Thyroid dysfunction in patients with chronic hepatitis C: Virus- or therapy-related?

Thyroid dysfunction in patients with chronic hepatitis C: Virus- or therapy-related?

Journal of Gastroenterology and Hepatology

Early View (Articles online in advance of print) May 2009

"These data suggest that IFN-based therapy, which is known to induce, reveal or exacerbate various autoimmune disorders, probably causes TD in CHC patients. Additionally, it is noteworthy that TD development in our study population was linked to the onset of other autoimmune disorders (except thyroid) during combination treatment....The actual relevance of the TD to chronic HCV infection per se remains to be fully clarified in large prospective, controlled studies......A clear conclusion concerning the relation of HCV infection per se and TD can not be made, because the control group included only 33 patients."

Elena Vezali, Ioannis Elefsiniotis, Constantinos Mihas, Evangelos Konstantinou and George Saroglou

University Department of Internal Medicine, Hepatology Unit, 'Elena Venizelou' Hospital, Athens, Greece

Correspondence to Dr Elena Vezali, Agias Paraskevis 7A, Lykovrisi, 14123 Athens, Greece. Email: avezalis@otenet.gr

ABSTRACT

Background and Aims: Thyroid dysfunction (TD) represents an extrahepatic manifestation of chronic hepatitis C (CHC). Moreover, the currently approved treatment of CHC is often associated with TD. However, it remains debatable if TD is mainly virus- or treatment-related. The aim of this study was to assess the incidence and features of TD, and to identify its predictors in treated and untreated CHC patients.

Methods: Ninety-four patients with CHC and normal thyroid function were evaluated long-term for TD: 33 were untreated (control group) and 61 were treated with pegylated interferon alpha (PEG-IFN-α) plus ribavirin (treatment group). Mean follow up was 80.1 and 39.4 months, respectively.

Results: All patients in the control group remained euthyroid, while 13 treated patients (21.3%) developed TD (P < 0.001). Eleven of these were diagnosed with hypothyroidism and two with hyperthyroidism, which then converted to hypothyroidism. In the majority of cases (9/13, 69.2%) TD did not reverse after treatment discontinuation and required hormone replacement therapy. Pretreatment virological parameters did not predict TD, according to multiple logistic regression analysis. TD was not associated with total dose of PEG-IFN-α or ribavirin, viral kinetics or with virological outcome, but it was linked to development of other therapy-related autoimmune disorders (odds ratio, 8.29).

Conclusion: Antiviral therapy of CHC possibly induces de novo or exacerbates pre-existing silent TD. TD does not seem to correlate with any pretreatment virological parameter; it is probably not related to dose or treatment duration, nor linked to viral kinetics or virological outcome. The role of chronic hepatitis C per se in TD remains to be determined.

Introduction

Hepatitis C virus (HCV) infection is a major global health problem, being the second most common chronic viral infection in the world with a global prevalence of approximately 3%. HCV is both a hepatotropic as well as a lymphotropic virus and chronic infection is known to be responsible for both hepatic and extrahepatic diseases.1,2 A high prevalence of thyroid autoimmunity and hypothyroidism,3,4 as well as of papillary thyroid carcinoma,5 has been reported in patients with chronic HCV infection.

The current standard treatment for chronic hepatitis C (CHC) is the combination of pegylated interferon alpha (PEG-IFN-α) and ribavirin,6–8 which leads to a sustained virological response (SVR) rates of 54–80%.9,10 Despite their efficacy, both standard IFN-α and PEG-IFN-α have a well-known side-effect profile, including thyroid dysfunction (TD).11,12 IFN-α seems to act through major histocompatibility complex class I antigens to produce antithyroid antibodies and TD,13 the pattern of which includes autoimmune hypothyroidism, hyperthyroidism and thyroiditis. Also, recent data suggest that, in addition to its immunomodulatory mechanism, IFN-α also precipitates thyroiditis by direct thyrotoxic effects.14 The original data on other autoimmune disorders related to antiviral therapy are limited, only observations from other studies looking at different end-points.

It has been shown previously that TD induced by standard IFN-α therapy has no correlation with SVR rate.15 However, the relationship between rapid virological response (RVR), early virological response (EVR), end of treatment response (ETR) and thyroid function remains unclear. Additionally, the long-term reversibility of TD after PEG-IFN-α therapy remains controversial, while the follow-up duration of most studies is limited to 2 years.

The primary aim of this study was to investigate the association between chronic HCV infection treated with PEG-IFN-α and ribavirin combination therapy or not treated, and TD development. Secondary aims were to determine the possible prognostic factors for TD development, the relationship between TD and baseline as well as on-treatment parameters such us pretreatment viral load, HCV genotype, treatment duration and total dose of PEG-IFN-α and ribavirin, RVR, EVR, ETR, SVR and the spontaneous long-term reversibility of TD.

