Quality of Life Research (2021) 30:2853–2861 https://doi.org/10.1007/s11136-021-02850-0 Health‑related quality of life in patients with autoimmune hepatitis Maurice Michel1,2 · Francesca Spinelli1,2 · Annette Grambihler1,2 · Christian Labenz1,2 · Michael Nagel1,2 · Leonard Kaps1,2 · Yvonne Huber1,2 · Peter R. Galle1,2 · Marcus‑Alexander Wörns1,2 · Jörn M. Schattenberg1,2 Accepted: 16 April 2021 / Published online: 12 May 2021 © The Author(s) 2021 Abstract Background Autoimmune hepatitis (AIH) is a rare chronic liver disease. Impaired health-related quality of life (HRQL) contributes to the overall disease burden. At current, only limited data related to the impact of treatment response on HRQL are available. Objective The aim of the study was to determine the impact of biochemical remission on HRQL. Methods Patients with AIH were prospectively enrolled between July 2018 and June 2019. A liver disease-specific tool, the chronic liver disease questionnaire (CLDQ) and the generic EQ-5D-5L were used to quantify HRQL. Treatment response was assessed biochemically by measurement of immunoglobulin G, ALT and AST. The cohort was divided into two groups according to their biochemical remission status in either complete vs. incomplete remission. Clinical as well as laboratory parameters and comorbidities were analysed using univariable and multivariable analysis to identify predictors of poor HRQL. Results A total of 116 AIH patients were included (median age: 55; 77.6% female), of which 9.5% had liver cirrhosis. In this cohort, 38 (38.4%) showed a complete and 61 (61.6%) an incomplete biochemical remission at study entry. The HRQL was significantly higher in patients with a complete as compared to an incomplete biochemical remission (CLDQ overall score: 5.66 ± 1.15 vs. 5.10 ± 1.35; p = 0.03). In contrast, the generic EQ-5D-5L UI-value was not different between the groups. Multivariable analysis identified AST (p = 0.02) and an incomplete biochemical remission (p = 0.04) as independent predic- tors of reduced HRQL (CLDQ total value). Conclusion Patients with a complete biochemical remission had a significantly higher HRQL. Liver-related quality of life in patients living with AIH is dependent on the response to immunosuppressive treatment. Keywords AIH · Biochemical remission · HRQL · CLDQ · EQ-5D-5L Abbreviations HRQL Health-related quality of life AIH Autoimmune hepatitis IgG Immunoglobulin G ALT Alanine transaminase IQR I nterquartile range ALP A lkaline phosphatase PBC Primary biliary cholangitis AST Aspartate transaminase PSC Primary sclerosing cholangitis gGT Gamma-glutamyltransferase CLDQ C hronic Liver Disease Questionnaire EQ-5D-5L European Quality-of-Life 5-Dimension Introduction 5-Level Autoimmune hepatitis (AIH) is a rare, chronic liver disease characterized by chronic inflammation targeting hepatocytes * Jörn M. Schattenberg with a prevalence ranging from 16 to 18 cases per 100,000 joern.schattenberg@unimedizin-mainz.de inhabitants in Europe [1]. If left untreated, AIH can progress 1 Metabolic Liver Research Program, I. Department to end stage liver disease or acute liver failure with liver of Medicine, University Medical Centre Mainz, Mainz, transplantation being required [2, 3]. Immunosuppressive Germany therapy has significantly improved the prognosis of patients 2 I. Department of Medicine, University Medical Centre with AIH [4]. The combination of corticosteroids with Mainz, Mainz, Germany Vol.:(012 3456789) 2854 Quality of Life Research (2021) 30:2853–2861 azathioprine is recommended for first line treatment. Fre- or primary sclerosing cholangitis (PSC) in addition to the quently, tapering of immunosuppressive therapy over time diagnosis of AIH. For inclusion, patients had to be at least can be successfully performed without losing remission [5, 18 years of age. Patients with an expected life expectancy of 6]. By preventing recurrent flares, disease progression can < 6 months were not included, in particular patients with a be halted in most patients, although long-term mortality still history of hepatocellular carcinoma (HCC) or malignancies remains higher compared to the general population [7]. In were not included. Coexisting liver disease other than AIH the management of AIH, biochemical remission defined by or AIH-overlap (i.e. viral hepatitis) were excluded. Labora- normalization of ALT, AST and immunoglobulin G (IgG) is tory values were obtained on the day of inclusion. Medical considered a complete response to treatment [8, 9]. and treatment history were recorded and available from the Beyond biochemical remission, reported or perceived electronic health care records. Immunosuppressive therapy health is critical for the management of patients with AIH, was recorded at the time of the HRQL assessment by the and health-related quality of life (HRQL) is an integral part treating hepatologists. To investigate the impact of actual of the general concept of being healthful. HRQL—the deter- treatment and response on HRQL, only