Clinicopathologic Features of Adult-onset Still's Disease Complicated by Severe Liver Injury

Abstract

Adult-onset Still's disease (AOSD) is a systemic inflammatory disorder. Severe liver injury has rarely been reported, although liver enzyme elevation is a common complication of AOSD. We herein report four cases of relapsed AOSD with severe liver disorder by tapering or terminating corticosteroids. Liver specimens revealed robust infiltration of inflammatory cells throughout the lobule, especially cluster of differentiation (CD) 8-positive cells. Relapsed AOSD was refractory to corticosteroid reintroduction and required immunosuppressants. Severe liver injury with AOSD is pathologically characterized by extensive lobular infiltration of CD8-positive cells, and we should consider additive immunosuppressive agents on corticosteroids for treatment.

Introduction

Adult-onset Still's disease (AOSD) is an idiopathic systemic inflammatory disease characterized by a spiking fever, arthritis, skin rash, sore throat, lymphadenopathy, and hepatosplenomegaly. AOSD is difficult to diagnose due to the lack of specific clinical manifestations and serum biomarkers. Although several factors, including genetics, infections, and a dysregulated immune system, have been proposed to be involved in the pathogenesis of AOSD, the precise details are not fully understood. Elevated liver enzymes are complicated in 50-80% of patients with AOSD (1,2), and case reports of severe liver injury have been scattered (3-6). However, the pathogenesis and mechanism of liver injury are unknown and yet to be elucidated. The histopathological findings of liver injury complicated with AOSD have been reported to be (1) periportal inflammatory infiltrates, (2) periportal fibrosis, (3) Kupffer cell hyperplasia, and (4) focal hepatitis with submassive necrosis; however, none of these findings is specific to AOSD (7-9). Liver injury complicated with AOSD can be caused by drug-induced liver damage associated with treatment or other coexisting liver diseases; thus, its diagnosis requires a careful review of the clinical course, discontinuation of suspected drugs, and exclusion of other liver diseases. Here, we review four cases of severe liver injury accompanied by AOSD and identify the clinical and histopathological features.

Case Reports

These four patients were diagnosed with AOSD based on Yamaguchi's criteria (10), and they were all women 26-54 years old with severe liver injury. The positive findings of Yamaguchi's criteria, blood test results, and treatments before and after their admission to our hospital are summarized in Table. Detailed histories for each case are described below. Severe liver damage existed at the AOSD diagnosis in one case and emerged concurrently with relapsed AOSD in the other three cases. Figure 1～2344 showed the clinical course starting at the relapse of AOSD and present liver histopathological images. Liver biopsy specimens were subjected to Hematoxylin and Eosin (HE) staining and immunohistochemical analyses for multiple lymphocyte markers with the following antibodies: anti-cluster of differentiation (CD) 3 (clone PS1, prediluted; Nichirei Biosciences, Tokyo, Japan), anti-CD4 (clone 4B12, prediluted; Leica Biosystems, Nussloch, Germany), anti-CD8 (clone 4B11, prediluted, Leica Biosystems), anti-CD20 (clone IgG2a, dilution 1:500; Dako Cytomation, Glostrup, Denmark), and anti-CD68 (clone PG-M1, dilution 1:100; Dako Cytomation).

Table.

Clinical Features of Four AOSD Cases with Severe Liver Damage.

Case		1	2	3	4
Age/Sex		30/F	39/F	26/F	54/F
Positive findings in Yamaguchi’s criteria	Major criteria	Fever rash	Fever rash Leukocytosis	Fever rash Leukocytosis	Fever rash Arthralgia
	Minor criteria	Splenomegaly Elevated liver enzyme Negative RF/ANA	Elevated liver enzyme Negative RF/ANA	Elevated liver Enzyme negative RF	Lymphadenopathy Elevated liver enzyme Negative RF/ANA
Laboratory data on admission	AST (IU/L)	731	1,238	1,538	637
	ALT (IU/L)	1,623	1,928	1,669	1,675
	T-Bil (mg/dL)	2.0	12.6	0.8	3.0
	PT-INR	1.03	1.40	1.41	1.21
	Ferritin (ng/mL)	2,390.9	2,399.9	4,926.6	1,989.7
Prior treatment		Oral PSL (terminated)	Oral PSL (terminated)	Oral PSL (tapering)	Oral PSL (tapering)
Additional treatment	High-dose intravenous mPSL	Done	Done	Done	Done
	Oral PSL	1.0 mg/kg/day	1.0 mg/kg/day	0.9 mg/kg/day	0.5 mg/kg/day
	Immuno-suppressant	Azathioprine	Azathioprine Cyclosporin	Cyclosporin	Cyclosporin