Discussion
The immunostimulatory effects of IFN-α have been well described, with the thyroid being the most common endocrine organ affected.16,17 The role of ribavirin in the development of TD is also under discussion.18 Although many studies have concentrated mostly on combination therapy with standard IFN-α and ribavirin, there are few data on the effects of PEG-IFN-α on the thyroid gland. Moreover, it is generally speculated that HCV infection itself may perpetuate the immune cascade, which leads to the appearance of autoimmune thyroid disorders, especially in genetically predisposed subjects.1–4

In our study, 21.3% of the treated patients who were found to have clinically evident TD during a short-term follow-up period (6–96; mean, 39.4 months), while none of the untreated patients with CHC developed TD despite the longer follow up (12–312; mean, 80.1 months). These data suggest that IFN-based therapy, which is known to induce, reveal or exacerbate various autoimmune disorders, probably causes TD in CHC patients. Additionally, it is noteworthy that TD development in our study population was linked to the onset of other autoimmune disorders (except thyroid) during combination treatment (OR, 8.29; 95% CI, 2.09–32.83; P = 0.003).

A clear conclusion concerning the relation of HCV infection per se and TD can not be made, because the control group included only 33 patients. Besides this, seventeen patients with CHC and previous evident TD were not included in the analysis (four of them were treated with PEG-IFN-α and 13 remained without antiviral treatment). Additionally, we did not study patients without hepatitis C to assess the de novo TD development in these subjects and can provide only data from the population-based studies reporting the prevalence of TD in the general population as 3.1–3.8%.19,20 So, the analysis may have been underpowered. This concern is lessened be the similar results in two previous population-based studies that excluded a specific role of HCV infection in determining the development of thyroid disease.3,21 Similarly, in the studies of Floreani et al.22 and of Marazuela et al.23, untreated chronic HCV infection was not linked to TD development. Another interesting observation, which supports the lack of a pathogenic role of HCV in inducing thyroid autoimmunity, comes from a 17-year follow-up study of 376 Irish women infected in 1977 with HCV genotype 1b by contaminated immunoglobulin.24 In that study, all women with CHC were euthyroid by clinical and laboratory parameters, despite the detection of thyroid-specific autoantibodies in a significant proportion (13.9%).

Another important observation is the relatively higher incidence of TD in our cohort compared to the corresponding rates reported previously from similar studies using standard IFN monotherapy (3–18%),13,25–27 standard IFN and ribavirin (10.7–12.5%),12,15,28 as well as PEG-IFN-α plus ribavirin therapy (12–12.8%).12,29 Therefore, in a recent study by Tran et al.30 the authors concluded that PEG-IFN-α does not aggravate the risk of TD in comparison to that with standard IFN, but the study included only 24 patients previously treated with standard IFN, and the follow-up period was relatively short at 12 months after treatment completion. Differences in study design, follow-up length, and possibly ethnic factors, as well as variable individual predisposition to the disease may explain this discrepancy.

Also, we found that TD development is not related to any virological (HCV genotype, pretreatment viral load) or pharmacological (PEG-IFN-α regimen, treatment duration, and total dose of PEG-IFN-α and ribavirin) parameters, nor the viral kinetics or therapy outcome (RVR, EVR, ETR and SVR). Many other studies have reached similar conclusions.12,15 The most common risk factors that predicted the development of thyroid disease during IFN-α therapy reported to date are female sex and the presence of thyroid autoantibodies (particularly antithyroid peroxidase antibodies) before initiation of therapy.15,26,27,29 In our study, TD was more common in women, but this difference did not reach statistical significance (P = 0.122). Women are more susceptible than men to develop IFN-α-related thyroid disease, having a relative risk three- to sevenfold higher, as reported in some but not all studies.31,32 Such a factor as a relatively small amount of patients in our cohort could play a role in the observed lack of association between female sex and TD development.

In the treatment group, genotype 3 was the most frequent (41%), followed by genotype 1 (36.1%). Nevertheless, there was a trend of greater incidence of TD in patients with genotype 1 infections versus patients with genotype 3, but this association was not statistically significant, probably because of the small number of cases. Only in one recent study by Friedrich-Rust et al., older age, HCV genotypes 1 and 2, and a low baseline viral load were associated with TD.29 In another larger cohort study,26 HCV genotype did not predict TD after controlling the confounding factors by using multivariate analysis. Further larger scale studies are needed to clarify the relation between HCV genotypes and IFN-α-induced thyroid dysfunction.

There are no data, including our study, that support the link of TD development with the total dose or duration of antiviral treatment. On the basis of these findings, clinicians should avoid dose reduction or premature discontinuation of combination treatment in those patients who develop or are predisposed to TD (e.g. euthyroid female CHC patients with detectable thyroid-specific autoantibodies at baseline), thus compromising the therapeutic response to antiviral treatment. It is opportune to perform regular TSH determinations during treatment and, in case of altered values, the decision to continue or interrupt the treatment should be individually considered. Future studies are needed in order to determine the optimal method and frequency of screening for TD during HCV therapy.