current treatments mination of how a patient feels and functions—can be meas- were considered. ured by different instruments in defined domains that include general health or disease-specific health aspects. HRQL is Definition of biochemical remission potentially influenced by disease activity, but also adverse effects from treatment or comorbidities. In patients with Biochemical remission at study entry was defined as either AIH, an increased prevalence of anxiety and major depres- complete (normalization of ALT, AST and IgG) or incom- sive symptoms have been observed [10]. While an effect plete (elevation of one or two, or persistently elevated lev- of these comorbidities on HRQL appears obvious, depres- els of the surrogates: ALT, AST and IgG) according to sion and anxiety are also associated with a greater degree published criteria [15]. Complete normalization had to be of non-adherence to immunosuppressive therapy [11] and maintained for at least 6 months before considering com- thus can accelerate disease progression and the burden of plete remission. Due to missing blood values, 17 patients disease. Adverse effects from the use of corticosteroids can were excluded for the final comparison between complete also potentially impair HRQL in AIH [12]. In a Japanese vs. incomplete biochemical remission (n = 99). The upper study exploring 265 patients with AIH, a negative impact of limit of normal (ULN) was defined as follows according corticosteroid use on the subdomain ‘worry’—but not over- to the reference ranges of the Institute of Clinical Chemis- all HRQL scores—was detected [13]. Studies that explored try and Laboratory Medicine at University Medical Centre budesonide did not find a difference for the use of budeson- Mainz: ALT in males ≤ 50 U/l, ALT in females ≤ 35 U/l, ide over corticosteroids with regards to HRQL [14]. AST in males ≤ 35 U/l, AST in females ≤ 31 U/l, IgG ≤ 16.0 Currently, there is only limited data available related to g/L irrespective of sex. Furthermore, the ULN of the remain- the impact of biochemical remission on HRQL in AIH. Yet, ing laboratory values was as follows: ALP in males ≤ 138, no study has focussed on correlating clinical and biochemi- ALP in females ≤ 111 U/l, gGT in males ≤ 64 U/l, gGT in cal parameters with HRQL. Hence, the aim of this study was females ≤ 36 U/l, bilirubin ≤ 1.2 mg/dl irrespective of sex. to analyse HRQL using a liver disease-specific tool, compare these to generic measures of quality of life and determine Health‑related quality of life (HRQL) assessment the effects of biochemical remission and identify surrogate markers of HRQL. HRQL was assessed by the validated German version of the Chronic Liver Disease Questionnaire (CLDQ) and the European Quality-of-Life 5-Dimension 5-Level (EQ- Materials and methods 5D-5L) questionnaire under standardized conditions. The CLDQ compromises a total of 29 items and it is divided Study population and population subtypes into 6 subscale scores (abdominal symptoms, fatigue, sys- temic symptoms, activity, emotional functioning, worry) A total of 116 patients with AIH were enrolled between July and an overall score. The results of the scores are presented 2018 and July 2019 in this non-interventional and cross- on a scale from 1 to 7, with 1 indicating worst HRQL (bad) sectional cohort study after informed consent was obtained and 7 indicating best HRQL (good). The CLDQ is more at the outpatient clinic of the Metabolic Liver Research Pro- specific regarding liver related health issues [16]. The gram of the University Medical Centre Mainz in Germany. EQ-5D-5L represents a general assessment of the quality Inclusion criteria were AIH and AIH-overlap according of life [17], which has also been used in AIH previously to the EASL clinical practice guidelines [8]. AIH-overlap and other liver related pathologies [12]. It consists of five included patients with primary biliary cholangitis (PBC) dimensions, including mobility, self-care, usual activities, 1 3 Quality of Life Research (2021) 30:2853–2861 2855 pain/discomfort and anxiety/depression with each dimen- Results sion having five response levels: no problems, slight prob- lems, moderate problems, severe problems and unable to/ Study population extreme problems. The response levels are represented by numbers from 1 to 5, with 1 indicating “no problems” and A total of 116 patients with AIH were enrolled prospec- 5 indicating “extreme problems”. The EQ visual analogue tively. The majority of patients were female (n = 90, 77.6%) scale (VAS) is part of the EQ-5D-5L score and records the with a median age of 55 years (IQR 41; 64) at inclusion. respondent’s overall current health, ranging from 0 (the Baseline characteristics are summarized in Table 1. In the worst health you can imagine) to 100 (the best health you entire cohort, median ALT (30.5 U/I (IQR 21; 44.25)) was can imagine). From these five subdimensions, the utility below ULN, while median AST (33 U/I (IQR 26; 42)) and index (UI) was