Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Presented in part at the annual meeting of the American Association for the Study of Liver Diseases, November 12, 2005, San Francisco, California.
Autoimmune hepatitis is classically a disease of young women. Our aims were to determine its occurrence, clinical phenotype, and outcome in elderly patients and contrast findings to young adults. Two-hundred-and-five white North American adults with definite type 1 autoimmune hepatitis were grouped according to age at presentation and the groups compared. Forty-seven patients (23%) were aged ≥60 years (median age, 68 years), and 31 patients (15%) were aged ≤30 years (median age, 25 years). The patients ≥60 years had a higher frequency of cirrhosis at presentation than the patients ≤30 years (33% versus 10%, P = .03). They also had thyroid or rheumatic diseases more commonly (42% vs. 13%, P = .006). HLA DR3 occurred more frequently in the patients ≤30 years than in those ≥60 years (58% vs. 23%, P = .004), and HLA DR4 occurred more often in the patients ≥60 years (47% vs. 13%, P = .003). Patients aged ≥60 years failed corticosteroid treatment less commonly than those aged ≤30 years (5% vs. 24%, P = .03). Autoimmune hepatitis occurred in patients aged 18-30 years (15%), 31-39 years (15%), 40-49 years (21%), 50-59 years (25%), and ≥60 years (23%). Differences in age distribution, HLA frequencies, and treatment outcome occurred after age ≥40 years. In conclusion, elderly patients have a greater frequency of cirrhosis at presentation and HLA DR4 than patients ≤30 years, and they have a lower occurrence of treatment failure. Transitions in clinical and genetic phenotypes occur after age ≥40 years. Genetic susceptibilities may favor etiologic factors that are age-related. (HEPATOLOGY 2006;43:532–538.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Autoimmune hepatitis has been classically described as a disease of young women,1–6 and early experiences suggested a bimodal age distribution.7, 8 Initial perceptions that peak occurrences were between ages 10 years and 30 years and between 40 years and 50 years were probably affected by referral patterns to tertiary medical centers.9, 10 Current experiences indicate that the disease occurs as commonly across all age ranges and that it may be underdiagnosed in the elderly.11–13
The median time to diagnosis is more than twice as long in patients who are 65 years old or older,13 and the frequency of cirrhosis at presentation exceeds 30%.11, 12 These observations have suggested that the disease may be indolent but progressive in the elderly and that this indolence may delay its recognition and treatment.12 Furthermore, the consequences of therapy in the elderly population remain uncertain. Experiences that indicated a lower frequency of remission and more complicated course in the aged14 have been counterbalanced by those describing similar frequencies of remission between the young and old.11–13, 15
Genetic factors influence the occurrence, clinical expression, and treatment outcome of autoimmune hepatitis. In white North American and northern European patients, HLA DR3 and HLA DR4 are independent risk factors for the disease.16 These genetic markers also influence clinical expression and treatment outcome. HLA DR3 has been associated with an earlier age of disease onset than HLA DR4 and a poorer response to corticosteroid therapy.17, 18 In Japan where HLA DR3 is uncommon, patients with autoimmune hepatitis typically have HLA DR4, and they tend to be elderly with mild disease and excellent treatment outcomes.19, 20 The HLA phenotypes of different age groups with autoimmune hepatitis in North America have not been fully defined. HLA distinctions between the young and old in this geographic region may not only affect clinical features and outcome but generate hypotheses about etiologic mechanisms.21, 22
In this retrospective analysis of prospectively acquired data, we determined the frequency of definite type 1 autoimmune hepatitis in white North American adults within different age ranges, and we assessed age-related distinctions in clinical features, treatment outcome, and HLA phenotype for DR3 and DR4. In this fashion, we evaluated age as a factor in disease expression and outcome in adult patients and as a surrogate for a genetic propensity that might affect disease behavior or reflect its cause.
HLA, human leukocyte antigen; ANA, antinuclear antibodies; SMA, smooth muscle antibodies; AMA, antimitochondrial antibodies; LKM1, liver/kidney microsome type 1; AST, serum aspartate aminotransferase; MHC, major histocompatibility complex; NF-κB, nuclear factor-kappa B.
