Hormone replacement therapy and risk of hepatocellular carcinoma

Hepatocellular carcinoma (HCC) remains one of the leading causes of cancer burden globally; therefore, the prevention of HCC is a critical issue in public health.¹ In general, the most effective way to prevent HCC development is based on the major etiology in the carinogenesis of HCC; for example, antiviral treatment for chronic hepatitis B virus (HBV) or hepatitis C virus (HCV) infection.^{2, 3} However, although antiviral treatment can significantly reduce HCC risk amongst patients with chronic viral hepatitis, the risk is not completely eliminated. Therefore, other methods which can further lower HCC risk remain highly expected. For example, due to its anti-inflammatory properties, aspirin has been previously investigated for its possible chemopreventive effect in HBV- or HCV-related HCC,^{4, 5} even though antiviral treatment remains the first consideration for HCC prevention. In addition, some etiologies of HCC remain lacking effective therapies; for example, non-alcoholic fatty liver disease.⁶ In general, discovering another new way to prevent HCC development should be encouraged.

The epidemiological investigations of HCC disclose that the rates of both incidence and mortality are two to three times higher among men than among women in most regions.¹ With an obvious gender disparity, sex hormone may play an important role in the pathogenesis of HCC, therefore whether hormone replacement therapy (HRT) can reduce HCC risk in females has been highly discussed.^{7, 8} In cell and animal experiments, some genetic, biochemical, or immunological mechanisms have been explored to explain the possible HCC chemoprevention effect of HRT.⁹ For example, in HBV-related HCC, the HBV X protein and estrogen receptor-alpha complex could downregulate the mechanisms of HCC initiation or progression.¹⁰ In HCV-related HCC, estrogen was found to inhibit mature HCV production through estrogen receptor-alpha,¹¹ and the risk of HCC development may thus be reduced. However, even though experimental studies support the preventive effect of HRT, the clinical study results are not always consistent in the prevention of HCC.^{7, 12} Importantly, current clinical evidence mainly comes from observational studies, and the study findings cannot be directly deferred to a causal relationship.

In this issue of Advances in Digestive Medicine, Chang et al¹³ aimed to investigate the chemoprevention effect of HRT on HCC risk and overall survival in women with chronic hepatitis C, and some interesting findings were disclosed after a long period of follow-up. In this retrospective population-based cohort study using data from Taiwan's National Health Insurance Research Database, 1022 patients who received HRT and 1022 matched controls were recruited, and HRT was independently associated with a reduced risk of HCC (adjusted hazard ratio 0.49). Moreover, the patient mortality rate was lower in the HRT treated group, when compared to that of the control group. However, several concerns in the study design should be considered for adjustments in the future studies. First, several important risk factors of HCC development were not included for analysis. For example, the degree of liver fibrosis, HCV virological data, and the time interval between antiviral therapy and the study index date. Second, the indication bias coming from HRT should be carefully avoided. The blood estrogen levels in matched controls who did not receive HRT, particularly in young patients, could be within the normal ranges. In other words, the blood hormone levels in patients receiving HRT could be even lower than those in the young controls. Therefore, the study design to examine the hypothesis that HRT for hormone-lacking patients might reduce HCC risk, could not be precise. Third, HCV eradication therapy has come very easy to achieve a sustained virological response in the era of direct-acting antiviral agents nowadays; therefore, the study data for most patients (>70%) who did not receive any HCV antiviral therapy, may need to be revised. In the future, the role of HRT in the prevention of HCC should be tested amongst patients who have received HCV eradication therapy. Fourth, the patient proportions of underlying comorbidities, such as coronary arterial disease, were significantly different between the two study groups. Therefore, the data regarding patient survival might not be comparable between the two study groups, and should not be only simply examined by a log-rank test. A sufficient adjustment for the baseline patient characteristics can improve the survival analysis.

Last but not the least, the safety concern should be fully discussed before applying HRT in clinical practice. With the concerns of adverse effects from HRT, current guidelines only recommend HRT for symptomatic postmenopausal women.¹⁴ About three decades ago, preventive HRT was advocated for the prevention of osteoporotic fractures and coronary arterial diseases according to the results from observational studies. However, in the following years, serious harm from preventive HRT was found in high-quality randomized clinical trials, with an increased risk for severe diseases, such as breast cancer, cardiovascular diseases, and dementia.¹⁵ In the study by Chang et al, the cumulative incidence of HCC was quite high amongst HCV-infected patients, particularly in the control group. Whether the reduced HCC development rate resulted in the higher overall survival rate in the HRT-treated group should be further clarified. Although a randomized clinical trial may resolve this dilemma, it will be difficult to be conducted due to the concern of HRT-related harm. However, for patients in a special population, in which the risk of HCC-related mortality is higher than that of HRT-related mortality, a well-designed HRT trial may remain considered. Anyway, according to the current evidence, although HRT may be related to a decreased risk of HCC, other preventive ways without potential serious adverse effects should be first considered for patients without symptomatic postmenopausal disorders.

The author declares no conflict of interest.