calculated. In order to calculate the UI, gGT (38 U/I (IQR 22.5; 84.5) were slightly above the ULN. the EQ-5D-5L value set for Germany was used [18]. In Likewise, median bilirubin (total bilirubin: 0.72 mg/dl; this study, the calculated UI ranges from − 0.21 (worst IQR 0.50, 1.01) and alkaline phosphatase (ALP; 87 U/I; health) to 1 (best health). Higher values on both VAS and IQR 64; 117) were within the ULN. A total of 19 (16.4%) UI indicate a higher HRQL. In contrast, a higher score patients were diagnosed with AIH-overlap syndrome with on the subdomains of EQ-5D-5L is associated with a primary biliary cholangitis (PBC) in 11 (9.5%) and primary lower HRQL. Both questionnaires were completed by all sclerosing cholangitis (PSC) in 8 (6.9%) patients. Compen- enrolled patients (n = 116) during the visit at the outpatient sated liver cirrhosis was present in 11 patients (9.5%). Most study centre. patients were treated with azathioprine (87%), while 58.8% received corticosteroids. A complete biochemical remis- sion was seen in 38.4% (n = 38) of patients. In turn, 61.6% Ethics (n = 61) showed an incomplete biochemical remission, as indicated by only partial normalization (either one or two All patients provided written informed consent. The study of ALT/AST and IgG) or persistently elevated blood levels was conducted according to the ethical guidelines of the of ALT, AST and IgG. 1975 Declaration of Helsinki (6th revision, 2008). The study protocol was approved by the ethics committee of the Landesärztekammer Rhineland-Palatine on 25th May Comparison of patients with complete 2010 with amendment on 07th May 2013 (Nr. 873.199.10 and incomplete biochemical remission 7208). The median age between patients with complete and incom- plete remission was comparable between both groups. Statistical analysis Besides the surrogate markers of biochemical remission, gGT and ALP blood levels were significantly higher in the Descriptive analysis of data is expressed as either mean incomplete remission group. More patients with incomplete or median with standard deviations or interquartile ranges remission were on a dual therapy with azathioprine and ster- (IQR). To compare groups and to calculate differences oids and showed a higher use of steroids in comparison to between two groups with quantitative values, Mann–Whit- patients with a complete remission (Table 1). ney U rank test was used. For the comparison of two or more patient groups a Chi-square test was applied. All tests were two-tailed, statistically significant values were defined as Health‑related quality of life in AIH p < 0.05. Univariable regression analysis was used to exam- ine associations between two variables. All variables with The mean CLDQ total value in the study cohort was p < 0.1 and based on clinical importance were then included 5.31 ± 1.25. The subdomain fatigue of the CLDQ showed into a multivariable linear regression model. To avoid mul- the lowest score (4.34 ± 1.66) with no significant differ- ticollinearity, the variables AST and incomplete remission ence between male or female. In contrast, the subdomain were independently analysed in two multivariable models, activity of the CLDQ showed the highest mean score with respectively. Because the data analysis was exploratory, no 5.84 ± 1.31 (Table 2), and it significantly differed between adjustment for multiple testing was performed. Due to the male and female (m: 6.37 ± 0.89 vs. f: 5.69 ± 1.37; p = 0.01) large number of tests, p-values should be interpreted with (Supplementary Table 1). In the EQ-5D-5L, the subdomain caution and in connection with effect estimates. For all data pain/discomfort showed the highest mean score (1.97 ± 0.98) analysis and statistical tests, IBM SPSS Statistic Version while the domain self-care had the lowest mean score 23.0 (Armonk, NY: IBM Corp.) was used. (1.23 ± 0.66). The mean VAS score and UI-value were 1 3 2 856 Quality of Life Research (2021) 30:2853–2861 Table 1 Clinical characteristics, demographic data and differences between complete and incomplete biochemical remission Variables Total cohort (n = 116) Complete remission (n = 38) Incomplete remis- p-value sion (n = 61) n (%) or median (25th; 75th) n (%) or median (25th; 75th) n (%) or median (25th; 75th) Age at inclusion 55 (41; 64) 54 (40.5; 73) 52 (41; 64) 0.18 Sex, female 90 (77.6) 31 (81.6) 46 (75.4) 0.32 Time since diagnosis (> 10 years) 48 (41.4) 17 (44.7) 27 (44.3) 0.56 Disease duration (years) 6 (2.5; 13) 8 (3.2;13.7) 6 (2; 14) 0.64 Type 2 diabetes 15 (12.9) 4 (10.5) 10 (16.4) 0.42 Arterial hypertension 28 (24.1) 10 (26.3) 14 (22.9) 0.70 Laboratory ALT (U/l) n = 106 30.5 (21; 44.25) 22.5 (14.5; 29.5) 38 (31; 74) < 0.001 AST (U/I) n = 107 33 (26; 42) 26 (22; 28.7) 39 (34; 58) < 0.001 ALP (U/l) n = 93 87 (64; 117) 69.5 (57.7; 68.2) 101 (73; 132)a 0.004 gGT (U/l) n = 105 38 (22.5; 84.5) 24.5 (20.7; 29.7) 66 (29; 87) < 0.001 Total bilirubin (g/dl) n = 101 0.72 (0.50; 1.01) 0.74 (0.6; 0.86) 0.63 (0.5; 0.9) 0.82 IgG (g/dl) n = 99 