrawn, with no recurrent elevation of the liver enzymes. This case demonstrates that there can be an incongruency between liver enzymes an histopathology in ChILI. As in our patient, this can have significant implications in managemen. Furthermore, this case is unique as there are no other reported cases in the literature of chronic ChILI.

Background

The use of immune checkpoint inhibitors (ICI) is increasing, as such immune-related adverse events, including immune checkpoint inhibitor-induced liver injury (ChILI), are seen more frequently in clinical practice. The incidence of ChILI has been reported to range from 0.7% to 16%.¹ The rates are higher for combination therapy and the use of anti-cytotoxic T lymphocyte-associated protein-4 (anti-CTLA-4) drugs.^{1 2} The typical clinical pattern occurs 6–14 weeks post commencement of immunotherapy and can present as cholestatic (37% of cases), hepatocellular (38%) or mixed pattern (25%).³ The grading of ChILI as per the Common Terminology Criteria for Adverse Events (CTCAE), is based solely on liver enzymes, with Grade (G) 3 and 4 deemed to be severe (figure 1).^{4 5} The grading system was generated from oncology clinical trials and expert opinion, and differs from other systems to grade severity of drug-induced liver injury which incorporate markers of liver dysfunction.⁶ First-line treatment, involving cessation of ICI and high doses of steroids, is effective in most cases.

Figure 1. Common Terminology Criteria for Adverse Events (CTCAE). AST, aspartate aminotransferase; ALT, alanine aminotransferase; ULN, upper limit of normal.

We report a case of ChILI which is unique due to both the chronic nature of the hepatitis and the discordance in severity of liver injury between liver enzymes and histopathology.

Case presentation

A man in his early 70s was referred to the hepatology clinic for assessment of raised liver enzymes following treatment for metastatic melanoma stage IV (cT0, cNx and cM1b) diagnosed 2 years earlier.

His treatment course is described in figure 2. He received ipilimumab and nivolumab for four cycles, followed by nivolumab monotherapy.

Figure 2. Alanine aminotransferase levels following immunotherapy for metastatic melanoma. G, grade; IN, ipilimumab and nivolumab; N, nivolumab; WPW, weaned per week.

Treatment was complicated by several immune-mediated adverse events. After his second cycle of ipilimumab/nivolumab, he developed a G3 dermatitis and G1 hepatitis (figure 1), both of which improved following treatment with prednisone. Shortly after transitioning to nivolumab monotherapy, he developed G2 colitis, which again improved with prednisone. He developed a polyarthritis that was treated with ibuprofen 13 months post immunotherapy commencement. Around the same time, the alanine aminotransferase (ALT) began to rise and peaked at 127 U/L, consistent with G2 hepatitis. He was treated with prednisone 1 mg/kg/day with tapering over the following 4 months with improvement; however, not normalisation of the ALT. The ALT was 67 U/L when prednisone was ceased. Repeat blood tests 2 months later demonstrated a rise in the ALT to 109 U/L which increased further over the next 2 months to a peak of 144 U/L.

He was referred to hepatology services and a biopsy was performed at which time the ALT was 125 U/L, alkaline phosphatase 90 U/L, bilirubin 13 µmol/L, international normalised ratio (INR) 1.1. Histopathology showed a severe hepatitis with interface (periportal) and perivenular (zone 3) necroinflammatory activity, consistent with persistent ChILI (figure 3). Focal bridging necrosis was also seen. The trichrome stain shows mild periportal fibrosis, stage 1 of 4 by Laennec.

Figure 3. Liver parenchyma histopathology H&E stain of liver parenchyma. There is severe hepatitis with interface (periportal) and perivenular (zone 3) necroinflammatory activity. Focal bridging necrosis is also present.

Treatment

In line with Clinical Practice Guidelines, he was started on 40 mg prednisone and mycophenolate mofetil (MMF) 500 mg two times per day (BD) with sulfamethoxazole and trimethoprim for pneumocystis jiroveci pneumonia prophylaxis and pantoprazole for gastric protection.⁵ A dose of 500 mg BD was initially started to monitor for side effects such as bone marrow suppression. The ALT improved but plateaued at 62 U/L 2 months after commencing therapy. Given the lack of normalisation of the ALT, MMF was increased to 1 g BD (the maximal recommended dose in autoimmune hepatitis and liver transplant), although the prednisone dose had to be reduced due to side effects of blurred vision (he had a history of central serous retinopathy). Prednisone taper continued by 5 mg per week and the ALT ultimately normalised. The prednisone was ceased 4 months post initiation with the ALT at 32 U/L. MMF was reduced to 500 mg BD 4 months post initiation and ceased 2 weeks later. There was no rebound in ALT after discontinuation.

Outcome and follow-up

Eleven months following the withdrawal of immunosuppression the ALT, bilirubin and INR have remained persistently normal. The blurred vision resolved with cessation of prednisone. Over 1 year post ceasing immunotherapy staging scans have shown no evidence of melanoma recurrence. He has since been on a family holiday to Paris and continues to feel well.

