Type 1 diabetes mellitus in patients with chronic hepatitis C before and after interferon therapy

Authors


Dr P. Fabris, Department of Infectious Diseases and Tropical Medicine, S. Bortolo Hospital, Viale Rodolfi 37, Vicenza, Italy.
E-mail: pfabris@yahoo.com

Summary

Type 1 diabetes mellitus is the result of an autoimmune process characterized by pancreatic beta cell destruction. It has been reported that chronic hepatitis C infection is associated with type 2 diabetes mellitus, but not with type 1. Although the prevalence of markers of pancreatic autoimmunity in hepatitis C virus-positive patients is not significantly different to that reported in the general population, it increases during alpha-interferon therapy from 3 to 7%, probably due to the immunostimulatory effects of this cytokine. To date, 31 case reports of type 1 diabetes mellitus related to interferon treatment have been published. Type 1 diabetes mellitus occurs more frequently in patients treated for chronic hepatitis C than for other conditions and is irreversible in most cases. In 50% of these patients, markers of pancreatic autoimmunity predated treatment, the majority of cases having a genetic predisposition. Thus, in predisposed individuals, alpha-interferon can either induce or accelerate a diabetogenic process already underway. We suggest that islet cell autoantibodies and glutamic acid decarboxylase autoantibodies should be investigated before and during interferon treatment in order to identify subjects at high risk of developing type 1 diabetes mellitus.

Type 1 diabetes mellitus

Type 1 diabetes mellitus (formerly type I, insulin-dependent diabetes mellitus or juvenile diabetes) is the result of an autoimmune process characterized by progressive beta cell destruction leading to absolute insulin deficiency.1 The onset of the disease is characterized by islet infiltration of T lymphocytes and monocytes and by hyperexpression of major histocompatibility complex class I antigens.2, 3 Type 1 diabetes mellitus, once viewed as an acute condition, results from a longstanding autoimmune process which precedes the overt clinical manifestation. Only when most of the beta cells have been destroyed do the classical clinical manifestations of hyperglycaemia and ketosis occur.1 A genetic predisposition and environmental factors are necessary for the development of type 1 diabetes mellitus.1 A genetic predisposition to type 1 diabetes mellitus maps to the region of chromosome 6 that contains the highly polymorphic class II HLA genes which determine immune responsiveness. Analysis of DNA sequences from diabetics indicates that the alleles of DRB1 and DQB1 determine both disease susceptibility and resistance, and that the structure of the DQ molecule, in particular residue 57 of the beta chain, specifies the autoimmune response against beta cells.4 A genetic predisposition is necessary but not sufficient for the development of type 1 diabetes mellitus. The onset of the disease is preceded and accompanied by circulating autoantibodies to the endocrine pancreas. Those characterized to date include islet cell autoantibodies, insulin autoantibodies, glutamic acid decarboxylase autoantibodies (GADAbs) and tyrosine-like phosphatase autoantibodies (IA2Abs). Islet cell autoantibodies react with an unidentified antigen(s) present within islet cells; the other three autoantibodies target molecularly defined beta cell antigens.5 One or more of these autoantibodies is present in over 90% of patients with newly diagnosed type 1 diabetes mellitus.1 They can also be present in the normal population (schoolchildren, first-degree relatives of patients with type 1 diabetes mellitus, patients with autoimmune diseases), where the risk of overt diabetes is directly correlated with the titre of islet cell autoantibodies and/or with the number of these markers.5, 6 Pancreatic biopsies performed in recent-onset type 1 diabetes mellitus patients have shown a close relationship between the presence and titre of islet autoantibodies and the extent of the histological lesion.7