12 (10.4; 15.7) 10.2 (8.9; 12.3) 13.2 (11.5; 17.7) < 0.001 Thrombocytes (1000/µ) n = 99 237 (187; 289) 238 (194; 302) 240 (187; 284)b 0.71 Medication Steroids 67 (58.8) 15 (39.5) 42 (68.9) 0.004 Prednisolone 42 (37.2) 9 (23.7) 26 (42.6) 0.11 Budesonide 25 (21.6) 6 (15.8) 16 (26.3) 0.17 Azathioprine 101 (87) 34 (89.5) 53 (86.9) 0.76 Azathioprine monotherapy 42 (36.2) 23 (60.6) 16 (26.2) 0.08 Azathioprine and steroid therapy 59 (50.8) 11 (28.9) 37 (60.7) 0.001 AIH-related comorbidities AIH-overlap syndrome 19 (16.4) 4 (10.6) 11 (18) 0.24 PBC 11 (9.5) 2 (5.3) 6 (9.8) 0.33 PSC 8 (6.9) 2 (5.3) 5 (8.2) 0.43 Liver cirrhosis 11 (9.5) 4 (10.5) 5 (8.2) 0.48 Data are expressed as numbers, median, percentage (%) or interquartile ranges (IQR) ALT alanine-aminotransaminase, AST aspartat-aminotransaminase, ALP alkaline phosphatase, gGT gamma-GT, IgG immunoglobulin G p-values refer to the comparison between complete vs. incomplete remission. Boldface indicates statistical significance. A p-value < 0.05 was considered statistically significant a Measured in 56 patients b Measured in 57 patients 71.2 ± 20.5 and 0.86 ± 0.18, respectively. In women, the EQ- (CR: 5.66 ± 1.15 vs. IR: 5.10 ± 1.35, p = 0.03). The CLDQ- 5D-5L UI-value and the subdomain pain/discomfort were subdomains abdominal symptoms (CR: 6.03 ± 1.38 vs. lower compared to men. Overall, no difference in HRQL IR: 5.36 ± 1.53, p = 0.01), activity (CR: 6.15 ± 1.23 vs. IR: between patients with AIH and AIH-overlap syndrome were 5.61 ± 1.44, p = 0.03), emotional function (CR: 5.44 ± 1.40 detectable (Supplementary Table 2). vs. IR: 4.88 ± 1.41, p = 0.04) and worry (CR: 5.83 ± 1.47 vs. IR: 5.34 ± 1.43, p = 0.03) were significantly higher in Health‑related quality of life in complete vs. the group with complete biochemical remission (Fig. 1). incomplete biochemical remission Although the subdomain fatigue had the lowest score over- all, no significant difference was seen between patients with Next, the cohort was divided into two groups—complete complete or incomplete biochemical remission. vs. incomplete biochemical remission—according to the Using the generic HRQL EQ-5D-5L questionnaire, only blood levels of ALT, AST and IgG at the day of presentation. the subdomain ‘usual activities’ showed a significant dif- The total CLDQ score was significantly higher in complete ference between the two groups (CR: 1.34 ± 0.71 vs. IR: remission (CR) as compared to incomplete remission (IR) 1.80 ± 1.2, p = 0.03) (Fig. 2). Patients with an incomplete 1 3 Quality of Life Research (2021) 30:2853–2861 2857 Table 2 Mean scores of each questionnaire factors that remained independently associated with a worse Variable Total cohort (n = 116) HRQL (Table 3). In turn, type 2 diabetes (β: − 0.37; 95% CI − 0.54, − 0.20) was the only factor to be associated with CLDQ an impaired HRQL as captured by EQ-5D-5L UI-value in CLDQ abdominal symptoms 5.57 ± 1.50 an univariable analysis. In a multivariable linear regression CLDQ fatigue 4.34 ± 1.66 analysis, type 2 diabetes (β: − 0.35; 95% CI − 0.56, − 0.15) CLDQ systemic symptoms 5.45 ± 1.31 remained independently associated with a worse HRQL CLDQ activity 5.84 ± 1.31 (Table 4). CLDQ emotional function 5.10 ± 1.38 CLDQ worry 5.55 ± 1.39 CLDQ total value 5.31 ± 1.25 Discussion EQ-5D-5L EQ-5D-5L mobility 1.49 ± 0.88 In this study, we assessed HRQL using two validated ques- EQ-5D-5L self-care 1.23 ± 0.66 tionnaires—the liver-specific CLDQ and the generic EQ- EQ-5D-5L usual activities 1.60 ± 0.95 5D-5L—in patients with AIH. Each questionnaire is vali- EQ-5D-5L pain/discomfort 1.97 ± 0.98 dated in German and carries several subdomains to address EQ-5D-5L anxiety/depression 1.63 ± 0.94 different aspects of a patients’ well-being. The central find- EQ-5D-5L VAS 71.2 ± 20.5 ing of the current analysis is that patients with a complete EQ-5D-5L UI-value 0.86 ± 0.18 biochemical remission have a higher HRQL in comparison Data are expressed as means with standard deviation to an incomplete remission in this cohort, when using a liver disease-specific tool. While the CLDQ overall score as well as several subdomains exhibited a positive correlation with remission exhibited a numerically lower VAS of 70 ± 21.3 biochemical remission, no correlation with the general and lower UI-value of 0.83 ± 0.22, with, however, no statisti- HRQL tool—EQ-5D-5L—was detectable. One important cally significant difference. The complete statistical analysis aspect is that the CLDQ was designed to capture liver-spe- of both questionnaires is provided in Supplementary Table 3. cific aspects of the quality of life, which are not assessed by the generic/general questionnaires [16]. Additionally, Predictors of poor health‑related quality of life we observed a significant correlation of several liver related in AIH blood and clinical parameters with the total value of CLDQ, but not the UI-value of EQ-5D-5L. In a univariable analysis, an incomplete remission (β: The CLDQ overall score was 5.31 among all AIH − 0.21; 95% CI − 