ALT: alanine aminotransferase, ANA: antinuclear antibody, AOSD: adult-onset Still's disease, AST: aspartate aminotransferase, mPSL: methylprednisolone, PSL: prednisolone, PT-INR: prothrombin time-international normalized ratio, RF: rheumatoid factor, T-Bil: total bilirubin

Figure 1.

Clinical course and liver histopathology in Case 1. (a) Chronological changes in ALT and ferritin and a summary of treatments. (b) Liver pathology revealed moderate lymphocytic infiltration in the portal areas and lobules. The infiltrating inflammatory cells were predominantly CD8-positive and were widely distributed throughout the liver parenchyma (original magnification ×200). Scale bar: 100 μm. ALT: alanine aminotransferase, CD: cluster of differentiation, HE: Hematoxylin and Eosin staining

Figure 2.

Clinical course and liver histopathology in Case 2. (a) Chronological changes in ALT and ferritin and a summary of treatments. (b) HE of the liver tissue revealed spillover infiltration of inflammatory cells throughout the lobule as well as in the portal area. Immunostaining revealed that the infiltrating inflammatory cells were predominantly CD8-positive cells (original magnification ×200). Scale bar: 100 μm. ALT: alanine aminotransferase, CD: cluster of differentiation, HE: Hematoxylin and Eosin staining

Figure 3.

Clinical course and liver histopathology in Case 3. (a) Chronological changes in ALT and ferritin and a summary of treatments. (b) HE of the liver specimen revealed diffuse inflammatory cell infiltration in the lobules. The infiltrating inflammatory cells were predominantly CD8-positive and were widely distributed throughout the lobule (original magnification ×200). Scale bar: 100 μm. ALT: alanine aminotransferase, CD: cluster of differentiation, HE: Hematoxylin and Eosin staining

Figure 4.

Clinical course and liver histopathology in Case 4. (a) Chronological changes in ALT and ferritin and a summary of treatments. (b) HE of liver tissue revealed extensive infiltration of inflammatory cells throughout the lobule. The inflammatory cell infiltrate consisted mainly of CD8-positive cells that were widely distributed in the parenchyma (original magnification ×200). Scale bar: 100 μm. ALT: alanine aminotransferase, CD: cluster of differentiation, HE: Hematoxylin and Eosin staining

Case 1

A 30-year-old woman consulted her local physician with a persistent high fever lasting for 2 months. She was diagnosed with AOSD based on fever, anterior chest erythema, leukocytosis, splenomegaly, and hyperferritinemia, and prednisolone (PSL) at 30 mg/day was initiated. The intermittent fever continued, while the skin rash disappeared and the inflammatory response decreased. Her temperature increased to over 40°C on the 25th day after the visit, and hepatic disorder emerged (Fig. 1a). The fever and inflammatory findings disappeared after intravenous methylprednisolone (mPSL) at 1,000 mg/day and subsequent oral PSL at 45 mg/day administration; however, liver enzyme levels were persistently high, and she was transferred to our hospital. Laboratory tests (Supplementary material 1) showed liver damage with aspartate transferase (AST) of 731 IU/L, alanine aminotransferase (ALT) of 1,623 IU/L, lactate dehydrogenase (LDH) of 435 IU/L, total bilirubin (T-Bil) of 2.0 mg/dL, prothrombin time (PT) of 96%, PT-international normalized ratio (PT-INR) of 1.03, and ferritin of 2,390.9 ng/mL. Antinuclear antibodies, anti-mitochondrial antibodies, and hepatitis virus markers were all negative. A bone marrow examination revealed hemophagocytosis. The patient was diagnosed with AOSD-associated liver injury complicated by hemophagocytic syndrome following the clinical course, bone marrow hemophagocytosis, and negative findings for autoimmune hepatitis (AIH).