Finally, TD was not spontaneously reversible following the withdrawal of combination therapy in the majority of patients, and needed hormone replacement therapy. Two patients who developed hyperthyroidism during treatment converted to overt hypothyroidism during the following 6 months. Data on the evolution of thyroid disease after IFN-α and ribavirin treatment are controversial. Complete recovery of thyroid function within a few months after therapy withdrawal has been reported.30 However, in most studies the reversal of TD was only partial.18,27,33 Late-onset hypothyroidism (38 months after HCV treatment initiation) was observed in one patient of our study population. The occurrence of hypothyroidism several years after IFN-α withdrawal has been described previously,18,27 which suggests progression of the thyroid autoimmune process after treatment discontinuation, especially with the use of the long-acting PEG-IFN molecules.

Limitations of this study are its retrospective and observational nature, small amount of patients included and the absence of data concerning thyroid-specific autoantibodies at baseline. Screening of antithyroid antibodies was not included in the standard pretreatment evaluation of our study because, a few years ago, their detection in the presence of normal thyroid function (normal TSH, FT3 and FT4 levels) would not have been considered a contraindication for antiviral treatment. Currently, there is a large body of evidence to show that TD occurs mainly in euthyroid patients with an underlying predisposition, such as the presence of thyroid-specific autoantibodies.23,27,34–36 Because the study included only 94 patients, our results should be interpreted carefully and confirmed by larger prospective controlled trials.

In conclusion, our long-term observational study demonstrates that TD occurs in approximately one-fifth of CHC patients treated with the currently proposed combination treatment. Permanent hypothyroidism, which frequently needs hormone replacement therapy, represents the most common thyroid disorder in these patients. TD probably cannot be predicted by any pretreatment virological parameter (genotype, viral load), as well as it seems not to be associated with the PEG-IFN-α formulations, the total dose of PEG-IFN-α and ribavirin, or the duration of therapy, nor to be linked to the early viral kinetics or virological outcome in treated patients. The actual relevance of the TD to chronic HCV infection per se remains to be fully clarified in large prospective, controlled studies.

Results

The baseline demographics and clinical and laboratory characteristics of the study population are presented in Table 1. The two groups were comparable according to pretreatment demographics, clinical and laboratory parameters. In the treatment group, 32 patients (52.5%) received PEG-IFN-α-2a plus ribavirin and 29 (47.5%) received PEG-IFN-α-2b plus ribavirin. The most prevalent genotype was 3 (41.0%) in the treatment group and 1 in the control group (45.5%). Overall, 50 of 61 treated patients (82%) completed the treatment course, while in 11 patients (18%) therapy was discontinued, primarily because of PEG-IFN-α-related adverse events. The mean follow-up period for the treated patients was 39.4 (range, 6–96) months, and for the patients in the control group it was 80.1 (range, 12–312) months. Follow-up characteristics of the study population are presented in Table 2.

Thirteen patients (21.3%) from the treatment group developed TD, whereas all patients in the control group remained euthyroid (P < 0.001). Among 13 patients with TD, 11 were diagnosed with hypothyroidism and two with hyperthyroidism, which later converted to hypothyroidism. Mean time of diagnosis of TD was 9.5 (range, 3–38) months after initiation of therapy. Specifically, there were eight cases of on-treatment TD (3–12 months after treatment initiation), two cases of TD which appeared early after treatment discontinuation (both 1 month after treatment completion) and three cases of late-onset TD, which appeared 6, 6.5 and 26 months, respectively, after treatment completion. In nine of 13 subjects (69.2%), TD did not spontaneously reverse after PEG-IFN-α and ribavirin were discontinued, and all these patients needed hormone replacement therapy. The incidence of TD in women (nine of 13, 69.2%) was higher than that observed in men (four of 13, 30.8%), but this difference did not reach statistical significance (P = 0.122). Moreover, 10 treated patients (16.4%) developed at least one more autoimmune disease during the treatment course (four psoriasis, two mixed cryoglobulinemia, two rheumatoid arthritis-like syndrome, one lichen planus and one idiopathic thrombocytopenia).

Multivariate logistic regression analysis evaluated the potential relationship of 13 independent variables (sex, age, BMI, HCV genotype, pretreatment viral load, treatment regimen and duration, total dose of PEG-IFN-α and ribavirin, RVR, EVR, ETR and SVR) with the development of TD during treatment. As shown in Table 3, none of these predicted TD. The only correlation revealed by our study was that of TD development and the onset of other autoimmune disorders during combination therapy (P = 0.003; OR, 8.29; 95% CI, 2.09–32.83). After cessation of combination therapy, thyroid function returned to normal values spontaneously in four of 13 patients (30.8%). In the remaining nine patients (69.2%), TD did not reverse, and all of them needed hormone replacement therapy.