0.43, − 0.02), blood levels of AST (β: patients. Interestingly, this overall score was comparable − 0.21; 95% CI − 0.41, − 0.02) and type 2 diabetes (β: to patients with compensated cirrhosis and covert hepatic − 0.21; 95% CI − 0.39, − 0.03) were associated with an encephalopathy that were assessed in the same outpatient impaired HRQL. In a multivariable linear regression analy- setting [19] and also comparable to a Japanese study that sis, AST (β: − 0.22; 95% CI − 0.42, − 0.03) and incomplete compared AIH with healthy controls [13]. Furthermore, remission (β: − 0.21; 95% CI − 0.38, − 0.02) were the two fatigue is a prominent symptom in chronic liver disease and Fig. 1 Scores of CLDQ total value and its respective subdo- mains in complete and incom- plete biochemical remission. Scores are represented as mean values ± standard deviation. A p-value < 0.05 is considered sta- tistically significant (*p < 0.05). Please note: A higher value on CLDQ indicates a better HRQL with 7 being the highest pos- sible score to obtain 1 3 2858 Quality of Life Research (2021) 30:2853–2861 Fig. 2 Scores of EQ-5D-5L and its respective subdomains in complete and incomplete biochemical remission. Scores are represented as mean values ± standard deviation. A p-value < 0.05 is considered sta- tistically significant (*p < 0.05). Please note: A higher value on EQ-5D-5L subdomains indicates a worse HRQL with 5 being the highest possible score to obtain Table 3 Uni- and multivariable analyses of clinical and laboratory parameters with CLDQ total value Variable CLDQ total value Univariable analysis Multivariable analysisa Multivariable analysisb β 95% CI p β 95% CI p β 95% CI p Sex − 0.11 − 0.31, 0.06 0.19 − 0.13 − 0.32, 0.07 0.19 − 0.08 − 0.34, 0.06 0.16 Age at study entry − 0.07 − 0.27, 0.10 0.37 Disease duration − 0.07 − 0.28, 0.11 0.37 Time since diagnosis (> 10 years) − 0.16 − 0.34, 0.02 0.09 − 0.14 − 0.34, 0.04 0.13 − 0.14 − 0.34, 0.06 0.16 AST − 0.21 − 0.41, − 0.02 0.03 − 0.22 − 0.42, − 0.03 0.02 ALT − 0.05 − 0.25, 0.15 0.64 gGT − 0.09 − 0.29, 0.11 0.38 IgG 0.01 − 0.21, 0.22 0.95 Steroids − 0.09 − 0.28, 0.10 0.33 Azathioprine monotherapy 0.03 − 0.16, 0.22 0.75 Azathioprine and steroid therapy − 0.10 − 0.30, 0.10 0.32 Incomplete remission − 0.21 − 0.43, − 0.02 0.03 − 0.21 − 0.38, − 0.02 0.04 Liver cirrhosis − 0.11 − 0.29, 0.08 0.25 AIH-overlap − 0.02 − 0.20, 0.17 0.89 Type 2 diabetes − 0.21 − 0.39, − 0.03 0.02 − 0.17 − 0.36, 0.02 0.08 − 0.18 − 0.38, 0.02 0.08 Arterial hypertension − 0.15 − 0.33, 0.03 0.11 Sex: 1 for male, 2 for female; Time since diagnosis (> 10 years): 1 for No, 2 for Yes; Steroids 1 for No, 2 for Yes; Azathioprine monotherapy: 1 for No, 2 for Yes; Azathioprine and steroid therapy: 1 for No, 2 for Yes; Incomplete remission: 1 for No, 2 for Yes; Liver Cirrhosis: 1 for No, 2 for Yes; AIH-Overlap: 1 for No, 2 for Yes; Type 2 Diabetes: 1 for No, 2 for Yes; Arterial Hypertension: 1 for No, 2 for Yes; ALT, Alanine- Aminotransaminase; AST, Aspartat-Aminotransaminase; gGT, Gamma-GT; IgG, Immunoglobulin G At first, univariable analysis of data was done. With all factors showing a p-value < 0.1 and the clinical parameter ‘sex’, a multivariable linear regression model was built. Confidence Interval (CI) and Beta (β) show standardized values, respectively. Boldface indicates statistical signifi- cance. A p-value < 0.05 was considered statistically significant a Multivariable linear regression model including the variables: sex, time since diagnosis, AST, type 2 diabetes b Multivariable linear regression model including the variables: sex, time since diagnosis, incomplete remission, type 2 diabetes occurs in a variety of aetiologies including viral hepatitis, captured fatigue as the most relevant and burdening symp- cholestatic liver disease or non-alcoholic fatty liver disease tom in AIH and the CLDQ subdomain fatigue exhibited the [20, 21]. In the current study, the employed questionnaires lowest score. Similar findings were replicated in a study that 1 3 Quality of Life Research (2021) 30:2853–2861 2859 Table 4 Uni- and multivariable Variable UI-value analysis of clinical and laboratory parameters with Univariable analysis Multivariable a nalysisa EQ-5D-5L UI-value β 95% CI p β 95% CI p Sex − 0.12 − 0.31, 0.06 0.19 − 0.03 − 0.24, 0.17 0.74 Age at study entry − 0.10 − 0.29, 0.08 0.27 Disease duration − 0.01 − 0.20, 0.18 0.92 Time since diagnosis − 0.06 − 0.25, 0.13 0.52 (> 10 years) AST − 0.11 − 0.31, 0.09 0.27 ALT − 0.01 − 0.20, 0.20 0.99 gGT -0.03 − 0.23, 0.19 0.79 IgG 0.08 − 0.13, 0.30 0.43 Steroids − 0.14 − 0.33, 0.05 0.14 Azathioprine monotherapy 0.10 − 0.08, 0.29 0.28 Azathioprine and steroid therapy − 0.14 − 0.35, 0.06 0.16 Incomplete remission − 0.17 − 0.40, − 0.03 0.09 − 0.15 − 0.36, 0.04 0.12 Liver cirrhosis − 0.03 − 0.21, 0.16 0.79 AIH-overlap 0.09 − 0.09, 0.28 0.33 Type 2 diabetes − 0.37 − 0.54, − 0.20 < 0.001 − 0.35 − 0.56, − 0.15 0.001 Arterial hypertension − 0.15 − 0.38, 0.03 0.10 − 0.08 − 0.22, 0.17 0.42 Sex: 1 for male, 2 for