Drug-induced autoimmune-like hepatitis

Download video file ^{(19.1MB, mp4)}

INTRODUCTION

The term “drug-induced autoimmune-like hepatitis” (DI-ALH) refers to an increasingly recognized phenotype of DILI associated with a variety of drugs and herbal or dietary supplements, small molecules, biological agents, and vaccines (Table 1) that share several features and, in some patients, exhibits a histological pattern of injury that closely resembles idiopathic autoimmune hepatitis (AIH). However, this phenotypic presentation of DILI is rare compared to other DILI presentations, ⁹ and its frequency depends on the definition used. In a recent study analyzing cases collected in the Spanish and LatinDILI registries over the years, DI-ALH, which was defined by the presence of at least 2 out of 3 features (detectable titers of autoantibodies, elevated IgG, and compatible histology), represented 2.3% of all DILI cases. ¹ Among patients with AIH from the Mayo Clinic database, the 9.2% frequency of DI-AIH reported was calculated on a retrospective basis. ² However, the true incidence of DI-ALH is probably underestimated because a significant proportion of cases are misclassified as idiopathic autoimmune hepatitis. Drugs classically associated with DI-ALH are α-methyldopa, nitrofurantoin, minocycline, and more recently, infliximab and statins. ¹

TABLE 1.

Agents most frequently implicated in cases of drug-induced autoimmune-like hepatitis (summary table of the main papers published up to date)

Agents	Patients n	Age (y) mean (range)	Females %	Time to onset months patients (%)	Jaundice %	Pattern of liver injury, % Hep	Hypersensivity features %	Positive autoantibodies %	High IgG %	Typical or suggestive AIH histology %	IMS treatment %	Relapse %	Follow-up (mo) Mean (range)	References
Nitrofurantoin	61	67 (36–91)	100	< 1 m: 5 (12.5) 1–6 m: 8 (16) 6–12 m: 5 (10) > 12 m: 32 (64)	66	69	11 (peripheral eosinophilia, 2%)	ANA: 75 ASMA: 60	46	45	84	5	33 (3–60)	Garcia-Cortes et al. 1 Björnsson et al. 2 De Boer et al. 3 Björnsson et al. 4
Infliximab	58	45 (28–69)	81	<1 m: 2 (4) 1–6 m: 43 (93.5) 6–12 m: 0 > 12 m: 1 (2)	15	64	NA	ANA: 79 ASMA: 2	28	33	62	0	113 (6–182)	De Boer et al. 3 Rodrigues et al. 5 Ghabril et al. 6 Björnsson et al. 7
Minocycline	43	20 (15–61)	77	< 1 m: 1 (3) 1–6 m: 10 (31) 6–12 m: 7 (22) > 12 m: 14 (44)	44	82	11 (peripheral eosinophilia, 7%)	ANA: 79 ASMA: 41	59	73	93	20	62 (4.5–211)	Garcia-Cortes et al. 1 Björnsson et al. 2 De Boer et al. 3
Statins a	32	59 (35-85)	62.5	<1 m: 0 1–6 m: 16 (50) 6–12 m: 4 (12.5) > 12 m: 11 (34)	NA	NA	NA	ANA: 75 ASMA: 44	46	71	62	31	59 (2–216)	Garcia-Cortes et al. 1 Björnsson et al. 8
Adalimumab	13	41(33–54)	92	<1 m: 0 1–6 m: 7 (64) 6–12 m: 3 (27) > 12 m: (18)	0	100	NA	ANA: 77 ASMA: NA	46	70	69	9	33 (5–96)	Rodrigues et al. 5 Ghabril et al. 6 Björnsson et al. 8
Khat	11	35(24–52)	0	NA	NA	NA	NA	ANA: 27 ASMA: 44	91	91	100	0	7.5 (1–28)	Björnsson et al. 8
Methyldopa	10	29 (18–43)	100	< 1 m: 1 (9) 1–6 m: 9 (82) 6–12 m: 1 (9) > 12 m: 0	82	100	27 (Peripheral eosinophilia, 17%)	ANA: 40 ASMA: 40	50	NA	27	20	NA	De Boer et al. 3
Tinospora cordifolia	8	44 (36–68)	87	NA	NA	NA	NA	ANA: 62 ASMA: 12	71	100	62	0	7 (3–12)	Björnsson et al. 8
Imatinib	8	43 (17–61)	75	NA	NA	NA	NA	ANA: 62 ASMA: 12	43	100	87.5	0	21 (3–60)	Björnsson et al. 8
Hydralazine	7	60 (42–76)	71	< 1 m: 3 (43) 1–6 m: 3 (43) 6–12 m: 1 (14) > 12 m: 0	43	43	14	ANA: 50 ASMA: 25	25	NA	57	28.5	NA	De Boer et al. 3
Etanercept	4	23.5 (9–50)	100	< 1 m: 1 (25) 1–6 m: 2 (50) 6–12 m: 1 (25) > 12 m: 0	50	100	NA	ANA: 100 ASMA: 50	100	100	100	0	15.5 (3–28)	Ghabril et al. 6 Björnsson et al. 8

Note: All percentages have been calculated according to available information.

^aStatin: 7 fluvastatin, 13 atorvastatin, 5 rosuvastatin, 6 simvastatin and 1 pravastatin.

Abbreviations: AIH, autoimmune hepatitis; ANA, antinuclear antibodies; ASMA, anti-smooth muscle antibodies; Hep, hepatocellular; IMS, immunosuppressive; m, months; n, number; NA, not available.