Materials and Methods
Two-hundred-and-five adult white North American patients who had been fully characterized by one investigator (A.J.C.) in accordance with a previously published clinical protocol constituted the study population.23 All patients had been referred to the Autoimmune Hepatitis Treatment Program of the Mayo Clinic, and they had been evaluated between 1975 and 2005. Study patients were selected from 302 patients who had been seen during this same interval because they satisfied the following additional criteria. Patients had to be ≥18 years old at presentation to be considered adult and included in the study population. They had to satisfy the pretreatment diagnostic criteria of the International Autoimmune Hepatitis Group for definite autoimmune hepatitis,24 and the definite diagnosis had to be supported by the international scoring system for autoimmune hepatitis.24 All patients had to be Caucasoid to evaluate the genetic predispositions associated with the disease, and each required HLA typing by DNA-based techniques. The mean diagnostic score prior to therapy for this study population was 18.5 points (range, 16-22 points; median score, 19 points).24
One-hundred-seventy-five patients were women (85%), and the mean age was 48±1 years (range, 18-83 years; median, 49 years). All patients had been assessed for antinuclear antibodies (ANA), smooth muscle antibodies (SMA), and antimitochondrial antibodies (AMA) by indirect immunofluorescence or enzyme-linked immunosorbent assays based on recombinant antigens, as previously described, and each patient was seronegative for AMA.25 Antibodies to liver/kidney microsome type 1 (anti-LKM1) had been sought in 198 patients (97%), and they were absent in each instance. The 7 patients who were not tested for anti-LKM1 had SMA and ANA (3 patients), SMA (2 patients), and ANA (2 patients). All patients satisfied serological criteria for type 1 autoimmune hepatitis.26
Liver tissue examinations were performed at presentation in 203 of the 205 patients (99%), and the histological findings of autoimmune hepatitis were documented in each, including 53 patients (26%) who satisfied histological criteria for cirrhosis.27 Concurrent immune diseases were sought systematically in each patient according to protocol,23 and HLA DR3 and DR4 were determined in each individual by restriction fragment length polymorphism (84 patients) or polymerase chain reaction with sequence specific primers (121 patients) according to methods reported previously.17, 18 Our study had been approved by the Institutional Review Board of the Mayo Clinic.
Patients were categorized according to age at presentation. Forty-seven patients (23%) were aged at least 60 years, and they constituted the elderly population. Their mean age was 69±1 years (range, 60-83 years; median age, 68 years), and 43 patients were women (91%). Thirty-one patients were aged 30 years or less (15%), and they constituted the young population. Their mean age was 25±1 years (range, 18-30 years; median age, 25 years), and 25 patients were women (81%). Individuals not classifiable as elderly or young at presentation were grouped as patients aged 31-39 years, 40-49 years, and 50-59 years. Each age group was used to assess the distribution of the disease across age ranges in adults and age-related differences in clinical and HLA phenotypes.
One hundred seventy-nine patients (87%) were treated with conventional corticosteroid regimens, and the treatment outcomes between the patients aged ≥60 years and those aged ≤30 years were compared.23 One hundred fifteen patients received prednisone in combination with azathioprine (64%), and 64 patients received a higher dose of prednisone alone (36%). Each schedule had been shown previously to be equally effective in treating autoimmune hepatitis.28 Twenty-six patients (13%) either did not satisfy criteria for treatment (17 patients) or received investigational therapies (9 patients). These 26 patients were not included in the analyses of outcome.
Treatment was continued until the pre-defined end points of remission or treatment failure had been achieved.29Remission connoted absence of symptoms, improvement of serum aspartate aminotransferase (AST) levels to normal or near normal (less than twice the upper limits of normal), and histological improvement to minimal or no inflammatory activity. Medication was then withdrawn in a gradual fashion over a 6 week period. Treatment failure connoted worsening of clinical, laboratory and/or histological features despite compliance with therapy. Doses of medication were then increased in accordance with a previously published protocol.29 Reappearance of symptoms and increase in the serum AST level to more than threefold normal after drug withdrawal indicated relapse, and these findings justified re-treatment. Absence of symptoms and serum AST levels that were normal or below the relapse threshold after drug withdrawal connoted a sustained remission. Clinical, laboratory or histological findings during therapy that indicated residual disease activity which had not worsened or improved sufficiently to satisfy end point criteria warranted continued treatment.