female; Time since diagnosis (> 10 years): 1 for No, 2 for Yes; Steroids 1 for No, 2 for Yes; Azathioprine monotherapy: 1 for No, 2 for Yes; Azathioprine and steroid therapy: 1 for No, 2 for Yes; Incomplete remission: 1 for No, 2 for Yes; Liver Cirrhosis: 1 for No, 2 for Yes; AIH-Overlap: 1 No, 2 for Yes; Type 2 Diabetes: 1 for No, 2 for Yes; Arterial Hypertension: 1 for No, 2 for Yes; ALT, Alanine- Aminotransaminase; AST, Aspartat-Aminotransaminase; gGT, Gamma-GT; IgG, Immunoglobulin G At first, univariable analysis of data was done. With all factors showing a p-value < 0.1 and the clinical parameter ‘sex’, a multivariable linear regression model was built. Confidence Interval (CI) and Beta (β) show standardized values, respectively. Boldface indicates statistical significance. A p-value < 0.05 was considered statistically significant a Multivariable linear regression model including the variables: sex, incomplete remission, type 2 diabetes, arterial hypertension employed the modified fatigue impact score (MFIS) and indicating that women perceive and experience HRQL observed a strong impact of AIH on fatigue [14]. A recent impairment stronger than men. Similar findings have been study highlighted that fatigue in patients with decompen- reported previously in other chronic liver diseases [21] and sated liver cirrhosis was associated with anaemia [19]. In the cirrhosis [19] and this is likely related to a different self- present study, we did not enrol patients with decompensated perception of the body and general health [23]. cirrhosis and no correlation between haemoglobin, haema- Next, we explored surrogate markers of disease activity, tocrit or erythrocytes and HRQL was detectable (data not as well as clinical parameters and their impact on CLDQ shown). More intriguingly, the subscale fatigue was indepen- total value and EQ-5D-5L UI-value. On univariable analy- dently affected by the status of biochemical remission. Yet, ses, AST was inversely correlated with CLDQ total value. the mechanisms underlying fatigue in AIH remain poorly Elevated AST blood levels have been linked to a progres- understood [22], and further investigations are needed to sion of disease and treatment failure [24, 25]. Furthermore, unveil the underlying pathophysiology. Another aspect is the patients exhibiting an incomplete biochemical remission had comparison of mental vs. physical health and a recent study an overall worse CLDQ total value. More notably, using a in AIH reported that mental well-being is more severely multivariable approach, AST and incomplete remission were impaired compared to physical well-being [10]. the only independent predictors of an impaired HRQL in this The EQ-5D-5L was employed to assess general/generic cohort as captured by CLDQ. This could be in parts related HRQL and the corresponding UI-value can be used to com- to disease progression and its liver-related complications. In pare it to other disease states. In the EQ-5D-5L, females fact, a German study previously showed that the severity of reported a significantly lower UI-value, as well as a lower depression and anxiety are associated with concerns about score in the subdomain pain/discomfort compared to males disease progression in AIH, consequently leading to a lower 1 3 2860 Quality of Life Research (2021) 30:2853–2861 HRQL [10]. Furthermore, age did not significantly corre- In conclusion, the current study provides evidence that a lated with CLDQ total value which is in line with findings complete biochemical remission is corresponding to a higher of a Polish study that did not describe an impact of age using HRQL. The current study identified surrogate markers that generic HRQL tools [14]. Interestingly, the EQ-5D-5L was correlated with HRQL and can therefore help physicians more likely to detect non liver related aspects to be associ- identify potential HRQL aspects in their patients. However, ated with a lower HRQL. future studies are warranted to verify these results. Although previous studies have observed a negative impact of corticosteroids on HRQL [12], we did not observe Supplementary Information The online version contains supplemen- a significant difference between patients receiving corticos- tary material available at https://d oi.o rg/1 0.1 007/s 11136-0 21-0 2850-0. teroids vs. no corticosteroids in the current analysis. This is Acknowledgements Dr. Maurice Michel is supported by the Clinician interesting and is most likely related to differences between Scientist Fellowship “Else Kröner Research College: 2018_Kolleg.05”. baseline characteristics and treatment intensity of the two This work is part of the doctoral thesis of Mrs. Francesca Spinelli at cohorts. Importantly, most of the patients in the current anal- the University Medical Centre Mainz. We also thank the patients who ysis were treated with azathioprine and low dose of corticos- participated in the study. teroids or even corticosteroid free, thus providing potential Author contributions MM, FS, and