HOW TO DISTINGUISH DRUG-INDUCED AUTOIMMUNE-LIKE HEPATITIS FROM IDIOPATHIC AUTOIMMUNE HEPATITIS?

Distinguishing DI-ALH from idiopathic AIH is important because DI-ALH may not require immunosuppression or if steroid therapy is initiated, it may be discontinued without recurrence of injury. However, DI-ALH remains a poorly defined disorder, and even its existence is questioned by some experts in the field of AIH, who believe that the relationship between drugs and AIH is “casual” rather than “causal.” Distinguishing DILI with autoimmune characteristics from AIH is hampered by the lack of pathognomonic features and the absence of specific diagnostic criteria. ¹⁰ Although recent genetic studies in patients with DILI and the identification of new laboratory biomarkers (ie, polyreactive IgG, and IgM autoantibody signatures) ¹¹ and alanine aminotransferase kinetics in response to corticosteroids ¹² have been proposed to help distinguish DI-AIH from AIH, the data are pending of further validation and the diagnosis of both entities still relies on the exclusion of competing causes. ¹³ Neither the RUCAM clinical method nor the more recent causality assessment tool RECAM (an electronic updated version of RUCAM), ¹⁴ nor the scoring system for the diagnosis of autoimmune hepatitis ¹⁵ were designed for this specific phenotype.

Clinically, the 2 entities may be indistinguishable. Although AIH causes chronic (nonresolving) inflammation in the liver, it may begin with an acute onset, ¹⁰ similar to DILI with autoimmune features. While DI-ALH mimics the morphological pattern of AIH and the microscopic findings that might help distinguish between the 2 conditions are largely unknown, advanced fibrosis (ie, cirrhosis) appears to be seen only in AIH but not in DI-ALH, so this feature would further suggest “idiopathic” AIH. However, cases of cirrhosis at the time of diagnosis of DI-ALH have been reported in a context where the histological picture was attributable to the drug (ie, 1 subject on intermittent treatment with nitrofurantoin and another on long-term therapy with cyproterone acetate). ¹ Subclinical liver damage would probably allow continued treatment with the offending drug, eventually leading to cirrhosis. Nevertheless, a liver biopsy is often required in particular in cases where liver tests do not improve after cessation of the implicated agent or if a suspected case of DILI has autoimmune features; the findings can be used to further classify and manage the case. ⁹ Importantly, the biopsy should be taken as early as possible, as some degree of fibrosis may develop over time as inflammation persists in DILI, confounding the diagnosis.

A recent international expert conference on DI-ALH, organized by the Drug and Herbal & Dietary Supplement-Induced Liver Injury Consortium and the International Autoimmune Hepatitis Group, with the aim of agreeing on terminology, management strategies, and future directions for research, published its conclusions and recommendations ¹³ and proposed an algorithm for clinicians when approaching cases of liver injury with autoimmune features, a modification of which is shown in Figure 1.

FIGURE 1.

Management algoritm of liver disease with DILI suspicion and autoimmune features. Abbreviations: AIH, autoimmune hepatitis; ANA, antinuclear antibody; anti-LKM1, anti-liver-kidney microsomal type 1 antibody; anti-SLA/LP, anti-soluble liver antigen/liver pancreas antigen; ASMA, anti-smooth muscle; DI-ALH, drug-induced autoimmune-like hepatitis; HDS, herbal and dietary supplements. ¹³

THERAPEUTIC APPROACH TO DI-ALH

Withdrawal of the offending agent is the most important first step in any suspected case of DILI. Evidence for the use of steroids in DI-ALH is largely based on case reports or case series. However, the decision to start steroid therapy should not be a straightforward one, as published cases of DI-ALH have reported high rates of spontaneous recovery after discontinuation, so a chance for spontaneous resolution should be given, especially in milder cases. However, persistent or even worsening liver injury may occur despite the withdrawal of the offending drug. In this context, corticosteroids are justified in cases of hepatocellular or mixed liver injury, although the decision should be individualized. ¹³ While the timing of initiation of immunosuppressive therapy in DI-ALH is unknown, an international collaborative study of 2 prospective DILI registries to assess the safety and potential benefit of steroids in DILI found that cases treated with steroids were enriched for the DI-ALH phenotype. The authors used propensity score matching analysis and found that the increase in the rate of normalization of liver biochemistry was more evident in patients with severe DILI based on nR-Hy’s law who did not resolve biochemical abnormalities within 30 days of stopping the implicated drug. ¹⁶ The optimal dose and duration of steroid therapy in DI-ALH have not yet been stablished. ¹³ However, a short course (1–2 mo) of tapering steroids and then keeping the patient under close observation seems reasonable. Although relapse and chronic injury are less common in DI-ALH, the likelihood of relapse increases over time, reaching 50% after more than 4 years of follow-up in registry studies, especially in patients without eosinophilia and with very high levels of transaminases and total bilirubin at presentation, ¹ contrary to what is commonly reported.^7,8 The results of these studies, however, call for long-term follow-up of patients diagnosed with DIALH.^7,8

In conclusion, a systematic approach to any patient who develops liver injury with autoimmune features and is exposed to drugs or other agents would allow for confident management of DI-ALH. Hopefully, collaborative research to better characterize DILI phenotypes and the use of various omics technologies will ultimately lead to accurate diagnosis and management of this entity.