Hepatitis virus infections and diabetes mellitus

Recent epidemiological studies have shown a relationship between hepatitis C virus (HCV) infection and diabetes mellitus, indicating that diabetes could represent another extra-hepatic manifestation of HCV infection.8–13 However, in these studies, HCV infection appeared to be mainly linked to type 2 diabetes mellitus and not type 1. In a US cross-sectional survey by Mehta et al., the prevalence of diabetes was significantly higher in HCV-positive than HCV-negative subjects, but none of the 19 subjects with type 1 diabetes mellitus had evidence of exposure to HCV.12 Similar findings were obtained by Caronia et al., who documented an association between non-insulin-dependent diabetes mellitus and HCV infection.14 In the same study, the prevalence of diabetes mellitus was closely associated with Child–Pugh scores and with increasing age. Furthermore, Hassan et al. interestingly found a significant synergy between heavy alcohol consumption, HCV infection and diabetes mellitus, suggesting a common pathway for hepatocarcinogenesis.15

Because the pathogenesis of type 2 diabetes mellitus is related to impaired insulin secretion, peripheral insensitivity to insulin and/or hepatic glucose production, the association between HCV and type 2 diabetes mellitus could simply reflect the underlying liver disease.13, 16 In fact, Konrad et al., measuring insulin sensitivity, glucose effectiveness and first- and second-phase insulin in 13 patients with HCV chronic hepatitis, found that the severity of the HCV-related liver disease was associated with a deterioration in insulin sensitivity, resulting in an impaired glucose homeostasis.17 The lack of association between HCV infection and type 1 diabetes mellitus becomes apparent in studies evaluating the prevalence of the main markers of pancreatic autoimmunity in patients with chronic hepatitis C, with and without diabetes mellitus. In the study by Mason et al., GADAbs were detected in two (8%) of the 25 HCV-infected diabetic patients.16 By contrast, none of the 63 diabetic patients with chronic hepatitis C studied by Ando et al. had GADAbs.18 Similar findings were obtained in studies on HCV-positive patients without diabetes mellitus. Hieronimus et al. found only one (2%) GADAb-positive case among 47 nondiabetic HCV-infected patients.19 Similarly, in another study by Piquer et al., the prevalence of GADAbs, IA2Abs and islet cell autoantibodies among 277 nondiabetic HCV subjects was 1, 0 and 0%, respectively, which was comparable to that observed in the control population.20 Nine different studies (Table 1) have assessed the frequency of pancreatic autoimmunity in patients selected for interferon treatment. Of the 440 HCV-positive patients tested, only 12 (3%) were found to be positive for at least one marker of pancreatic autoimmunity.20–28 The message emerging from the studies alluded to above is that patients with chronic hepatitis C do not have a significantly higher prevalence of pancreatic autoantibodies compared with that reported in the normal population.20, 29

Table 1.  Markers of pancreatic autoimmunity in patients with chronic viral hepatitis before and after interferon (IFN) therapy
ReferenceNo. patientsIndication for IFN therapyType of IFN (3 times weekly)Duration of therapyAntibodies before treatmentAntibodies after treatmentEvolution to diabetes
  1. * At baseline the patients had both ICA and GAD antibodies. Their titre progressively increased and the patients developed type 1 diabetes mellitus after 5 months of therapy.28

  2. HBV, hepatitis B virus; HCV, hepatitis C virus; CH, chronic hepatitis; ICA, islet cell autoantibodies; IAA, insulin autoantibodies; GADAb, glutamic acid decarboxylase autoantibodies; GCA, glucagon-producing cells antibodies; IA2Ab, tyrosine-like phosphatase autoantibodies.