The liver injury did not improve with 3 days of intravenous mPSL at 1,000 mg/day and oral PSL at 45 mg (1.0 mg/kg)/day; thus, azathioprine (AZT) was additionally initiated. After that, the liver enzymes decreased, and both PSL and AZT were able to be tapered.

Liver pathology revealed moderate lymphocytic infiltration and minimal fibrosis in the portal areas and lobules, with slight single-cell necrosis. The infiltrating inflammatory cells were predominantly CD8-positive and were widely distributed throughout the liver parenchyma (Fig. 1b).

Case 2

A 39-year-old woman had repeated episodes of a fever and redness of the extremities, which appeared simultaneously and resolved spontaneously. She visited a local doctor and was diagnosed with a severe hepatic disorder, leading to hospitalization. Laboratory tests revealed white blood cells (WBCs) of 16,270 /μL, neutrophils of 93.3%, AST of 1,290 IU/L, ALT of 1,386 IU/L, LDH of 503 IU/L, T-Bil of 3.3 mg/dL, PT of 71%, PT-INR of 1.24, and ferritin of 1,297.3 ng/mL. High-dose intravenous mPSL therapy was followed by oral PSL at 40 mg/day, which was tapered as the hepatic damage resolved. Although she underwent a liver biopsy, the cause of the liver damage was unidentifiable, and the PSL was terminated. Afterwards, a fever and skin rash appeared occasionally, and hyperferritinemia and mild liver injury coexisted, leading to an AOSD diagnosis one year later. The patient was admitted to our hospital four years after her last admission due to repeated symptom flare-ups complicated by a severe liver disorder (Fig. 2a). Laboratory tests (Supplementary material 2) revealed a WBC count of 16,270 /μL, neutrophils of 88.0%, AST of 1,238 IU/L, ALT of 1,928 IU/L, LDH of 395 IU/L, T-Bil of 12.6 mg/dL, PT of 56%, PT-INR of 1.4, and ferritin of 2,399.9 ng/mL. The patient was negative for antinuclear antibodies, rheumatoid factor, anti-mitochondrial antibodies, and hepatitis virus markers. No findings specific to AIH were obtained, and liver damage associated with AOSD was diagnosed based on hematology and drug history.

Three days of intravenous mPSL at 1,000 mg/day followed by oral PSL at 45 mg (1 mg/kg)/day were initiated; however, the patient developed exacerbated liver damage. The liver injury was not ameliorated after adding AZT at 100 mg/day to PSL, and the immunosuppressive agent was switched from AZT to cyclosporin (CyA). Afterwards, liver damage improved, and the dose was gradually tapered.

HE staining of the liver tissue revealed spillover infiltration of inflammatory cells, such as lymphocytes and neutrophils, throughout the lobule and in the portal area and spotty necrotic and degenerated hepatocytes. Immunostaining revealed that the infiltrating inflammatory cells were predominantly CD8-positive cells (Fig. 2b).

Case 3

A 26-year-old woman consulted her primary physician and dermatologist with complaints of a skin rash and fever over 38°C that persisted for several days. She was prescribed antipyretic, antibiotics, and PSL at 15 mg/day, and her symptoms resolved. However, she was hospitalized for the first time due to ocular conjunctival jaundice and markedly elevated liver enzymes in blood tests. The laboratory tests revealed a WBC count of 14,990 /μL, neutrophils of 91.7%, AST of 1,077 IU/L, ALT of 2,084 IU/L, LDH of 578 IU/L, T-Bil of 6.0 mg/dL, PT of 51%, PT-INR of 1.47, and ferritin of 4,985.5 ng/mL. The patient began taking PSL at 50 mg/day because drug-induced liver injury was suspected based on the clinical course and liver biopsy findings. PSL was tapered over seven months after the liver injury improved. The patient presented with a fever exceeding 39°C, general malaise, and redness on both upper limbs 4 months after terminating PSL (Fig. 3a). She was referred and admitted to our hospital due to severe liver injury after being prescribed PSL at 30 mg/day. On admission, diffuse erythema was seen on the bilateral upper extremities. Blood tests revealed AST of 1,538 IU/L, ALT of 1,669 IU/L, LDH of 1,355 IU/L, T-Bil of 0.8 mg/dL, PT of 54%, PT-INR of 1.41, and ferritin of 4,926.6 ng/mL (Supplementary material 3). Antinuclear antibodies were mildly positive at a 1:40 dilution. Anti-mitochondrial antibodies and hepatitis virus markers were negative. A diagnosis of AOSD and associated liver injury was made based on Yamaguchi's criteria, because the histological results of the skin biopsy were consistent with AOSD.