JMS performed research. MM, FS, insight on these different findings. However, patients with a AG, CL, YH, MN, LK, MAW, PRG, JMS designed the experiments and complete biochemical remission were lower on steroids, and analysed the data. MAW, PRG, and JMS contributed reagents/materi- mostly received an azathioprine monotherapy in comparison als/analysis tools. MM and JMS wrote the paper and did statistical to patients with an incomplete remission. Therefore, a lower analysis. All authors approved the final version of the manuscript and the authorship list. JMS is the guarantor of the article. use of steroids in the complete remission group could likely contribute to an overall lower disease burden due to lower Funding Open Access funding enabled and organized by Projekt drug-related side effects. DEAL. Supported by internal research funds of the University Medi- We included a small cohort of 19 patients that were diag- cal Centre Mainz. nosed with an AIH-overlap syndrome. In this subgroup, HRQL was not significantly different compared to non- Declarations overlap patients. These findings differed from observations in the larger analysis in the UK-AIH cohort, when patients Conflict of interest JMS has acted as consultant to Boehringer Ingel-heim, BMS, Genfit, Gilead Sciences, Intercept Pharmaceuticals, Mad- with AIH and PBC or PSC overlap exhibited a lower EQ- rigal, Novartis, Novo Nordisk, Nordic Bioscience, Pfizer, Roche, Sa- 5D-5L UI-value [12]. These differences can be multifactorial nofi, Siemens Healthcare GmbH. Research Funding: Gilead Sciences, and we have previously observed a country-specific impact Boehringer Ingelheim. Speakers Bureau: Falk Foundation MSD Sharp between the UK and Germany when exploring HRQL in & Dohme GmbH. The other authors have no conflicts of interest to declare. NAFLD [23]. The current study has several limitations. Most impor- Open Access This article is licensed under a Creative Commons Attri- tantly, the two HRQL tools were not specifically validated bution 4.0 International License, which permits use, sharing, adapta- for AIH. Therefore, the cause for the differences that are cap- tion, distribution and reproduction in any medium or format, as long tured between the liver-specific tool (CLDQ) and the generic as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes tool (EQ-5D-5L) can only be speculated on. Nevertheless, were made. The images or other third party material in this article are CLDQ has been widely adopted to study HRQL in differ- included in the article’s Creative Commons licence, unless indicated ent aetiologies of chronic liver disease [26]. In a Japanese otherwise in a credit line to the material. If material is not included in study, EQ-5D-5L scores were significantly impaired in AIH the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will patients and were highly associated with corticosteroid use need to obtain permission directly from the copyright holder. To view a [12]. Importantly and as detailed above, the current analysis copy of this licence, visit http://c reati vecom mons.o rg/l icens es/b y/4.0 /. was performed in patients that received combined therapy with low-dose steroids. The generalizability of the current study is limited as it constitutes a single-centre analysis and thus the results have to be interpreted with caution. The References findings are specific for the outpatient setting at the study centre. Also, HRQL can be impacted by country and cul- 1. Feld, J. J., & Heathcote, E. J. (2003). Epidemiology of autoim-mune liver disease. Journal of Gastroenterology and Hepatology, ture-specific differences [21] and thus it is difficult to merge 18, 1118–1128 data from independent study centres. Yet, one strength of 2. Krawitt, E. L. (2006). Autoimmune hepatitis. New England Jour- the analysis is the large cohort of 116 AIH patients and the nal of Medicine, 354, 54–66 standardized use of two different HRQL questionnaires with 3. van Gerven, N. M. F., Verwer, B. J., Witte, B. I., van Erpecum, K. J., van Buuren, H. R., Maijers, I., Visscher, A. P., Verschuren, one focussing on liver disease and one on general HRQL. 1 3 Quality of Life Research (2021) 30:2853–2861 2861 E. C., van Hoek, B., Coenraad, M. J., & Beuers, U. H. (2014). and management of autoimmune hepatitis. Hepatology, 51, Epidemiology and clinical characteristics of autoimmune hepatitis 2193–2213 in the Netherlands. Scandinavian Journal of Gastroenterology, 49, 1 6. Younossi, Z. M., Guyatt, G., Kiwi, M., Boparai, N., & King, D. 1245–1254 (1999). Development of a disease specific questionnaire to meas- 4. Lamers, M. M. H., van Oijen, M. G. H., Pronk, M., & Drenth, J. P. ure health related quality of life in patients with chronic liver (2010). Treatment options for autoimmune hepatitis: A systematic disease. Gut, 45, 295–300 review of randomized controlled trials. Journal of Hepatology, 53, 1 7. Janssen, M. F., Birnie, E., Haagsma, J. A., & Bonsel, G. J. (2008). 