2046HCV-CHα-2a6 monthsGADAb positive 0GADAb positive 0None
α-2b IA2Ab positive 0IA2Ab positive 0 
2132HBV-CH17 α-2a 4.5 MU4 monthsICA positive 0ICA positive 0None
15 α-lymph 5 MU/m26 months   
2258HCV-CHα-2a, α-2b, α-leuc6 monthsICA positive 0ICA positive 0None
2348HCV- CH31 α-2b, 29 α-lymph6–12 monthsICA positive 0ICA positive 0None
12HBV-CH  IAA positive 2IAA positive 8 
2440HCV-CHα-2a, α-2b6 monthsGADAb positive 0GADAb positive 1None
2547HCV-CHα-2b 3 MU12 monthsICA positive 2ICA positive 6Type 1 diabetes in one case*
2670HCV-CHα-2b 5–3 MU6 monthsICA positive 1ICA positive 1Type 1 diabetes in one case*
  GCA positive 2GCA positive 2 
  GADAb positive 2GADAb positive 2 
2775HCV-CH55 α-2b8.3 months*GADAb positive 3GADAb positive 8None
7 α-2a IA2Ab positive 0IA2Ab positive 1 
13 α-lymph    
2856HCV-CHαICA positive 0ICA positive 1None
  GADAb positive 0GADAb positive 2 
  IAA positive 0IAA positive 3 
  IA2Ab positive 0IA2Ab positive 1 

Role of alpha-interferon in the onset of type 1 diabetes mellitus

In transgenic mice in which beta cells express alpha-interferon, a syndrome resembling type 1 diabetes has been reported.30 These mice develop a hypo-insulinaemic diabetes associated with a mixed cellular inflammation centred on the islets. The development of the disease could be prevented by the use of a neutralizing antibody to alpha-interferon. However, the effect of alpha-interferon in diabetogenesis in the experimental animal model is still controversial, with some studies indicating that alpha-interferon may improve glucose tolerance and prevent or even cure the disease in non-obese diabetic mice.31, 32

In a human study, alpha-interferon administered for 3 months in patients with chronic hepatitis C did not have measurable effects on insulin sensitivity and glucose tolerance.33 In newly diagnosed type 1 diabetes mellitus in man, alpha-interferon was detectable in the islets by immunocytochemistry.3 Huang et al., using a reverse transcriptase–polymerase chain reaction in the pancreas and islets from patients with and without type 1 diabetes, found that among the cytokines evaluated, including alpha-interferon, gamma-interferon, beta-interferon, tumour necrosis factor-alpha, interleukin-1-beta, interleukin-2, interleukin-4 and interleukin-6, only alpha-interferon was significantly over-expressed in diabetic patients.34 These results are in agreement with the hypothesis that alpha-interferon is involved in the development of type 1 diabetes in man.

Alpha-interferon is widely used for its anti-viral, antiproliferative and immunomodulatory activities. It increases HLA class I antigen expression, natural killer cell and T cell activities. Alpha-interferon has been suggested as an important cofactor in the development of Th-1 immune reaction, which can contribute to the development of autoimmune diseases by the activation of CD4+ lymphocytes secreting interleukin-2, gamma-interferon and tumour necrosis factor-beta.35 These cytokines provide help for the generation of CD8+ cytotoxic T cells. Organ-specific manifestations have been reported during alpha-interferon therapy and the thyroid represents the main target for autoimmunity associated with alpha-interferon therapy. Patients positive for thyroid autoantibodies can develop clinical thyroid dysfunction, but also those initially seronegative for these autoantibodies, but with a genetic predisposition, can seroconvert during or after interferon therapy, and develop clinical or subclinical thyroid autoimmune dysfunction.36–38

In the context of diabetes, our group, in 1992, documented the first case of type 1 diabetes mellitus developing during alpha-interferon therapy.39 This patient was positive for GADAbs and insulin autoantibodies before treatment; at the time of diagnosis their titre was increased and islet cell autoantibodies became detectable. Furthermore, the immunogenetic background of this individual was that typically conferring susceptibility to autoimmune type 1 diabetes mellitus (Table 2). We then suggested that alpha-interferon may induce the development of overt autoimmune diabetes mellitus in genetically and immunologically predisposed individuals.

Table 2.  Main clinical, immunological and genetic features of 31 case reports documenting the development of type 1 diabetes mellitus during or soon after interferon (IFN) treatment (1992–2002)
ReferenceSex/ ageIndication for IFN treatmentDiabetes mellitus
in relatives
Type of IFN
(total dose, MU)
Latency before
type 1 diabetes
AutoantibodiesHLA-DR
DQB1
DQA1
Before IFNAt onset of type 1 diabetes
  1. HCV, hepatitis C virus; HBV, hepatitis B virus; CH, chronic hepatitis; ICA, islet cell autoantibodies; GADAb, glutamic acid decarboxylase autoantibodies; IAA, insulin autoantibodies; IA2Ab, tyrosine-like phosphatase autoantibodies; IL-2, interleukin-2.