The patient was treated with 1,000 mg/day intravenous mPSL for 3 days and 50 mg (0.9 mg/kg)/day oral PSL afterwards. The hepatic injury improved temporarily but flared up when the PSL dose was reduced to 30 mg/day. CyA was started due to the patient's resistance to steroid therapy, and the liver damage improved.

HE staining of the liver biopsy specimen revealed diffuse infiltration of inflammatory cells, mainly lymphocytes, as well as collapsed necrosis and single-cell necrosis in the lobules. The infiltrating inflammatory cells were predominantly CD8-positive and were widely distributed throughout the lobule (Fig. 3b). No findings specific to AIH, such as plasma cell infiltration, were observed.

Case 4

A 54-year-old woman had been diagnosed with AOSD by her primary physician 1 year prior due to a fever, arthralgia, swollen cervical lymph nodes, and skin rash. Her symptoms improved after taking oral PSL at 20 mg/day, which was gradually tapered. The fever and skin rash flared up again 3 months after PSL discontinuation, so the patient resumed PSL at 30 mg/day with a diagnosis of relapsed AOSD (Fig. 4a). Hepatic disorder appeared one month after PSL resumption and gradually worsened, while the fever and skin rash improved. She was admitted to our hospital three months after restarting PSL. The laboratory tests (Supplementary material 4) revealed liver injury with AST of 637 IU/L, ALT of 1,675 IU/L, LDH of 442 IU/L, T-Bil of 3.0 mg/dL, PT of 71%, PT-INR of 1.21, and ferritin of 1,989.7 ng/mL. The patient was negative for antinuclear antibodies, rheumatoid factor, anti-mitochondrial antibodies, and hepatitis virus markers.

High-dose intravenous mPSL therapy and oral PSL at 30 mg (0.5 mg/kg)/day were introduced; however, liver injury showed no improvement, and CyA was added. The liver disorder improved rapidly, and both drugs gradually decreased.

HE staining of liver tissue revealed extensive inflammatory cell infiltration, such as lymphocytes and neutrophils, throughout the lobule. Hepatocyte swelling, severe perivenular necrosis, and single-cell necrosis were also observed. The liver biopsy specimen revealed no findings specific to AIH, such as plasma cell infiltration. The inflammatory cell infiltrate consisted mainly of CD8-positive cells that were widely distributed in the parenchyma (Fig. 4b).

Discussion

Although AOSD is known to be frequently associated with liver injury, severe cases are uncommon. This study evaluated the histopathological and clinical features of four patients with severe liver damage who required hospitalization.

AOSD is a systemic inflammatory disease characterized by a spiking fever, arthritis, skin rash, sore throat, lymphadenopathy, and hepatosplenomegaly. It is classified into three patterns based on its clinical course: monocyclic, polycyclic, and chronic articular. The monocyclic type is a single episode with systemic symptoms lasting less than one year, followed by permanent remission. The polycyclic type has repeated systemic flares characterized by remissions ranging from weeks to years between relapses. The chronic articular type is dominated by persistent symptoms lasting for one year or more (11,12). Case 1 was monocyclic, and Cases 2, 3, and 4 were polycyclic in the present study. Although AOSD is slightly more common in women than in men, and certainly all patients in this study were women, it is undefined whether sex differences play a crucial role in the onset of liver injury associated with AOSD or not. Although the pathogenesis of AOSD remains unclear, numerous recent reports have suggested that AOSD is a complex condition involving both the innate and adaptive immune systems triggered by second-hit, unknown factors, such as infection, which ultimately leads to aberrant inflammatory response activation and AOSD development (12). Innate immune cell activation, such as macrophages (3,13) and neutrophils (14), and the impaired cytotoxic function as well as the decreased frequency of natural killer (NK) cells (15) are thought to be involved in AOSD pathogenesis. Hyperactivation of Th1 cells in the adaptive immune system is known to cause excessive interferon-gamma (IFNγ) production and cytotoxic T-cell activation. Intracellular IFNγ expression in AOSD has also been reported to be increased in cytotoxic T cells and NK cells (16).