The Fisher exact test was used to compare dichotomous variables, and the unpaired t test was used to compare differences in the means of continuous variables. Nonparametric variables in independent samples were compared by the Mann-Whitney test. Because the variables for comparison had been formulated a priori and then assessed systematically in each study group, an unadjusted P value of .05 was used to determine statistical significance. Data are presented as the mean±standard error of the mean in tables and text.
Clinical and HLA Distinctions.
Patients aged ≥60 years had a higher frequency of cirrhosis at presentation than patients aged ≤30 years (33% vs. 10%, P = .03) (Table 1). They also had concurrent thyroid disorders or rheumatic conditions more commonly than patients aged ≤30 years (42% vs. 13%, P = .006) (Table 2). In contrast, the patients aged ≤30 years more commonly had chronic ulcerative colitis or autoimmune hemolysis than the patients aged ≥60 years (13% vs. 0%, P = .02). Both groups were otherwise similar in gender distribution, duration of symptoms prior to presentation, and laboratory indices at presentation (Table 1).
Table 1. Clinical and HLA Findings in the Young and Elderly Age Groups at Presentation
Table 2. Concurrent Immune Diseases According to Age Group
Concurrent Immune Diseases
Patients ≥60 Years (N = 47)
Patients ≤30 Years (N =31)
NOTE. Numbers in parentheses are percentages.
Systemic lupus erythematosus
Type 1 diabetes mellitus
Total concurrent diseases
HLA DR4 occurred more frequently in the patients aged ≥60 years than in those aged ≤30 years (47% vs. 13%, P = .003), and HLA DR3 occurred more commonly in the patients aged ≤30 years than in those aged ≥60 years (58% vs. 23%, P = .004) (Table 1).
Patients aged ≥60 years failed corticosteroid treatment less commonly than those aged ≤30 years (5% vs. 24%, P = .03) (Table 3). Remission, relapse, sustained remission after drug withdrawal, requirement for continued treatment, death from hepatic failure or need for liver transplantation (5% vs. 21%, P = .06) occurred as frequently in both groups. The duration of follow-up was significantly longer in the younger patients (Table 3), but follow-up for at least 5 years was achieved with similar frequency in each age group (44% vs. 59%, P = .3).
Table 3. Treatment Outcomes in the Young and Elderly Age Groups
Treated Patients ≥60 Years (N = 41)
Treated Patients ≤30 Years (N = 29)
NOTE. Numbers in parentheses are percentages.
Significantly different from each other at level of
Age-related Frequencies of Treatment Outcomes and HLA Phenotypes.
Autoimmune hepatitis occurred in patients aged 18-30 years (15%), 31-39 years (15%), 40-49 years (21%), 50-59 years (25%), and ≥60 years (23%). The distribution of disease in each age range tended to increase after 40 years (Fig. 1). Patients aged 18-30 years failed corticosteroid therapy more frequently than those aged 50-59 years (24% vs. 7%, P = .04) and those aged ≥60 years (24% vs. 5%, P = .03) (Table 4). These patients also had a significantly higher occurrence of HLA DR3 than those aged >60 years (59% vs. 24%, P = .006) as did the patients aged 31-39 years (54% vs. 24%, P = .02) (Table 4).
Table 4. Treatment Outcomes and HLA Phenotypes in Each Treated Age Group
In contrast, the patients aged 18-30 years had a significantly lower frequency of HLA DR4 than did the patients aged 50-59 years (14% vs. 45%, P = .01) and those aged ≥60 years (14% vs. 44%, P = .009). Similarly, the patients aged 31-39 years had lower occurrences of HLA DR4 than did the patients aged 50-59 years (11% vs. 45%, P = .004) and those aged ≥60 years (11% vs. 44%, P = .004).
Differences in treatment outcome and frequencies of HLA DR3 and DR4 became evident at the age of 40 years and above (Fig. 2). Patients aged less than 40 years had greater frequencies of treatment failure (23% vs. 6%, P = .003) and death from hepatic failure or requirement for liver transplantation (19% vs. 6%, P = .02) than patients aged ≥40 years (Table 4). The treated patients aged <40 years also had higher occurrences of HLA DR3 (56% vs. 36%, P = .01) and lower occurrences of HLA DR4 (12% vs. 40%, P = .000002) than similarly treated patients aged ≥40 years.