29M/59HCV-positive CHNegativeα-2b (80 mg/week) + ribavirin (1200 mg)10 monthsICA positiveDR2, DR4
    GADAb positiveIA2Ab negative 
29M/54HCV-positive CHType 2α-2b (240 MU) + ribavirin (194 g)6 monthsICA positive
    GADAb positiveIA2Ab negative 
39M/61HCV-positive CHType 2α-2b (216 MU)6 monthsICA negativeICA positive4/11
    GADAb positiveGADAb positive 
    IA2Ab negativeIA2Ab negative 
    IAA positiveIAA positive 
50M/48HBV-positive CHNegativeα-2b (120 MU)2 monthsICA positiveDR3/DR4 negative
51F/41Hairy cell leukaemiaNegativeα-2b (120 MU)16 monthsICA negativeICA positiveDR4/DR7/DRW53
    37kDAb negative37kDAb negativeDQ3/DQ2
52M/31HBV-positive CHType 2α-2a (250 MU)48 monthsICA positiveDR4, DRW53
  β (100 U)  GADAb positive 
53M/54Kaposi's sarcoma      
HIV positiveα-2a (65 MU)10 daysICA negativeICA negative
    IAA negativeIAA negative 
54F/56HCV-positive CHType 1α-2a (396 MU)11 monthsICA positiveDR3/W52
55M/66Renal cell carcinomaNegativeα (660 MU)11 monthsICA negative
     IAA positive 
56M/51HCV-positive CHNegativeα-2a (1350 MU)5 monthsICA negativeICA negativeDR4, DR12
    IAA negativeIAA negativeDQ1, DQ4
57M/37HCV-positive CHα (384 MU)4 monthsGADAb positiveGADAb positiveDR9, DR53
      DQ3
58M/53Renal cell carcinomaNegativeα-2a (702)15 months
  IL-2 (464)    
59M/56HCV-positive CHType 2α (216 MU)3 weeksICA positive
    after cessation   
60M/50HCV-positive CHType 1α (672 MU)7 weeksICA negativeICA negative
    IAA negativeIAA negative 
61F/59HCV-positive CH
Negative
α (288 MU)
4 months