Yamaguchi's criteria (10) are widely used for diagnosing AOSD, with a sensitivity of 93.5% and a specificity of 92.1%; however, AOSD is difficult to diagnose due to the lack of specific clinical manifestations and serum biomarkers. In addition, a definitive diagnosis requires the exclusion of neoplastic, infectious, autoimmune, and autoinflammatory diseases. It is difficult to determine the cause of liver damage, including AOSD exacerbation, therapeutic drug-induced liver injury, or other liver diseases, in addition to the difficulty in diagnosing AOSD itself. Previously reported liver biopsy findings in active AOSD have been nonspecific, and the distinction was based on the clinical course, discontinuation of the suspected drug, and exclusion of other causes. The present study carefully evaluated liver biopsy specimens from AOSD patients with severe liver damage. Significant CD8-positive T-cell infiltration throughout the liver parenchyma was observed in all cases, whereas CD4-positive T-cell and CD20-positive B-cell infiltrations were mild. It was reported that CD8-positive cells are not always predominant in AIH and other chronic active hepatitis, and CD4-positive and CD20-positive cells are frequently infiltrated the portal area (17,18), which is significantly different from the cases in this study. Kim et al. histologically evaluated the skin, lymph nodes, bone marrow, and liver of patients with AOSD (19). CD8-positive cells were dominant in the skin, bone marrow, and liver, whereas CD4-positive cells were dominant in the lymph nodes. The proportions of neutrophils and CD8-positive T cells were significantly increased in peripheral blood leukocyte fractions of AOSD, whereas NK and CD4-positive T cells were decreased (20,21). CD68-positive cell infiltration into the sinusoids was observed, while CD8-positive cells were significantly increased in this study. These findings suggest that the effect of macrophages on liver injury may be restricted, although macrophage activation may be involved in the development of AOSD. CD4-positive and CD25-high regulatory T cells are decreased in AIH and in cases of immune-related adverse events (irAEs), which may interact with the disease's pathogenesis (22,23). Here, we detected a relative reduction in CD4-positive cells in liver tissue, suggesting that a decrease in regulatory T cells may also have been involved in the pathogenesis of AOSD liver injury. CD8-positive cell infiltration throughout the liver lobules might be a pathologic feature of AOSD liver injury; thus, their cytotoxic effect may be involved in the liver pathogenesis of AOSD. Liver injury associated with AOSD should be considered when extensive CD8-positive lymphocyte infiltration into the parenchyma is observed in liver biopsy specimens from patients with liver injury of unknown etiology.

All four patients had been introduced to corticosteroids and had their doses tapered before the onset of severe liver injury. As their symptoms, such as a fever, skin rash, and leukocytosis, improved during the process of corticosteroid tapering, the treatment response seemed to be favorable. Severe liver damage that was uncontrolled by high-dose steroids alone developed after steroid discontinuation; therefore, the addition of an immunosuppressive agent was required. The interval from steroid therapy completion to liver injury development ranged from one month to four years, and some patients had no other objective findings or symptoms. A review of 17 cases of acute liver failure (ALF) associated with AOSD noted that ALF could occur at any time, e.g. at the time of the AOSD diagnosis, during immunosuppressive therapy tapering, or a few years after the diagnosis (8). In addition, ALF that emerges during immunosuppressant tapering tends to respond more poorly to steroid therapy than ALF diagnosed simultaneously with AOSD (8). Furthermore, the more frequently AOSD with ALF relapses, the less responsive it is to intravenous methylprednisolone therapy at high doses (8), which was consistent with our cases.