Discussion

Our case is unique in two ways. First, there was a discrepancy in severity of hepatitis between the liver enzymes and histopathology. Based on the liver enzymes at the time of biopsy, he only met the criteria for G2 ChILI, but histopathology showed a severe hepatitis with areas of bridging necrosis. As per the National Cancer Institute’s CTCAE, the grading of ChILI is based solely on liver enzymes, with only G3 and G4 regarded as severe (figure 1).^{4 5} Clinical practice guidelines recommend high dose steroids for the treatment of G3 or G4 ChILI, but clinician discretion can be used regarding the treatment of G2 ChILI.⁵ The CTCAE assumes that ALT will accurately reflect the degree of liver inflammation. However, ALT is an imperfect test to measure liver inflammation which has clearly been documented across a diverse range of primary liver diseases including autoimmune hepatitis, viral hepatitis and metabolic associated steatotic liver disease. Our case highlights the potential for discrepancy between biochemical and histological findings in ChILI. Although the mechanism of liver injury is not well established in ChILI, the fact that this disconnect is seen across a diverse range of liver diseases suggests that this is a general limitation of being over-reliant on ALT in assessment of severity of hepatitis.7,9

Liver biopsy is infrequently performed for ChILI, but can be important in the diagnosis as autoantibodies associated with autoimmune hepatitis are generally not present and immunoglobulin levels are normal.¹⁰ Due to the lack of normalisation of the ALT with prednisone, biopsy was useful in our case to exclude other aetiologies such as drug-induced liver injury from ibuprofen. Usually, a biopsy is not done to determine the severity of the disease. However, a recent study has proposed a novel management algorithm of ChILI, in which severity of liver injury on biopsy is used to guide steroid management in patients whose liver enzymes have not improved after discontinuing ICI.¹¹ Our case challenges the grading method currently employed for ChILI severity and lends support to an approach involving early biopsy to gauge severity and necessity for therapy. This of course, must be weighed against the risks of percutaneous liver biopsy including pain (3%–5%), bleeding (0.8%–1.7%) and vasovagal reactions (1%–3%) as well as timely access to biopsy.^{12 13} The European Association for the Study of the Liver suggests liver biopsy should be considered in cases with severe drug-induced liver injury or refractory hepatotoxicity to standard immunosuppression with corticosteroids, but the decision should be made by a multidisciplinary team.⁴

Second, our patient’s course was unique due to the chronic nature of his hepatitis. This was most likely a consequence of an incomplete response to prednisone. Although most patients with ChILI respond to steroid treatment, primary non-response or relapse on withdrawal of steroids occurs in 20%–30% of patients.¹³ In these cases, additional second or even third line immunosuppression agents are often required. There are no randomised studies; however, MMF is listed in clinical guidelines for second line therapy for steroid resistant ChILI, with an alternative being azathioprine (1–2 mg/kg/day).⁹

In our case, the patient’s initial course of prednisone was tapered and ceased prior to complete normalisation of liver enzymes (ALT 67 U/L). In clinical practice, there may be many reasons why steroids are discontinued before liver enzymes have fully normalised. These may include an assumption that low level ALT elevation is not clinically significant, especially if recommended protocols for steroid tapering have been followed; an assumption that ChILI is an acute phenomenon which will naturally wane over time after drug exposure; or that the residual mild ALT elevation is caused by steroid induced hepatic steatosis.⁵ This case demonstrates that even mild residual ALT elevation, several months after ICI exposure, can reflect significant ongoing liver injury and that unless adequately treated, ChILI can persist. As such, it is important that treatment be continued until the liver enzymes have normalised completely.

Patient’s perspective.

Having experienced several side effects from my dual agent immunotherapy regimen, when I first was advised of my grade 2 hepatitis, I was not surprised but worried that it was impacting my liver given the importance of this organ. After 4 months of treatment with the prednisone, I was delighted to be medication free as I experienced a lot of side effects from the prednisone, the worst being insomnia.

It was therefore very discouraging to learn that my ALT was again worsening as I was symptom-free and feeling good. I had resumed my athletic endeavours, including gym workouts, running and playing hockey. I was further dismayed and shocked to learn that my liver biopsy demonstrated a worse picture than my bloodwork; it led to a lot of questions and discussion with the team. Being an outlier or unique case is not a comforting fact. It made me question if I should have stopped my immunotherapy sooner given there are no clear guidelines for optimal duration. The decision to continue immunotherapy was at all times a joint decision with myself and my team. Although I experienced more side effects, I am so grateful that I responded to both the prednisone and mycophenolate treatment and that my ALT values are currently in the normal range. The liver biopsy did facilitate more prompt treatment in the end. Best of all, my CT scans continue to show no evidence of melanoma recurrence in my head, neck, chest, abdomen and pelvis.

Learning points.

• Severity of checkpoint inhibitor induced liver injury (ChILI) may not be accurately reflected by Common Terminology Criteria for Adverse Events criteria based on liver enzymes, and a liver biopsy may be helpful to more accurately grade severity.

• Although many cases of ChILI may be self-limiting or resolve following withdrawal of immunotherapy and/or short courses of immunosuppression, there is the potential for immune checkpoint inhibitor hepatitis to evolve into a chronic immune-mediated hepatitis.

• All cases of ChILI should be treated until liver enzymes normalise before being considered resolved.

• Hepatology should be consulted for steroid-resistant or refractory ChILI.