GADAb positive
DR4, DRW13, DRW53, DRW52
    ICA positiveICA positiveDQW3, DQW6
    IAA negativeIAA negative 
62F/–HCV-positive CH
renal transplanted
α-2b (110 MU)7 months after cessationICA negativeDR1, DR7
63M/29HCV-positive CHType 1α-2b (300 MU)5 monthsICA positiveICA positive 
    GADAb positiveGADAb positiveDR4/8
    IA2Ab negativeIA2Ab negativeN-Asp/Asp
    IAA negativeIAA negativeArg+/Arg+
64F/53HCV-positive CHα (480 MU)7 months after cessationGADAb positiveDR4, DR8
     DQ1, 4
      (DQB10401 0601)
65M/36HCV-positive CHNegativeα (240 MU)4 monthsICA negativeICA negativeDR4
    GADAb positiveGADAb positive 
    IA2Ab positiveIA2Ab positive 
66M/57HCV-positive CH
Negative
β (183 MU)
2 months
ICA negative
GADAb negative
ICA negative
GADAb positive
DR4 DRB1*0405, DQB104,
DPB0501
    IAA negativeIAA negative 
67M/37HCV-positive CHType 2α-2b (950 MU) + ribavirin (210 g)7.5 monthsGADAb positiveGADAb positive 
    ICA negativeICA positiveDR3
68M/57HCV-positive CHα-2b (530 MU)3.7 monthsICA negativeICA negativeDR4, DR14, DRB1*0405/1401
    GADAb negativeGADAb positiveDQB1*0401/0503 DRPB1*0201/0501
    IAA negativeIAA positive 
69M/29HCV-positive CHα (522 MU)8.5 monthsICA negativeICA positiveDR3
    GADAb negativeGADAb positiveDQA1 *0501, DRB*03011
    IA2Ab negativeIA2Ab positiveDQB1 *0201
70M/53HCV-positive CHNegativeα-2b (450 MU) + ribavirin (236 g)5 monthsICA negative
    GADAb positive 
71M/35HCV-positive CHType 2α-2b (144 MU) + ribavirin (144 g)4 monthsGADAb negativeGADAb positive
72M/41HCV-positive CHNegativeα-2b (108 MU) + ribavirin (90 g)3 monthsGADAb negativeGADAb negativeDRB1 *0101/041
    IA2Ab negativeIA2Ab negative 
    IAA negativeIAA negative 
72M/36HCV-positive CHNegativeα-2b (108 MU) + ribavirin (90 g)3 monthsGADAb positiveGADAb positive
    IA2Ab negativeIA2Ab negative 
    IAA negativeIAA negative 
73F/40HCV positive and HIV positiveNegativeα-2b (240 MU) + ribavirin (180 g)6 monthsICA negative
    GADAb positiveGADAb positive 
    IA2Ab negative 
73M/23HCV-positive CHNegativeα-2b (198 MU)5 monthsICA negative
    GADAb negative 
73M/59HBV-positive CHNegativeα-2b (27 MU)2 weeksICA negative
    GADAb negative 
73M/40HCV-positive CHNegativeα-2b (81 MU) +  ribavirin (60 g)2 monthsICA negativeICA negative
    GADAb positiveGADAb positive 
    IAA negativeIAA negative 

Two large retrospective studies investigated the side-effects of interferon in patients with chronic hepatitis C. In an Italian study by Fattovich et al., de novo diabetes mellitus was documented in 10 of 11 241 (0.08%) treated patients.40 In the second study, performed in Japan, five of 667 interferon-treated patients developed diabetes, with an incidence of 0.7%.41 These incidences are significantly higher than the annual incidence rates reported in the general populations of Italy and Japan.42, 43 Unfortunately, autoantibodies to beta cells were not investigated and therefore the distinction between autoimmune and non-autoimmune diabetes mellitus could not be made in these two studies.

The sequential study of endocrine-specific autoantibodies is needed to monitor the development of type 1 diabetes mellitus, and possibly to gain insight into the pathogenesis of diabetes mellitus, in this group of patients, especially in view of the fact that alpha-interferon, in addition to its immunomodulatory properties, can increase insulin resistance and induce hyperglycaemia.44–49

The natural history of type 1 diabetes mellitus during interferon treatment

To date, the effect of alpha-interferon therapy on pancreatic autoimmunity has been evaluated in nine studies: one in patients with chronic hepatitis B, a second in chronically infected hepatitis B or C patients, and the remaining seven in patients with chronic hepatitis C.20–28 These studies are difficult to compare, as different types of interferon and different total dosages and durations were used, and different markers of pancreatic autoimmunity were evaluated (Table 1). In total, 484 patients (440 with chronic hepatitis C and 44 with chronic hepatitis B) were studied. The prevalence of at least one marker of pancreatic autoimmunity was documented before interferon commencement in 12 (3%) of the 440 patients with chronic hepatitis C and in one (2%) of the 44 patients with chronic hepatitis B; this frequency being similar to that in normal controls. After alpha-interferon treatment, the prevalence of pancreatic autoantibodies increased to 7% (34/440) in patients with chronic hepatitis C and to 5% (2/44) in patients with chronic hepatitis B (Table 1). These data indicate that alpha-interferon therapy is able to stimulate pancreatic autoimmunity, and some of these positive cases went on to develop type 1 diabetes mellitus. To date, 31 cases (25 males, six females, mean age 46.7 years, range 23–66 years) of type 1 diabetes mellitus during or soon after interferon therapy have been reported.39, 50–73 (Table 2).