Eleven of the 179 patients who underwent corticosteroid therapy (6%) died of non-hepatic causes, including cardiac disease (9 patients), ovarian cancer (1 patient), and renal failure (1 patient) during 104±8 months of observation (median follow-up, 77 months). These patients were older than 40 years (mean age, 64±4 years; range, 42-82 years; median, 68 years), and 7 patients were older than 60 years (Table 4).
Our study indicates that elderly patients with type 1 autoimmune hepatitis have a different clinical and HLA phenotype than young adult patients and that their disease is less refractory to corticosteroid treatment. The patients aged ≥60 years had a higher frequency of cirrhosis at presentation than those young adults aged ≤30 years, and they had concurrent thyroid or rheumatic diseases more often. Furthermore, they more commonly had HLA DR4 than the young adult patients, and they had a lower frequency of treatment failure. In contrast, the young adult patients had HLA DR3 more commonly than patients aged ≥60 years, and they deteriorated more often during corticosteroid treatment.
These findings extend earlier observations relating clinical features and treatment results to HLA status17–18 by demonstrating that the differences in clinical phenotype and outcome emerge gradually with age progression and that these transitions occur mainly at age 40 years and higher. The age-related differences in the nature of the immune diseases associated with autoimmune hepatitis also support the hypothesis that a transition occurs during adult life in immune response and targets. These insights provide a time perspective by which to understand HLA influences on the clinical manifestations and behavior of autoimmune hepatitis. They also afford opportunities to explore age-related differences in immune reactivity and etiologic risk factors and to develop customized management strategies.
Our findings suggest that elderly patients may progress to cirrhosis more rapidly than young adults with type 1 autoimmune hepatitis or that their liver disease is more indolent and therefore more advanced at the time of detection. The latter possibility is favored because the elderly patients did not have laboratory evidence of more severe inflammatory activity at presentation than the young adults and they had a low frequency of death from liver failure or requirement for liver transplantation. Rapid progression to cirrhosis is usually associated with laboratory indices that reflect severe disease activity,30, 31 and it is frequently associated with a poor outcome.7 Other experiences that have indicated delay in the diagnosis and treatment of the elderly with autoimmune hepatitis have also recognized the high frequency of cirrhosis in these patients.11, 12 These observations support the hypothesis that the elderly have an indolent aggressive liver disease that frequently progresses to cirrhosis prior to diagnosis.12
Autoimmune hepatitis may have an acute or abrupt onset which reflects an exacerbated pre-existent chronic disease or a de novo inflammatory process.32 Earlier studies have indicated that 47%-71% of elderly patients present as an acute icteric hepatitis.11, 13 This presentation, however, has not been uniformly distinctive since it has not been recognized in some series.12 Our retrospective study was unable to evaluate this clinical feature directly, but our elderly patients tended to present more quickly after the onset of symptoms than patients aged ≤30 years (Table 1).
Presentations within 1 month (8% vs. 0%, P = .1) and within 6 months of symptomatic onset (34% vs. 13%, P = .06) may be a reflection of the abruptness of disease onset, and they tended to occur more commonly in the elderly than the young patients. These differences, however, were not statistically significant, and we cannot conclude that the elderly patients presented in a different fashion than the patients aged ≤30 years. Furthermore, the abruptness of the presentation does not distinguish between a preexistent indolent disease that has exacerbated and a de novo occurrence of autoimmune hepatitis, and each possibility must be considered in the elderly.