Patients with AOSD should therefore continue to undergo periodic blood tests, even if they are asymptomatic, after the completion of steroids. In addition, possible resistance to steroid treatments should be considered when liver injury occurs in patients with AOSD. The first clinical practice guideline was established in Japan in 2017, although no internationally accepted recommendation for managing AOSD exists. Systemic corticosteroids and methotrexate (MTX) were strongly encouraged as effective treatments, while other conventional synthetic disease-modifying anti-rheumatic drugs (csDMARDs) and biologic agents were weakly recommended (24). Nonsteroidal anti-inflammatory drugs (NSAIDs) are considered adjuncts in the diagnostic process, with response rates of 16-18% (25). Corticosteroids are used as first-line agents following the AOSD diagnosis. The effectiveness of corticosteroids is approximately 65% (25,26), which is more significant in the monocyclic and polycyclic patterns of AOSD (25,26) and worse in AOSD patients with a high proportion of CD8-positive T cells in the peripheral blood (20). csDMARDs should be added when corticosteroids fail to control the disease or in cases of steroid dependence; however, no evidence has shown that csDMARDs are effective for AOSD liver injury. MTX has been reported to control the disease well in all forms and to be effective in 40-70% of patients with steroid-dependent AOSD (27), but exacerbated liver injury due to MTX is a concern, because MTX frequently causes drug-induced liver injury. Some reports suggest that CyA may be as effective as MTX in monocyclic and polycyclic AOSD (25,28). AZT is efficacious in some patients with systemic forms (25), and treatment with biological agents, such as interleukin (IL)-1, IL-6, and IL-18 (29-32) inhibitors, can be considered if the patient is refractory to the above therapies. The present study revealed that corticosteroid-resistant high-grade liver injury was successfully treated with csDMARDs, including CyA. CyA is a calcineurin inhibitor and selectively inhibits T-cell activation by blocking nuclear translocation of the nuclear factor of activated T cells. As mentioned above, excessive IFNγ secreted by hyperactivated Th1 cells was reportedly involved in AOSD pathogenesis. This is consistent with our finding that robust CD8-positive cytotoxic T cells infiltrated the hepatic lobules in cases of severe liver injury complicated by AOSD. In addition, hepatic irAEs during immune checkpoint inhibitor treatment presented robust CD8-positive cell infiltration within the hepatic parenchyma (17,33), and immunosuppressive agents, including calcineurin inhibitors, are often required to manage irAE liver injury in addition to corticosteroids. Therefore, calcineurin inhibitors are reasonable treatment options for liver injury associated with cytotoxic T cells, including AOSD and irAEs, to suppress T-cell activation. IL-6 expression is enhanced in immune cells, such as T lymphocytes and macrophages, promoting immune responses and inflammatory processes in AOSD (11). The inhibition of disease progression by IL-6 inhibitors, including tocilizumab, is also a therapeutic option.

One limitation of this study is that it only included AOSD cases accompanied by severe liver injury. Thus, future studies should investigate the similarity of histopathological findings in livers of different severities (ALF or mild liver injury) by increasing the number of AOSD cases. Furthermore, all four cases were evaluated for their liver pathology at the time of recurrence of liver injury, so it was impossible to determine whether the initial and recurrent liver injury had a similar pathology or not. Although the infiltration of CD8-positive cells into the liver may increase with prolonged inflammation, CD8-positive cells are not always predominant in AIH and other chronic active hepatitis cases (17,18). Consequently, CD8-positive cell infiltration throughout the liver lobules might be a pathologic feature of AOSD liver injury. In addition, in more than half of the cases, a liver biopsy was performed after corticosteroid treatment, which might have modulated the activation of macrophages or neutrophils. Only Case 2 underwent a liver biopsy before steroids were administered, confirming the finding that CD8-positive cells had more prominent infiltration than CD68-positive cells without treatment. We should therefore evaluate more cases of AOSD complicated with severe liver injury in the future.

In conclusion, periodic visits and blood tests should be continued even after remission, because patients with AOSD can develop severe liver damage during tapering or after discontinuation of steroids. The pathology of liver injury associated with AOSD shows robust CD8-positive T lymphocyte infiltration throughout the liver lobules, which can be helpful for making an AOSD diagnosis. We should consider the addition of immunosuppressive agents, including calcineurin inhibitors, with the prospect of possible resistance to corticosteroids for the treatment of AOSD complicated by severe liver injury.

All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The analysis of the data in this study was approved by the Ethics Committee of Kyushu University Hospital (No. 27-377 and 28-432), and written informed consent was waived because of the retrospective design. Consent for publication was obtained from all patients.

The authors state that they have no Conflict of Interest (COI).

Supplementary Material

Table S1

Case 1 laboratory data on admission

Table S2

Case 2 laboratory data on admission

Table S3

Case 3 laboratory data on admission

Table S4

Case 4 laboratory data on admission