Twenty-five patients were treated with interferon for chronic hepatitis C, three for chronic hepatitis B, and in three cases interferon was used as an adjuvant therapy for cancer. Nine patients were treated with the combination therapy (alpha-interferon + ribavirin), one patient with alpha-interferon in combination with interleukin-2, one patient with alpha- and beta-interferon, and one patient with beta-interferon alone. The remaining patients were treated with alpha-interferon monotherapy. Family history was positive for diabetes mellitus in nine cases (type 1 in three cases, type 2 in six cases), negative in 16 cases, and data were not available in six cases. The total dose of interferon administered before the onset of diabetes varied from 65 to 1350 MU. The latency of diabetes after interferon therapy commencement ranged from 10 days to 4 years. Before treatment, nine of 18 (50%) patients tested were positive for at least one marker of pancreatic autoimmunity. At the onset of diabetes mellitus, 23 (77%) of the 30 cases evaluated had at least one marker of pancreatic autoimmunity (18 were positive for GADAbs, 11 were positive for islet cell autoantibodies, and three cases were positive for other markers). HLA haplotypes conferring susceptibility to autoimmune diabetes were detected in 16 (89%) of the 18 subjects studied (see Table 2).39, 51, 52, 54, 60, 62–69

The onset of the disease was characterized in most patients by polyuria, polydipsia, weight loss, and in some cases by ketoacidosis and hyperosmolar coma.47, 53, 57, 61, 65–69 Transient insulin dependency was observed in eight cases51, 53, 58–62, 73 and permanent insulin treatment was required in the other reported cases. These data demonstrate that in some cases the autoimmune attack is at least partially reversible with the interruption of interferon therapy. It is important to underline that five patients initially negative for pancreatic autoimmunity seroconverted during therapy. Of particular interest is the case reported by Bosi et al., in which diabetes ran a fulminant course after 8.5 months of alpha-interferon therapy in a genetically predisposed subject.69 In this case, autoantibodies against beta cells were negative at baseline and up to 10 weeks before the onset of the disease: they appeared at a high titre when the disease was clinically manifest. Such a fulminant onset is unusual, even in type 1 diabetes of early childhood where autoantibodies are detectable months before the clinical onset of the disease.74

In summary: (a) a small proportion of patients with chronic HCV hepatitis is positive for one or more marker of pancreatic autoimmunity; (b) these patients are at relative risk of developing diabetes mellitus if treated with alpha-interferon; (c) a small number of patients can develop de novo pancreatic autoimmunity and fall in the group of patients at risk of developing diabetes mellitus; (d) a timely suspension of alpha-interferon therapy is rarely accompanied by regression of clinical diabetes mellitus; (e) no correlation has been documented between the response to anti-viral therapy and the development of diabetes mellitus.

Conclusions

Type 1 diabetes mellitus is an autoimmune disease mediated by the destruction of pancreatic beta cells, and autoantibodies to pancreatic beta cells are markers of the ongoing autoimmune destruction.5 To date, longitudinal studies in normal populations have focused on groups at high risk for type 1 diabetes mellitus, namely the first-degree relatives of type 1 diabetics and patients with other autoimmune diseases in whom the risk increases with the number of markers of pancreatic autoimmunity.75

The present review identifies patients with chronic hepatitis C positive for autoantibodies against beta cells as an additional group of subjects at high risk of diabetes mellitus. Based on the observations discussed above, the suggestion emerges that patients with chronic hepatitis C should be screened for GADAbs and/or islet cell autoantibodies before and during interferon therapy. Seropositive cases should be further tested for IA2Abs and insulin autoantibodies in order to define the risk of future disease. Diabetes mellitus-prone patients should be informed of their immunological status and the potential risk of developing diabetes mellitus during alpha-interferon therapy. Initiation or suspension of the treatment should be evaluated, considering the risk of diabetes and the benefit of the treatment.

Ancillary