Our elderly patients were distinguished from the young adults with type 1 autoimmune hepatitis by having a significantly greater frequency of HLA DR4 and lower occurrence of HLA DR3. They also had concurrent thyroid or rheumatic diseases more commonly and a lower frequency of treatment failure. The mechanisms by which the HLA phenotype influences disease occurrence and behavior are uncertain, but analyses of amino acid sequence variations encoded by the susceptibility alleles suggest that disease risk and outcome relate to the configuration of the antigen binding groove of the class II molecule of the major histocompatibility complex (MHC).33 This molecule cradles the triggering antigen, and the complex activates the CD4+ T helper cells that propagate the immune response. Multiple antigens may have structural or conformational similarities that favor their presentation in these class II MHC molecules, and numerous homologous peptides may trigger the same disease.34
HLA DR4 is associated with 38 alleles, of which 10 are common in the normal white North American and northern European populations.35 In contrast, HLA DR3 is associated with 18 alleles, of which only one is common in these same normal ethnic populations. Individuals with HLA DR4 can present a more diverse range of antigens than individuals with HLA DR3, and this diversity of antigen presentation may in turn contribute to the greater frequency of concurrent immune diseases found in the DR4-positive patients.36 Alternatively, the restricted range of peptides that can trigger the disease in patients with HLA DR3 may implicate an etiologic factor that is less frequently encountered in the elderly and more likely to generate a vigorous immune response in the young. This transition in susceptibility occurred at the age of 40 years and above. By defining the susceptibility alleles for type 1 autoimmune hepatitis within different subgroups of the same population and between populations with different ethnic and geographic characteristics, indigenous etiologic agents that naturally select genetically susceptible individuals to develop the disease may be discovered.21, 22, 33
Our elderly patients entered remission, relapsed after drug withdrawal, and sustained a long term remission without medication as commonly as the young adults, and these findings are in accord with other experiences that have emphasized the responsiveness of the elderly to conventional treatments.11–13, 15 Histological cirrhosis in the absence of clinical decompensation is not associated with an impaired treatment response,37, 38 and our findings re-emphasize these earlier observations.
We now extend these observations by indicating that the elderly patients have a lower frequency of treatment failure than the young adults with the disease (Table 3). The low occurrence of refractory disease in this age group may in turn account for the low frequency of death from hepatic failure or requirement for liver transplantation observed in this population. The duration of follow-up was significantly longer in the patients aged ≤30 years, possibly because natural life-spans limited the opportunity for comparable intervals of observation in the elderly (Table 3). Follow-up for at least 5 years, however, had been maintained with similar frequency in the elderly and young adult groups (44% vs. 59%, P = .3), and differences in the frequency of treatment failure, which typically occurs within 6 months after institution of therapy,39 could not be ascribed to the observation intervals.
Treatment failure is uncommon in patients with HLA DR4,17, 18 and the antigenic triggers of the immune response and the cytokine milieu propagated by immunocytes activated by the DR4-antigen complex may favor corticosteroid responsiveness. The pathogenic mechanisms in patients with HLA DR4 may be more dependent on cytokines modified by corticosteroid therapy than patients with HLA DR3.40, 41 Alternatively, the responsiveness of the elderly to corticosteroid therapy may reflect in part age-related changes in immune reactivity. Aging alters immune responsiveness by decreasing the expression of HLA class II molecules42 and reducing stimulation and proliferation of antigen-stimulated T cells.43–45 These effects might make the pathogenic process less severe and more responsive to corticosteroid therapy. The age-related immune transitions may affect the treatment outcome of type 1 autoimmune hepatitis since the disease is mainly cell mediated.46 Future studies must assess age-related differences in cell-mediated responses and etiologic triggers. Patients aged ≥40 years may have exposures to antigens that are different from the young and optimally presented by the DR4 molecule of the MHC.
We cannot exclude referral bias as a factor in our analysis, and our findings may not accurately reflect those in the general population. Referral bias may affect estimations of the duration of illness before presentation, and it may select patients with severe disease. These selection factors, however, should be evident in both the young and elderly adult patients in our referral practice, and we would not expect the age-related differences in the frequencies of clinical phenotype, histological cirrhosis, treatment outcome, and HLA status to be largely an artifact of this bias. Furthermore, the consistent responsiveness of our young adult and elderly patients to conventional corticosteroid therapy indicates that a selection bias for severe disease does not impact adversely on our treatment results.
In summary, autoimmune hepatitis affects adults of all ages, and it is common in individuals aged ≥60 years. Elderly patients have a higher frequency of cirrhosis at presentation than patients aged ≤30 years, but they enter remission as commonly during corticosteroid treatment. They fail treatment less often, and the elderly may have a lower occurrence of death from hepatic failure or need for liver transplantation. Patients aged ≥60 years have HLA DR4 more frequently and HLA DR3 less often than patients aged ≤30 years, and this genetic distinction may affect the frequencies of concurrent thyroid or rheumatic diseases and treatment failure. Distinctions in clinical and HLA phenotypes occur at the age of 40 years and higher, and there may be age-related transitions in etiologic triggers and immune responsiveness. Autoimmune hepatitis should be considered in all elderly patients with chronic liver disease, and the diagnosis warrants treatment with conventional corticosteroid regimens.