191–198 Comparing the standard EQ-5D three-level system with a five- 5. Schramm, C. (2019). Autoimmune hepatitis beyond steroids: level version. Value in Health, 11, 275–284 Effective trial design and attention to quality of life. Clinical Liver 1 8. Ludwig, K., Graf von der Schulenburg, J.-M., & Greiner, W. Disease (Hoboken), 14, 33–36 (2018). German value set for the EQ-5D-5L. Pharmacoeconom- 6. Sebode, M., Hartl, J., Vergani, D., Lohse, A. W., & International ics, 36, 663–674 Autoimmune Hepatitis Group (IAIHG). (2018). Autoimmune 19. Labenz, C., Toenges, G., Schattenberg, J. M., Nagel, M., Sprinzl, hepatitis: From current knowledge and clinical practice to future M. F., Nguyen-Tat, M., Zimmermann, T., Huber, Y., Marquardt, research agenda. Liver International, 38, 15–22 J. U., Galle, P. R., & Wörns, M. A. (2019). Clinical predictors for 7. Hoeroldt, B., McFarlane, E., Dube, A., Basumani, P., Karajeh, M., poor quality of life in patients with covert hepatic encephalopathy. Campbell, M. J., & Gleeson, D. (2011). Long-term outcomes of Journal of Clinical Gastroenterology, 53, e303–e307 patients with autoimmune hepatitis managed at a nontransplant 20. van der Plas, S. M., Hansen, B. E., de Boer, J. B., Stijnen, T., Pass- center. Gastroenterology, 140, 1980–1989 chier, J., Rob, A., & Schalm, S. W. (2007). Generic and disease- 8. EASL Clinical Practice Guidelines. (2015). Autoimmune hepati- specific health related quality of life of liver patients with various tis. Journal of Hepatology, 63, 971–1004 aetiologies: A survey. Quality of Life Research, 16, 375–388 9. Mack, C. L., Adams, D., Assis, D. N., Kerkar, N., Manns, M. P., 2 1. Huber, Y., Boyle, M., Hallsworth, K., Tiniakos, D., Straub, B. K., Mayo, M. J., Vierling, J. M., Alsawas, M., Murad, M. H., & Czaja, Labenz, C., Ruckes, C., Galle, P. R., Romero-Gómez, M., Anstee, A. J. (2019). Diagnosis and management of autoimmune hepatitis Q. M., & Schattenberg, J. M. (2019). Health-related quality of in adults and children: 2019 practice guidance and guidelines from life in nonalcoholic fatty liver disease associates with hepatic the American Association for the study of liver diseases. Hepatol- inflammation. Clinical Gastroenterology and Hepatology, 17, ogy, 72, 671–722 2085–2092.e1 10. Schramm, C., Wahl, I., Weiler-Normann, C., Voigt, K., Wiegard, 2 2. Dyson, J. K., de Martin, E., Dalekos, G. N., Drenth, J. P., Herkel, C., Glaubke, C., Brähler, E., Löwe, B., Lohse, A. W., & Rose, M. J., Hubscher, S. G., Kelly, D., Lenzi, M., Milkiewicz, P., Oo, Y. (2014). Health-related quality of life, depression, and anxiety in H., & Heneghan, M. A. (2019). Review article: Unanswered clini- patients with autoimmune hepatitis. Journal of Hepatology, 60, cal and research questions in autoimmune hepatitis-conclusions of 618–624 the International Autoimmune Hepatitis Group Research Work- 1 1. Sockalingam, S., Blank, D., Abdelhamid, N., Abbey, S. E., & shop. Alimentary Pharmacology & Therapeutics, 49, 528–536 Hirschfield, G. M. (2012). Identifying opportunities to improve 23. Montali, L., Frigerio, A., Riva, P., & Invernizzi, P. (2011). ‘It’s as management of autoimmune hepatitis: Evaluation of drug if PBC didn’t exist’: The illness experience of women affected by adherence and psychosocial factors. Journal of Hepatology, 57, primary biliary cirrhosis. Psychology & Health, 26, 1429–1445 1299–1304 2 4. Sahebjam, F., & Vierling, J. M. (2015). Autoimmune hepatitis. 12. Wong, L. L., Fisher, H. F., Stocken, D. D., Rice, S., Khanna, A., Frontiers in Medicine, 9, 187–219 Heneghan, M. A., Oo, Y. H., Mells, G., Kendrick, S., Dyson, J. 25. Al-Chalabi, T., Underhill, J. A., Portmann, B. C., McFarlane, I. K., & Jones, D. E. (2018). The impact of autoimmune hepatitis G., & Heneghan, M. A. (2008). Effects of serum aspartate ami- and its treatment on health utility. Hepatology, 68, 1487–1497 notransferase levels in patients with autoimmune hepatitis influ- 13. Takahashi, A., Moriya, K., Ohira, H., Arinaga-Hino, T., Zeniya, ence disease course and outcome. Clinical Gastroenterology and M., Torimura, T., Abe, M., Takaki, A., Kang, J. H., Inui, A., & Hepatology, 6, 1389–1395 Fujisawa, T. (2018). Health-related quality of life in patients with 2 6. Orr, J. G., Homer, T., Ternent, L., Newton, J., McNeil, C. J., Hud- autoimmune hepatitis: A questionnaire survey. PLoS One, 13, son, M., & Jones, D. E. (2014). Health related quality of life in e0204772 people with advanced chronic liver disease. Journal of Hepatol- 14. Janik, M. K., Wunsch, E., Raszeja-Wyszomirska, J., Moskwa, M., ogy, 61, 1158–1165 Kruk, B., Krawczyk, M., & Milkiewicz, P. (2019). Autoimmune hepatitis exerts a profound, negative effect on health-related qual- Publisher’s Note Springer Nature remains neutral with regard to ity of life: A prospective, single-centre study. Liver International, jurisdictional claims in published maps and institutional affiliations. 39, 215–221 15. Manns, M. P., Czaja, A. J., Gorham, J. D., Krawitt, E. L., Mieli- Vergani, G., Vergani, D., & Vierling, J. M. (2010). Diagnosis 1 3